kernel-release 5.7.11-2 (aarch64) 2020-12059
9999

Submitter benbullard79 [@T] cox.net
Platform rolling
Repository main
URL https://abf.openmandriva.org/build_lists/811188
Packages
bpftool-5.7.11-2.aarch64.binary
cross-aarch64-openmandriva-linux-android-kernel-release-headers-5.7.11-2.aarch64.binary
cross-aarch64-openmandriva-linux-musl-kernel-release-headers-5.7.11-2.aarch64.binary
cross-aarch64-openmandriva-linux-uclibc-kernel-release-headers-5.7.11-2.aarch64.binary
cross-armv7hnl-openmandriva-linux-gnueabihf-kernel-release-headers-5.7.11-2.aarch64.binary
cross-armv7hnl-openmandriva-linux-musleabihf-kernel-release-headers-5.7.11-2.aarch64.binary
cross-armv7hnl-openmandriva-linux-uclibceabihf-kernel-release-headers-5.7.11-2.aarch64.binary
cross-armv7l-openmandriva-linux-androideabi-kernel-release-headers-5.7.11-2.aarch64.binary
cross-armv8l-openmandriva-linux-androideabi-kernel-release-headers-5.7.11-2.aarch64.binary
cross-i686-openmandriva-linux-gnu-kernel-release-headers-5.7.11-2.aarch64.binary
cross-i686-openmandriva-linux-musl-kernel-release-headers-5.7.11-2.aarch64.binary
cross-i686-openmandriva-linux-uclibc-kernel-release-headers-5.7.11-2.aarch64.binary
cross-riscv32-openmandriva-linux-gnu-kernel-release-headers-5.7.11-2.aarch64.binary
cross-riscv32-openmandriva-linux-musl-kernel-release-headers-5.7.11-2.aarch64.binary
cross-riscv32-openmandriva-linux-uclibc-kernel-release-headers-5.7.11-2.aarch64.binary
cross-riscv64-openmandriva-linux-gnu-kernel-release-headers-5.7.11-2.aarch64.binary
cross-riscv64-openmandriva-linux-musl-kernel-release-headers-5.7.11-2.aarch64.binary
cross-riscv64-openmandriva-linux-uclibc-kernel-release-headers-5.7.11-2.aarch64.binary
cross-x86_64-openmandriva-linux-android-kernel-release-headers-5.7.11-2.aarch64.binary
cross-x86_64-openmandriva-linux-gnu-kernel-release-headers-5.7.11-2.aarch64.binary
cross-x86_64-openmandriva-linux-gnux32-kernel-release-headers-5.7.11-2.aarch64.binary
cross-x86_64-openmandriva-linux-musl-kernel-release-headers-5.7.11-2.aarch64.binary
cross-x86_64-openmandriva-linux-muslx32-kernel-release-headers-5.7.11-2.aarch64.binary
cross-x86_64-openmandriva-linux-uclibc-kernel-release-headers-5.7.11-2.aarch64.binary
cross-x86_64-openmandriva-linux-uclibcx32-kernel-release-headers-5.7.11-2.aarch64.binary
kernel-release-5.7.11-2.aarch64.source
kernel-release-headers-1:5.7.11-2.aarch64.binary
kernel-release-server-5.7.11-2.aarch64.binary
kernel-release-server-devel-5.7.11-2.aarch64.binary
kernel-release-source-5.7.11-2.aarch64.binary
lib64bpf0-5.7.11-2.aarch64.binary
lib64bpf-devel-5.7.11-2.aarch64.binary
Build Date 2020-07-31 02:21:44 +0000 UTC
Last Updated 2020-08-08 22:49:05.310177737 +0000 UTC
$ git diff --patch-with-stat --summary 899acbd766b157ef2446990460790c768e705dbe..8b71359564ed514682b6057824394ec385a64627

 .abf.yml                                           |     6 +-
 0001-Fix-for-compilation-with-clang.patch          |    42 -
 0001-cpu-5.2-merge-graysky-s-patchset.patch        |     6 +-
 ...ld-LLVMLinux-Set-compiler-flags-for-clang.patch |    59 -
 ...inux-Remove-warning-from-COMPATIBLE_IOCTL.patch |    37 -
 0002-lib-Add-zstd-modules.patch                    | 13301 -------------------
 0003-btrfs-Add-zstd-support.patch                  |   740 --
 ...Linux-Add-support-for-generating-LLVM-bit.patch |    94 -
 ...Linux-Make-asm-offset-generation-work-wit.patch |   121 -
 0004-squashfs-Add-zstd-support.patch               |   306 -
 0005-crypto-Add-zstd-support.patch                 |   393 -
 ...VMLinux-Remove-nested-function-from-bcach.patch |    65 -
 0006-apparmor-LLVMLinux-Remove-VLAIS.patch         |   108 -
 ...inux-Remove-VLAIS-from-exofs-FIXME-Check-.patch |    91 -
 ...Linux-Use-Oz-instead-of-Os-when-using-cla.patch |    35 -
 ...x86-boot-LLVMLinux-Work-around-clang-PR39.patch |    71 -
 ...REAM-xen-LLVMLinux-Remove-VLAIS-from-xen-.patch |    76 -
 ...REAM-arm-LLVMLinux-Provide-__aeabi_-symbo.patch |    72 -
 ...REAM-arm-firmware-LLVMLinux-replace-naked.patch |    47 -
 ...ux-Remove-unreachable-from-naked-function.patch |    28 -
 ...nux-Fix-a-cast-to-type-not-present-in-uni.patch |    36 -
 ...nux-Fix-an-inline-asm-input-output-type-m.patch |    41 -
 ...nux-Silence-variable-self-assignment-warn.patch |   123 -
 ...nux-Silence-unicode-warnings-when-preproc.patch |    44 -
 ...e-attributes-error-and-warning-with-clang.patch |    30 -
 ...ed-references-to-acpi_idle_driver-on-aarc.patch |    51 -
 ...ware-LLVMLinux-fix-EFI-libstub-with-clang.patch |    46 -
 ...pto-LLVMLinux-Fix-inline-assembly-for-cla.patch |    46 -
 ...MLinux-Make-spin_lock_prefetch-asm-code-c.patch |    34 -
 ...on-t-use-attribute-externally_visible-whe.patch |    33 -
 ...LLVMLinux-Fix-building-x86_64-AES-extensi.patch |    32 -
 ...ux-Qualify-mul-as-mulq-to-make-clang-happ.patch |    33 -
 ...Linux-Add-Werror-to-cc-option-in-order-to.patch |    55 -
 ...LLVMLinux-Check-for-compiler-support-of-f.patch |    37 -
 ...-6602f080cb28745259e2fab1a4cf55eeb5894f93.patch |   166 +
 0107-intel_idle-tweak-cpuidle-cstates.patch        |    78 +-
 ...Don-t-declare-__force_order-in-kaslr_64.c.patch |    37 +
 7a8b64d17e35810dc3176fe61208b45c15d25402.patch     |    77 +
 9d55bebd9816903b821a403a69a94190442ac043.patch     |    68 +
 ...ble-building-ashmem-and-binder-as-modules.patch |   114 +-
 arm64-common.config                                |     6 +
 btrfs-zstd-backport-to-4.12-api.patch              |    17 -
 common-desktop.config                              |     2 +
 common-server.config                               |     7 +-
 common.config                                      |    32 +-
 export-symbols-needed-by-android-drivers.patch     |    75 +-
 extra-wifi-drivers-compile.patch                   |    50 +
 extra-wifi-drivers-port-to-5.6.patch               |   412 +
 kernel-5.6-kvm-gcc10.patch                         |    12 +
 ...l-5.7-fewer-conditions-for-ARM64_PTR_AUTH.patch |    27 +
 ...l-5.8-nouveau-write-combining-only-on-x86.patch |    72 +
 kernel-release.spec                                |   412 +-
 linux-4.7-intel-dvi-duallink.patch                 |    18 +-
 linux-4.8.1-buildfix.patch                         |    33 -
 linux-5.5-corsair-strafe-quirks.patch              |    22 +
 linux-5.5.tar.sign                                 |    19 -
 linux-5.6-fix-disassembler-4args-detection.patch   |    13 +
 smpboot-no-stack-protector-for-gcc10.patch         |    13 +
 socket.h-include-bitsperlong.h.patch               |    11 +
 ...ernel-and-ramfs-comp-and-decomp-with-zstd.patch |   407 +-
 uksm-5.4.patch => uksm-5.6.patch                   |   118 +-
 uksm-5.7.patch                                     |  6936 ++++++++++
 ...tualBox-guest-shared-folder-vboxsf-support.diff |  3370 -----
 ...86-Add-support-for-ZSTD-compressed-kernel.patch |   102 -
 vbox-6.1-fix-build-on-znver1-hosts.patch           |     6 +-
 x86_64-common.config                               |     3 +-
 66 files changed, 8565 insertions(+), 20509 deletions(-)
 delete mode 100644 0001-Fix-for-compilation-with-clang.patch
 delete mode 100644 0001-kbuild-LLVMLinux-Set-compiler-flags-for-clang.patch
 delete mode 100644 0002-fs-LLVMLinux-Remove-warning-from-COMPATIBLE_IOCTL.patch
 delete mode 100644 0002-lib-Add-zstd-modules.patch
 delete mode 100644 0003-btrfs-Add-zstd-support.patch
 delete mode 100644 0003-kbuild-LLVMLinux-Add-support-for-generating-LLVM-bit.patch
 delete mode 100644 0004-kbuild-LLVMLinux-Make-asm-offset-generation-work-wit.patch
 delete mode 100644 0004-squashfs-Add-zstd-support.patch
 delete mode 100644 0005-crypto-Add-zstd-support.patch
 delete mode 100644 0005-md-sysfs-LLVMLinux-Remove-nested-function-from-bcach.patch
 delete mode 100644 0006-apparmor-LLVMLinux-Remove-VLAIS.patch
 delete mode 100644 0007-exofs-LLVMLinux-Remove-VLAIS-from-exofs-FIXME-Check-.patch
 delete mode 100644 0011-Kbuild-LLVMLinux-Use-Oz-instead-of-Os-when-using-cla.patch
 delete mode 100644 0012-WORKAROUND-x86-boot-LLVMLinux-Work-around-clang-PR39.patch
 delete mode 100644 0013-DO-NOT-UPSTREAM-xen-LLVMLinux-Remove-VLAIS-from-xen-.patch
 delete mode 100644 0014-DO-NOT-UPSTREAM-arm-LLVMLinux-Provide-__aeabi_-symbo.patch
 delete mode 100644 0015-DO-NOT-UPSTREAM-arm-firmware-LLVMLinux-replace-naked.patch
 delete mode 100644 0016-arm-LLVMLinux-Remove-unreachable-from-naked-function.patch
 delete mode 100644 0017-MIPS-LLVMLinux-Fix-a-cast-to-type-not-present-in-uni.patch
 delete mode 100644 0018-MIPS-LLVMLinux-Fix-an-inline-asm-input-output-type-m.patch
 delete mode 100644 0019-MIPS-LLVMLinux-Silence-variable-self-assignment-warn.patch
 delete mode 100644 0020-MIPS-LLVMLinux-Silence-unicode-warnings-when-preproc.patch
 delete mode 100644 0021-Don-t-use-attributes-error-and-warning-with-clang.patch
 delete mode 100644 0022-Fix-undefined-references-to-acpi_idle_driver-on-aarc.patch
 delete mode 100644 0023-HACK-firmware-LLVMLinux-fix-EFI-libstub-with-clang.patch
 delete mode 100644 0024-aarch64-crypto-LLVMLinux-Fix-inline-assembly-for-cla.patch
 delete mode 100644 0025-aarch64-LLVMLinux-Make-spin_lock_prefetch-asm-code-c.patch
 delete mode 100644 0026-LLVMLinux-Don-t-use-attribute-externally_visible-whe.patch
 delete mode 100644 0027-x86-crypto-LLVMLinux-Fix-building-x86_64-AES-extensi.patch
 delete mode 100644 0028-x86-LLVMLinux-Qualify-mul-as-mulq-to-make-clang-happ.patch
 delete mode 100644 0029-kbuild-LLVMLinux-Add-Werror-to-cc-option-in-order-to.patch
 delete mode 100644 0030-x86-kbuild-LLVMLinux-Check-for-compiler-support-of-f.patch
 create mode 100644 01-revert-6602f080cb28745259e2fab1a4cf55eeb5894f93.patch
 create mode 100644 2-0001-x86-Don-t-declare-__force_order-in-kaslr_64.c.patch
 create mode 100644 7a8b64d17e35810dc3176fe61208b45c15d25402.patch
 create mode 100644 9d55bebd9816903b821a403a69a94190442ac043.patch
 delete mode 100644 btrfs-zstd-backport-to-4.12-api.patch
 create mode 100644 extra-wifi-drivers-compile.patch
 create mode 100644 extra-wifi-drivers-port-to-5.6.patch
 create mode 100644 kernel-5.6-kvm-gcc10.patch
 create mode 100644 kernel-5.7-fewer-conditions-for-ARM64_PTR_AUTH.patch
 create mode 100644 kernel-5.8-nouveau-write-combining-only-on-x86.patch
 delete mode 100644 linux-4.8.1-buildfix.patch
 create mode 100644 linux-5.5-corsair-strafe-quirks.patch
 delete mode 100644 linux-5.5.tar.sign
 create mode 100644 linux-5.6-fix-disassembler-4args-detection.patch
 create mode 100644 smpboot-no-stack-protector-for-gcc10.patch
 create mode 100644 socket.h-include-bitsperlong.h.patch
 rename v2-1-2-lib-Add-support-for-ZSTD-compressed-kernel.patch => support-kernel-and-ramfs-comp-and-decomp-with-zstd.patch (59%)
 rename uksm-5.4.patch => uksm-5.6.patch (98%)
 create mode 100644 uksm-5.7.patch
 delete mode 100644 v19-fs-Add-VirtualBox-guest-shared-folder-vboxsf-support.diff
 delete mode 100644 v2-2-2-x86-Add-support-for-ZSTD-compressed-kernel.patch

diff --git a/.abf.yml b/.abf.yml
index 8d4126e..cef6999 100644
--- a/.abf.yml
+++ b/.abf.yml
@@ -1,4 +1,6 @@
 sources:
-  linux-5.5.tar.xz: 527d2b01ecb9223ee041ba4b5d703cfe739783b2
+  extra-wifi-drivers-20200301.tar.zst: 3390c738c7d91250714ce0f88d26371e93bc40b8
   saa716x-driver.tar.xz: f9b6ef1cd6f1f71f53d9a8aadfba2cf6b5c3d7b6
-  patch-5.5.2.xz: b547b768b4e11b254cfd53f87d6ce7417b9c3356
+  linux-5.7.tar.xz: 07e40057b78f1c9dd2b042056325d99fcf9f8a08
+  linux-5.7.tar.sign: 9bdcc9e48ee0eca0baa8118b9a10ba2eda64363e
+  patch-5.7.11.xz: 22e11d00f94d641f0167db08f61a8feb2f231a25
diff --git a/0001-Fix-for-compilation-with-clang.patch b/0001-Fix-for-compilation-with-clang.patch
deleted file mode 100644
index b5da460..0000000
--- a/0001-Fix-for-compilation-with-clang.patch
+++ /dev/null
@@ -1,42 +0,0 @@
->From 14e1f86ad297cc3c2b575135699b8c08849d1c6a Mon Sep 17 00:00:00 2001
-From: Hans Petter Selasky <hps@selasky.org>
-Date: Tue, 25 Apr 2017 09:30:47 +0200
-Subject: [PATCH] Fix for compilation with clang
-
-The clang compiler treats the memory address of a structure field as
-tautologically true when compared to non-NULL. This breaks the macros
-in llist.h when compiled using clang. Fix this by not comparing the
-address of a structure field to NULL when determining the end of the
-llist. Refer to -Wtautological-pointer-compare in clang for more
-information.
-
-Signed-off-by: Hans Petter Selasky <hps@selasky.org>
----
- include/linux/llist.h | 4 ++--
- 1 file changed, 2 insertions(+), 2 deletions(-)
-
-diff --git a/include/linux/llist.h b/include/linux/llist.h
-index 171baa90f6f6..76a6d618de75 100644
---- a/include/linux/llist.h
-+++ b/include/linux/llist.h
-@@ -126,7 +126,7 @@ static inline void init_llist_head(struct llist_head *list)
-  */
- #define llist_for_each_entry(pos, node, member)				\
- 	for ((pos) = llist_entry((node), typeof(*(pos)), member);	\
--	     &(pos)->member != NULL;					\
-+	     (pos) != llist_entry(NULL, typeof(*(pos)), member);	\
- 	     (pos) = llist_entry((pos)->member.next, typeof(*(pos)), member))
- 
- /**
-@@ -148,7 +148,7 @@ static inline void init_llist_head(struct llist_head *list)
-  */
- #define llist_for_each_entry_safe(pos, n, node, member)			       \
- 	for (pos = llist_entry((node), typeof(*pos), member);		       \
--	     &pos->member != NULL &&					       \
-+	     (pos) != llist_entry(NULL, typeof(*(pos)), member) &&	       \
- 	        (n = llist_entry(pos->member.next, typeof(*n), member), true); \
- 	     pos = n)
- 
--- 
-2.11.1
-
diff --git a/0001-cpu-5.2-merge-graysky-s-patchset.patch b/0001-cpu-5.2-merge-graysky-s-patchset.patch
index 38091dc..ad3c196 100644
--- a/0001-cpu-5.2-merge-graysky-s-patchset.patch
+++ b/0001-cpu-5.2-merge-graysky-s-patchset.patch
@@ -1,6 +1,6 @@
-diff -up linux-5.5-rc1/arch/x86/include/asm/module.h.17~ linux-5.5-rc1/arch/x86/include/asm/module.h
---- linux-5.5-rc1/arch/x86/include/asm/module.h.17~	2019-12-08 23:57:55.000000000 +0100
-+++ linux-5.5-rc1/arch/x86/include/asm/module.h	2019-12-10 17:00:36.851313621 +0100
+diff -up linux-5.5-rc1/arch/x86/include/asm/vermagic.h.17~ linux-5.5-rc1/arch/x86/include/asm/vermagic.h
+--- linux-5.5-rc1/arch/x86/include/asm/vermagic.h.17~	2019-12-08 23:57:55.000000000 +0100
++++ linux-5.5-rc1/arch/x86/include/asm/vermagic.h	2019-12-10 17:00:36.851313621 +0100
 @@ -27,6 +27,30 @@ struct mod_arch_specific {
  #define MODULE_PROC_FAMILY "586MMX "
  #elif defined CONFIG_MCORE2
diff --git a/0001-kbuild-LLVMLinux-Set-compiler-flags-for-clang.patch b/0001-kbuild-LLVMLinux-Set-compiler-flags-for-clang.patch
deleted file mode 100644
index e906bf9..0000000
--- a/0001-kbuild-LLVMLinux-Set-compiler-flags-for-clang.patch
+++ /dev/null
@@ -1,59 +0,0 @@
---- linux-4.10/Makefile.1000~	2017-04-25 20:31:43.347811443 +0200
-+++ linux-4.10/Makefile	2017-04-25 20:32:20.593089243 +0200
-@@ -375,6 +375,19 @@ LDFLAGS_vmlinux =
- CFLAGS_GCOV	:= -fprofile-arcs -ftest-coverage -fno-tree-loop-im $(call cc-disable-warning,maybe-uninitialized,)
- CFLAGS_KCOV	:= $(call cc-option,-fsanitize-coverage=trace-pc,)
- 
-+ifeq ($(cc-name),clang)
-+ifneq ($(CROSS_COMPILE),)
-+CLANG_TARGET	:= -target $(notdir $(CROSS_COMPILE:%-=%))
-+GCC_TOOLCHAIN	:= $(dir $(CROSS_COMPILE))
-+endif
-+ifneq ($(GCC_TOOLCHAIN),)
-+CLANG_GCC_TC	:= -gcc-toolchain $(GCC_TOOLCHAIN)
-+endif
-+ifneq ($(IA),1)
-+CLANG_IA_FLAG	:= -fno-integrated-as
-+endif
-+CLANG_FLAGS	:= $(CLANG_TARGET) $(CLANG_GCC_TC) $(CLANG_IA_FLAG) -meabi gnu
-+endif
- 
- # Use USERINCLUDE when you must reference the UAPI directories only.
- USERINCLUDE    := \
-@@ -400,13 +413,13 @@ KBUILD_CPPFLAGS := -D__KERNEL__
- KBUILD_CFLAGS   := -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs \
- 		   -fno-strict-aliasing -fno-common \
- 		   -Werror-implicit-function-declaration \
--		   -Wno-format-security \
-+		   -Wno-format-security $(CLANG_FLAGS) \
- 		   -std=gnu89 $(call cc-option,-fno-PIE)
- 
- 
- KBUILD_AFLAGS_KERNEL :=
- KBUILD_CFLAGS_KERNEL :=
--KBUILD_AFLAGS   := -D__ASSEMBLY__ $(call cc-option,-fno-PIE)
-+KBUILD_AFLAGS   := -D__ASSEMBLY__ $(call cc-option,-fno-PIE) $(CLANG_FLAGS)
- KBUILD_AFLAGS_MODULE  := -DMODULE
- KBUILD_CFLAGS_MODULE  := -DMODULE
- KBUILD_LDFLAGS_MODULE := -T $(srctree)/scripts/module-common.lds
-@@ -700,9 +713,12 @@ KBUILD_CFLAGS += $(stackp-flag)
- ifeq ($(cc-name),clang)
- KBUILD_CPPFLAGS += $(call cc-option,-Qunused-arguments,)
- KBUILD_CPPFLAGS += $(call cc-option,-Wno-unknown-warning-option,)
-+KBUILD_CPPFLAGS += $(call cc-option,-fno-integrated-as,)
- KBUILD_CFLAGS += $(call cc-disable-warning, unused-variable)
- KBUILD_CFLAGS += $(call cc-disable-warning, format-invalid-specifier)
- KBUILD_CFLAGS += $(call cc-disable-warning, gnu)
-+KBUILD_CFLAGS += $(call cc-disable-warning, asm-operand-widths)
-+KBUILD_CFLAGS += $(call cc-disable-warning, initializer-overrides)
- # Quiet clang warning: comparison of unsigned expression < 0 is always false
- KBUILD_CFLAGS += $(call cc-disable-warning, tautological-compare)
- # CLANG uses a _MergedGlobals as optimization, but this breaks modpost, as the
-@@ -710,6 +726,7 @@ KBUILD_CFLAGS += $(call cc-disable-warni
- # See modpost pattern 2
- KBUILD_CFLAGS += $(call cc-option, -mno-global-merge,)
- KBUILD_CFLAGS += $(call cc-option, -fcatch-undefined-behavior)
-+KBUILD_CFLAGS += $(call cc-option, -fno-builtin)
- else
- 
- # These warnings generated too much noise in a regular build.
diff --git a/0002-fs-LLVMLinux-Remove-warning-from-COMPATIBLE_IOCTL.patch b/0002-fs-LLVMLinux-Remove-warning-from-COMPATIBLE_IOCTL.patch
deleted file mode 100644
index dc11701..0000000
--- a/0002-fs-LLVMLinux-Remove-warning-from-COMPATIBLE_IOCTL.patch
+++ /dev/null
@@ -1,37 +0,0 @@
-From 73207dae13dbf5219a244376e6d3d6bc59b0a56b Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?= <bero@linaro.org>
-Date: Thu, 23 Feb 2017 00:06:22 +0100
-Subject: [PATCH 02/30] fs, LLVMLinux: Remove warning from COMPATIBLE_IOCTL
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-cmd in COMPATIBLE_IOCTL is always a u32, so cast it so there isn't a warning
-about an overflow in XFORM.
-
-Author: Mark Charlebois <charlebm@gmail.com>
-Signed-off-by: Mark Charlebois <charlebm@gmail.com>
-Signed-off-by: Behan Webster <behanw@converseincode.com>
-Signed-off-by: Bernhard Rosenkränzer <bero@linaro.org>
-Forward-ported-by: Bernhard Rosenkränzer <bero@linaro.org>
-Acked-by: Arnd Bergmann <arnd@arndb.de>
----
- fs/compat_ioctl.c | 2 +-
- 1 file changed, 1 insertion(+), 1 deletion(-)
-
-diff --git a/fs/compat_ioctl.c b/fs/compat_ioctl.c
-index 11d087b2b28e..6116d5275a3e 100644
---- a/fs/compat_ioctl.c
-+++ b/fs/compat_ioctl.c
-@@ -833,7 +833,7 @@ static int compat_ioctl_preallocate(struct file *file,
-  */
- #define XFORM(i) (((i) ^ ((i) << 27) ^ ((i) << 17)) & 0xffffffff)
- 
--#define COMPATIBLE_IOCTL(cmd) XFORM(cmd),
-+#define COMPATIBLE_IOCTL(cmd) XFORM((u32)cmd),
- /* ioctl should not be warned about even if it's not implemented.
-    Valid reasons to use this:
-    - It is implemented with ->compat_ioctl on some device, but programs
--- 
-2.11.0
-
diff --git a/0002-lib-Add-zstd-modules.patch b/0002-lib-Add-zstd-modules.patch
deleted file mode 100644
index eb8b8b2..0000000
--- a/0002-lib-Add-zstd-modules.patch
+++ /dev/null
@@ -1,13301 +0,0 @@
-From b7f044163968d724be55bf4841fd80babe036dc2 Mon Sep 17 00:00:00 2001
-From: Nick Terrell <terrelln@fb.com>
-Date: Mon, 17 Jul 2017 17:08:19 -0700
-Subject: [PATCH v5 2/5] lib: Add zstd modules
-
-Add zstd compression and decompression kernel modules.
-zstd offers a wide varity of compression speed and quality trade-offs.
-It can compress at speeds approaching lz4, and quality approaching lzma.
-zstd decompressions at speeds more than twice as fast as zlib, and
-decompression speed remains roughly the same across all compression levels.
-
-The code was ported from the upstream zstd source repository. The
-`linux/zstd.h` header was modified to match linux kernel style.
-The cross-platform and allocation code was stripped out. Instead zstd
-requires the caller to pass a preallocated workspace. The source files
-were clang-formatted [1] to match the Linux Kernel style as much as
-possible. Otherwise, the code was unmodified. We would like to avoid
-as much further manual modification to the source code as possible, so it
-will be easier to keep the kernel zstd up to date.
-
-I benchmarked zstd compression as a special character device. I ran zstd
-and zlib compression at several levels, as well as performing no
-compression, which measure the time spent copying the data to kernel space.
-Data is passed to the compresser 4096 B at a time. The benchmark file is
-located in the upstream zstd source repository under
-`contrib/linux-kernel/zstd_compress_test.c` [2].
-
-I ran the benchmarks on a Ubuntu 14.04 VM with 2 cores and 4 GiB of RAM.
-The VM is running on a MacBook Pro with a 3.1 GHz Intel Core i7 processor,
-16 GB of RAM, and a SSD. I benchmarked using `silesia.tar` [3], which is
-211,988,480 B large. Run the following commands for the benchmark:
-
-    sudo modprobe zstd_compress_test
-    sudo mknod zstd_compress_test c 245 0
-    sudo cp silesia.tar zstd_compress_test
-
-The time is reported by the time of the userland `cp`.
-The MB/s is computed with
-
-    1,536,217,008 B / time(buffer size, hash)
-
-which includes the time to copy from userland.
-The Adjusted MB/s is computed with
-
-    1,536,217,088 B / (time(buffer size, hash) - time(buffer size, none)).
-
-The memory reported is the amount of memory the compressor requests.
-
-| Method   | Size (B) | Time (s) | Ratio | MB/s    | Adj MB/s | Mem (MB) |
-|----------|----------|----------|-------|---------|----------|----------|
-| none     | 11988480 |    0.100 |     1 | 2119.88 |        - |        - |
-| zstd -1  | 73645762 |    1.044 | 2.878 |  203.05 |   224.56 |     1.23 |
-| zstd -3  | 66988878 |    1.761 | 3.165 |  120.38 |   127.63 |     2.47 |
-| zstd -5  | 65001259 |    2.563 | 3.261 |   82.71 |    86.07 |     2.86 |
-| zstd -10 | 60165346 |   13.242 | 3.523 |   16.01 |    16.13 |    13.22 |
-| zstd -15 | 58009756 |   47.601 | 3.654 |    4.45 |     4.46 |    21.61 |
-| zstd -19 | 54014593 |  102.835 | 3.925 |    2.06 |     2.06 |    60.15 |
-| zlib -1  | 77260026 |    2.895 | 2.744 |   73.23 |    75.85 |     0.27 |
-| zlib -3  | 72972206 |    4.116 | 2.905 |   51.50 |    52.79 |     0.27 |
-| zlib -6  | 68190360 |    9.633 | 3.109 |   22.01 |    22.24 |     0.27 |
-| zlib -9  | 67613382 |   22.554 | 3.135 |    9.40 |     9.44 |     0.27 |
-
-I benchmarked zstd decompression using the same method on the same machine.
-The benchmark file is located in the upstream zstd repo under
-`contrib/linux-kernel/zstd_decompress_test.c` [4]. The memory reported is
-the amount of memory required to decompress data compressed with the given
-compression level. If you know the maximum size of your input, you can
-reduce the memory usage of decompression irrespective of the compression
-level.
-
-| Method   | Time (s) | MB/s    | Adjusted MB/s | Memory (MB) |
-|----------|----------|---------|---------------|-------------|
-| none     |    0.025 | 8479.54 |             - |           - |
-| zstd -1  |    0.358 |  592.15 |        636.60 |        0.84 |
-| zstd -3  |    0.396 |  535.32 |        571.40 |        1.46 |
-| zstd -5  |    0.396 |  535.32 |        571.40 |        1.46 |
-| zstd -10 |    0.374 |  566.81 |        607.42 |        2.51 |
-| zstd -15 |    0.379 |  559.34 |        598.84 |        4.61 |
-| zstd -19 |    0.412 |  514.54 |        547.77 |        8.80 |
-| zlib -1  |    0.940 |  225.52 |        231.68 |        0.04 |
-| zlib -3  |    0.883 |  240.08 |        247.07 |        0.04 |
-| zlib -6  |    0.844 |  251.17 |        258.84 |        0.04 |
-| zlib -9  |    0.837 |  253.27 |        287.64 |        0.04 |
-
-Tested in userland using the test-suite in the zstd repo under
-`contrib/linux-kernel/test/UserlandTest.cpp` [5] by mocking the kernel
-functions. Fuzz tested using libfuzzer [6] with the fuzz harnesses under
-`contrib/linux-kernel/test/{RoundTripCrash.c,DecompressCrash.c}` [7] [8]
-with ASAN, UBSAN, and MSAN. Additionaly, it was tested while testing the
-BtrFS and SquashFS patches coming next.
-
-[1] https://clang.llvm.org/docs/ClangFormat.html
-[2] https://github.com/facebook/zstd/blob/dev/contrib/linux-kernel/zstd_compress_test.c
-[3] http://sun.aei.polsl.pl/~sdeor/index.php?page=silesia
-[4] https://github.com/facebook/zstd/blob/dev/contrib/linux-kernel/zstd_decompress_test.c
-[5] https://github.com/facebook/zstd/blob/dev/contrib/linux-kernel/test/UserlandTest.cpp
-[6] http://llvm.org/docs/LibFuzzer.html
-[7] https://github.com/facebook/zstd/blob/dev/contrib/linux-kernel/test/RoundTripCrash.c
-[8] https://github.com/facebook/zstd/blob/dev/contrib/linux-kernel/test/DecompressCrash.c
-
-zstd source repository: https://github.com/facebook/zstd
-
-Signed-off-by: Nick Terrell <terrelln@fb.com>
----
-v1 -> v2:
-- Use div_u64() for division of u64s
-- Reduce stack usage of ZSTD_compressSequences(), ZSTD_buildSeqTable(),
-  ZSTD_decompressSequencesLong(), FSE_buildDTable(), FSE_decompress_wksp(),
-  HUF_writeCTable(), HUF_readStats(), HUF_readCTable(),
-  HUF_compressWeights(), HUF_readDTableX2(), and HUF_readDTableX4()
-- No function uses more than 400 B of stack space
-
-v2 -> v3:
-- Work around gcc-7 bug https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81388
-- Fix bug in dictionary compression from upstream commit cc1522351f
-
-v3 -> v4:
-- Fix minor compiler warnings
-
-v4 -> v5:
-- Fix rare compression bug from upstream commit 308047eb5d
-- Fix bug introduced in v3 when working around the gcc-7 bug
-
- include/linux/zstd.h      | 1157 +++++++++++++++
- lib/Kconfig               |    8 +
- lib/Makefile              |    2 +
- lib/zstd/Makefile         |   18 +
- lib/zstd/bitstream.h      |  374 +++++
- lib/zstd/compress.c       | 3484 +++++++++++++++++++++++++++++++++++++++++++++
- lib/zstd/decompress.c     | 2528 ++++++++++++++++++++++++++++++++
- lib/zstd/entropy_common.c |  243 ++++
- lib/zstd/error_private.h  |   53 +
- lib/zstd/fse.h            |  575 ++++++++
- lib/zstd/fse_compress.c   |  795 +++++++++++
- lib/zstd/fse_decompress.c |  332 +++++
- lib/zstd/huf.h            |  212 +++
- lib/zstd/huf_compress.c   |  770 ++++++++++
- lib/zstd/huf_decompress.c |  960 +++++++++++++
- lib/zstd/mem.h            |  151 ++
- lib/zstd/zstd_common.c    |   75 +
- lib/zstd/zstd_internal.h  |  263 ++++
- lib/zstd/zstd_opt.h       | 1014 +++++++++++++
- 19 files changed, 13014 insertions(+)
- create mode 100644 include/linux/zstd.h
- create mode 100644 lib/zstd/Makefile
- create mode 100644 lib/zstd/bitstream.h
- create mode 100644 lib/zstd/compress.c
- create mode 100644 lib/zstd/decompress.c
- create mode 100644 lib/zstd/entropy_common.c
- create mode 100644 lib/zstd/error_private.h
- create mode 100644 lib/zstd/fse.h
- create mode 100644 lib/zstd/fse_compress.c
- create mode 100644 lib/zstd/fse_decompress.c
- create mode 100644 lib/zstd/huf.h
- create mode 100644 lib/zstd/huf_compress.c
- create mode 100644 lib/zstd/huf_decompress.c
- create mode 100644 lib/zstd/mem.h
- create mode 100644 lib/zstd/zstd_common.c
- create mode 100644 lib/zstd/zstd_internal.h
- create mode 100644 lib/zstd/zstd_opt.h
-
-diff --git a/include/linux/zstd.h b/include/linux/zstd.h
-new file mode 100644
-index 0000000..249575e
---- /dev/null
-+++ b/include/linux/zstd.h
-@@ -0,0 +1,1157 @@
-+/*
-+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
-+ * All rights reserved.
-+ *
-+ * This source code is licensed under the BSD-style license found in the
-+ * LICENSE file in the root directory of https://github.com/facebook/zstd.
-+ * An additional grant of patent rights can be found in the PATENTS file in the
-+ * same directory.
-+ *
-+ * This program is free software; you can redistribute it and/or modify it under
-+ * the terms of the GNU General Public License version 2 as published by the
-+ * Free Software Foundation. This program is dual-licensed; you may select
-+ * either version 2 of the GNU General Public License ("GPL") or BSD license
-+ * ("BSD").
-+ */
-+
-+#ifndef ZSTD_H
-+#define ZSTD_H
-+
-+/* ======   Dependency   ======*/
-+#include <linux/types.h>   /* size_t */
-+
-+
-+/*-*****************************************************************************
-+ * Introduction
-+ *
-+ * zstd, short for Zstandard, is a fast lossless compression algorithm,
-+ * targeting real-time compression scenarios at zlib-level and better
-+ * compression ratios. The zstd compression library provides in-memory
-+ * compression and decompression functions. The library supports compression
-+ * levels from 1 up to ZSTD_maxCLevel() which is 22. Levels >= 20, labeled
-+ * ultra, should be used with caution, as they require more memory.
-+ * Compression can be done in:
-+ *  - a single step, reusing a context (described as Explicit memory management)
-+ *  - unbounded multiple steps (described as Streaming compression)
-+ * The compression ratio achievable on small data can be highly improved using
-+ * compression with a dictionary in:
-+ *  - a single step (described as Simple dictionary API)
-+ *  - a single step, reusing a dictionary (described as Fast dictionary API)
-+ ******************************************************************************/
-+
-+/*======  Helper functions  ======*/
-+
-+/**
-+ * enum ZSTD_ErrorCode - zstd error codes
-+ *
-+ * Functions that return size_t can be checked for errors using ZSTD_isError()
-+ * and the ZSTD_ErrorCode can be extracted using ZSTD_getErrorCode().
-+ */
-+typedef enum {
-+	ZSTD_error_no_error,
-+	ZSTD_error_GENERIC,
-+	ZSTD_error_prefix_unknown,
-+	ZSTD_error_version_unsupported,
-+	ZSTD_error_parameter_unknown,
-+	ZSTD_error_frameParameter_unsupported,
-+	ZSTD_error_frameParameter_unsupportedBy32bits,
-+	ZSTD_error_frameParameter_windowTooLarge,
-+	ZSTD_error_compressionParameter_unsupported,
-+	ZSTD_error_init_missing,
-+	ZSTD_error_memory_allocation,
-+	ZSTD_error_stage_wrong,
-+	ZSTD_error_dstSize_tooSmall,
-+	ZSTD_error_srcSize_wrong,
-+	ZSTD_error_corruption_detected,
-+	ZSTD_error_checksum_wrong,
-+	ZSTD_error_tableLog_tooLarge,
-+	ZSTD_error_maxSymbolValue_tooLarge,
-+	ZSTD_error_maxSymbolValue_tooSmall,
-+	ZSTD_error_dictionary_corrupted,
-+	ZSTD_error_dictionary_wrong,
-+	ZSTD_error_dictionaryCreation_failed,
-+	ZSTD_error_maxCode
-+} ZSTD_ErrorCode;
-+
-+/**
-+ * ZSTD_maxCLevel() - maximum compression level available
-+ *
-+ * Return: Maximum compression level available.
-+ */
-+int ZSTD_maxCLevel(void);
-+/**
-+ * ZSTD_compressBound() - maximum compressed size in worst case scenario
-+ * @srcSize: The size of the data to compress.
-+ *
-+ * Return:   The maximum compressed size in the worst case scenario.
-+ */
-+size_t ZSTD_compressBound(size_t srcSize);
-+/**
-+ * ZSTD_isError() - tells if a size_t function result is an error code
-+ * @code:  The function result to check for error.
-+ *
-+ * Return: Non-zero iff the code is an error.
-+ */
-+static __attribute__((unused)) unsigned int ZSTD_isError(size_t code)
-+{
-+	return code > (size_t)-ZSTD_error_maxCode;
-+}
-+/**
-+ * ZSTD_getErrorCode() - translates an error function result to a ZSTD_ErrorCode
-+ * @functionResult: The result of a function for which ZSTD_isError() is true.
-+ *
-+ * Return:          The ZSTD_ErrorCode corresponding to the functionResult or 0
-+ *                  if the functionResult isn't an error.
-+ */
-+static __attribute__((unused)) ZSTD_ErrorCode ZSTD_getErrorCode(
-+	size_t functionResult)
-+{
-+	if (!ZSTD_isError(functionResult))
-+		return (ZSTD_ErrorCode)0;
-+	return (ZSTD_ErrorCode)(0 - functionResult);
-+}
-+
-+/**
-+ * enum ZSTD_strategy - zstd compression search strategy
-+ *
-+ * From faster to stronger.
-+ */
-+typedef enum {
-+	ZSTD_fast,
-+	ZSTD_dfast,
-+	ZSTD_greedy,
-+	ZSTD_lazy,
-+	ZSTD_lazy2,
-+	ZSTD_btlazy2,
-+	ZSTD_btopt,
-+	ZSTD_btopt2
-+} ZSTD_strategy;
-+
-+/**
-+ * struct ZSTD_compressionParameters - zstd compression parameters
-+ * @windowLog:    Log of the largest match distance. Larger means more
-+ *                compression, and more memory needed during decompression.
-+ * @chainLog:     Fully searched segment. Larger means more compression, slower,
-+ *                and more memory (useless for fast).
-+ * @hashLog:      Dispatch table. Larger means more compression,
-+ *                slower, and more memory.
-+ * @searchLog:    Number of searches. Larger means more compression and slower.
-+ * @searchLength: Match length searched. Larger means faster decompression,
-+ *                sometimes less compression.
-+ * @targetLength: Acceptable match size for optimal parser (only). Larger means
-+ *                more compression, and slower.
-+ * @strategy:     The zstd compression strategy.
-+ */
-+typedef struct {
-+	unsigned int windowLog;
-+	unsigned int chainLog;
-+	unsigned int hashLog;
-+	unsigned int searchLog;
-+	unsigned int searchLength;
-+	unsigned int targetLength;
-+	ZSTD_strategy strategy;
-+} ZSTD_compressionParameters;
-+
-+/**
-+ * struct ZSTD_frameParameters - zstd frame parameters
-+ * @contentSizeFlag: Controls whether content size will be present in the frame
-+ *                   header (when known).
-+ * @checksumFlag:    Controls whether a 32-bit checksum is generated at the end
-+ *                   of the frame for error detection.
-+ * @noDictIDFlag:    Controls whether dictID will be saved into the frame header
-+ *                   when using dictionary compression.
-+ *
-+ * The default value is all fields set to 0.
-+ */
-+typedef struct {
-+	unsigned int contentSizeFlag;
-+	unsigned int checksumFlag;
-+	unsigned int noDictIDFlag;
-+} ZSTD_frameParameters;
-+
-+/**
-+ * struct ZSTD_parameters - zstd parameters
-+ * @cParams: The compression parameters.
-+ * @fParams: The frame parameters.
-+ */
-+typedef struct {
-+	ZSTD_compressionParameters cParams;
-+	ZSTD_frameParameters fParams;
-+} ZSTD_parameters;
-+
-+/**
-+ * ZSTD_getCParams() - returns ZSTD_compressionParameters for selected level
-+ * @compressionLevel: The compression level from 1 to ZSTD_maxCLevel().
-+ * @estimatedSrcSize: The estimated source size to compress or 0 if unknown.
-+ * @dictSize:         The dictionary size or 0 if a dictionary isn't being used.
-+ *
-+ * Return:            The selected ZSTD_compressionParameters.
-+ */
-+ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel,
-+	unsigned long long estimatedSrcSize, size_t dictSize);
-+
-+/**
-+ * ZSTD_getParams() - returns ZSTD_parameters for selected level
-+ * @compressionLevel: The compression level from 1 to ZSTD_maxCLevel().
-+ * @estimatedSrcSize: The estimated source size to compress or 0 if unknown.
-+ * @dictSize:         The dictionary size or 0 if a dictionary isn't being used.
-+ *
-+ * The same as ZSTD_getCParams() except also selects the default frame
-+ * parameters (all zero).
-+ *
-+ * Return:            The selected ZSTD_parameters.
-+ */
-+ZSTD_parameters ZSTD_getParams(int compressionLevel,
-+	unsigned long long estimatedSrcSize, size_t dictSize);
-+
-+/*-*************************************
-+ * Explicit memory management
-+ **************************************/
-+
-+/**
-+ * ZSTD_CCtxWorkspaceBound() - amount of memory needed to initialize a ZSTD_CCtx
-+ * @cParams: The compression parameters to be used for compression.
-+ *
-+ * If multiple compression parameters might be used, the caller must call
-+ * ZSTD_CCtxWorkspaceBound() for each set of parameters and use the maximum
-+ * size.
-+ *
-+ * Return:   A lower bound on the size of the workspace that is passed to
-+ *           ZSTD_initCCtx().
-+ */
-+size_t ZSTD_CCtxWorkspaceBound(ZSTD_compressionParameters cParams);
-+
-+/**
-+ * struct ZSTD_CCtx - the zstd compression context
-+ *
-+ * When compressing many times it is recommended to allocate a context just once
-+ * and reuse it for each successive compression operation.
-+ */
-+typedef struct ZSTD_CCtx_s ZSTD_CCtx;
-+/**
-+ * ZSTD_initCCtx() - initialize a zstd compression context
-+ * @workspace:     The workspace to emplace the context into. It must outlive
-+ *                 the returned context.
-+ * @workspaceSize: The size of workspace. Use ZSTD_CCtxWorkspaceBound() to
-+ *                 determine how large the workspace must be.
-+ *
-+ * Return:         A compression context emplaced into workspace.
-+ */
-+ZSTD_CCtx *ZSTD_initCCtx(void *workspace, size_t workspaceSize);
-+
-+/**
-+ * ZSTD_compressCCtx() - compress src into dst
-+ * @ctx:         The context. Must have been initialized with a workspace at
-+ *               least as large as ZSTD_CCtxWorkspaceBound(params.cParams).
-+ * @dst:         The buffer to compress src into.
-+ * @dstCapacity: The size of the destination buffer. May be any size, but
-+ *               ZSTD_compressBound(srcSize) is guaranteed to be large enough.
-+ * @src:         The data to compress.
-+ * @srcSize:     The size of the data to compress.
-+ * @params:      The parameters to use for compression. See ZSTD_getParams().
-+ *
-+ * Return:       The compressed size or an error, which can be checked using
-+ *               ZSTD_isError().
-+ */
-+size_t ZSTD_compressCCtx(ZSTD_CCtx *ctx, void *dst, size_t dstCapacity,
-+	const void *src, size_t srcSize, ZSTD_parameters params);
-+
-+/**
-+ * ZSTD_DCtxWorkspaceBound() - amount of memory needed to initialize a ZSTD_DCtx
-+ *
-+ * Return: A lower bound on the size of the workspace that is passed to
-+ *         ZSTD_initDCtx().
-+ */
-+size_t ZSTD_DCtxWorkspaceBound(void);
-+
-+/**
-+ * struct ZSTD_DCtx - the zstd decompression context
-+ *
-+ * When decompressing many times it is recommended to allocate a context just
-+ * once and reuse it for each successive decompression operation.
-+ */
-+typedef struct ZSTD_DCtx_s ZSTD_DCtx;
-+/**
-+ * ZSTD_initDCtx() - initialize a zstd decompression context
-+ * @workspace:     The workspace to emplace the context into. It must outlive
-+ *                 the returned context.
-+ * @workspaceSize: The size of workspace. Use ZSTD_DCtxWorkspaceBound() to
-+ *                 determine how large the workspace must be.
-+ *
-+ * Return:         A decompression context emplaced into workspace.
-+ */
-+ZSTD_DCtx *ZSTD_initDCtx(void *workspace, size_t workspaceSize);
-+
-+/**
-+ * ZSTD_decompressDCtx() - decompress zstd compressed src into dst
-+ * @ctx:         The decompression context.
-+ * @dst:         The buffer to decompress src into.
-+ * @dstCapacity: The size of the destination buffer. Must be at least as large
-+ *               as the decompressed size. If the caller cannot upper bound the
-+ *               decompressed size, then it's better to use the streaming API.
-+ * @src:         The zstd compressed data to decompress. Multiple concatenated
-+ *               frames and skippable frames are allowed.
-+ * @srcSize:     The exact size of the data to decompress.
-+ *
-+ * Return:       The decompressed size or an error, which can be checked using
-+ *               ZSTD_isError().
-+ */
-+size_t ZSTD_decompressDCtx(ZSTD_DCtx *ctx, void *dst, size_t dstCapacity,
-+	const void *src, size_t srcSize);
-+
-+/*-************************
-+ * Simple dictionary API
-+ **************************/
-+
-+/**
-+ * ZSTD_compress_usingDict() - compress src into dst using a dictionary
-+ * @ctx:         The context. Must have been initialized with a workspace at
-+ *               least as large as ZSTD_CCtxWorkspaceBound(params.cParams).
-+ * @dst:         The buffer to compress src into.
-+ * @dstCapacity: The size of the destination buffer. May be any size, but
-+ *               ZSTD_compressBound(srcSize) is guaranteed to be large enough.
-+ * @src:         The data to compress.
-+ * @srcSize:     The size of the data to compress.
-+ * @dict:        The dictionary to use for compression.
-+ * @dictSize:    The size of the dictionary.
-+ * @params:      The parameters to use for compression. See ZSTD_getParams().
-+ *
-+ * Compression using a predefined dictionary. The same dictionary must be used
-+ * during decompression.
-+ *
-+ * Return:       The compressed size or an error, which can be checked using
-+ *               ZSTD_isError().
-+ */
-+size_t ZSTD_compress_usingDict(ZSTD_CCtx *ctx, void *dst, size_t dstCapacity,
-+	const void *src, size_t srcSize, const void *dict, size_t dictSize,
-+	ZSTD_parameters params);
-+
-+/**
-+ * ZSTD_decompress_usingDict() - decompress src into dst using a dictionary
-+ * @ctx:         The decompression context.
-+ * @dst:         The buffer to decompress src into.
-+ * @dstCapacity: The size of the destination buffer. Must be at least as large
-+ *               as the decompressed size. If the caller cannot upper bound the
-+ *               decompressed size, then it's better to use the streaming API.
-+ * @src:         The zstd compressed data to decompress. Multiple concatenated
-+ *               frames and skippable frames are allowed.
-+ * @srcSize:     The exact size of the data to decompress.
-+ * @dict:        The dictionary to use for decompression. The same dictionary
-+ *               must've been used to compress the data.
-+ * @dictSize:    The size of the dictionary.
-+ *
-+ * Return:       The decompressed size or an error, which can be checked using
-+ *               ZSTD_isError().
-+ */
-+size_t ZSTD_decompress_usingDict(ZSTD_DCtx *ctx, void *dst, size_t dstCapacity,
-+	const void *src, size_t srcSize, const void *dict, size_t dictSize);
-+
-+/*-**************************
-+ * Fast dictionary API
-+ ***************************/
-+
-+/**
-+ * ZSTD_CDictWorkspaceBound() - memory needed to initialize a ZSTD_CDict
-+ * @cParams: The compression parameters to be used for compression.
-+ *
-+ * Return:   A lower bound on the size of the workspace that is passed to
-+ *           ZSTD_initCDict().
-+ */
-+size_t ZSTD_CDictWorkspaceBound(ZSTD_compressionParameters cParams);
-+
-+/**
-+ * struct ZSTD_CDict - a digested dictionary to be used for compression
-+ */
-+typedef struct ZSTD_CDict_s ZSTD_CDict;
-+
-+/**
-+ * ZSTD_initCDict() - initialize a digested dictionary for compression
-+ * @dictBuffer:    The dictionary to digest. The buffer is referenced by the
-+ *                 ZSTD_CDict so it must outlive the returned ZSTD_CDict.
-+ * @dictSize:      The size of the dictionary.
-+ * @params:        The parameters to use for compression. See ZSTD_getParams().
-+ * @workspace:     The workspace. It must outlive the returned ZSTD_CDict.
-+ * @workspaceSize: The workspace size. Must be at least
-+ *                 ZSTD_CDictWorkspaceBound(params.cParams).
-+ *
-+ * When compressing multiple messages / blocks with the same dictionary it is
-+ * recommended to load it just once. The ZSTD_CDict merely references the
-+ * dictBuffer, so it must outlive the returned ZSTD_CDict.
-+ *
-+ * Return:         The digested dictionary emplaced into workspace.
-+ */
-+ZSTD_CDict *ZSTD_initCDict(const void *dictBuffer, size_t dictSize,
-+	ZSTD_parameters params, void *workspace, size_t workspaceSize);
-+
-+/**
-+ * ZSTD_compress_usingCDict() - compress src into dst using a ZSTD_CDict
-+ * @ctx:         The context. Must have been initialized with a workspace at
-+ *               least as large as ZSTD_CCtxWorkspaceBound(cParams) where
-+ *               cParams are the compression parameters used to initialize the
-+ *               cdict.
-+ * @dst:         The buffer to compress src into.
-+ * @dstCapacity: The size of the destination buffer. May be any size, but
-+ *               ZSTD_compressBound(srcSize) is guaranteed to be large enough.
-+ * @src:         The data to compress.
-+ * @srcSize:     The size of the data to compress.
-+ * @cdict:       The digested dictionary to use for compression.
-+ * @params:      The parameters to use for compression. See ZSTD_getParams().
-+ *
-+ * Compression using a digested dictionary. The same dictionary must be used
-+ * during decompression.
-+ *
-+ * Return:       The compressed size or an error, which can be checked using
-+ *               ZSTD_isError().
-+ */
-+size_t ZSTD_compress_usingCDict(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity,
-+	const void *src, size_t srcSize, const ZSTD_CDict *cdict);
-+
-+
-+/**
-+ * ZSTD_DDictWorkspaceBound() - memory needed to initialize a ZSTD_DDict
-+ *
-+ * Return:  A lower bound on the size of the workspace that is passed to
-+ *          ZSTD_initDDict().
-+ */
-+size_t ZSTD_DDictWorkspaceBound(void);
-+
-+/**
-+ * struct ZSTD_DDict - a digested dictionary to be used for decompression
-+ */
-+typedef struct ZSTD_DDict_s ZSTD_DDict;
-+
-+/**
-+ * ZSTD_initDDict() - initialize a digested dictionary for decompression
-+ * @dictBuffer:    The dictionary to digest. The buffer is referenced by the
-+ *                 ZSTD_DDict so it must outlive the returned ZSTD_DDict.
-+ * @dictSize:      The size of the dictionary.
-+ * @workspace:     The workspace. It must outlive the returned ZSTD_DDict.
-+ * @workspaceSize: The workspace size. Must be at least
-+ *                 ZSTD_DDictWorkspaceBound().
-+ *
-+ * When decompressing multiple messages / blocks with the same dictionary it is
-+ * recommended to load it just once. The ZSTD_DDict merely references the
-+ * dictBuffer, so it must outlive the returned ZSTD_DDict.
-+ *
-+ * Return:         The digested dictionary emplaced into workspace.
-+ */
-+ZSTD_DDict *ZSTD_initDDict(const void *dictBuffer, size_t dictSize,
-+	void *workspace, size_t workspaceSize);
-+
-+/**
-+ * ZSTD_decompress_usingDDict() - decompress src into dst using a ZSTD_DDict
-+ * @ctx:         The decompression context.
-+ * @dst:         The buffer to decompress src into.
-+ * @dstCapacity: The size of the destination buffer. Must be at least as large
-+ *               as the decompressed size. If the caller cannot upper bound the
-+ *               decompressed size, then it's better to use the streaming API.
-+ * @src:         The zstd compressed data to decompress. Multiple concatenated
-+ *               frames and skippable frames are allowed.
-+ * @srcSize:     The exact size of the data to decompress.
-+ * @ddict:       The digested dictionary to use for decompression. The same
-+ *               dictionary must've been used to compress the data.
-+ *
-+ * Return:       The decompressed size or an error, which can be checked using
-+ *               ZSTD_isError().
-+ */
-+size_t ZSTD_decompress_usingDDict(ZSTD_DCtx *dctx, void *dst,
-+	size_t dstCapacity, const void *src, size_t srcSize,
-+	const ZSTD_DDict *ddict);
-+
-+
-+/*-**************************
-+ * Streaming
-+ ***************************/
-+
-+/**
-+ * struct ZSTD_inBuffer - input buffer for streaming
-+ * @src:  Start of the input buffer.
-+ * @size: Size of the input buffer.
-+ * @pos:  Position where reading stopped. Will be updated.
-+ *        Necessarily 0 <= pos <= size.
-+ */
-+typedef struct ZSTD_inBuffer_s {
-+	const void *src;
-+	size_t size;
-+	size_t pos;
-+} ZSTD_inBuffer;
-+
-+/**
-+ * struct ZSTD_outBuffer - output buffer for streaming
-+ * @dst:  Start of the output buffer.
-+ * @size: Size of the output buffer.
-+ * @pos:  Position where writing stopped. Will be updated.
-+ *        Necessarily 0 <= pos <= size.
-+ */
-+typedef struct ZSTD_outBuffer_s {
-+	void *dst;
-+	size_t size;
-+	size_t pos;
-+} ZSTD_outBuffer;
-+
-+
-+
-+/*-*****************************************************************************
-+ * Streaming compression - HowTo
-+ *
-+ * A ZSTD_CStream object is required to track streaming operation.
-+ * Use ZSTD_initCStream() to initialize a ZSTD_CStream object.
-+ * ZSTD_CStream objects can be reused multiple times on consecutive compression
-+ * operations. It is recommended to re-use ZSTD_CStream in situations where many
-+ * streaming operations will be achieved consecutively. Use one separate
-+ * ZSTD_CStream per thread for parallel execution.
-+ *
-+ * Use ZSTD_compressStream() repetitively to consume input stream.
-+ * The function will automatically update both `pos` fields.
-+ * Note that it may not consume the entire input, in which case `pos < size`,
-+ * and it's up to the caller to present again remaining data.
-+ * It returns a hint for the preferred number of bytes to use as an input for
-+ * the next function call.
-+ *
-+ * At any moment, it's possible to flush whatever data remains within internal
-+ * buffer, using ZSTD_flushStream(). `output->pos` will be updated. There might
-+ * still be some content left within the internal buffer if `output->size` is
-+ * too small. It returns the number of bytes left in the internal buffer and
-+ * must be called until it returns 0.
-+ *
-+ * ZSTD_endStream() instructs to finish a frame. It will perform a flush and
-+ * write frame epilogue. The epilogue is required for decoders to consider a
-+ * frame completed. Similar to ZSTD_flushStream(), it may not be able to flush
-+ * the full content if `output->size` is too small. In which case, call again
-+ * ZSTD_endStream() to complete the flush. It returns the number of bytes left
-+ * in the internal buffer and must be called until it returns 0.
-+ ******************************************************************************/
-+
-+/**
-+ * ZSTD_CStreamWorkspaceBound() - memory needed to initialize a ZSTD_CStream
-+ * @cParams: The compression parameters to be used for compression.
-+ *
-+ * Return:   A lower bound on the size of the workspace that is passed to
-+ *           ZSTD_initCStream() and ZSTD_initCStream_usingCDict().
-+ */
-+size_t ZSTD_CStreamWorkspaceBound(ZSTD_compressionParameters cParams);
-+
-+/**
-+ * struct ZSTD_CStream - the zstd streaming compression context
-+ */
-+typedef struct ZSTD_CStream_s ZSTD_CStream;
-+
-+/*===== ZSTD_CStream management functions =====*/
-+/**
-+ * ZSTD_initCStream() - initialize a zstd streaming compression context
-+ * @params:         The zstd compression parameters.
-+ * @pledgedSrcSize: If params.fParams.contentSizeFlag == 1 then the caller must
-+ *                  pass the source size (zero means empty source). Otherwise,
-+ *                  the caller may optionally pass the source size, or zero if
-+ *                  unknown.
-+ * @workspace:      The workspace to emplace the context into. It must outlive
-+ *                  the returned context.
-+ * @workspaceSize:  The size of workspace.
-+ *                  Use ZSTD_CStreamWorkspaceBound(params.cParams) to determine
-+ *                  how large the workspace must be.
-+ *
-+ * Return:          The zstd streaming compression context.
-+ */
-+ZSTD_CStream *ZSTD_initCStream(ZSTD_parameters params,
-+	unsigned long long pledgedSrcSize, void *workspace,
-+	size_t workspaceSize);
-+
-+/**
-+ * ZSTD_initCStream_usingCDict() - initialize a streaming compression context
-+ * @cdict:          The digested dictionary to use for compression.
-+ * @pledgedSrcSize: Optionally the source size, or zero if unknown.
-+ * @workspace:      The workspace to emplace the context into. It must outlive
-+ *                  the returned context.
-+ * @workspaceSize:  The size of workspace. Call ZSTD_CStreamWorkspaceBound()
-+ *                  with the cParams used to initialize the cdict to determine
-+ *                  how large the workspace must be.
-+ *
-+ * Return:          The zstd streaming compression context.
-+ */
-+ZSTD_CStream *ZSTD_initCStream_usingCDict(const ZSTD_CDict *cdict,
-+	unsigned long long pledgedSrcSize, void *workspace,
-+	size_t workspaceSize);
-+
-+/*===== Streaming compression functions =====*/
-+/**
-+ * ZSTD_resetCStream() - reset the context using parameters from creation
-+ * @zcs:            The zstd streaming compression context to reset.
-+ * @pledgedSrcSize: Optionally the source size, or zero if unknown.
-+ *
-+ * Resets the context using the parameters from creation. Skips dictionary
-+ * loading, since it can be reused. If `pledgedSrcSize` is non-zero the frame
-+ * content size is always written into the frame header.
-+ *
-+ * Return:          Zero or an error, which can be checked using ZSTD_isError().
-+ */
-+size_t ZSTD_resetCStream(ZSTD_CStream *zcs, unsigned long long pledgedSrcSize);
-+/**
-+ * ZSTD_compressStream() - streaming compress some of input into output
-+ * @zcs:    The zstd streaming compression context.
-+ * @output: Destination buffer. `output->pos` is updated to indicate how much
-+ *          compressed data was written.
-+ * @input:  Source buffer. `input->pos` is updated to indicate how much data was
-+ *          read. Note that it may not consume the entire input, in which case
-+ *          `input->pos < input->size`, and it's up to the caller to present
-+ *          remaining data again.
-+ *
-+ * The `input` and `output` buffers may be any size. Guaranteed to make some
-+ * forward progress if `input` and `output` are not empty.
-+ *
-+ * Return:  A hint for the number of bytes to use as the input for the next
-+ *          function call or an error, which can be checked using
-+ *          ZSTD_isError().
-+ */
-+size_t ZSTD_compressStream(ZSTD_CStream *zcs, ZSTD_outBuffer *output,
-+	ZSTD_inBuffer *input);
-+/**
-+ * ZSTD_flushStream() - flush internal buffers into output
-+ * @zcs:    The zstd streaming compression context.
-+ * @output: Destination buffer. `output->pos` is updated to indicate how much
-+ *          compressed data was written.
-+ *
-+ * ZSTD_flushStream() must be called until it returns 0, meaning all the data
-+ * has been flushed. Since ZSTD_flushStream() causes a block to be ended,
-+ * calling it too often will degrade the compression ratio.
-+ *
-+ * Return:  The number of bytes still present within internal buffers or an
-+ *          error, which can be checked using ZSTD_isError().
-+ */
-+size_t ZSTD_flushStream(ZSTD_CStream *zcs, ZSTD_outBuffer *output);
-+/**
-+ * ZSTD_endStream() - flush internal buffers into output and end the frame
-+ * @zcs:    The zstd streaming compression context.
-+ * @output: Destination buffer. `output->pos` is updated to indicate how much
-+ *          compressed data was written.
-+ *
-+ * ZSTD_endStream() must be called until it returns 0, meaning all the data has
-+ * been flushed and the frame epilogue has been written.
-+ *
-+ * Return:  The number of bytes still present within internal buffers or an
-+ *          error, which can be checked using ZSTD_isError().
-+ */
-+size_t ZSTD_endStream(ZSTD_CStream *zcs, ZSTD_outBuffer *output);
-+
-+/**
-+ * ZSTD_CStreamInSize() - recommended size for the input buffer
-+ *
-+ * Return: The recommended size for the input buffer.
-+ */
-+size_t ZSTD_CStreamInSize(void);
-+/**
-+ * ZSTD_CStreamOutSize() - recommended size for the output buffer
-+ *
-+ * When the output buffer is at least this large, it is guaranteed to be large
-+ * enough to flush at least one complete compressed block.
-+ *
-+ * Return: The recommended size for the output buffer.
-+ */
-+size_t ZSTD_CStreamOutSize(void);
-+
-+
-+
-+/*-*****************************************************************************
-+ * Streaming decompression - HowTo
-+ *
-+ * A ZSTD_DStream object is required to track streaming operations.
-+ * Use ZSTD_initDStream() to initialize a ZSTD_DStream object.
-+ * ZSTD_DStream objects can be re-used multiple times.
-+ *
-+ * Use ZSTD_decompressStream() repetitively to consume your input.
-+ * The function will update both `pos` fields.
-+ * If `input->pos < input->size`, some input has not been consumed.
-+ * It's up to the caller to present again remaining data.
-+ * If `output->pos < output->size`, decoder has flushed everything it could.
-+ * Returns 0 iff a frame is completely decoded and fully flushed.
-+ * Otherwise it returns a suggested next input size that will never load more
-+ * than the current frame.
-+ ******************************************************************************/
-+
-+/**
-+ * ZSTD_DStreamWorkspaceBound() - memory needed to initialize a ZSTD_DStream
-+ * @maxWindowSize: The maximum window size allowed for compressed frames.
-+ *
-+ * Return:         A lower bound on the size of the workspace that is passed to
-+ *                 ZSTD_initDStream() and ZSTD_initDStream_usingDDict().
-+ */
-+size_t ZSTD_DStreamWorkspaceBound(size_t maxWindowSize);
-+
-+/**
-+ * struct ZSTD_DStream - the zstd streaming decompression context
-+ */
-+typedef struct ZSTD_DStream_s ZSTD_DStream;
-+/*===== ZSTD_DStream management functions =====*/
-+/**
-+ * ZSTD_initDStream() - initialize a zstd streaming decompression context
-+ * @maxWindowSize: The maximum window size allowed for compressed frames.
-+ * @workspace:     The workspace to emplace the context into. It must outlive
-+ *                 the returned context.
-+ * @workspaceSize: The size of workspace.
-+ *                 Use ZSTD_DStreamWorkspaceBound(maxWindowSize) to determine
-+ *                 how large the workspace must be.
-+ *
-+ * Return:         The zstd streaming decompression context.
-+ */
-+ZSTD_DStream *ZSTD_initDStream(size_t maxWindowSize, void *workspace,
-+	size_t workspaceSize);
-+/**
-+ * ZSTD_initDStream_usingDDict() - initialize streaming decompression context
-+ * @maxWindowSize: The maximum window size allowed for compressed frames.
-+ * @ddict:         The digested dictionary to use for decompression.
-+ * @workspace:     The workspace to emplace the context into. It must outlive
-+ *                 the returned context.
-+ * @workspaceSize: The size of workspace.
-+ *                 Use ZSTD_DStreamWorkspaceBound(maxWindowSize) to determine
-+ *                 how large the workspace must be.
-+ *
-+ * Return:         The zstd streaming decompression context.
-+ */
-+ZSTD_DStream *ZSTD_initDStream_usingDDict(size_t maxWindowSize,
-+	const ZSTD_DDict *ddict, void *workspace, size_t workspaceSize);
-+
-+/*===== Streaming decompression functions =====*/
-+/**
-+ * ZSTD_resetDStream() - reset the context using parameters from creation
-+ * @zds:   The zstd streaming decompression context to reset.
-+ *
-+ * Resets the context using the parameters from creation. Skips dictionary
-+ * loading, since it can be reused.
-+ *
-+ * Return: Zero or an error, which can be checked using ZSTD_isError().
-+ */
-+size_t ZSTD_resetDStream(ZSTD_DStream *zds);
-+/**
-+ * ZSTD_decompressStream() - streaming decompress some of input into output
-+ * @zds:    The zstd streaming decompression context.
-+ * @output: Destination buffer. `output.pos` is updated to indicate how much
-+ *          decompressed data was written.
-+ * @input:  Source buffer. `input.pos` is updated to indicate how much data was
-+ *          read. Note that it may not consume the entire input, in which case
-+ *          `input.pos < input.size`, and it's up to the caller to present
-+ *          remaining data again.
-+ *
-+ * The `input` and `output` buffers may be any size. Guaranteed to make some
-+ * forward progress if `input` and `output` are not empty.
-+ * ZSTD_decompressStream() will not consume the last byte of the frame until
-+ * the entire frame is flushed.
-+ *
-+ * Return:  Returns 0 iff a frame is completely decoded and fully flushed.
-+ *          Otherwise returns a hint for the number of bytes to use as the input
-+ *          for the next function call or an error, which can be checked using
-+ *          ZSTD_isError(). The size hint will never load more than the frame.
-+ */
-+size_t ZSTD_decompressStream(ZSTD_DStream *zds, ZSTD_outBuffer *output,
-+	ZSTD_inBuffer *input);
-+
-+/**
-+ * ZSTD_DStreamInSize() - recommended size for the input buffer
-+ *
-+ * Return: The recommended size for the input buffer.
-+ */
-+size_t ZSTD_DStreamInSize(void);
-+/**
-+ * ZSTD_DStreamOutSize() - recommended size for the output buffer
-+ *
-+ * When the output buffer is at least this large, it is guaranteed to be large
-+ * enough to flush at least one complete decompressed block.
-+ *
-+ * Return: The recommended size for the output buffer.
-+ */
-+size_t ZSTD_DStreamOutSize(void);
-+
-+
-+/* --- Constants ---*/
-+#define ZSTD_MAGICNUMBER            0xFD2FB528   /* >= v0.8.0 */
-+#define ZSTD_MAGIC_SKIPPABLE_START  0x184D2A50U
-+
-+#define ZSTD_CONTENTSIZE_UNKNOWN (0ULL - 1)
-+#define ZSTD_CONTENTSIZE_ERROR   (0ULL - 2)
-+
-+#define ZSTD_WINDOWLOG_MAX_32  27
-+#define ZSTD_WINDOWLOG_MAX_64  27
-+#define ZSTD_WINDOWLOG_MAX \
-+	((unsigned int)(sizeof(size_t) == 4 \
-+		? ZSTD_WINDOWLOG_MAX_32 \
-+		: ZSTD_WINDOWLOG_MAX_64))
-+#define ZSTD_WINDOWLOG_MIN 10
-+#define ZSTD_HASHLOG_MAX ZSTD_WINDOWLOG_MAX
-+#define ZSTD_HASHLOG_MIN        6
-+#define ZSTD_CHAINLOG_MAX     (ZSTD_WINDOWLOG_MAX+1)
-+#define ZSTD_CHAINLOG_MIN      ZSTD_HASHLOG_MIN
-+#define ZSTD_HASHLOG3_MAX      17
-+#define ZSTD_SEARCHLOG_MAX    (ZSTD_WINDOWLOG_MAX-1)
-+#define ZSTD_SEARCHLOG_MIN      1
-+/* only for ZSTD_fast, other strategies are limited to 6 */
-+#define ZSTD_SEARCHLENGTH_MAX   7
-+/* only for ZSTD_btopt, other strategies are limited to 4 */
-+#define ZSTD_SEARCHLENGTH_MIN   3
-+#define ZSTD_TARGETLENGTH_MIN   4
-+#define ZSTD_TARGETLENGTH_MAX 999
-+
-+/* for static allocation */
-+#define ZSTD_FRAMEHEADERSIZE_MAX 18
-+#define ZSTD_FRAMEHEADERSIZE_MIN  6
-+static const size_t ZSTD_frameHeaderSize_prefix = 5;
-+static const size_t ZSTD_frameHeaderSize_min = ZSTD_FRAMEHEADERSIZE_MIN;
-+static const size_t ZSTD_frameHeaderSize_max = ZSTD_FRAMEHEADERSIZE_MAX;
-+/* magic number + skippable frame length */
-+static const size_t ZSTD_skippableHeaderSize = 8;
-+
-+
-+/*-*************************************
-+ * Compressed size functions
-+ **************************************/
-+
-+/**
-+ * ZSTD_findFrameCompressedSize() - returns the size of a compressed frame
-+ * @src:     Source buffer. It should point to the start of a zstd encoded frame
-+ *           or a skippable frame.
-+ * @srcSize: The size of the source buffer. It must be at least as large as the
-+ *           size of the frame.
-+ *
-+ * Return:   The compressed size of the frame pointed to by `src` or an error,
-+ *           which can be check with ZSTD_isError().
-+ *           Suitable to pass to ZSTD_decompress() or similar functions.
-+ */
-+size_t ZSTD_findFrameCompressedSize(const void *src, size_t srcSize);
-+
-+/*-*************************************
-+ * Decompressed size functions
-+ **************************************/
-+/**
-+ * ZSTD_getFrameContentSize() - returns the content size in a zstd frame header
-+ * @src:     It should point to the start of a zstd encoded frame.
-+ * @srcSize: The size of the source buffer. It must be at least as large as the
-+ *           frame header. `ZSTD_frameHeaderSize_max` is always large enough.
-+ *
-+ * Return:   The frame content size stored in the frame header if known.
-+ *           `ZSTD_CONTENTSIZE_UNKNOWN` if the content size isn't stored in the
-+ *           frame header. `ZSTD_CONTENTSIZE_ERROR` on invalid input.
-+ */
-+unsigned long long ZSTD_getFrameContentSize(const void *src, size_t srcSize);
-+
-+/**
-+ * ZSTD_findDecompressedSize() - returns decompressed size of a series of frames
-+ * @src:     It should point to the start of a series of zstd encoded and/or
-+ *           skippable frames.
-+ * @srcSize: The exact size of the series of frames.
-+ *
-+ * If any zstd encoded frame in the series doesn't have the frame content size
-+ * set, `ZSTD_CONTENTSIZE_UNKNOWN` is returned. But frame content size is always
-+ * set when using ZSTD_compress(). The decompressed size can be very large.
-+ * If the source is untrusted, the decompressed size could be wrong or
-+ * intentionally modified. Always ensure the result fits within the
-+ * application's authorized limits. ZSTD_findDecompressedSize() handles multiple
-+ * frames, and so it must traverse the input to read each frame header. This is
-+ * efficient as most of the data is skipped, however it does mean that all frame
-+ * data must be present and valid.
-+ *
-+ * Return:   Decompressed size of all the data contained in the frames if known.
-+ *           `ZSTD_CONTENTSIZE_UNKNOWN` if the decompressed size is unknown.
-+ *           `ZSTD_CONTENTSIZE_ERROR` if an error occurred.
-+ */
-+unsigned long long ZSTD_findDecompressedSize(const void *src, size_t srcSize);
-+
-+/*-*************************************
-+ * Advanced compression functions
-+ **************************************/
-+/**
-+ * ZSTD_checkCParams() - ensure parameter values remain within authorized range
-+ * @cParams: The zstd compression parameters.
-+ *
-+ * Return:   Zero or an error, which can be checked using ZSTD_isError().
-+ */
-+size_t ZSTD_checkCParams(ZSTD_compressionParameters cParams);
-+
-+/**
-+ * ZSTD_adjustCParams() - optimize parameters for a given srcSize and dictSize
-+ * @srcSize:  Optionally the estimated source size, or zero if unknown.
-+ * @dictSize: Optionally the estimated dictionary size, or zero if unknown.
-+ *
-+ * Return:    The optimized parameters.
-+ */
-+ZSTD_compressionParameters ZSTD_adjustCParams(
-+	ZSTD_compressionParameters cParams, unsigned long long srcSize,
-+	size_t dictSize);
-+
-+/*--- Advanced decompression functions ---*/
-+
-+/**
-+ * ZSTD_isFrame() - returns true iff the buffer starts with a valid frame
-+ * @buffer: The source buffer to check.
-+ * @size:   The size of the source buffer, must be at least 4 bytes.
-+ *
-+ * Return: True iff the buffer starts with a zstd or skippable frame identifier.
-+ */
-+unsigned int ZSTD_isFrame(const void *buffer, size_t size);
-+
-+/**
-+ * ZSTD_getDictID_fromDict() - returns the dictionary id stored in a dictionary
-+ * @dict:     The dictionary buffer.
-+ * @dictSize: The size of the dictionary buffer.
-+ *
-+ * Return:    The dictionary id stored within the dictionary or 0 if the
-+ *            dictionary is not a zstd dictionary. If it returns 0 the
-+ *            dictionary can still be loaded as a content-only dictionary.
-+ */
-+unsigned int ZSTD_getDictID_fromDict(const void *dict, size_t dictSize);
-+
-+/**
-+ * ZSTD_getDictID_fromDDict() - returns the dictionary id stored in a ZSTD_DDict
-+ * @ddict: The ddict to find the id of.
-+ *
-+ * Return: The dictionary id stored within `ddict` or 0 if the dictionary is not
-+ *         a zstd dictionary. If it returns 0 `ddict` will be loaded as a
-+ *         content-only dictionary.
-+ */
-+unsigned int ZSTD_getDictID_fromDDict(const ZSTD_DDict *ddict);
-+
-+/**
-+ * ZSTD_getDictID_fromFrame() - returns the dictionary id stored in a zstd frame
-+ * @src:     Source buffer. It must be a zstd encoded frame.
-+ * @srcSize: The size of the source buffer. It must be at least as large as the
-+ *           frame header. `ZSTD_frameHeaderSize_max` is always large enough.
-+ *
-+ * Return:   The dictionary id required to decompress the frame stored within
-+ *           `src` or 0 if the dictionary id could not be decoded. It can return
-+ *           0 if the frame does not require a dictionary, the dictionary id
-+ *           wasn't stored in the frame, `src` is not a zstd frame, or `srcSize`
-+ *           is too small.
-+ */
-+unsigned int ZSTD_getDictID_fromFrame(const void *src, size_t srcSize);
-+
-+/**
-+ * struct ZSTD_frameParams - zstd frame parameters stored in the frame header
-+ * @frameContentSize: The frame content size, or 0 if not present.
-+ * @windowSize:       The window size, or 0 if the frame is a skippable frame.
-+ * @dictID:           The dictionary id, or 0 if not present.
-+ * @checksumFlag:     Whether a checksum was used.
-+ */
-+typedef struct {
-+	unsigned long long frameContentSize;
-+	unsigned int windowSize;
-+	unsigned int dictID;
-+	unsigned int checksumFlag;
-+} ZSTD_frameParams;
-+
-+/**
-+ * ZSTD_getFrameParams() - extracts parameters from a zstd or skippable frame
-+ * @fparamsPtr: On success the frame parameters are written here.
-+ * @src:        The source buffer. It must point to a zstd or skippable frame.
-+ * @srcSize:    The size of the source buffer. `ZSTD_frameHeaderSize_max` is
-+ *              always large enough to succeed.
-+ *
-+ * Return:      0 on success. If more data is required it returns how many bytes
-+ *              must be provided to make forward progress. Otherwise it returns
-+ *              an error, which can be checked using ZSTD_isError().
-+ */
-+size_t ZSTD_getFrameParams(ZSTD_frameParams *fparamsPtr, const void *src,
-+	size_t srcSize);
-+
-+/*-*****************************************************************************
-+ * Buffer-less and synchronous inner streaming functions
-+ *
-+ * This is an advanced API, giving full control over buffer management, for
-+ * users which need direct control over memory.
-+ * But it's also a complex one, with many restrictions (documented below).
-+ * Prefer using normal streaming API for an easier experience
-+ ******************************************************************************/
-+
-+/*-*****************************************************************************
-+ * Buffer-less streaming compression (synchronous mode)
-+ *
-+ * A ZSTD_CCtx object is required to track streaming operations.
-+ * Use ZSTD_initCCtx() to initialize a context.
-+ * ZSTD_CCtx object can be re-used multiple times within successive compression
-+ * operations.
-+ *
-+ * Start by initializing a context.
-+ * Use ZSTD_compressBegin(), or ZSTD_compressBegin_usingDict() for dictionary
-+ * compression,
-+ * or ZSTD_compressBegin_advanced(), for finer parameter control.
-+ * It's also possible to duplicate a reference context which has already been
-+ * initialized, using ZSTD_copyCCtx()
-+ *
-+ * Then, consume your input using ZSTD_compressContinue().
-+ * There are some important considerations to keep in mind when using this
-+ * advanced function :
-+ * - ZSTD_compressContinue() has no internal buffer. It uses externally provided
-+ *   buffer only.
-+ * - Interface is synchronous : input is consumed entirely and produce 1+
-+ *   (or more) compressed blocks.
-+ * - Caller must ensure there is enough space in `dst` to store compressed data
-+ *   under worst case scenario. Worst case evaluation is provided by
-+ *   ZSTD_compressBound().
-+ *   ZSTD_compressContinue() doesn't guarantee recover after a failed
-+ *   compression.
-+ * - ZSTD_compressContinue() presumes prior input ***is still accessible and
-+ *   unmodified*** (up to maximum distance size, see WindowLog).
-+ *   It remembers all previous contiguous blocks, plus one separated memory
-+ *   segment (which can itself consists of multiple contiguous blocks)
-+ * - ZSTD_compressContinue() detects that prior input has been overwritten when
-+ *   `src` buffer overlaps. In which case, it will "discard" the relevant memory
-+ *   section from its history.
-+ *
-+ * Finish a frame with ZSTD_compressEnd(), which will write the last block(s)
-+ * and optional checksum. It's possible to use srcSize==0, in which case, it
-+ * will write a final empty block to end the frame. Without last block mark,
-+ * frames will be considered unfinished (corrupted) by decoders.
-+ *
-+ * `ZSTD_CCtx` object can be re-used (ZSTD_compressBegin()) to compress some new
-+ * frame.
-+ ******************************************************************************/
-+
-+/*=====   Buffer-less streaming compression functions  =====*/
-+size_t ZSTD_compressBegin(ZSTD_CCtx *cctx, int compressionLevel);
-+size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx *cctx, const void *dict,
-+	size_t dictSize, int compressionLevel);
-+size_t ZSTD_compressBegin_advanced(ZSTD_CCtx *cctx, const void *dict,
-+	size_t dictSize, ZSTD_parameters params,
-+	unsigned long long pledgedSrcSize);
-+size_t ZSTD_copyCCtx(ZSTD_CCtx *cctx, const ZSTD_CCtx *preparedCCtx,
-+	unsigned long long pledgedSrcSize);
-+size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx *cctx, const ZSTD_CDict *cdict,
-+	unsigned long long pledgedSrcSize);
-+size_t ZSTD_compressContinue(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity,
-+	const void *src, size_t srcSize);
-+size_t ZSTD_compressEnd(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity,
-+	const void *src, size_t srcSize);
-+
-+
-+
-+/*-*****************************************************************************
-+ * Buffer-less streaming decompression (synchronous mode)
-+ *
-+ * A ZSTD_DCtx object is required to track streaming operations.
-+ * Use ZSTD_initDCtx() to initialize a context.
-+ * A ZSTD_DCtx object can be re-used multiple times.
-+ *
-+ * First typical operation is to retrieve frame parameters, using
-+ * ZSTD_getFrameParams(). It fills a ZSTD_frameParams structure which provide
-+ * important information to correctly decode the frame, such as the minimum
-+ * rolling buffer size to allocate to decompress data (`windowSize`), and the
-+ * dictionary ID used.
-+ * Note: content size is optional, it may not be present. 0 means unknown.
-+ * Note that these values could be wrong, either because of data malformation,
-+ * or because an attacker is spoofing deliberate false information. As a
-+ * consequence, check that values remain within valid application range,
-+ * especially `windowSize`, before allocation. Each application can set its own
-+ * limit, depending on local restrictions. For extended interoperability, it is
-+ * recommended to support at least 8 MB.
-+ * Frame parameters are extracted from the beginning of the compressed frame.
-+ * Data fragment must be large enough to ensure successful decoding, typically
-+ * `ZSTD_frameHeaderSize_max` bytes.
-+ * Result: 0: successful decoding, the `ZSTD_frameParams` structure is filled.
-+ *        >0: `srcSize` is too small, provide at least this many bytes.
-+ *        errorCode, which can be tested using ZSTD_isError().
-+ *
-+ * Start decompression, with ZSTD_decompressBegin() or
-+ * ZSTD_decompressBegin_usingDict(). Alternatively, you can copy a prepared
-+ * context, using ZSTD_copyDCtx().
-+ *
-+ * Then use ZSTD_nextSrcSizeToDecompress() and ZSTD_decompressContinue()
-+ * alternatively.
-+ * ZSTD_nextSrcSizeToDecompress() tells how many bytes to provide as 'srcSize'
-+ * to ZSTD_decompressContinue().
-+ * ZSTD_decompressContinue() requires this _exact_ amount of bytes, or it will
-+ * fail.
-+ *
-+ * The result of ZSTD_decompressContinue() is the number of bytes regenerated
-+ * within 'dst' (necessarily <= dstCapacity). It can be zero, which is not an
-+ * error; it just means ZSTD_decompressContinue() has decoded some metadata
-+ * item. It can also be an error code, which can be tested with ZSTD_isError().
-+ *
-+ * ZSTD_decompressContinue() needs previous data blocks during decompression, up
-+ * to `windowSize`. They should preferably be located contiguously, prior to
-+ * current block. Alternatively, a round buffer of sufficient size is also
-+ * possible. Sufficient size is determined by frame parameters.
-+ * ZSTD_decompressContinue() is very sensitive to contiguity, if 2 blocks don't
-+ * follow each other, make sure that either the compressor breaks contiguity at
-+ * the same place, or that previous contiguous segment is large enough to
-+ * properly handle maximum back-reference.
-+ *
-+ * A frame is fully decoded when ZSTD_nextSrcSizeToDecompress() returns zero.
-+ * Context can then be reset to start a new decompression.
-+ *
-+ * Note: it's possible to know if next input to present is a header or a block,
-+ * using ZSTD_nextInputType(). This information is not required to properly
-+ * decode a frame.
-+ *
-+ * == Special case: skippable frames ==
-+ *
-+ * Skippable frames allow integration of user-defined data into a flow of
-+ * concatenated frames. Skippable frames will be ignored (skipped) by a
-+ * decompressor. The format of skippable frames is as follows:
-+ * a) Skippable frame ID - 4 Bytes, Little endian format, any value from
-+ *    0x184D2A50 to 0x184D2A5F
-+ * b) Frame Size - 4 Bytes, Little endian format, unsigned 32-bits
-+ * c) Frame Content - any content (User Data) of length equal to Frame Size
-+ * For skippable frames ZSTD_decompressContinue() always returns 0.
-+ * For skippable frames ZSTD_getFrameParams() returns fparamsPtr->windowLog==0
-+ * what means that a frame is skippable.
-+ * Note: If fparamsPtr->frameContentSize==0, it is ambiguous: the frame might
-+ *       actually be a zstd encoded frame with no content. For purposes of
-+ *       decompression, it is valid in both cases to skip the frame using
-+ *       ZSTD_findFrameCompressedSize() to find its size in bytes.
-+ * It also returns frame size as fparamsPtr->frameContentSize.
-+ ******************************************************************************/
-+
-+/*=====   Buffer-less streaming decompression functions  =====*/
-+size_t ZSTD_decompressBegin(ZSTD_DCtx *dctx);
-+size_t ZSTD_decompressBegin_usingDict(ZSTD_DCtx *dctx, const void *dict,
-+	size_t dictSize);
-+void   ZSTD_copyDCtx(ZSTD_DCtx *dctx, const ZSTD_DCtx *preparedDCtx);
-+size_t ZSTD_nextSrcSizeToDecompress(ZSTD_DCtx *dctx);
-+size_t ZSTD_decompressContinue(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity,
-+	const void *src, size_t srcSize);
-+typedef enum {
-+	ZSTDnit_frameHeader,
-+	ZSTDnit_blockHeader,
-+	ZSTDnit_block,
-+	ZSTDnit_lastBlock,
-+	ZSTDnit_checksum,
-+	ZSTDnit_skippableFrame
-+} ZSTD_nextInputType_e;
-+ZSTD_nextInputType_e ZSTD_nextInputType(ZSTD_DCtx *dctx);
-+
-+/*-*****************************************************************************
-+ * Block functions
-+ *
-+ * Block functions produce and decode raw zstd blocks, without frame metadata.
-+ * Frame metadata cost is typically ~18 bytes, which can be non-negligible for
-+ * very small blocks (< 100 bytes). User will have to take in charge required
-+ * information to regenerate data, such as compressed and content sizes.
-+ *
-+ * A few rules to respect:
-+ * - Compressing and decompressing require a context structure
-+ *   + Use ZSTD_initCCtx() and ZSTD_initDCtx()
-+ * - It is necessary to init context before starting
-+ *   + compression : ZSTD_compressBegin()
-+ *   + decompression : ZSTD_decompressBegin()
-+ *   + variants _usingDict() are also allowed
-+ *   + copyCCtx() and copyDCtx() work too
-+ * - Block size is limited, it must be <= ZSTD_getBlockSizeMax()
-+ *   + If you need to compress more, cut data into multiple blocks
-+ *   + Consider using the regular ZSTD_compress() instead, as frame metadata
-+ *     costs become negligible when source size is large.
-+ * - When a block is considered not compressible enough, ZSTD_compressBlock()
-+ *   result will be zero. In which case, nothing is produced into `dst`.
-+ *   + User must test for such outcome and deal directly with uncompressed data
-+ *   + ZSTD_decompressBlock() doesn't accept uncompressed data as input!!!
-+ *   + In case of multiple successive blocks, decoder must be informed of
-+ *     uncompressed block existence to follow proper history. Use
-+ *     ZSTD_insertBlock() in such a case.
-+ ******************************************************************************/
-+
-+/* Define for static allocation */
-+#define ZSTD_BLOCKSIZE_ABSOLUTEMAX (128 * 1024)
-+/*=====   Raw zstd block functions  =====*/
-+size_t ZSTD_getBlockSizeMax(ZSTD_CCtx *cctx);
-+size_t ZSTD_compressBlock(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity,
-+	const void *src, size_t srcSize);
-+size_t ZSTD_decompressBlock(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity,
-+	const void *src, size_t srcSize);
-+size_t ZSTD_insertBlock(ZSTD_DCtx *dctx, const void *blockStart,
-+	size_t blockSize);
-+
-+#endif  /* ZSTD_H */
-diff --git a/lib/Kconfig b/lib/Kconfig
-index 5e7541f..0d49ed0 100644
---- a/lib/Kconfig
-+++ b/lib/Kconfig
-@@ -249,6 +249,14 @@ config LZ4HC_COMPRESS
- config LZ4_DECOMPRESS
- 	tristate
-
-+config ZSTD_COMPRESS
-+	select XXHASH
-+	tristate
-+
-+config ZSTD_DECOMPRESS
-+	select XXHASH
-+	tristate
-+
- source "lib/xz/Kconfig"
-
- #
-diff --git a/lib/Makefile b/lib/Makefile
-index d06b68a..d5c8a4f 100644
---- a/lib/Makefile
-+++ b/lib/Makefile
-@@ -116,6 +116,8 @@ obj-$(CONFIG_LZO_DECOMPRESS) += lzo/
- obj-$(CONFIG_LZ4_COMPRESS) += lz4/
- obj-$(CONFIG_LZ4HC_COMPRESS) += lz4/
- obj-$(CONFIG_LZ4_DECOMPRESS) += lz4/
-+obj-$(CONFIG_ZSTD_COMPRESS) += zstd/
-+obj-$(CONFIG_ZSTD_DECOMPRESS) += zstd/
- obj-$(CONFIG_XZ_DEC) += xz/
- obj-$(CONFIG_RAID6_PQ) += raid6/
-
-diff --git a/lib/zstd/Makefile b/lib/zstd/Makefile
-new file mode 100644
-index 0000000..dd0a359
---- /dev/null
-+++ b/lib/zstd/Makefile
-@@ -0,0 +1,18 @@
-+obj-$(CONFIG_ZSTD_COMPRESS) += zstd_compress.o
-+obj-$(CONFIG_ZSTD_DECOMPRESS) += zstd_decompress.o
-+
-+ccflags-y += -O3
-+
-+# Object files unique to zstd_compress and zstd_decompress
-+zstd_compress-y := fse_compress.o huf_compress.o compress.o
-+zstd_decompress-y := huf_decompress.o decompress.o
-+
-+# These object files are shared between the modules.
-+# Always add them to zstd_compress.
-+# Unless both zstd_compress and zstd_decompress are built in
-+# then also add them to zstd_decompress.
-+zstd_compress-y += entropy_common.o fse_decompress.o zstd_common.o
-+
-+ifneq ($(CONFIG_ZSTD_COMPRESS)$(CONFIG_ZSTD_DECOMPRESS),yy)
-+	zstd_decompress-y += entropy_common.o fse_decompress.o zstd_common.o
-+endif
-diff --git a/lib/zstd/bitstream.h b/lib/zstd/bitstream.h
-new file mode 100644
-index 0000000..a826b99
---- /dev/null
-+++ b/lib/zstd/bitstream.h
-@@ -0,0 +1,374 @@
-+/*
-+ * bitstream
-+ * Part of FSE library
-+ * header file (to include)
-+ * Copyright (C) 2013-2016, Yann Collet.
-+ *
-+ * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
-+ *
-+ * Redistribution and use in source and binary forms, with or without
-+ * modification, are permitted provided that the following conditions are
-+ * met:
-+ *
-+ *   * Redistributions of source code must retain the above copyright
-+ * notice, this list of conditions and the following disclaimer.
-+ *   * Redistributions in binary form must reproduce the above
-+ * copyright notice, this list of conditions and the following disclaimer
-+ * in the documentation and/or other materials provided with the
-+ * distribution.
-+ *
-+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
-+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
-+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
-+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
-+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
-+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
-+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
-+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
-+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
-+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
-+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-+ *
-+ * This program is free software; you can redistribute it and/or modify it under
-+ * the terms of the GNU General Public License version 2 as published by the
-+ * Free Software Foundation. This program is dual-licensed; you may select
-+ * either version 2 of the GNU General Public License ("GPL") or BSD license
-+ * ("BSD").
-+ *
-+ * You can contact the author at :
-+ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
-+ */
-+#ifndef BITSTREAM_H_MODULE
-+#define BITSTREAM_H_MODULE
-+
-+/*
-+*  This API consists of small unitary functions, which must be inlined for best performance.
-+*  Since link-time-optimization is not available for all compilers,
-+*  these functions are defined into a .h to be included.
-+*/
-+
-+/*-****************************************
-+*  Dependencies
-+******************************************/
-+#include "error_private.h" /* error codes and messages */
-+#include "mem.h"	   /* unaligned access routines */
-+
-+/*=========================================
-+*  Target specific
-+=========================================*/
-+#define STREAM_ACCUMULATOR_MIN_32 25
-+#define STREAM_ACCUMULATOR_MIN_64 57
-+#define STREAM_ACCUMULATOR_MIN ((U32)(ZSTD_32bits() ? STREAM_ACCUMULATOR_MIN_32 : STREAM_ACCUMULATOR_MIN_64))
-+
-+/*-******************************************
-+*  bitStream encoding API (write forward)
-+********************************************/
-+/* bitStream can mix input from multiple sources.
-+*  A critical property of these streams is that they encode and decode in **reverse** direction.
-+*  So the first bit sequence you add will be the last to be read, like a LIFO stack.
-+*/
-+typedef struct {
-+	size_t bitContainer;
-+	int bitPos;
-+	char *startPtr;
-+	char *ptr;
-+	char *endPtr;
-+} BIT_CStream_t;
-+
-+ZSTD_STATIC size_t BIT_initCStream(BIT_CStream_t *bitC, void *dstBuffer, size_t dstCapacity);
-+ZSTD_STATIC void BIT_addBits(BIT_CStream_t *bitC, size_t value, unsigned nbBits);
-+ZSTD_STATIC void BIT_flushBits(BIT_CStream_t *bitC);
-+ZSTD_STATIC size_t BIT_closeCStream(BIT_CStream_t *bitC);
-+
-+/* Start with initCStream, providing the size of buffer to write into.
-+*  bitStream will never write outside of this buffer.
-+*  `dstCapacity` must be >= sizeof(bitD->bitContainer), otherwise @return will be an error code.
-+*
-+*  bits are first added to a local register.
-+*  Local register is size_t, hence 64-bits on 64-bits systems, or 32-bits on 32-bits systems.
-+*  Writing data into memory is an explicit operation, performed by the flushBits function.
-+*  Hence keep track how many bits are potentially stored into local register to avoid register overflow.
-+*  After a flushBits, a maximum of 7 bits might still be stored into local register.
-+*
-+*  Avoid storing elements of more than 24 bits if you want compatibility with 32-bits bitstream readers.
-+*
-+*  Last operation is to close the bitStream.
-+*  The function returns the final size of CStream in bytes.
-+*  If data couldn't fit into `dstBuffer`, it will return a 0 ( == not storable)
-+*/
-+
-+/*-********************************************
-+*  bitStream decoding API (read backward)
-+**********************************************/
-+typedef struct {
-+	size_t bitContainer;
-+	unsigned bitsConsumed;
-+	const char *ptr;
-+	const char *start;
-+} BIT_DStream_t;
-+
-+typedef enum {
-+	BIT_DStream_unfinished = 0,
-+	BIT_DStream_endOfBuffer = 1,
-+	BIT_DStream_completed = 2,
-+	BIT_DStream_overflow = 3
-+} BIT_DStream_status; /* result of BIT_reloadDStream() */
-+/* 1,2,4,8 would be better for bitmap combinations, but slows down performance a bit ... :( */
-+
-+ZSTD_STATIC size_t BIT_initDStream(BIT_DStream_t *bitD, const void *srcBuffer, size_t srcSize);
-+ZSTD_STATIC size_t BIT_readBits(BIT_DStream_t *bitD, unsigned nbBits);
-+ZSTD_STATIC BIT_DStream_status BIT_reloadDStream(BIT_DStream_t *bitD);
-+ZSTD_STATIC unsigned BIT_endOfDStream(const BIT_DStream_t *bitD);
-+
-+/* Start by invoking BIT_initDStream().
-+*  A chunk of the bitStream is then stored into a local register.
-+*  Local register size is 64-bits on 64-bits systems, 32-bits on 32-bits systems (size_t).
-+*  You can then retrieve bitFields stored into the local register, **in reverse order**.
-+*  Local register is explicitly reloaded from memory by the BIT_reloadDStream() method.
-+*  A reload guarantee a minimum of ((8*sizeof(bitD->bitContainer))-7) bits when its result is BIT_DStream_unfinished.
-+*  Otherwise, it can be less than that, so proceed accordingly.
-+*  Checking if DStream has reached its end can be performed with BIT_endOfDStream().
-+*/
-+
-+/*-****************************************
-+*  unsafe API
-+******************************************/
-+ZSTD_STATIC void BIT_addBitsFast(BIT_CStream_t *bitC, size_t value, unsigned nbBits);
-+/* faster, but works only if value is "clean", meaning all high bits above nbBits are 0 */
-+
-+ZSTD_STATIC void BIT_flushBitsFast(BIT_CStream_t *bitC);
-+/* unsafe version; does not check buffer overflow */
-+
-+ZSTD_STATIC size_t BIT_readBitsFast(BIT_DStream_t *bitD, unsigned nbBits);
-+/* faster, but works only if nbBits >= 1 */
-+
-+/*-**************************************************************
-+*  Internal functions
-+****************************************************************/
-+ZSTD_STATIC unsigned BIT_highbit32(register U32 val) { return 31 - __builtin_clz(val); }
-+
-+/*=====    Local Constants   =====*/
-+static const unsigned BIT_mask[] = {0,       1,       3,       7,	0xF,      0x1F,     0x3F,     0x7F,      0xFF,
-+				    0x1FF,   0x3FF,   0x7FF,   0xFFF,    0x1FFF,   0x3FFF,   0x7FFF,   0xFFFF,    0x1FFFF,
-+				    0x3FFFF, 0x7FFFF, 0xFFFFF, 0x1FFFFF, 0x3FFFFF, 0x7FFFFF, 0xFFFFFF, 0x1FFFFFF, 0x3FFFFFF}; /* up to 26 bits */
-+
-+/*-**************************************************************
-+*  bitStream encoding
-+****************************************************************/
-+/*! BIT_initCStream() :
-+ *  `dstCapacity` must be > sizeof(void*)
-+ *  @return : 0 if success,
-+			  otherwise an error code (can be tested using ERR_isError() ) */
-+ZSTD_STATIC size_t BIT_initCStream(BIT_CStream_t *bitC, void *startPtr, size_t dstCapacity)
-+{
-+	bitC->bitContainer = 0;
-+	bitC->bitPos = 0;
-+	bitC->startPtr = (char *)startPtr;
-+	bitC->ptr = bitC->startPtr;
-+	bitC->endPtr = bitC->startPtr + dstCapacity - sizeof(bitC->ptr);
-+	if (dstCapacity <= sizeof(bitC->ptr))
-+		return ERROR(dstSize_tooSmall);
-+	return 0;
-+}
-+
-+/*! BIT_addBits() :
-+	can add up to 26 bits into `bitC`.
-+	Does not check for register overflow ! */
-+ZSTD_STATIC void BIT_addBits(BIT_CStream_t *bitC, size_t value, unsigned nbBits)
-+{
-+	bitC->bitContainer |= (value & BIT_mask[nbBits]) << bitC->bitPos;
-+	bitC->bitPos += nbBits;
-+}
-+
-+/*! BIT_addBitsFast() :
-+ *  works only if `value` is _clean_, meaning all high bits above nbBits are 0 */
-+ZSTD_STATIC void BIT_addBitsFast(BIT_CStream_t *bitC, size_t value, unsigned nbBits)
-+{
-+	bitC->bitContainer |= value << bitC->bitPos;
-+	bitC->bitPos += nbBits;
-+}
-+
-+/*! BIT_flushBitsFast() :
-+ *  unsafe version; does not check buffer overflow */
-+ZSTD_STATIC void BIT_flushBitsFast(BIT_CStream_t *bitC)
-+{
-+	size_t const nbBytes = bitC->bitPos >> 3;
-+	ZSTD_writeLEST(bitC->ptr, bitC->bitContainer);
-+	bitC->ptr += nbBytes;
-+	bitC->bitPos &= 7;
-+	bitC->bitContainer >>= nbBytes * 8; /* if bitPos >= sizeof(bitContainer)*8 --> undefined behavior */
-+}
-+
-+/*! BIT_flushBits() :
-+ *  safe version; check for buffer overflow, and prevents it.
-+ *  note : does not signal buffer overflow. This will be revealed later on using BIT_closeCStream() */
-+ZSTD_STATIC void BIT_flushBits(BIT_CStream_t *bitC)
-+{
-+	size_t const nbBytes = bitC->bitPos >> 3;
-+	ZSTD_writeLEST(bitC->ptr, bitC->bitContainer);
-+	bitC->ptr += nbBytes;
-+	if (bitC->ptr > bitC->endPtr)
-+		bitC->ptr = bitC->endPtr;
-+	bitC->bitPos &= 7;
-+	bitC->bitContainer >>= nbBytes * 8; /* if bitPos >= sizeof(bitContainer)*8 --> undefined behavior */
-+}
-+
-+/*! BIT_closeCStream() :
-+ *  @return : size of CStream, in bytes,
-+			  or 0 if it could not fit into dstBuffer */
-+ZSTD_STATIC size_t BIT_closeCStream(BIT_CStream_t *bitC)
-+{
-+	BIT_addBitsFast(bitC, 1, 1); /* endMark */
-+	BIT_flushBits(bitC);
-+
-+	if (bitC->ptr >= bitC->endPtr)
-+		return 0; /* doesn't fit within authorized budget : cancel */
-+
-+	return (bitC->ptr - bitC->startPtr) + (bitC->bitPos > 0);
-+}
-+
-+/*-********************************************************
-+* bitStream decoding
-+**********************************************************/
-+/*! BIT_initDStream() :
-+*   Initialize a BIT_DStream_t.
-+*   `bitD` : a pointer to an already allocated BIT_DStream_t structure.
-+*   `srcSize` must be the *exact* size of the bitStream, in bytes.
-+*   @return : size of stream (== srcSize) or an errorCode if a problem is detected
-+*/
-+ZSTD_STATIC size_t BIT_initDStream(BIT_DStream_t *bitD, const void *srcBuffer, size_t srcSize)
-+{
-+	if (srcSize < 1) {
-+		memset(bitD, 0, sizeof(*bitD));
-+		return ERROR(srcSize_wrong);
-+	}
-+
-+	if (srcSize >= sizeof(bitD->bitContainer)) { /* normal case */
-+		bitD->start = (const char *)srcBuffer;
-+		bitD->ptr = (const char *)srcBuffer + srcSize - sizeof(bitD->bitContainer);
-+		bitD->bitContainer = ZSTD_readLEST(bitD->ptr);
-+		{
-+			BYTE const lastByte = ((const BYTE *)srcBuffer)[srcSize - 1];
-+			bitD->bitsConsumed = lastByte ? 8 - BIT_highbit32(lastByte) : 0; /* ensures bitsConsumed is always set */
-+			if (lastByte == 0)
-+				return ERROR(GENERIC); /* endMark not present */
-+		}
-+	} else {
-+		bitD->start = (const char *)srcBuffer;
-+		bitD->ptr = bitD->start;
-+		bitD->bitContainer = *(const BYTE *)(bitD->start);
-+		switch (srcSize) {
-+		case 7: bitD->bitContainer += (size_t)(((const BYTE *)(srcBuffer))[6]) << (sizeof(bitD->bitContainer) * 8 - 16);
-+		case 6: bitD->bitContainer += (size_t)(((const BYTE *)(srcBuffer))[5]) << (sizeof(bitD->bitContainer) * 8 - 24);
-+		case 5: bitD->bitContainer += (size_t)(((const BYTE *)(srcBuffer))[4]) << (sizeof(bitD->bitContainer) * 8 - 32);
-+		case 4: bitD->bitContainer += (size_t)(((const BYTE *)(srcBuffer))[3]) << 24;
-+		case 3: bitD->bitContainer += (size_t)(((const BYTE *)(srcBuffer))[2]) << 16;
-+		case 2: bitD->bitContainer += (size_t)(((const BYTE *)(srcBuffer))[1]) << 8;
-+		default:;
-+		}
-+		{
-+			BYTE const lastByte = ((const BYTE *)srcBuffer)[srcSize - 1];
-+			bitD->bitsConsumed = lastByte ? 8 - BIT_highbit32(lastByte) : 0;
-+			if (lastByte == 0)
-+				return ERROR(GENERIC); /* endMark not present */
-+		}
-+		bitD->bitsConsumed += (U32)(sizeof(bitD->bitContainer) - srcSize) * 8;
-+	}
-+
-+	return srcSize;
-+}
-+
-+ZSTD_STATIC size_t BIT_getUpperBits(size_t bitContainer, U32 const start) { return bitContainer >> start; }
-+
-+ZSTD_STATIC size_t BIT_getMiddleBits(size_t bitContainer, U32 const start, U32 const nbBits) { return (bitContainer >> start) & BIT_mask[nbBits]; }
-+
-+ZSTD_STATIC size_t BIT_getLowerBits(size_t bitContainer, U32 const nbBits) { return bitContainer & BIT_mask[nbBits]; }
-+
-+/*! BIT_lookBits() :
-+ *  Provides next n bits from local register.
-+ *  local register is not modified.
-+ *  On 32-bits, maxNbBits==24.
-+ *  On 64-bits, maxNbBits==56.
-+ *  @return : value extracted
-+ */
-+ZSTD_STATIC size_t BIT_lookBits(const BIT_DStream_t *bitD, U32 nbBits)
-+{
-+	U32 const bitMask = sizeof(bitD->bitContainer) * 8 - 1;
-+	return ((bitD->bitContainer << (bitD->bitsConsumed & bitMask)) >> 1) >> ((bitMask - nbBits) & bitMask);
-+}
-+
-+/*! BIT_lookBitsFast() :
-+*   unsafe version; only works only if nbBits >= 1 */
-+ZSTD_STATIC size_t BIT_lookBitsFast(const BIT_DStream_t *bitD, U32 nbBits)
-+{
-+	U32 const bitMask = sizeof(bitD->bitContainer) * 8 - 1;
-+	return (bitD->bitContainer << (bitD->bitsConsumed & bitMask)) >> (((bitMask + 1) - nbBits) & bitMask);
-+}
-+
-+ZSTD_STATIC void BIT_skipBits(BIT_DStream_t *bitD, U32 nbBits) { bitD->bitsConsumed += nbBits; }
-+
-+/*! BIT_readBits() :
-+ *  Read (consume) next n bits from local register and update.
-+ *  Pay attention to not read more than nbBits contained into local register.
-+ *  @return : extracted value.
-+ */
-+ZSTD_STATIC size_t BIT_readBits(BIT_DStream_t *bitD, U32 nbBits)
-+{
-+	size_t const value = BIT_lookBits(bitD, nbBits);
-+	BIT_skipBits(bitD, nbBits);
-+	return value;
-+}
-+
-+/*! BIT_readBitsFast() :
-+*   unsafe version; only works only if nbBits >= 1 */
-+ZSTD_STATIC size_t BIT_readBitsFast(BIT_DStream_t *bitD, U32 nbBits)
-+{
-+	size_t const value = BIT_lookBitsFast(bitD, nbBits);
-+	BIT_skipBits(bitD, nbBits);
-+	return value;
-+}
-+
-+/*! BIT_reloadDStream() :
-+*   Refill `bitD` from buffer previously set in BIT_initDStream() .
-+*   This function is safe, it guarantees it will not read beyond src buffer.
-+*   @return : status of `BIT_DStream_t` internal register.
-+			  if status == BIT_DStream_unfinished, internal register is filled with >= (sizeof(bitD->bitContainer)*8 - 7) bits */
-+ZSTD_STATIC BIT_DStream_status BIT_reloadDStream(BIT_DStream_t *bitD)
-+{
-+	if (bitD->bitsConsumed > (sizeof(bitD->bitContainer) * 8)) /* should not happen => corruption detected */
-+		return BIT_DStream_overflow;
-+
-+	if (bitD->ptr >= bitD->start + sizeof(bitD->bitContainer)) {
-+		bitD->ptr -= bitD->bitsConsumed >> 3;
-+		bitD->bitsConsumed &= 7;
-+		bitD->bitContainer = ZSTD_readLEST(bitD->ptr);
-+		return BIT_DStream_unfinished;
-+	}
-+	if (bitD->ptr == bitD->start) {
-+		if (bitD->bitsConsumed < sizeof(bitD->bitContainer) * 8)
-+			return BIT_DStream_endOfBuffer;
-+		return BIT_DStream_completed;
-+	}
-+	{
-+		U32 nbBytes = bitD->bitsConsumed >> 3;
-+		BIT_DStream_status result = BIT_DStream_unfinished;
-+		if (bitD->ptr - nbBytes < bitD->start) {
-+			nbBytes = (U32)(bitD->ptr - bitD->start); /* ptr > start */
-+			result = BIT_DStream_endOfBuffer;
-+		}
-+		bitD->ptr -= nbBytes;
-+		bitD->bitsConsumed -= nbBytes * 8;
-+		bitD->bitContainer = ZSTD_readLEST(bitD->ptr); /* reminder : srcSize > sizeof(bitD) */
-+		return result;
-+	}
-+}
-+
-+/*! BIT_endOfDStream() :
-+*   @return Tells if DStream has exactly reached its end (all bits consumed).
-+*/
-+ZSTD_STATIC unsigned BIT_endOfDStream(const BIT_DStream_t *DStream)
-+{
-+	return ((DStream->ptr == DStream->start) && (DStream->bitsConsumed == sizeof(DStream->bitContainer) * 8));
-+}
-+
-+#endif /* BITSTREAM_H_MODULE */
-diff --git a/lib/zstd/compress.c b/lib/zstd/compress.c
-new file mode 100644
-index 0000000..f9166cf
---- /dev/null
-+++ b/lib/zstd/compress.c
-@@ -0,0 +1,3484 @@
-+/**
-+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
-+ * All rights reserved.
-+ *
-+ * This source code is licensed under the BSD-style license found in the
-+ * LICENSE file in the root directory of https://github.com/facebook/zstd.
-+ * An additional grant of patent rights can be found in the PATENTS file in the
-+ * same directory.
-+ *
-+ * This program is free software; you can redistribute it and/or modify it under
-+ * the terms of the GNU General Public License version 2 as published by the
-+ * Free Software Foundation. This program is dual-licensed; you may select
-+ * either version 2 of the GNU General Public License ("GPL") or BSD license
-+ * ("BSD").
-+ */
-+
-+/*-*************************************
-+*  Dependencies
-+***************************************/
-+#include "fse.h"
-+#include "huf.h"
-+#include "mem.h"
-+#include "zstd_internal.h" /* includes zstd.h */
-+#include <linux/kernel.h>
-+#include <linux/module.h>
-+#include <linux/string.h> /* memset */
-+
-+/*-*************************************
-+*  Constants
-+***************************************/
-+static const U32 g_searchStrength = 8; /* control skip over incompressible data */
-+#define HASH_READ_SIZE 8
-+typedef enum { ZSTDcs_created = 0, ZSTDcs_init, ZSTDcs_ongoing, ZSTDcs_ending } ZSTD_compressionStage_e;
-+
-+/*-*************************************
-+*  Helper functions
-+***************************************/
-+size_t ZSTD_compressBound(size_t srcSize) { return FSE_compressBound(srcSize) + 12; }
-+
-+/*-*************************************
-+*  Sequence storage
-+***************************************/
-+static void ZSTD_resetSeqStore(seqStore_t *ssPtr)
-+{
-+	ssPtr->lit = ssPtr->litStart;
-+	ssPtr->sequences = ssPtr->sequencesStart;
-+	ssPtr->longLengthID = 0;
-+}
-+
-+/*-*************************************
-+*  Context memory management
-+***************************************/
-+struct ZSTD_CCtx_s {
-+	const BYTE *nextSrc;  /* next block here to continue on curr prefix */
-+	const BYTE *base;     /* All regular indexes relative to this position */
-+	const BYTE *dictBase; /* extDict indexes relative to this position */
-+	U32 dictLimit;	/* below that point, need extDict */
-+	U32 lowLimit;	 /* below that point, no more data */
-+	U32 nextToUpdate;     /* index from which to continue dictionary update */
-+	U32 nextToUpdate3;    /* index from which to continue dictionary update */
-+	U32 hashLog3;	 /* dispatch table : larger == faster, more memory */
-+	U32 loadedDictEnd;    /* index of end of dictionary */
-+	U32 forceWindow;      /* force back-references to respect limit of 1<<wLog, even for dictionary */
-+	U32 forceRawDict;     /* Force loading dictionary in "content-only" mode (no header analysis) */
-+	ZSTD_compressionStage_e stage;
-+	U32 rep[ZSTD_REP_NUM];
-+	U32 repToConfirm[ZSTD_REP_NUM];
-+	U32 dictID;
-+	ZSTD_parameters params;
-+	void *workSpace;
-+	size_t workSpaceSize;
-+	size_t blockSize;
-+	U64 frameContentSize;
-+	struct xxh64_state xxhState;
-+	ZSTD_customMem customMem;
-+
-+	seqStore_t seqStore; /* sequences storage ptrs */
-+	U32 *hashTable;
-+	U32 *hashTable3;
-+	U32 *chainTable;
-+	HUF_CElt *hufTable;
-+	U32 flagStaticTables;
-+	HUF_repeat flagStaticHufTable;
-+	FSE_CTable offcodeCTable[FSE_CTABLE_SIZE_U32(OffFSELog, MaxOff)];
-+	FSE_CTable matchlengthCTable[FSE_CTABLE_SIZE_U32(MLFSELog, MaxML)];
-+	FSE_CTable litlengthCTable[FSE_CTABLE_SIZE_U32(LLFSELog, MaxLL)];
-+	unsigned tmpCounters[HUF_COMPRESS_WORKSPACE_SIZE_U32];
-+};
-+
-+size_t ZSTD_CCtxWorkspaceBound(ZSTD_compressionParameters cParams)
-+{
-+	size_t const blockSize = MIN(ZSTD_BLOCKSIZE_ABSOLUTEMAX, (size_t)1 << cParams.windowLog);
-+	U32 const divider = (cParams.searchLength == 3) ? 3 : 4;
-+	size_t const maxNbSeq = blockSize / divider;
-+	size_t const tokenSpace = blockSize + 11 * maxNbSeq;
-+	size_t const chainSize = (cParams.strategy == ZSTD_fast) ? 0 : (1 << cParams.chainLog);
-+	size_t const hSize = ((size_t)1) << cParams.hashLog;
-+	U32 const hashLog3 = (cParams.searchLength > 3) ? 0 : MIN(ZSTD_HASHLOG3_MAX, cParams.windowLog);
-+	size_t const h3Size = ((size_t)1) << hashLog3;
-+	size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32);
-+	size_t const optSpace =
-+	    ((MaxML + 1) + (MaxLL + 1) + (MaxOff + 1) + (1 << Litbits)) * sizeof(U32) + (ZSTD_OPT_NUM + 1) * (sizeof(ZSTD_match_t) + sizeof(ZSTD_optimal_t));
-+	size_t const workspaceSize = tableSpace + (256 * sizeof(U32)) /* huffTable */ + tokenSpace +
-+				     (((cParams.strategy == ZSTD_btopt) || (cParams.strategy == ZSTD_btopt2)) ? optSpace : 0);
-+
-+	return ZSTD_ALIGN(sizeof(ZSTD_stack)) + ZSTD_ALIGN(sizeof(ZSTD_CCtx)) + ZSTD_ALIGN(workspaceSize);
-+}
-+
-+static ZSTD_CCtx *ZSTD_createCCtx_advanced(ZSTD_customMem customMem)
-+{
-+	ZSTD_CCtx *cctx;
-+	if (!customMem.customAlloc || !customMem.customFree)
-+		return NULL;
-+	cctx = (ZSTD_CCtx *)ZSTD_malloc(sizeof(ZSTD_CCtx), customMem);
-+	if (!cctx)
-+		return NULL;
-+	memset(cctx, 0, sizeof(ZSTD_CCtx));
-+	cctx->customMem = customMem;
-+	return cctx;
-+}
-+
-+ZSTD_CCtx *ZSTD_initCCtx(void *workspace, size_t workspaceSize)
-+{
-+	ZSTD_customMem const stackMem = ZSTD_initStack(workspace, workspaceSize);
-+	ZSTD_CCtx *cctx = ZSTD_createCCtx_advanced(stackMem);
-+	if (cctx) {
-+		cctx->workSpace = ZSTD_stackAllocAll(cctx->customMem.opaque, &cctx->workSpaceSize);
-+	}
-+	return cctx;
-+}
-+
-+size_t ZSTD_freeCCtx(ZSTD_CCtx *cctx)
-+{
-+	if (cctx == NULL)
-+		return 0; /* support free on NULL */
-+	ZSTD_free(cctx->workSpace, cctx->customMem);
-+	ZSTD_free(cctx, cctx->customMem);
-+	return 0; /* reserved as a potential error code in the future */
-+}
-+
-+const seqStore_t *ZSTD_getSeqStore(const ZSTD_CCtx *ctx) /* hidden interface */ { return &(ctx->seqStore); }
-+
-+static ZSTD_parameters ZSTD_getParamsFromCCtx(const ZSTD_CCtx *cctx) { return cctx->params; }
-+
-+/** ZSTD_checkParams() :
-+	ensure param values remain within authorized range.
-+	@return : 0, or an error code if one value is beyond authorized range */
-+size_t ZSTD_checkCParams(ZSTD_compressionParameters cParams)
-+{
-+#define CLAMPCHECK(val, min, max)                                       \
-+	{                                                               \
-+		if ((val < min) | (val > max))                          \
-+			return ERROR(compressionParameter_unsupported); \
-+	}
-+	CLAMPCHECK(cParams.windowLog, ZSTD_WINDOWLOG_MIN, ZSTD_WINDOWLOG_MAX);
-+	CLAMPCHECK(cParams.chainLog, ZSTD_CHAINLOG_MIN, ZSTD_CHAINLOG_MAX);
-+	CLAMPCHECK(cParams.hashLog, ZSTD_HASHLOG_MIN, ZSTD_HASHLOG_MAX);
-+	CLAMPCHECK(cParams.searchLog, ZSTD_SEARCHLOG_MIN, ZSTD_SEARCHLOG_MAX);
-+	CLAMPCHECK(cParams.searchLength, ZSTD_SEARCHLENGTH_MIN, ZSTD_SEARCHLENGTH_MAX);
-+	CLAMPCHECK(cParams.targetLength, ZSTD_TARGETLENGTH_MIN, ZSTD_TARGETLENGTH_MAX);
-+	if ((U32)(cParams.strategy) > (U32)ZSTD_btopt2)
-+		return ERROR(compressionParameter_unsupported);
-+	return 0;
-+}
-+
-+/** ZSTD_cycleLog() :
-+ *  condition for correct operation : hashLog > 1 */
-+static U32 ZSTD_cycleLog(U32 hashLog, ZSTD_strategy strat)
-+{
-+	U32 const btScale = ((U32)strat >= (U32)ZSTD_btlazy2);
-+	return hashLog - btScale;
-+}
-+
-+/** ZSTD_adjustCParams() :
-+	optimize `cPar` for a given input (`srcSize` and `dictSize`).
-+	mostly downsizing to reduce memory consumption and initialization.
-+	Both `srcSize` and `dictSize` are optional (use 0 if unknown),
-+	but if both are 0, no optimization can be done.
-+	Note : cPar is considered validated at this stage. Use ZSTD_checkParams() to ensure that. */
-+ZSTD_compressionParameters ZSTD_adjustCParams(ZSTD_compressionParameters cPar, unsigned long long srcSize, size_t dictSize)
-+{
-+	if (srcSize + dictSize == 0)
-+		return cPar; /* no size information available : no adjustment */
-+
-+	/* resize params, to use less memory when necessary */
-+	{
-+		U32 const minSrcSize = (srcSize == 0) ? 500 : 0;
-+		U64 const rSize = srcSize + dictSize + minSrcSize;
-+		if (rSize < ((U64)1 << ZSTD_WINDOWLOG_MAX)) {
-+			U32 const srcLog = MAX(ZSTD_HASHLOG_MIN, ZSTD_highbit32((U32)(rSize)-1) + 1);
-+			if (cPar.windowLog > srcLog)
-+				cPar.windowLog = srcLog;
-+		}
-+	}
-+	if (cPar.hashLog > cPar.windowLog)
-+		cPar.hashLog = cPar.windowLog;
-+	{
-+		U32 const cycleLog = ZSTD_cycleLog(cPar.chainLog, cPar.strategy);
-+		if (cycleLog > cPar.windowLog)
-+			cPar.chainLog -= (cycleLog - cPar.windowLog);
-+	}
-+
-+	if (cPar.windowLog < ZSTD_WINDOWLOG_ABSOLUTEMIN)
-+		cPar.windowLog = ZSTD_WINDOWLOG_ABSOLUTEMIN; /* required for frame header */
-+
-+	return cPar;
-+}
-+
-+static U32 ZSTD_equivalentParams(ZSTD_parameters param1, ZSTD_parameters param2)
-+{
-+	return (param1.cParams.hashLog == param2.cParams.hashLog) & (param1.cParams.chainLog == param2.cParams.chainLog) &
-+	       (param1.cParams.strategy == param2.cParams.strategy) & ((param1.cParams.searchLength == 3) == (param2.cParams.searchLength == 3));
-+}
-+
-+/*! ZSTD_continueCCtx() :
-+	reuse CCtx without reset (note : requires no dictionary) */
-+static size_t ZSTD_continueCCtx(ZSTD_CCtx *cctx, ZSTD_parameters params, U64 frameContentSize)
-+{
-+	U32 const end = (U32)(cctx->nextSrc - cctx->base);
-+	cctx->params = params;
-+	cctx->frameContentSize = frameContentSize;
-+	cctx->lowLimit = end;
-+	cctx->dictLimit = end;
-+	cctx->nextToUpdate = end + 1;
-+	cctx->stage = ZSTDcs_init;
-+	cctx->dictID = 0;
-+	cctx->loadedDictEnd = 0;
-+	{
-+		int i;
-+		for (i = 0; i < ZSTD_REP_NUM; i++)
-+			cctx->rep[i] = repStartValue[i];
-+	}
-+	cctx->seqStore.litLengthSum = 0; /* force reset of btopt stats */
-+	xxh64_reset(&cctx->xxhState, 0);
-+	return 0;
-+}
-+
-+typedef enum { ZSTDcrp_continue, ZSTDcrp_noMemset, ZSTDcrp_fullReset } ZSTD_compResetPolicy_e;
-+
-+/*! ZSTD_resetCCtx_advanced() :
-+	note : `params` must be validated */
-+static size_t ZSTD_resetCCtx_advanced(ZSTD_CCtx *zc, ZSTD_parameters params, U64 frameContentSize, ZSTD_compResetPolicy_e const crp)
-+{
-+	if (crp == ZSTDcrp_continue)
-+		if (ZSTD_equivalentParams(params, zc->params)) {
-+			zc->flagStaticTables = 0;
-+			zc->flagStaticHufTable = HUF_repeat_none;
-+			return ZSTD_continueCCtx(zc, params, frameContentSize);
-+		}
-+
-+	{
-+		size_t const blockSize = MIN(ZSTD_BLOCKSIZE_ABSOLUTEMAX, (size_t)1 << params.cParams.windowLog);
-+		U32 const divider = (params.cParams.searchLength == 3) ? 3 : 4;
-+		size_t const maxNbSeq = blockSize / divider;
-+		size_t const tokenSpace = blockSize + 11 * maxNbSeq;
-+		size_t const chainSize = (params.cParams.strategy == ZSTD_fast) ? 0 : (1 << params.cParams.chainLog);
-+		size_t const hSize = ((size_t)1) << params.cParams.hashLog;
-+		U32 const hashLog3 = (params.cParams.searchLength > 3) ? 0 : MIN(ZSTD_HASHLOG3_MAX, params.cParams.windowLog);
-+		size_t const h3Size = ((size_t)1) << hashLog3;
-+		size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32);
-+		void *ptr;
-+
-+		/* Check if workSpace is large enough, alloc a new one if needed */
-+		{
-+			size_t const optSpace = ((MaxML + 1) + (MaxLL + 1) + (MaxOff + 1) + (1 << Litbits)) * sizeof(U32) +
-+						(ZSTD_OPT_NUM + 1) * (sizeof(ZSTD_match_t) + sizeof(ZSTD_optimal_t));
-+			size_t const neededSpace = tableSpace + (256 * sizeof(U32)) /* huffTable */ + tokenSpace +
-+						   (((params.cParams.strategy == ZSTD_btopt) || (params.cParams.strategy == ZSTD_btopt2)) ? optSpace : 0);
-+			if (zc->workSpaceSize < neededSpace) {
-+				ZSTD_free(zc->workSpace, zc->customMem);
-+				zc->workSpace = ZSTD_malloc(neededSpace, zc->customMem);
-+				if (zc->workSpace == NULL)
-+					return ERROR(memory_allocation);
-+				zc->workSpaceSize = neededSpace;
-+			}
-+		}
-+
-+		if (crp != ZSTDcrp_noMemset)
-+			memset(zc->workSpace, 0, tableSpace); /* reset tables only */
-+		xxh64_reset(&zc->xxhState, 0);
-+		zc->hashLog3 = hashLog3;
-+		zc->hashTable = (U32 *)(zc->workSpace);
-+		zc->chainTable = zc->hashTable + hSize;
-+		zc->hashTable3 = zc->chainTable + chainSize;
-+		ptr = zc->hashTable3 + h3Size;
-+		zc->hufTable = (HUF_CElt *)ptr;
-+		zc->flagStaticTables = 0;
-+		zc->flagStaticHufTable = HUF_repeat_none;
-+		ptr = ((U32 *)ptr) + 256; /* note : HUF_CElt* is incomplete type, size is simulated using U32 */
-+
-+		zc->nextToUpdate = 1;
-+		zc->nextSrc = NULL;
-+		zc->base = NULL;
-+		zc->dictBase = NULL;
-+		zc->dictLimit = 0;
-+		zc->lowLimit = 0;
-+		zc->params = params;
-+		zc->blockSize = blockSize;
-+		zc->frameContentSize = frameContentSize;
-+		{
-+			int i;
-+			for (i = 0; i < ZSTD_REP_NUM; i++)
-+				zc->rep[i] = repStartValue[i];
-+		}
-+
-+		if ((params.cParams.strategy == ZSTD_btopt) || (params.cParams.strategy == ZSTD_btopt2)) {
-+			zc->seqStore.litFreq = (U32 *)ptr;
-+			zc->seqStore.litLengthFreq = zc->seqStore.litFreq + (1 << Litbits);
-+			zc->seqStore.matchLengthFreq = zc->seqStore.litLengthFreq + (MaxLL + 1);
-+			zc->seqStore.offCodeFreq = zc->seqStore.matchLengthFreq + (MaxML + 1);
-+			ptr = zc->seqStore.offCodeFreq + (MaxOff + 1);
-+			zc->seqStore.matchTable = (ZSTD_match_t *)ptr;
-+			ptr = zc->seqStore.matchTable + ZSTD_OPT_NUM + 1;
-+			zc->seqStore.priceTable = (ZSTD_optimal_t *)ptr;
-+			ptr = zc->seqStore.priceTable + ZSTD_OPT_NUM + 1;
-+			zc->seqStore.litLengthSum = 0;
-+		}
-+		zc->seqStore.sequencesStart = (seqDef *)ptr;
-+		ptr = zc->seqStore.sequencesStart + maxNbSeq;
-+		zc->seqStore.llCode = (BYTE *)ptr;
-+		zc->seqStore.mlCode = zc->seqStore.llCode + maxNbSeq;
-+		zc->seqStore.ofCode = zc->seqStore.mlCode + maxNbSeq;
-+		zc->seqStore.litStart = zc->seqStore.ofCode + maxNbSeq;
-+
-+		zc->stage = ZSTDcs_init;
-+		zc->dictID = 0;
-+		zc->loadedDictEnd = 0;
-+
-+		return 0;
-+	}
-+}
-+
-+/* ZSTD_invalidateRepCodes() :
-+ * ensures next compression will not use repcodes from previous block.
-+ * Note : only works with regular variant;
-+ *        do not use with extDict variant ! */
-+void ZSTD_invalidateRepCodes(ZSTD_CCtx *cctx)
-+{
-+	int i;
-+	for (i = 0; i < ZSTD_REP_NUM; i++)
-+		cctx->rep[i] = 0;
-+}
-+
-+/*! ZSTD_copyCCtx() :
-+*   Duplicate an existing context `srcCCtx` into another one `dstCCtx`.
-+*   Only works during stage ZSTDcs_init (i.e. after creation, but before first call to ZSTD_compressContinue()).
-+*   @return : 0, or an error code */
-+size_t ZSTD_copyCCtx(ZSTD_CCtx *dstCCtx, const ZSTD_CCtx *srcCCtx, unsigned long long pledgedSrcSize)
-+{
-+	if (srcCCtx->stage != ZSTDcs_init)
-+		return ERROR(stage_wrong);
-+
-+	memcpy(&dstCCtx->customMem, &srcCCtx->customMem, sizeof(ZSTD_customMem));
-+	{
-+		ZSTD_parameters params = srcCCtx->params;
-+		params.fParams.contentSizeFlag = (pledgedSrcSize > 0);
-+		ZSTD_resetCCtx_advanced(dstCCtx, params, pledgedSrcSize, ZSTDcrp_noMemset);
-+	}
-+
-+	/* copy tables */
-+	{
-+		size_t const chainSize = (srcCCtx->params.cParams.strategy == ZSTD_fast) ? 0 : (1 << srcCCtx->params.cParams.chainLog);
-+		size_t const hSize = ((size_t)1) << srcCCtx->params.cParams.hashLog;
-+		size_t const h3Size = (size_t)1 << srcCCtx->hashLog3;
-+		size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32);
-+		memcpy(dstCCtx->workSpace, srcCCtx->workSpace, tableSpace);
-+	}
-+
-+	/* copy dictionary offsets */
-+	dstCCtx->nextToUpdate = srcCCtx->nextToUpdate;
-+	dstCCtx->nextToUpdate3 = srcCCtx->nextToUpdate3;
-+	dstCCtx->nextSrc = srcCCtx->nextSrc;
-+	dstCCtx->base = srcCCtx->base;
-+	dstCCtx->dictBase = srcCCtx->dictBase;
-+	dstCCtx->dictLimit = srcCCtx->dictLimit;
-+	dstCCtx->lowLimit = srcCCtx->lowLimit;
-+	dstCCtx->loadedDictEnd = srcCCtx->loadedDictEnd;
-+	dstCCtx->dictID = srcCCtx->dictID;
-+
-+	/* copy entropy tables */
-+	dstCCtx->flagStaticTables = srcCCtx->flagStaticTables;
-+	dstCCtx->flagStaticHufTable = srcCCtx->flagStaticHufTable;
-+	if (srcCCtx->flagStaticTables) {
-+		memcpy(dstCCtx->litlengthCTable, srcCCtx->litlengthCTable, sizeof(dstCCtx->litlengthCTable));
-+		memcpy(dstCCtx->matchlengthCTable, srcCCtx->matchlengthCTable, sizeof(dstCCtx->matchlengthCTable));
-+		memcpy(dstCCtx->offcodeCTable, srcCCtx->offcodeCTable, sizeof(dstCCtx->offcodeCTable));
-+	}
-+	if (srcCCtx->flagStaticHufTable) {
-+		memcpy(dstCCtx->hufTable, srcCCtx->hufTable, 256 * 4);
-+	}
-+
-+	return 0;
-+}
-+
-+/*! ZSTD_reduceTable() :
-+*   reduce table indexes by `reducerValue` */
-+static void ZSTD_reduceTable(U32 *const table, U32 const size, U32 const reducerValue)
-+{
-+	U32 u;
-+	for (u = 0; u < size; u++) {
-+		if (table[u] < reducerValue)
-+			table[u] = 0;
-+		else
-+			table[u] -= reducerValue;
-+	}
-+}
-+
-+/*! ZSTD_reduceIndex() :
-+*   rescale all indexes to avoid future overflow (indexes are U32) */
-+static void ZSTD_reduceIndex(ZSTD_CCtx *zc, const U32 reducerValue)
-+{
-+	{
-+		U32 const hSize = 1 << zc->params.cParams.hashLog;
-+		ZSTD_reduceTable(zc->hashTable, hSize, reducerValue);
-+	}
-+
-+	{
-+		U32 const chainSize = (zc->params.cParams.strategy == ZSTD_fast) ? 0 : (1 << zc->params.cParams.chainLog);
-+		ZSTD_reduceTable(zc->chainTable, chainSize, reducerValue);
-+	}
-+
-+	{
-+		U32 const h3Size = (zc->hashLog3) ? 1 << zc->hashLog3 : 0;
-+		ZSTD_reduceTable(zc->hashTable3, h3Size, reducerValue);
-+	}
-+}
-+
-+/*-*******************************************************
-+*  Block entropic compression
-+*********************************************************/
-+
-+/* See doc/zstd_compression_format.md for detailed format description */
-+
-+size_t ZSTD_noCompressBlock(void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-+{
-+	if (srcSize + ZSTD_blockHeaderSize > dstCapacity)
-+		return ERROR(dstSize_tooSmall);
-+	memcpy((BYTE *)dst + ZSTD_blockHeaderSize, src, srcSize);
-+	ZSTD_writeLE24(dst, (U32)(srcSize << 2) + (U32)bt_raw);
-+	return ZSTD_blockHeaderSize + srcSize;
-+}
-+
-+static size_t ZSTD_noCompressLiterals(void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-+{
-+	BYTE *const ostart = (BYTE * const)dst;
-+	U32 const flSize = 1 + (srcSize > 31) + (srcSize > 4095);
-+
-+	if (srcSize + flSize > dstCapacity)
-+		return ERROR(dstSize_tooSmall);
-+
-+	switch (flSize) {
-+	case 1: /* 2 - 1 - 5 */ ostart[0] = (BYTE)((U32)set_basic + (srcSize << 3)); break;
-+	case 2: /* 2 - 2 - 12 */ ZSTD_writeLE16(ostart, (U16)((U32)set_basic + (1 << 2) + (srcSize << 4))); break;
-+	default: /*note : should not be necessary : flSize is within {1,2,3} */
-+	case 3: /* 2 - 2 - 20 */ ZSTD_writeLE32(ostart, (U32)((U32)set_basic + (3 << 2) + (srcSize << 4))); break;
-+	}
-+
-+	memcpy(ostart + flSize, src, srcSize);
-+	return srcSize + flSize;
-+}
-+
-+static size_t ZSTD_compressRleLiteralsBlock(void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-+{
-+	BYTE *const ostart = (BYTE * const)dst;
-+	U32 const flSize = 1 + (srcSize > 31) + (srcSize > 4095);
-+
-+	(void)dstCapacity; /* dstCapacity already guaranteed to be >=4, hence large enough */
-+
-+	switch (flSize) {
-+	case 1: /* 2 - 1 - 5 */ ostart[0] = (BYTE)((U32)set_rle + (srcSize << 3)); break;
-+	case 2: /* 2 - 2 - 12 */ ZSTD_writeLE16(ostart, (U16)((U32)set_rle + (1 << 2) + (srcSize << 4))); break;
-+	default: /*note : should not be necessary : flSize is necessarily within {1,2,3} */
-+	case 3: /* 2 - 2 - 20 */ ZSTD_writeLE32(ostart, (U32)((U32)set_rle + (3 << 2) + (srcSize << 4))); break;
-+	}
-+
-+	ostart[flSize] = *(const BYTE *)src;
-+	return flSize + 1;
-+}
-+
-+static size_t ZSTD_minGain(size_t srcSize) { return (srcSize >> 6) + 2; }
-+
-+static size_t ZSTD_compressLiterals(ZSTD_CCtx *zc, void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-+{
-+	size_t const minGain = ZSTD_minGain(srcSize);
-+	size_t const lhSize = 3 + (srcSize >= 1 KB) + (srcSize >= 16 KB);
-+	BYTE *const ostart = (BYTE *)dst;
-+	U32 singleStream = srcSize < 256;
-+	symbolEncodingType_e hType = set_compressed;
-+	size_t cLitSize;
-+
-+/* small ? don't even attempt compression (speed opt) */
-+#define LITERAL_NOENTROPY 63
-+	{
-+		size_t const minLitSize = zc->flagStaticHufTable == HUF_repeat_valid ? 6 : LITERAL_NOENTROPY;
-+		if (srcSize <= minLitSize)
-+			return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
-+	}
-+
-+	if (dstCapacity < lhSize + 1)
-+		return ERROR(dstSize_tooSmall); /* not enough space for compression */
-+	{
-+		HUF_repeat repeat = zc->flagStaticHufTable;
-+		int const preferRepeat = zc->params.cParams.strategy < ZSTD_lazy ? srcSize <= 1024 : 0;
-+		if (repeat == HUF_repeat_valid && lhSize == 3)
-+			singleStream = 1;
-+		cLitSize = singleStream ? HUF_compress1X_repeat(ostart + lhSize, dstCapacity - lhSize, src, srcSize, 255, 11, zc->tmpCounters,
-+								sizeof(zc->tmpCounters), zc->hufTable, &repeat, preferRepeat)
-+					: HUF_compress4X_repeat(ostart + lhSize, dstCapacity - lhSize, src, srcSize, 255, 11, zc->tmpCounters,
-+								sizeof(zc->tmpCounters), zc->hufTable, &repeat, preferRepeat);
-+		if (repeat != HUF_repeat_none) {
-+			hType = set_repeat;
-+		} /* reused the existing table */
-+		else {
-+			zc->flagStaticHufTable = HUF_repeat_check;
-+		} /* now have a table to reuse */
-+	}
-+
-+	if ((cLitSize == 0) | (cLitSize >= srcSize - minGain)) {
-+		zc->flagStaticHufTable = HUF_repeat_none;
-+		return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
-+	}
-+	if (cLitSize == 1) {
-+		zc->flagStaticHufTable = HUF_repeat_none;
-+		return ZSTD_compressRleLiteralsBlock(dst, dstCapacity, src, srcSize);
-+	}
-+
-+	/* Build header */
-+	switch (lhSize) {
-+	case 3: /* 2 - 2 - 10 - 10 */
-+	{
-+		U32 const lhc = hType + ((!singleStream) << 2) + ((U32)srcSize << 4) + ((U32)cLitSize << 14);
-+		ZSTD_writeLE24(ostart, lhc);
-+		break;
-+	}
-+	case 4: /* 2 - 2 - 14 - 14 */
-+	{
-+		U32 const lhc = hType + (2 << 2) + ((U32)srcSize << 4) + ((U32)cLitSize << 18);
-+		ZSTD_writeLE32(ostart, lhc);
-+		break;
-+	}
-+	default: /* should not be necessary, lhSize is only {3,4,5} */
-+	case 5:  /* 2 - 2 - 18 - 18 */
-+	{
-+		U32 const lhc = hType + (3 << 2) + ((U32)srcSize << 4) + ((U32)cLitSize << 22);
-+		ZSTD_writeLE32(ostart, lhc);
-+		ostart[4] = (BYTE)(cLitSize >> 10);
-+		break;
-+	}
-+	}
-+	return lhSize + cLitSize;
-+}
-+
-+static const BYTE LL_Code[64] = {0,  1,  2,  3,  4,  5,  6,  7,  8,  9,  10, 11, 12, 13, 14, 15, 16, 16, 17, 17, 18, 18,
-+				 19, 19, 20, 20, 20, 20, 21, 21, 21, 21, 22, 22, 22, 22, 22, 22, 22, 22, 23, 23, 23, 23,
-+				 23, 23, 23, 23, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24};
-+
-+static const BYTE ML_Code[128] = {0,  1,  2,  3,  4,  5,  6,  7,  8,  9,  10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25,
-+				  26, 27, 28, 29, 30, 31, 32, 32, 33, 33, 34, 34, 35, 35, 36, 36, 36, 36, 37, 37, 37, 37, 38, 38, 38, 38,
-+				  38, 38, 38, 38, 39, 39, 39, 39, 39, 39, 39, 39, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40, 40,
-+				  40, 40, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 42, 42, 42, 42, 42, 42, 42, 42,
-+				  42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42};
-+
-+void ZSTD_seqToCodes(const seqStore_t *seqStorePtr)
-+{
-+	BYTE const LL_deltaCode = 19;
-+	BYTE const ML_deltaCode = 36;
-+	const seqDef *const sequences = seqStorePtr->sequencesStart;
-+	BYTE *const llCodeTable = seqStorePtr->llCode;
-+	BYTE *const ofCodeTable = seqStorePtr->ofCode;
-+	BYTE *const mlCodeTable = seqStorePtr->mlCode;
-+	U32 const nbSeq = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
-+	U32 u;
-+	for (u = 0; u < nbSeq; u++) {
-+		U32 const llv = sequences[u].litLength;
-+		U32 const mlv = sequences[u].matchLength;
-+		llCodeTable[u] = (llv > 63) ? (BYTE)ZSTD_highbit32(llv) + LL_deltaCode : LL_Code[llv];
-+		ofCodeTable[u] = (BYTE)ZSTD_highbit32(sequences[u].offset);
-+		mlCodeTable[u] = (mlv > 127) ? (BYTE)ZSTD_highbit32(mlv) + ML_deltaCode : ML_Code[mlv];
-+	}
-+	if (seqStorePtr->longLengthID == 1)
-+		llCodeTable[seqStorePtr->longLengthPos] = MaxLL;
-+	if (seqStorePtr->longLengthID == 2)
-+		mlCodeTable[seqStorePtr->longLengthPos] = MaxML;
-+}
-+
-+ZSTD_STATIC size_t ZSTD_compressSequences_internal(ZSTD_CCtx *zc, void *dst, size_t dstCapacity)
-+{
-+	const int longOffsets = zc->params.cParams.windowLog > STREAM_ACCUMULATOR_MIN;
-+	const seqStore_t *seqStorePtr = &(zc->seqStore);
-+	FSE_CTable *CTable_LitLength = zc->litlengthCTable;
-+	FSE_CTable *CTable_OffsetBits = zc->offcodeCTable;
-+	FSE_CTable *CTable_MatchLength = zc->matchlengthCTable;
-+	U32 LLtype, Offtype, MLtype; /* compressed, raw or rle */
-+	const seqDef *const sequences = seqStorePtr->sequencesStart;
-+	const BYTE *const ofCodeTable = seqStorePtr->ofCode;
-+	const BYTE *const llCodeTable = seqStorePtr->llCode;
-+	const BYTE *const mlCodeTable = seqStorePtr->mlCode;
-+	BYTE *const ostart = (BYTE *)dst;
-+	BYTE *const oend = ostart + dstCapacity;
-+	BYTE *op = ostart;
-+	size_t const nbSeq = seqStorePtr->sequences - seqStorePtr->sequencesStart;
-+	BYTE *seqHead;
-+
-+	U32 *count;
-+	S16 *norm;
-+	U32 *workspace;
-+	size_t workspaceSize = sizeof(zc->tmpCounters);
-+	{
-+		size_t spaceUsed32 = 0;
-+		count = (U32 *)zc->tmpCounters + spaceUsed32;
-+		spaceUsed32 += MaxSeq + 1;
-+		norm = (S16 *)((U32 *)zc->tmpCounters + spaceUsed32);
-+		spaceUsed32 += ALIGN(sizeof(S16) * (MaxSeq + 1), sizeof(U32)) >> 2;
-+
-+		workspace = (U32 *)zc->tmpCounters + spaceUsed32;
-+		workspaceSize -= (spaceUsed32 << 2);
-+	}
-+
-+	/* Compress literals */
-+	{
-+		const BYTE *const literals = seqStorePtr->litStart;
-+		size_t const litSize = seqStorePtr->lit - literals;
-+		size_t const cSize = ZSTD_compressLiterals(zc, op, dstCapacity, literals, litSize);
-+		if (ZSTD_isError(cSize))
-+			return cSize;
-+		op += cSize;
-+	}
-+
-+	/* Sequences Header */
-+	if ((oend - op) < 3 /*max nbSeq Size*/ + 1 /*seqHead */)
-+		return ERROR(dstSize_tooSmall);
-+	if (nbSeq < 0x7F)
-+		*op++ = (BYTE)nbSeq;
-+	else if (nbSeq < LONGNBSEQ)
-+		op[0] = (BYTE)((nbSeq >> 8) + 0x80), op[1] = (BYTE)nbSeq, op += 2;
-+	else
-+		op[0] = 0xFF, ZSTD_writeLE16(op + 1, (U16)(nbSeq - LONGNBSEQ)), op += 3;
-+	if (nbSeq == 0)
-+		return op - ostart;
-+
-+	/* seqHead : flags for FSE encoding type */
-+	seqHead = op++;
-+
-+#define MIN_SEQ_FOR_DYNAMIC_FSE 64
-+#define MAX_SEQ_FOR_STATIC_FSE 1000
-+
-+	/* convert length/distances into codes */
-+	ZSTD_seqToCodes(seqStorePtr);
-+
-+	/* CTable for Literal Lengths */
-+	{
-+		U32 max = MaxLL;
-+		size_t const mostFrequent = FSE_countFast_wksp(count, &max, llCodeTable, nbSeq, workspace);
-+		if ((mostFrequent == nbSeq) && (nbSeq > 2)) {
-+			*op++ = llCodeTable[0];
-+			FSE_buildCTable_rle(CTable_LitLength, (BYTE)max);
-+			LLtype = set_rle;
-+		} else if ((zc->flagStaticTables) && (nbSeq < MAX_SEQ_FOR_STATIC_FSE)) {
-+			LLtype = set_repeat;
-+		} else if ((nbSeq < MIN_SEQ_FOR_DYNAMIC_FSE) || (mostFrequent < (nbSeq >> (LL_defaultNormLog - 1)))) {
-+			FSE_buildCTable_wksp(CTable_LitLength, LL_defaultNorm, MaxLL, LL_defaultNormLog, workspace, workspaceSize);
-+			LLtype = set_basic;
-+		} else {
-+			size_t nbSeq_1 = nbSeq;
-+			const U32 tableLog = FSE_optimalTableLog(LLFSELog, nbSeq, max);
-+			if (count[llCodeTable[nbSeq - 1]] > 1) {
-+				count[llCodeTable[nbSeq - 1]]--;
-+				nbSeq_1--;
-+			}
-+			FSE_normalizeCount(norm, tableLog, count, nbSeq_1, max);
-+			{
-+				size_t const NCountSize = FSE_writeNCount(op, oend - op, norm, max, tableLog); /* overflow protected */
-+				if (FSE_isError(NCountSize))
-+					return NCountSize;
-+				op += NCountSize;
-+			}
-+			FSE_buildCTable_wksp(CTable_LitLength, norm, max, tableLog, workspace, workspaceSize);
-+			LLtype = set_compressed;
-+		}
-+	}
-+
-+	/* CTable for Offsets */
-+	{
-+		U32 max = MaxOff;
-+		size_t const mostFrequent = FSE_countFast_wksp(count, &max, ofCodeTable, nbSeq, workspace);
-+		if ((mostFrequent == nbSeq) && (nbSeq > 2)) {
-+			*op++ = ofCodeTable[0];
-+			FSE_buildCTable_rle(CTable_OffsetBits, (BYTE)max);
-+			Offtype = set_rle;
-+		} else if ((zc->flagStaticTables) && (nbSeq < MAX_SEQ_FOR_STATIC_FSE)) {
-+			Offtype = set_repeat;
-+		} else if ((nbSeq < MIN_SEQ_FOR_DYNAMIC_FSE) || (mostFrequent < (nbSeq >> (OF_defaultNormLog - 1)))) {
-+			FSE_buildCTable_wksp(CTable_OffsetBits, OF_defaultNorm, MaxOff, OF_defaultNormLog, workspace, workspaceSize);
-+			Offtype = set_basic;
-+		} else {
-+			size_t nbSeq_1 = nbSeq;
-+			const U32 tableLog = FSE_optimalTableLog(OffFSELog, nbSeq, max);
-+			if (count[ofCodeTable[nbSeq - 1]] > 1) {
-+				count[ofCodeTable[nbSeq - 1]]--;
-+				nbSeq_1--;
-+			}
-+			FSE_normalizeCount(norm, tableLog, count, nbSeq_1, max);
-+			{
-+				size_t const NCountSize = FSE_writeNCount(op, oend - op, norm, max, tableLog); /* overflow protected */
-+				if (FSE_isError(NCountSize))
-+					return NCountSize;
-+				op += NCountSize;
-+			}
-+			FSE_buildCTable_wksp(CTable_OffsetBits, norm, max, tableLog, workspace, workspaceSize);
-+			Offtype = set_compressed;
-+		}
-+	}
-+
-+	/* CTable for MatchLengths */
-+	{
-+		U32 max = MaxML;
-+		size_t const mostFrequent = FSE_countFast_wksp(count, &max, mlCodeTable, nbSeq, workspace);
-+		if ((mostFrequent == nbSeq) && (nbSeq > 2)) {
-+			*op++ = *mlCodeTable;
-+			FSE_buildCTable_rle(CTable_MatchLength, (BYTE)max);
-+			MLtype = set_rle;
-+		} else if ((zc->flagStaticTables) && (nbSeq < MAX_SEQ_FOR_STATIC_FSE)) {
-+			MLtype = set_repeat;
-+		} else if ((nbSeq < MIN_SEQ_FOR_DYNAMIC_FSE) || (mostFrequent < (nbSeq >> (ML_defaultNormLog - 1)))) {
-+			FSE_buildCTable_wksp(CTable_MatchLength, ML_defaultNorm, MaxML, ML_defaultNormLog, workspace, workspaceSize);
-+			MLtype = set_basic;
-+		} else {
-+			size_t nbSeq_1 = nbSeq;
-+			const U32 tableLog = FSE_optimalTableLog(MLFSELog, nbSeq, max);
-+			if (count[mlCodeTable[nbSeq - 1]] > 1) {
-+				count[mlCodeTable[nbSeq - 1]]--;
-+				nbSeq_1--;
-+			}
-+			FSE_normalizeCount(norm, tableLog, count, nbSeq_1, max);
-+			{
-+				size_t const NCountSize = FSE_writeNCount(op, oend - op, norm, max, tableLog); /* overflow protected */
-+				if (FSE_isError(NCountSize))
-+					return NCountSize;
-+				op += NCountSize;
-+			}
-+			FSE_buildCTable_wksp(CTable_MatchLength, norm, max, tableLog, workspace, workspaceSize);
-+			MLtype = set_compressed;
-+		}
-+	}
-+
-+	*seqHead = (BYTE)((LLtype << 6) + (Offtype << 4) + (MLtype << 2));
-+	zc->flagStaticTables = 0;
-+
-+	/* Encoding Sequences */
-+	{
-+		BIT_CStream_t blockStream;
-+		FSE_CState_t stateMatchLength;
-+		FSE_CState_t stateOffsetBits;
-+		FSE_CState_t stateLitLength;
-+
-+		CHECK_E(BIT_initCStream(&blockStream, op, oend - op), dstSize_tooSmall); /* not enough space remaining */
-+
-+		/* first symbols */
-+		FSE_initCState2(&stateMatchLength, CTable_MatchLength, mlCodeTable[nbSeq - 1]);
-+		FSE_initCState2(&stateOffsetBits, CTable_OffsetBits, ofCodeTable[nbSeq - 1]);
-+		FSE_initCState2(&stateLitLength, CTable_LitLength, llCodeTable[nbSeq - 1]);
-+		BIT_addBits(&blockStream, sequences[nbSeq - 1].litLength, LL_bits[llCodeTable[nbSeq - 1]]);
-+		if (ZSTD_32bits())
-+			BIT_flushBits(&blockStream);
-+		BIT_addBits(&blockStream, sequences[nbSeq - 1].matchLength, ML_bits[mlCodeTable[nbSeq - 1]]);
-+		if (ZSTD_32bits())
-+			BIT_flushBits(&blockStream);
-+		if (longOffsets) {
-+			U32 const ofBits = ofCodeTable[nbSeq - 1];
-+			int const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN - 1);
-+			if (extraBits) {
-+				BIT_addBits(&blockStream, sequences[nbSeq - 1].offset, extraBits);
-+				BIT_flushBits(&blockStream);
-+			}
-+			BIT_addBits(&blockStream, sequences[nbSeq - 1].offset >> extraBits, ofBits - extraBits);
-+		} else {
-+			BIT_addBits(&blockStream, sequences[nbSeq - 1].offset, ofCodeTable[nbSeq - 1]);
-+		}
-+		BIT_flushBits(&blockStream);
-+
-+		{
-+			size_t n;
-+			for (n = nbSeq - 2; n < nbSeq; n--) { /* intentional underflow */
-+				BYTE const llCode = llCodeTable[n];
-+				BYTE const ofCode = ofCodeTable[n];
-+				BYTE const mlCode = mlCodeTable[n];
-+				U32 const llBits = LL_bits[llCode];
-+				U32 const ofBits = ofCode; /* 32b*/ /* 64b*/
-+				U32 const mlBits = ML_bits[mlCode];
-+				/* (7)*/							    /* (7)*/
-+				FSE_encodeSymbol(&blockStream, &stateOffsetBits, ofCode); /* 15 */  /* 15 */
-+				FSE_encodeSymbol(&blockStream, &stateMatchLength, mlCode); /* 24 */ /* 24 */
-+				if (ZSTD_32bits())
-+					BIT_flushBits(&blockStream);				  /* (7)*/
-+				FSE_encodeSymbol(&blockStream, &stateLitLength, llCode); /* 16 */ /* 33 */
-+				if (ZSTD_32bits() || (ofBits + mlBits + llBits >= 64 - 7 - (LLFSELog + MLFSELog + OffFSELog)))
-+					BIT_flushBits(&blockStream); /* (7)*/
-+				BIT_addBits(&blockStream, sequences[n].litLength, llBits);
-+				if (ZSTD_32bits() && ((llBits + mlBits) > 24))
-+					BIT_flushBits(&blockStream);
-+				BIT_addBits(&blockStream, sequences[n].matchLength, mlBits);
-+				if (ZSTD_32bits())
-+					BIT_flushBits(&blockStream); /* (7)*/
-+				if (longOffsets) {
-+					int const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN - 1);
-+					if (extraBits) {
-+						BIT_addBits(&blockStream, sequences[n].offset, extraBits);
-+						BIT_flushBits(&blockStream); /* (7)*/
-+					}
-+					BIT_addBits(&blockStream, sequences[n].offset >> extraBits, ofBits - extraBits); /* 31 */
-+				} else {
-+					BIT_addBits(&blockStream, sequences[n].offset, ofBits); /* 31 */
-+				}
-+				BIT_flushBits(&blockStream); /* (7)*/
-+			}
-+		}
-+
-+		FSE_flushCState(&blockStream, &stateMatchLength);
-+		FSE_flushCState(&blockStream, &stateOffsetBits);
-+		FSE_flushCState(&blockStream, &stateLitLength);
-+
-+		{
-+			size_t const streamSize = BIT_closeCStream(&blockStream);
-+			if (streamSize == 0)
-+				return ERROR(dstSize_tooSmall); /* not enough space */
-+			op += streamSize;
-+		}
-+	}
-+	return op - ostart;
-+}
-+
-+ZSTD_STATIC size_t ZSTD_compressSequences(ZSTD_CCtx *zc, void *dst, size_t dstCapacity, size_t srcSize)
-+{
-+	size_t const cSize = ZSTD_compressSequences_internal(zc, dst, dstCapacity);
-+	size_t const minGain = ZSTD_minGain(srcSize);
-+	size_t const maxCSize = srcSize - minGain;
-+	/* If the srcSize <= dstCapacity, then there is enough space to write a
-+	 * raw uncompressed block. Since we ran out of space, the block must not
-+	 * be compressible, so fall back to a raw uncompressed block.
-+	 */
-+	int const uncompressibleError = cSize == ERROR(dstSize_tooSmall) && srcSize <= dstCapacity;
-+	int i;
-+
-+	if (ZSTD_isError(cSize) && !uncompressibleError)
-+		return cSize;
-+	if (cSize >= maxCSize || uncompressibleError) {
-+		zc->flagStaticHufTable = HUF_repeat_none;
-+		return 0;
-+	}
-+	/* confirm repcodes */
-+	for (i = 0; i < ZSTD_REP_NUM; i++)
-+		zc->rep[i] = zc->repToConfirm[i];
-+	return cSize;
-+}
-+
-+/*! ZSTD_storeSeq() :
-+	Store a sequence (literal length, literals, offset code and match length code) into seqStore_t.
-+	`offsetCode` : distance to match, or 0 == repCode.
-+	`matchCode` : matchLength - MINMATCH
-+*/
-+ZSTD_STATIC void ZSTD_storeSeq(seqStore_t *seqStorePtr, size_t litLength, const void *literals, U32 offsetCode, size_t matchCode)
-+{
-+	/* copy Literals */
-+	ZSTD_wildcopy(seqStorePtr->lit, literals, litLength);
-+	seqStorePtr->lit += litLength;
-+
-+	/* literal Length */
-+	if (litLength > 0xFFFF) {
-+		seqStorePtr->longLengthID = 1;
-+		seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
-+	}
-+	seqStorePtr->sequences[0].litLength = (U16)litLength;
-+
-+	/* match offset */
-+	seqStorePtr->sequences[0].offset = offsetCode + 1;
-+
-+	/* match Length */
-+	if (matchCode > 0xFFFF) {
-+		seqStorePtr->longLengthID = 2;
-+		seqStorePtr->longLengthPos = (U32)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
-+	}
-+	seqStorePtr->sequences[0].matchLength = (U16)matchCode;
-+
-+	seqStorePtr->sequences++;
-+}
-+
-+/*-*************************************
-+*  Match length counter
-+***************************************/
-+static unsigned ZSTD_NbCommonBytes(register size_t val)
-+{
-+	if (ZSTD_isLittleEndian()) {
-+		if (ZSTD_64bits()) {
-+			return (__builtin_ctzll((U64)val) >> 3);
-+		} else { /* 32 bits */
-+			return (__builtin_ctz((U32)val) >> 3);
-+		}
-+	} else { /* Big Endian CPU */
-+		if (ZSTD_64bits()) {
-+			return (__builtin_clzll(val) >> 3);
-+		} else { /* 32 bits */
-+			return (__builtin_clz((U32)val) >> 3);
-+		}
-+	}
-+}
-+
-+static size_t ZSTD_count(const BYTE *pIn, const BYTE *pMatch, const BYTE *const pInLimit)
-+{
-+	const BYTE *const pStart = pIn;
-+	const BYTE *const pInLoopLimit = pInLimit - (sizeof(size_t) - 1);
-+
-+	while (pIn < pInLoopLimit) {
-+		size_t const diff = ZSTD_readST(pMatch) ^ ZSTD_readST(pIn);
-+		if (!diff) {
-+			pIn += sizeof(size_t);
-+			pMatch += sizeof(size_t);
-+			continue;
-+		}
-+		pIn += ZSTD_NbCommonBytes(diff);
-+		return (size_t)(pIn - pStart);
-+	}
-+	if (ZSTD_64bits())
-+		if ((pIn < (pInLimit - 3)) && (ZSTD_read32(pMatch) == ZSTD_read32(pIn))) {
-+			pIn += 4;
-+			pMatch += 4;
-+		}
-+	if ((pIn < (pInLimit - 1)) && (ZSTD_read16(pMatch) == ZSTD_read16(pIn))) {
-+		pIn += 2;
-+		pMatch += 2;
-+	}
-+	if ((pIn < pInLimit) && (*pMatch == *pIn))
-+		pIn++;
-+	return (size_t)(pIn - pStart);
-+}
-+
-+/** ZSTD_count_2segments() :
-+*   can count match length with `ip` & `match` in 2 different segments.
-+*   convention : on reaching mEnd, match count continue starting from iStart
-+*/
-+static size_t ZSTD_count_2segments(const BYTE *ip, const BYTE *match, const BYTE *iEnd, const BYTE *mEnd, const BYTE *iStart)
-+{
-+	const BYTE *const vEnd = MIN(ip + (mEnd - match), iEnd);
-+	size_t const matchLength = ZSTD_count(ip, match, vEnd);
-+	if (match + matchLength != mEnd)
-+		return matchLength;
-+	return matchLength + ZSTD_count(ip + matchLength, iStart, iEnd);
-+}
-+
-+/*-*************************************
-+*  Hashes
-+***************************************/
-+static const U32 prime3bytes = 506832829U;
-+static U32 ZSTD_hash3(U32 u, U32 h) { return ((u << (32 - 24)) * prime3bytes) >> (32 - h); }
-+ZSTD_STATIC size_t ZSTD_hash3Ptr(const void *ptr, U32 h) { return ZSTD_hash3(ZSTD_readLE32(ptr), h); } /* only in zstd_opt.h */
-+
-+static const U32 prime4bytes = 2654435761U;
-+static U32 ZSTD_hash4(U32 u, U32 h) { return (u * prime4bytes) >> (32 - h); }
-+static size_t ZSTD_hash4Ptr(const void *ptr, U32 h) { return ZSTD_hash4(ZSTD_read32(ptr), h); }
-+
-+static const U64 prime5bytes = 889523592379ULL;
-+static size_t ZSTD_hash5(U64 u, U32 h) { return (size_t)(((u << (64 - 40)) * prime5bytes) >> (64 - h)); }
-+static size_t ZSTD_hash5Ptr(const void *p, U32 h) { return ZSTD_hash5(ZSTD_readLE64(p), h); }
-+
-+static const U64 prime6bytes = 227718039650203ULL;
-+static size_t ZSTD_hash6(U64 u, U32 h) { return (size_t)(((u << (64 - 48)) * prime6bytes) >> (64 - h)); }
-+static size_t ZSTD_hash6Ptr(const void *p, U32 h) { return ZSTD_hash6(ZSTD_readLE64(p), h); }
-+
-+static const U64 prime7bytes = 58295818150454627ULL;
-+static size_t ZSTD_hash7(U64 u, U32 h) { return (size_t)(((u << (64 - 56)) * prime7bytes) >> (64 - h)); }
-+static size_t ZSTD_hash7Ptr(const void *p, U32 h) { return ZSTD_hash7(ZSTD_readLE64(p), h); }
-+
-+static const U64 prime8bytes = 0xCF1BBCDCB7A56463ULL;
-+static size_t ZSTD_hash8(U64 u, U32 h) { return (size_t)(((u)*prime8bytes) >> (64 - h)); }
-+static size_t ZSTD_hash8Ptr(const void *p, U32 h) { return ZSTD_hash8(ZSTD_readLE64(p), h); }
-+
-+static size_t ZSTD_hashPtr(const void *p, U32 hBits, U32 mls)
-+{
-+	switch (mls) {
-+	// case 3: return ZSTD_hash3Ptr(p, hBits);
-+	default:
-+	case 4: return ZSTD_hash4Ptr(p, hBits);
-+	case 5: return ZSTD_hash5Ptr(p, hBits);
-+	case 6: return ZSTD_hash6Ptr(p, hBits);
-+	case 7: return ZSTD_hash7Ptr(p, hBits);
-+	case 8: return ZSTD_hash8Ptr(p, hBits);
-+	}
-+}
-+
-+/*-*************************************
-+*  Fast Scan
-+***************************************/
-+static void ZSTD_fillHashTable(ZSTD_CCtx *zc, const void *end, const U32 mls)
-+{
-+	U32 *const hashTable = zc->hashTable;
-+	U32 const hBits = zc->params.cParams.hashLog;
-+	const BYTE *const base = zc->base;
-+	const BYTE *ip = base + zc->nextToUpdate;
-+	const BYTE *const iend = ((const BYTE *)end) - HASH_READ_SIZE;
-+	const size_t fastHashFillStep = 3;
-+
-+	while (ip <= iend) {
-+		hashTable[ZSTD_hashPtr(ip, hBits, mls)] = (U32)(ip - base);
-+		ip += fastHashFillStep;
-+	}
-+}
-+
-+FORCE_INLINE
-+void ZSTD_compressBlock_fast_generic(ZSTD_CCtx *cctx, const void *src, size_t srcSize, const U32 mls)
-+{
-+	U32 *const hashTable = cctx->hashTable;
-+	U32 const hBits = cctx->params.cParams.hashLog;
-+	seqStore_t *seqStorePtr = &(cctx->seqStore);
-+	const BYTE *const base = cctx->base;
-+	const BYTE *const istart = (const BYTE *)src;
-+	const BYTE *ip = istart;
-+	const BYTE *anchor = istart;
-+	const U32 lowestIndex = cctx->dictLimit;
-+	const BYTE *const lowest = base + lowestIndex;
-+	const BYTE *const iend = istart + srcSize;
-+	const BYTE *const ilimit = iend - HASH_READ_SIZE;
-+	U32 offset_1 = cctx->rep[0], offset_2 = cctx->rep[1];
-+	U32 offsetSaved = 0;
-+
-+	/* init */
-+	ip += (ip == lowest);
-+	{
-+		U32 const maxRep = (U32)(ip - lowest);
-+		if (offset_2 > maxRep)
-+			offsetSaved = offset_2, offset_2 = 0;
-+		if (offset_1 > maxRep)
-+			offsetSaved = offset_1, offset_1 = 0;
-+	}
-+
-+	/* Main Search Loop */
-+	while (ip < ilimit) { /* < instead of <=, because repcode check at (ip+1) */
-+		size_t mLength;
-+		size_t const h = ZSTD_hashPtr(ip, hBits, mls);
-+		U32 const curr = (U32)(ip - base);
-+		U32 const matchIndex = hashTable[h];
-+		const BYTE *match = base + matchIndex;
-+		hashTable[h] = curr; /* update hash table */
-+
-+		if ((offset_1 > 0) & (ZSTD_read32(ip + 1 - offset_1) == ZSTD_read32(ip + 1))) {
-+			mLength = ZSTD_count(ip + 1 + 4, ip + 1 + 4 - offset_1, iend) + 4;
-+			ip++;
-+			ZSTD_storeSeq(seqStorePtr, ip - anchor, anchor, 0, mLength - MINMATCH);
-+		} else {
-+			U32 offset;
-+			if ((matchIndex <= lowestIndex) || (ZSTD_read32(match) != ZSTD_read32(ip))) {
-+				ip += ((ip - anchor) >> g_searchStrength) + 1;
-+				continue;
-+			}
-+			mLength = ZSTD_count(ip + 4, match + 4, iend) + 4;
-+			offset = (U32)(ip - match);
-+			while (((ip > anchor) & (match > lowest)) && (ip[-1] == match[-1])) {
-+				ip--;
-+				match--;
-+				mLength++;
-+			} /* catch up */
-+			offset_2 = offset_1;
-+			offset_1 = offset;
-+
-+			ZSTD_storeSeq(seqStorePtr, ip - anchor, anchor, offset + ZSTD_REP_MOVE, mLength - MINMATCH);
-+		}
-+
-+		/* match found */
-+		ip += mLength;
-+		anchor = ip;
-+
-+		if (ip <= ilimit) {
-+			/* Fill Table */
-+			hashTable[ZSTD_hashPtr(base + curr + 2, hBits, mls)] = curr + 2; /* here because curr+2 could be > iend-8 */
-+			hashTable[ZSTD_hashPtr(ip - 2, hBits, mls)] = (U32)(ip - 2 - base);
-+			/* check immediate repcode */
-+			while ((ip <= ilimit) && ((offset_2 > 0) & (ZSTD_read32(ip) == ZSTD_read32(ip - offset_2)))) {
-+				/* store sequence */
-+				size_t const rLength = ZSTD_count(ip + 4, ip + 4 - offset_2, iend) + 4;
-+				{
-+					U32 const tmpOff = offset_2;
-+					offset_2 = offset_1;
-+					offset_1 = tmpOff;
-+				} /* swap offset_2 <=> offset_1 */
-+				hashTable[ZSTD_hashPtr(ip, hBits, mls)] = (U32)(ip - base);
-+				ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, rLength - MINMATCH);
-+				ip += rLength;
-+				anchor = ip;
-+				continue; /* faster when present ... (?) */
-+			}
-+		}
-+	}
-+
-+	/* save reps for next block */
-+	cctx->repToConfirm[0] = offset_1 ? offset_1 : offsetSaved;
-+	cctx->repToConfirm[1] = offset_2 ? offset_2 : offsetSaved;
-+
-+	/* Last Literals */
-+	{
-+		size_t const lastLLSize = iend - anchor;
-+		memcpy(seqStorePtr->lit, anchor, lastLLSize);
-+		seqStorePtr->lit += lastLLSize;
-+	}
-+}
-+
-+static void ZSTD_compressBlock_fast(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-+{
-+	const U32 mls = ctx->params.cParams.searchLength;
-+	switch (mls) {
-+	default: /* includes case 3 */
-+	case 4: ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 4); return;
-+	case 5: ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 5); return;
-+	case 6: ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 6); return;
-+	case 7: ZSTD_compressBlock_fast_generic(ctx, src, srcSize, 7); return;
-+	}
-+}
-+
-+static void ZSTD_compressBlock_fast_extDict_generic(ZSTD_CCtx *ctx, const void *src, size_t srcSize, const U32 mls)
-+{
-+	U32 *hashTable = ctx->hashTable;
-+	const U32 hBits = ctx->params.cParams.hashLog;
-+	seqStore_t *seqStorePtr = &(ctx->seqStore);
-+	const BYTE *const base = ctx->base;
-+	const BYTE *const dictBase = ctx->dictBase;
-+	const BYTE *const istart = (const BYTE *)src;
-+	const BYTE *ip = istart;
-+	const BYTE *anchor = istart;
-+	const U32 lowestIndex = ctx->lowLimit;
-+	const BYTE *const dictStart = dictBase + lowestIndex;
-+	const U32 dictLimit = ctx->dictLimit;
-+	const BYTE *const lowPrefixPtr = base + dictLimit;
-+	const BYTE *const dictEnd = dictBase + dictLimit;
-+	const BYTE *const iend = istart + srcSize;
-+	const BYTE *const ilimit = iend - 8;
-+	U32 offset_1 = ctx->rep[0], offset_2 = ctx->rep[1];
-+
-+	/* Search Loop */
-+	while (ip < ilimit) { /* < instead of <=, because (ip+1) */
-+		const size_t h = ZSTD_hashPtr(ip, hBits, mls);
-+		const U32 matchIndex = hashTable[h];
-+		const BYTE *matchBase = matchIndex < dictLimit ? dictBase : base;
-+		const BYTE *match = matchBase + matchIndex;
-+		const U32 curr = (U32)(ip - base);
-+		const U32 repIndex = curr + 1 - offset_1; /* offset_1 expected <= curr +1 */
-+		const BYTE *repBase = repIndex < dictLimit ? dictBase : base;
-+		const BYTE *repMatch = repBase + repIndex;
-+		size_t mLength;
-+		hashTable[h] = curr; /* update hash table */
-+
-+		if ((((U32)((dictLimit - 1) - repIndex) >= 3) /* intentional underflow */ & (repIndex > lowestIndex)) &&
-+		    (ZSTD_read32(repMatch) == ZSTD_read32(ip + 1))) {
-+			const BYTE *repMatchEnd = repIndex < dictLimit ? dictEnd : iend;
-+			mLength = ZSTD_count_2segments(ip + 1 + EQUAL_READ32, repMatch + EQUAL_READ32, iend, repMatchEnd, lowPrefixPtr) + EQUAL_READ32;
-+			ip++;
-+			ZSTD_storeSeq(seqStorePtr, ip - anchor, anchor, 0, mLength - MINMATCH);
-+		} else {
-+			if ((matchIndex < lowestIndex) || (ZSTD_read32(match) != ZSTD_read32(ip))) {
-+				ip += ((ip - anchor) >> g_searchStrength) + 1;
-+				continue;
-+			}
-+			{
-+				const BYTE *matchEnd = matchIndex < dictLimit ? dictEnd : iend;
-+				const BYTE *lowMatchPtr = matchIndex < dictLimit ? dictStart : lowPrefixPtr;
-+				U32 offset;
-+				mLength = ZSTD_count_2segments(ip + EQUAL_READ32, match + EQUAL_READ32, iend, matchEnd, lowPrefixPtr) + EQUAL_READ32;
-+				while (((ip > anchor) & (match > lowMatchPtr)) && (ip[-1] == match[-1])) {
-+					ip--;
-+					match--;
-+					mLength++;
-+				} /* catch up */
-+				offset = curr - matchIndex;
-+				offset_2 = offset_1;
-+				offset_1 = offset;
-+				ZSTD_storeSeq(seqStorePtr, ip - anchor, anchor, offset + ZSTD_REP_MOVE, mLength - MINMATCH);
-+			}
-+		}
-+
-+		/* found a match : store it */
-+		ip += mLength;
-+		anchor = ip;
-+
-+		if (ip <= ilimit) {
-+			/* Fill Table */
-+			hashTable[ZSTD_hashPtr(base + curr + 2, hBits, mls)] = curr + 2;
-+			hashTable[ZSTD_hashPtr(ip - 2, hBits, mls)] = (U32)(ip - 2 - base);
-+			/* check immediate repcode */
-+			while (ip <= ilimit) {
-+				U32 const curr2 = (U32)(ip - base);
-+				U32 const repIndex2 = curr2 - offset_2;
-+				const BYTE *repMatch2 = repIndex2 < dictLimit ? dictBase + repIndex2 : base + repIndex2;
-+				if ((((U32)((dictLimit - 1) - repIndex2) >= 3) & (repIndex2 > lowestIndex)) /* intentional overflow */
-+				    && (ZSTD_read32(repMatch2) == ZSTD_read32(ip))) {
-+					const BYTE *const repEnd2 = repIndex2 < dictLimit ? dictEnd : iend;
-+					size_t repLength2 =
-+					    ZSTD_count_2segments(ip + EQUAL_READ32, repMatch2 + EQUAL_READ32, iend, repEnd2, lowPrefixPtr) + EQUAL_READ32;
-+					U32 tmpOffset = offset_2;
-+					offset_2 = offset_1;
-+					offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */
-+					ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, repLength2 - MINMATCH);
-+					hashTable[ZSTD_hashPtr(ip, hBits, mls)] = curr2;
-+					ip += repLength2;
-+					anchor = ip;
-+					continue;
-+				}
-+				break;
-+			}
-+		}
-+	}
-+
-+	/* save reps for next block */
-+	ctx->repToConfirm[0] = offset_1;
-+	ctx->repToConfirm[1] = offset_2;
-+
-+	/* Last Literals */
-+	{
-+		size_t const lastLLSize = iend - anchor;
-+		memcpy(seqStorePtr->lit, anchor, lastLLSize);
-+		seqStorePtr->lit += lastLLSize;
-+	}
-+}
-+
-+static void ZSTD_compressBlock_fast_extDict(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-+{
-+	U32 const mls = ctx->params.cParams.searchLength;
-+	switch (mls) {
-+	default: /* includes case 3 */
-+	case 4: ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 4); return;
-+	case 5: ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 5); return;
-+	case 6: ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 6); return;
-+	case 7: ZSTD_compressBlock_fast_extDict_generic(ctx, src, srcSize, 7); return;
-+	}
-+}
-+
-+/*-*************************************
-+*  Double Fast
-+***************************************/
-+static void ZSTD_fillDoubleHashTable(ZSTD_CCtx *cctx, const void *end, const U32 mls)
-+{
-+	U32 *const hashLarge = cctx->hashTable;
-+	U32 const hBitsL = cctx->params.cParams.hashLog;
-+	U32 *const hashSmall = cctx->chainTable;
-+	U32 const hBitsS = cctx->params.cParams.chainLog;
-+	const BYTE *const base = cctx->base;
-+	const BYTE *ip = base + cctx->nextToUpdate;
-+	const BYTE *const iend = ((const BYTE *)end) - HASH_READ_SIZE;
-+	const size_t fastHashFillStep = 3;
-+
-+	while (ip <= iend) {
-+		hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = (U32)(ip - base);
-+		hashLarge[ZSTD_hashPtr(ip, hBitsL, 8)] = (U32)(ip - base);
-+		ip += fastHashFillStep;
-+	}
-+}
-+
-+FORCE_INLINE
-+void ZSTD_compressBlock_doubleFast_generic(ZSTD_CCtx *cctx, const void *src, size_t srcSize, const U32 mls)
-+{
-+	U32 *const hashLong = cctx->hashTable;
-+	const U32 hBitsL = cctx->params.cParams.hashLog;
-+	U32 *const hashSmall = cctx->chainTable;
-+	const U32 hBitsS = cctx->params.cParams.chainLog;
-+	seqStore_t *seqStorePtr = &(cctx->seqStore);
-+	const BYTE *const base = cctx->base;
-+	const BYTE *const istart = (const BYTE *)src;
-+	const BYTE *ip = istart;
-+	const BYTE *anchor = istart;
-+	const U32 lowestIndex = cctx->dictLimit;
-+	const BYTE *const lowest = base + lowestIndex;
-+	const BYTE *const iend = istart + srcSize;
-+	const BYTE *const ilimit = iend - HASH_READ_SIZE;
-+	U32 offset_1 = cctx->rep[0], offset_2 = cctx->rep[1];
-+	U32 offsetSaved = 0;
-+
-+	/* init */
-+	ip += (ip == lowest);
-+	{
-+		U32 const maxRep = (U32)(ip - lowest);
-+		if (offset_2 > maxRep)
-+			offsetSaved = offset_2, offset_2 = 0;
-+		if (offset_1 > maxRep)
-+			offsetSaved = offset_1, offset_1 = 0;
-+	}
-+
-+	/* Main Search Loop */
-+	while (ip < ilimit) { /* < instead of <=, because repcode check at (ip+1) */
-+		size_t mLength;
-+		size_t const h2 = ZSTD_hashPtr(ip, hBitsL, 8);
-+		size_t const h = ZSTD_hashPtr(ip, hBitsS, mls);
-+		U32 const curr = (U32)(ip - base);
-+		U32 const matchIndexL = hashLong[h2];
-+		U32 const matchIndexS = hashSmall[h];
-+		const BYTE *matchLong = base + matchIndexL;
-+		const BYTE *match = base + matchIndexS;
-+		hashLong[h2] = hashSmall[h] = curr; /* update hash tables */
-+
-+		if ((offset_1 > 0) & (ZSTD_read32(ip + 1 - offset_1) == ZSTD_read32(ip + 1))) { /* note : by construction, offset_1 <= curr */
-+			mLength = ZSTD_count(ip + 1 + 4, ip + 1 + 4 - offset_1, iend) + 4;
-+			ip++;
-+			ZSTD_storeSeq(seqStorePtr, ip - anchor, anchor, 0, mLength - MINMATCH);
-+		} else {
-+			U32 offset;
-+			if ((matchIndexL > lowestIndex) && (ZSTD_read64(matchLong) == ZSTD_read64(ip))) {
-+				mLength = ZSTD_count(ip + 8, matchLong + 8, iend) + 8;
-+				offset = (U32)(ip - matchLong);
-+				while (((ip > anchor) & (matchLong > lowest)) && (ip[-1] == matchLong[-1])) {
-+					ip--;
-+					matchLong--;
-+					mLength++;
-+				} /* catch up */
-+			} else if ((matchIndexS > lowestIndex) && (ZSTD_read32(match) == ZSTD_read32(ip))) {
-+				size_t const h3 = ZSTD_hashPtr(ip + 1, hBitsL, 8);
-+				U32 const matchIndex3 = hashLong[h3];
-+				const BYTE *match3 = base + matchIndex3;
-+				hashLong[h3] = curr + 1;
-+				if ((matchIndex3 > lowestIndex) && (ZSTD_read64(match3) == ZSTD_read64(ip + 1))) {
-+					mLength = ZSTD_count(ip + 9, match3 + 8, iend) + 8;
-+					ip++;
-+					offset = (U32)(ip - match3);
-+					while (((ip > anchor) & (match3 > lowest)) && (ip[-1] == match3[-1])) {
-+						ip--;
-+						match3--;
-+						mLength++;
-+					} /* catch up */
-+				} else {
-+					mLength = ZSTD_count(ip + 4, match + 4, iend) + 4;
-+					offset = (U32)(ip - match);
-+					while (((ip > anchor) & (match > lowest)) && (ip[-1] == match[-1])) {
-+						ip--;
-+						match--;
-+						mLength++;
-+					} /* catch up */
-+				}
-+			} else {
-+				ip += ((ip - anchor) >> g_searchStrength) + 1;
-+				continue;
-+			}
-+
-+			offset_2 = offset_1;
-+			offset_1 = offset;
-+
-+			ZSTD_storeSeq(seqStorePtr, ip - anchor, anchor, offset + ZSTD_REP_MOVE, mLength - MINMATCH);
-+		}
-+
-+		/* match found */
-+		ip += mLength;
-+		anchor = ip;
-+
-+		if (ip <= ilimit) {
-+			/* Fill Table */
-+			hashLong[ZSTD_hashPtr(base + curr + 2, hBitsL, 8)] = hashSmall[ZSTD_hashPtr(base + curr + 2, hBitsS, mls)] =
-+			    curr + 2; /* here because curr+2 could be > iend-8 */
-+			hashLong[ZSTD_hashPtr(ip - 2, hBitsL, 8)] = hashSmall[ZSTD_hashPtr(ip - 2, hBitsS, mls)] = (U32)(ip - 2 - base);
-+
-+			/* check immediate repcode */
-+			while ((ip <= ilimit) && ((offset_2 > 0) & (ZSTD_read32(ip) == ZSTD_read32(ip - offset_2)))) {
-+				/* store sequence */
-+				size_t const rLength = ZSTD_count(ip + 4, ip + 4 - offset_2, iend) + 4;
-+				{
-+					U32 const tmpOff = offset_2;
-+					offset_2 = offset_1;
-+					offset_1 = tmpOff;
-+				} /* swap offset_2 <=> offset_1 */
-+				hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = (U32)(ip - base);
-+				hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = (U32)(ip - base);
-+				ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, rLength - MINMATCH);
-+				ip += rLength;
-+				anchor = ip;
-+				continue; /* faster when present ... (?) */
-+			}
-+		}
-+	}
-+
-+	/* save reps for next block */
-+	cctx->repToConfirm[0] = offset_1 ? offset_1 : offsetSaved;
-+	cctx->repToConfirm[1] = offset_2 ? offset_2 : offsetSaved;
-+
-+	/* Last Literals */
-+	{
-+		size_t const lastLLSize = iend - anchor;
-+		memcpy(seqStorePtr->lit, anchor, lastLLSize);
-+		seqStorePtr->lit += lastLLSize;
-+	}
-+}
-+
-+static void ZSTD_compressBlock_doubleFast(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-+{
-+	const U32 mls = ctx->params.cParams.searchLength;
-+	switch (mls) {
-+	default: /* includes case 3 */
-+	case 4: ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 4); return;
-+	case 5: ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 5); return;
-+	case 6: ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 6); return;
-+	case 7: ZSTD_compressBlock_doubleFast_generic(ctx, src, srcSize, 7); return;
-+	}
-+}
-+
-+static void ZSTD_compressBlock_doubleFast_extDict_generic(ZSTD_CCtx *ctx, const void *src, size_t srcSize, const U32 mls)
-+{
-+	U32 *const hashLong = ctx->hashTable;
-+	U32 const hBitsL = ctx->params.cParams.hashLog;
-+	U32 *const hashSmall = ctx->chainTable;
-+	U32 const hBitsS = ctx->params.cParams.chainLog;
-+	seqStore_t *seqStorePtr = &(ctx->seqStore);
-+	const BYTE *const base = ctx->base;
-+	const BYTE *const dictBase = ctx->dictBase;
-+	const BYTE *const istart = (const BYTE *)src;
-+	const BYTE *ip = istart;
-+	const BYTE *anchor = istart;
-+	const U32 lowestIndex = ctx->lowLimit;
-+	const BYTE *const dictStart = dictBase + lowestIndex;
-+	const U32 dictLimit = ctx->dictLimit;
-+	const BYTE *const lowPrefixPtr = base + dictLimit;
-+	const BYTE *const dictEnd = dictBase + dictLimit;
-+	const BYTE *const iend = istart + srcSize;
-+	const BYTE *const ilimit = iend - 8;
-+	U32 offset_1 = ctx->rep[0], offset_2 = ctx->rep[1];
-+
-+	/* Search Loop */
-+	while (ip < ilimit) { /* < instead of <=, because (ip+1) */
-+		const size_t hSmall = ZSTD_hashPtr(ip, hBitsS, mls);
-+		const U32 matchIndex = hashSmall[hSmall];
-+		const BYTE *matchBase = matchIndex < dictLimit ? dictBase : base;
-+		const BYTE *match = matchBase + matchIndex;
-+
-+		const size_t hLong = ZSTD_hashPtr(ip, hBitsL, 8);
-+		const U32 matchLongIndex = hashLong[hLong];
-+		const BYTE *matchLongBase = matchLongIndex < dictLimit ? dictBase : base;
-+		const BYTE *matchLong = matchLongBase + matchLongIndex;
-+
-+		const U32 curr = (U32)(ip - base);
-+		const U32 repIndex = curr + 1 - offset_1; /* offset_1 expected <= curr +1 */
-+		const BYTE *repBase = repIndex < dictLimit ? dictBase : base;
-+		const BYTE *repMatch = repBase + repIndex;
-+		size_t mLength;
-+		hashSmall[hSmall] = hashLong[hLong] = curr; /* update hash table */
-+
-+		if ((((U32)((dictLimit - 1) - repIndex) >= 3) /* intentional underflow */ & (repIndex > lowestIndex)) &&
-+		    (ZSTD_read32(repMatch) == ZSTD_read32(ip + 1))) {
-+			const BYTE *repMatchEnd = repIndex < dictLimit ? dictEnd : iend;
-+			mLength = ZSTD_count_2segments(ip + 1 + 4, repMatch + 4, iend, repMatchEnd, lowPrefixPtr) + 4;
-+			ip++;
-+			ZSTD_storeSeq(seqStorePtr, ip - anchor, anchor, 0, mLength - MINMATCH);
-+		} else {
-+			if ((matchLongIndex > lowestIndex) && (ZSTD_read64(matchLong) == ZSTD_read64(ip))) {
-+				const BYTE *matchEnd = matchLongIndex < dictLimit ? dictEnd : iend;
-+				const BYTE *lowMatchPtr = matchLongIndex < dictLimit ? dictStart : lowPrefixPtr;
-+				U32 offset;
-+				mLength = ZSTD_count_2segments(ip + 8, matchLong + 8, iend, matchEnd, lowPrefixPtr) + 8;
-+				offset = curr - matchLongIndex;
-+				while (((ip > anchor) & (matchLong > lowMatchPtr)) && (ip[-1] == matchLong[-1])) {
-+					ip--;
-+					matchLong--;
-+					mLength++;
-+				} /* catch up */
-+				offset_2 = offset_1;
-+				offset_1 = offset;
-+				ZSTD_storeSeq(seqStorePtr, ip - anchor, anchor, offset + ZSTD_REP_MOVE, mLength - MINMATCH);
-+
-+			} else if ((matchIndex > lowestIndex) && (ZSTD_read32(match) == ZSTD_read32(ip))) {
-+				size_t const h3 = ZSTD_hashPtr(ip + 1, hBitsL, 8);
-+				U32 const matchIndex3 = hashLong[h3];
-+				const BYTE *const match3Base = matchIndex3 < dictLimit ? dictBase : base;
-+				const BYTE *match3 = match3Base + matchIndex3;
-+				U32 offset;
-+				hashLong[h3] = curr + 1;
-+				if ((matchIndex3 > lowestIndex) && (ZSTD_read64(match3) == ZSTD_read64(ip + 1))) {
-+					const BYTE *matchEnd = matchIndex3 < dictLimit ? dictEnd : iend;
-+					const BYTE *lowMatchPtr = matchIndex3 < dictLimit ? dictStart : lowPrefixPtr;
-+					mLength = ZSTD_count_2segments(ip + 9, match3 + 8, iend, matchEnd, lowPrefixPtr) + 8;
-+					ip++;
-+					offset = curr + 1 - matchIndex3;
-+					while (((ip > anchor) & (match3 > lowMatchPtr)) && (ip[-1] == match3[-1])) {
-+						ip--;
-+						match3--;
-+						mLength++;
-+					} /* catch up */
-+				} else {
-+					const BYTE *matchEnd = matchIndex < dictLimit ? dictEnd : iend;
-+					const BYTE *lowMatchPtr = matchIndex < dictLimit ? dictStart : lowPrefixPtr;
-+					mLength = ZSTD_count_2segments(ip + 4, match + 4, iend, matchEnd, lowPrefixPtr) + 4;
-+					offset = curr - matchIndex;
-+					while (((ip > anchor) & (match > lowMatchPtr)) && (ip[-1] == match[-1])) {
-+						ip--;
-+						match--;
-+						mLength++;
-+					} /* catch up */
-+				}
-+				offset_2 = offset_1;
-+				offset_1 = offset;
-+				ZSTD_storeSeq(seqStorePtr, ip - anchor, anchor, offset + ZSTD_REP_MOVE, mLength - MINMATCH);
-+
-+			} else {
-+				ip += ((ip - anchor) >> g_searchStrength) + 1;
-+				continue;
-+			}
-+		}
-+
-+		/* found a match : store it */
-+		ip += mLength;
-+		anchor = ip;
-+
-+		if (ip <= ilimit) {
-+			/* Fill Table */
-+			hashSmall[ZSTD_hashPtr(base + curr + 2, hBitsS, mls)] = curr + 2;
-+			hashLong[ZSTD_hashPtr(base + curr + 2, hBitsL, 8)] = curr + 2;
-+			hashSmall[ZSTD_hashPtr(ip - 2, hBitsS, mls)] = (U32)(ip - 2 - base);
-+			hashLong[ZSTD_hashPtr(ip - 2, hBitsL, 8)] = (U32)(ip - 2 - base);
-+			/* check immediate repcode */
-+			while (ip <= ilimit) {
-+				U32 const curr2 = (U32)(ip - base);
-+				U32 const repIndex2 = curr2 - offset_2;
-+				const BYTE *repMatch2 = repIndex2 < dictLimit ? dictBase + repIndex2 : base + repIndex2;
-+				if ((((U32)((dictLimit - 1) - repIndex2) >= 3) & (repIndex2 > lowestIndex)) /* intentional overflow */
-+				    && (ZSTD_read32(repMatch2) == ZSTD_read32(ip))) {
-+					const BYTE *const repEnd2 = repIndex2 < dictLimit ? dictEnd : iend;
-+					size_t const repLength2 =
-+					    ZSTD_count_2segments(ip + EQUAL_READ32, repMatch2 + EQUAL_READ32, iend, repEnd2, lowPrefixPtr) + EQUAL_READ32;
-+					U32 tmpOffset = offset_2;
-+					offset_2 = offset_1;
-+					offset_1 = tmpOffset; /* swap offset_2 <=> offset_1 */
-+					ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, repLength2 - MINMATCH);
-+					hashSmall[ZSTD_hashPtr(ip, hBitsS, mls)] = curr2;
-+					hashLong[ZSTD_hashPtr(ip, hBitsL, 8)] = curr2;
-+					ip += repLength2;
-+					anchor = ip;
-+					continue;
-+				}
-+				break;
-+			}
-+		}
-+	}
-+
-+	/* save reps for next block */
-+	ctx->repToConfirm[0] = offset_1;
-+	ctx->repToConfirm[1] = offset_2;
-+
-+	/* Last Literals */
-+	{
-+		size_t const lastLLSize = iend - anchor;
-+		memcpy(seqStorePtr->lit, anchor, lastLLSize);
-+		seqStorePtr->lit += lastLLSize;
-+	}
-+}
-+
-+static void ZSTD_compressBlock_doubleFast_extDict(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-+{
-+	U32 const mls = ctx->params.cParams.searchLength;
-+	switch (mls) {
-+	default: /* includes case 3 */
-+	case 4: ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 4); return;
-+	case 5: ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 5); return;
-+	case 6: ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 6); return;
-+	case 7: ZSTD_compressBlock_doubleFast_extDict_generic(ctx, src, srcSize, 7); return;
-+	}
-+}
-+
-+/*-*************************************
-+*  Binary Tree search
-+***************************************/
-+/** ZSTD_insertBt1() : add one or multiple positions to tree.
-+*   ip : assumed <= iend-8 .
-+*   @return : nb of positions added */
-+static U32 ZSTD_insertBt1(ZSTD_CCtx *zc, const BYTE *const ip, const U32 mls, const BYTE *const iend, U32 nbCompares, U32 extDict)
-+{
-+	U32 *const hashTable = zc->hashTable;
-+	U32 const hashLog = zc->params.cParams.hashLog;
-+	size_t const h = ZSTD_hashPtr(ip, hashLog, mls);
-+	U32 *const bt = zc->chainTable;
-+	U32 const btLog = zc->params.cParams.chainLog - 1;
-+	U32 const btMask = (1 << btLog) - 1;
-+	U32 matchIndex = hashTable[h];
-+	size_t commonLengthSmaller = 0, commonLengthLarger = 0;
-+	const BYTE *const base = zc->base;
-+	const BYTE *const dictBase = zc->dictBase;
-+	const U32 dictLimit = zc->dictLimit;
-+	const BYTE *const dictEnd = dictBase + dictLimit;
-+	const BYTE *const prefixStart = base + dictLimit;
-+	const BYTE *match;
-+	const U32 curr = (U32)(ip - base);
-+	const U32 btLow = btMask >= curr ? 0 : curr - btMask;
-+	U32 *smallerPtr = bt + 2 * (curr & btMask);
-+	U32 *largerPtr = smallerPtr + 1;
-+	U32 dummy32; /* to be nullified at the end */
-+	U32 const windowLow = zc->lowLimit;
-+	U32 matchEndIdx = curr + 8;
-+	size_t bestLength = 8;
-+
-+	hashTable[h] = curr; /* Update Hash Table */
-+
-+	while (nbCompares-- && (matchIndex > windowLow)) {
-+		U32 *const nextPtr = bt + 2 * (matchIndex & btMask);
-+		size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */
-+
-+		if ((!extDict) || (matchIndex + matchLength >= dictLimit)) {
-+			match = base + matchIndex;
-+			if (match[matchLength] == ip[matchLength])
-+				matchLength += ZSTD_count(ip + matchLength + 1, match + matchLength + 1, iend) + 1;
-+		} else {
-+			match = dictBase + matchIndex;
-+			matchLength += ZSTD_count_2segments(ip + matchLength, match + matchLength, iend, dictEnd, prefixStart);
-+			if (matchIndex + matchLength >= dictLimit)
-+				match = base + matchIndex; /* to prepare for next usage of match[matchLength] */
-+		}
-+
-+		if (matchLength > bestLength) {
-+			bestLength = matchLength;
-+			if (matchLength > matchEndIdx - matchIndex)
-+				matchEndIdx = matchIndex + (U32)matchLength;
-+		}
-+
-+		if (ip + matchLength == iend) /* equal : no way to know if inf or sup */
-+			break;		      /* drop , to guarantee consistency ; miss a bit of compression, but other solutions can corrupt the tree */
-+
-+		if (match[matchLength] < ip[matchLength]) { /* necessarily within correct buffer */
-+			/* match is smaller than curr */
-+			*smallerPtr = matchIndex;	  /* update smaller idx */
-+			commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */
-+			if (matchIndex <= btLow) {
-+				smallerPtr = &dummy32;
-+				break;
-+			}			  /* beyond tree size, stop the search */
-+			smallerPtr = nextPtr + 1; /* new "smaller" => larger of match */
-+			matchIndex = nextPtr[1];  /* new matchIndex larger than previous (closer to curr) */
-+		} else {
-+			/* match is larger than curr */
-+			*largerPtr = matchIndex;
-+			commonLengthLarger = matchLength;
-+			if (matchIndex <= btLow) {
-+				largerPtr = &dummy32;
-+				break;
-+			} /* beyond tree size, stop the search */
-+			largerPtr = nextPtr;
-+			matchIndex = nextPtr[0];
-+		}
-+	}
-+
-+	*smallerPtr = *largerPtr = 0;
-+	if (bestLength > 384)
-+		return MIN(192, (U32)(bestLength - 384)); /* speed optimization */
-+	if (matchEndIdx > curr + 8)
-+		return matchEndIdx - curr - 8;
-+	return 1;
-+}
-+
-+static size_t ZSTD_insertBtAndFindBestMatch(ZSTD_CCtx *zc, const BYTE *const ip, const BYTE *const iend, size_t *offsetPtr, U32 nbCompares, const U32 mls,
-+					    U32 extDict)
-+{
-+	U32 *const hashTable = zc->hashTable;
-+	U32 const hashLog = zc->params.cParams.hashLog;
-+	size_t const h = ZSTD_hashPtr(ip, hashLog, mls);
-+	U32 *const bt = zc->chainTable;
-+	U32 const btLog = zc->params.cParams.chainLog - 1;
-+	U32 const btMask = (1 << btLog) - 1;
-+	U32 matchIndex = hashTable[h];
-+	size_t commonLengthSmaller = 0, commonLengthLarger = 0;
-+	const BYTE *const base = zc->base;
-+	const BYTE *const dictBase = zc->dictBase;
-+	const U32 dictLimit = zc->dictLimit;
-+	const BYTE *const dictEnd = dictBase + dictLimit;
-+	const BYTE *const prefixStart = base + dictLimit;
-+	const U32 curr = (U32)(ip - base);
-+	const U32 btLow = btMask >= curr ? 0 : curr - btMask;
-+	const U32 windowLow = zc->lowLimit;
-+	U32 *smallerPtr = bt + 2 * (curr & btMask);
-+	U32 *largerPtr = bt + 2 * (curr & btMask) + 1;
-+	U32 matchEndIdx = curr + 8;
-+	U32 dummy32; /* to be nullified at the end */
-+	size_t bestLength = 0;
-+
-+	hashTable[h] = curr; /* Update Hash Table */
-+
-+	while (nbCompares-- && (matchIndex > windowLow)) {
-+		U32 *const nextPtr = bt + 2 * (matchIndex & btMask);
-+		size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */
-+		const BYTE *match;
-+
-+		if ((!extDict) || (matchIndex + matchLength >= dictLimit)) {
-+			match = base + matchIndex;
-+			if (match[matchLength] == ip[matchLength])
-+				matchLength += ZSTD_count(ip + matchLength + 1, match + matchLength + 1, iend) + 1;
-+		} else {
-+			match = dictBase + matchIndex;
-+			matchLength += ZSTD_count_2segments(ip + matchLength, match + matchLength, iend, dictEnd, prefixStart);
-+			if (matchIndex + matchLength >= dictLimit)
-+				match = base + matchIndex; /* to prepare for next usage of match[matchLength] */
-+		}
-+
-+		if (matchLength > bestLength) {
-+			if (matchLength > matchEndIdx - matchIndex)
-+				matchEndIdx = matchIndex + (U32)matchLength;
-+			if ((4 * (int)(matchLength - bestLength)) > (int)(ZSTD_highbit32(curr - matchIndex + 1) - ZSTD_highbit32((U32)offsetPtr[0] + 1)))
-+				bestLength = matchLength, *offsetPtr = ZSTD_REP_MOVE + curr - matchIndex;
-+			if (ip + matchLength == iend) /* equal : no way to know if inf or sup */
-+				break;		      /* drop, to guarantee consistency (miss a little bit of compression) */
-+		}
-+
-+		if (match[matchLength] < ip[matchLength]) {
-+			/* match is smaller than curr */
-+			*smallerPtr = matchIndex;	  /* update smaller idx */
-+			commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */
-+			if (matchIndex <= btLow) {
-+				smallerPtr = &dummy32;
-+				break;
-+			}			  /* beyond tree size, stop the search */
-+			smallerPtr = nextPtr + 1; /* new "smaller" => larger of match */
-+			matchIndex = nextPtr[1];  /* new matchIndex larger than previous (closer to curr) */
-+		} else {
-+			/* match is larger than curr */
-+			*largerPtr = matchIndex;
-+			commonLengthLarger = matchLength;
-+			if (matchIndex <= btLow) {
-+				largerPtr = &dummy32;
-+				break;
-+			} /* beyond tree size, stop the search */
-+			largerPtr = nextPtr;
-+			matchIndex = nextPtr[0];
-+		}
-+	}
-+
-+	*smallerPtr = *largerPtr = 0;
-+
-+	zc->nextToUpdate = (matchEndIdx > curr + 8) ? matchEndIdx - 8 : curr + 1;
-+	return bestLength;
-+}
-+
-+static void ZSTD_updateTree(ZSTD_CCtx *zc, const BYTE *const ip, const BYTE *const iend, const U32 nbCompares, const U32 mls)
-+{
-+	const BYTE *const base = zc->base;
-+	const U32 target = (U32)(ip - base);
-+	U32 idx = zc->nextToUpdate;
-+
-+	while (idx < target)
-+		idx += ZSTD_insertBt1(zc, base + idx, mls, iend, nbCompares, 0);
-+}
-+
-+/** ZSTD_BtFindBestMatch() : Tree updater, providing best match */
-+static size_t ZSTD_BtFindBestMatch(ZSTD_CCtx *zc, const BYTE *const ip, const BYTE *const iLimit, size_t *offsetPtr, const U32 maxNbAttempts, const U32 mls)
-+{
-+	if (ip < zc->base + zc->nextToUpdate)
-+		return 0; /* skipped area */
-+	ZSTD_updateTree(zc, ip, iLimit, maxNbAttempts, mls);
-+	return ZSTD_insertBtAndFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, mls, 0);
-+}
-+
-+static size_t ZSTD_BtFindBestMatch_selectMLS(ZSTD_CCtx *zc, /* Index table will be updated */
-+					     const BYTE *ip, const BYTE *const iLimit, size_t *offsetPtr, const U32 maxNbAttempts, const U32 matchLengthSearch)
-+{
-+	switch (matchLengthSearch) {
-+	default: /* includes case 3 */
-+	case 4: return ZSTD_BtFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4);
-+	case 5: return ZSTD_BtFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5);
-+	case 7:
-+	case 6: return ZSTD_BtFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6);
-+	}
-+}
-+
-+static void ZSTD_updateTree_extDict(ZSTD_CCtx *zc, const BYTE *const ip, const BYTE *const iend, const U32 nbCompares, const U32 mls)
-+{
-+	const BYTE *const base = zc->base;
-+	const U32 target = (U32)(ip - base);
-+	U32 idx = zc->nextToUpdate;
-+
-+	while (idx < target)
-+		idx += ZSTD_insertBt1(zc, base + idx, mls, iend, nbCompares, 1);
-+}
-+
-+/** Tree updater, providing best match */
-+static size_t ZSTD_BtFindBestMatch_extDict(ZSTD_CCtx *zc, const BYTE *const ip, const BYTE *const iLimit, size_t *offsetPtr, const U32 maxNbAttempts,
-+					   const U32 mls)
-+{
-+	if (ip < zc->base + zc->nextToUpdate)
-+		return 0; /* skipped area */
-+	ZSTD_updateTree_extDict(zc, ip, iLimit, maxNbAttempts, mls);
-+	return ZSTD_insertBtAndFindBestMatch(zc, ip, iLimit, offsetPtr, maxNbAttempts, mls, 1);
-+}
-+
-+static size_t ZSTD_BtFindBestMatch_selectMLS_extDict(ZSTD_CCtx *zc, /* Index table will be updated */
-+						     const BYTE *ip, const BYTE *const iLimit, size_t *offsetPtr, const U32 maxNbAttempts,
-+						     const U32 matchLengthSearch)
-+{
-+	switch (matchLengthSearch) {
-+	default: /* includes case 3 */
-+	case 4: return ZSTD_BtFindBestMatch_extDict(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4);
-+	case 5: return ZSTD_BtFindBestMatch_extDict(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5);
-+	case 7:
-+	case 6: return ZSTD_BtFindBestMatch_extDict(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6);
-+	}
-+}
-+
-+/* *********************************
-+*  Hash Chain
-+***********************************/
-+#define NEXT_IN_CHAIN(d, mask) chainTable[(d)&mask]
-+
-+/* Update chains up to ip (excluded)
-+   Assumption : always within prefix (i.e. not within extDict) */
-+FORCE_INLINE
-+U32 ZSTD_insertAndFindFirstIndex(ZSTD_CCtx *zc, const BYTE *ip, U32 mls)
-+{
-+	U32 *const hashTable = zc->hashTable;
-+	const U32 hashLog = zc->params.cParams.hashLog;
-+	U32 *const chainTable = zc->chainTable;
-+	const U32 chainMask = (1 << zc->params.cParams.chainLog) - 1;
-+	const BYTE *const base = zc->base;
-+	const U32 target = (U32)(ip - base);
-+	U32 idx = zc->nextToUpdate;
-+
-+	while (idx < target) { /* catch up */
-+		size_t const h = ZSTD_hashPtr(base + idx, hashLog, mls);
-+		NEXT_IN_CHAIN(idx, chainMask) = hashTable[h];
-+		hashTable[h] = idx;
-+		idx++;
-+	}
-+
-+	zc->nextToUpdate = target;
-+	return hashTable[ZSTD_hashPtr(ip, hashLog, mls)];
-+}
-+
-+/* inlining is important to hardwire a hot branch (template emulation) */
-+FORCE_INLINE
-+size_t ZSTD_HcFindBestMatch_generic(ZSTD_CCtx *zc, /* Index table will be updated */
-+				    const BYTE *const ip, const BYTE *const iLimit, size_t *offsetPtr, const U32 maxNbAttempts, const U32 mls,
-+				    const U32 extDict)
-+{
-+	U32 *const chainTable = zc->chainTable;
-+	const U32 chainSize = (1 << zc->params.cParams.chainLog);
-+	const U32 chainMask = chainSize - 1;
-+	const BYTE *const base = zc->base;
-+	const BYTE *const dictBase = zc->dictBase;
-+	const U32 dictLimit = zc->dictLimit;
-+	const BYTE *const prefixStart = base + dictLimit;
-+	const BYTE *const dictEnd = dictBase + dictLimit;
-+	const U32 lowLimit = zc->lowLimit;
-+	const U32 curr = (U32)(ip - base);
-+	const U32 minChain = curr > chainSize ? curr - chainSize : 0;
-+	int nbAttempts = maxNbAttempts;
-+	size_t ml = EQUAL_READ32 - 1;
-+
-+	/* HC4 match finder */
-+	U32 matchIndex = ZSTD_insertAndFindFirstIndex(zc, ip, mls);
-+
-+	for (; (matchIndex > lowLimit) & (nbAttempts > 0); nbAttempts--) {
-+		const BYTE *match;
-+		size_t currMl = 0;
-+		if ((!extDict) || matchIndex >= dictLimit) {
-+			match = base + matchIndex;
-+			if (match[ml] == ip[ml]) /* potentially better */
-+				currMl = ZSTD_count(ip, match, iLimit);
-+		} else {
-+			match = dictBase + matchIndex;
-+			if (ZSTD_read32(match) == ZSTD_read32(ip)) /* assumption : matchIndex <= dictLimit-4 (by table construction) */
-+				currMl = ZSTD_count_2segments(ip + EQUAL_READ32, match + EQUAL_READ32, iLimit, dictEnd, prefixStart) + EQUAL_READ32;
-+		}
-+
-+		/* save best solution */
-+		if (currMl > ml) {
-+			ml = currMl;
-+			*offsetPtr = curr - matchIndex + ZSTD_REP_MOVE;
-+			if (ip + currMl == iLimit)
-+				break; /* best possible, and avoid read overflow*/
-+		}
-+
-+		if (matchIndex <= minChain)
-+			break;
-+		matchIndex = NEXT_IN_CHAIN(matchIndex, chainMask);
-+	}
-+
-+	return ml;
-+}
-+
-+FORCE_INLINE size_t ZSTD_HcFindBestMatch_selectMLS(ZSTD_CCtx *zc, const BYTE *ip, const BYTE *const iLimit, size_t *offsetPtr, const U32 maxNbAttempts,
-+						   const U32 matchLengthSearch)
-+{
-+	switch (matchLengthSearch) {
-+	default: /* includes case 3 */
-+	case 4: return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4, 0);
-+	case 5: return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5, 0);
-+	case 7:
-+	case 6: return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6, 0);
-+	}
-+}
-+
-+FORCE_INLINE size_t ZSTD_HcFindBestMatch_extDict_selectMLS(ZSTD_CCtx *zc, const BYTE *ip, const BYTE *const iLimit, size_t *offsetPtr, const U32 maxNbAttempts,
-+							   const U32 matchLengthSearch)
-+{
-+	switch (matchLengthSearch) {
-+	default: /* includes case 3 */
-+	case 4: return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 4, 1);
-+	case 5: return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 5, 1);
-+	case 7:
-+	case 6: return ZSTD_HcFindBestMatch_generic(zc, ip, iLimit, offsetPtr, maxNbAttempts, 6, 1);
-+	}
-+}
-+
-+/* *******************************
-+*  Common parser - lazy strategy
-+*********************************/
-+FORCE_INLINE
-+void ZSTD_compressBlock_lazy_generic(ZSTD_CCtx *ctx, const void *src, size_t srcSize, const U32 searchMethod, const U32 depth)
-+{
-+	seqStore_t *seqStorePtr = &(ctx->seqStore);
-+	const BYTE *const istart = (const BYTE *)src;
-+	const BYTE *ip = istart;
-+	const BYTE *anchor = istart;
-+	const BYTE *const iend = istart + srcSize;
-+	const BYTE *const ilimit = iend - 8;
-+	const BYTE *const base = ctx->base + ctx->dictLimit;
-+
-+	U32 const maxSearches = 1 << ctx->params.cParams.searchLog;
-+	U32 const mls = ctx->params.cParams.searchLength;
-+
-+	typedef size_t (*searchMax_f)(ZSTD_CCtx * zc, const BYTE *ip, const BYTE *iLimit, size_t *offsetPtr, U32 maxNbAttempts, U32 matchLengthSearch);
-+	searchMax_f const searchMax = searchMethod ? ZSTD_BtFindBestMatch_selectMLS : ZSTD_HcFindBestMatch_selectMLS;
-+	U32 offset_1 = ctx->rep[0], offset_2 = ctx->rep[1], savedOffset = 0;
-+
-+	/* init */
-+	ip += (ip == base);
-+	ctx->nextToUpdate3 = ctx->nextToUpdate;
-+	{
-+		U32 const maxRep = (U32)(ip - base);
-+		if (offset_2 > maxRep)
-+			savedOffset = offset_2, offset_2 = 0;
-+		if (offset_1 > maxRep)
-+			savedOffset = offset_1, offset_1 = 0;
-+	}
-+
-+	/* Match Loop */
-+	while (ip < ilimit) {
-+		size_t matchLength = 0;
-+		size_t offset = 0;
-+		const BYTE *start = ip + 1;
-+
-+		/* check repCode */
-+		if ((offset_1 > 0) & (ZSTD_read32(ip + 1) == ZSTD_read32(ip + 1 - offset_1))) {
-+			/* repcode : we take it */
-+			matchLength = ZSTD_count(ip + 1 + EQUAL_READ32, ip + 1 + EQUAL_READ32 - offset_1, iend) + EQUAL_READ32;
-+			if (depth == 0)
-+				goto _storeSequence;
-+		}
-+
-+		/* first search (depth 0) */
-+		{
-+			size_t offsetFound = 99999999;
-+			size_t const ml2 = searchMax(ctx, ip, iend, &offsetFound, maxSearches, mls);
-+			if (ml2 > matchLength)
-+				matchLength = ml2, start = ip, offset = offsetFound;
-+		}
-+
-+		if (matchLength < EQUAL_READ32) {
-+			ip += ((ip - anchor) >> g_searchStrength) + 1; /* jump faster over incompressible sections */
-+			continue;
-+		}
-+
-+		/* let's try to find a better solution */
-+		if (depth >= 1)
-+			while (ip < ilimit) {
-+				ip++;
-+				if ((offset) && ((offset_1 > 0) & (ZSTD_read32(ip) == ZSTD_read32(ip - offset_1)))) {
-+					size_t const mlRep = ZSTD_count(ip + EQUAL_READ32, ip + EQUAL_READ32 - offset_1, iend) + EQUAL_READ32;
-+					int const gain2 = (int)(mlRep * 3);
-+					int const gain1 = (int)(matchLength * 3 - ZSTD_highbit32((U32)offset + 1) + 1);
-+					if ((mlRep >= EQUAL_READ32) && (gain2 > gain1))
-+						matchLength = mlRep, offset = 0, start = ip;
-+				}
-+				{
-+					size_t offset2 = 99999999;
-+					size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);
-+					int const gain2 = (int)(ml2 * 4 - ZSTD_highbit32((U32)offset2 + 1)); /* raw approx */
-+					int const gain1 = (int)(matchLength * 4 - ZSTD_highbit32((U32)offset + 1) + 4);
-+					if ((ml2 >= EQUAL_READ32) && (gain2 > gain1)) {
-+						matchLength = ml2, offset = offset2, start = ip;
-+						continue; /* search a better one */
-+					}
-+				}
-+
-+				/* let's find an even better one */
-+				if ((depth == 2) && (ip < ilimit)) {
-+					ip++;
-+					if ((offset) && ((offset_1 > 0) & (ZSTD_read32(ip) == ZSTD_read32(ip - offset_1)))) {
-+						size_t const ml2 = ZSTD_count(ip + EQUAL_READ32, ip + EQUAL_READ32 - offset_1, iend) + EQUAL_READ32;
-+						int const gain2 = (int)(ml2 * 4);
-+						int const gain1 = (int)(matchLength * 4 - ZSTD_highbit32((U32)offset + 1) + 1);
-+						if ((ml2 >= EQUAL_READ32) && (gain2 > gain1))
-+							matchLength = ml2, offset = 0, start = ip;
-+					}
-+					{
-+						size_t offset2 = 99999999;
-+						size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);
-+						int const gain2 = (int)(ml2 * 4 - ZSTD_highbit32((U32)offset2 + 1)); /* raw approx */
-+						int const gain1 = (int)(matchLength * 4 - ZSTD_highbit32((U32)offset + 1) + 7);
-+						if ((ml2 >= EQUAL_READ32) && (gain2 > gain1)) {
-+							matchLength = ml2, offset = offset2, start = ip;
-+							continue;
-+						}
-+					}
-+				}
-+				break; /* nothing found : store previous solution */
-+			}
-+
-+		/* NOTE:
-+		 * start[-offset+ZSTD_REP_MOVE-1] is undefined behavior.
-+		 * (-offset+ZSTD_REP_MOVE-1) is unsigned, and is added to start, which
-+		 * overflows the pointer, which is undefined behavior.
-+		 */
-+		/* catch up */
-+		if (offset) {
-+			while ((start > anchor) && (start > base + offset - ZSTD_REP_MOVE) &&
-+			       (start[-1] == (start-offset+ZSTD_REP_MOVE)[-1])) /* only search for offset within prefix */
-+			{
-+				start--;
-+				matchLength++;
-+			}
-+			offset_2 = offset_1;
-+			offset_1 = (U32)(offset - ZSTD_REP_MOVE);
-+		}
-+
-+	/* store sequence */
-+_storeSequence:
-+		{
-+			size_t const litLength = start - anchor;
-+			ZSTD_storeSeq(seqStorePtr, litLength, anchor, (U32)offset, matchLength - MINMATCH);
-+			anchor = ip = start + matchLength;
-+		}
-+
-+		/* check immediate repcode */
-+		while ((ip <= ilimit) && ((offset_2 > 0) & (ZSTD_read32(ip) == ZSTD_read32(ip - offset_2)))) {
-+			/* store sequence */
-+			matchLength = ZSTD_count(ip + EQUAL_READ32, ip + EQUAL_READ32 - offset_2, iend) + EQUAL_READ32;
-+			offset = offset_2;
-+			offset_2 = offset_1;
-+			offset_1 = (U32)offset; /* swap repcodes */
-+			ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, matchLength - MINMATCH);
-+			ip += matchLength;
-+			anchor = ip;
-+			continue; /* faster when present ... (?) */
-+		}
-+	}
-+
-+	/* Save reps for next block */
-+	ctx->repToConfirm[0] = offset_1 ? offset_1 : savedOffset;
-+	ctx->repToConfirm[1] = offset_2 ? offset_2 : savedOffset;
-+
-+	/* Last Literals */
-+	{
-+		size_t const lastLLSize = iend - anchor;
-+		memcpy(seqStorePtr->lit, anchor, lastLLSize);
-+		seqStorePtr->lit += lastLLSize;
-+	}
-+}
-+
-+static void ZSTD_compressBlock_btlazy2(ZSTD_CCtx *ctx, const void *src, size_t srcSize) { ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 1, 2); }
-+
-+static void ZSTD_compressBlock_lazy2(ZSTD_CCtx *ctx, const void *src, size_t srcSize) { ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 0, 2); }
-+
-+static void ZSTD_compressBlock_lazy(ZSTD_CCtx *ctx, const void *src, size_t srcSize) { ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 0, 1); }
-+
-+static void ZSTD_compressBlock_greedy(ZSTD_CCtx *ctx, const void *src, size_t srcSize) { ZSTD_compressBlock_lazy_generic(ctx, src, srcSize, 0, 0); }
-+
-+FORCE_INLINE
-+void ZSTD_compressBlock_lazy_extDict_generic(ZSTD_CCtx *ctx, const void *src, size_t srcSize, const U32 searchMethod, const U32 depth)
-+{
-+	seqStore_t *seqStorePtr = &(ctx->seqStore);
-+	const BYTE *const istart = (const BYTE *)src;
-+	const BYTE *ip = istart;
-+	const BYTE *anchor = istart;
-+	const BYTE *const iend = istart + srcSize;
-+	const BYTE *const ilimit = iend - 8;
-+	const BYTE *const base = ctx->base;
-+	const U32 dictLimit = ctx->dictLimit;
-+	const U32 lowestIndex = ctx->lowLimit;
-+	const BYTE *const prefixStart = base + dictLimit;
-+	const BYTE *const dictBase = ctx->dictBase;
-+	const BYTE *const dictEnd = dictBase + dictLimit;
-+	const BYTE *const dictStart = dictBase + ctx->lowLimit;
-+
-+	const U32 maxSearches = 1 << ctx->params.cParams.searchLog;
-+	const U32 mls = ctx->params.cParams.searchLength;
-+
-+	typedef size_t (*searchMax_f)(ZSTD_CCtx * zc, const BYTE *ip, const BYTE *iLimit, size_t *offsetPtr, U32 maxNbAttempts, U32 matchLengthSearch);
-+	searchMax_f searchMax = searchMethod ? ZSTD_BtFindBestMatch_selectMLS_extDict : ZSTD_HcFindBestMatch_extDict_selectMLS;
-+
-+	U32 offset_1 = ctx->rep[0], offset_2 = ctx->rep[1];
-+
-+	/* init */
-+	ctx->nextToUpdate3 = ctx->nextToUpdate;
-+	ip += (ip == prefixStart);
-+
-+	/* Match Loop */
-+	while (ip < ilimit) {
-+		size_t matchLength = 0;
-+		size_t offset = 0;
-+		const BYTE *start = ip + 1;
-+		U32 curr = (U32)(ip - base);
-+
-+		/* check repCode */
-+		{
-+			const U32 repIndex = (U32)(curr + 1 - offset_1);
-+			const BYTE *const repBase = repIndex < dictLimit ? dictBase : base;
-+			const BYTE *const repMatch = repBase + repIndex;
-+			if (((U32)((dictLimit - 1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */
-+				if (ZSTD_read32(ip + 1) == ZSTD_read32(repMatch)) {
-+					/* repcode detected we should take it */
-+					const BYTE *const repEnd = repIndex < dictLimit ? dictEnd : iend;
-+					matchLength =
-+					    ZSTD_count_2segments(ip + 1 + EQUAL_READ32, repMatch + EQUAL_READ32, iend, repEnd, prefixStart) + EQUAL_READ32;
-+					if (depth == 0)
-+						goto _storeSequence;
-+				}
-+		}
-+
-+		/* first search (depth 0) */
-+		{
-+			size_t offsetFound = 99999999;
-+			size_t const ml2 = searchMax(ctx, ip, iend, &offsetFound, maxSearches, mls);
-+			if (ml2 > matchLength)
-+				matchLength = ml2, start = ip, offset = offsetFound;
-+		}
-+
-+		if (matchLength < EQUAL_READ32) {
-+			ip += ((ip - anchor) >> g_searchStrength) + 1; /* jump faster over incompressible sections */
-+			continue;
-+		}
-+
-+		/* let's try to find a better solution */
-+		if (depth >= 1)
-+			while (ip < ilimit) {
-+				ip++;
-+				curr++;
-+				/* check repCode */
-+				if (offset) {
-+					const U32 repIndex = (U32)(curr - offset_1);
-+					const BYTE *const repBase = repIndex < dictLimit ? dictBase : base;
-+					const BYTE *const repMatch = repBase + repIndex;
-+					if (((U32)((dictLimit - 1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */
-+						if (ZSTD_read32(ip) == ZSTD_read32(repMatch)) {
-+							/* repcode detected */
-+							const BYTE *const repEnd = repIndex < dictLimit ? dictEnd : iend;
-+							size_t const repLength =
-+							    ZSTD_count_2segments(ip + EQUAL_READ32, repMatch + EQUAL_READ32, iend, repEnd, prefixStart) +
-+							    EQUAL_READ32;
-+							int const gain2 = (int)(repLength * 3);
-+							int const gain1 = (int)(matchLength * 3 - ZSTD_highbit32((U32)offset + 1) + 1);
-+							if ((repLength >= EQUAL_READ32) && (gain2 > gain1))
-+								matchLength = repLength, offset = 0, start = ip;
-+						}
-+				}
-+
-+				/* search match, depth 1 */
-+				{
-+					size_t offset2 = 99999999;
-+					size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);
-+					int const gain2 = (int)(ml2 * 4 - ZSTD_highbit32((U32)offset2 + 1)); /* raw approx */
-+					int const gain1 = (int)(matchLength * 4 - ZSTD_highbit32((U32)offset + 1) + 4);
-+					if ((ml2 >= EQUAL_READ32) && (gain2 > gain1)) {
-+						matchLength = ml2, offset = offset2, start = ip;
-+						continue; /* search a better one */
-+					}
-+				}
-+
-+				/* let's find an even better one */
-+				if ((depth == 2) && (ip < ilimit)) {
-+					ip++;
-+					curr++;
-+					/* check repCode */
-+					if (offset) {
-+						const U32 repIndex = (U32)(curr - offset_1);
-+						const BYTE *const repBase = repIndex < dictLimit ? dictBase : base;
-+						const BYTE *const repMatch = repBase + repIndex;
-+						if (((U32)((dictLimit - 1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */
-+							if (ZSTD_read32(ip) == ZSTD_read32(repMatch)) {
-+								/* repcode detected */
-+								const BYTE *const repEnd = repIndex < dictLimit ? dictEnd : iend;
-+								size_t repLength = ZSTD_count_2segments(ip + EQUAL_READ32, repMatch + EQUAL_READ32, iend,
-+													repEnd, prefixStart) +
-+										   EQUAL_READ32;
-+								int gain2 = (int)(repLength * 4);
-+								int gain1 = (int)(matchLength * 4 - ZSTD_highbit32((U32)offset + 1) + 1);
-+								if ((repLength >= EQUAL_READ32) && (gain2 > gain1))
-+									matchLength = repLength, offset = 0, start = ip;
-+							}
-+					}
-+
-+					/* search match, depth 2 */
-+					{
-+						size_t offset2 = 99999999;
-+						size_t const ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);
-+						int const gain2 = (int)(ml2 * 4 - ZSTD_highbit32((U32)offset2 + 1)); /* raw approx */
-+						int const gain1 = (int)(matchLength * 4 - ZSTD_highbit32((U32)offset + 1) + 7);
-+						if ((ml2 >= EQUAL_READ32) && (gain2 > gain1)) {
-+							matchLength = ml2, offset = offset2, start = ip;
-+							continue;
-+						}
-+					}
-+				}
-+				break; /* nothing found : store previous solution */
-+			}
-+
-+		/* catch up */
-+		if (offset) {
-+			U32 const matchIndex = (U32)((start - base) - (offset - ZSTD_REP_MOVE));
-+			const BYTE *match = (matchIndex < dictLimit) ? dictBase + matchIndex : base + matchIndex;
-+			const BYTE *const mStart = (matchIndex < dictLimit) ? dictStart : prefixStart;
-+			while ((start > anchor) && (match > mStart) && (start[-1] == match[-1])) {
-+				start--;
-+				match--;
-+				matchLength++;
-+			} /* catch up */
-+			offset_2 = offset_1;
-+			offset_1 = (U32)(offset - ZSTD_REP_MOVE);
-+		}
-+
-+	/* store sequence */
-+	_storeSequence : {
-+		size_t const litLength = start - anchor;
-+		ZSTD_storeSeq(seqStorePtr, litLength, anchor, (U32)offset, matchLength - MINMATCH);
-+		anchor = ip = start + matchLength;
-+	}
-+
-+		/* check immediate repcode */
-+		while (ip <= ilimit) {
-+			const U32 repIndex = (U32)((ip - base) - offset_2);
-+			const BYTE *const repBase = repIndex < dictLimit ? dictBase : base;
-+			const BYTE *const repMatch = repBase + repIndex;
-+			if (((U32)((dictLimit - 1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */
-+				if (ZSTD_read32(ip) == ZSTD_read32(repMatch)) {
-+					/* repcode detected we should take it */
-+					const BYTE *const repEnd = repIndex < dictLimit ? dictEnd : iend;
-+					matchLength =
-+					    ZSTD_count_2segments(ip + EQUAL_READ32, repMatch + EQUAL_READ32, iend, repEnd, prefixStart) + EQUAL_READ32;
-+					offset = offset_2;
-+					offset_2 = offset_1;
-+					offset_1 = (U32)offset; /* swap offset history */
-+					ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, matchLength - MINMATCH);
-+					ip += matchLength;
-+					anchor = ip;
-+					continue; /* faster when present ... (?) */
-+				}
-+			break;
-+		}
-+	}
-+
-+	/* Save reps for next block */
-+	ctx->repToConfirm[0] = offset_1;
-+	ctx->repToConfirm[1] = offset_2;
-+
-+	/* Last Literals */
-+	{
-+		size_t const lastLLSize = iend - anchor;
-+		memcpy(seqStorePtr->lit, anchor, lastLLSize);
-+		seqStorePtr->lit += lastLLSize;
-+	}
-+}
-+
-+void ZSTD_compressBlock_greedy_extDict(ZSTD_CCtx *ctx, const void *src, size_t srcSize) { ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 0, 0); }
-+
-+static void ZSTD_compressBlock_lazy_extDict(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-+{
-+	ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 0, 1);
-+}
-+
-+static void ZSTD_compressBlock_lazy2_extDict(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-+{
-+	ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 0, 2);
-+}
-+
-+static void ZSTD_compressBlock_btlazy2_extDict(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-+{
-+	ZSTD_compressBlock_lazy_extDict_generic(ctx, src, srcSize, 1, 2);
-+}
-+
-+/* The optimal parser */
-+#include "zstd_opt.h"
-+
-+static void ZSTD_compressBlock_btopt(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-+{
-+#ifdef ZSTD_OPT_H_91842398743
-+	ZSTD_compressBlock_opt_generic(ctx, src, srcSize, 0);
-+#else
-+	(void)ctx;
-+	(void)src;
-+	(void)srcSize;
-+	return;
-+#endif
-+}
-+
-+static void ZSTD_compressBlock_btopt2(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-+{
-+#ifdef ZSTD_OPT_H_91842398743
-+	ZSTD_compressBlock_opt_generic(ctx, src, srcSize, 1);
-+#else
-+	(void)ctx;
-+	(void)src;
-+	(void)srcSize;
-+	return;
-+#endif
-+}
-+
-+static void ZSTD_compressBlock_btopt_extDict(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-+{
-+#ifdef ZSTD_OPT_H_91842398743
-+	ZSTD_compressBlock_opt_extDict_generic(ctx, src, srcSize, 0);
-+#else
-+	(void)ctx;
-+	(void)src;
-+	(void)srcSize;
-+	return;
-+#endif
-+}
-+
-+static void ZSTD_compressBlock_btopt2_extDict(ZSTD_CCtx *ctx, const void *src, size_t srcSize)
-+{
-+#ifdef ZSTD_OPT_H_91842398743
-+	ZSTD_compressBlock_opt_extDict_generic(ctx, src, srcSize, 1);
-+#else
-+	(void)ctx;
-+	(void)src;
-+	(void)srcSize;
-+	return;
-+#endif
-+}
-+
-+typedef void (*ZSTD_blockCompressor)(ZSTD_CCtx *ctx, const void *src, size_t srcSize);
-+
-+static ZSTD_blockCompressor ZSTD_selectBlockCompressor(ZSTD_strategy strat, int extDict)
-+{
-+	static const ZSTD_blockCompressor blockCompressor[2][8] = {
-+	    {ZSTD_compressBlock_fast, ZSTD_compressBlock_doubleFast, ZSTD_compressBlock_greedy, ZSTD_compressBlock_lazy, ZSTD_compressBlock_lazy2,
-+	     ZSTD_compressBlock_btlazy2, ZSTD_compressBlock_btopt, ZSTD_compressBlock_btopt2},
-+	    {ZSTD_compressBlock_fast_extDict, ZSTD_compressBlock_doubleFast_extDict, ZSTD_compressBlock_greedy_extDict, ZSTD_compressBlock_lazy_extDict,
-+	     ZSTD_compressBlock_lazy2_extDict, ZSTD_compressBlock_btlazy2_extDict, ZSTD_compressBlock_btopt_extDict, ZSTD_compressBlock_btopt2_extDict}};
-+
-+	return blockCompressor[extDict][(U32)strat];
-+}
-+
-+static size_t ZSTD_compressBlock_internal(ZSTD_CCtx *zc, void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-+{
-+	ZSTD_blockCompressor const blockCompressor = ZSTD_selectBlockCompressor(zc->params.cParams.strategy, zc->lowLimit < zc->dictLimit);
-+	const BYTE *const base = zc->base;
-+	const BYTE *const istart = (const BYTE *)src;
-+	const U32 curr = (U32)(istart - base);
-+	if (srcSize < MIN_CBLOCK_SIZE + ZSTD_blockHeaderSize + 1)
-+		return 0; /* don't even attempt compression below a certain srcSize */
-+	ZSTD_resetSeqStore(&(zc->seqStore));
-+	if (curr > zc->nextToUpdate + 384)
-+		zc->nextToUpdate = curr - MIN(192, (U32)(curr - zc->nextToUpdate - 384)); /* update tree not updated after finding very long rep matches */
-+	blockCompressor(zc, src, srcSize);
-+	return ZSTD_compressSequences(zc, dst, dstCapacity, srcSize);
-+}
-+
-+/*! ZSTD_compress_generic() :
-+*   Compress a chunk of data into one or multiple blocks.
-+*   All blocks will be terminated, all input will be consumed.
-+*   Function will issue an error if there is not enough `dstCapacity` to hold the compressed content.
-+*   Frame is supposed already started (header already produced)
-+*   @return : compressed size, or an error code
-+*/
-+static size_t ZSTD_compress_generic(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, U32 lastFrameChunk)
-+{
-+	size_t blockSize = cctx->blockSize;
-+	size_t remaining = srcSize;
-+	const BYTE *ip = (const BYTE *)src;
-+	BYTE *const ostart = (BYTE *)dst;
-+	BYTE *op = ostart;
-+	U32 const maxDist = 1 << cctx->params.cParams.windowLog;
-+
-+	if (cctx->params.fParams.checksumFlag && srcSize)
-+		xxh64_update(&cctx->xxhState, src, srcSize);
-+
-+	while (remaining) {
-+		U32 const lastBlock = lastFrameChunk & (blockSize >= remaining);
-+		size_t cSize;
-+
-+		if (dstCapacity < ZSTD_blockHeaderSize + MIN_CBLOCK_SIZE)
-+			return ERROR(dstSize_tooSmall); /* not enough space to store compressed block */
-+		if (remaining < blockSize)
-+			blockSize = remaining;
-+
-+		/* preemptive overflow correction */
-+		if (cctx->lowLimit > (3U << 29)) {
-+			U32 const cycleMask = (1 << ZSTD_cycleLog(cctx->params.cParams.hashLog, cctx->params.cParams.strategy)) - 1;
-+			U32 const curr = (U32)(ip - cctx->base);
-+			U32 const newCurr = (curr & cycleMask) + (1 << cctx->params.cParams.windowLog);
-+			U32 const correction = curr - newCurr;
-+			ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX_64 <= 30);
-+			ZSTD_reduceIndex(cctx, correction);
-+			cctx->base += correction;
-+			cctx->dictBase += correction;
-+			cctx->lowLimit -= correction;
-+			cctx->dictLimit -= correction;
-+			if (cctx->nextToUpdate < correction)
-+				cctx->nextToUpdate = 0;
-+			else
-+				cctx->nextToUpdate -= correction;
-+		}
-+
-+		if ((U32)(ip + blockSize - cctx->base) > cctx->loadedDictEnd + maxDist) {
-+			/* enforce maxDist */
-+			U32 const newLowLimit = (U32)(ip + blockSize - cctx->base) - maxDist;
-+			if (cctx->lowLimit < newLowLimit)
-+				cctx->lowLimit = newLowLimit;
-+			if (cctx->dictLimit < cctx->lowLimit)
-+				cctx->dictLimit = cctx->lowLimit;
-+		}
-+
-+		cSize = ZSTD_compressBlock_internal(cctx, op + ZSTD_blockHeaderSize, dstCapacity - ZSTD_blockHeaderSize, ip, blockSize);
-+		if (ZSTD_isError(cSize))
-+			return cSize;
-+
-+		if (cSize == 0) { /* block is not compressible */
-+			U32 const cBlockHeader24 = lastBlock + (((U32)bt_raw) << 1) + (U32)(blockSize << 3);
-+			if (blockSize + ZSTD_blockHeaderSize > dstCapacity)
-+				return ERROR(dstSize_tooSmall);
-+			ZSTD_writeLE32(op, cBlockHeader24); /* no pb, 4th byte will be overwritten */
-+			memcpy(op + ZSTD_blockHeaderSize, ip, blockSize);
-+			cSize = ZSTD_blockHeaderSize + blockSize;
-+		} else {
-+			U32 const cBlockHeader24 = lastBlock + (((U32)bt_compressed) << 1) + (U32)(cSize << 3);
-+			ZSTD_writeLE24(op, cBlockHeader24);
-+			cSize += ZSTD_blockHeaderSize;
-+		}
-+
-+		remaining -= blockSize;
-+		dstCapacity -= cSize;
-+		ip += blockSize;
-+		op += cSize;
-+	}
-+
-+	if (lastFrameChunk && (op > ostart))
-+		cctx->stage = ZSTDcs_ending;
-+	return op - ostart;
-+}
-+
-+static size_t ZSTD_writeFrameHeader(void *dst, size_t dstCapacity, ZSTD_parameters params, U64 pledgedSrcSize, U32 dictID)
-+{
-+	BYTE *const op = (BYTE *)dst;
-+	U32 const dictIDSizeCode = (dictID > 0) + (dictID >= 256) + (dictID >= 65536); /* 0-3 */
-+	U32 const checksumFlag = params.fParams.checksumFlag > 0;
-+	U32 const windowSize = 1U << params.cParams.windowLog;
-+	U32 const singleSegment = params.fParams.contentSizeFlag && (windowSize >= pledgedSrcSize);
-+	BYTE const windowLogByte = (BYTE)((params.cParams.windowLog - ZSTD_WINDOWLOG_ABSOLUTEMIN) << 3);
-+	U32 const fcsCode =
-+	    params.fParams.contentSizeFlag ? (pledgedSrcSize >= 256) + (pledgedSrcSize >= 65536 + 256) + (pledgedSrcSize >= 0xFFFFFFFFU) : 0; /* 0-3 */
-+	BYTE const frameHeaderDecriptionByte = (BYTE)(dictIDSizeCode + (checksumFlag << 2) + (singleSegment << 5) + (fcsCode << 6));
-+	size_t pos;
-+
-+	if (dstCapacity < ZSTD_frameHeaderSize_max)
-+		return ERROR(dstSize_tooSmall);
-+
-+	ZSTD_writeLE32(dst, ZSTD_MAGICNUMBER);
-+	op[4] = frameHeaderDecriptionByte;
-+	pos = 5;
-+	if (!singleSegment)
-+		op[pos++] = windowLogByte;
-+	switch (dictIDSizeCode) {
-+	default: /* impossible */
-+	case 0: break;
-+	case 1:
-+		op[pos] = (BYTE)(dictID);
-+		pos++;
-+		break;
-+	case 2:
-+		ZSTD_writeLE16(op + pos, (U16)dictID);
-+		pos += 2;
-+		break;
-+	case 3:
-+		ZSTD_writeLE32(op + pos, dictID);
-+		pos += 4;
-+		break;
-+	}
-+	switch (fcsCode) {
-+	default: /* impossible */
-+	case 0:
-+		if (singleSegment)
-+			op[pos++] = (BYTE)(pledgedSrcSize);
-+		break;
-+	case 1:
-+		ZSTD_writeLE16(op + pos, (U16)(pledgedSrcSize - 256));
-+		pos += 2;
-+		break;
-+	case 2:
-+		ZSTD_writeLE32(op + pos, (U32)(pledgedSrcSize));
-+		pos += 4;
-+		break;
-+	case 3:
-+		ZSTD_writeLE64(op + pos, (U64)(pledgedSrcSize));
-+		pos += 8;
-+		break;
-+	}
-+	return pos;
-+}
-+
-+static size_t ZSTD_compressContinue_internal(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, U32 frame, U32 lastFrameChunk)
-+{
-+	const BYTE *const ip = (const BYTE *)src;
-+	size_t fhSize = 0;
-+
-+	if (cctx->stage == ZSTDcs_created)
-+		return ERROR(stage_wrong); /* missing init (ZSTD_compressBegin) */
-+
-+	if (frame && (cctx->stage == ZSTDcs_init)) {
-+		fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, cctx->params, cctx->frameContentSize, cctx->dictID);
-+		if (ZSTD_isError(fhSize))
-+			return fhSize;
-+		dstCapacity -= fhSize;
-+		dst = (char *)dst + fhSize;
-+		cctx->stage = ZSTDcs_ongoing;
-+	}
-+
-+	/* Check if blocks follow each other */
-+	if (src != cctx->nextSrc) {
-+		/* not contiguous */
-+		ptrdiff_t const delta = cctx->nextSrc - ip;
-+		cctx->lowLimit = cctx->dictLimit;
-+		cctx->dictLimit = (U32)(cctx->nextSrc - cctx->base);
-+		cctx->dictBase = cctx->base;
-+		cctx->base -= delta;
-+		cctx->nextToUpdate = cctx->dictLimit;
-+		if (cctx->dictLimit - cctx->lowLimit < HASH_READ_SIZE)
-+			cctx->lowLimit = cctx->dictLimit; /* too small extDict */
-+	}
-+
-+	/* if input and dictionary overlap : reduce dictionary (area presumed modified by input) */
-+	if ((ip + srcSize > cctx->dictBase + cctx->lowLimit) & (ip < cctx->dictBase + cctx->dictLimit)) {
-+		ptrdiff_t const highInputIdx = (ip + srcSize) - cctx->dictBase;
-+		U32 const lowLimitMax = (highInputIdx > (ptrdiff_t)cctx->dictLimit) ? cctx->dictLimit : (U32)highInputIdx;
-+		cctx->lowLimit = lowLimitMax;
-+	}
-+
-+	cctx->nextSrc = ip + srcSize;
-+
-+	if (srcSize) {
-+		size_t const cSize = frame ? ZSTD_compress_generic(cctx, dst, dstCapacity, src, srcSize, lastFrameChunk)
-+					   : ZSTD_compressBlock_internal(cctx, dst, dstCapacity, src, srcSize);
-+		if (ZSTD_isError(cSize))
-+			return cSize;
-+		return cSize + fhSize;
-+	} else
-+		return fhSize;
-+}
-+
-+size_t ZSTD_compressContinue(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-+{
-+	return ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 1, 0);
-+}
-+
-+size_t ZSTD_getBlockSizeMax(ZSTD_CCtx *cctx) { return MIN(ZSTD_BLOCKSIZE_ABSOLUTEMAX, 1 << cctx->params.cParams.windowLog); }
-+
-+size_t ZSTD_compressBlock(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-+{
-+	size_t const blockSizeMax = ZSTD_getBlockSizeMax(cctx);
-+	if (srcSize > blockSizeMax)
-+		return ERROR(srcSize_wrong);
-+	return ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 0, 0);
-+}
-+
-+/*! ZSTD_loadDictionaryContent() :
-+ *  @return : 0, or an error code
-+ */
-+static size_t ZSTD_loadDictionaryContent(ZSTD_CCtx *zc, const void *src, size_t srcSize)
-+{
-+	const BYTE *const ip = (const BYTE *)src;
-+	const BYTE *const iend = ip + srcSize;
-+
-+	/* input becomes curr prefix */
-+	zc->lowLimit = zc->dictLimit;
-+	zc->dictLimit = (U32)(zc->nextSrc - zc->base);
-+	zc->dictBase = zc->base;
-+	zc->base += ip - zc->nextSrc;
-+	zc->nextToUpdate = zc->dictLimit;
-+	zc->loadedDictEnd = zc->forceWindow ? 0 : (U32)(iend - zc->base);
-+
-+	zc->nextSrc = iend;
-+	if (srcSize <= HASH_READ_SIZE)
-+		return 0;
-+
-+	switch (zc->params.cParams.strategy) {
-+	case ZSTD_fast: ZSTD_fillHashTable(zc, iend, zc->params.cParams.searchLength); break;
-+
-+	case ZSTD_dfast: ZSTD_fillDoubleHashTable(zc, iend, zc->params.cParams.searchLength); break;
-+
-+	case ZSTD_greedy:
-+	case ZSTD_lazy:
-+	case ZSTD_lazy2:
-+		if (srcSize >= HASH_READ_SIZE)
-+			ZSTD_insertAndFindFirstIndex(zc, iend - HASH_READ_SIZE, zc->params.cParams.searchLength);
-+		break;
-+
-+	case ZSTD_btlazy2:
-+	case ZSTD_btopt:
-+	case ZSTD_btopt2:
-+		if (srcSize >= HASH_READ_SIZE)
-+			ZSTD_updateTree(zc, iend - HASH_READ_SIZE, iend, 1 << zc->params.cParams.searchLog, zc->params.cParams.searchLength);
-+		break;
-+
-+	default:
-+		return ERROR(GENERIC); /* strategy doesn't exist; impossible */
-+	}
-+
-+	zc->nextToUpdate = (U32)(iend - zc->base);
-+	return 0;
-+}
-+
-+/* Dictionaries that assign zero probability to symbols that show up causes problems
-+   when FSE encoding.  Refuse dictionaries that assign zero probability to symbols
-+   that we may encounter during compression.
-+   NOTE: This behavior is not standard and could be improved in the future. */
-+static size_t ZSTD_checkDictNCount(short *normalizedCounter, unsigned dictMaxSymbolValue, unsigned maxSymbolValue)
-+{
-+	U32 s;
-+	if (dictMaxSymbolValue < maxSymbolValue)
-+		return ERROR(dictionary_corrupted);
-+	for (s = 0; s <= maxSymbolValue; ++s) {
-+		if (normalizedCounter[s] == 0)
-+			return ERROR(dictionary_corrupted);
-+	}
-+	return 0;
-+}
-+
-+/* Dictionary format :
-+ * See :
-+ * https://github.com/facebook/zstd/blob/master/doc/zstd_compression_format.md#dictionary-format
-+ */
-+/*! ZSTD_loadZstdDictionary() :
-+ * @return : 0, or an error code
-+ *  assumptions : magic number supposed already checked
-+ *                dictSize supposed > 8
-+ */
-+static size_t ZSTD_loadZstdDictionary(ZSTD_CCtx *cctx, const void *dict, size_t dictSize)
-+{
-+	const BYTE *dictPtr = (const BYTE *)dict;
-+	const BYTE *const dictEnd = dictPtr + dictSize;
-+	short offcodeNCount[MaxOff + 1];
-+	unsigned offcodeMaxValue = MaxOff;
-+
-+	dictPtr += 4; /* skip magic number */
-+	cctx->dictID = cctx->params.fParams.noDictIDFlag ? 0 : ZSTD_readLE32(dictPtr);
-+	dictPtr += 4;
-+
-+	{
-+		size_t const hufHeaderSize = HUF_readCTable_wksp(cctx->hufTable, 255, dictPtr, dictEnd - dictPtr, cctx->tmpCounters, sizeof(cctx->tmpCounters));
-+		if (HUF_isError(hufHeaderSize))
-+			return ERROR(dictionary_corrupted);
-+		dictPtr += hufHeaderSize;
-+	}
-+
-+	{
-+		unsigned offcodeLog;
-+		size_t const offcodeHeaderSize = FSE_readNCount(offcodeNCount, &offcodeMaxValue, &offcodeLog, dictPtr, dictEnd - dictPtr);
-+		if (FSE_isError(offcodeHeaderSize))
-+			return ERROR(dictionary_corrupted);
-+		if (offcodeLog > OffFSELog)
-+			return ERROR(dictionary_corrupted);
-+		/* Defer checking offcodeMaxValue because we need to know the size of the dictionary content */
-+		CHECK_E(FSE_buildCTable_wksp(cctx->offcodeCTable, offcodeNCount, offcodeMaxValue, offcodeLog, cctx->tmpCounters, sizeof(cctx->tmpCounters)),
-+			dictionary_corrupted);
-+		dictPtr += offcodeHeaderSize;
-+	}
-+
-+	{
-+		short matchlengthNCount[MaxML + 1];
-+		unsigned matchlengthMaxValue = MaxML, matchlengthLog;
-+		size_t const matchlengthHeaderSize = FSE_readNCount(matchlengthNCount, &matchlengthMaxValue, &matchlengthLog, dictPtr, dictEnd - dictPtr);
-+		if (FSE_isError(matchlengthHeaderSize))
-+			return ERROR(dictionary_corrupted);
-+		if (matchlengthLog > MLFSELog)
-+			return ERROR(dictionary_corrupted);
-+		/* Every match length code must have non-zero probability */
-+		CHECK_F(ZSTD_checkDictNCount(matchlengthNCount, matchlengthMaxValue, MaxML));
-+		CHECK_E(
-+		    FSE_buildCTable_wksp(cctx->matchlengthCTable, matchlengthNCount, matchlengthMaxValue, matchlengthLog, cctx->tmpCounters, sizeof(cctx->tmpCounters)),
-+		    dictionary_corrupted);
-+		dictPtr += matchlengthHeaderSize;
-+	}
-+
-+	{
-+		short litlengthNCount[MaxLL + 1];
-+		unsigned litlengthMaxValue = MaxLL, litlengthLog;
-+		size_t const litlengthHeaderSize = FSE_readNCount(litlengthNCount, &litlengthMaxValue, &litlengthLog, dictPtr, dictEnd - dictPtr);
-+		if (FSE_isError(litlengthHeaderSize))
-+			return ERROR(dictionary_corrupted);
-+		if (litlengthLog > LLFSELog)
-+			return ERROR(dictionary_corrupted);
-+		/* Every literal length code must have non-zero probability */
-+		CHECK_F(ZSTD_checkDictNCount(litlengthNCount, litlengthMaxValue, MaxLL));
-+		CHECK_E(FSE_buildCTable_wksp(cctx->litlengthCTable, litlengthNCount, litlengthMaxValue, litlengthLog, cctx->tmpCounters, sizeof(cctx->tmpCounters)),
-+			dictionary_corrupted);
-+		dictPtr += litlengthHeaderSize;
-+	}
-+
-+	if (dictPtr + 12 > dictEnd)
-+		return ERROR(dictionary_corrupted);
-+	cctx->rep[0] = ZSTD_readLE32(dictPtr + 0);
-+	cctx->rep[1] = ZSTD_readLE32(dictPtr + 4);
-+	cctx->rep[2] = ZSTD_readLE32(dictPtr + 8);
-+	dictPtr += 12;
-+
-+	{
-+		size_t const dictContentSize = (size_t)(dictEnd - dictPtr);
-+		U32 offcodeMax = MaxOff;
-+		if (dictContentSize <= ((U32)-1) - 128 KB) {
-+			U32 const maxOffset = (U32)dictContentSize + 128 KB; /* The maximum offset that must be supported */
-+			offcodeMax = ZSTD_highbit32(maxOffset);		     /* Calculate minimum offset code required to represent maxOffset */
-+		}
-+		/* All offset values <= dictContentSize + 128 KB must be representable */
-+		CHECK_F(ZSTD_checkDictNCount(offcodeNCount, offcodeMaxValue, MIN(offcodeMax, MaxOff)));
-+		/* All repCodes must be <= dictContentSize and != 0*/
-+		{
-+			U32 u;
-+			for (u = 0; u < 3; u++) {
-+				if (cctx->rep[u] == 0)
-+					return ERROR(dictionary_corrupted);
-+				if (cctx->rep[u] > dictContentSize)
-+					return ERROR(dictionary_corrupted);
-+			}
-+		}
-+
-+		cctx->flagStaticTables = 1;
-+		cctx->flagStaticHufTable = HUF_repeat_valid;
-+		return ZSTD_loadDictionaryContent(cctx, dictPtr, dictContentSize);
-+	}
-+}
-+
-+/** ZSTD_compress_insertDictionary() :
-+*   @return : 0, or an error code */
-+static size_t ZSTD_compress_insertDictionary(ZSTD_CCtx *cctx, const void *dict, size_t dictSize)
-+{
-+	if ((dict == NULL) || (dictSize <= 8))
-+		return 0;
-+
-+	/* dict as pure content */
-+	if ((ZSTD_readLE32(dict) != ZSTD_DICT_MAGIC) || (cctx->forceRawDict))
-+		return ZSTD_loadDictionaryContent(cctx, dict, dictSize);
-+
-+	/* dict as zstd dictionary */
-+	return ZSTD_loadZstdDictionary(cctx, dict, dictSize);
-+}
-+
-+/*! ZSTD_compressBegin_internal() :
-+*   @return : 0, or an error code */
-+static size_t ZSTD_compressBegin_internal(ZSTD_CCtx *cctx, const void *dict, size_t dictSize, ZSTD_parameters params, U64 pledgedSrcSize)
-+{
-+	ZSTD_compResetPolicy_e const crp = dictSize ? ZSTDcrp_fullReset : ZSTDcrp_continue;
-+	CHECK_F(ZSTD_resetCCtx_advanced(cctx, params, pledgedSrcSize, crp));
-+	return ZSTD_compress_insertDictionary(cctx, dict, dictSize);
-+}
-+
-+/*! ZSTD_compressBegin_advanced() :
-+*   @return : 0, or an error code */
-+size_t ZSTD_compressBegin_advanced(ZSTD_CCtx *cctx, const void *dict, size_t dictSize, ZSTD_parameters params, unsigned long long pledgedSrcSize)
-+{
-+	/* compression parameters verification and optimization */
-+	CHECK_F(ZSTD_checkCParams(params.cParams));
-+	return ZSTD_compressBegin_internal(cctx, dict, dictSize, params, pledgedSrcSize);
-+}
-+
-+size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx *cctx, const void *dict, size_t dictSize, int compressionLevel)
-+{
-+	ZSTD_parameters const params = ZSTD_getParams(compressionLevel, 0, dictSize);
-+	return ZSTD_compressBegin_internal(cctx, dict, dictSize, params, 0);
-+}
-+
-+size_t ZSTD_compressBegin(ZSTD_CCtx *cctx, int compressionLevel) { return ZSTD_compressBegin_usingDict(cctx, NULL, 0, compressionLevel); }
-+
-+/*! ZSTD_writeEpilogue() :
-+*   Ends a frame.
-+*   @return : nb of bytes written into dst (or an error code) */
-+static size_t ZSTD_writeEpilogue(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity)
-+{
-+	BYTE *const ostart = (BYTE *)dst;
-+	BYTE *op = ostart;
-+	size_t fhSize = 0;
-+
-+	if (cctx->stage == ZSTDcs_created)
-+		return ERROR(stage_wrong); /* init missing */
-+
-+	/* special case : empty frame */
-+	if (cctx->stage == ZSTDcs_init) {
-+		fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, cctx->params, 0, 0);
-+		if (ZSTD_isError(fhSize))
-+			return fhSize;
-+		dstCapacity -= fhSize;
-+		op += fhSize;
-+		cctx->stage = ZSTDcs_ongoing;
-+	}
-+
-+	if (cctx->stage != ZSTDcs_ending) {
-+		/* write one last empty block, make it the "last" block */
-+		U32 const cBlockHeader24 = 1 /* last block */ + (((U32)bt_raw) << 1) + 0;
-+		if (dstCapacity < 4)
-+			return ERROR(dstSize_tooSmall);
-+		ZSTD_writeLE32(op, cBlockHeader24);
-+		op += ZSTD_blockHeaderSize;
-+		dstCapacity -= ZSTD_blockHeaderSize;
-+	}
-+
-+	if (cctx->params.fParams.checksumFlag) {
-+		U32 const checksum = (U32)xxh64_digest(&cctx->xxhState);
-+		if (dstCapacity < 4)
-+			return ERROR(dstSize_tooSmall);
-+		ZSTD_writeLE32(op, checksum);
-+		op += 4;
-+	}
-+
-+	cctx->stage = ZSTDcs_created; /* return to "created but no init" status */
-+	return op - ostart;
-+}
-+
-+size_t ZSTD_compressEnd(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-+{
-+	size_t endResult;
-+	size_t const cSize = ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 1, 1);
-+	if (ZSTD_isError(cSize))
-+		return cSize;
-+	endResult = ZSTD_writeEpilogue(cctx, (char *)dst + cSize, dstCapacity - cSize);
-+	if (ZSTD_isError(endResult))
-+		return endResult;
-+	return cSize + endResult;
-+}
-+
-+static size_t ZSTD_compress_internal(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, const void *dict, size_t dictSize,
-+				     ZSTD_parameters params)
-+{
-+	CHECK_F(ZSTD_compressBegin_internal(cctx, dict, dictSize, params, srcSize));
-+	return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize);
-+}
-+
-+size_t ZSTD_compress_usingDict(ZSTD_CCtx *ctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, const void *dict, size_t dictSize,
-+			       ZSTD_parameters params)
-+{
-+	return ZSTD_compress_internal(ctx, dst, dstCapacity, src, srcSize, dict, dictSize, params);
-+}
-+
-+size_t ZSTD_compressCCtx(ZSTD_CCtx *ctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, ZSTD_parameters params)
-+{
-+	return ZSTD_compress_internal(ctx, dst, dstCapacity, src, srcSize, NULL, 0, params);
-+}
-+
-+/* =====  Dictionary API  ===== */
-+
-+struct ZSTD_CDict_s {
-+	void *dictBuffer;
-+	const void *dictContent;
-+	size_t dictContentSize;
-+	ZSTD_CCtx *refContext;
-+}; /* typedef'd tp ZSTD_CDict within "zstd.h" */
-+
-+size_t ZSTD_CDictWorkspaceBound(ZSTD_compressionParameters cParams) { return ZSTD_CCtxWorkspaceBound(cParams) + ZSTD_ALIGN(sizeof(ZSTD_CDict)); }
-+
-+static ZSTD_CDict *ZSTD_createCDict_advanced(const void *dictBuffer, size_t dictSize, unsigned byReference, ZSTD_parameters params, ZSTD_customMem customMem)
-+{
-+	if (!customMem.customAlloc || !customMem.customFree)
-+		return NULL;
-+
-+	{
-+		ZSTD_CDict *const cdict = (ZSTD_CDict *)ZSTD_malloc(sizeof(ZSTD_CDict), customMem);
-+		ZSTD_CCtx *const cctx = ZSTD_createCCtx_advanced(customMem);
-+
-+		if (!cdict || !cctx) {
-+			ZSTD_free(cdict, customMem);
-+			ZSTD_freeCCtx(cctx);
-+			return NULL;
-+		}
-+
-+		if ((byReference) || (!dictBuffer) || (!dictSize)) {
-+			cdict->dictBuffer = NULL;
-+			cdict->dictContent = dictBuffer;
-+		} else {
-+			void *const internalBuffer = ZSTD_malloc(dictSize, customMem);
-+			if (!internalBuffer) {
-+				ZSTD_free(cctx, customMem);
-+				ZSTD_free(cdict, customMem);
-+				return NULL;
-+			}
-+			memcpy(internalBuffer, dictBuffer, dictSize);
-+			cdict->dictBuffer = internalBuffer;
-+			cdict->dictContent = internalBuffer;
-+		}
-+
-+		{
-+			size_t const errorCode = ZSTD_compressBegin_advanced(cctx, cdict->dictContent, dictSize, params, 0);
-+			if (ZSTD_isError(errorCode)) {
-+				ZSTD_free(cdict->dictBuffer, customMem);
-+				ZSTD_free(cdict, customMem);
-+				ZSTD_freeCCtx(cctx);
-+				return NULL;
-+			}
-+		}
-+
-+		cdict->refContext = cctx;
-+		cdict->dictContentSize = dictSize;
-+		return cdict;
-+	}
-+}
-+
-+ZSTD_CDict *ZSTD_initCDict(const void *dict, size_t dictSize, ZSTD_parameters params, void *workspace, size_t workspaceSize)
-+{
-+	ZSTD_customMem const stackMem = ZSTD_initStack(workspace, workspaceSize);
-+	return ZSTD_createCDict_advanced(dict, dictSize, 1, params, stackMem);
-+}
-+
-+size_t ZSTD_freeCDict(ZSTD_CDict *cdict)
-+{
-+	if (cdict == NULL)
-+		return 0; /* support free on NULL */
-+	{
-+		ZSTD_customMem const cMem = cdict->refContext->customMem;
-+		ZSTD_freeCCtx(cdict->refContext);
-+		ZSTD_free(cdict->dictBuffer, cMem);
-+		ZSTD_free(cdict, cMem);
-+		return 0;
-+	}
-+}
-+
-+static ZSTD_parameters ZSTD_getParamsFromCDict(const ZSTD_CDict *cdict) { return ZSTD_getParamsFromCCtx(cdict->refContext); }
-+
-+size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx *cctx, const ZSTD_CDict *cdict, unsigned long long pledgedSrcSize)
-+{
-+	if (cdict->dictContentSize)
-+		CHECK_F(ZSTD_copyCCtx(cctx, cdict->refContext, pledgedSrcSize))
-+	else {
-+		ZSTD_parameters params = cdict->refContext->params;
-+		params.fParams.contentSizeFlag = (pledgedSrcSize > 0);
-+		CHECK_F(ZSTD_compressBegin_advanced(cctx, NULL, 0, params, pledgedSrcSize));
-+	}
-+	return 0;
-+}
-+
-+/*! ZSTD_compress_usingCDict() :
-+*   Compression using a digested Dictionary.
-+*   Faster startup than ZSTD_compress_usingDict(), recommended when same dictionary is used multiple times.
-+*   Note that compression level is decided during dictionary creation */
-+size_t ZSTD_compress_usingCDict(ZSTD_CCtx *cctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, const ZSTD_CDict *cdict)
-+{
-+	CHECK_F(ZSTD_compressBegin_usingCDict(cctx, cdict, srcSize));
-+
-+	if (cdict->refContext->params.fParams.contentSizeFlag == 1) {
-+		cctx->params.fParams.contentSizeFlag = 1;
-+		cctx->frameContentSize = srcSize;
-+	} else {
-+		cctx->params.fParams.contentSizeFlag = 0;
-+	}
-+
-+	return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize);
-+}
-+
-+/* ******************************************************************
-+*  Streaming
-+********************************************************************/
-+
-+typedef enum { zcss_init, zcss_load, zcss_flush, zcss_final } ZSTD_cStreamStage;
-+
-+struct ZSTD_CStream_s {
-+	ZSTD_CCtx *cctx;
-+	ZSTD_CDict *cdictLocal;
-+	const ZSTD_CDict *cdict;
-+	char *inBuff;
-+	size_t inBuffSize;
-+	size_t inToCompress;
-+	size_t inBuffPos;
-+	size_t inBuffTarget;
-+	size_t blockSize;
-+	char *outBuff;
-+	size_t outBuffSize;
-+	size_t outBuffContentSize;
-+	size_t outBuffFlushedSize;
-+	ZSTD_cStreamStage stage;
-+	U32 checksum;
-+	U32 frameEnded;
-+	U64 pledgedSrcSize;
-+	U64 inputProcessed;
-+	ZSTD_parameters params;
-+	ZSTD_customMem customMem;
-+}; /* typedef'd to ZSTD_CStream within "zstd.h" */
-+
-+size_t ZSTD_CStreamWorkspaceBound(ZSTD_compressionParameters cParams)
-+{
-+	size_t const inBuffSize = (size_t)1 << cParams.windowLog;
-+	size_t const blockSize = MIN(ZSTD_BLOCKSIZE_ABSOLUTEMAX, inBuffSize);
-+	size_t const outBuffSize = ZSTD_compressBound(blockSize) + 1;
-+
-+	return ZSTD_CCtxWorkspaceBound(cParams) + ZSTD_ALIGN(sizeof(ZSTD_CStream)) + ZSTD_ALIGN(inBuffSize) + ZSTD_ALIGN(outBuffSize);
-+}
-+
-+ZSTD_CStream *ZSTD_createCStream_advanced(ZSTD_customMem customMem)
-+{
-+	ZSTD_CStream *zcs;
-+
-+	if (!customMem.customAlloc || !customMem.customFree)
-+		return NULL;
-+
-+	zcs = (ZSTD_CStream *)ZSTD_malloc(sizeof(ZSTD_CStream), customMem);
-+	if (zcs == NULL)
-+		return NULL;
-+	memset(zcs, 0, sizeof(ZSTD_CStream));
-+	memcpy(&zcs->customMem, &customMem, sizeof(ZSTD_customMem));
-+	zcs->cctx = ZSTD_createCCtx_advanced(customMem);
-+	if (zcs->cctx == NULL) {
-+		ZSTD_freeCStream(zcs);
-+		return NULL;
-+	}
-+	return zcs;
-+}
-+
-+size_t ZSTD_freeCStream(ZSTD_CStream *zcs)
-+{
-+	if (zcs == NULL)
-+		return 0; /* support free on NULL */
-+	{
-+		ZSTD_customMem const cMem = zcs->customMem;
-+		ZSTD_freeCCtx(zcs->cctx);
-+		zcs->cctx = NULL;
-+		ZSTD_freeCDict(zcs->cdictLocal);
-+		zcs->cdictLocal = NULL;
-+		ZSTD_free(zcs->inBuff, cMem);
-+		zcs->inBuff = NULL;
-+		ZSTD_free(zcs->outBuff, cMem);
-+		zcs->outBuff = NULL;
-+		ZSTD_free(zcs, cMem);
-+		return 0;
-+	}
-+}
-+
-+/*======   Initialization   ======*/
-+
-+size_t ZSTD_CStreamInSize(void) { return ZSTD_BLOCKSIZE_ABSOLUTEMAX; }
-+size_t ZSTD_CStreamOutSize(void) { return ZSTD_compressBound(ZSTD_BLOCKSIZE_ABSOLUTEMAX) + ZSTD_blockHeaderSize + 4 /* 32-bits hash */; }
-+
-+static size_t ZSTD_resetCStream_internal(ZSTD_CStream *zcs, unsigned long long pledgedSrcSize)
-+{
-+	if (zcs->inBuffSize == 0)
-+		return ERROR(stage_wrong); /* zcs has not been init at least once => can't reset */
-+
-+	if (zcs->cdict)
-+		CHECK_F(ZSTD_compressBegin_usingCDict(zcs->cctx, zcs->cdict, pledgedSrcSize))
-+	else
-+		CHECK_F(ZSTD_compressBegin_advanced(zcs->cctx, NULL, 0, zcs->params, pledgedSrcSize));
-+
-+	zcs->inToCompress = 0;
-+	zcs->inBuffPos = 0;
-+	zcs->inBuffTarget = zcs->blockSize;
-+	zcs->outBuffContentSize = zcs->outBuffFlushedSize = 0;
-+	zcs->stage = zcss_load;
-+	zcs->frameEnded = 0;
-+	zcs->pledgedSrcSize = pledgedSrcSize;
-+	zcs->inputProcessed = 0;
-+	return 0; /* ready to go */
-+}
-+
-+size_t ZSTD_resetCStream(ZSTD_CStream *zcs, unsigned long long pledgedSrcSize)
-+{
-+
-+	zcs->params.fParams.contentSizeFlag = (pledgedSrcSize > 0);
-+
-+	return ZSTD_resetCStream_internal(zcs, pledgedSrcSize);
-+}
-+
-+static size_t ZSTD_initCStream_advanced(ZSTD_CStream *zcs, const void *dict, size_t dictSize, ZSTD_parameters params, unsigned long long pledgedSrcSize)
-+{
-+	/* allocate buffers */
-+	{
-+		size_t const neededInBuffSize = (size_t)1 << params.cParams.windowLog;
-+		if (zcs->inBuffSize < neededInBuffSize) {
-+			zcs->inBuffSize = neededInBuffSize;
-+			ZSTD_free(zcs->inBuff, zcs->customMem);
-+			zcs->inBuff = (char *)ZSTD_malloc(neededInBuffSize, zcs->customMem);
-+			if (zcs->inBuff == NULL)
-+				return ERROR(memory_allocation);
-+		}
-+		zcs->blockSize = MIN(ZSTD_BLOCKSIZE_ABSOLUTEMAX, neededInBuffSize);
-+	}
-+	if (zcs->outBuffSize < ZSTD_compressBound(zcs->blockSize) + 1) {
-+		zcs->outBuffSize = ZSTD_compressBound(zcs->blockSize) + 1;
-+		ZSTD_free(zcs->outBuff, zcs->customMem);
-+		zcs->outBuff = (char *)ZSTD_malloc(zcs->outBuffSize, zcs->customMem);
-+		if (zcs->outBuff == NULL)
-+			return ERROR(memory_allocation);
-+	}
-+
-+	if (dict && dictSize >= 8) {
-+		ZSTD_freeCDict(zcs->cdictLocal);
-+		zcs->cdictLocal = ZSTD_createCDict_advanced(dict, dictSize, 0, params, zcs->customMem);
-+		if (zcs->cdictLocal == NULL)
-+			return ERROR(memory_allocation);
-+		zcs->cdict = zcs->cdictLocal;
-+	} else
-+		zcs->cdict = NULL;
-+
-+	zcs->checksum = params.fParams.checksumFlag > 0;
-+	zcs->params = params;
-+
-+	return ZSTD_resetCStream_internal(zcs, pledgedSrcSize);
-+}
-+
-+ZSTD_CStream *ZSTD_initCStream(ZSTD_parameters params, unsigned long long pledgedSrcSize, void *workspace, size_t workspaceSize)
-+{
-+	ZSTD_customMem const stackMem = ZSTD_initStack(workspace, workspaceSize);
-+	ZSTD_CStream *const zcs = ZSTD_createCStream_advanced(stackMem);
-+	if (zcs) {
-+		size_t const code = ZSTD_initCStream_advanced(zcs, NULL, 0, params, pledgedSrcSize);
-+		if (ZSTD_isError(code)) {
-+			return NULL;
-+		}
-+	}
-+	return zcs;
-+}
-+
-+ZSTD_CStream *ZSTD_initCStream_usingCDict(const ZSTD_CDict *cdict, unsigned long long pledgedSrcSize, void *workspace, size_t workspaceSize)
-+{
-+	ZSTD_parameters const params = ZSTD_getParamsFromCDict(cdict);
-+	ZSTD_CStream *const zcs = ZSTD_initCStream(params, pledgedSrcSize, workspace, workspaceSize);
-+	if (zcs) {
-+		zcs->cdict = cdict;
-+		if (ZSTD_isError(ZSTD_resetCStream_internal(zcs, pledgedSrcSize))) {
-+			return NULL;
-+		}
-+	}
-+	return zcs;
-+}
-+
-+/*======   Compression   ======*/
-+
-+typedef enum { zsf_gather, zsf_flush, zsf_end } ZSTD_flush_e;
-+
-+ZSTD_STATIC size_t ZSTD_limitCopy(void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-+{
-+	size_t const length = MIN(dstCapacity, srcSize);
-+	memcpy(dst, src, length);
-+	return length;
-+}
-+
-+static size_t ZSTD_compressStream_generic(ZSTD_CStream *zcs, void *dst, size_t *dstCapacityPtr, const void *src, size_t *srcSizePtr, ZSTD_flush_e const flush)
-+{
-+	U32 someMoreWork = 1;
-+	const char *const istart = (const char *)src;
-+	const char *const iend = istart + *srcSizePtr;
-+	const char *ip = istart;
-+	char *const ostart = (char *)dst;
-+	char *const oend = ostart + *dstCapacityPtr;
-+	char *op = ostart;
-+
-+	while (someMoreWork) {
-+		switch (zcs->stage) {
-+		case zcss_init:
-+			return ERROR(init_missing); /* call ZBUFF_compressInit() first ! */
-+
-+		case zcss_load:
-+			/* complete inBuffer */
-+			{
-+				size_t const toLoad = zcs->inBuffTarget - zcs->inBuffPos;
-+				size_t const loaded = ZSTD_limitCopy(zcs->inBuff + zcs->inBuffPos, toLoad, ip, iend - ip);
-+				zcs->inBuffPos += loaded;
-+				ip += loaded;
-+				if ((zcs->inBuffPos == zcs->inToCompress) || (!flush && (toLoad != loaded))) {
-+					someMoreWork = 0;
-+					break; /* not enough input to get a full block : stop there, wait for more */
-+				}
-+			}
-+			/* compress curr block (note : this stage cannot be stopped in the middle) */
-+			{
-+				void *cDst;
-+				size_t cSize;
-+				size_t const iSize = zcs->inBuffPos - zcs->inToCompress;
-+				size_t oSize = oend - op;
-+				if (oSize >= ZSTD_compressBound(iSize))
-+					cDst = op; /* compress directly into output buffer (avoid flush stage) */
-+				else
-+					cDst = zcs->outBuff, oSize = zcs->outBuffSize;
-+				cSize = (flush == zsf_end) ? ZSTD_compressEnd(zcs->cctx, cDst, oSize, zcs->inBuff + zcs->inToCompress, iSize)
-+							   : ZSTD_compressContinue(zcs->cctx, cDst, oSize, zcs->inBuff + zcs->inToCompress, iSize);
-+				if (ZSTD_isError(cSize))
-+					return cSize;
-+				if (flush == zsf_end)
-+					zcs->frameEnded = 1;
-+				/* prepare next block */
-+				zcs->inBuffTarget = zcs->inBuffPos + zcs->blockSize;
-+				if (zcs->inBuffTarget > zcs->inBuffSize)
-+					zcs->inBuffPos = 0, zcs->inBuffTarget = zcs->blockSize; /* note : inBuffSize >= blockSize */
-+				zcs->inToCompress = zcs->inBuffPos;
-+				if (cDst == op) {
-+					op += cSize;
-+					break;
-+				} /* no need to flush */
-+				zcs->outBuffContentSize = cSize;
-+				zcs->outBuffFlushedSize = 0;
-+				zcs->stage = zcss_flush; /* pass-through to flush stage */
-+			}
-+
-+		case zcss_flush: {
-+			size_t const toFlush = zcs->outBuffContentSize - zcs->outBuffFlushedSize;
-+			size_t const flushed = ZSTD_limitCopy(op, oend - op, zcs->outBuff + zcs->outBuffFlushedSize, toFlush);
-+			op += flushed;
-+			zcs->outBuffFlushedSize += flushed;
-+			if (toFlush != flushed) {
-+				someMoreWork = 0;
-+				break;
-+			} /* dst too small to store flushed data : stop there */
-+			zcs->outBuffContentSize = zcs->outBuffFlushedSize = 0;
-+			zcs->stage = zcss_load;
-+			break;
-+		}
-+
-+		case zcss_final:
-+			someMoreWork = 0; /* do nothing */
-+			break;
-+
-+		default:
-+			return ERROR(GENERIC); /* impossible */
-+		}
-+	}
-+
-+	*srcSizePtr = ip - istart;
-+	*dstCapacityPtr = op - ostart;
-+	zcs->inputProcessed += *srcSizePtr;
-+	if (zcs->frameEnded)
-+		return 0;
-+	{
-+		size_t hintInSize = zcs->inBuffTarget - zcs->inBuffPos;
-+		if (hintInSize == 0)
-+			hintInSize = zcs->blockSize;
-+		return hintInSize;
-+	}
-+}
-+
-+size_t ZSTD_compressStream(ZSTD_CStream *zcs, ZSTD_outBuffer *output, ZSTD_inBuffer *input)
-+{
-+	size_t sizeRead = input->size - input->pos;
-+	size_t sizeWritten = output->size - output->pos;
-+	size_t const result =
-+	    ZSTD_compressStream_generic(zcs, (char *)(output->dst) + output->pos, &sizeWritten, (const char *)(input->src) + input->pos, &sizeRead, zsf_gather);
-+	input->pos += sizeRead;
-+	output->pos += sizeWritten;
-+	return result;
-+}
-+
-+/*======   Finalize   ======*/
-+
-+/*! ZSTD_flushStream() :
-+*   @return : amount of data remaining to flush */
-+size_t ZSTD_flushStream(ZSTD_CStream *zcs, ZSTD_outBuffer *output)
-+{
-+	size_t srcSize = 0;
-+	size_t sizeWritten = output->size - output->pos;
-+	size_t const result = ZSTD_compressStream_generic(zcs, (char *)(output->dst) + output->pos, &sizeWritten, &srcSize,
-+							  &srcSize, /* use a valid src address instead of NULL */
-+							  zsf_flush);
-+	output->pos += sizeWritten;
-+	if (ZSTD_isError(result))
-+		return result;
-+	return zcs->outBuffContentSize - zcs->outBuffFlushedSize; /* remaining to flush */
-+}
-+
-+size_t ZSTD_endStream(ZSTD_CStream *zcs, ZSTD_outBuffer *output)
-+{
-+	BYTE *const ostart = (BYTE *)(output->dst) + output->pos;
-+	BYTE *const oend = (BYTE *)(output->dst) + output->size;
-+	BYTE *op = ostart;
-+
-+	if ((zcs->pledgedSrcSize) && (zcs->inputProcessed != zcs->pledgedSrcSize))
-+		return ERROR(srcSize_wrong); /* pledgedSrcSize not respected */
-+
-+	if (zcs->stage != zcss_final) {
-+		/* flush whatever remains */
-+		size_t srcSize = 0;
-+		size_t sizeWritten = output->size - output->pos;
-+		size_t const notEnded =
-+		    ZSTD_compressStream_generic(zcs, ostart, &sizeWritten, &srcSize, &srcSize, zsf_end); /* use a valid src address instead of NULL */
-+		size_t const remainingToFlush = zcs->outBuffContentSize - zcs->outBuffFlushedSize;
-+		op += sizeWritten;
-+		if (remainingToFlush) {
-+			output->pos += sizeWritten;
-+			return remainingToFlush + ZSTD_BLOCKHEADERSIZE /* final empty block */ + (zcs->checksum * 4);
-+		}
-+		/* create epilogue */
-+		zcs->stage = zcss_final;
-+		zcs->outBuffContentSize = !notEnded ? 0 : ZSTD_compressEnd(zcs->cctx, zcs->outBuff, zcs->outBuffSize, NULL,
-+									   0); /* write epilogue, including final empty block, into outBuff */
-+	}
-+
-+	/* flush epilogue */
-+	{
-+		size_t const toFlush = zcs->outBuffContentSize - zcs->outBuffFlushedSize;
-+		size_t const flushed = ZSTD_limitCopy(op, oend - op, zcs->outBuff + zcs->outBuffFlushedSize, toFlush);
-+		op += flushed;
-+		zcs->outBuffFlushedSize += flushed;
-+		output->pos += op - ostart;
-+		if (toFlush == flushed)
-+			zcs->stage = zcss_init; /* end reached */
-+		return toFlush - flushed;
-+	}
-+}
-+
-+/*-=====  Pre-defined compression levels  =====-*/
-+
-+#define ZSTD_DEFAULT_CLEVEL 1
-+#define ZSTD_MAX_CLEVEL 22
-+int ZSTD_maxCLevel(void) { return ZSTD_MAX_CLEVEL; }
-+
-+static const ZSTD_compressionParameters ZSTD_defaultCParameters[4][ZSTD_MAX_CLEVEL + 1] = {
-+    {
-+	/* "default" */
-+	/* W,  C,  H,  S,  L, TL, strat */
-+	{18, 12, 12, 1, 7, 16, ZSTD_fast},    /* level  0 - never used */
-+	{19, 13, 14, 1, 7, 16, ZSTD_fast},    /* level  1 */
-+	{19, 15, 16, 1, 6, 16, ZSTD_fast},    /* level  2 */
-+	{20, 16, 17, 1, 5, 16, ZSTD_dfast},   /* level  3.*/
-+	{20, 18, 18, 1, 5, 16, ZSTD_dfast},   /* level  4.*/
-+	{20, 15, 18, 3, 5, 16, ZSTD_greedy},  /* level  5 */
-+	{21, 16, 19, 2, 5, 16, ZSTD_lazy},    /* level  6 */
-+	{21, 17, 20, 3, 5, 16, ZSTD_lazy},    /* level  7 */
-+	{21, 18, 20, 3, 5, 16, ZSTD_lazy2},   /* level  8 */
-+	{21, 20, 20, 3, 5, 16, ZSTD_lazy2},   /* level  9 */
-+	{21, 19, 21, 4, 5, 16, ZSTD_lazy2},   /* level 10 */
-+	{22, 20, 22, 4, 5, 16, ZSTD_lazy2},   /* level 11 */
-+	{22, 20, 22, 5, 5, 16, ZSTD_lazy2},   /* level 12 */
-+	{22, 21, 22, 5, 5, 16, ZSTD_lazy2},   /* level 13 */
-+	{22, 21, 22, 6, 5, 16, ZSTD_lazy2},   /* level 14 */
-+	{22, 21, 21, 5, 5, 16, ZSTD_btlazy2}, /* level 15 */
-+	{23, 22, 22, 5, 5, 16, ZSTD_btlazy2}, /* level 16 */
-+	{23, 21, 22, 4, 5, 24, ZSTD_btopt},   /* level 17 */
-+	{23, 23, 22, 6, 5, 32, ZSTD_btopt},   /* level 18 */
-+	{23, 23, 22, 6, 3, 48, ZSTD_btopt},   /* level 19 */
-+	{25, 25, 23, 7, 3, 64, ZSTD_btopt2},  /* level 20 */
-+	{26, 26, 23, 7, 3, 256, ZSTD_btopt2}, /* level 21 */
-+	{27, 27, 25, 9, 3, 512, ZSTD_btopt2}, /* level 22 */
-+    },
-+    {
-+	/* for srcSize <= 256 KB */
-+	/* W,  C,  H,  S,  L,  T, strat */
-+	{0, 0, 0, 0, 0, 0, ZSTD_fast},	 /* level  0 - not used */
-+	{18, 13, 14, 1, 6, 8, ZSTD_fast},      /* level  1 */
-+	{18, 14, 13, 1, 5, 8, ZSTD_dfast},     /* level  2 */
-+	{18, 16, 15, 1, 5, 8, ZSTD_dfast},     /* level  3 */
-+	{18, 15, 17, 1, 5, 8, ZSTD_greedy},    /* level  4.*/
-+	{18, 16, 17, 4, 5, 8, ZSTD_greedy},    /* level  5.*/
-+	{18, 16, 17, 3, 5, 8, ZSTD_lazy},      /* level  6.*/
-+	{18, 17, 17, 4, 4, 8, ZSTD_lazy},      /* level  7 */
-+	{18, 17, 17, 4, 4, 8, ZSTD_lazy2},     /* level  8 */
-+	{18, 17, 17, 5, 4, 8, ZSTD_lazy2},     /* level  9 */
-+	{18, 17, 17, 6, 4, 8, ZSTD_lazy2},     /* level 10 */
-+	{18, 18, 17, 6, 4, 8, ZSTD_lazy2},     /* level 11.*/
-+	{18, 18, 17, 7, 4, 8, ZSTD_lazy2},     /* level 12.*/
-+	{18, 19, 17, 6, 4, 8, ZSTD_btlazy2},   /* level 13 */
-+	{18, 18, 18, 4, 4, 16, ZSTD_btopt},    /* level 14.*/
-+	{18, 18, 18, 4, 3, 16, ZSTD_btopt},    /* level 15.*/
-+	{18, 19, 18, 6, 3, 32, ZSTD_btopt},    /* level 16.*/
-+	{18, 19, 18, 8, 3, 64, ZSTD_btopt},    /* level 17.*/
-+	{18, 19, 18, 9, 3, 128, ZSTD_btopt},   /* level 18.*/
-+	{18, 19, 18, 10, 3, 256, ZSTD_btopt},  /* level 19.*/
-+	{18, 19, 18, 11, 3, 512, ZSTD_btopt2}, /* level 20.*/
-+	{18, 19, 18, 12, 3, 512, ZSTD_btopt2}, /* level 21.*/
-+	{18, 19, 18, 13, 3, 512, ZSTD_btopt2}, /* level 22.*/
-+    },
-+    {
-+	/* for srcSize <= 128 KB */
-+	/* W,  C,  H,  S,  L,  T, strat */
-+	{17, 12, 12, 1, 7, 8, ZSTD_fast},      /* level  0 - not used */
-+	{17, 12, 13, 1, 6, 8, ZSTD_fast},      /* level  1 */
-+	{17, 13, 16, 1, 5, 8, ZSTD_fast},      /* level  2 */
-+	{17, 16, 16, 2, 5, 8, ZSTD_dfast},     /* level  3 */
-+	{17, 13, 15, 3, 4, 8, ZSTD_greedy},    /* level  4 */
-+	{17, 15, 17, 4, 4, 8, ZSTD_greedy},    /* level  5 */
-+	{17, 16, 17, 3, 4, 8, ZSTD_lazy},      /* level  6 */
-+	{17, 15, 17, 4, 4, 8, ZSTD_lazy2},     /* level  7 */
-+	{17, 17, 17, 4, 4, 8, ZSTD_lazy2},     /* level  8 */
-+	{17, 17, 17, 5, 4, 8, ZSTD_lazy2},     /* level  9 */
-+	{17, 17, 17, 6, 4, 8, ZSTD_lazy2},     /* level 10 */
-+	{17, 17, 17, 7, 4, 8, ZSTD_lazy2},     /* level 11 */
-+	{17, 17, 17, 8, 4, 8, ZSTD_lazy2},     /* level 12 */
-+	{17, 18, 17, 6, 4, 8, ZSTD_btlazy2},   /* level 13.*/
-+	{17, 17, 17, 7, 3, 8, ZSTD_btopt},     /* level 14.*/
-+	{17, 17, 17, 7, 3, 16, ZSTD_btopt},    /* level 15.*/
-+	{17, 18, 17, 7, 3, 32, ZSTD_btopt},    /* level 16.*/
-+	{17, 18, 17, 7, 3, 64, ZSTD_btopt},    /* level 17.*/
-+	{17, 18, 17, 7, 3, 256, ZSTD_btopt},   /* level 18.*/
-+	{17, 18, 17, 8, 3, 256, ZSTD_btopt},   /* level 19.*/
-+	{17, 18, 17, 9, 3, 256, ZSTD_btopt2},  /* level 20.*/
-+	{17, 18, 17, 10, 3, 256, ZSTD_btopt2}, /* level 21.*/
-+	{17, 18, 17, 11, 3, 512, ZSTD_btopt2}, /* level 22.*/
-+    },
-+    {
-+	/* for srcSize <= 16 KB */
-+	/* W,  C,  H,  S,  L,  T, strat */
-+	{14, 12, 12, 1, 7, 6, ZSTD_fast},      /* level  0 - not used */
-+	{14, 14, 14, 1, 6, 6, ZSTD_fast},      /* level  1 */
-+	{14, 14, 14, 1, 4, 6, ZSTD_fast},      /* level  2 */
-+	{14, 14, 14, 1, 4, 6, ZSTD_dfast},     /* level  3.*/
-+	{14, 14, 14, 4, 4, 6, ZSTD_greedy},    /* level  4.*/
-+	{14, 14, 14, 3, 4, 6, ZSTD_lazy},      /* level  5.*/
-+	{14, 14, 14, 4, 4, 6, ZSTD_lazy2},     /* level  6 */
-+	{14, 14, 14, 5, 4, 6, ZSTD_lazy2},     /* level  7 */
-+	{14, 14, 14, 6, 4, 6, ZSTD_lazy2},     /* level  8.*/
-+	{14, 15, 14, 6, 4, 6, ZSTD_btlazy2},   /* level  9.*/
-+	{14, 15, 14, 3, 3, 6, ZSTD_btopt},     /* level 10.*/
-+	{14, 15, 14, 6, 3, 8, ZSTD_btopt},     /* level 11.*/
-+	{14, 15, 14, 6, 3, 16, ZSTD_btopt},    /* level 12.*/
-+	{14, 15, 14, 6, 3, 24, ZSTD_btopt},    /* level 13.*/
-+	{14, 15, 15, 6, 3, 48, ZSTD_btopt},    /* level 14.*/
-+	{14, 15, 15, 6, 3, 64, ZSTD_btopt},    /* level 15.*/
-+	{14, 15, 15, 6, 3, 96, ZSTD_btopt},    /* level 16.*/
-+	{14, 15, 15, 6, 3, 128, ZSTD_btopt},   /* level 17.*/
-+	{14, 15, 15, 6, 3, 256, ZSTD_btopt},   /* level 18.*/
-+	{14, 15, 15, 7, 3, 256, ZSTD_btopt},   /* level 19.*/
-+	{14, 15, 15, 8, 3, 256, ZSTD_btopt2},  /* level 20.*/
-+	{14, 15, 15, 9, 3, 256, ZSTD_btopt2},  /* level 21.*/
-+	{14, 15, 15, 10, 3, 256, ZSTD_btopt2}, /* level 22.*/
-+    },
-+};
-+
-+/*! ZSTD_getCParams() :
-+*   @return ZSTD_compressionParameters structure for a selected compression level, `srcSize` and `dictSize`.
-+*   Size values are optional, provide 0 if not known or unused */
-+ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long long srcSize, size_t dictSize)
-+{
-+	ZSTD_compressionParameters cp;
-+	size_t const addedSize = srcSize ? 0 : 500;
-+	U64 const rSize = srcSize + dictSize ? srcSize + dictSize + addedSize : (U64)-1;
-+	U32 const tableID = (rSize <= 256 KB) + (rSize <= 128 KB) + (rSize <= 16 KB); /* intentional underflow for srcSizeHint == 0 */
-+	if (compressionLevel <= 0)
-+		compressionLevel = ZSTD_DEFAULT_CLEVEL; /* 0 == default; no negative compressionLevel yet */
-+	if (compressionLevel > ZSTD_MAX_CLEVEL)
-+		compressionLevel = ZSTD_MAX_CLEVEL;
-+	cp = ZSTD_defaultCParameters[tableID][compressionLevel];
-+	if (ZSTD_32bits()) { /* auto-correction, for 32-bits mode */
-+		if (cp.windowLog > ZSTD_WINDOWLOG_MAX)
-+			cp.windowLog = ZSTD_WINDOWLOG_MAX;
-+		if (cp.chainLog > ZSTD_CHAINLOG_MAX)
-+			cp.chainLog = ZSTD_CHAINLOG_MAX;
-+		if (cp.hashLog > ZSTD_HASHLOG_MAX)
-+			cp.hashLog = ZSTD_HASHLOG_MAX;
-+	}
-+	cp = ZSTD_adjustCParams(cp, srcSize, dictSize);
-+	return cp;
-+}
-+
-+/*! ZSTD_getParams() :
-+*   same as ZSTD_getCParams(), but @return a `ZSTD_parameters` object (instead of `ZSTD_compressionParameters`).
-+*   All fields of `ZSTD_frameParameters` are set to default (0) */
-+ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long srcSize, size_t dictSize)
-+{
-+	ZSTD_parameters params;
-+	ZSTD_compressionParameters const cParams = ZSTD_getCParams(compressionLevel, srcSize, dictSize);
-+	memset(&params, 0, sizeof(params));
-+	params.cParams = cParams;
-+	return params;
-+}
-+
-+EXPORT_SYMBOL(ZSTD_maxCLevel);
-+EXPORT_SYMBOL(ZSTD_compressBound);
-+
-+EXPORT_SYMBOL(ZSTD_CCtxWorkspaceBound);
-+EXPORT_SYMBOL(ZSTD_initCCtx);
-+EXPORT_SYMBOL(ZSTD_compressCCtx);
-+EXPORT_SYMBOL(ZSTD_compress_usingDict);
-+
-+EXPORT_SYMBOL(ZSTD_CDictWorkspaceBound);
-+EXPORT_SYMBOL(ZSTD_initCDict);
-+EXPORT_SYMBOL(ZSTD_compress_usingCDict);
-+
-+EXPORT_SYMBOL(ZSTD_CStreamWorkspaceBound);
-+EXPORT_SYMBOL(ZSTD_initCStream);
-+EXPORT_SYMBOL(ZSTD_initCStream_usingCDict);
-+EXPORT_SYMBOL(ZSTD_resetCStream);
-+EXPORT_SYMBOL(ZSTD_compressStream);
-+EXPORT_SYMBOL(ZSTD_flushStream);
-+EXPORT_SYMBOL(ZSTD_endStream);
-+EXPORT_SYMBOL(ZSTD_CStreamInSize);
-+EXPORT_SYMBOL(ZSTD_CStreamOutSize);
-+
-+EXPORT_SYMBOL(ZSTD_getCParams);
-+EXPORT_SYMBOL(ZSTD_getParams);
-+EXPORT_SYMBOL(ZSTD_checkCParams);
-+EXPORT_SYMBOL(ZSTD_adjustCParams);
-+
-+EXPORT_SYMBOL(ZSTD_compressBegin);
-+EXPORT_SYMBOL(ZSTD_compressBegin_usingDict);
-+EXPORT_SYMBOL(ZSTD_compressBegin_advanced);
-+EXPORT_SYMBOL(ZSTD_copyCCtx);
-+EXPORT_SYMBOL(ZSTD_compressBegin_usingCDict);
-+EXPORT_SYMBOL(ZSTD_compressContinue);
-+EXPORT_SYMBOL(ZSTD_compressEnd);
-+
-+EXPORT_SYMBOL(ZSTD_getBlockSizeMax);
-+EXPORT_SYMBOL(ZSTD_compressBlock);
-+
-+MODULE_LICENSE("Dual BSD/GPL");
-+MODULE_DESCRIPTION("Zstd Compressor");
-diff --git a/lib/zstd/decompress.c b/lib/zstd/decompress.c
-new file mode 100644
-index 0000000..b178467
---- /dev/null
-+++ b/lib/zstd/decompress.c
-@@ -0,0 +1,2528 @@
-+/**
-+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
-+ * All rights reserved.
-+ *
-+ * This source code is licensed under the BSD-style license found in the
-+ * LICENSE file in the root directory of https://github.com/facebook/zstd.
-+ * An additional grant of patent rights can be found in the PATENTS file in the
-+ * same directory.
-+ *
-+ * This program is free software; you can redistribute it and/or modify it under
-+ * the terms of the GNU General Public License version 2 as published by the
-+ * Free Software Foundation. This program is dual-licensed; you may select
-+ * either version 2 of the GNU General Public License ("GPL") or BSD license
-+ * ("BSD").
-+ */
-+
-+/* ***************************************************************
-+*  Tuning parameters
-+*****************************************************************/
-+/*!
-+*  MAXWINDOWSIZE_DEFAULT :
-+*  maximum window size accepted by DStream, by default.
-+*  Frames requiring more memory will be rejected.
-+*/
-+#ifndef ZSTD_MAXWINDOWSIZE_DEFAULT
-+#define ZSTD_MAXWINDOWSIZE_DEFAULT ((1 << ZSTD_WINDOWLOG_MAX) + 1) /* defined within zstd.h */
-+#endif
-+
-+/*-*******************************************************
-+*  Dependencies
-+*********************************************************/
-+#include "fse.h"
-+#include "huf.h"
-+#include "mem.h" /* low level memory routines */
-+#include "zstd_internal.h"
-+#include <linux/kernel.h>
-+#include <linux/module.h>
-+#include <linux/string.h> /* memcpy, memmove, memset */
-+
-+#define ZSTD_PREFETCH(ptr) __builtin_prefetch(ptr, 0, 0)
-+
-+/*-*************************************
-+*  Macros
-+***************************************/
-+#define ZSTD_isError ERR_isError /* for inlining */
-+#define FSE_isError ERR_isError
-+#define HUF_isError ERR_isError
-+
-+/*_*******************************************************
-+*  Memory operations
-+**********************************************************/
-+static void ZSTD_copy4(void *dst, const void *src) { memcpy(dst, src, 4); }
-+
-+/*-*************************************************************
-+*   Context management
-+***************************************************************/
-+typedef enum {
-+	ZSTDds_getFrameHeaderSize,
-+	ZSTDds_decodeFrameHeader,
-+	ZSTDds_decodeBlockHeader,
-+	ZSTDds_decompressBlock,
-+	ZSTDds_decompressLastBlock,
-+	ZSTDds_checkChecksum,
-+	ZSTDds_decodeSkippableHeader,
-+	ZSTDds_skipFrame
-+} ZSTD_dStage;
-+
-+typedef struct {
-+	FSE_DTable LLTable[FSE_DTABLE_SIZE_U32(LLFSELog)];
-+	FSE_DTable OFTable[FSE_DTABLE_SIZE_U32(OffFSELog)];
-+	FSE_DTable MLTable[FSE_DTABLE_SIZE_U32(MLFSELog)];
-+	HUF_DTable hufTable[HUF_DTABLE_SIZE(HufLog)]; /* can accommodate HUF_decompress4X */
-+	U64 workspace[HUF_DECOMPRESS_WORKSPACE_SIZE_U32 / 2];
-+	U32 rep[ZSTD_REP_NUM];
-+} ZSTD_entropyTables_t;
-+
-+struct ZSTD_DCtx_s {
-+	const FSE_DTable *LLTptr;
-+	const FSE_DTable *MLTptr;
-+	const FSE_DTable *OFTptr;
-+	const HUF_DTable *HUFptr;
-+	ZSTD_entropyTables_t entropy;
-+	const void *previousDstEnd; /* detect continuity */
-+	const void *base;	   /* start of curr segment */
-+	const void *vBase;	  /* virtual start of previous segment if it was just before curr one */
-+	const void *dictEnd;	/* end of previous segment */
-+	size_t expected;
-+	ZSTD_frameParams fParams;
-+	blockType_e bType; /* used in ZSTD_decompressContinue(), to transfer blockType between header decoding and block decoding stages */
-+	ZSTD_dStage stage;
-+	U32 litEntropy;
-+	U32 fseEntropy;
-+	struct xxh64_state xxhState;
-+	size_t headerSize;
-+	U32 dictID;
-+	const BYTE *litPtr;
-+	ZSTD_customMem customMem;
-+	size_t litSize;
-+	size_t rleSize;
-+	BYTE litBuffer[ZSTD_BLOCKSIZE_ABSOLUTEMAX + WILDCOPY_OVERLENGTH];
-+	BYTE headerBuffer[ZSTD_FRAMEHEADERSIZE_MAX];
-+}; /* typedef'd to ZSTD_DCtx within "zstd.h" */
-+
-+size_t ZSTD_DCtxWorkspaceBound(void) { return ZSTD_ALIGN(sizeof(ZSTD_stack)) + ZSTD_ALIGN(sizeof(ZSTD_DCtx)); }
-+
-+size_t ZSTD_decompressBegin(ZSTD_DCtx *dctx)
-+{
-+	dctx->expected = ZSTD_frameHeaderSize_prefix;
-+	dctx->stage = ZSTDds_getFrameHeaderSize;
-+	dctx->previousDstEnd = NULL;
-+	dctx->base = NULL;
-+	dctx->vBase = NULL;
-+	dctx->dictEnd = NULL;
-+	dctx->entropy.hufTable[0] = (HUF_DTable)((HufLog)*0x1000001); /* cover both little and big endian */
-+	dctx->litEntropy = dctx->fseEntropy = 0;
-+	dctx->dictID = 0;
-+	ZSTD_STATIC_ASSERT(sizeof(dctx->entropy.rep) == sizeof(repStartValue));
-+	memcpy(dctx->entropy.rep, repStartValue, sizeof(repStartValue)); /* initial repcodes */
-+	dctx->LLTptr = dctx->entropy.LLTable;
-+	dctx->MLTptr = dctx->entropy.MLTable;
-+	dctx->OFTptr = dctx->entropy.OFTable;
-+	dctx->HUFptr = dctx->entropy.hufTable;
-+	return 0;
-+}
-+
-+ZSTD_DCtx *ZSTD_createDCtx_advanced(ZSTD_customMem customMem)
-+{
-+	ZSTD_DCtx *dctx;
-+
-+	if (!customMem.customAlloc || !customMem.customFree)
-+		return NULL;
-+
-+	dctx = (ZSTD_DCtx *)ZSTD_malloc(sizeof(ZSTD_DCtx), customMem);
-+	if (!dctx)
-+		return NULL;
-+	memcpy(&dctx->customMem, &customMem, sizeof(customMem));
-+	ZSTD_decompressBegin(dctx);
-+	return dctx;
-+}
-+
-+ZSTD_DCtx *ZSTD_initDCtx(void *workspace, size_t workspaceSize)
-+{
-+	ZSTD_customMem const stackMem = ZSTD_initStack(workspace, workspaceSize);
-+	return ZSTD_createDCtx_advanced(stackMem);
-+}
-+
-+size_t ZSTD_freeDCtx(ZSTD_DCtx *dctx)
-+{
-+	if (dctx == NULL)
-+		return 0; /* support free on NULL */
-+	ZSTD_free(dctx, dctx->customMem);
-+	return 0; /* reserved as a potential error code in the future */
-+}
-+
-+void ZSTD_copyDCtx(ZSTD_DCtx *dstDCtx, const ZSTD_DCtx *srcDCtx)
-+{
-+	size_t const workSpaceSize = (ZSTD_BLOCKSIZE_ABSOLUTEMAX + WILDCOPY_OVERLENGTH) + ZSTD_frameHeaderSize_max;
-+	memcpy(dstDCtx, srcDCtx, sizeof(ZSTD_DCtx) - workSpaceSize); /* no need to copy workspace */
-+}
-+
-+static void ZSTD_refDDict(ZSTD_DCtx *dstDCtx, const ZSTD_DDict *ddict);
-+
-+/*-*************************************************************
-+*   Decompression section
-+***************************************************************/
-+
-+/*! ZSTD_isFrame() :
-+ *  Tells if the content of `buffer` starts with a valid Frame Identifier.
-+ *  Note : Frame Identifier is 4 bytes. If `size < 4`, @return will always be 0.
-+ *  Note 2 : Legacy Frame Identifiers are considered valid only if Legacy Support is enabled.
-+ *  Note 3 : Skippable Frame Identifiers are considered valid. */
-+unsigned ZSTD_isFrame(const void *buffer, size_t size)
-+{
-+	if (size < 4)
-+		return 0;
-+	{
-+		U32 const magic = ZSTD_readLE32(buffer);
-+		if (magic == ZSTD_MAGICNUMBER)
-+			return 1;
-+		if ((magic & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START)
-+			return 1;
-+	}
-+	return 0;
-+}
-+
-+/** ZSTD_frameHeaderSize() :
-+*   srcSize must be >= ZSTD_frameHeaderSize_prefix.
-+*   @return : size of the Frame Header */
-+static size_t ZSTD_frameHeaderSize(const void *src, size_t srcSize)
-+{
-+	if (srcSize < ZSTD_frameHeaderSize_prefix)
-+		return ERROR(srcSize_wrong);
-+	{
-+		BYTE const fhd = ((const BYTE *)src)[4];
-+		U32 const dictID = fhd & 3;
-+		U32 const singleSegment = (fhd >> 5) & 1;
-+		U32 const fcsId = fhd >> 6;
-+		return ZSTD_frameHeaderSize_prefix + !singleSegment + ZSTD_did_fieldSize[dictID] + ZSTD_fcs_fieldSize[fcsId] + (singleSegment && !fcsId);
-+	}
-+}
-+
-+/** ZSTD_getFrameParams() :
-+*   decode Frame Header, or require larger `srcSize`.
-+*   @return : 0, `fparamsPtr` is correctly filled,
-+*            >0, `srcSize` is too small, result is expected `srcSize`,
-+*             or an error code, which can be tested using ZSTD_isError() */
-+size_t ZSTD_getFrameParams(ZSTD_frameParams *fparamsPtr, const void *src, size_t srcSize)
-+{
-+	const BYTE *ip = (const BYTE *)src;
-+
-+	if (srcSize < ZSTD_frameHeaderSize_prefix)
-+		return ZSTD_frameHeaderSize_prefix;
-+	if (ZSTD_readLE32(src) != ZSTD_MAGICNUMBER) {
-+		if ((ZSTD_readLE32(src) & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) {
-+			if (srcSize < ZSTD_skippableHeaderSize)
-+				return ZSTD_skippableHeaderSize; /* magic number + skippable frame length */
-+			memset(fparamsPtr, 0, sizeof(*fparamsPtr));
-+			fparamsPtr->frameContentSize = ZSTD_readLE32((const char *)src + 4);
-+			fparamsPtr->windowSize = 0; /* windowSize==0 means a frame is skippable */
-+			return 0;
-+		}
-+		return ERROR(prefix_unknown);
-+	}
-+
-+	/* ensure there is enough `srcSize` to fully read/decode frame header */
-+	{
-+		size_t const fhsize = ZSTD_frameHeaderSize(src, srcSize);
-+		if (srcSize < fhsize)
-+			return fhsize;
-+	}
-+
-+	{
-+		BYTE const fhdByte = ip[4];
-+		size_t pos = 5;
-+		U32 const dictIDSizeCode = fhdByte & 3;
-+		U32 const checksumFlag = (fhdByte >> 2) & 1;
-+		U32 const singleSegment = (fhdByte >> 5) & 1;
-+		U32 const fcsID = fhdByte >> 6;
-+		U32 const windowSizeMax = 1U << ZSTD_WINDOWLOG_MAX;
-+		U32 windowSize = 0;
-+		U32 dictID = 0;
-+		U64 frameContentSize = 0;
-+		if ((fhdByte & 0x08) != 0)
-+			return ERROR(frameParameter_unsupported); /* reserved bits, which must be zero */
-+		if (!singleSegment) {
-+			BYTE const wlByte = ip[pos++];
-+			U32 const windowLog = (wlByte >> 3) + ZSTD_WINDOWLOG_ABSOLUTEMIN;
-+			if (windowLog > ZSTD_WINDOWLOG_MAX)
-+				return ERROR(frameParameter_windowTooLarge); /* avoids issue with 1 << windowLog */
-+			windowSize = (1U << windowLog);
-+			windowSize += (windowSize >> 3) * (wlByte & 7);
-+		}
-+
-+		switch (dictIDSizeCode) {
-+		default: /* impossible */
-+		case 0: break;
-+		case 1:
-+			dictID = ip[pos];
-+			pos++;
-+			break;
-+		case 2:
-+			dictID = ZSTD_readLE16(ip + pos);
-+			pos += 2;
-+			break;
-+		case 3:
-+			dictID = ZSTD_readLE32(ip + pos);
-+			pos += 4;
-+			break;
-+		}
-+		switch (fcsID) {
-+		default: /* impossible */
-+		case 0:
-+			if (singleSegment)
-+				frameContentSize = ip[pos];
-+			break;
-+		case 1: frameContentSize = ZSTD_readLE16(ip + pos) + 256; break;
-+		case 2: frameContentSize = ZSTD_readLE32(ip + pos); break;
-+		case 3: frameContentSize = ZSTD_readLE64(ip + pos); break;
-+		}
-+		if (!windowSize)
-+			windowSize = (U32)frameContentSize;
-+		if (windowSize > windowSizeMax)
-+			return ERROR(frameParameter_windowTooLarge);
-+		fparamsPtr->frameContentSize = frameContentSize;
-+		fparamsPtr->windowSize = windowSize;
-+		fparamsPtr->dictID = dictID;
-+		fparamsPtr->checksumFlag = checksumFlag;
-+	}
-+	return 0;
-+}
-+
-+/** ZSTD_getFrameContentSize() :
-+*   compatible with legacy mode
-+*   @return : decompressed size of the single frame pointed to be `src` if known, otherwise
-+*             - ZSTD_CONTENTSIZE_UNKNOWN if the size cannot be determined
-+*             - ZSTD_CONTENTSIZE_ERROR if an error occurred (e.g. invalid magic number, srcSize too small) */
-+unsigned long long ZSTD_getFrameContentSize(const void *src, size_t srcSize)
-+{
-+	{
-+		ZSTD_frameParams fParams;
-+		if (ZSTD_getFrameParams(&fParams, src, srcSize) != 0)
-+			return ZSTD_CONTENTSIZE_ERROR;
-+		if (fParams.windowSize == 0) {
-+			/* Either skippable or empty frame, size == 0 either way */
-+			return 0;
-+		} else if (fParams.frameContentSize != 0) {
-+			return fParams.frameContentSize;
-+		} else {
-+			return ZSTD_CONTENTSIZE_UNKNOWN;
-+		}
-+	}
-+}
-+
-+/** ZSTD_findDecompressedSize() :
-+ *  compatible with legacy mode
-+ *  `srcSize` must be the exact length of some number of ZSTD compressed and/or
-+ *      skippable frames
-+ *  @return : decompressed size of the frames contained */
-+unsigned long long ZSTD_findDecompressedSize(const void *src, size_t srcSize)
-+{
-+	{
-+		unsigned long long totalDstSize = 0;
-+		while (srcSize >= ZSTD_frameHeaderSize_prefix) {
-+			const U32 magicNumber = ZSTD_readLE32(src);
-+
-+			if ((magicNumber & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) {
-+				size_t skippableSize;
-+				if (srcSize < ZSTD_skippableHeaderSize)
-+					return ERROR(srcSize_wrong);
-+				skippableSize = ZSTD_readLE32((const BYTE *)src + 4) + ZSTD_skippableHeaderSize;
-+				if (srcSize < skippableSize) {
-+					return ZSTD_CONTENTSIZE_ERROR;
-+				}
-+
-+				src = (const BYTE *)src + skippableSize;
-+				srcSize -= skippableSize;
-+				continue;
-+			}
-+
-+			{
-+				unsigned long long const ret = ZSTD_getFrameContentSize(src, srcSize);
-+				if (ret >= ZSTD_CONTENTSIZE_ERROR)
-+					return ret;
-+
-+				/* check for overflow */
-+				if (totalDstSize + ret < totalDstSize)
-+					return ZSTD_CONTENTSIZE_ERROR;
-+				totalDstSize += ret;
-+			}
-+			{
-+				size_t const frameSrcSize = ZSTD_findFrameCompressedSize(src, srcSize);
-+				if (ZSTD_isError(frameSrcSize)) {
-+					return ZSTD_CONTENTSIZE_ERROR;
-+				}
-+
-+				src = (const BYTE *)src + frameSrcSize;
-+				srcSize -= frameSrcSize;
-+			}
-+		}
-+
-+		if (srcSize) {
-+			return ZSTD_CONTENTSIZE_ERROR;
-+		}
-+
-+		return totalDstSize;
-+	}
-+}
-+
-+/** ZSTD_decodeFrameHeader() :
-+*   `headerSize` must be the size provided by ZSTD_frameHeaderSize().
-+*   @return : 0 if success, or an error code, which can be tested using ZSTD_isError() */
-+static size_t ZSTD_decodeFrameHeader(ZSTD_DCtx *dctx, const void *src, size_t headerSize)
-+{
-+	size_t const result = ZSTD_getFrameParams(&(dctx->fParams), src, headerSize);
-+	if (ZSTD_isError(result))
-+		return result; /* invalid header */
-+	if (result > 0)
-+		return ERROR(srcSize_wrong); /* headerSize too small */
-+	if (dctx->fParams.dictID && (dctx->dictID != dctx->fParams.dictID))
-+		return ERROR(dictionary_wrong);
-+	if (dctx->fParams.checksumFlag)
-+		xxh64_reset(&dctx->xxhState, 0);
-+	return 0;
-+}
-+
-+typedef struct {
-+	blockType_e blockType;
-+	U32 lastBlock;
-+	U32 origSize;
-+} blockProperties_t;
-+
-+/*! ZSTD_getcBlockSize() :
-+*   Provides the size of compressed block from block header `src` */
-+size_t ZSTD_getcBlockSize(const void *src, size_t srcSize, blockProperties_t *bpPtr)
-+{
-+	if (srcSize < ZSTD_blockHeaderSize)
-+		return ERROR(srcSize_wrong);
-+	{
-+		U32 const cBlockHeader = ZSTD_readLE24(src);
-+		U32 const cSize = cBlockHeader >> 3;
-+		bpPtr->lastBlock = cBlockHeader & 1;
-+		bpPtr->blockType = (blockType_e)((cBlockHeader >> 1) & 3);
-+		bpPtr->origSize = cSize; /* only useful for RLE */
-+		if (bpPtr->blockType == bt_rle)
-+			return 1;
-+		if (bpPtr->blockType == bt_reserved)
-+			return ERROR(corruption_detected);
-+		return cSize;
-+	}
-+}
-+
-+static size_t ZSTD_copyRawBlock(void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-+{
-+	if (srcSize > dstCapacity)
-+		return ERROR(dstSize_tooSmall);
-+	memcpy(dst, src, srcSize);
-+	return srcSize;
-+}
-+
-+static size_t ZSTD_setRleBlock(void *dst, size_t dstCapacity, const void *src, size_t srcSize, size_t regenSize)
-+{
-+	if (srcSize != 1)
-+		return ERROR(srcSize_wrong);
-+	if (regenSize > dstCapacity)
-+		return ERROR(dstSize_tooSmall);
-+	memset(dst, *(const BYTE *)src, regenSize);
-+	return regenSize;
-+}
-+
-+/*! ZSTD_decodeLiteralsBlock() :
-+	@return : nb of bytes read from src (< srcSize ) */
-+size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx *dctx, const void *src, size_t srcSize) /* note : srcSize < BLOCKSIZE */
-+{
-+	if (srcSize < MIN_CBLOCK_SIZE)
-+		return ERROR(corruption_detected);
-+
-+	{
-+		const BYTE *const istart = (const BYTE *)src;
-+		symbolEncodingType_e const litEncType = (symbolEncodingType_e)(istart[0] & 3);
-+
-+		switch (litEncType) {
-+		case set_repeat:
-+			if (dctx->litEntropy == 0)
-+				return ERROR(dictionary_corrupted);
-+		/* fall-through */
-+		case set_compressed:
-+			if (srcSize < 5)
-+				return ERROR(corruption_detected); /* srcSize >= MIN_CBLOCK_SIZE == 3; here we need up to 5 for case 3 */
-+			{
-+				size_t lhSize, litSize, litCSize;
-+				U32 singleStream = 0;
-+				U32 const lhlCode = (istart[0] >> 2) & 3;
-+				U32 const lhc = ZSTD_readLE32(istart);
-+				switch (lhlCode) {
-+				case 0:
-+				case 1:
-+				default: /* note : default is impossible, since lhlCode into [0..3] */
-+					/* 2 - 2 - 10 - 10 */
-+					singleStream = !lhlCode;
-+					lhSize = 3;
-+					litSize = (lhc >> 4) & 0x3FF;
-+					litCSize = (lhc >> 14) & 0x3FF;
-+					break;
-+				case 2:
-+					/* 2 - 2 - 14 - 14 */
-+					lhSize = 4;
-+					litSize = (lhc >> 4) & 0x3FFF;
-+					litCSize = lhc >> 18;
-+					break;
-+				case 3:
-+					/* 2 - 2 - 18 - 18 */
-+					lhSize = 5;
-+					litSize = (lhc >> 4) & 0x3FFFF;
-+					litCSize = (lhc >> 22) + (istart[4] << 10);
-+					break;
-+				}
-+				if (litSize > ZSTD_BLOCKSIZE_ABSOLUTEMAX)
-+					return ERROR(corruption_detected);
-+				if (litCSize + lhSize > srcSize)
-+					return ERROR(corruption_detected);
-+
-+				if (HUF_isError(
-+					(litEncType == set_repeat)
-+					    ? (singleStream ? HUF_decompress1X_usingDTable(dctx->litBuffer, litSize, istart + lhSize, litCSize, dctx->HUFptr)
-+							    : HUF_decompress4X_usingDTable(dctx->litBuffer, litSize, istart + lhSize, litCSize, dctx->HUFptr))
-+					    : (singleStream
-+						   ? HUF_decompress1X2_DCtx_wksp(dctx->entropy.hufTable, dctx->litBuffer, litSize, istart + lhSize, litCSize,
-+										 dctx->entropy.workspace, sizeof(dctx->entropy.workspace))
-+						   : HUF_decompress4X_hufOnly_wksp(dctx->entropy.hufTable, dctx->litBuffer, litSize, istart + lhSize, litCSize,
-+										   dctx->entropy.workspace, sizeof(dctx->entropy.workspace)))))
-+					return ERROR(corruption_detected);
-+
-+				dctx->litPtr = dctx->litBuffer;
-+				dctx->litSize = litSize;
-+				dctx->litEntropy = 1;
-+				if (litEncType == set_compressed)
-+					dctx->HUFptr = dctx->entropy.hufTable;
-+				memset(dctx->litBuffer + dctx->litSize, 0, WILDCOPY_OVERLENGTH);
-+				return litCSize + lhSize;
-+			}
-+
-+		case set_basic: {
-+			size_t litSize, lhSize;
-+			U32 const lhlCode = ((istart[0]) >> 2) & 3;
-+			switch (lhlCode) {
-+			case 0:
-+			case 2:
-+			default: /* note : default is impossible, since lhlCode into [0..3] */
-+				lhSize = 1;
-+				litSize = istart[0] >> 3;
-+				break;
-+			case 1:
-+				lhSize = 2;
-+				litSize = ZSTD_readLE16(istart) >> 4;
-+				break;
-+			case 3:
-+				lhSize = 3;
-+				litSize = ZSTD_readLE24(istart) >> 4;
-+				break;
-+			}
-+
-+			if (lhSize + litSize + WILDCOPY_OVERLENGTH > srcSize) { /* risk reading beyond src buffer with wildcopy */
-+				if (litSize + lhSize > srcSize)
-+					return ERROR(corruption_detected);
-+				memcpy(dctx->litBuffer, istart + lhSize, litSize);
-+				dctx->litPtr = dctx->litBuffer;
-+				dctx->litSize = litSize;
-+				memset(dctx->litBuffer + dctx->litSize, 0, WILDCOPY_OVERLENGTH);
-+				return lhSize + litSize;
-+			}
-+			/* direct reference into compressed stream */
-+			dctx->litPtr = istart + lhSize;
-+			dctx->litSize = litSize;
-+			return lhSize + litSize;
-+		}
-+
-+		case set_rle: {
-+			U32 const lhlCode = ((istart[0]) >> 2) & 3;
-+			size_t litSize, lhSize;
-+			switch (lhlCode) {
-+			case 0:
-+			case 2:
-+			default: /* note : default is impossible, since lhlCode into [0..3] */
-+				lhSize = 1;
-+				litSize = istart[0] >> 3;
-+				break;
-+			case 1:
-+				lhSize = 2;
-+				litSize = ZSTD_readLE16(istart) >> 4;
-+				break;
-+			case 3:
-+				lhSize = 3;
-+				litSize = ZSTD_readLE24(istart) >> 4;
-+				if (srcSize < 4)
-+					return ERROR(corruption_detected); /* srcSize >= MIN_CBLOCK_SIZE == 3; here we need lhSize+1 = 4 */
-+				break;
-+			}
-+			if (litSize > ZSTD_BLOCKSIZE_ABSOLUTEMAX)
-+				return ERROR(corruption_detected);
-+			memset(dctx->litBuffer, istart[lhSize], litSize + WILDCOPY_OVERLENGTH);
-+			dctx->litPtr = dctx->litBuffer;
-+			dctx->litSize = litSize;
-+			return lhSize + 1;
-+		}
-+		default:
-+			return ERROR(corruption_detected); /* impossible */
-+		}
-+	}
-+}
-+
-+typedef union {
-+	FSE_decode_t realData;
-+	U32 alignedBy4;
-+} FSE_decode_t4;
-+
-+static const FSE_decode_t4 LL_defaultDTable[(1 << LL_DEFAULTNORMLOG) + 1] = {
-+    {{LL_DEFAULTNORMLOG, 1, 1}}, /* header : tableLog, fastMode, fastMode */
-+    {{0, 0, 4}},		 /* 0 : base, symbol, bits */
-+    {{16, 0, 4}},
-+    {{32, 1, 5}},
-+    {{0, 3, 5}},
-+    {{0, 4, 5}},
-+    {{0, 6, 5}},
-+    {{0, 7, 5}},
-+    {{0, 9, 5}},
-+    {{0, 10, 5}},
-+    {{0, 12, 5}},
-+    {{0, 14, 6}},
-+    {{0, 16, 5}},
-+    {{0, 18, 5}},
-+    {{0, 19, 5}},
-+    {{0, 21, 5}},
-+    {{0, 22, 5}},
-+    {{0, 24, 5}},
-+    {{32, 25, 5}},
-+    {{0, 26, 5}},
-+    {{0, 27, 6}},
-+    {{0, 29, 6}},
-+    {{0, 31, 6}},
-+    {{32, 0, 4}},
-+    {{0, 1, 4}},
-+    {{0, 2, 5}},
-+    {{32, 4, 5}},
-+    {{0, 5, 5}},
-+    {{32, 7, 5}},
-+    {{0, 8, 5}},
-+    {{32, 10, 5}},
-+    {{0, 11, 5}},
-+    {{0, 13, 6}},
-+    {{32, 16, 5}},
-+    {{0, 17, 5}},
-+    {{32, 19, 5}},
-+    {{0, 20, 5}},
-+    {{32, 22, 5}},
-+    {{0, 23, 5}},
-+    {{0, 25, 4}},
-+    {{16, 25, 4}},
-+    {{32, 26, 5}},
-+    {{0, 28, 6}},
-+    {{0, 30, 6}},
-+    {{48, 0, 4}},
-+    {{16, 1, 4}},
-+    {{32, 2, 5}},
-+    {{32, 3, 5}},
-+    {{32, 5, 5}},
-+    {{32, 6, 5}},
-+    {{32, 8, 5}},
-+    {{32, 9, 5}},
-+    {{32, 11, 5}},
-+    {{32, 12, 5}},
-+    {{0, 15, 6}},
-+    {{32, 17, 5}},
-+    {{32, 18, 5}},
-+    {{32, 20, 5}},
-+    {{32, 21, 5}},
-+    {{32, 23, 5}},
-+    {{32, 24, 5}},
-+    {{0, 35, 6}},
-+    {{0, 34, 6}},
-+    {{0, 33, 6}},
-+    {{0, 32, 6}},
-+}; /* LL_defaultDTable */
-+
-+static const FSE_decode_t4 ML_defaultDTable[(1 << ML_DEFAULTNORMLOG) + 1] = {
-+    {{ML_DEFAULTNORMLOG, 1, 1}}, /* header : tableLog, fastMode, fastMode */
-+    {{0, 0, 6}},		 /* 0 : base, symbol, bits */
-+    {{0, 1, 4}},
-+    {{32, 2, 5}},
-+    {{0, 3, 5}},
-+    {{0, 5, 5}},
-+    {{0, 6, 5}},
-+    {{0, 8, 5}},
-+    {{0, 10, 6}},
-+    {{0, 13, 6}},
-+    {{0, 16, 6}},
-+    {{0, 19, 6}},
-+    {{0, 22, 6}},
-+    {{0, 25, 6}},
-+    {{0, 28, 6}},
-+    {{0, 31, 6}},
-+    {{0, 33, 6}},
-+    {{0, 35, 6}},
-+    {{0, 37, 6}},
-+    {{0, 39, 6}},
-+    {{0, 41, 6}},
-+    {{0, 43, 6}},
-+    {{0, 45, 6}},
-+    {{16, 1, 4}},
-+    {{0, 2, 4}},
-+    {{32, 3, 5}},
-+    {{0, 4, 5}},
-+    {{32, 6, 5}},
-+    {{0, 7, 5}},
-+    {{0, 9, 6}},
-+    {{0, 12, 6}},
-+    {{0, 15, 6}},
-+    {{0, 18, 6}},
-+    {{0, 21, 6}},
-+    {{0, 24, 6}},
-+    {{0, 27, 6}},
-+    {{0, 30, 6}},
-+    {{0, 32, 6}},
-+    {{0, 34, 6}},
-+    {{0, 36, 6}},
-+    {{0, 38, 6}},
-+    {{0, 40, 6}},
-+    {{0, 42, 6}},
-+    {{0, 44, 6}},
-+    {{32, 1, 4}},
-+    {{48, 1, 4}},
-+    {{16, 2, 4}},
-+    {{32, 4, 5}},
-+    {{32, 5, 5}},
-+    {{32, 7, 5}},
-+    {{32, 8, 5}},
-+    {{0, 11, 6}},
-+    {{0, 14, 6}},
-+    {{0, 17, 6}},
-+    {{0, 20, 6}},
-+    {{0, 23, 6}},
-+    {{0, 26, 6}},
-+    {{0, 29, 6}},
-+    {{0, 52, 6}},
-+    {{0, 51, 6}},
-+    {{0, 50, 6}},
-+    {{0, 49, 6}},
-+    {{0, 48, 6}},
-+    {{0, 47, 6}},
-+    {{0, 46, 6}},
-+}; /* ML_defaultDTable */
-+
-+static const FSE_decode_t4 OF_defaultDTable[(1 << OF_DEFAULTNORMLOG) + 1] = {
-+    {{OF_DEFAULTNORMLOG, 1, 1}}, /* header : tableLog, fastMode, fastMode */
-+    {{0, 0, 5}},		 /* 0 : base, symbol, bits */
-+    {{0, 6, 4}},
-+    {{0, 9, 5}},
-+    {{0, 15, 5}},
-+    {{0, 21, 5}},
-+    {{0, 3, 5}},
-+    {{0, 7, 4}},
-+    {{0, 12, 5}},
-+    {{0, 18, 5}},
-+    {{0, 23, 5}},
-+    {{0, 5, 5}},
-+    {{0, 8, 4}},
-+    {{0, 14, 5}},
-+    {{0, 20, 5}},
-+    {{0, 2, 5}},
-+    {{16, 7, 4}},
-+    {{0, 11, 5}},
-+    {{0, 17, 5}},
-+    {{0, 22, 5}},
-+    {{0, 4, 5}},
-+    {{16, 8, 4}},
-+    {{0, 13, 5}},
-+    {{0, 19, 5}},
-+    {{0, 1, 5}},
-+    {{16, 6, 4}},
-+    {{0, 10, 5}},
-+    {{0, 16, 5}},
-+    {{0, 28, 5}},
-+    {{0, 27, 5}},
-+    {{0, 26, 5}},
-+    {{0, 25, 5}},
-+    {{0, 24, 5}},
-+}; /* OF_defaultDTable */
-+
-+/*! ZSTD_buildSeqTable() :
-+	@return : nb bytes read from src,
-+			  or an error code if it fails, testable with ZSTD_isError()
-+*/
-+static size_t ZSTD_buildSeqTable(FSE_DTable *DTableSpace, const FSE_DTable **DTablePtr, symbolEncodingType_e type, U32 max, U32 maxLog, const void *src,
-+				 size_t srcSize, const FSE_decode_t4 *defaultTable, U32 flagRepeatTable, void *workspace, size_t workspaceSize)
-+{
-+	const void *const tmpPtr = defaultTable; /* bypass strict aliasing */
-+	switch (type) {
-+	case set_rle:
-+		if (!srcSize)
-+			return ERROR(srcSize_wrong);
-+		if ((*(const BYTE *)src) > max)
-+			return ERROR(corruption_detected);
-+		FSE_buildDTable_rle(DTableSpace, *(const BYTE *)src);
-+		*DTablePtr = DTableSpace;
-+		return 1;
-+	case set_basic: *DTablePtr = (const FSE_DTable *)tmpPtr; return 0;
-+	case set_repeat:
-+		if (!flagRepeatTable)
-+			return ERROR(corruption_detected);
-+		return 0;
-+	default: /* impossible */
-+	case set_compressed: {
-+		U32 tableLog;
-+		S16 *norm = (S16 *)workspace;
-+		size_t const spaceUsed32 = ALIGN(sizeof(S16) * (MaxSeq + 1), sizeof(U32)) >> 2;
-+
-+		if ((spaceUsed32 << 2) > workspaceSize)
-+			return ERROR(GENERIC);
-+		workspace = (U32 *)workspace + spaceUsed32;
-+		workspaceSize -= (spaceUsed32 << 2);
-+		{
-+			size_t const headerSize = FSE_readNCount(norm, &max, &tableLog, src, srcSize);
-+			if (FSE_isError(headerSize))
-+				return ERROR(corruption_detected);
-+			if (tableLog > maxLog)
-+				return ERROR(corruption_detected);
-+			FSE_buildDTable_wksp(DTableSpace, norm, max, tableLog, workspace, workspaceSize);
-+			*DTablePtr = DTableSpace;
-+			return headerSize;
-+		}
-+	}
-+	}
-+}
-+
-+size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx *dctx, int *nbSeqPtr, const void *src, size_t srcSize)
-+{
-+	const BYTE *const istart = (const BYTE *const)src;
-+	const BYTE *const iend = istart + srcSize;
-+	const BYTE *ip = istart;
-+
-+	/* check */
-+	if (srcSize < MIN_SEQUENCES_SIZE)
-+		return ERROR(srcSize_wrong);
-+
-+	/* SeqHead */
-+	{
-+		int nbSeq = *ip++;
-+		if (!nbSeq) {
-+			*nbSeqPtr = 0;
-+			return 1;
-+		}
-+		if (nbSeq > 0x7F) {
-+			if (nbSeq == 0xFF) {
-+				if (ip + 2 > iend)
-+					return ERROR(srcSize_wrong);
-+				nbSeq = ZSTD_readLE16(ip) + LONGNBSEQ, ip += 2;
-+			} else {
-+				if (ip >= iend)
-+					return ERROR(srcSize_wrong);
-+				nbSeq = ((nbSeq - 0x80) << 8) + *ip++;
-+			}
-+		}
-+		*nbSeqPtr = nbSeq;
-+	}
-+
-+	/* FSE table descriptors */
-+	if (ip + 4 > iend)
-+		return ERROR(srcSize_wrong); /* minimum possible size */
-+	{
-+		symbolEncodingType_e const LLtype = (symbolEncodingType_e)(*ip >> 6);
-+		symbolEncodingType_e const OFtype = (symbolEncodingType_e)((*ip >> 4) & 3);
-+		symbolEncodingType_e const MLtype = (symbolEncodingType_e)((*ip >> 2) & 3);
-+		ip++;
-+
-+		/* Build DTables */
-+		{
-+			size_t const llhSize = ZSTD_buildSeqTable(dctx->entropy.LLTable, &dctx->LLTptr, LLtype, MaxLL, LLFSELog, ip, iend - ip,
-+								  LL_defaultDTable, dctx->fseEntropy, dctx->entropy.workspace, sizeof(dctx->entropy.workspace));
-+			if (ZSTD_isError(llhSize))
-+				return ERROR(corruption_detected);
-+			ip += llhSize;
-+		}
-+		{
-+			size_t const ofhSize = ZSTD_buildSeqTable(dctx->entropy.OFTable, &dctx->OFTptr, OFtype, MaxOff, OffFSELog, ip, iend - ip,
-+								  OF_defaultDTable, dctx->fseEntropy, dctx->entropy.workspace, sizeof(dctx->entropy.workspace));
-+			if (ZSTD_isError(ofhSize))
-+				return ERROR(corruption_detected);
-+			ip += ofhSize;
-+		}
-+		{
-+			size_t const mlhSize = ZSTD_buildSeqTable(dctx->entropy.MLTable, &dctx->MLTptr, MLtype, MaxML, MLFSELog, ip, iend - ip,
-+								  ML_defaultDTable, dctx->fseEntropy, dctx->entropy.workspace, sizeof(dctx->entropy.workspace));
-+			if (ZSTD_isError(mlhSize))
-+				return ERROR(corruption_detected);
-+			ip += mlhSize;
-+		}
-+	}
-+
-+	return ip - istart;
-+}
-+
-+typedef struct {
-+	size_t litLength;
-+	size_t matchLength;
-+	size_t offset;
-+	const BYTE *match;
-+} seq_t;
-+
-+typedef struct {
-+	BIT_DStream_t DStream;
-+	FSE_DState_t stateLL;
-+	FSE_DState_t stateOffb;
-+	FSE_DState_t stateML;
-+	size_t prevOffset[ZSTD_REP_NUM];
-+	const BYTE *base;
-+	size_t pos;
-+	uPtrDiff gotoDict;
-+} seqState_t;
-+
-+FORCE_NOINLINE
-+size_t ZSTD_execSequenceLast7(BYTE *op, BYTE *const oend, seq_t sequence, const BYTE **litPtr, const BYTE *const litLimit, const BYTE *const base,
-+			      const BYTE *const vBase, const BYTE *const dictEnd)
-+{
-+	BYTE *const oLitEnd = op + sequence.litLength;
-+	size_t const sequenceLength = sequence.litLength + sequence.matchLength;
-+	BYTE *const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */
-+	BYTE *const oend_w = oend - WILDCOPY_OVERLENGTH;
-+	const BYTE *const iLitEnd = *litPtr + sequence.litLength;
-+	const BYTE *match = oLitEnd - sequence.offset;
-+
-+	/* check */
-+	if (oMatchEnd > oend)
-+		return ERROR(dstSize_tooSmall); /* last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend */
-+	if (iLitEnd > litLimit)
-+		return ERROR(corruption_detected); /* over-read beyond lit buffer */
-+	if (oLitEnd <= oend_w)
-+		return ERROR(GENERIC); /* Precondition */
-+
-+	/* copy literals */
-+	if (op < oend_w) {
-+		ZSTD_wildcopy(op, *litPtr, oend_w - op);
-+		*litPtr += oend_w - op;
-+		op = oend_w;
-+	}
-+	while (op < oLitEnd)
-+		*op++ = *(*litPtr)++;
-+
-+	/* copy Match */
-+	if (sequence.offset > (size_t)(oLitEnd - base)) {
-+		/* offset beyond prefix */
-+		if (sequence.offset > (size_t)(oLitEnd - vBase))
-+			return ERROR(corruption_detected);
-+		match = dictEnd - (base - match);
-+		if (match + sequence.matchLength <= dictEnd) {
-+			memmove(oLitEnd, match, sequence.matchLength);
-+			return sequenceLength;
-+		}
-+		/* span extDict & currPrefixSegment */
-+		{
-+			size_t const length1 = dictEnd - match;
-+			memmove(oLitEnd, match, length1);
-+			op = oLitEnd + length1;
-+			sequence.matchLength -= length1;
-+			match = base;
-+		}
-+	}
-+	while (op < oMatchEnd)
-+		*op++ = *match++;
-+	return sequenceLength;
-+}
-+
-+static seq_t ZSTD_decodeSequence(seqState_t *seqState)
-+{
-+	seq_t seq;
-+
-+	U32 const llCode = FSE_peekSymbol(&seqState->stateLL);
-+	U32 const mlCode = FSE_peekSymbol(&seqState->stateML);
-+	U32 const ofCode = FSE_peekSymbol(&seqState->stateOffb); /* <= maxOff, by table construction */
-+
-+	U32 const llBits = LL_bits[llCode];
-+	U32 const mlBits = ML_bits[mlCode];
-+	U32 const ofBits = ofCode;
-+	U32 const totalBits = llBits + mlBits + ofBits;
-+
-+	static const U32 LL_base[MaxLL + 1] = {0,  1,  2,  3,  4,  5,  6,  7,  8,    9,     10,    11,    12,    13,     14,     15,     16,     18,
-+					       20, 22, 24, 28, 32, 40, 48, 64, 0x80, 0x100, 0x200, 0x400, 0x800, 0x1000, 0x2000, 0x4000, 0x8000, 0x10000};
-+
-+	static const U32 ML_base[MaxML + 1] = {3,  4,  5,  6,  7,  8,  9,  10,   11,    12,    13,    14,    15,     16,     17,     18,     19,     20,
-+					       21, 22, 23, 24, 25, 26, 27, 28,   29,    30,    31,    32,    33,     34,     35,     37,     39,     41,
-+					       43, 47, 51, 59, 67, 83, 99, 0x83, 0x103, 0x203, 0x403, 0x803, 0x1003, 0x2003, 0x4003, 0x8003, 0x10003};
-+
-+	static const U32 OF_base[MaxOff + 1] = {0,       1,	1,	5,	0xD,      0x1D,      0x3D,      0x7D,      0xFD,     0x1FD,
-+						0x3FD,   0x7FD,    0xFFD,    0x1FFD,   0x3FFD,   0x7FFD,    0xFFFD,    0x1FFFD,   0x3FFFD,  0x7FFFD,
-+						0xFFFFD, 0x1FFFFD, 0x3FFFFD, 0x7FFFFD, 0xFFFFFD, 0x1FFFFFD, 0x3FFFFFD, 0x7FFFFFD, 0xFFFFFFD};
-+
-+	/* sequence */
-+	{
-+		size_t offset;
-+		if (!ofCode)
-+			offset = 0;
-+		else {
-+			offset = OF_base[ofCode] + BIT_readBitsFast(&seqState->DStream, ofBits); /* <=  (ZSTD_WINDOWLOG_MAX-1) bits */
-+			if (ZSTD_32bits())
-+				BIT_reloadDStream(&seqState->DStream);
-+		}
-+
-+		if (ofCode <= 1) {
-+			offset += (llCode == 0);
-+			if (offset) {
-+				size_t temp = (offset == 3) ? seqState->prevOffset[0] - 1 : seqState->prevOffset[offset];
-+				temp += !temp; /* 0 is not valid; input is corrupted; force offset to 1 */
-+				if (offset != 1)
-+					seqState->prevOffset[2] = seqState->prevOffset[1];
-+				seqState->prevOffset[1] = seqState->prevOffset[0];
-+				seqState->prevOffset[0] = offset = temp;
-+			} else {
-+				offset = seqState->prevOffset[0];
-+			}
-+		} else {
-+			seqState->prevOffset[2] = seqState->prevOffset[1];
-+			seqState->prevOffset[1] = seqState->prevOffset[0];
-+			seqState->prevOffset[0] = offset;
-+		}
-+		seq.offset = offset;
-+	}
-+
-+	seq.matchLength = ML_base[mlCode] + ((mlCode > 31) ? BIT_readBitsFast(&seqState->DStream, mlBits) : 0); /* <=  16 bits */
-+	if (ZSTD_32bits() && (mlBits + llBits > 24))
-+		BIT_reloadDStream(&seqState->DStream);
-+
-+	seq.litLength = LL_base[llCode] + ((llCode > 15) ? BIT_readBitsFast(&seqState->DStream, llBits) : 0); /* <=  16 bits */
-+	if (ZSTD_32bits() || (totalBits > 64 - 7 - (LLFSELog + MLFSELog + OffFSELog)))
-+		BIT_reloadDStream(&seqState->DStream);
-+
-+	/* ANS state update */
-+	FSE_updateState(&seqState->stateLL, &seqState->DStream); /* <=  9 bits */
-+	FSE_updateState(&seqState->stateML, &seqState->DStream); /* <=  9 bits */
-+	if (ZSTD_32bits())
-+		BIT_reloadDStream(&seqState->DStream);		   /* <= 18 bits */
-+	FSE_updateState(&seqState->stateOffb, &seqState->DStream); /* <=  8 bits */
-+
-+	seq.match = NULL;
-+
-+	return seq;
-+}
-+
-+FORCE_INLINE
-+size_t ZSTD_execSequence(BYTE *op, BYTE *const oend, seq_t sequence, const BYTE **litPtr, const BYTE *const litLimit, const BYTE *const base,
-+			 const BYTE *const vBase, const BYTE *const dictEnd)
-+{
-+	BYTE *const oLitEnd = op + sequence.litLength;
-+	size_t const sequenceLength = sequence.litLength + sequence.matchLength;
-+	BYTE *const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */
-+	BYTE *const oend_w = oend - WILDCOPY_OVERLENGTH;
-+	const BYTE *const iLitEnd = *litPtr + sequence.litLength;
-+	const BYTE *match = oLitEnd - sequence.offset;
-+
-+	/* check */
-+	if (oMatchEnd > oend)
-+		return ERROR(dstSize_tooSmall); /* last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend */
-+	if (iLitEnd > litLimit)
-+		return ERROR(corruption_detected); /* over-read beyond lit buffer */
-+	if (oLitEnd > oend_w)
-+		return ZSTD_execSequenceLast7(op, oend, sequence, litPtr, litLimit, base, vBase, dictEnd);
-+
-+	/* copy Literals */
-+	ZSTD_copy8(op, *litPtr);
-+	if (sequence.litLength > 8)
-+		ZSTD_wildcopy(op + 8, (*litPtr) + 8,
-+			      sequence.litLength - 8); /* note : since oLitEnd <= oend-WILDCOPY_OVERLENGTH, no risk of overwrite beyond oend */
-+	op = oLitEnd;
-+	*litPtr = iLitEnd; /* update for next sequence */
-+
-+	/* copy Match */
-+	if (sequence.offset > (size_t)(oLitEnd - base)) {
-+		/* offset beyond prefix */
-+		if (sequence.offset > (size_t)(oLitEnd - vBase))
-+			return ERROR(corruption_detected);
-+		match = dictEnd + (match - base);
-+		if (match + sequence.matchLength <= dictEnd) {
-+			memmove(oLitEnd, match, sequence.matchLength);
-+			return sequenceLength;
-+		}
-+		/* span extDict & currPrefixSegment */
-+		{
-+			size_t const length1 = dictEnd - match;
-+			memmove(oLitEnd, match, length1);
-+			op = oLitEnd + length1;
-+			sequence.matchLength -= length1;
-+			match = base;
-+			if (op > oend_w || sequence.matchLength < MINMATCH) {
-+				U32 i;
-+				for (i = 0; i < sequence.matchLength; ++i)
-+					op[i] = match[i];
-+				return sequenceLength;
-+			}
-+		}
-+	}
-+	/* Requirement: op <= oend_w && sequence.matchLength >= MINMATCH */
-+
-+	/* match within prefix */
-+	if (sequence.offset < 8) {
-+		/* close range match, overlap */
-+		static const U32 dec32table[] = {0, 1, 2, 1, 4, 4, 4, 4};   /* added */
-+		static const int dec64table[] = {8, 8, 8, 7, 8, 9, 10, 11}; /* subtracted */
-+		int const sub2 = dec64table[sequence.offset];
-+		op[0] = match[0];
-+		op[1] = match[1];
-+		op[2] = match[2];
-+		op[3] = match[3];
-+		match += dec32table[sequence.offset];
-+		ZSTD_copy4(op + 4, match);
-+		match -= sub2;
-+	} else {
-+		ZSTD_copy8(op, match);
-+	}
-+	op += 8;
-+	match += 8;
-+
-+	if (oMatchEnd > oend - (16 - MINMATCH)) {
-+		if (op < oend_w) {
-+			ZSTD_wildcopy(op, match, oend_w - op);
-+			match += oend_w - op;
-+			op = oend_w;
-+		}
-+		while (op < oMatchEnd)
-+			*op++ = *match++;
-+	} else {
-+		ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength - 8); /* works even if matchLength < 8 */
-+	}
-+	return sequenceLength;
-+}
-+
-+static size_t ZSTD_decompressSequences(ZSTD_DCtx *dctx, void *dst, size_t maxDstSize, const void *seqStart, size_t seqSize)
-+{
-+	const BYTE *ip = (const BYTE *)seqStart;
-+	const BYTE *const iend = ip + seqSize;
-+	BYTE *const ostart = (BYTE * const)dst;
-+	BYTE *const oend = ostart + maxDstSize;
-+	BYTE *op = ostart;
-+	const BYTE *litPtr = dctx->litPtr;
-+	const BYTE *const litEnd = litPtr + dctx->litSize;
-+	const BYTE *const base = (const BYTE *)(dctx->base);
-+	const BYTE *const vBase = (const BYTE *)(dctx->vBase);
-+	const BYTE *const dictEnd = (const BYTE *)(dctx->dictEnd);
-+	int nbSeq;
-+
-+	/* Build Decoding Tables */
-+	{
-+		size_t const seqHSize = ZSTD_decodeSeqHeaders(dctx, &nbSeq, ip, seqSize);
-+		if (ZSTD_isError(seqHSize))
-+			return seqHSize;
-+		ip += seqHSize;
-+	}
-+
-+	/* Regen sequences */
-+	if (nbSeq) {
-+		seqState_t seqState;
-+		dctx->fseEntropy = 1;
-+		{
-+			U32 i;
-+			for (i = 0; i < ZSTD_REP_NUM; i++)
-+				seqState.prevOffset[i] = dctx->entropy.rep[i];
-+		}
-+		CHECK_E(BIT_initDStream(&seqState.DStream, ip, iend - ip), corruption_detected);
-+		FSE_initDState(&seqState.stateLL, &seqState.DStream, dctx->LLTptr);
-+		FSE_initDState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr);
-+		FSE_initDState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);
-+
-+		for (; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && nbSeq;) {
-+			nbSeq--;
-+			{
-+				seq_t const sequence = ZSTD_decodeSequence(&seqState);
-+				size_t const oneSeqSize = ZSTD_execSequence(op, oend, sequence, &litPtr, litEnd, base, vBase, dictEnd);
-+				if (ZSTD_isError(oneSeqSize))
-+					return oneSeqSize;
-+				op += oneSeqSize;
-+			}
-+		}
-+
-+		/* check if reached exact end */
-+		if (nbSeq)
-+			return ERROR(corruption_detected);
-+		/* save reps for next block */
-+		{
-+			U32 i;
-+			for (i = 0; i < ZSTD_REP_NUM; i++)
-+				dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]);
-+		}
-+	}
-+
-+	/* last literal segment */
-+	{
-+		size_t const lastLLSize = litEnd - litPtr;
-+		if (lastLLSize > (size_t)(oend - op))
-+			return ERROR(dstSize_tooSmall);
-+		memcpy(op, litPtr, lastLLSize);
-+		op += lastLLSize;
-+	}
-+
-+	return op - ostart;
-+}
-+
-+FORCE_INLINE seq_t ZSTD_decodeSequenceLong_generic(seqState_t *seqState, int const longOffsets)
-+{
-+	seq_t seq;
-+
-+	U32 const llCode = FSE_peekSymbol(&seqState->stateLL);
-+	U32 const mlCode = FSE_peekSymbol(&seqState->stateML);
-+	U32 const ofCode = FSE_peekSymbol(&seqState->stateOffb); /* <= maxOff, by table construction */
-+
-+	U32 const llBits = LL_bits[llCode];
-+	U32 const mlBits = ML_bits[mlCode];
-+	U32 const ofBits = ofCode;
-+	U32 const totalBits = llBits + mlBits + ofBits;
-+
-+	static const U32 LL_base[MaxLL + 1] = {0,  1,  2,  3,  4,  5,  6,  7,  8,    9,     10,    11,    12,    13,     14,     15,     16,     18,
-+					       20, 22, 24, 28, 32, 40, 48, 64, 0x80, 0x100, 0x200, 0x400, 0x800, 0x1000, 0x2000, 0x4000, 0x8000, 0x10000};
-+
-+	static const U32 ML_base[MaxML + 1] = {3,  4,  5,  6,  7,  8,  9,  10,   11,    12,    13,    14,    15,     16,     17,     18,     19,     20,
-+					       21, 22, 23, 24, 25, 26, 27, 28,   29,    30,    31,    32,    33,     34,     35,     37,     39,     41,
-+					       43, 47, 51, 59, 67, 83, 99, 0x83, 0x103, 0x203, 0x403, 0x803, 0x1003, 0x2003, 0x4003, 0x8003, 0x10003};
-+
-+	static const U32 OF_base[MaxOff + 1] = {0,       1,	1,	5,	0xD,      0x1D,      0x3D,      0x7D,      0xFD,     0x1FD,
-+						0x3FD,   0x7FD,    0xFFD,    0x1FFD,   0x3FFD,   0x7FFD,    0xFFFD,    0x1FFFD,   0x3FFFD,  0x7FFFD,
-+						0xFFFFD, 0x1FFFFD, 0x3FFFFD, 0x7FFFFD, 0xFFFFFD, 0x1FFFFFD, 0x3FFFFFD, 0x7FFFFFD, 0xFFFFFFD};
-+
-+	/* sequence */
-+	{
-+		size_t offset;
-+		if (!ofCode)
-+			offset = 0;
-+		else {
-+			if (longOffsets) {
-+				int const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN);
-+				offset = OF_base[ofCode] + (BIT_readBitsFast(&seqState->DStream, ofBits - extraBits) << extraBits);
-+				if (ZSTD_32bits() || extraBits)
-+					BIT_reloadDStream(&seqState->DStream);
-+				if (extraBits)
-+					offset += BIT_readBitsFast(&seqState->DStream, extraBits);
-+			} else {
-+				offset = OF_base[ofCode] + BIT_readBitsFast(&seqState->DStream, ofBits); /* <=  (ZSTD_WINDOWLOG_MAX-1) bits */
-+				if (ZSTD_32bits())
-+					BIT_reloadDStream(&seqState->DStream);
-+			}
-+		}
-+
-+		if (ofCode <= 1) {
-+			offset += (llCode == 0);
-+			if (offset) {
-+				size_t temp = (offset == 3) ? seqState->prevOffset[0] - 1 : seqState->prevOffset[offset];
-+				temp += !temp; /* 0 is not valid; input is corrupted; force offset to 1 */
-+				if (offset != 1)
-+					seqState->prevOffset[2] = seqState->prevOffset[1];
-+				seqState->prevOffset[1] = seqState->prevOffset[0];
-+				seqState->prevOffset[0] = offset = temp;
-+			} else {
-+				offset = seqState->prevOffset[0];
-+			}
-+		} else {
-+			seqState->prevOffset[2] = seqState->prevOffset[1];
-+			seqState->prevOffset[1] = seqState->prevOffset[0];
-+			seqState->prevOffset[0] = offset;
-+		}
-+		seq.offset = offset;
-+	}
-+
-+	seq.matchLength = ML_base[mlCode] + ((mlCode > 31) ? BIT_readBitsFast(&seqState->DStream, mlBits) : 0); /* <=  16 bits */
-+	if (ZSTD_32bits() && (mlBits + llBits > 24))
-+		BIT_reloadDStream(&seqState->DStream);
-+
-+	seq.litLength = LL_base[llCode] + ((llCode > 15) ? BIT_readBitsFast(&seqState->DStream, llBits) : 0); /* <=  16 bits */
-+	if (ZSTD_32bits() || (totalBits > 64 - 7 - (LLFSELog + MLFSELog + OffFSELog)))
-+		BIT_reloadDStream(&seqState->DStream);
-+
-+	{
-+		size_t const pos = seqState->pos + seq.litLength;
-+		seq.match = seqState->base + pos - seq.offset; /* single memory segment */
-+		if (seq.offset > pos)
-+			seq.match += seqState->gotoDict; /* separate memory segment */
-+		seqState->pos = pos + seq.matchLength;
-+	}
-+
-+	/* ANS state update */
-+	FSE_updateState(&seqState->stateLL, &seqState->DStream); /* <=  9 bits */
-+	FSE_updateState(&seqState->stateML, &seqState->DStream); /* <=  9 bits */
-+	if (ZSTD_32bits())
-+		BIT_reloadDStream(&seqState->DStream);		   /* <= 18 bits */
-+	FSE_updateState(&seqState->stateOffb, &seqState->DStream); /* <=  8 bits */
-+
-+	return seq;
-+}
-+
-+static seq_t ZSTD_decodeSequenceLong(seqState_t *seqState, unsigned const windowSize)
-+{
-+	if (ZSTD_highbit32(windowSize) > STREAM_ACCUMULATOR_MIN) {
-+		return ZSTD_decodeSequenceLong_generic(seqState, 1);
-+	} else {
-+		return ZSTD_decodeSequenceLong_generic(seqState, 0);
-+	}
-+}
-+
-+FORCE_INLINE
-+size_t ZSTD_execSequenceLong(BYTE *op, BYTE *const oend, seq_t sequence, const BYTE **litPtr, const BYTE *const litLimit, const BYTE *const base,
-+			     const BYTE *const vBase, const BYTE *const dictEnd)
-+{
-+	BYTE *const oLitEnd = op + sequence.litLength;
-+	size_t const sequenceLength = sequence.litLength + sequence.matchLength;
-+	BYTE *const oMatchEnd = op + sequenceLength; /* risk : address space overflow (32-bits) */
-+	BYTE *const oend_w = oend - WILDCOPY_OVERLENGTH;
-+	const BYTE *const iLitEnd = *litPtr + sequence.litLength;
-+	const BYTE *match = sequence.match;
-+
-+	/* check */
-+	if (oMatchEnd > oend)
-+		return ERROR(dstSize_tooSmall); /* last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend */
-+	if (iLitEnd > litLimit)
-+		return ERROR(corruption_detected); /* over-read beyond lit buffer */
-+	if (oLitEnd > oend_w)
-+		return ZSTD_execSequenceLast7(op, oend, sequence, litPtr, litLimit, base, vBase, dictEnd);
-+
-+	/* copy Literals */
-+	ZSTD_copy8(op, *litPtr);
-+	if (sequence.litLength > 8)
-+		ZSTD_wildcopy(op + 8, (*litPtr) + 8,
-+			      sequence.litLength - 8); /* note : since oLitEnd <= oend-WILDCOPY_OVERLENGTH, no risk of overwrite beyond oend */
-+	op = oLitEnd;
-+	*litPtr = iLitEnd; /* update for next sequence */
-+
-+	/* copy Match */
-+	if (sequence.offset > (size_t)(oLitEnd - base)) {
-+		/* offset beyond prefix */
-+		if (sequence.offset > (size_t)(oLitEnd - vBase))
-+			return ERROR(corruption_detected);
-+		if (match + sequence.matchLength <= dictEnd) {
-+			memmove(oLitEnd, match, sequence.matchLength);
-+			return sequenceLength;
-+		}
-+		/* span extDict & currPrefixSegment */
-+		{
-+			size_t const length1 = dictEnd - match;
-+			memmove(oLitEnd, match, length1);
-+			op = oLitEnd + length1;
-+			sequence.matchLength -= length1;
-+			match = base;
-+			if (op > oend_w || sequence.matchLength < MINMATCH) {
-+				U32 i;
-+				for (i = 0; i < sequence.matchLength; ++i)
-+					op[i] = match[i];
-+				return sequenceLength;
-+			}
-+		}
-+	}
-+	/* Requirement: op <= oend_w && sequence.matchLength >= MINMATCH */
-+
-+	/* match within prefix */
-+	if (sequence.offset < 8) {
-+		/* close range match, overlap */
-+		static const U32 dec32table[] = {0, 1, 2, 1, 4, 4, 4, 4};   /* added */
-+		static const int dec64table[] = {8, 8, 8, 7, 8, 9, 10, 11}; /* subtracted */
-+		int const sub2 = dec64table[sequence.offset];
-+		op[0] = match[0];
-+		op[1] = match[1];
-+		op[2] = match[2];
-+		op[3] = match[3];
-+		match += dec32table[sequence.offset];
-+		ZSTD_copy4(op + 4, match);
-+		match -= sub2;
-+	} else {
-+		ZSTD_copy8(op, match);
-+	}
-+	op += 8;
-+	match += 8;
-+
-+	if (oMatchEnd > oend - (16 - MINMATCH)) {
-+		if (op < oend_w) {
-+			ZSTD_wildcopy(op, match, oend_w - op);
-+			match += oend_w - op;
-+			op = oend_w;
-+		}
-+		while (op < oMatchEnd)
-+			*op++ = *match++;
-+	} else {
-+		ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength - 8); /* works even if matchLength < 8 */
-+	}
-+	return sequenceLength;
-+}
-+
-+static size_t ZSTD_decompressSequencesLong(ZSTD_DCtx *dctx, void *dst, size_t maxDstSize, const void *seqStart, size_t seqSize)
-+{
-+	const BYTE *ip = (const BYTE *)seqStart;
-+	const BYTE *const iend = ip + seqSize;
-+	BYTE *const ostart = (BYTE * const)dst;
-+	BYTE *const oend = ostart + maxDstSize;
-+	BYTE *op = ostart;
-+	const BYTE *litPtr = dctx->litPtr;
-+	const BYTE *const litEnd = litPtr + dctx->litSize;
-+	const BYTE *const base = (const BYTE *)(dctx->base);
-+	const BYTE *const vBase = (const BYTE *)(dctx->vBase);
-+	const BYTE *const dictEnd = (const BYTE *)(dctx->dictEnd);
-+	unsigned const windowSize = dctx->fParams.windowSize;
-+	int nbSeq;
-+
-+	/* Build Decoding Tables */
-+	{
-+		size_t const seqHSize = ZSTD_decodeSeqHeaders(dctx, &nbSeq, ip, seqSize);
-+		if (ZSTD_isError(seqHSize))
-+			return seqHSize;
-+		ip += seqHSize;
-+	}
-+
-+	/* Regen sequences */
-+	if (nbSeq) {
-+#define STORED_SEQS 4
-+#define STOSEQ_MASK (STORED_SEQS - 1)
-+#define ADVANCED_SEQS 4
-+		seq_t *sequences = (seq_t *)dctx->entropy.workspace;
-+		int const seqAdvance = MIN(nbSeq, ADVANCED_SEQS);
-+		seqState_t seqState;
-+		int seqNb;
-+		ZSTD_STATIC_ASSERT(sizeof(dctx->entropy.workspace) >= sizeof(seq_t) * STORED_SEQS);
-+		dctx->fseEntropy = 1;
-+		{
-+			U32 i;
-+			for (i = 0; i < ZSTD_REP_NUM; i++)
-+				seqState.prevOffset[i] = dctx->entropy.rep[i];
-+		}
-+		seqState.base = base;
-+		seqState.pos = (size_t)(op - base);
-+		seqState.gotoDict = (uPtrDiff)dictEnd - (uPtrDiff)base; /* cast to avoid undefined behaviour */
-+		CHECK_E(BIT_initDStream(&seqState.DStream, ip, iend - ip), corruption_detected);
-+		FSE_initDState(&seqState.stateLL, &seqState.DStream, dctx->LLTptr);
-+		FSE_initDState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr);
-+		FSE_initDState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);
-+
-+		/* prepare in advance */
-+		for (seqNb = 0; (BIT_reloadDStream(&seqState.DStream) <= BIT_DStream_completed) && seqNb < seqAdvance; seqNb++) {
-+			sequences[seqNb] = ZSTD_decodeSequenceLong(&seqState, windowSize);
-+		}
-+		if (seqNb < seqAdvance)
-+			return ERROR(corruption_detected);
-+
-+		/* decode and decompress */
-+		for (; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && seqNb < nbSeq; seqNb++) {
-+			seq_t const sequence = ZSTD_decodeSequenceLong(&seqState, windowSize);
-+			size_t const oneSeqSize =
-+			    ZSTD_execSequenceLong(op, oend, sequences[(seqNb - ADVANCED_SEQS) & STOSEQ_MASK], &litPtr, litEnd, base, vBase, dictEnd);
-+			if (ZSTD_isError(oneSeqSize))
-+				return oneSeqSize;
-+			ZSTD_PREFETCH(sequence.match);
-+			sequences[seqNb & STOSEQ_MASK] = sequence;
-+			op += oneSeqSize;
-+		}
-+		if (seqNb < nbSeq)
-+			return ERROR(corruption_detected);
-+
-+		/* finish queue */
-+		seqNb -= seqAdvance;
-+		for (; seqNb < nbSeq; seqNb++) {
-+			size_t const oneSeqSize = ZSTD_execSequenceLong(op, oend, sequences[seqNb & STOSEQ_MASK], &litPtr, litEnd, base, vBase, dictEnd);
-+			if (ZSTD_isError(oneSeqSize))
-+				return oneSeqSize;
-+			op += oneSeqSize;
-+		}
-+
-+		/* save reps for next block */
-+		{
-+			U32 i;
-+			for (i = 0; i < ZSTD_REP_NUM; i++)
-+				dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]);
-+		}
-+	}
-+
-+	/* last literal segment */
-+	{
-+		size_t const lastLLSize = litEnd - litPtr;
-+		if (lastLLSize > (size_t)(oend - op))
-+			return ERROR(dstSize_tooSmall);
-+		memcpy(op, litPtr, lastLLSize);
-+		op += lastLLSize;
-+	}
-+
-+	return op - ostart;
-+}
-+
-+static size_t ZSTD_decompressBlock_internal(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-+{ /* blockType == blockCompressed */
-+	const BYTE *ip = (const BYTE *)src;
-+
-+	if (srcSize >= ZSTD_BLOCKSIZE_ABSOLUTEMAX)
-+		return ERROR(srcSize_wrong);
-+
-+	/* Decode literals section */
-+	{
-+		size_t const litCSize = ZSTD_decodeLiteralsBlock(dctx, src, srcSize);
-+		if (ZSTD_isError(litCSize))
-+			return litCSize;
-+		ip += litCSize;
-+		srcSize -= litCSize;
-+	}
-+	if (sizeof(size_t) > 4) /* do not enable prefetching on 32-bits x86, as it's performance detrimental */
-+				/* likely because of register pressure */
-+				/* if that's the correct cause, then 32-bits ARM should be affected differently */
-+				/* it would be good to test this on ARM real hardware, to see if prefetch version improves speed */
-+		if (dctx->fParams.windowSize > (1 << 23))
-+			return ZSTD_decompressSequencesLong(dctx, dst, dstCapacity, ip, srcSize);
-+	return ZSTD_decompressSequences(dctx, dst, dstCapacity, ip, srcSize);
-+}
-+
-+static void ZSTD_checkContinuity(ZSTD_DCtx *dctx, const void *dst)
-+{
-+	if (dst != dctx->previousDstEnd) { /* not contiguous */
-+		dctx->dictEnd = dctx->previousDstEnd;
-+		dctx->vBase = (const char *)dst - ((const char *)(dctx->previousDstEnd) - (const char *)(dctx->base));
-+		dctx->base = dst;
-+		dctx->previousDstEnd = dst;
-+	}
-+}
-+
-+size_t ZSTD_decompressBlock(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-+{
-+	size_t dSize;
-+	ZSTD_checkContinuity(dctx, dst);
-+	dSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize);
-+	dctx->previousDstEnd = (char *)dst + dSize;
-+	return dSize;
-+}
-+
-+/** ZSTD_insertBlock() :
-+	insert `src` block into `dctx` history. Useful to track uncompressed blocks. */
-+size_t ZSTD_insertBlock(ZSTD_DCtx *dctx, const void *blockStart, size_t blockSize)
-+{
-+	ZSTD_checkContinuity(dctx, blockStart);
-+	dctx->previousDstEnd = (const char *)blockStart + blockSize;
-+	return blockSize;
-+}
-+
-+size_t ZSTD_generateNxBytes(void *dst, size_t dstCapacity, BYTE byte, size_t length)
-+{
-+	if (length > dstCapacity)
-+		return ERROR(dstSize_tooSmall);
-+	memset(dst, byte, length);
-+	return length;
-+}
-+
-+/** ZSTD_findFrameCompressedSize() :
-+ *  compatible with legacy mode
-+ *  `src` must point to the start of a ZSTD frame, ZSTD legacy frame, or skippable frame
-+ *  `srcSize` must be at least as large as the frame contained
-+ *  @return : the compressed size of the frame starting at `src` */
-+size_t ZSTD_findFrameCompressedSize(const void *src, size_t srcSize)
-+{
-+	if (srcSize >= ZSTD_skippableHeaderSize && (ZSTD_readLE32(src) & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) {
-+		return ZSTD_skippableHeaderSize + ZSTD_readLE32((const BYTE *)src + 4);
-+	} else {
-+		const BYTE *ip = (const BYTE *)src;
-+		const BYTE *const ipstart = ip;
-+		size_t remainingSize = srcSize;
-+		ZSTD_frameParams fParams;
-+
-+		size_t const headerSize = ZSTD_frameHeaderSize(ip, remainingSize);
-+		if (ZSTD_isError(headerSize))
-+			return headerSize;
-+
-+		/* Frame Header */
-+		{
-+			size_t const ret = ZSTD_getFrameParams(&fParams, ip, remainingSize);
-+			if (ZSTD_isError(ret))
-+				return ret;
-+			if (ret > 0)
-+				return ERROR(srcSize_wrong);
-+		}
-+
-+		ip += headerSize;
-+		remainingSize -= headerSize;
-+
-+		/* Loop on each block */
-+		while (1) {
-+			blockProperties_t blockProperties;
-+			size_t const cBlockSize = ZSTD_getcBlockSize(ip, remainingSize, &blockProperties);
-+			if (ZSTD_isError(cBlockSize))
-+				return cBlockSize;
-+
-+			if (ZSTD_blockHeaderSize + cBlockSize > remainingSize)
-+				return ERROR(srcSize_wrong);
-+
-+			ip += ZSTD_blockHeaderSize + cBlockSize;
-+			remainingSize -= ZSTD_blockHeaderSize + cBlockSize;
-+
-+			if (blockProperties.lastBlock)
-+				break;
-+		}
-+
-+		if (fParams.checksumFlag) { /* Frame content checksum */
-+			if (remainingSize < 4)
-+				return ERROR(srcSize_wrong);
-+			ip += 4;
-+			remainingSize -= 4;
-+		}
-+
-+		return ip - ipstart;
-+	}
-+}
-+
-+/*! ZSTD_decompressFrame() :
-+*   @dctx must be properly initialized */
-+static size_t ZSTD_decompressFrame(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void **srcPtr, size_t *srcSizePtr)
-+{
-+	const BYTE *ip = (const BYTE *)(*srcPtr);
-+	BYTE *const ostart = (BYTE * const)dst;
-+	BYTE *const oend = ostart + dstCapacity;
-+	BYTE *op = ostart;
-+	size_t remainingSize = *srcSizePtr;
-+
-+	/* check */
-+	if (remainingSize < ZSTD_frameHeaderSize_min + ZSTD_blockHeaderSize)
-+		return ERROR(srcSize_wrong);
-+
-+	/* Frame Header */
-+	{
-+		size_t const frameHeaderSize = ZSTD_frameHeaderSize(ip, ZSTD_frameHeaderSize_prefix);
-+		if (ZSTD_isError(frameHeaderSize))
-+			return frameHeaderSize;
-+		if (remainingSize < frameHeaderSize + ZSTD_blockHeaderSize)
-+			return ERROR(srcSize_wrong);
-+		CHECK_F(ZSTD_decodeFrameHeader(dctx, ip, frameHeaderSize));
-+		ip += frameHeaderSize;
-+		remainingSize -= frameHeaderSize;
-+	}
-+
-+	/* Loop on each block */
-+	while (1) {
-+		size_t decodedSize;
-+		blockProperties_t blockProperties;
-+		size_t const cBlockSize = ZSTD_getcBlockSize(ip, remainingSize, &blockProperties);
-+		if (ZSTD_isError(cBlockSize))
-+			return cBlockSize;
-+
-+		ip += ZSTD_blockHeaderSize;
-+		remainingSize -= ZSTD_blockHeaderSize;
-+		if (cBlockSize > remainingSize)
-+			return ERROR(srcSize_wrong);
-+
-+		switch (blockProperties.blockType) {
-+		case bt_compressed: decodedSize = ZSTD_decompressBlock_internal(dctx, op, oend - op, ip, cBlockSize); break;
-+		case bt_raw: decodedSize = ZSTD_copyRawBlock(op, oend - op, ip, cBlockSize); break;
-+		case bt_rle: decodedSize = ZSTD_generateNxBytes(op, oend - op, *ip, blockProperties.origSize); break;
-+		case bt_reserved:
-+		default: return ERROR(corruption_detected);
-+		}
-+
-+		if (ZSTD_isError(decodedSize))
-+			return decodedSize;
-+		if (dctx->fParams.checksumFlag)
-+			xxh64_update(&dctx->xxhState, op, decodedSize);
-+		op += decodedSize;
-+		ip += cBlockSize;
-+		remainingSize -= cBlockSize;
-+		if (blockProperties.lastBlock)
-+			break;
-+	}
-+
-+	if (dctx->fParams.checksumFlag) { /* Frame content checksum verification */
-+		U32 const checkCalc = (U32)xxh64_digest(&dctx->xxhState);
-+		U32 checkRead;
-+		if (remainingSize < 4)
-+			return ERROR(checksum_wrong);
-+		checkRead = ZSTD_readLE32(ip);
-+		if (checkRead != checkCalc)
-+			return ERROR(checksum_wrong);
-+		ip += 4;
-+		remainingSize -= 4;
-+	}
-+
-+	/* Allow caller to get size read */
-+	*srcPtr = ip;
-+	*srcSizePtr = remainingSize;
-+	return op - ostart;
-+}
-+
-+static const void *ZSTD_DDictDictContent(const ZSTD_DDict *ddict);
-+static size_t ZSTD_DDictDictSize(const ZSTD_DDict *ddict);
-+
-+static size_t ZSTD_decompressMultiFrame(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, const void *dict, size_t dictSize,
-+					const ZSTD_DDict *ddict)
-+{
-+	void *const dststart = dst;
-+
-+	if (ddict) {
-+		if (dict) {
-+			/* programmer error, these two cases should be mutually exclusive */
-+			return ERROR(GENERIC);
-+		}
-+
-+		dict = ZSTD_DDictDictContent(ddict);
-+		dictSize = ZSTD_DDictDictSize(ddict);
-+	}
-+
-+	while (srcSize >= ZSTD_frameHeaderSize_prefix) {
-+		U32 magicNumber;
-+
-+		magicNumber = ZSTD_readLE32(src);
-+		if (magicNumber != ZSTD_MAGICNUMBER) {
-+			if ((magicNumber & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) {
-+				size_t skippableSize;
-+				if (srcSize < ZSTD_skippableHeaderSize)
-+					return ERROR(srcSize_wrong);
-+				skippableSize = ZSTD_readLE32((const BYTE *)src + 4) + ZSTD_skippableHeaderSize;
-+				if (srcSize < skippableSize) {
-+					return ERROR(srcSize_wrong);
-+				}
-+
-+				src = (const BYTE *)src + skippableSize;
-+				srcSize -= skippableSize;
-+				continue;
-+			} else {
-+				return ERROR(prefix_unknown);
-+			}
-+		}
-+
-+		if (ddict) {
-+			/* we were called from ZSTD_decompress_usingDDict */
-+			ZSTD_refDDict(dctx, ddict);
-+		} else {
-+			/* this will initialize correctly with no dict if dict == NULL, so
-+			 * use this in all cases but ddict */
-+			CHECK_F(ZSTD_decompressBegin_usingDict(dctx, dict, dictSize));
-+		}
-+		ZSTD_checkContinuity(dctx, dst);
-+
-+		{
-+			const size_t res = ZSTD_decompressFrame(dctx, dst, dstCapacity, &src, &srcSize);
-+			if (ZSTD_isError(res))
-+				return res;
-+			/* don't need to bounds check this, ZSTD_decompressFrame will have
-+			 * already */
-+			dst = (BYTE *)dst + res;
-+			dstCapacity -= res;
-+		}
-+	}
-+
-+	if (srcSize)
-+		return ERROR(srcSize_wrong); /* input not entirely consumed */
-+
-+	return (BYTE *)dst - (BYTE *)dststart;
-+}
-+
-+size_t ZSTD_decompress_usingDict(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, const void *dict, size_t dictSize)
-+{
-+	return ZSTD_decompressMultiFrame(dctx, dst, dstCapacity, src, srcSize, dict, dictSize, NULL);
-+}
-+
-+size_t ZSTD_decompressDCtx(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-+{
-+	return ZSTD_decompress_usingDict(dctx, dst, dstCapacity, src, srcSize, NULL, 0);
-+}
-+
-+/*-**************************************
-+*   Advanced Streaming Decompression API
-+*   Bufferless and synchronous
-+****************************************/
-+size_t ZSTD_nextSrcSizeToDecompress(ZSTD_DCtx *dctx) { return dctx->expected; }
-+
-+ZSTD_nextInputType_e ZSTD_nextInputType(ZSTD_DCtx *dctx)
-+{
-+	switch (dctx->stage) {
-+	default: /* should not happen */
-+	case ZSTDds_getFrameHeaderSize:
-+	case ZSTDds_decodeFrameHeader: return ZSTDnit_frameHeader;
-+	case ZSTDds_decodeBlockHeader: return ZSTDnit_blockHeader;
-+	case ZSTDds_decompressBlock: return ZSTDnit_block;
-+	case ZSTDds_decompressLastBlock: return ZSTDnit_lastBlock;
-+	case ZSTDds_checkChecksum: return ZSTDnit_checksum;
-+	case ZSTDds_decodeSkippableHeader:
-+	case ZSTDds_skipFrame: return ZSTDnit_skippableFrame;
-+	}
-+}
-+
-+int ZSTD_isSkipFrame(ZSTD_DCtx *dctx) { return dctx->stage == ZSTDds_skipFrame; } /* for zbuff */
-+
-+/** ZSTD_decompressContinue() :
-+*   @return : nb of bytes generated into `dst` (necessarily <= `dstCapacity)
-+*             or an error code, which can be tested using ZSTD_isError() */
-+size_t ZSTD_decompressContinue(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-+{
-+	/* Sanity check */
-+	if (srcSize != dctx->expected)
-+		return ERROR(srcSize_wrong);
-+	if (dstCapacity)
-+		ZSTD_checkContinuity(dctx, dst);
-+
-+	switch (dctx->stage) {
-+	case ZSTDds_getFrameHeaderSize:
-+		if (srcSize != ZSTD_frameHeaderSize_prefix)
-+			return ERROR(srcSize_wrong);					/* impossible */
-+		if ((ZSTD_readLE32(src) & 0xFFFFFFF0U) == ZSTD_MAGIC_SKIPPABLE_START) { /* skippable frame */
-+			memcpy(dctx->headerBuffer, src, ZSTD_frameHeaderSize_prefix);
-+			dctx->expected = ZSTD_skippableHeaderSize - ZSTD_frameHeaderSize_prefix; /* magic number + skippable frame length */
-+			dctx->stage = ZSTDds_decodeSkippableHeader;
-+			return 0;
-+		}
-+		dctx->headerSize = ZSTD_frameHeaderSize(src, ZSTD_frameHeaderSize_prefix);
-+		if (ZSTD_isError(dctx->headerSize))
-+			return dctx->headerSize;
-+		memcpy(dctx->headerBuffer, src, ZSTD_frameHeaderSize_prefix);
-+		if (dctx->headerSize > ZSTD_frameHeaderSize_prefix) {
-+			dctx->expected = dctx->headerSize - ZSTD_frameHeaderSize_prefix;
-+			dctx->stage = ZSTDds_decodeFrameHeader;
-+			return 0;
-+		}
-+		dctx->expected = 0; /* not necessary to copy more */
-+
-+	case ZSTDds_decodeFrameHeader:
-+		memcpy(dctx->headerBuffer + ZSTD_frameHeaderSize_prefix, src, dctx->expected);
-+		CHECK_F(ZSTD_decodeFrameHeader(dctx, dctx->headerBuffer, dctx->headerSize));
-+		dctx->expected = ZSTD_blockHeaderSize;
-+		dctx->stage = ZSTDds_decodeBlockHeader;
-+		return 0;
-+
-+	case ZSTDds_decodeBlockHeader: {
-+		blockProperties_t bp;
-+		size_t const cBlockSize = ZSTD_getcBlockSize(src, ZSTD_blockHeaderSize, &bp);
-+		if (ZSTD_isError(cBlockSize))
-+			return cBlockSize;
-+		dctx->expected = cBlockSize;
-+		dctx->bType = bp.blockType;
-+		dctx->rleSize = bp.origSize;
-+		if (cBlockSize) {
-+			dctx->stage = bp.lastBlock ? ZSTDds_decompressLastBlock : ZSTDds_decompressBlock;
-+			return 0;
-+		}
-+		/* empty block */
-+		if (bp.lastBlock) {
-+			if (dctx->fParams.checksumFlag) {
-+				dctx->expected = 4;
-+				dctx->stage = ZSTDds_checkChecksum;
-+			} else {
-+				dctx->expected = 0; /* end of frame */
-+				dctx->stage = ZSTDds_getFrameHeaderSize;
-+			}
-+		} else {
-+			dctx->expected = 3; /* go directly to next header */
-+			dctx->stage = ZSTDds_decodeBlockHeader;
-+		}
-+		return 0;
-+	}
-+	case ZSTDds_decompressLastBlock:
-+	case ZSTDds_decompressBlock: {
-+		size_t rSize;
-+		switch (dctx->bType) {
-+		case bt_compressed: rSize = ZSTD_decompressBlock_internal(dctx, dst, dstCapacity, src, srcSize); break;
-+		case bt_raw: rSize = ZSTD_copyRawBlock(dst, dstCapacity, src, srcSize); break;
-+		case bt_rle: rSize = ZSTD_setRleBlock(dst, dstCapacity, src, srcSize, dctx->rleSize); break;
-+		case bt_reserved: /* should never happen */
-+		default: return ERROR(corruption_detected);
-+		}
-+		if (ZSTD_isError(rSize))
-+			return rSize;
-+		if (dctx->fParams.checksumFlag)
-+			xxh64_update(&dctx->xxhState, dst, rSize);
-+
-+		if (dctx->stage == ZSTDds_decompressLastBlock) { /* end of frame */
-+			if (dctx->fParams.checksumFlag) {	/* another round for frame checksum */
-+				dctx->expected = 4;
-+				dctx->stage = ZSTDds_checkChecksum;
-+			} else {
-+				dctx->expected = 0; /* ends here */
-+				dctx->stage = ZSTDds_getFrameHeaderSize;
-+			}
-+		} else {
-+			dctx->stage = ZSTDds_decodeBlockHeader;
-+			dctx->expected = ZSTD_blockHeaderSize;
-+			dctx->previousDstEnd = (char *)dst + rSize;
-+		}
-+		return rSize;
-+	}
-+	case ZSTDds_checkChecksum: {
-+		U32 const h32 = (U32)xxh64_digest(&dctx->xxhState);
-+		U32 const check32 = ZSTD_readLE32(src); /* srcSize == 4, guaranteed by dctx->expected */
-+		if (check32 != h32)
-+			return ERROR(checksum_wrong);
-+		dctx->expected = 0;
-+		dctx->stage = ZSTDds_getFrameHeaderSize;
-+		return 0;
-+	}
-+	case ZSTDds_decodeSkippableHeader: {
-+		memcpy(dctx->headerBuffer + ZSTD_frameHeaderSize_prefix, src, dctx->expected);
-+		dctx->expected = ZSTD_readLE32(dctx->headerBuffer + 4);
-+		dctx->stage = ZSTDds_skipFrame;
-+		return 0;
-+	}
-+	case ZSTDds_skipFrame: {
-+		dctx->expected = 0;
-+		dctx->stage = ZSTDds_getFrameHeaderSize;
-+		return 0;
-+	}
-+	default:
-+		return ERROR(GENERIC); /* impossible */
-+	}
-+}
-+
-+static size_t ZSTD_refDictContent(ZSTD_DCtx *dctx, const void *dict, size_t dictSize)
-+{
-+	dctx->dictEnd = dctx->previousDstEnd;
-+	dctx->vBase = (const char *)dict - ((const char *)(dctx->previousDstEnd) - (const char *)(dctx->base));
-+	dctx->base = dict;
-+	dctx->previousDstEnd = (const char *)dict + dictSize;
-+	return 0;
-+}
-+
-+/* ZSTD_loadEntropy() :
-+ * dict : must point at beginning of a valid zstd dictionary
-+ * @return : size of entropy tables read */
-+static size_t ZSTD_loadEntropy(ZSTD_entropyTables_t *entropy, const void *const dict, size_t const dictSize)
-+{
-+	const BYTE *dictPtr = (const BYTE *)dict;
-+	const BYTE *const dictEnd = dictPtr + dictSize;
-+
-+	if (dictSize <= 8)
-+		return ERROR(dictionary_corrupted);
-+	dictPtr += 8; /* skip header = magic + dictID */
-+
-+	{
-+		size_t const hSize = HUF_readDTableX4_wksp(entropy->hufTable, dictPtr, dictEnd - dictPtr, entropy->workspace, sizeof(entropy->workspace));
-+		if (HUF_isError(hSize))
-+			return ERROR(dictionary_corrupted);
-+		dictPtr += hSize;
-+	}
-+
-+	{
-+		short offcodeNCount[MaxOff + 1];
-+		U32 offcodeMaxValue = MaxOff, offcodeLog;
-+		size_t const offcodeHeaderSize = FSE_readNCount(offcodeNCount, &offcodeMaxValue, &offcodeLog, dictPtr, dictEnd - dictPtr);
-+		if (FSE_isError(offcodeHeaderSize))
-+			return ERROR(dictionary_corrupted);
-+		if (offcodeLog > OffFSELog)
-+			return ERROR(dictionary_corrupted);
-+		CHECK_E(FSE_buildDTable_wksp(entropy->OFTable, offcodeNCount, offcodeMaxValue, offcodeLog, entropy->workspace, sizeof(entropy->workspace)), dictionary_corrupted);
-+		dictPtr += offcodeHeaderSize;
-+	}
-+
-+	{
-+		short matchlengthNCount[MaxML + 1];
-+		unsigned matchlengthMaxValue = MaxML, matchlengthLog;
-+		size_t const matchlengthHeaderSize = FSE_readNCount(matchlengthNCount, &matchlengthMaxValue, &matchlengthLog, dictPtr, dictEnd - dictPtr);
-+		if (FSE_isError(matchlengthHeaderSize))
-+			return ERROR(dictionary_corrupted);
-+		if (matchlengthLog > MLFSELog)
-+			return ERROR(dictionary_corrupted);
-+		CHECK_E(FSE_buildDTable_wksp(entropy->MLTable, matchlengthNCount, matchlengthMaxValue, matchlengthLog, entropy->workspace, sizeof(entropy->workspace)), dictionary_corrupted);
-+		dictPtr += matchlengthHeaderSize;
-+	}
-+
-+	{
-+		short litlengthNCount[MaxLL + 1];
-+		unsigned litlengthMaxValue = MaxLL, litlengthLog;
-+		size_t const litlengthHeaderSize = FSE_readNCount(litlengthNCount, &litlengthMaxValue, &litlengthLog, dictPtr, dictEnd - dictPtr);
-+		if (FSE_isError(litlengthHeaderSize))
-+			return ERROR(dictionary_corrupted);
-+		if (litlengthLog > LLFSELog)
-+			return ERROR(dictionary_corrupted);
-+		CHECK_E(FSE_buildDTable_wksp(entropy->LLTable, litlengthNCount, litlengthMaxValue, litlengthLog, entropy->workspace, sizeof(entropy->workspace)), dictionary_corrupted);
-+		dictPtr += litlengthHeaderSize;
-+	}
-+
-+	if (dictPtr + 12 > dictEnd)
-+		return ERROR(dictionary_corrupted);
-+	{
-+		int i;
-+		size_t const dictContentSize = (size_t)(dictEnd - (dictPtr + 12));
-+		for (i = 0; i < 3; i++) {
-+			U32 const rep = ZSTD_readLE32(dictPtr);
-+			dictPtr += 4;
-+			if (rep == 0 || rep >= dictContentSize)
-+				return ERROR(dictionary_corrupted);
-+			entropy->rep[i] = rep;
-+		}
-+	}
-+
-+	return dictPtr - (const BYTE *)dict;
-+}
-+
-+static size_t ZSTD_decompress_insertDictionary(ZSTD_DCtx *dctx, const void *dict, size_t dictSize)
-+{
-+	if (dictSize < 8)
-+		return ZSTD_refDictContent(dctx, dict, dictSize);
-+	{
-+		U32 const magic = ZSTD_readLE32(dict);
-+		if (magic != ZSTD_DICT_MAGIC) {
-+			return ZSTD_refDictContent(dctx, dict, dictSize); /* pure content mode */
-+		}
-+	}
-+	dctx->dictID = ZSTD_readLE32((const char *)dict + 4);
-+
-+	/* load entropy tables */
-+	{
-+		size_t const eSize = ZSTD_loadEntropy(&dctx->entropy, dict, dictSize);
-+		if (ZSTD_isError(eSize))
-+			return ERROR(dictionary_corrupted);
-+		dict = (const char *)dict + eSize;
-+		dictSize -= eSize;
-+	}
-+	dctx->litEntropy = dctx->fseEntropy = 1;
-+
-+	/* reference dictionary content */
-+	return ZSTD_refDictContent(dctx, dict, dictSize);
-+}
-+
-+size_t ZSTD_decompressBegin_usingDict(ZSTD_DCtx *dctx, const void *dict, size_t dictSize)
-+{
-+	CHECK_F(ZSTD_decompressBegin(dctx));
-+	if (dict && dictSize)
-+		CHECK_E(ZSTD_decompress_insertDictionary(dctx, dict, dictSize), dictionary_corrupted);
-+	return 0;
-+}
-+
-+/* ======   ZSTD_DDict   ====== */
-+
-+struct ZSTD_DDict_s {
-+	void *dictBuffer;
-+	const void *dictContent;
-+	size_t dictSize;
-+	ZSTD_entropyTables_t entropy;
-+	U32 dictID;
-+	U32 entropyPresent;
-+	ZSTD_customMem cMem;
-+}; /* typedef'd to ZSTD_DDict within "zstd.h" */
-+
-+size_t ZSTD_DDictWorkspaceBound(void) { return ZSTD_ALIGN(sizeof(ZSTD_stack)) + ZSTD_ALIGN(sizeof(ZSTD_DDict)); }
-+
-+static const void *ZSTD_DDictDictContent(const ZSTD_DDict *ddict) { return ddict->dictContent; }
-+
-+static size_t ZSTD_DDictDictSize(const ZSTD_DDict *ddict) { return ddict->dictSize; }
-+
-+static void ZSTD_refDDict(ZSTD_DCtx *dstDCtx, const ZSTD_DDict *ddict)
-+{
-+	ZSTD_decompressBegin(dstDCtx); /* init */
-+	if (ddict) {		       /* support refDDict on NULL */
-+		dstDCtx->dictID = ddict->dictID;
-+		dstDCtx->base = ddict->dictContent;
-+		dstDCtx->vBase = ddict->dictContent;
-+		dstDCtx->dictEnd = (const BYTE *)ddict->dictContent + ddict->dictSize;
-+		dstDCtx->previousDstEnd = dstDCtx->dictEnd;
-+		if (ddict->entropyPresent) {
-+			dstDCtx->litEntropy = 1;
-+			dstDCtx->fseEntropy = 1;
-+			dstDCtx->LLTptr = ddict->entropy.LLTable;
-+			dstDCtx->MLTptr = ddict->entropy.MLTable;
-+			dstDCtx->OFTptr = ddict->entropy.OFTable;
-+			dstDCtx->HUFptr = ddict->entropy.hufTable;
-+			dstDCtx->entropy.rep[0] = ddict->entropy.rep[0];
-+			dstDCtx->entropy.rep[1] = ddict->entropy.rep[1];
-+			dstDCtx->entropy.rep[2] = ddict->entropy.rep[2];
-+		} else {
-+			dstDCtx->litEntropy = 0;
-+			dstDCtx->fseEntropy = 0;
-+		}
-+	}
-+}
-+
-+static size_t ZSTD_loadEntropy_inDDict(ZSTD_DDict *ddict)
-+{
-+	ddict->dictID = 0;
-+	ddict->entropyPresent = 0;
-+	if (ddict->dictSize < 8)
-+		return 0;
-+	{
-+		U32 const magic = ZSTD_readLE32(ddict->dictContent);
-+		if (magic != ZSTD_DICT_MAGIC)
-+			return 0; /* pure content mode */
-+	}
-+	ddict->dictID = ZSTD_readLE32((const char *)ddict->dictContent + 4);
-+
-+	/* load entropy tables */
-+	CHECK_E(ZSTD_loadEntropy(&ddict->entropy, ddict->dictContent, ddict->dictSize), dictionary_corrupted);
-+	ddict->entropyPresent = 1;
-+	return 0;
-+}
-+
-+static ZSTD_DDict *ZSTD_createDDict_advanced(const void *dict, size_t dictSize, unsigned byReference, ZSTD_customMem customMem)
-+{
-+	if (!customMem.customAlloc || !customMem.customFree)
-+		return NULL;
-+
-+	{
-+		ZSTD_DDict *const ddict = (ZSTD_DDict *)ZSTD_malloc(sizeof(ZSTD_DDict), customMem);
-+		if (!ddict)
-+			return NULL;
-+		ddict->cMem = customMem;
-+
-+		if ((byReference) || (!dict) || (!dictSize)) {
-+			ddict->dictBuffer = NULL;
-+			ddict->dictContent = dict;
-+		} else {
-+			void *const internalBuffer = ZSTD_malloc(dictSize, customMem);
-+			if (!internalBuffer) {
-+				ZSTD_freeDDict(ddict);
-+				return NULL;
-+			}
-+			memcpy(internalBuffer, dict, dictSize);
-+			ddict->dictBuffer = internalBuffer;
-+			ddict->dictContent = internalBuffer;
-+		}
-+		ddict->dictSize = dictSize;
-+		ddict->entropy.hufTable[0] = (HUF_DTable)((HufLog)*0x1000001); /* cover both little and big endian */
-+		/* parse dictionary content */
-+		{
-+			size_t const errorCode = ZSTD_loadEntropy_inDDict(ddict);
-+			if (ZSTD_isError(errorCode)) {
-+				ZSTD_freeDDict(ddict);
-+				return NULL;
-+			}
-+		}
-+
-+		return ddict;
-+	}
-+}
-+
-+/*! ZSTD_initDDict() :
-+*   Create a digested dictionary, to start decompression without startup delay.
-+*   `dict` content is copied inside DDict.
-+*   Consequently, `dict` can be released after `ZSTD_DDict` creation */
-+ZSTD_DDict *ZSTD_initDDict(const void *dict, size_t dictSize, void *workspace, size_t workspaceSize)
-+{
-+	ZSTD_customMem const stackMem = ZSTD_initStack(workspace, workspaceSize);
-+	return ZSTD_createDDict_advanced(dict, dictSize, 1, stackMem);
-+}
-+
-+size_t ZSTD_freeDDict(ZSTD_DDict *ddict)
-+{
-+	if (ddict == NULL)
-+		return 0; /* support free on NULL */
-+	{
-+		ZSTD_customMem const cMem = ddict->cMem;
-+		ZSTD_free(ddict->dictBuffer, cMem);
-+		ZSTD_free(ddict, cMem);
-+		return 0;
-+	}
-+}
-+
-+/*! ZSTD_getDictID_fromDict() :
-+ *  Provides the dictID stored within dictionary.
-+ *  if @return == 0, the dictionary is not conformant with Zstandard specification.
-+ *  It can still be loaded, but as a content-only dictionary. */
-+unsigned ZSTD_getDictID_fromDict(const void *dict, size_t dictSize)
-+{
-+	if (dictSize < 8)
-+		return 0;
-+	if (ZSTD_readLE32(dict) != ZSTD_DICT_MAGIC)
-+		return 0;
-+	return ZSTD_readLE32((const char *)dict + 4);
-+}
-+
-+/*! ZSTD_getDictID_fromDDict() :
-+ *  Provides the dictID of the dictionary loaded into `ddict`.
-+ *  If @return == 0, the dictionary is not conformant to Zstandard specification, or empty.
-+ *  Non-conformant dictionaries can still be loaded, but as content-only dictionaries. */
-+unsigned ZSTD_getDictID_fromDDict(const ZSTD_DDict *ddict)
-+{
-+	if (ddict == NULL)
-+		return 0;
-+	return ZSTD_getDictID_fromDict(ddict->dictContent, ddict->dictSize);
-+}
-+
-+/*! ZSTD_getDictID_fromFrame() :
-+ *  Provides the dictID required to decompressed the frame stored within `src`.
-+ *  If @return == 0, the dictID could not be decoded.
-+ *  This could for one of the following reasons :
-+ *  - The frame does not require a dictionary to be decoded (most common case).
-+ *  - The frame was built with dictID intentionally removed. Whatever dictionary is necessary is a hidden information.
-+ *    Note : this use case also happens when using a non-conformant dictionary.
-+ *  - `srcSize` is too small, and as a result, the frame header could not be decoded (only possible if `srcSize < ZSTD_FRAMEHEADERSIZE_MAX`).
-+ *  - This is not a Zstandard frame.
-+ *  When identifying the exact failure cause, it's possible to used ZSTD_getFrameParams(), which will provide a more precise error code. */
-+unsigned ZSTD_getDictID_fromFrame(const void *src, size_t srcSize)
-+{
-+	ZSTD_frameParams zfp = {0, 0, 0, 0};
-+	size_t const hError = ZSTD_getFrameParams(&zfp, src, srcSize);
-+	if (ZSTD_isError(hError))
-+		return 0;
-+	return zfp.dictID;
-+}
-+
-+/*! ZSTD_decompress_usingDDict() :
-+*   Decompression using a pre-digested Dictionary
-+*   Use dictionary without significant overhead. */
-+size_t ZSTD_decompress_usingDDict(ZSTD_DCtx *dctx, void *dst, size_t dstCapacity, const void *src, size_t srcSize, const ZSTD_DDict *ddict)
-+{
-+	/* pass content and size in case legacy frames are encountered */
-+	return ZSTD_decompressMultiFrame(dctx, dst, dstCapacity, src, srcSize, NULL, 0, ddict);
-+}
-+
-+/*=====================================
-+*   Streaming decompression
-+*====================================*/
-+
-+typedef enum { zdss_init, zdss_loadHeader, zdss_read, zdss_load, zdss_flush } ZSTD_dStreamStage;
-+
-+/* *** Resource management *** */
-+struct ZSTD_DStream_s {
-+	ZSTD_DCtx *dctx;
-+	ZSTD_DDict *ddictLocal;
-+	const ZSTD_DDict *ddict;
-+	ZSTD_frameParams fParams;
-+	ZSTD_dStreamStage stage;
-+	char *inBuff;
-+	size_t inBuffSize;
-+	size_t inPos;
-+	size_t maxWindowSize;
-+	char *outBuff;
-+	size_t outBuffSize;
-+	size_t outStart;
-+	size_t outEnd;
-+	size_t blockSize;
-+	BYTE headerBuffer[ZSTD_FRAMEHEADERSIZE_MAX]; /* tmp buffer to store frame header */
-+	size_t lhSize;
-+	ZSTD_customMem customMem;
-+	void *legacyContext;
-+	U32 previousLegacyVersion;
-+	U32 legacyVersion;
-+	U32 hostageByte;
-+}; /* typedef'd to ZSTD_DStream within "zstd.h" */
-+
-+size_t ZSTD_DStreamWorkspaceBound(size_t maxWindowSize)
-+{
-+	size_t const blockSize = MIN(maxWindowSize, ZSTD_BLOCKSIZE_ABSOLUTEMAX);
-+	size_t const inBuffSize = blockSize;
-+	size_t const outBuffSize = maxWindowSize + blockSize + WILDCOPY_OVERLENGTH * 2;
-+	return ZSTD_DCtxWorkspaceBound() + ZSTD_ALIGN(sizeof(ZSTD_DStream)) + ZSTD_ALIGN(inBuffSize) + ZSTD_ALIGN(outBuffSize);
-+}
-+
-+static ZSTD_DStream *ZSTD_createDStream_advanced(ZSTD_customMem customMem)
-+{
-+	ZSTD_DStream *zds;
-+
-+	if (!customMem.customAlloc || !customMem.customFree)
-+		return NULL;
-+
-+	zds = (ZSTD_DStream *)ZSTD_malloc(sizeof(ZSTD_DStream), customMem);
-+	if (zds == NULL)
-+		return NULL;
-+	memset(zds, 0, sizeof(ZSTD_DStream));
-+	memcpy(&zds->customMem, &customMem, sizeof(ZSTD_customMem));
-+	zds->dctx = ZSTD_createDCtx_advanced(customMem);
-+	if (zds->dctx == NULL) {
-+		ZSTD_freeDStream(zds);
-+		return NULL;
-+	}
-+	zds->stage = zdss_init;
-+	zds->maxWindowSize = ZSTD_MAXWINDOWSIZE_DEFAULT;
-+	return zds;
-+}
-+
-+ZSTD_DStream *ZSTD_initDStream(size_t maxWindowSize, void *workspace, size_t workspaceSize)
-+{
-+	ZSTD_customMem const stackMem = ZSTD_initStack(workspace, workspaceSize);
-+	ZSTD_DStream *zds = ZSTD_createDStream_advanced(stackMem);
-+	if (!zds) {
-+		return NULL;
-+	}
-+
-+	zds->maxWindowSize = maxWindowSize;
-+	zds->stage = zdss_loadHeader;
-+	zds->lhSize = zds->inPos = zds->outStart = zds->outEnd = 0;
-+	ZSTD_freeDDict(zds->ddictLocal);
-+	zds->ddictLocal = NULL;
-+	zds->ddict = zds->ddictLocal;
-+	zds->legacyVersion = 0;
-+	zds->hostageByte = 0;
-+
-+	{
-+		size_t const blockSize = MIN(zds->maxWindowSize, ZSTD_BLOCKSIZE_ABSOLUTEMAX);
-+		size_t const neededOutSize = zds->maxWindowSize + blockSize + WILDCOPY_OVERLENGTH * 2;
-+
-+		zds->inBuff = (char *)ZSTD_malloc(blockSize, zds->customMem);
-+		zds->inBuffSize = blockSize;
-+		zds->outBuff = (char *)ZSTD_malloc(neededOutSize, zds->customMem);
-+		zds->outBuffSize = neededOutSize;
-+		if (zds->inBuff == NULL || zds->outBuff == NULL) {
-+			ZSTD_freeDStream(zds);
-+			return NULL;
-+		}
-+	}
-+	return zds;
-+}
-+
-+ZSTD_DStream *ZSTD_initDStream_usingDDict(size_t maxWindowSize, const ZSTD_DDict *ddict, void *workspace, size_t workspaceSize)
-+{
-+	ZSTD_DStream *zds = ZSTD_initDStream(maxWindowSize, workspace, workspaceSize);
-+	if (zds) {
-+		zds->ddict = ddict;
-+	}
-+	return zds;
-+}
-+
-+size_t ZSTD_freeDStream(ZSTD_DStream *zds)
-+{
-+	if (zds == NULL)
-+		return 0; /* support free on null */
-+	{
-+		ZSTD_customMem const cMem = zds->customMem;
-+		ZSTD_freeDCtx(zds->dctx);
-+		zds->dctx = NULL;
-+		ZSTD_freeDDict(zds->ddictLocal);
-+		zds->ddictLocal = NULL;
-+		ZSTD_free(zds->inBuff, cMem);
-+		zds->inBuff = NULL;
-+		ZSTD_free(zds->outBuff, cMem);
-+		zds->outBuff = NULL;
-+		ZSTD_free(zds, cMem);
-+		return 0;
-+	}
-+}
-+
-+/* *** Initialization *** */
-+
-+size_t ZSTD_DStreamInSize(void) { return ZSTD_BLOCKSIZE_ABSOLUTEMAX + ZSTD_blockHeaderSize; }
-+size_t ZSTD_DStreamOutSize(void) { return ZSTD_BLOCKSIZE_ABSOLUTEMAX; }
-+
-+size_t ZSTD_resetDStream(ZSTD_DStream *zds)
-+{
-+	zds->stage = zdss_loadHeader;
-+	zds->lhSize = zds->inPos = zds->outStart = zds->outEnd = 0;
-+	zds->legacyVersion = 0;
-+	zds->hostageByte = 0;
-+	return ZSTD_frameHeaderSize_prefix;
-+}
-+
-+/* *****   Decompression   ***** */
-+
-+ZSTD_STATIC size_t ZSTD_limitCopy(void *dst, size_t dstCapacity, const void *src, size_t srcSize)
-+{
-+	size_t const length = MIN(dstCapacity, srcSize);
-+	memcpy(dst, src, length);
-+	return length;
-+}
-+
-+size_t ZSTD_decompressStream(ZSTD_DStream *zds, ZSTD_outBuffer *output, ZSTD_inBuffer *input)
-+{
-+	const char *const istart = (const char *)(input->src) + input->pos;
-+	const char *const iend = (const char *)(input->src) + input->size;
-+	const char *ip = istart;
-+	char *const ostart = (char *)(output->dst) + output->pos;
-+	char *const oend = (char *)(output->dst) + output->size;
-+	char *op = ostart;
-+	U32 someMoreWork = 1;
-+
-+	while (someMoreWork) {
-+		switch (zds->stage) {
-+		case zdss_init:
-+			ZSTD_resetDStream(zds); /* transparent reset on starting decoding a new frame */
-+						/* fall-through */
-+
-+		case zdss_loadHeader: {
-+			size_t const hSize = ZSTD_getFrameParams(&zds->fParams, zds->headerBuffer, zds->lhSize);
-+			if (ZSTD_isError(hSize))
-+				return hSize;
-+			if (hSize != 0) {				   /* need more input */
-+				size_t const toLoad = hSize - zds->lhSize; /* if hSize!=0, hSize > zds->lhSize */
-+				if (toLoad > (size_t)(iend - ip)) {	/* not enough input to load full header */
-+					memcpy(zds->headerBuffer + zds->lhSize, ip, iend - ip);
-+					zds->lhSize += iend - ip;
-+					input->pos = input->size;
-+					return (MAX(ZSTD_frameHeaderSize_min, hSize) - zds->lhSize) +
-+					       ZSTD_blockHeaderSize; /* remaining header bytes + next block header */
-+				}
-+				memcpy(zds->headerBuffer + zds->lhSize, ip, toLoad);
-+				zds->lhSize = hSize;
-+				ip += toLoad;
-+				break;
-+			}
-+
-+			/* check for single-pass mode opportunity */
-+			if (zds->fParams.frameContentSize && zds->fParams.windowSize /* skippable frame if == 0 */
-+			    && (U64)(size_t)(oend - op) >= zds->fParams.frameContentSize) {
-+				size_t const cSize = ZSTD_findFrameCompressedSize(istart, iend - istart);
-+				if (cSize <= (size_t)(iend - istart)) {
-+					size_t const decompressedSize = ZSTD_decompress_usingDDict(zds->dctx, op, oend - op, istart, cSize, zds->ddict);
-+					if (ZSTD_isError(decompressedSize))
-+						return decompressedSize;
-+					ip = istart + cSize;
-+					op += decompressedSize;
-+					zds->dctx->expected = 0;
-+					zds->stage = zdss_init;
-+					someMoreWork = 0;
-+					break;
-+				}
-+			}
-+
-+			/* Consume header */
-+			ZSTD_refDDict(zds->dctx, zds->ddict);
-+			{
-+				size_t const h1Size = ZSTD_nextSrcSizeToDecompress(zds->dctx); /* == ZSTD_frameHeaderSize_prefix */
-+				CHECK_F(ZSTD_decompressContinue(zds->dctx, NULL, 0, zds->headerBuffer, h1Size));
-+				{
-+					size_t const h2Size = ZSTD_nextSrcSizeToDecompress(zds->dctx);
-+					CHECK_F(ZSTD_decompressContinue(zds->dctx, NULL, 0, zds->headerBuffer + h1Size, h2Size));
-+				}
-+			}
-+
-+			zds->fParams.windowSize = MAX(zds->fParams.windowSize, 1U << ZSTD_WINDOWLOG_ABSOLUTEMIN);
-+			if (zds->fParams.windowSize > zds->maxWindowSize)
-+				return ERROR(frameParameter_windowTooLarge);
-+
-+			/* Buffers are preallocated, but double check */
-+			{
-+				size_t const blockSize = MIN(zds->maxWindowSize, ZSTD_BLOCKSIZE_ABSOLUTEMAX);
-+				size_t const neededOutSize = zds->maxWindowSize + blockSize + WILDCOPY_OVERLENGTH * 2;
-+				if (zds->inBuffSize < blockSize) {
-+					return ERROR(GENERIC);
-+				}
-+				if (zds->outBuffSize < neededOutSize) {
-+					return ERROR(GENERIC);
-+				}
-+				zds->blockSize = blockSize;
-+			}
-+			zds->stage = zdss_read;
-+		}
-+		/* pass-through */
-+
-+		case zdss_read: {
-+			size_t const neededInSize = ZSTD_nextSrcSizeToDecompress(zds->dctx);
-+			if (neededInSize == 0) { /* end of frame */
-+				zds->stage = zdss_init;
-+				someMoreWork = 0;
-+				break;
-+			}
-+			if ((size_t)(iend - ip) >= neededInSize) { /* decode directly from src */
-+				const int isSkipFrame = ZSTD_isSkipFrame(zds->dctx);
-+				size_t const decodedSize = ZSTD_decompressContinue(zds->dctx, zds->outBuff + zds->outStart,
-+										   (isSkipFrame ? 0 : zds->outBuffSize - zds->outStart), ip, neededInSize);
-+				if (ZSTD_isError(decodedSize))
-+					return decodedSize;
-+				ip += neededInSize;
-+				if (!decodedSize && !isSkipFrame)
-+					break; /* this was just a header */
-+				zds->outEnd = zds->outStart + decodedSize;
-+				zds->stage = zdss_flush;
-+				break;
-+			}
-+			if (ip == iend) {
-+				someMoreWork = 0;
-+				break;
-+			} /* no more input */
-+			zds->stage = zdss_load;
-+			/* pass-through */
-+		}
-+
-+		case zdss_load: {
-+			size_t const neededInSize = ZSTD_nextSrcSizeToDecompress(zds->dctx);
-+			size_t const toLoad = neededInSize - zds->inPos; /* should always be <= remaining space within inBuff */
-+			size_t loadedSize;
-+			if (toLoad > zds->inBuffSize - zds->inPos)
-+				return ERROR(corruption_detected); /* should never happen */
-+			loadedSize = ZSTD_limitCopy(zds->inBuff + zds->inPos, toLoad, ip, iend - ip);
-+			ip += loadedSize;
-+			zds->inPos += loadedSize;
-+			if (loadedSize < toLoad) {
-+				someMoreWork = 0;
-+				break;
-+			} /* not enough input, wait for more */
-+
-+			/* decode loaded input */
-+			{
-+				const int isSkipFrame = ZSTD_isSkipFrame(zds->dctx);
-+				size_t const decodedSize = ZSTD_decompressContinue(zds->dctx, zds->outBuff + zds->outStart, zds->outBuffSize - zds->outStart,
-+										   zds->inBuff, neededInSize);
-+				if (ZSTD_isError(decodedSize))
-+					return decodedSize;
-+				zds->inPos = 0; /* input is consumed */
-+				if (!decodedSize && !isSkipFrame) {
-+					zds->stage = zdss_read;
-+					break;
-+				} /* this was just a header */
-+				zds->outEnd = zds->outStart + decodedSize;
-+				zds->stage = zdss_flush;
-+				/* pass-through */
-+			}
-+		}
-+
-+		case zdss_flush: {
-+			size_t const toFlushSize = zds->outEnd - zds->outStart;
-+			size_t const flushedSize = ZSTD_limitCopy(op, oend - op, zds->outBuff + zds->outStart, toFlushSize);
-+			op += flushedSize;
-+			zds->outStart += flushedSize;
-+			if (flushedSize == toFlushSize) { /* flush completed */
-+				zds->stage = zdss_read;
-+				if (zds->outStart + zds->blockSize > zds->outBuffSize)
-+					zds->outStart = zds->outEnd = 0;
-+				break;
-+			}
-+			/* cannot complete flush */
-+			someMoreWork = 0;
-+			break;
-+		}
-+		default:
-+			return ERROR(GENERIC); /* impossible */
-+		}
-+	}
-+
-+	/* result */
-+	input->pos += (size_t)(ip - istart);
-+	output->pos += (size_t)(op - ostart);
-+	{
-+		size_t nextSrcSizeHint = ZSTD_nextSrcSizeToDecompress(zds->dctx);
-+		if (!nextSrcSizeHint) {			    /* frame fully decoded */
-+			if (zds->outEnd == zds->outStart) { /* output fully flushed */
-+				if (zds->hostageByte) {
-+					if (input->pos >= input->size) {
-+						zds->stage = zdss_read;
-+						return 1;
-+					}	     /* can't release hostage (not present) */
-+					input->pos++; /* release hostage */
-+				}
-+				return 0;
-+			}
-+			if (!zds->hostageByte) { /* output not fully flushed; keep last byte as hostage; will be released when all output is flushed */
-+				input->pos--;    /* note : pos > 0, otherwise, impossible to finish reading last block */
-+				zds->hostageByte = 1;
-+			}
-+			return 1;
-+		}
-+		nextSrcSizeHint += ZSTD_blockHeaderSize * (ZSTD_nextInputType(zds->dctx) == ZSTDnit_block); /* preload header of next block */
-+		if (zds->inPos > nextSrcSizeHint)
-+			return ERROR(GENERIC); /* should never happen */
-+		nextSrcSizeHint -= zds->inPos; /* already loaded*/
-+		return nextSrcSizeHint;
-+	}
-+}
-+
-+EXPORT_SYMBOL(ZSTD_DCtxWorkspaceBound);
-+EXPORT_SYMBOL(ZSTD_initDCtx);
-+EXPORT_SYMBOL(ZSTD_decompressDCtx);
-+EXPORT_SYMBOL(ZSTD_decompress_usingDict);
-+
-+EXPORT_SYMBOL(ZSTD_DDictWorkspaceBound);
-+EXPORT_SYMBOL(ZSTD_initDDict);
-+EXPORT_SYMBOL(ZSTD_decompress_usingDDict);
-+
-+EXPORT_SYMBOL(ZSTD_DStreamWorkspaceBound);
-+EXPORT_SYMBOL(ZSTD_initDStream);
-+EXPORT_SYMBOL(ZSTD_initDStream_usingDDict);
-+EXPORT_SYMBOL(ZSTD_resetDStream);
-+EXPORT_SYMBOL(ZSTD_decompressStream);
-+EXPORT_SYMBOL(ZSTD_DStreamInSize);
-+EXPORT_SYMBOL(ZSTD_DStreamOutSize);
-+
-+EXPORT_SYMBOL(ZSTD_findFrameCompressedSize);
-+EXPORT_SYMBOL(ZSTD_getFrameContentSize);
-+EXPORT_SYMBOL(ZSTD_findDecompressedSize);
-+
-+EXPORT_SYMBOL(ZSTD_isFrame);
-+EXPORT_SYMBOL(ZSTD_getDictID_fromDict);
-+EXPORT_SYMBOL(ZSTD_getDictID_fromDDict);
-+EXPORT_SYMBOL(ZSTD_getDictID_fromFrame);
-+
-+EXPORT_SYMBOL(ZSTD_getFrameParams);
-+EXPORT_SYMBOL(ZSTD_decompressBegin);
-+EXPORT_SYMBOL(ZSTD_decompressBegin_usingDict);
-+EXPORT_SYMBOL(ZSTD_copyDCtx);
-+EXPORT_SYMBOL(ZSTD_nextSrcSizeToDecompress);
-+EXPORT_SYMBOL(ZSTD_decompressContinue);
-+EXPORT_SYMBOL(ZSTD_nextInputType);
-+
-+EXPORT_SYMBOL(ZSTD_decompressBlock);
-+EXPORT_SYMBOL(ZSTD_insertBlock);
-+
-+MODULE_LICENSE("Dual BSD/GPL");
-+MODULE_DESCRIPTION("Zstd Decompressor");
-diff --git a/lib/zstd/entropy_common.c b/lib/zstd/entropy_common.c
-new file mode 100644
-index 0000000..2b0a643
---- /dev/null
-+++ b/lib/zstd/entropy_common.c
-@@ -0,0 +1,243 @@
-+/*
-+ * Common functions of New Generation Entropy library
-+ * Copyright (C) 2016, Yann Collet.
-+ *
-+ * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
-+ *
-+ * Redistribution and use in source and binary forms, with or without
-+ * modification, are permitted provided that the following conditions are
-+ * met:
-+ *
-+ *   * Redistributions of source code must retain the above copyright
-+ * notice, this list of conditions and the following disclaimer.
-+ *   * Redistributions in binary form must reproduce the above
-+ * copyright notice, this list of conditions and the following disclaimer
-+ * in the documentation and/or other materials provided with the
-+ * distribution.
-+ *
-+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
-+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
-+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
-+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
-+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
-+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
-+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
-+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
-+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
-+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
-+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-+ *
-+ * This program is free software; you can redistribute it and/or modify it under
-+ * the terms of the GNU General Public License version 2 as published by the
-+ * Free Software Foundation. This program is dual-licensed; you may select
-+ * either version 2 of the GNU General Public License ("GPL") or BSD license
-+ * ("BSD").
-+ *
-+ * You can contact the author at :
-+ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
-+ */
-+
-+/* *************************************
-+*  Dependencies
-+***************************************/
-+#include "error_private.h" /* ERR_*, ERROR */
-+#include "fse.h"
-+#include "huf.h"
-+#include "mem.h"
-+
-+/*===   Version   ===*/
-+unsigned FSE_versionNumber(void) { return FSE_VERSION_NUMBER; }
-+
-+/*===   Error Management   ===*/
-+unsigned FSE_isError(size_t code) { return ERR_isError(code); }
-+
-+unsigned HUF_isError(size_t code) { return ERR_isError(code); }
-+
-+/*-**************************************************************
-+*  FSE NCount encoding-decoding
-+****************************************************************/
-+size_t FSE_readNCount(short *normalizedCounter, unsigned *maxSVPtr, unsigned *tableLogPtr, const void *headerBuffer, size_t hbSize)
-+{
-+	const BYTE *const istart = (const BYTE *)headerBuffer;
-+	const BYTE *const iend = istart + hbSize;
-+	const BYTE *ip = istart;
-+	int nbBits;
-+	int remaining;
-+	int threshold;
-+	U32 bitStream;
-+	int bitCount;
-+	unsigned charnum = 0;
-+	int previous0 = 0;
-+
-+	if (hbSize < 4)
-+		return ERROR(srcSize_wrong);
-+	bitStream = ZSTD_readLE32(ip);
-+	nbBits = (bitStream & 0xF) + FSE_MIN_TABLELOG; /* extract tableLog */
-+	if (nbBits > FSE_TABLELOG_ABSOLUTE_MAX)
-+		return ERROR(tableLog_tooLarge);
-+	bitStream >>= 4;
-+	bitCount = 4;
-+	*tableLogPtr = nbBits;
-+	remaining = (1 << nbBits) + 1;
-+	threshold = 1 << nbBits;
-+	nbBits++;
-+
-+	while ((remaining > 1) & (charnum <= *maxSVPtr)) {
-+		if (previous0) {
-+			unsigned n0 = charnum;
-+			while ((bitStream & 0xFFFF) == 0xFFFF) {
-+				n0 += 24;
-+				if (ip < iend - 5) {
-+					ip += 2;
-+					bitStream = ZSTD_readLE32(ip) >> bitCount;
-+				} else {
-+					bitStream >>= 16;
-+					bitCount += 16;
-+				}
-+			}
-+			while ((bitStream & 3) == 3) {
-+				n0 += 3;
-+				bitStream >>= 2;
-+				bitCount += 2;
-+			}
-+			n0 += bitStream & 3;
-+			bitCount += 2;
-+			if (n0 > *maxSVPtr)
-+				return ERROR(maxSymbolValue_tooSmall);
-+			while (charnum < n0)
-+				normalizedCounter[charnum++] = 0;
-+			if ((ip <= iend - 7) || (ip + (bitCount >> 3) <= iend - 4)) {
-+				ip += bitCount >> 3;
-+				bitCount &= 7;
-+				bitStream = ZSTD_readLE32(ip) >> bitCount;
-+			} else {
-+				bitStream >>= 2;
-+			}
-+		}
-+		{
-+			int const max = (2 * threshold - 1) - remaining;
-+			int count;
-+
-+			if ((bitStream & (threshold - 1)) < (U32)max) {
-+				count = bitStream & (threshold - 1);
-+				bitCount += nbBits - 1;
-+			} else {
-+				count = bitStream & (2 * threshold - 1);
-+				if (count >= threshold)
-+					count -= max;
-+				bitCount += nbBits;
-+			}
-+
-+			count--;				 /* extra accuracy */
-+			remaining -= count < 0 ? -count : count; /* -1 means +1 */
-+			normalizedCounter[charnum++] = (short)count;
-+			previous0 = !count;
-+			while (remaining < threshold) {
-+				nbBits--;
-+				threshold >>= 1;
-+			}
-+
-+			if ((ip <= iend - 7) || (ip + (bitCount >> 3) <= iend - 4)) {
-+				ip += bitCount >> 3;
-+				bitCount &= 7;
-+			} else {
-+				bitCount -= (int)(8 * (iend - 4 - ip));
-+				ip = iend - 4;
-+			}
-+			bitStream = ZSTD_readLE32(ip) >> (bitCount & 31);
-+		}
-+	} /* while ((remaining>1) & (charnum<=*maxSVPtr)) */
-+	if (remaining != 1)
-+		return ERROR(corruption_detected);
-+	if (bitCount > 32)
-+		return ERROR(corruption_detected);
-+	*maxSVPtr = charnum - 1;
-+
-+	ip += (bitCount + 7) >> 3;
-+	return ip - istart;
-+}
-+
-+/*! HUF_readStats() :
-+	Read compact Huffman tree, saved by HUF_writeCTable().
-+	`huffWeight` is destination buffer.
-+	`rankStats` is assumed to be a table of at least HUF_TABLELOG_MAX U32.
-+	@return : size read from `src` , or an error Code .
-+	Note : Needed by HUF_readCTable() and HUF_readDTableX?() .
-+*/
-+size_t HUF_readStats_wksp(BYTE *huffWeight, size_t hwSize, U32 *rankStats, U32 *nbSymbolsPtr, U32 *tableLogPtr, const void *src, size_t srcSize, void *workspace, size_t workspaceSize)
-+{
-+	U32 weightTotal;
-+	const BYTE *ip = (const BYTE *)src;
-+	size_t iSize;
-+	size_t oSize;
-+
-+	if (!srcSize)
-+		return ERROR(srcSize_wrong);
-+	iSize = ip[0];
-+	/* memset(huffWeight, 0, hwSize);   */ /* is not necessary, even though some analyzer complain ... */
-+
-+	if (iSize >= 128) { /* special header */
-+		oSize = iSize - 127;
-+		iSize = ((oSize + 1) / 2);
-+		if (iSize + 1 > srcSize)
-+			return ERROR(srcSize_wrong);
-+		if (oSize >= hwSize)
-+			return ERROR(corruption_detected);
-+		ip += 1;
-+		{
-+			U32 n;
-+			for (n = 0; n < oSize; n += 2) {
-+				huffWeight[n] = ip[n / 2] >> 4;
-+				huffWeight[n + 1] = ip[n / 2] & 15;
-+			}
-+		}
-+	} else {						 /* header compressed with FSE (normal case) */
-+		if (iSize + 1 > srcSize)
-+			return ERROR(srcSize_wrong);
-+		oSize = FSE_decompress_wksp(huffWeight, hwSize - 1, ip + 1, iSize, 6, workspace, workspaceSize); /* max (hwSize-1) values decoded, as last one is implied */
-+		if (FSE_isError(oSize))
-+			return oSize;
-+	}
-+
-+	/* collect weight stats */
-+	memset(rankStats, 0, (HUF_TABLELOG_MAX + 1) * sizeof(U32));
-+	weightTotal = 0;
-+	{
-+		U32 n;
-+		for (n = 0; n < oSize; n++) {
-+			if (huffWeight[n] >= HUF_TABLELOG_MAX)
-+				return ERROR(corruption_detected);
-+			rankStats[huffWeight[n]]++;
-+			weightTotal += (1 << huffWeight[n]) >> 1;
-+		}
-+	}
-+	if (weightTotal == 0)
-+		return ERROR(corruption_detected);
-+
-+	/* get last non-null symbol weight (implied, total must be 2^n) */
-+	{
-+		U32 const tableLog = BIT_highbit32(weightTotal) + 1;
-+		if (tableLog > HUF_TABLELOG_MAX)
-+			return ERROR(corruption_detected);
-+		*tableLogPtr = tableLog;
-+		/* determine last weight */
-+		{
-+			U32 const total = 1 << tableLog;
-+			U32 const rest = total - weightTotal;
-+			U32 const verif = 1 << BIT_highbit32(rest);
-+			U32 const lastWeight = BIT_highbit32(rest) + 1;
-+			if (verif != rest)
-+				return ERROR(corruption_detected); /* last value must be a clean power of 2 */
-+			huffWeight[oSize] = (BYTE)lastWeight;
-+			rankStats[lastWeight]++;
-+		}
-+	}
-+
-+	/* check tree construction validity */
-+	if ((rankStats[1] < 2) || (rankStats[1] & 1))
-+		return ERROR(corruption_detected); /* by construction : at least 2 elts of rank 1, must be even */
-+
-+	/* results */
-+	*nbSymbolsPtr = (U32)(oSize + 1);
-+	return iSize + 1;
-+}
-diff --git a/lib/zstd/error_private.h b/lib/zstd/error_private.h
-new file mode 100644
-index 0000000..1a60b31
---- /dev/null
-+++ b/lib/zstd/error_private.h
-@@ -0,0 +1,53 @@
-+/**
-+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
-+ * All rights reserved.
-+ *
-+ * This source code is licensed under the BSD-style license found in the
-+ * LICENSE file in the root directory of https://github.com/facebook/zstd.
-+ * An additional grant of patent rights can be found in the PATENTS file in the
-+ * same directory.
-+ *
-+ * This program is free software; you can redistribute it and/or modify it under
-+ * the terms of the GNU General Public License version 2 as published by the
-+ * Free Software Foundation. This program is dual-licensed; you may select
-+ * either version 2 of the GNU General Public License ("GPL") or BSD license
-+ * ("BSD").
-+ */
-+
-+/* Note : this module is expected to remain private, do not expose it */
-+
-+#ifndef ERROR_H_MODULE
-+#define ERROR_H_MODULE
-+
-+/* ****************************************
-+*  Dependencies
-+******************************************/
-+#include <linux/types.h> /* size_t */
-+#include <linux/zstd.h>  /* enum list */
-+
-+/* ****************************************
-+*  Compiler-specific
-+******************************************/
-+#define ERR_STATIC static __attribute__((unused))
-+
-+/*-****************************************
-+*  Customization (error_public.h)
-+******************************************/
-+typedef ZSTD_ErrorCode ERR_enum;
-+#define PREFIX(name) ZSTD_error_##name
-+
-+/*-****************************************
-+*  Error codes handling
-+******************************************/
-+#define ERROR(name) ((size_t)-PREFIX(name))
-+
-+ERR_STATIC unsigned ERR_isError(size_t code) { return (code > ERROR(maxCode)); }
-+
-+ERR_STATIC ERR_enum ERR_getErrorCode(size_t code)
-+{
-+	if (!ERR_isError(code))
-+		return (ERR_enum)0;
-+	return (ERR_enum)(0 - code);
-+}
-+
-+#endif /* ERROR_H_MODULE */
-diff --git a/lib/zstd/fse.h b/lib/zstd/fse.h
-new file mode 100644
-index 0000000..7460ab0
---- /dev/null
-+++ b/lib/zstd/fse.h
-@@ -0,0 +1,575 @@
-+/*
-+ * FSE : Finite State Entropy codec
-+ * Public Prototypes declaration
-+ * Copyright (C) 2013-2016, Yann Collet.
-+ *
-+ * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
-+ *
-+ * Redistribution and use in source and binary forms, with or without
-+ * modification, are permitted provided that the following conditions are
-+ * met:
-+ *
-+ *   * Redistributions of source code must retain the above copyright
-+ * notice, this list of conditions and the following disclaimer.
-+ *   * Redistributions in binary form must reproduce the above
-+ * copyright notice, this list of conditions and the following disclaimer
-+ * in the documentation and/or other materials provided with the
-+ * distribution.
-+ *
-+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
-+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
-+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
-+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
-+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
-+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
-+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
-+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
-+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
-+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
-+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-+ *
-+ * This program is free software; you can redistribute it and/or modify it under
-+ * the terms of the GNU General Public License version 2 as published by the
-+ * Free Software Foundation. This program is dual-licensed; you may select
-+ * either version 2 of the GNU General Public License ("GPL") or BSD license
-+ * ("BSD").
-+ *
-+ * You can contact the author at :
-+ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
-+ */
-+#ifndef FSE_H
-+#define FSE_H
-+
-+/*-*****************************************
-+*  Dependencies
-+******************************************/
-+#include <linux/types.h> /* size_t, ptrdiff_t */
-+
-+/*-*****************************************
-+*  FSE_PUBLIC_API : control library symbols visibility
-+******************************************/
-+#define FSE_PUBLIC_API
-+
-+/*------   Version   ------*/
-+#define FSE_VERSION_MAJOR 0
-+#define FSE_VERSION_MINOR 9
-+#define FSE_VERSION_RELEASE 0
-+
-+#define FSE_LIB_VERSION FSE_VERSION_MAJOR.FSE_VERSION_MINOR.FSE_VERSION_RELEASE
-+#define FSE_QUOTE(str) #str
-+#define FSE_EXPAND_AND_QUOTE(str) FSE_QUOTE(str)
-+#define FSE_VERSION_STRING FSE_EXPAND_AND_QUOTE(FSE_LIB_VERSION)
-+
-+#define FSE_VERSION_NUMBER (FSE_VERSION_MAJOR * 100 * 100 + FSE_VERSION_MINOR * 100 + FSE_VERSION_RELEASE)
-+FSE_PUBLIC_API unsigned FSE_versionNumber(void); /**< library version number; to be used when checking dll version */
-+
-+/*-*****************************************
-+*  Tool functions
-+******************************************/
-+FSE_PUBLIC_API size_t FSE_compressBound(size_t size); /* maximum compressed size */
-+
-+/* Error Management */
-+FSE_PUBLIC_API unsigned FSE_isError(size_t code); /* tells if a return value is an error code */
-+
-+/*-*****************************************
-+*  FSE detailed API
-+******************************************/
-+/*!
-+FSE_compress() does the following:
-+1. count symbol occurrence from source[] into table count[]
-+2. normalize counters so that sum(count[]) == Power_of_2 (2^tableLog)
-+3. save normalized counters to memory buffer using writeNCount()
-+4. build encoding table 'CTable' from normalized counters
-+5. encode the data stream using encoding table 'CTable'
-+
-+FSE_decompress() does the following:
-+1. read normalized counters with readNCount()
-+2. build decoding table 'DTable' from normalized counters
-+3. decode the data stream using decoding table 'DTable'
-+
-+The following API allows targeting specific sub-functions for advanced tasks.
-+For example, it's possible to compress several blocks using the same 'CTable',
-+or to save and provide normalized distribution using external method.
-+*/
-+
-+/* *** COMPRESSION *** */
-+/*! FSE_optimalTableLog():
-+	dynamically downsize 'tableLog' when conditions are met.
-+	It saves CPU time, by using smaller tables, while preserving or even improving compression ratio.
-+	@return : recommended tableLog (necessarily <= 'maxTableLog') */
-+FSE_PUBLIC_API unsigned FSE_optimalTableLog(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue);
-+
-+/*! FSE_normalizeCount():
-+	normalize counts so that sum(count[]) == Power_of_2 (2^tableLog)
-+	'normalizedCounter' is a table of short, of minimum size (maxSymbolValue+1).
-+	@return : tableLog,
-+			  or an errorCode, which can be tested using FSE_isError() */
-+FSE_PUBLIC_API size_t FSE_normalizeCount(short *normalizedCounter, unsigned tableLog, const unsigned *count, size_t srcSize, unsigned maxSymbolValue);
-+
-+/*! FSE_NCountWriteBound():
-+	Provides the maximum possible size of an FSE normalized table, given 'maxSymbolValue' and 'tableLog'.
-+	Typically useful for allocation purpose. */
-+FSE_PUBLIC_API size_t FSE_NCountWriteBound(unsigned maxSymbolValue, unsigned tableLog);
-+
-+/*! FSE_writeNCount():
-+	Compactly save 'normalizedCounter' into 'buffer'.
-+	@return : size of the compressed table,
-+			  or an errorCode, which can be tested using FSE_isError(). */
-+FSE_PUBLIC_API size_t FSE_writeNCount(void *buffer, size_t bufferSize, const short *normalizedCounter, unsigned maxSymbolValue, unsigned tableLog);
-+
-+/*! Constructor and Destructor of FSE_CTable.
-+	Note that FSE_CTable size depends on 'tableLog' and 'maxSymbolValue' */
-+typedef unsigned FSE_CTable; /* don't allocate that. It's only meant to be more restrictive than void* */
-+
-+/*! FSE_compress_usingCTable():
-+	Compress `src` using `ct` into `dst` which must be already allocated.
-+	@return : size of compressed data (<= `dstCapacity`),
-+			  or 0 if compressed data could not fit into `dst`,
-+			  or an errorCode, which can be tested using FSE_isError() */
-+FSE_PUBLIC_API size_t FSE_compress_usingCTable(void *dst, size_t dstCapacity, const void *src, size_t srcSize, const FSE_CTable *ct);
-+
-+/*!
-+Tutorial :
-+----------
-+The first step is to count all symbols. FSE_count() does this job very fast.
-+Result will be saved into 'count', a table of unsigned int, which must be already allocated, and have 'maxSymbolValuePtr[0]+1' cells.
-+'src' is a table of bytes of size 'srcSize'. All values within 'src' MUST be <= maxSymbolValuePtr[0]
-+maxSymbolValuePtr[0] will be updated, with its real value (necessarily <= original value)
-+FSE_count() will return the number of occurrence of the most frequent symbol.
-+This can be used to know if there is a single symbol within 'src', and to quickly evaluate its compressibility.
-+If there is an error, the function will return an ErrorCode (which can be tested using FSE_isError()).
-+
-+The next step is to normalize the frequencies.
-+FSE_normalizeCount() will ensure that sum of frequencies is == 2 ^'tableLog'.
-+It also guarantees a minimum of 1 to any Symbol with frequency >= 1.
-+You can use 'tableLog'==0 to mean "use default tableLog value".
-+If you are unsure of which tableLog value to use, you can ask FSE_optimalTableLog(),
-+which will provide the optimal valid tableLog given sourceSize, maxSymbolValue, and a user-defined maximum (0 means "default").
-+
-+The result of FSE_normalizeCount() will be saved into a table,
-+called 'normalizedCounter', which is a table of signed short.
-+'normalizedCounter' must be already allocated, and have at least 'maxSymbolValue+1' cells.
-+The return value is tableLog if everything proceeded as expected.
-+It is 0 if there is a single symbol within distribution.
-+If there is an error (ex: invalid tableLog value), the function will return an ErrorCode (which can be tested using FSE_isError()).
-+
-+'normalizedCounter' can be saved in a compact manner to a memory area using FSE_writeNCount().
-+'buffer' must be already allocated.
-+For guaranteed success, buffer size must be at least FSE_headerBound().
-+The result of the function is the number of bytes written into 'buffer'.
-+If there is an error, the function will return an ErrorCode (which can be tested using FSE_isError(); ex : buffer size too small).
-+
-+'normalizedCounter' can then be used to create the compression table 'CTable'.
-+The space required by 'CTable' must be already allocated, using FSE_createCTable().
-+You can then use FSE_buildCTable() to fill 'CTable'.
-+If there is an error, both functions will return an ErrorCode (which can be tested using FSE_isError()).
-+
-+'CTable' can then be used to compress 'src', with FSE_compress_usingCTable().
-+Similar to FSE_count(), the convention is that 'src' is assumed to be a table of char of size 'srcSize'
-+The function returns the size of compressed data (without header), necessarily <= `dstCapacity`.
-+If it returns '0', compressed data could not fit into 'dst'.
-+If there is an error, the function will return an ErrorCode (which can be tested using FSE_isError()).
-+*/
-+
-+/* *** DECOMPRESSION *** */
-+
-+/*! FSE_readNCount():
-+	Read compactly saved 'normalizedCounter' from 'rBuffer'.
-+	@return : size read from 'rBuffer',
-+			  or an errorCode, which can be tested using FSE_isError().
-+			  maxSymbolValuePtr[0] and tableLogPtr[0] will also be updated with their respective values */
-+FSE_PUBLIC_API size_t FSE_readNCount(short *normalizedCounter, unsigned *maxSymbolValuePtr, unsigned *tableLogPtr, const void *rBuffer, size_t rBuffSize);
-+
-+/*! Constructor and Destructor of FSE_DTable.
-+	Note that its size depends on 'tableLog' */
-+typedef unsigned FSE_DTable; /* don't allocate that. It's just a way to be more restrictive than void* */
-+
-+/*! FSE_buildDTable():
-+	Builds 'dt', which must be already allocated, using FSE_createDTable().
-+	return : 0, or an errorCode, which can be tested using FSE_isError() */
-+FSE_PUBLIC_API size_t FSE_buildDTable_wksp(FSE_DTable *dt, const short *normalizedCounter, unsigned maxSymbolValue, unsigned tableLog, void *workspace, size_t workspaceSize);
-+
-+/*! FSE_decompress_usingDTable():
-+	Decompress compressed source `cSrc` of size `cSrcSize` using `dt`
-+	into `dst` which must be already allocated.
-+	@return : size of regenerated data (necessarily <= `dstCapacity`),
-+			  or an errorCode, which can be tested using FSE_isError() */
-+FSE_PUBLIC_API size_t FSE_decompress_usingDTable(void *dst, size_t dstCapacity, const void *cSrc, size_t cSrcSize, const FSE_DTable *dt);
-+
-+/*!
-+Tutorial :
-+----------
-+(Note : these functions only decompress FSE-compressed blocks.
-+ If block is uncompressed, use memcpy() instead
-+ If block is a single repeated byte, use memset() instead )
-+
-+The first step is to obtain the normalized frequencies of symbols.
-+This can be performed by FSE_readNCount() if it was saved using FSE_writeNCount().
-+'normalizedCounter' must be already allocated, and have at least 'maxSymbolValuePtr[0]+1' cells of signed short.
-+In practice, that means it's necessary to know 'maxSymbolValue' beforehand,
-+or size the table to handle worst case situations (typically 256).
-+FSE_readNCount() will provide 'tableLog' and 'maxSymbolValue'.
-+The result of FSE_readNCount() is the number of bytes read from 'rBuffer'.
-+Note that 'rBufferSize' must be at least 4 bytes, even if useful information is less than that.
-+If there is an error, the function will return an error code, which can be tested using FSE_isError().
-+
-+The next step is to build the decompression tables 'FSE_DTable' from 'normalizedCounter'.
-+This is performed by the function FSE_buildDTable().
-+The space required by 'FSE_DTable' must be already allocated using FSE_createDTable().
-+If there is an error, the function will return an error code, which can be tested using FSE_isError().
-+
-+`FSE_DTable` can then be used to decompress `cSrc`, with FSE_decompress_usingDTable().
-+`cSrcSize` must be strictly correct, otherwise decompression will fail.
-+FSE_decompress_usingDTable() result will tell how many bytes were regenerated (<=`dstCapacity`).
-+If there is an error, the function will return an error code, which can be tested using FSE_isError(). (ex: dst buffer too small)
-+*/
-+
-+/* *** Dependency *** */
-+#include "bitstream.h"
-+
-+/* *****************************************
-+*  Static allocation
-+*******************************************/
-+/* FSE buffer bounds */
-+#define FSE_NCOUNTBOUND 512
-+#define FSE_BLOCKBOUND(size) (size + (size >> 7))
-+#define FSE_COMPRESSBOUND(size) (FSE_NCOUNTBOUND + FSE_BLOCKBOUND(size)) /* Macro version, useful for static allocation */
-+
-+/* It is possible to statically allocate FSE CTable/DTable as a table of FSE_CTable/FSE_DTable using below macros */
-+#define FSE_CTABLE_SIZE_U32(maxTableLog, maxSymbolValue) (1 + (1 << (maxTableLog - 1)) + ((maxSymbolValue + 1) * 2))
-+#define FSE_DTABLE_SIZE_U32(maxTableLog) (1 + (1 << maxTableLog))
-+
-+/* *****************************************
-+*  FSE advanced API
-+*******************************************/
-+/* FSE_count_wksp() :
-+ * Same as FSE_count(), but using an externally provided scratch buffer.
-+ * `workSpace` size must be table of >= `1024` unsigned
-+ */
-+size_t FSE_count_wksp(unsigned *count, unsigned *maxSymbolValuePtr, const void *source, size_t sourceSize, unsigned *workSpace);
-+
-+/* FSE_countFast_wksp() :
-+ * Same as FSE_countFast(), but using an externally provided scratch buffer.
-+ * `workSpace` must be a table of minimum `1024` unsigned
-+ */
-+size_t FSE_countFast_wksp(unsigned *count, unsigned *maxSymbolValuePtr, const void *src, size_t srcSize, unsigned *workSpace);
-+
-+/*! FSE_count_simple
-+ * Same as FSE_countFast(), but does not use any additional memory (not even on stack).
-+ * This function is unsafe, and will segfault if any value within `src` is `> *maxSymbolValuePtr` (presuming it's also the size of `count`).
-+*/
-+size_t FSE_count_simple(unsigned *count, unsigned *maxSymbolValuePtr, const void *src, size_t srcSize);
-+
-+unsigned FSE_optimalTableLog_internal(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue, unsigned minus);
-+/**< same as FSE_optimalTableLog(), which used `minus==2` */
-+
-+size_t FSE_buildCTable_raw(FSE_CTable *ct, unsigned nbBits);
-+/**< build a fake FSE_CTable, designed for a flat distribution, where each symbol uses nbBits */
-+
-+size_t FSE_buildCTable_rle(FSE_CTable *ct, unsigned char symbolValue);
-+/**< build a fake FSE_CTable, designed to compress always the same symbolValue */
-+
-+/* FSE_buildCTable_wksp() :
-+ * Same as FSE_buildCTable(), but using an externally allocated scratch buffer (`workSpace`).
-+ * `wkspSize` must be >= `(1<<tableLog)`.
-+ */
-+size_t FSE_buildCTable_wksp(FSE_CTable *ct, const short *normalizedCounter, unsigned maxSymbolValue, unsigned tableLog, void *workSpace, size_t wkspSize);
-+
-+size_t FSE_buildDTable_raw(FSE_DTable *dt, unsigned nbBits);
-+/**< build a fake FSE_DTable, designed to read a flat distribution where each symbol uses nbBits */
-+
-+size_t FSE_buildDTable_rle(FSE_DTable *dt, unsigned char symbolValue);
-+/**< build a fake FSE_DTable, designed to always generate the same symbolValue */
-+
-+size_t FSE_decompress_wksp(void *dst, size_t dstCapacity, const void *cSrc, size_t cSrcSize, unsigned maxLog, void *workspace, size_t workspaceSize);
-+/**< same as FSE_decompress(), using an externally allocated `workSpace` produced with `FSE_DTABLE_SIZE_U32(maxLog)` */
-+
-+/* *****************************************
-+*  FSE symbol compression API
-+*******************************************/
-+/*!
-+   This API consists of small unitary functions, which highly benefit from being inlined.
-+   Hence their body are included in next section.
-+*/
-+typedef struct {
-+	ptrdiff_t value;
-+	const void *stateTable;
-+	const void *symbolTT;
-+	unsigned stateLog;
-+} FSE_CState_t;
-+
-+static void FSE_initCState(FSE_CState_t *CStatePtr, const FSE_CTable *ct);
-+
-+static void FSE_encodeSymbol(BIT_CStream_t *bitC, FSE_CState_t *CStatePtr, unsigned symbol);
-+
-+static void FSE_flushCState(BIT_CStream_t *bitC, const FSE_CState_t *CStatePtr);
-+
-+/**<
-+These functions are inner components of FSE_compress_usingCTable().
-+They allow the creation of custom streams, mixing multiple tables and bit sources.
-+
-+A key property to keep in mind is that encoding and decoding are done **in reverse direction**.
-+So the first symbol you will encode is the last you will decode, like a LIFO stack.
-+
-+You will need a few variables to track your CStream. They are :
-+
-+FSE_CTable    ct;         // Provided by FSE_buildCTable()
-+BIT_CStream_t bitStream;  // bitStream tracking structure
-+FSE_CState_t  state;      // State tracking structure (can have several)
-+
-+
-+The first thing to do is to init bitStream and state.
-+	size_t errorCode = BIT_initCStream(&bitStream, dstBuffer, maxDstSize);
-+	FSE_initCState(&state, ct);
-+
-+Note that BIT_initCStream() can produce an error code, so its result should be tested, using FSE_isError();
-+You can then encode your input data, byte after byte.
-+FSE_encodeSymbol() outputs a maximum of 'tableLog' bits at a time.
-+Remember decoding will be done in reverse direction.
-+	FSE_encodeByte(&bitStream, &state, symbol);
-+
-+At any time, you can also add any bit sequence.
-+Note : maximum allowed nbBits is 25, for compatibility with 32-bits decoders
-+	BIT_addBits(&bitStream, bitField, nbBits);
-+
-+The above methods don't commit data to memory, they just store it into local register, for speed.
-+Local register size is 64-bits on 64-bits systems, 32-bits on 32-bits systems (size_t).
-+Writing data to memory is a manual operation, performed by the flushBits function.
-+	BIT_flushBits(&bitStream);
-+
-+Your last FSE encoding operation shall be to flush your last state value(s).
-+	FSE_flushState(&bitStream, &state);
-+
-+Finally, you must close the bitStream.
-+The function returns the size of CStream in bytes.
-+If data couldn't fit into dstBuffer, it will return a 0 ( == not compressible)
-+If there is an error, it returns an errorCode (which can be tested using FSE_isError()).
-+	size_t size = BIT_closeCStream(&bitStream);
-+*/
-+
-+/* *****************************************
-+*  FSE symbol decompression API
-+*******************************************/
-+typedef struct {
-+	size_t state;
-+	const void *table; /* precise table may vary, depending on U16 */
-+} FSE_DState_t;
-+
-+static void FSE_initDState(FSE_DState_t *DStatePtr, BIT_DStream_t *bitD, const FSE_DTable *dt);
-+
-+static unsigned char FSE_decodeSymbol(FSE_DState_t *DStatePtr, BIT_DStream_t *bitD);
-+
-+static unsigned FSE_endOfDState(const FSE_DState_t *DStatePtr);
-+
-+/**<
-+Let's now decompose FSE_decompress_usingDTable() into its unitary components.
-+You will decode FSE-encoded symbols from the bitStream,
-+and also any other bitFields you put in, **in reverse order**.
-+
-+You will need a few variables to track your bitStream. They are :
-+
-+BIT_DStream_t DStream;    // Stream context
-+FSE_DState_t  DState;     // State context. Multiple ones are possible
-+FSE_DTable*   DTablePtr;  // Decoding table, provided by FSE_buildDTable()
-+
-+The first thing to do is to init the bitStream.
-+	errorCode = BIT_initDStream(&DStream, srcBuffer, srcSize);
-+
-+You should then retrieve your initial state(s)
-+(in reverse flushing order if you have several ones) :
-+	errorCode = FSE_initDState(&DState, &DStream, DTablePtr);
-+
-+You can then decode your data, symbol after symbol.
-+For information the maximum number of bits read by FSE_decodeSymbol() is 'tableLog'.
-+Keep in mind that symbols are decoded in reverse order, like a LIFO stack (last in, first out).
-+	unsigned char symbol = FSE_decodeSymbol(&DState, &DStream);
-+
-+You can retrieve any bitfield you eventually stored into the bitStream (in reverse order)
-+Note : maximum allowed nbBits is 25, for 32-bits compatibility
-+	size_t bitField = BIT_readBits(&DStream, nbBits);
-+
-+All above operations only read from local register (which size depends on size_t).
-+Refueling the register from memory is manually performed by the reload method.
-+	endSignal = FSE_reloadDStream(&DStream);
-+
-+BIT_reloadDStream() result tells if there is still some more data to read from DStream.
-+BIT_DStream_unfinished : there is still some data left into the DStream.
-+BIT_DStream_endOfBuffer : Dstream reached end of buffer. Its container may no longer be completely filled.
-+BIT_DStream_completed : Dstream reached its exact end, corresponding in general to decompression completed.
-+BIT_DStream_tooFar : Dstream went too far. Decompression result is corrupted.
-+
-+When reaching end of buffer (BIT_DStream_endOfBuffer), progress slowly, notably if you decode multiple symbols per loop,
-+to properly detect the exact end of stream.
-+After each decoded symbol, check if DStream is fully consumed using this simple test :
-+	BIT_reloadDStream(&DStream) >= BIT_DStream_completed
-+
-+When it's done, verify decompression is fully completed, by checking both DStream and the relevant states.
-+Checking if DStream has reached its end is performed by :
-+	BIT_endOfDStream(&DStream);
-+Check also the states. There might be some symbols left there, if some high probability ones (>50%) are possible.
-+	FSE_endOfDState(&DState);
-+*/
-+
-+/* *****************************************
-+*  FSE unsafe API
-+*******************************************/
-+static unsigned char FSE_decodeSymbolFast(FSE_DState_t *DStatePtr, BIT_DStream_t *bitD);
-+/* faster, but works only if nbBits is always >= 1 (otherwise, result will be corrupted) */
-+
-+/* *****************************************
-+*  Implementation of inlined functions
-+*******************************************/
-+typedef struct {
-+	int deltaFindState;
-+	U32 deltaNbBits;
-+} FSE_symbolCompressionTransform; /* total 8 bytes */
-+
-+ZSTD_STATIC void FSE_initCState(FSE_CState_t *statePtr, const FSE_CTable *ct)
-+{
-+	const void *ptr = ct;
-+	const U16 *u16ptr = (const U16 *)ptr;
-+	const U32 tableLog = ZSTD_read16(ptr);
-+	statePtr->value = (ptrdiff_t)1 << tableLog;
-+	statePtr->stateTable = u16ptr + 2;
-+	statePtr->symbolTT = ((const U32 *)ct + 1 + (tableLog ? (1 << (tableLog - 1)) : 1));
-+	statePtr->stateLog = tableLog;
-+}
-+
-+/*! FSE_initCState2() :
-+*   Same as FSE_initCState(), but the first symbol to include (which will be the last to be read)
-+*   uses the smallest state value possible, saving the cost of this symbol */
-+ZSTD_STATIC void FSE_initCState2(FSE_CState_t *statePtr, const FSE_CTable *ct, U32 symbol)
-+{
-+	FSE_initCState(statePtr, ct);
-+	{
-+		const FSE_symbolCompressionTransform symbolTT = ((const FSE_symbolCompressionTransform *)(statePtr->symbolTT))[symbol];
-+		const U16 *stateTable = (const U16 *)(statePtr->stateTable);
-+		U32 nbBitsOut = (U32)((symbolTT.deltaNbBits + (1 << 15)) >> 16);
-+		statePtr->value = (nbBitsOut << 16) - symbolTT.deltaNbBits;
-+		statePtr->value = stateTable[(statePtr->value >> nbBitsOut) + symbolTT.deltaFindState];
-+	}
-+}
-+
-+ZSTD_STATIC void FSE_encodeSymbol(BIT_CStream_t *bitC, FSE_CState_t *statePtr, U32 symbol)
-+{
-+	const FSE_symbolCompressionTransform symbolTT = ((const FSE_symbolCompressionTransform *)(statePtr->symbolTT))[symbol];
-+	const U16 *const stateTable = (const U16 *)(statePtr->stateTable);
-+	U32 nbBitsOut = (U32)((statePtr->value + symbolTT.deltaNbBits) >> 16);
-+	BIT_addBits(bitC, statePtr->value, nbBitsOut);
-+	statePtr->value = stateTable[(statePtr->value >> nbBitsOut) + symbolTT.deltaFindState];
-+}
-+
-+ZSTD_STATIC void FSE_flushCState(BIT_CStream_t *bitC, const FSE_CState_t *statePtr)
-+{
-+	BIT_addBits(bitC, statePtr->value, statePtr->stateLog);
-+	BIT_flushBits(bitC);
-+}
-+
-+/* ======    Decompression    ====== */
-+
-+typedef struct {
-+	U16 tableLog;
-+	U16 fastMode;
-+} FSE_DTableHeader; /* sizeof U32 */
-+
-+typedef struct {
-+	unsigned short newState;
-+	unsigned char symbol;
-+	unsigned char nbBits;
-+} FSE_decode_t; /* size == U32 */
-+
-+ZSTD_STATIC void FSE_initDState(FSE_DState_t *DStatePtr, BIT_DStream_t *bitD, const FSE_DTable *dt)
-+{
-+	const void *ptr = dt;
-+	const FSE_DTableHeader *const DTableH = (const FSE_DTableHeader *)ptr;
-+	DStatePtr->state = BIT_readBits(bitD, DTableH->tableLog);
-+	BIT_reloadDStream(bitD);
-+	DStatePtr->table = dt + 1;
-+}
-+
-+ZSTD_STATIC BYTE FSE_peekSymbol(const FSE_DState_t *DStatePtr)
-+{
-+	FSE_decode_t const DInfo = ((const FSE_decode_t *)(DStatePtr->table))[DStatePtr->state];
-+	return DInfo.symbol;
-+}
-+
-+ZSTD_STATIC void FSE_updateState(FSE_DState_t *DStatePtr, BIT_DStream_t *bitD)
-+{
-+	FSE_decode_t const DInfo = ((const FSE_decode_t *)(DStatePtr->table))[DStatePtr->state];
-+	U32 const nbBits = DInfo.nbBits;
-+	size_t const lowBits = BIT_readBits(bitD, nbBits);
-+	DStatePtr->state = DInfo.newState + lowBits;
-+}
-+
-+ZSTD_STATIC BYTE FSE_decodeSymbol(FSE_DState_t *DStatePtr, BIT_DStream_t *bitD)
-+{
-+	FSE_decode_t const DInfo = ((const FSE_decode_t *)(DStatePtr->table))[DStatePtr->state];
-+	U32 const nbBits = DInfo.nbBits;
-+	BYTE const symbol = DInfo.symbol;
-+	size_t const lowBits = BIT_readBits(bitD, nbBits);
-+
-+	DStatePtr->state = DInfo.newState + lowBits;
-+	return symbol;
-+}
-+
-+/*! FSE_decodeSymbolFast() :
-+	unsafe, only works if no symbol has a probability > 50% */
-+ZSTD_STATIC BYTE FSE_decodeSymbolFast(FSE_DState_t *DStatePtr, BIT_DStream_t *bitD)
-+{
-+	FSE_decode_t const DInfo = ((const FSE_decode_t *)(DStatePtr->table))[DStatePtr->state];
-+	U32 const nbBits = DInfo.nbBits;
-+	BYTE const symbol = DInfo.symbol;
-+	size_t const lowBits = BIT_readBitsFast(bitD, nbBits);
-+
-+	DStatePtr->state = DInfo.newState + lowBits;
-+	return symbol;
-+}
-+
-+ZSTD_STATIC unsigned FSE_endOfDState(const FSE_DState_t *DStatePtr) { return DStatePtr->state == 0; }
-+
-+/* **************************************************************
-+*  Tuning parameters
-+****************************************************************/
-+/*!MEMORY_USAGE :
-+*  Memory usage formula : N->2^N Bytes (examples : 10 -> 1KB; 12 -> 4KB ; 16 -> 64KB; 20 -> 1MB; etc.)
-+*  Increasing memory usage improves compression ratio
-+*  Reduced memory usage can improve speed, due to cache effect
-+*  Recommended max value is 14, for 16KB, which nicely fits into Intel x86 L1 cache */
-+#ifndef FSE_MAX_MEMORY_USAGE
-+#define FSE_MAX_MEMORY_USAGE 14
-+#endif
-+#ifndef FSE_DEFAULT_MEMORY_USAGE
-+#define FSE_DEFAULT_MEMORY_USAGE 13
-+#endif
-+
-+/*!FSE_MAX_SYMBOL_VALUE :
-+*  Maximum symbol value authorized.
-+*  Required for proper stack allocation */
-+#ifndef FSE_MAX_SYMBOL_VALUE
-+#define FSE_MAX_SYMBOL_VALUE 255
-+#endif
-+
-+/* **************************************************************
-+*  template functions type & suffix
-+****************************************************************/
-+#define FSE_FUNCTION_TYPE BYTE
-+#define FSE_FUNCTION_EXTENSION
-+#define FSE_DECODE_TYPE FSE_decode_t
-+
-+/* ***************************************************************
-+*  Constants
-+*****************************************************************/
-+#define FSE_MAX_TABLELOG (FSE_MAX_MEMORY_USAGE - 2)
-+#define FSE_MAX_TABLESIZE (1U << FSE_MAX_TABLELOG)
-+#define FSE_MAXTABLESIZE_MASK (FSE_MAX_TABLESIZE - 1)
-+#define FSE_DEFAULT_TABLELOG (FSE_DEFAULT_MEMORY_USAGE - 2)
-+#define FSE_MIN_TABLELOG 5
-+
-+#define FSE_TABLELOG_ABSOLUTE_MAX 15
-+#if FSE_MAX_TABLELOG > FSE_TABLELOG_ABSOLUTE_MAX
-+#error "FSE_MAX_TABLELOG > FSE_TABLELOG_ABSOLUTE_MAX is not supported"
-+#endif
-+
-+#define FSE_TABLESTEP(tableSize) ((tableSize >> 1) + (tableSize >> 3) + 3)
-+
-+#endif /* FSE_H */
-diff --git a/lib/zstd/fse_compress.c b/lib/zstd/fse_compress.c
-new file mode 100644
-index 0000000..ef3d174
---- /dev/null
-+++ b/lib/zstd/fse_compress.c
-@@ -0,0 +1,795 @@
-+/*
-+ * FSE : Finite State Entropy encoder
-+ * Copyright (C) 2013-2015, Yann Collet.
-+ *
-+ * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
-+ *
-+ * Redistribution and use in source and binary forms, with or without
-+ * modification, are permitted provided that the following conditions are
-+ * met:
-+ *
-+ *   * Redistributions of source code must retain the above copyright
-+ * notice, this list of conditions and the following disclaimer.
-+ *   * Redistributions in binary form must reproduce the above
-+ * copyright notice, this list of conditions and the following disclaimer
-+ * in the documentation and/or other materials provided with the
-+ * distribution.
-+ *
-+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
-+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
-+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
-+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
-+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
-+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
-+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
-+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
-+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
-+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
-+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-+ *
-+ * This program is free software; you can redistribute it and/or modify it under
-+ * the terms of the GNU General Public License version 2 as published by the
-+ * Free Software Foundation. This program is dual-licensed; you may select
-+ * either version 2 of the GNU General Public License ("GPL") or BSD license
-+ * ("BSD").
-+ *
-+ * You can contact the author at :
-+ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
-+ */
-+
-+/* **************************************************************
-+*  Compiler specifics
-+****************************************************************/
-+#define FORCE_INLINE static __always_inline
-+
-+/* **************************************************************
-+*  Includes
-+****************************************************************/
-+#include "bitstream.h"
-+#include "fse.h"
-+#include <linux/compiler.h>
-+#include <linux/kernel.h>
-+#include <linux/math64.h>
-+#include <linux/string.h> /* memcpy, memset */
-+
-+/* **************************************************************
-+*  Error Management
-+****************************************************************/
-+#define FSE_STATIC_ASSERT(c)                                   \
-+	{                                                      \
-+		enum { FSE_static_assert = 1 / (int)(!!(c)) }; \
-+	} /* use only *after* variable declarations */
-+
-+/* **************************************************************
-+*  Templates
-+****************************************************************/
-+/*
-+  designed to be included
-+  for type-specific functions (template emulation in C)
-+  Objective is to write these functions only once, for improved maintenance
-+*/
-+
-+/* safety checks */
-+#ifndef FSE_FUNCTION_EXTENSION
-+#error "FSE_FUNCTION_EXTENSION must be defined"
-+#endif
-+#ifndef FSE_FUNCTION_TYPE
-+#error "FSE_FUNCTION_TYPE must be defined"
-+#endif
-+
-+/* Function names */
-+#define FSE_CAT(X, Y) X##Y
-+#define FSE_FUNCTION_NAME(X, Y) FSE_CAT(X, Y)
-+#define FSE_TYPE_NAME(X, Y) FSE_CAT(X, Y)
-+
-+/* Function templates */
-+
-+/* FSE_buildCTable_wksp() :
-+ * Same as FSE_buildCTable(), but using an externally allocated scratch buffer (`workSpace`).
-+ * wkspSize should be sized to handle worst case situation, which is `1<<max_tableLog * sizeof(FSE_FUNCTION_TYPE)`
-+ * workSpace must also be properly aligned with FSE_FUNCTION_TYPE requirements
-+ */
-+size_t FSE_buildCTable_wksp(FSE_CTable *ct, const short *normalizedCounter, unsigned maxSymbolValue, unsigned tableLog, void *workspace, size_t workspaceSize)
-+{
-+	U32 const tableSize = 1 << tableLog;
-+	U32 const tableMask = tableSize - 1;
-+	void *const ptr = ct;
-+	U16 *const tableU16 = ((U16 *)ptr) + 2;
-+	void *const FSCT = ((U32 *)ptr) + 1 /* header */ + (tableLog ? tableSize >> 1 : 1);
-+	FSE_symbolCompressionTransform *const symbolTT = (FSE_symbolCompressionTransform *)(FSCT);
-+	U32 const step = FSE_TABLESTEP(tableSize);
-+	U32 highThreshold = tableSize - 1;
-+
-+	U32 *cumul;
-+	FSE_FUNCTION_TYPE *tableSymbol;
-+	size_t spaceUsed32 = 0;
-+
-+	cumul = (U32 *)workspace + spaceUsed32;
-+	spaceUsed32 += FSE_MAX_SYMBOL_VALUE + 2;
-+	tableSymbol = (FSE_FUNCTION_TYPE *)((U32 *)workspace + spaceUsed32);
-+	spaceUsed32 += ALIGN(sizeof(FSE_FUNCTION_TYPE) * ((size_t)1 << tableLog), sizeof(U32)) >> 2;
-+
-+	if ((spaceUsed32 << 2) > workspaceSize)
-+		return ERROR(tableLog_tooLarge);
-+	workspace = (U32 *)workspace + spaceUsed32;
-+	workspaceSize -= (spaceUsed32 << 2);
-+
-+	/* CTable header */
-+	tableU16[-2] = (U16)tableLog;
-+	tableU16[-1] = (U16)maxSymbolValue;
-+
-+	/* For explanations on how to distribute symbol values over the table :
-+	*  http://fastcompression.blogspot.fr/2014/02/fse-distributing-symbol-values.html */
-+
-+	/* symbol start positions */
-+	{
-+		U32 u;
-+		cumul[0] = 0;
-+		for (u = 1; u <= maxSymbolValue + 1; u++) {
-+			if (normalizedCounter[u - 1] == -1) { /* Low proba symbol */
-+				cumul[u] = cumul[u - 1] + 1;
-+				tableSymbol[highThreshold--] = (FSE_FUNCTION_TYPE)(u - 1);
-+			} else {
-+				cumul[u] = cumul[u - 1] + normalizedCounter[u - 1];
-+			}
-+		}
-+		cumul[maxSymbolValue + 1] = tableSize + 1;
-+	}
-+
-+	/* Spread symbols */
-+	{
-+		U32 position = 0;
-+		U32 symbol;
-+		for (symbol = 0; symbol <= maxSymbolValue; symbol++) {
-+			int nbOccurences;
-+			for (nbOccurences = 0; nbOccurences < normalizedCounter[symbol]; nbOccurences++) {
-+				tableSymbol[position] = (FSE_FUNCTION_TYPE)symbol;
-+				position = (position + step) & tableMask;
-+				while (position > highThreshold)
-+					position = (position + step) & tableMask; /* Low proba area */
-+			}
-+		}
-+
-+		if (position != 0)
-+			return ERROR(GENERIC); /* Must have gone through all positions */
-+	}
-+
-+	/* Build table */
-+	{
-+		U32 u;
-+		for (u = 0; u < tableSize; u++) {
-+			FSE_FUNCTION_TYPE s = tableSymbol[u];	/* note : static analyzer may not understand tableSymbol is properly initialized */
-+			tableU16[cumul[s]++] = (U16)(tableSize + u); /* TableU16 : sorted by symbol order; gives next state value */
-+		}
-+	}
-+
-+	/* Build Symbol Transformation Table */
-+	{
-+		unsigned total = 0;
-+		unsigned s;
-+		for (s = 0; s <= maxSymbolValue; s++) {
-+			switch (normalizedCounter[s]) {
-+			case 0: break;
-+
-+			case -1:
-+			case 1:
-+				symbolTT[s].deltaNbBits = (tableLog << 16) - (1 << tableLog);
-+				symbolTT[s].deltaFindState = total - 1;
-+				total++;
-+				break;
-+			default: {
-+				U32 const maxBitsOut = tableLog - BIT_highbit32(normalizedCounter[s] - 1);
-+				U32 const minStatePlus = normalizedCounter[s] << maxBitsOut;
-+				symbolTT[s].deltaNbBits = (maxBitsOut << 16) - minStatePlus;
-+				symbolTT[s].deltaFindState = total - normalizedCounter[s];
-+				total += normalizedCounter[s];
-+			}
-+			}
-+		}
-+	}
-+
-+	return 0;
-+}
-+
-+/*-**************************************************************
-+*  FSE NCount encoding-decoding
-+****************************************************************/
-+size_t FSE_NCountWriteBound(unsigned maxSymbolValue, unsigned tableLog)
-+{
-+	size_t const maxHeaderSize = (((maxSymbolValue + 1) * tableLog) >> 3) + 3;
-+	return maxSymbolValue ? maxHeaderSize : FSE_NCOUNTBOUND; /* maxSymbolValue==0 ? use default */
-+}
-+
-+static size_t FSE_writeNCount_generic(void *header, size_t headerBufferSize, const short *normalizedCounter, unsigned maxSymbolValue, unsigned tableLog,
-+				      unsigned writeIsSafe)
-+{
-+	BYTE *const ostart = (BYTE *)header;
-+	BYTE *out = ostart;
-+	BYTE *const oend = ostart + headerBufferSize;
-+	int nbBits;
-+	const int tableSize = 1 << tableLog;
-+	int remaining;
-+	int threshold;
-+	U32 bitStream;
-+	int bitCount;
-+	unsigned charnum = 0;
-+	int previous0 = 0;
-+
-+	bitStream = 0;
-+	bitCount = 0;
-+	/* Table Size */
-+	bitStream += (tableLog - FSE_MIN_TABLELOG) << bitCount;
-+	bitCount += 4;
-+
-+	/* Init */
-+	remaining = tableSize + 1; /* +1 for extra accuracy */
-+	threshold = tableSize;
-+	nbBits = tableLog + 1;
-+
-+	while (remaining > 1) { /* stops at 1 */
-+		if (previous0) {
-+			unsigned start = charnum;
-+			while (!normalizedCounter[charnum])
-+				charnum++;
-+			while (charnum >= start + 24) {
-+				start += 24;
-+				bitStream += 0xFFFFU << bitCount;
-+				if ((!writeIsSafe) && (out > oend - 2))
-+					return ERROR(dstSize_tooSmall); /* Buffer overflow */
-+				out[0] = (BYTE)bitStream;
-+				out[1] = (BYTE)(bitStream >> 8);
-+				out += 2;
-+				bitStream >>= 16;
-+			}
-+			while (charnum >= start + 3) {
-+				start += 3;
-+				bitStream += 3 << bitCount;
-+				bitCount += 2;
-+			}
-+			bitStream += (charnum - start) << bitCount;
-+			bitCount += 2;
-+			if (bitCount > 16) {
-+				if ((!writeIsSafe) && (out > oend - 2))
-+					return ERROR(dstSize_tooSmall); /* Buffer overflow */
-+				out[0] = (BYTE)bitStream;
-+				out[1] = (BYTE)(bitStream >> 8);
-+				out += 2;
-+				bitStream >>= 16;
-+				bitCount -= 16;
-+			}
-+		}
-+		{
-+			int count = normalizedCounter[charnum++];
-+			int const max = (2 * threshold - 1) - remaining;
-+			remaining -= count < 0 ? -count : count;
-+			count++; /* +1 for extra accuracy */
-+			if (count >= threshold)
-+				count += max; /* [0..max[ [max..threshold[ (...) [threshold+max 2*threshold[ */
-+			bitStream += count << bitCount;
-+			bitCount += nbBits;
-+			bitCount -= (count < max);
-+			previous0 = (count == 1);
-+			if (remaining < 1)
-+				return ERROR(GENERIC);
-+			while (remaining < threshold)
-+				nbBits--, threshold >>= 1;
-+		}
-+		if (bitCount > 16) {
-+			if ((!writeIsSafe) && (out > oend - 2))
-+				return ERROR(dstSize_tooSmall); /* Buffer overflow */
-+			out[0] = (BYTE)bitStream;
-+			out[1] = (BYTE)(bitStream >> 8);
-+			out += 2;
-+			bitStream >>= 16;
-+			bitCount -= 16;
-+		}
-+	}
-+
-+	/* flush remaining bitStream */
-+	if ((!writeIsSafe) && (out > oend - 2))
-+		return ERROR(dstSize_tooSmall); /* Buffer overflow */
-+	out[0] = (BYTE)bitStream;
-+	out[1] = (BYTE)(bitStream >> 8);
-+	out += (bitCount + 7) / 8;
-+
-+	if (charnum > maxSymbolValue + 1)
-+		return ERROR(GENERIC);
-+
-+	return (out - ostart);
-+}
-+
-+size_t FSE_writeNCount(void *buffer, size_t bufferSize, const short *normalizedCounter, unsigned maxSymbolValue, unsigned tableLog)
-+{
-+	if (tableLog > FSE_MAX_TABLELOG)
-+		return ERROR(tableLog_tooLarge); /* Unsupported */
-+	if (tableLog < FSE_MIN_TABLELOG)
-+		return ERROR(GENERIC); /* Unsupported */
-+
-+	if (bufferSize < FSE_NCountWriteBound(maxSymbolValue, tableLog))
-+		return FSE_writeNCount_generic(buffer, bufferSize, normalizedCounter, maxSymbolValue, tableLog, 0);
-+
-+	return FSE_writeNCount_generic(buffer, bufferSize, normalizedCounter, maxSymbolValue, tableLog, 1);
-+}
-+
-+/*-**************************************************************
-+*  Counting histogram
-+****************************************************************/
-+/*! FSE_count_simple
-+	This function counts byte values within `src`, and store the histogram into table `count`.
-+	It doesn't use any additional memory.
-+	But this function is unsafe : it doesn't check that all values within `src` can fit into `count`.
-+	For this reason, prefer using a table `count` with 256 elements.
-+	@return : count of most numerous element
-+*/
-+size_t FSE_count_simple(unsigned *count, unsigned *maxSymbolValuePtr, const void *src, size_t srcSize)
-+{
-+	const BYTE *ip = (const BYTE *)src;
-+	const BYTE *const end = ip + srcSize;
-+	unsigned maxSymbolValue = *maxSymbolValuePtr;
-+	unsigned max = 0;
-+
-+	memset(count, 0, (maxSymbolValue + 1) * sizeof(*count));
-+	if (srcSize == 0) {
-+		*maxSymbolValuePtr = 0;
-+		return 0;
-+	}
-+
-+	while (ip < end)
-+		count[*ip++]++;
-+
-+	while (!count[maxSymbolValue])
-+		maxSymbolValue--;
-+	*maxSymbolValuePtr = maxSymbolValue;
-+
-+	{
-+		U32 s;
-+		for (s = 0; s <= maxSymbolValue; s++)
-+			if (count[s] > max)
-+				max = count[s];
-+	}
-+
-+	return (size_t)max;
-+}
-+
-+/* FSE_count_parallel_wksp() :
-+ * Same as FSE_count_parallel(), but using an externally provided scratch buffer.
-+ * `workSpace` size must be a minimum of `1024 * sizeof(unsigned)`` */
-+static size_t FSE_count_parallel_wksp(unsigned *count, unsigned *maxSymbolValuePtr, const void *source, size_t sourceSize, unsigned checkMax,
-+				      unsigned *const workSpace)
-+{
-+	const BYTE *ip = (const BYTE *)source;
-+	const BYTE *const iend = ip + sourceSize;
-+	unsigned maxSymbolValue = *maxSymbolValuePtr;
-+	unsigned max = 0;
-+	U32 *const Counting1 = workSpace;
-+	U32 *const Counting2 = Counting1 + 256;
-+	U32 *const Counting3 = Counting2 + 256;
-+	U32 *const Counting4 = Counting3 + 256;
-+
-+	memset(Counting1, 0, 4 * 256 * sizeof(unsigned));
-+
-+	/* safety checks */
-+	if (!sourceSize) {
-+		memset(count, 0, maxSymbolValue + 1);
-+		*maxSymbolValuePtr = 0;
-+		return 0;
-+	}
-+	if (!maxSymbolValue)
-+		maxSymbolValue = 255; /* 0 == default */
-+
-+	/* by stripes of 16 bytes */
-+	{
-+		U32 cached = ZSTD_read32(ip);
-+		ip += 4;
-+		while (ip < iend - 15) {
-+			U32 c = cached;
-+			cached = ZSTD_read32(ip);
-+			ip += 4;
-+			Counting1[(BYTE)c]++;
-+			Counting2[(BYTE)(c >> 8)]++;
-+			Counting3[(BYTE)(c >> 16)]++;
-+			Counting4[c >> 24]++;
-+			c = cached;
-+			cached = ZSTD_read32(ip);
-+			ip += 4;
-+			Counting1[(BYTE)c]++;
-+			Counting2[(BYTE)(c >> 8)]++;
-+			Counting3[(BYTE)(c >> 16)]++;
-+			Counting4[c >> 24]++;
-+			c = cached;
-+			cached = ZSTD_read32(ip);
-+			ip += 4;
-+			Counting1[(BYTE)c]++;
-+			Counting2[(BYTE)(c >> 8)]++;
-+			Counting3[(BYTE)(c >> 16)]++;
-+			Counting4[c >> 24]++;
-+			c = cached;
-+			cached = ZSTD_read32(ip);
-+			ip += 4;
-+			Counting1[(BYTE)c]++;
-+			Counting2[(BYTE)(c >> 8)]++;
-+			Counting3[(BYTE)(c >> 16)]++;
-+			Counting4[c >> 24]++;
-+		}
-+		ip -= 4;
-+	}
-+
-+	/* finish last symbols */
-+	while (ip < iend)
-+		Counting1[*ip++]++;
-+
-+	if (checkMax) { /* verify stats will fit into destination table */
-+		U32 s;
-+		for (s = 255; s > maxSymbolValue; s--) {
-+			Counting1[s] += Counting2[s] + Counting3[s] + Counting4[s];
-+			if (Counting1[s])
-+				return ERROR(maxSymbolValue_tooSmall);
-+		}
-+	}
-+
-+	{
-+		U32 s;
-+		for (s = 0; s <= maxSymbolValue; s++) {
-+			count[s] = Counting1[s] + Counting2[s] + Counting3[s] + Counting4[s];
-+			if (count[s] > max)
-+				max = count[s];
-+		}
-+	}
-+
-+	while (!count[maxSymbolValue])
-+		maxSymbolValue--;
-+	*maxSymbolValuePtr = maxSymbolValue;
-+	return (size_t)max;
-+}
-+
-+/* FSE_countFast_wksp() :
-+ * Same as FSE_countFast(), but using an externally provided scratch buffer.
-+ * `workSpace` size must be table of >= `1024` unsigned */
-+size_t FSE_countFast_wksp(unsigned *count, unsigned *maxSymbolValuePtr, const void *source, size_t sourceSize, unsigned *workSpace)
-+{
-+	if (sourceSize < 1500)
-+		return FSE_count_simple(count, maxSymbolValuePtr, source, sourceSize);
-+	return FSE_count_parallel_wksp(count, maxSymbolValuePtr, source, sourceSize, 0, workSpace);
-+}
-+
-+/* FSE_count_wksp() :
-+ * Same as FSE_count(), but using an externally provided scratch buffer.
-+ * `workSpace` size must be table of >= `1024` unsigned */
-+size_t FSE_count_wksp(unsigned *count, unsigned *maxSymbolValuePtr, const void *source, size_t sourceSize, unsigned *workSpace)
-+{
-+	if (*maxSymbolValuePtr < 255)
-+		return FSE_count_parallel_wksp(count, maxSymbolValuePtr, source, sourceSize, 1, workSpace);
-+	*maxSymbolValuePtr = 255;
-+	return FSE_countFast_wksp(count, maxSymbolValuePtr, source, sourceSize, workSpace);
-+}
-+
-+/*-**************************************************************
-+*  FSE Compression Code
-+****************************************************************/
-+/*! FSE_sizeof_CTable() :
-+	FSE_CTable is a variable size structure which contains :
-+	`U16 tableLog;`
-+	`U16 maxSymbolValue;`
-+	`U16 nextStateNumber[1 << tableLog];`                         // This size is variable
-+	`FSE_symbolCompressionTransform symbolTT[maxSymbolValue+1];`  // This size is variable
-+Allocation is manual (C standard does not support variable-size structures).
-+*/
-+size_t FSE_sizeof_CTable(unsigned maxSymbolValue, unsigned tableLog)
-+{
-+	if (tableLog > FSE_MAX_TABLELOG)
-+		return ERROR(tableLog_tooLarge);
-+	return FSE_CTABLE_SIZE_U32(tableLog, maxSymbolValue) * sizeof(U32);
-+}
-+
-+/* provides the minimum logSize to safely represent a distribution */
-+static unsigned FSE_minTableLog(size_t srcSize, unsigned maxSymbolValue)
-+{
-+	U32 minBitsSrc = BIT_highbit32((U32)(srcSize - 1)) + 1;
-+	U32 minBitsSymbols = BIT_highbit32(maxSymbolValue) + 2;
-+	U32 minBits = minBitsSrc < minBitsSymbols ? minBitsSrc : minBitsSymbols;
-+	return minBits;
-+}
-+
-+unsigned FSE_optimalTableLog_internal(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue, unsigned minus)
-+{
-+	U32 maxBitsSrc = BIT_highbit32((U32)(srcSize - 1)) - minus;
-+	U32 tableLog = maxTableLog;
-+	U32 minBits = FSE_minTableLog(srcSize, maxSymbolValue);
-+	if (tableLog == 0)
-+		tableLog = FSE_DEFAULT_TABLELOG;
-+	if (maxBitsSrc < tableLog)
-+		tableLog = maxBitsSrc; /* Accuracy can be reduced */
-+	if (minBits > tableLog)
-+		tableLog = minBits; /* Need a minimum to safely represent all symbol values */
-+	if (tableLog < FSE_MIN_TABLELOG)
-+		tableLog = FSE_MIN_TABLELOG;
-+	if (tableLog > FSE_MAX_TABLELOG)
-+		tableLog = FSE_MAX_TABLELOG;
-+	return tableLog;
-+}
-+
-+unsigned FSE_optimalTableLog(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue)
-+{
-+	return FSE_optimalTableLog_internal(maxTableLog, srcSize, maxSymbolValue, 2);
-+}
-+
-+/* Secondary normalization method.
-+   To be used when primary method fails. */
-+
-+static size_t FSE_normalizeM2(short *norm, U32 tableLog, const unsigned *count, size_t total, U32 maxSymbolValue)
-+{
-+	short const NOT_YET_ASSIGNED = -2;
-+	U32 s;
-+	U32 distributed = 0;
-+	U32 ToDistribute;
-+
-+	/* Init */
-+	U32 const lowThreshold = (U32)(total >> tableLog);
-+	U32 lowOne = (U32)((total * 3) >> (tableLog + 1));
-+
-+	for (s = 0; s <= maxSymbolValue; s++) {
-+		if (count[s] == 0) {
-+			norm[s] = 0;
-+			continue;
-+		}
-+		if (count[s] <= lowThreshold) {
-+			norm[s] = -1;
-+			distributed++;
-+			total -= count[s];
-+			continue;
-+		}
-+		if (count[s] <= lowOne) {
-+			norm[s] = 1;
-+			distributed++;
-+			total -= count[s];
-+			continue;
-+		}
-+
-+		norm[s] = NOT_YET_ASSIGNED;
-+	}
-+	ToDistribute = (1 << tableLog) - distributed;
-+
-+	if ((total / ToDistribute) > lowOne) {
-+		/* risk of rounding to zero */
-+		lowOne = (U32)((total * 3) / (ToDistribute * 2));
-+		for (s = 0; s <= maxSymbolValue; s++) {
-+			if ((norm[s] == NOT_YET_ASSIGNED) && (count[s] <= lowOne)) {
-+				norm[s] = 1;
-+				distributed++;
-+				total -= count[s];
-+				continue;
-+			}
-+		}
-+		ToDistribute = (1 << tableLog) - distributed;
-+	}
-+
-+	if (distributed == maxSymbolValue + 1) {
-+		/* all values are pretty poor;
-+		   probably incompressible data (should have already been detected);
-+		   find max, then give all remaining points to max */
-+		U32 maxV = 0, maxC = 0;
-+		for (s = 0; s <= maxSymbolValue; s++)
-+			if (count[s] > maxC)
-+				maxV = s, maxC = count[s];
-+		norm[maxV] += (short)ToDistribute;
-+		return 0;
-+	}
-+
-+	if (total == 0) {
-+		/* all of the symbols were low enough for the lowOne or lowThreshold */
-+		for (s = 0; ToDistribute > 0; s = (s + 1) % (maxSymbolValue + 1))
-+			if (norm[s] > 0)
-+				ToDistribute--, norm[s]++;
-+		return 0;
-+	}
-+
-+	{
-+		U64 const vStepLog = 62 - tableLog;
-+		U64 const mid = (1ULL << (vStepLog - 1)) - 1;
-+		U64 const rStep = div_u64((((U64)1 << vStepLog) * ToDistribute) + mid, (U32)total); /* scale on remaining */
-+		U64 tmpTotal = mid;
-+		for (s = 0; s <= maxSymbolValue; s++) {
-+			if (norm[s] == NOT_YET_ASSIGNED) {
-+				U64 const end = tmpTotal + (count[s] * rStep);
-+				U32 const sStart = (U32)(tmpTotal >> vStepLog);
-+				U32 const sEnd = (U32)(end >> vStepLog);
-+				U32 const weight = sEnd - sStart;
-+				if (weight < 1)
-+					return ERROR(GENERIC);
-+				norm[s] = (short)weight;
-+				tmpTotal = end;
-+			}
-+		}
-+	}
-+
-+	return 0;
-+}
-+
-+size_t FSE_normalizeCount(short *normalizedCounter, unsigned tableLog, const unsigned *count, size_t total, unsigned maxSymbolValue)
-+{
-+	/* Sanity checks */
-+	if (tableLog == 0)
-+		tableLog = FSE_DEFAULT_TABLELOG;
-+	if (tableLog < FSE_MIN_TABLELOG)
-+		return ERROR(GENERIC); /* Unsupported size */
-+	if (tableLog > FSE_MAX_TABLELOG)
-+		return ERROR(tableLog_tooLarge); /* Unsupported size */
-+	if (tableLog < FSE_minTableLog(total, maxSymbolValue))
-+		return ERROR(GENERIC); /* Too small tableLog, compression potentially impossible */
-+
-+	{
-+		U32 const rtbTable[] = {0, 473195, 504333, 520860, 550000, 700000, 750000, 830000};
-+		U64 const scale = 62 - tableLog;
-+		U64 const step = div_u64((U64)1 << 62, (U32)total); /* <== here, one division ! */
-+		U64 const vStep = 1ULL << (scale - 20);
-+		int stillToDistribute = 1 << tableLog;
-+		unsigned s;
-+		unsigned largest = 0;
-+		short largestP = 0;
-+		U32 lowThreshold = (U32)(total >> tableLog);
-+
-+		for (s = 0; s <= maxSymbolValue; s++) {
-+			if (count[s] == total)
-+				return 0; /* rle special case */
-+			if (count[s] == 0) {
-+				normalizedCounter[s] = 0;
-+				continue;
-+			}
-+			if (count[s] <= lowThreshold) {
-+				normalizedCounter[s] = -1;
-+				stillToDistribute--;
-+			} else {
-+				short proba = (short)((count[s] * step) >> scale);
-+				if (proba < 8) {
-+					U64 restToBeat = vStep * rtbTable[proba];
-+					proba += (count[s] * step) - ((U64)proba << scale) > restToBeat;
-+				}
-+				if (proba > largestP)
-+					largestP = proba, largest = s;
-+				normalizedCounter[s] = proba;
-+				stillToDistribute -= proba;
-+			}
-+		}
-+		if (-stillToDistribute >= (normalizedCounter[largest] >> 1)) {
-+			/* corner case, need another normalization method */
-+			size_t const errorCode = FSE_normalizeM2(normalizedCounter, tableLog, count, total, maxSymbolValue);
-+			if (FSE_isError(errorCode))
-+				return errorCode;
-+		} else
-+			normalizedCounter[largest] += (short)stillToDistribute;
-+	}
-+
-+	return tableLog;
-+}
-+
-+/* fake FSE_CTable, for raw (uncompressed) input */
-+size_t FSE_buildCTable_raw(FSE_CTable *ct, unsigned nbBits)
-+{
-+	const unsigned tableSize = 1 << nbBits;
-+	const unsigned tableMask = tableSize - 1;
-+	const unsigned maxSymbolValue = tableMask;
-+	void *const ptr = ct;
-+	U16 *const tableU16 = ((U16 *)ptr) + 2;
-+	void *const FSCT = ((U32 *)ptr) + 1 /* header */ + (tableSize >> 1); /* assumption : tableLog >= 1 */
-+	FSE_symbolCompressionTransform *const symbolTT = (FSE_symbolCompressionTransform *)(FSCT);
-+	unsigned s;
-+
-+	/* Sanity checks */
-+	if (nbBits < 1)
-+		return ERROR(GENERIC); /* min size */
-+
-+	/* header */
-+	tableU16[-2] = (U16)nbBits;
-+	tableU16[-1] = (U16)maxSymbolValue;
-+
-+	/* Build table */
-+	for (s = 0; s < tableSize; s++)
-+		tableU16[s] = (U16)(tableSize + s);
-+
-+	/* Build Symbol Transformation Table */
-+	{
-+		const U32 deltaNbBits = (nbBits << 16) - (1 << nbBits);
-+		for (s = 0; s <= maxSymbolValue; s++) {
-+			symbolTT[s].deltaNbBits = deltaNbBits;
-+			symbolTT[s].deltaFindState = s - 1;
-+		}
-+	}
-+
-+	return 0;
-+}
-+
-+/* fake FSE_CTable, for rle input (always same symbol) */
-+size_t FSE_buildCTable_rle(FSE_CTable *ct, BYTE symbolValue)
-+{
-+	void *ptr = ct;
-+	U16 *tableU16 = ((U16 *)ptr) + 2;
-+	void *FSCTptr = (U32 *)ptr + 2;
-+	FSE_symbolCompressionTransform *symbolTT = (FSE_symbolCompressionTransform *)FSCTptr;
-+
-+	/* header */
-+	tableU16[-2] = (U16)0;
-+	tableU16[-1] = (U16)symbolValue;
-+
-+	/* Build table */
-+	tableU16[0] = 0;
-+	tableU16[1] = 0; /* just in case */
-+
-+	/* Build Symbol Transformation Table */
-+	symbolTT[symbolValue].deltaNbBits = 0;
-+	symbolTT[symbolValue].deltaFindState = 0;
-+
-+	return 0;
-+}
-+
-+static size_t FSE_compress_usingCTable_generic(void *dst, size_t dstSize, const void *src, size_t srcSize, const FSE_CTable *ct, const unsigned fast)
-+{
-+	const BYTE *const istart = (const BYTE *)src;
-+	const BYTE *const iend = istart + srcSize;
-+	const BYTE *ip = iend;
-+
-+	BIT_CStream_t bitC;
-+	FSE_CState_t CState1, CState2;
-+
-+	/* init */
-+	if (srcSize <= 2)
-+		return 0;
-+	{
-+		size_t const initError = BIT_initCStream(&bitC, dst, dstSize);
-+		if (FSE_isError(initError))
-+			return 0; /* not enough space available to write a bitstream */
-+	}
-+
-+#define FSE_FLUSHBITS(s) (fast ? BIT_flushBitsFast(s) : BIT_flushBits(s))
-+
-+	if (srcSize & 1) {
-+		FSE_initCState2(&CState1, ct, *--ip);
-+		FSE_initCState2(&CState2, ct, *--ip);
-+		FSE_encodeSymbol(&bitC, &CState1, *--ip);
-+		FSE_FLUSHBITS(&bitC);
-+	} else {
-+		FSE_initCState2(&CState2, ct, *--ip);
-+		FSE_initCState2(&CState1, ct, *--ip);
-+	}
-+
-+	/* join to mod 4 */
-+	srcSize -= 2;
-+	if ((sizeof(bitC.bitContainer) * 8 > FSE_MAX_TABLELOG * 4 + 7) && (srcSize & 2)) { /* test bit 2 */
-+		FSE_encodeSymbol(&bitC, &CState2, *--ip);
-+		FSE_encodeSymbol(&bitC, &CState1, *--ip);
-+		FSE_FLUSHBITS(&bitC);
-+	}
-+
-+	/* 2 or 4 encoding per loop */
-+	while (ip > istart) {
-+
-+		FSE_encodeSymbol(&bitC, &CState2, *--ip);
-+
-+		if (sizeof(bitC.bitContainer) * 8 < FSE_MAX_TABLELOG * 2 + 7) /* this test must be static */
-+			FSE_FLUSHBITS(&bitC);
-+
-+		FSE_encodeSymbol(&bitC, &CState1, *--ip);
-+
-+		if (sizeof(bitC.bitContainer) * 8 > FSE_MAX_TABLELOG * 4 + 7) { /* this test must be static */
-+			FSE_encodeSymbol(&bitC, &CState2, *--ip);
-+			FSE_encodeSymbol(&bitC, &CState1, *--ip);
-+		}
-+
-+		FSE_FLUSHBITS(&bitC);
-+	}
-+
-+	FSE_flushCState(&bitC, &CState2);
-+	FSE_flushCState(&bitC, &CState1);
-+	return BIT_closeCStream(&bitC);
-+}
-+
-+size_t FSE_compress_usingCTable(void *dst, size_t dstSize, const void *src, size_t srcSize, const FSE_CTable *ct)
-+{
-+	unsigned const fast = (dstSize >= FSE_BLOCKBOUND(srcSize));
-+
-+	if (fast)
-+		return FSE_compress_usingCTable_generic(dst, dstSize, src, srcSize, ct, 1);
-+	else
-+		return FSE_compress_usingCTable_generic(dst, dstSize, src, srcSize, ct, 0);
-+}
-+
-+size_t FSE_compressBound(size_t size) { return FSE_COMPRESSBOUND(size); }
-diff --git a/lib/zstd/fse_decompress.c b/lib/zstd/fse_decompress.c
-new file mode 100644
-index 0000000..a84300e
---- /dev/null
-+++ b/lib/zstd/fse_decompress.c
-@@ -0,0 +1,332 @@
-+/*
-+ * FSE : Finite State Entropy decoder
-+ * Copyright (C) 2013-2015, Yann Collet.
-+ *
-+ * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
-+ *
-+ * Redistribution and use in source and binary forms, with or without
-+ * modification, are permitted provided that the following conditions are
-+ * met:
-+ *
-+ *   * Redistributions of source code must retain the above copyright
-+ * notice, this list of conditions and the following disclaimer.
-+ *   * Redistributions in binary form must reproduce the above
-+ * copyright notice, this list of conditions and the following disclaimer
-+ * in the documentation and/or other materials provided with the
-+ * distribution.
-+ *
-+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
-+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
-+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
-+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
-+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
-+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
-+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
-+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
-+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
-+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
-+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-+ *
-+ * This program is free software; you can redistribute it and/or modify it under
-+ * the terms of the GNU General Public License version 2 as published by the
-+ * Free Software Foundation. This program is dual-licensed; you may select
-+ * either version 2 of the GNU General Public License ("GPL") or BSD license
-+ * ("BSD").
-+ *
-+ * You can contact the author at :
-+ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
-+ */
-+
-+/* **************************************************************
-+*  Compiler specifics
-+****************************************************************/
-+#define FORCE_INLINE static __always_inline
-+
-+/* **************************************************************
-+*  Includes
-+****************************************************************/
-+#include "bitstream.h"
-+#include "fse.h"
-+#include <linux/compiler.h>
-+#include <linux/kernel.h>
-+#include <linux/string.h> /* memcpy, memset */
-+
-+/* **************************************************************
-+*  Error Management
-+****************************************************************/
-+#define FSE_isError ERR_isError
-+#define FSE_STATIC_ASSERT(c)                                   \
-+	{                                                      \
-+		enum { FSE_static_assert = 1 / (int)(!!(c)) }; \
-+	} /* use only *after* variable declarations */
-+
-+/* check and forward error code */
-+#define CHECK_F(f)                  \
-+	{                           \
-+		size_t const e = f; \
-+		if (FSE_isError(e)) \
-+			return e;   \
-+	}
-+
-+/* **************************************************************
-+*  Templates
-+****************************************************************/
-+/*
-+  designed to be included
-+  for type-specific functions (template emulation in C)
-+  Objective is to write these functions only once, for improved maintenance
-+*/
-+
-+/* safety checks */
-+#ifndef FSE_FUNCTION_EXTENSION
-+#error "FSE_FUNCTION_EXTENSION must be defined"
-+#endif
-+#ifndef FSE_FUNCTION_TYPE
-+#error "FSE_FUNCTION_TYPE must be defined"
-+#endif
-+
-+/* Function names */
-+#define FSE_CAT(X, Y) X##Y
-+#define FSE_FUNCTION_NAME(X, Y) FSE_CAT(X, Y)
-+#define FSE_TYPE_NAME(X, Y) FSE_CAT(X, Y)
-+
-+/* Function templates */
-+
-+size_t FSE_buildDTable_wksp(FSE_DTable *dt, const short *normalizedCounter, unsigned maxSymbolValue, unsigned tableLog, void *workspace, size_t workspaceSize)
-+{
-+	void *const tdPtr = dt + 1; /* because *dt is unsigned, 32-bits aligned on 32-bits */
-+	FSE_DECODE_TYPE *const tableDecode = (FSE_DECODE_TYPE *)(tdPtr);
-+	U16 *symbolNext = (U16 *)workspace;
-+
-+	U32 const maxSV1 = maxSymbolValue + 1;
-+	U32 const tableSize = 1 << tableLog;
-+	U32 highThreshold = tableSize - 1;
-+
-+	/* Sanity Checks */
-+	if (workspaceSize < sizeof(U16) * (FSE_MAX_SYMBOL_VALUE + 1))
-+		return ERROR(tableLog_tooLarge);
-+	if (maxSymbolValue > FSE_MAX_SYMBOL_VALUE)
-+		return ERROR(maxSymbolValue_tooLarge);
-+	if (tableLog > FSE_MAX_TABLELOG)
-+		return ERROR(tableLog_tooLarge);
-+
-+	/* Init, lay down lowprob symbols */
-+	{
-+		FSE_DTableHeader DTableH;
-+		DTableH.tableLog = (U16)tableLog;
-+		DTableH.fastMode = 1;
-+		{
-+			S16 const largeLimit = (S16)(1 << (tableLog - 1));
-+			U32 s;
-+			for (s = 0; s < maxSV1; s++) {
-+				if (normalizedCounter[s] == -1) {
-+					tableDecode[highThreshold--].symbol = (FSE_FUNCTION_TYPE)s;
-+					symbolNext[s] = 1;
-+				} else {
-+					if (normalizedCounter[s] >= largeLimit)
-+						DTableH.fastMode = 0;
-+					symbolNext[s] = normalizedCounter[s];
-+				}
-+			}
-+		}
-+		memcpy(dt, &DTableH, sizeof(DTableH));
-+	}
-+
-+	/* Spread symbols */
-+	{
-+		U32 const tableMask = tableSize - 1;
-+		U32 const step = FSE_TABLESTEP(tableSize);
-+		U32 s, position = 0;
-+		for (s = 0; s < maxSV1; s++) {
-+			int i;
-+			for (i = 0; i < normalizedCounter[s]; i++) {
-+				tableDecode[position].symbol = (FSE_FUNCTION_TYPE)s;
-+				position = (position + step) & tableMask;
-+				while (position > highThreshold)
-+					position = (position + step) & tableMask; /* lowprob area */
-+			}
-+		}
-+		if (position != 0)
-+			return ERROR(GENERIC); /* position must reach all cells once, otherwise normalizedCounter is incorrect */
-+	}
-+
-+	/* Build Decoding table */
-+	{
-+		U32 u;
-+		for (u = 0; u < tableSize; u++) {
-+			FSE_FUNCTION_TYPE const symbol = (FSE_FUNCTION_TYPE)(tableDecode[u].symbol);
-+			U16 nextState = symbolNext[symbol]++;
-+			tableDecode[u].nbBits = (BYTE)(tableLog - BIT_highbit32((U32)nextState));
-+			tableDecode[u].newState = (U16)((nextState << tableDecode[u].nbBits) - tableSize);
-+		}
-+	}
-+
-+	return 0;
-+}
-+
-+/*-*******************************************************
-+*  Decompression (Byte symbols)
-+*********************************************************/
-+size_t FSE_buildDTable_rle(FSE_DTable *dt, BYTE symbolValue)
-+{
-+	void *ptr = dt;
-+	FSE_DTableHeader *const DTableH = (FSE_DTableHeader *)ptr;
-+	void *dPtr = dt + 1;
-+	FSE_decode_t *const cell = (FSE_decode_t *)dPtr;
-+
-+	DTableH->tableLog = 0;
-+	DTableH->fastMode = 0;
-+
-+	cell->newState = 0;
-+	cell->symbol = symbolValue;
-+	cell->nbBits = 0;
-+
-+	return 0;
-+}
-+
-+size_t FSE_buildDTable_raw(FSE_DTable *dt, unsigned nbBits)
-+{
-+	void *ptr = dt;
-+	FSE_DTableHeader *const DTableH = (FSE_DTableHeader *)ptr;
-+	void *dPtr = dt + 1;
-+	FSE_decode_t *const dinfo = (FSE_decode_t *)dPtr;
-+	const unsigned tableSize = 1 << nbBits;
-+	const unsigned tableMask = tableSize - 1;
-+	const unsigned maxSV1 = tableMask + 1;
-+	unsigned s;
-+
-+	/* Sanity checks */
-+	if (nbBits < 1)
-+		return ERROR(GENERIC); /* min size */
-+
-+	/* Build Decoding Table */
-+	DTableH->tableLog = (U16)nbBits;
-+	DTableH->fastMode = 1;
-+	for (s = 0; s < maxSV1; s++) {
-+		dinfo[s].newState = 0;
-+		dinfo[s].symbol = (BYTE)s;
-+		dinfo[s].nbBits = (BYTE)nbBits;
-+	}
-+
-+	return 0;
-+}
-+
-+FORCE_INLINE size_t FSE_decompress_usingDTable_generic(void *dst, size_t maxDstSize, const void *cSrc, size_t cSrcSize, const FSE_DTable *dt,
-+						       const unsigned fast)
-+{
-+	BYTE *const ostart = (BYTE *)dst;
-+	BYTE *op = ostart;
-+	BYTE *const omax = op + maxDstSize;
-+	BYTE *const olimit = omax - 3;
-+
-+	BIT_DStream_t bitD;
-+	FSE_DState_t state1;
-+	FSE_DState_t state2;
-+
-+	/* Init */
-+	CHECK_F(BIT_initDStream(&bitD, cSrc, cSrcSize));
-+
-+	FSE_initDState(&state1, &bitD, dt);
-+	FSE_initDState(&state2, &bitD, dt);
-+
-+#define FSE_GETSYMBOL(statePtr) fast ? FSE_decodeSymbolFast(statePtr, &bitD) : FSE_decodeSymbol(statePtr, &bitD)
-+
-+	/* 4 symbols per loop */
-+	for (; (BIT_reloadDStream(&bitD) == BIT_DStream_unfinished) & (op < olimit); op += 4) {
-+		op[0] = FSE_GETSYMBOL(&state1);
-+
-+		if (FSE_MAX_TABLELOG * 2 + 7 > sizeof(bitD.bitContainer) * 8) /* This test must be static */
-+			BIT_reloadDStream(&bitD);
-+
-+		op[1] = FSE_GETSYMBOL(&state2);
-+
-+		if (FSE_MAX_TABLELOG * 4 + 7 > sizeof(bitD.bitContainer) * 8) /* This test must be static */
-+		{
-+			if (BIT_reloadDStream(&bitD) > BIT_DStream_unfinished) {
-+				op += 2;
-+				break;
-+			}
-+		}
-+
-+		op[2] = FSE_GETSYMBOL(&state1);
-+
-+		if (FSE_MAX_TABLELOG * 2 + 7 > sizeof(bitD.bitContainer) * 8) /* This test must be static */
-+			BIT_reloadDStream(&bitD);
-+
-+		op[3] = FSE_GETSYMBOL(&state2);
-+	}
-+
-+	/* tail */
-+	/* note : BIT_reloadDStream(&bitD) >= FSE_DStream_partiallyFilled; Ends at exactly BIT_DStream_completed */
-+	while (1) {
-+		if (op > (omax - 2))
-+			return ERROR(dstSize_tooSmall);
-+		*op++ = FSE_GETSYMBOL(&state1);
-+		if (BIT_reloadDStream(&bitD) == BIT_DStream_overflow) {
-+			*op++ = FSE_GETSYMBOL(&state2);
-+			break;
-+		}
-+
-+		if (op > (omax - 2))
-+			return ERROR(dstSize_tooSmall);
-+		*op++ = FSE_GETSYMBOL(&state2);
-+		if (BIT_reloadDStream(&bitD) == BIT_DStream_overflow) {
-+			*op++ = FSE_GETSYMBOL(&state1);
-+			break;
-+		}
-+	}
-+
-+	return op - ostart;
-+}
-+
-+size_t FSE_decompress_usingDTable(void *dst, size_t originalSize, const void *cSrc, size_t cSrcSize, const FSE_DTable *dt)
-+{
-+	const void *ptr = dt;
-+	const FSE_DTableHeader *DTableH = (const FSE_DTableHeader *)ptr;
-+	const U32 fastMode = DTableH->fastMode;
-+
-+	/* select fast mode (static) */
-+	if (fastMode)
-+		return FSE_decompress_usingDTable_generic(dst, originalSize, cSrc, cSrcSize, dt, 1);
-+	return FSE_decompress_usingDTable_generic(dst, originalSize, cSrc, cSrcSize, dt, 0);
-+}
-+
-+size_t FSE_decompress_wksp(void *dst, size_t dstCapacity, const void *cSrc, size_t cSrcSize, unsigned maxLog, void *workspace, size_t workspaceSize)
-+{
-+	const BYTE *const istart = (const BYTE *)cSrc;
-+	const BYTE *ip = istart;
-+	unsigned tableLog;
-+	unsigned maxSymbolValue = FSE_MAX_SYMBOL_VALUE;
-+	size_t NCountLength;
-+
-+	FSE_DTable *dt;
-+	short *counting;
-+	size_t spaceUsed32 = 0;
-+
-+	FSE_STATIC_ASSERT(sizeof(FSE_DTable) == sizeof(U32));
-+
-+	dt = (FSE_DTable *)((U32 *)workspace + spaceUsed32);
-+	spaceUsed32 += FSE_DTABLE_SIZE_U32(maxLog);
-+	counting = (short *)((U32 *)workspace + spaceUsed32);
-+	spaceUsed32 += ALIGN(sizeof(short) * (FSE_MAX_SYMBOL_VALUE + 1), sizeof(U32)) >> 2;
-+
-+	if ((spaceUsed32 << 2) > workspaceSize)
-+		return ERROR(tableLog_tooLarge);
-+	workspace = (U32 *)workspace + spaceUsed32;
-+	workspaceSize -= (spaceUsed32 << 2);
-+
-+	/* normal FSE decoding mode */
-+	NCountLength = FSE_readNCount(counting, &maxSymbolValue, &tableLog, istart, cSrcSize);
-+	if (FSE_isError(NCountLength))
-+		return NCountLength;
-+	// if (NCountLength >= cSrcSize) return ERROR(srcSize_wrong);   /* too small input size; supposed to be already checked in NCountLength, only remaining
-+	// case : NCountLength==cSrcSize */
-+	if (tableLog > maxLog)
-+		return ERROR(tableLog_tooLarge);
-+	ip += NCountLength;
-+	cSrcSize -= NCountLength;
-+
-+	CHECK_F(FSE_buildDTable_wksp(dt, counting, maxSymbolValue, tableLog, workspace, workspaceSize));
-+
-+	return FSE_decompress_usingDTable(dst, dstCapacity, ip, cSrcSize, dt); /* always return, even if it is an error code */
-+}
-diff --git a/lib/zstd/huf.h b/lib/zstd/huf.h
-new file mode 100644
-index 0000000..2143da2
---- /dev/null
-+++ b/lib/zstd/huf.h
-@@ -0,0 +1,212 @@
-+/*
-+ * Huffman coder, part of New Generation Entropy library
-+ * header file
-+ * Copyright (C) 2013-2016, Yann Collet.
-+ *
-+ * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
-+ *
-+ * Redistribution and use in source and binary forms, with or without
-+ * modification, are permitted provided that the following conditions are
-+ * met:
-+ *
-+ *   * Redistributions of source code must retain the above copyright
-+ * notice, this list of conditions and the following disclaimer.
-+ *   * Redistributions in binary form must reproduce the above
-+ * copyright notice, this list of conditions and the following disclaimer
-+ * in the documentation and/or other materials provided with the
-+ * distribution.
-+ *
-+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
-+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
-+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
-+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
-+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
-+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
-+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
-+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
-+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
-+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
-+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-+ *
-+ * This program is free software; you can redistribute it and/or modify it under
-+ * the terms of the GNU General Public License version 2 as published by the
-+ * Free Software Foundation. This program is dual-licensed; you may select
-+ * either version 2 of the GNU General Public License ("GPL") or BSD license
-+ * ("BSD").
-+ *
-+ * You can contact the author at :
-+ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
-+ */
-+#ifndef HUF_H_298734234
-+#define HUF_H_298734234
-+
-+/* *** Dependencies *** */
-+#include <linux/types.h> /* size_t */
-+
-+/* ***   Tool functions *** */
-+#define HUF_BLOCKSIZE_MAX (128 * 1024) /**< maximum input size for a single block compressed with HUF_compress */
-+size_t HUF_compressBound(size_t size); /**< maximum compressed size (worst case) */
-+
-+/* Error Management */
-+unsigned HUF_isError(size_t code); /**< tells if a return value is an error code */
-+
-+/* ***   Advanced function   *** */
-+
-+/** HUF_compress4X_wksp() :
-+*   Same as HUF_compress2(), but uses externally allocated `workSpace`, which must be a table of >= 1024 unsigned */
-+size_t HUF_compress4X_wksp(void *dst, size_t dstSize, const void *src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog, void *workSpace,
-+			   size_t wkspSize); /**< `workSpace` must be a table of at least HUF_COMPRESS_WORKSPACE_SIZE_U32 unsigned */
-+
-+/* *** Dependencies *** */
-+#include "mem.h" /* U32 */
-+
-+/* *** Constants *** */
-+#define HUF_TABLELOG_MAX 12     /* max configured tableLog (for static allocation); can be modified up to HUF_ABSOLUTEMAX_TABLELOG */
-+#define HUF_TABLELOG_DEFAULT 11 /* tableLog by default, when not specified */
-+#define HUF_SYMBOLVALUE_MAX 255
-+
-+#define HUF_TABLELOG_ABSOLUTEMAX 15 /* absolute limit of HUF_MAX_TABLELOG. Beyond that value, code does not work */
-+#if (HUF_TABLELOG_MAX > HUF_TABLELOG_ABSOLUTEMAX)
-+#error "HUF_TABLELOG_MAX is too large !"
-+#endif
-+
-+/* ****************************************
-+*  Static allocation
-+******************************************/
-+/* HUF buffer bounds */
-+#define HUF_CTABLEBOUND 129
-+#define HUF_BLOCKBOUND(size) (size + (size >> 8) + 8)			 /* only true if incompressible pre-filtered with fast heuristic */
-+#define HUF_COMPRESSBOUND(size) (HUF_CTABLEBOUND + HUF_BLOCKBOUND(size)) /* Macro version, useful for static allocation */
-+
-+/* static allocation of HUF's Compression Table */
-+#define HUF_CREATE_STATIC_CTABLE(name, maxSymbolValue) \
-+	U32 name##hb[maxSymbolValue + 1];              \
-+	void *name##hv = &(name##hb);                  \
-+	HUF_CElt *name = (HUF_CElt *)(name##hv) /* no final ; */
-+
-+/* static allocation of HUF's DTable */
-+typedef U32 HUF_DTable;
-+#define HUF_DTABLE_SIZE(maxTableLog) (1 + (1 << (maxTableLog)))
-+#define HUF_CREATE_STATIC_DTABLEX2(DTable, maxTableLog) HUF_DTable DTable[HUF_DTABLE_SIZE((maxTableLog)-1)] = {((U32)((maxTableLog)-1) * 0x01000001)}
-+#define HUF_CREATE_STATIC_DTABLEX4(DTable, maxTableLog) HUF_DTable DTable[HUF_DTABLE_SIZE(maxTableLog)] = {((U32)(maxTableLog)*0x01000001)}
-+
-+/* The workspace must have alignment at least 4 and be at least this large */
-+#define HUF_COMPRESS_WORKSPACE_SIZE (6 << 10)
-+#define HUF_COMPRESS_WORKSPACE_SIZE_U32 (HUF_COMPRESS_WORKSPACE_SIZE / sizeof(U32))
-+
-+/* The workspace must have alignment at least 4 and be at least this large */
-+#define HUF_DECOMPRESS_WORKSPACE_SIZE (3 << 10)
-+#define HUF_DECOMPRESS_WORKSPACE_SIZE_U32 (HUF_DECOMPRESS_WORKSPACE_SIZE / sizeof(U32))
-+
-+/* ****************************************
-+*  Advanced decompression functions
-+******************************************/
-+size_t HUF_decompress4X_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, size_t workspaceSize); /**< decodes RLE and uncompressed */
-+size_t HUF_decompress4X_hufOnly_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace,
-+				size_t workspaceSize);							       /**< considers RLE and uncompressed as errors */
-+size_t HUF_decompress4X2_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace,
-+				   size_t workspaceSize); /**< single-symbol decoder */
-+size_t HUF_decompress4X4_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace,
-+				   size_t workspaceSize); /**< double-symbols decoder */
-+
-+/* ****************************************
-+*  HUF detailed API
-+******************************************/
-+/*!
-+HUF_compress() does the following:
-+1. count symbol occurrence from source[] into table count[] using FSE_count()
-+2. (optional) refine tableLog using HUF_optimalTableLog()
-+3. build Huffman table from count using HUF_buildCTable()
-+4. save Huffman table to memory buffer using HUF_writeCTable_wksp()
-+5. encode the data stream using HUF_compress4X_usingCTable()
-+
-+The following API allows targeting specific sub-functions for advanced tasks.
-+For example, it's possible to compress several blocks using the same 'CTable',
-+or to save and regenerate 'CTable' using external methods.
-+*/
-+/* FSE_count() : find it within "fse.h" */
-+unsigned HUF_optimalTableLog(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue);
-+typedef struct HUF_CElt_s HUF_CElt; /* incomplete type */
-+size_t HUF_writeCTable_wksp(void *dst, size_t maxDstSize, const HUF_CElt *CTable, unsigned maxSymbolValue, unsigned huffLog, void *workspace, size_t workspaceSize);
-+size_t HUF_compress4X_usingCTable(void *dst, size_t dstSize, const void *src, size_t srcSize, const HUF_CElt *CTable);
-+
-+typedef enum {
-+	HUF_repeat_none,  /**< Cannot use the previous table */
-+	HUF_repeat_check, /**< Can use the previous table but it must be checked. Note : The previous table must have been constructed by HUF_compress{1,
-+			     4}X_repeat */
-+	HUF_repeat_valid  /**< Can use the previous table and it is asumed to be valid */
-+} HUF_repeat;
-+/** HUF_compress4X_repeat() :
-+*   Same as HUF_compress4X_wksp(), but considers using hufTable if *repeat != HUF_repeat_none.
-+*   If it uses hufTable it does not modify hufTable or repeat.
-+*   If it doesn't, it sets *repeat = HUF_repeat_none, and it sets hufTable to the table used.
-+*   If preferRepeat then the old table will always be used if valid. */
-+size_t HUF_compress4X_repeat(void *dst, size_t dstSize, const void *src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog, void *workSpace,
-+			     size_t wkspSize, HUF_CElt *hufTable, HUF_repeat *repeat,
-+			     int preferRepeat); /**< `workSpace` must be a table of at least HUF_COMPRESS_WORKSPACE_SIZE_U32 unsigned */
-+
-+/** HUF_buildCTable_wksp() :
-+ *  Same as HUF_buildCTable(), but using externally allocated scratch buffer.
-+ *  `workSpace` must be aligned on 4-bytes boundaries, and be at least as large as a table of 1024 unsigned.
-+ */
-+size_t HUF_buildCTable_wksp(HUF_CElt *tree, const U32 *count, U32 maxSymbolValue, U32 maxNbBits, void *workSpace, size_t wkspSize);
-+
-+/*! HUF_readStats() :
-+	Read compact Huffman tree, saved by HUF_writeCTable().
-+	`huffWeight` is destination buffer.
-+	@return : size read from `src` , or an error Code .
-+	Note : Needed by HUF_readCTable() and HUF_readDTableXn() . */
-+size_t HUF_readStats_wksp(BYTE *huffWeight, size_t hwSize, U32 *rankStats, U32 *nbSymbolsPtr, U32 *tableLogPtr, const void *src, size_t srcSize,
-+			  void *workspace, size_t workspaceSize);
-+
-+/** HUF_readCTable() :
-+*   Loading a CTable saved with HUF_writeCTable() */
-+size_t HUF_readCTable_wksp(HUF_CElt *CTable, unsigned maxSymbolValue, const void *src, size_t srcSize, void *workspace, size_t workspaceSize);
-+
-+/*
-+HUF_decompress() does the following:
-+1. select the decompression algorithm (X2, X4) based on pre-computed heuristics
-+2. build Huffman table from save, using HUF_readDTableXn()
-+3. decode 1 or 4 segments in parallel using HUF_decompressSXn_usingDTable
-+*/
-+
-+/** HUF_selectDecoder() :
-+*   Tells which decoder is likely to decode faster,
-+*   based on a set of pre-determined metrics.
-+*   @return : 0==HUF_decompress4X2, 1==HUF_decompress4X4 .
-+*   Assumption : 0 < cSrcSize < dstSize <= 128 KB */
-+U32 HUF_selectDecoder(size_t dstSize, size_t cSrcSize);
-+
-+size_t HUF_readDTableX2_wksp(HUF_DTable *DTable, const void *src, size_t srcSize, void *workspace, size_t workspaceSize);
-+size_t HUF_readDTableX4_wksp(HUF_DTable *DTable, const void *src, size_t srcSize, void *workspace, size_t workspaceSize);
-+
-+size_t HUF_decompress4X_usingDTable(void *dst, size_t maxDstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable);
-+size_t HUF_decompress4X2_usingDTable(void *dst, size_t maxDstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable);
-+size_t HUF_decompress4X4_usingDTable(void *dst, size_t maxDstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable);
-+
-+/* single stream variants */
-+
-+size_t HUF_compress1X_wksp(void *dst, size_t dstSize, const void *src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog, void *workSpace,
-+			   size_t wkspSize); /**< `workSpace` must be a table of at least HUF_COMPRESS_WORKSPACE_SIZE_U32 unsigned */
-+size_t HUF_compress1X_usingCTable(void *dst, size_t dstSize, const void *src, size_t srcSize, const HUF_CElt *CTable);
-+/** HUF_compress1X_repeat() :
-+*   Same as HUF_compress1X_wksp(), but considers using hufTable if *repeat != HUF_repeat_none.
-+*   If it uses hufTable it does not modify hufTable or repeat.
-+*   If it doesn't, it sets *repeat = HUF_repeat_none, and it sets hufTable to the table used.
-+*   If preferRepeat then the old table will always be used if valid. */
-+size_t HUF_compress1X_repeat(void *dst, size_t dstSize, const void *src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog, void *workSpace,
-+			     size_t wkspSize, HUF_CElt *hufTable, HUF_repeat *repeat,
-+			     int preferRepeat); /**< `workSpace` must be a table of at least HUF_COMPRESS_WORKSPACE_SIZE_U32 unsigned */
-+
-+size_t HUF_decompress1X_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, size_t workspaceSize);
-+size_t HUF_decompress1X2_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace,
-+				   size_t workspaceSize); /**< single-symbol decoder */
-+size_t HUF_decompress1X4_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace,
-+				   size_t workspaceSize); /**< double-symbols decoder */
-+
-+size_t HUF_decompress1X_usingDTable(void *dst, size_t maxDstSize, const void *cSrc, size_t cSrcSize,
-+				    const HUF_DTable *DTable); /**< automatic selection of sing or double symbol decoder, based on DTable */
-+size_t HUF_decompress1X2_usingDTable(void *dst, size_t maxDstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable);
-+size_t HUF_decompress1X4_usingDTable(void *dst, size_t maxDstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable);
-+
-+#endif /* HUF_H_298734234 */
-diff --git a/lib/zstd/huf_compress.c b/lib/zstd/huf_compress.c
-new file mode 100644
-index 0000000..40055a7
---- /dev/null
-+++ b/lib/zstd/huf_compress.c
-@@ -0,0 +1,770 @@
-+/*
-+ * Huffman encoder, part of New Generation Entropy library
-+ * Copyright (C) 2013-2016, Yann Collet.
-+ *
-+ * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
-+ *
-+ * Redistribution and use in source and binary forms, with or without
-+ * modification, are permitted provided that the following conditions are
-+ * met:
-+ *
-+ *   * Redistributions of source code must retain the above copyright
-+ * notice, this list of conditions and the following disclaimer.
-+ *   * Redistributions in binary form must reproduce the above
-+ * copyright notice, this list of conditions and the following disclaimer
-+ * in the documentation and/or other materials provided with the
-+ * distribution.
-+ *
-+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
-+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
-+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
-+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
-+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
-+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
-+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
-+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
-+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
-+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
-+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-+ *
-+ * This program is free software; you can redistribute it and/or modify it under
-+ * the terms of the GNU General Public License version 2 as published by the
-+ * Free Software Foundation. This program is dual-licensed; you may select
-+ * either version 2 of the GNU General Public License ("GPL") or BSD license
-+ * ("BSD").
-+ *
-+ * You can contact the author at :
-+ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
-+ */
-+
-+/* **************************************************************
-+*  Includes
-+****************************************************************/
-+#include "bitstream.h"
-+#include "fse.h" /* header compression */
-+#include "huf.h"
-+#include <linux/kernel.h>
-+#include <linux/string.h> /* memcpy, memset */
-+
-+/* **************************************************************
-+*  Error Management
-+****************************************************************/
-+#define HUF_STATIC_ASSERT(c)                                   \
-+	{                                                      \
-+		enum { HUF_static_assert = 1 / (int)(!!(c)) }; \
-+	} /* use only *after* variable declarations */
-+#define CHECK_V_F(e, f)     \
-+	size_t const e = f; \
-+	if (ERR_isError(e)) \
-+	return f
-+#define CHECK_F(f)                        \
-+	{                                 \
-+		CHECK_V_F(_var_err__, f); \
-+	}
-+
-+/* **************************************************************
-+*  Utils
-+****************************************************************/
-+unsigned HUF_optimalTableLog(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue)
-+{
-+	return FSE_optimalTableLog_internal(maxTableLog, srcSize, maxSymbolValue, 1);
-+}
-+
-+/* *******************************************************
-+*  HUF : Huffman block compression
-+*********************************************************/
-+/* HUF_compressWeights() :
-+ * Same as FSE_compress(), but dedicated to huff0's weights compression.
-+ * The use case needs much less stack memory.
-+ * Note : all elements within weightTable are supposed to be <= HUF_TABLELOG_MAX.
-+ */
-+#define MAX_FSE_TABLELOG_FOR_HUFF_HEADER 6
-+size_t HUF_compressWeights_wksp(void *dst, size_t dstSize, const void *weightTable, size_t wtSize, void *workspace, size_t workspaceSize)
-+{
-+	BYTE *const ostart = (BYTE *)dst;
-+	BYTE *op = ostart;
-+	BYTE *const oend = ostart + dstSize;
-+
-+	U32 maxSymbolValue = HUF_TABLELOG_MAX;
-+	U32 tableLog = MAX_FSE_TABLELOG_FOR_HUFF_HEADER;
-+
-+	FSE_CTable *CTable;
-+	U32 *count;
-+	S16 *norm;
-+	size_t spaceUsed32 = 0;
-+
-+	HUF_STATIC_ASSERT(sizeof(FSE_CTable) == sizeof(U32));
-+
-+	CTable = (FSE_CTable *)((U32 *)workspace + spaceUsed32);
-+	spaceUsed32 += FSE_CTABLE_SIZE_U32(MAX_FSE_TABLELOG_FOR_HUFF_HEADER, HUF_TABLELOG_MAX);
-+	count = (U32 *)workspace + spaceUsed32;
-+	spaceUsed32 += HUF_TABLELOG_MAX + 1;
-+	norm = (S16 *)((U32 *)workspace + spaceUsed32);
-+	spaceUsed32 += ALIGN(sizeof(S16) * (HUF_TABLELOG_MAX + 1), sizeof(U32)) >> 2;
-+
-+	if ((spaceUsed32 << 2) > workspaceSize)
-+		return ERROR(tableLog_tooLarge);
-+	workspace = (U32 *)workspace + spaceUsed32;
-+	workspaceSize -= (spaceUsed32 << 2);
-+
-+	/* init conditions */
-+	if (wtSize <= 1)
-+		return 0; /* Not compressible */
-+
-+	/* Scan input and build symbol stats */
-+	{
-+		CHECK_V_F(maxCount, FSE_count_simple(count, &maxSymbolValue, weightTable, wtSize));
-+		if (maxCount == wtSize)
-+			return 1; /* only a single symbol in src : rle */
-+		if (maxCount == 1)
-+			return 0; /* each symbol present maximum once => not compressible */
-+	}
-+
-+	tableLog = FSE_optimalTableLog(tableLog, wtSize, maxSymbolValue);
-+	CHECK_F(FSE_normalizeCount(norm, tableLog, count, wtSize, maxSymbolValue));
-+
-+	/* Write table description header */
-+	{
-+		CHECK_V_F(hSize, FSE_writeNCount(op, oend - op, norm, maxSymbolValue, tableLog));
-+		op += hSize;
-+	}
-+
-+	/* Compress */
-+	CHECK_F(FSE_buildCTable_wksp(CTable, norm, maxSymbolValue, tableLog, workspace, workspaceSize));
-+	{
-+		CHECK_V_F(cSize, FSE_compress_usingCTable(op, oend - op, weightTable, wtSize, CTable));
-+		if (cSize == 0)
-+			return 0; /* not enough space for compressed data */
-+		op += cSize;
-+	}
-+
-+	return op - ostart;
-+}
-+
-+struct HUF_CElt_s {
-+	U16 val;
-+	BYTE nbBits;
-+}; /* typedef'd to HUF_CElt within "huf.h" */
-+
-+/*! HUF_writeCTable_wksp() :
-+	`CTable` : Huffman tree to save, using huf representation.
-+	@return : size of saved CTable */
-+size_t HUF_writeCTable_wksp(void *dst, size_t maxDstSize, const HUF_CElt *CTable, U32 maxSymbolValue, U32 huffLog, void *workspace, size_t workspaceSize)
-+{
-+	BYTE *op = (BYTE *)dst;
-+	U32 n;
-+
-+	BYTE *bitsToWeight;
-+	BYTE *huffWeight;
-+	size_t spaceUsed32 = 0;
-+
-+	bitsToWeight = (BYTE *)((U32 *)workspace + spaceUsed32);
-+	spaceUsed32 += ALIGN(HUF_TABLELOG_MAX + 1, sizeof(U32)) >> 2;
-+	huffWeight = (BYTE *)((U32 *)workspace + spaceUsed32);
-+	spaceUsed32 += ALIGN(HUF_SYMBOLVALUE_MAX, sizeof(U32)) >> 2;
-+
-+	if ((spaceUsed32 << 2) > workspaceSize)
-+		return ERROR(tableLog_tooLarge);
-+	workspace = (U32 *)workspace + spaceUsed32;
-+	workspaceSize -= (spaceUsed32 << 2);
-+
-+	/* check conditions */
-+	if (maxSymbolValue > HUF_SYMBOLVALUE_MAX)
-+		return ERROR(maxSymbolValue_tooLarge);
-+
-+	/* convert to weight */
-+	bitsToWeight[0] = 0;
-+	for (n = 1; n < huffLog + 1; n++)
-+		bitsToWeight[n] = (BYTE)(huffLog + 1 - n);
-+	for (n = 0; n < maxSymbolValue; n++)
-+		huffWeight[n] = bitsToWeight[CTable[n].nbBits];
-+
-+	/* attempt weights compression by FSE */
-+	{
-+		CHECK_V_F(hSize, HUF_compressWeights_wksp(op + 1, maxDstSize - 1, huffWeight, maxSymbolValue, workspace, workspaceSize));
-+		if ((hSize > 1) & (hSize < maxSymbolValue / 2)) { /* FSE compressed */
-+			op[0] = (BYTE)hSize;
-+			return hSize + 1;
-+		}
-+	}
-+
-+	/* write raw values as 4-bits (max : 15) */
-+	if (maxSymbolValue > (256 - 128))
-+		return ERROR(GENERIC); /* should not happen : likely means source cannot be compressed */
-+	if (((maxSymbolValue + 1) / 2) + 1 > maxDstSize)
-+		return ERROR(dstSize_tooSmall); /* not enough space within dst buffer */
-+	op[0] = (BYTE)(128 /*special case*/ + (maxSymbolValue - 1));
-+	huffWeight[maxSymbolValue] = 0; /* to be sure it doesn't cause msan issue in final combination */
-+	for (n = 0; n < maxSymbolValue; n += 2)
-+		op[(n / 2) + 1] = (BYTE)((huffWeight[n] << 4) + huffWeight[n + 1]);
-+	return ((maxSymbolValue + 1) / 2) + 1;
-+}
-+
-+size_t HUF_readCTable_wksp(HUF_CElt *CTable, U32 maxSymbolValue, const void *src, size_t srcSize, void *workspace, size_t workspaceSize)
-+{
-+	U32 *rankVal;
-+	BYTE *huffWeight;
-+	U32 tableLog = 0;
-+	U32 nbSymbols = 0;
-+	size_t readSize;
-+	size_t spaceUsed32 = 0;
-+
-+	rankVal = (U32 *)workspace + spaceUsed32;
-+	spaceUsed32 += HUF_TABLELOG_ABSOLUTEMAX + 1;
-+	huffWeight = (BYTE *)((U32 *)workspace + spaceUsed32);
-+	spaceUsed32 += ALIGN(HUF_SYMBOLVALUE_MAX + 1, sizeof(U32)) >> 2;
-+
-+	if ((spaceUsed32 << 2) > workspaceSize)
-+		return ERROR(tableLog_tooLarge);
-+	workspace = (U32 *)workspace + spaceUsed32;
-+	workspaceSize -= (spaceUsed32 << 2);
-+
-+	/* get symbol weights */
-+	readSize = HUF_readStats_wksp(huffWeight, HUF_SYMBOLVALUE_MAX + 1, rankVal, &nbSymbols, &tableLog, src, srcSize, workspace, workspaceSize);
-+	if (ERR_isError(readSize))
-+		return readSize;
-+
-+	/* check result */
-+	if (tableLog > HUF_TABLELOG_MAX)
-+		return ERROR(tableLog_tooLarge);
-+	if (nbSymbols > maxSymbolValue + 1)
-+		return ERROR(maxSymbolValue_tooSmall);
-+
-+	/* Prepare base value per rank */
-+	{
-+		U32 n, nextRankStart = 0;
-+		for (n = 1; n <= tableLog; n++) {
-+			U32 curr = nextRankStart;
-+			nextRankStart += (rankVal[n] << (n - 1));
-+			rankVal[n] = curr;
-+		}
-+	}
-+
-+	/* fill nbBits */
-+	{
-+		U32 n;
-+		for (n = 0; n < nbSymbols; n++) {
-+			const U32 w = huffWeight[n];
-+			CTable[n].nbBits = (BYTE)(tableLog + 1 - w);
-+		}
-+	}
-+
-+	/* fill val */
-+	{
-+		U16 nbPerRank[HUF_TABLELOG_MAX + 2] = {0}; /* support w=0=>n=tableLog+1 */
-+		U16 valPerRank[HUF_TABLELOG_MAX + 2] = {0};
-+		{
-+			U32 n;
-+			for (n = 0; n < nbSymbols; n++)
-+				nbPerRank[CTable[n].nbBits]++;
-+		}
-+		/* determine stating value per rank */
-+		valPerRank[tableLog + 1] = 0; /* for w==0 */
-+		{
-+			U16 min = 0;
-+			U32 n;
-+			for (n = tableLog; n > 0; n--) { /* start at n=tablelog <-> w=1 */
-+				valPerRank[n] = min;     /* get starting value within each rank */
-+				min += nbPerRank[n];
-+				min >>= 1;
-+			}
-+		}
-+		/* assign value within rank, symbol order */
-+		{
-+			U32 n;
-+			for (n = 0; n <= maxSymbolValue; n++)
-+				CTable[n].val = valPerRank[CTable[n].nbBits]++;
-+		}
-+	}
-+
-+	return readSize;
-+}
-+
-+typedef struct nodeElt_s {
-+	U32 count;
-+	U16 parent;
-+	BYTE byte;
-+	BYTE nbBits;
-+} nodeElt;
-+
-+static U32 HUF_setMaxHeight(nodeElt *huffNode, U32 lastNonNull, U32 maxNbBits)
-+{
-+	const U32 largestBits = huffNode[lastNonNull].nbBits;
-+	if (largestBits <= maxNbBits)
-+		return largestBits; /* early exit : no elt > maxNbBits */
-+
-+	/* there are several too large elements (at least >= 2) */
-+	{
-+		int totalCost = 0;
-+		const U32 baseCost = 1 << (largestBits - maxNbBits);
-+		U32 n = lastNonNull;
-+
-+		while (huffNode[n].nbBits > maxNbBits) {
-+			totalCost += baseCost - (1 << (largestBits - huffNode[n].nbBits));
-+			huffNode[n].nbBits = (BYTE)maxNbBits;
-+			n--;
-+		} /* n stops at huffNode[n].nbBits <= maxNbBits */
-+		while (huffNode[n].nbBits == maxNbBits)
-+			n--; /* n end at index of smallest symbol using < maxNbBits */
-+
-+		/* renorm totalCost */
-+		totalCost >>= (largestBits - maxNbBits); /* note : totalCost is necessarily a multiple of baseCost */
-+
-+		/* repay normalized cost */
-+		{
-+			U32 const noSymbol = 0xF0F0F0F0;
-+			U32 rankLast[HUF_TABLELOG_MAX + 2];
-+			int pos;
-+
-+			/* Get pos of last (smallest) symbol per rank */
-+			memset(rankLast, 0xF0, sizeof(rankLast));
-+			{
-+				U32 currNbBits = maxNbBits;
-+				for (pos = n; pos >= 0; pos--) {
-+					if (huffNode[pos].nbBits >= currNbBits)
-+						continue;
-+					currNbBits = huffNode[pos].nbBits; /* < maxNbBits */
-+					rankLast[maxNbBits - currNbBits] = pos;
-+				}
-+			}
-+
-+			while (totalCost > 0) {
-+				U32 nBitsToDecrease = BIT_highbit32(totalCost) + 1;
-+				for (; nBitsToDecrease > 1; nBitsToDecrease--) {
-+					U32 highPos = rankLast[nBitsToDecrease];
-+					U32 lowPos = rankLast[nBitsToDecrease - 1];
-+					if (highPos == noSymbol)
-+						continue;
-+					if (lowPos == noSymbol)
-+						break;
-+					{
-+						U32 const highTotal = huffNode[highPos].count;
-+						U32 const lowTotal = 2 * huffNode[lowPos].count;
-+						if (highTotal <= lowTotal)
-+							break;
-+					}
-+				}
-+				/* only triggered when no more rank 1 symbol left => find closest one (note : there is necessarily at least one !) */
-+				/* HUF_MAX_TABLELOG test just to please gcc 5+; but it should not be necessary */
-+				while ((nBitsToDecrease <= HUF_TABLELOG_MAX) && (rankLast[nBitsToDecrease] == noSymbol))
-+					nBitsToDecrease++;
-+				totalCost -= 1 << (nBitsToDecrease - 1);
-+				if (rankLast[nBitsToDecrease - 1] == noSymbol)
-+					rankLast[nBitsToDecrease - 1] = rankLast[nBitsToDecrease]; /* this rank is no longer empty */
-+				huffNode[rankLast[nBitsToDecrease]].nbBits++;
-+				if (rankLast[nBitsToDecrease] == 0) /* special case, reached largest symbol */
-+					rankLast[nBitsToDecrease] = noSymbol;
-+				else {
-+					rankLast[nBitsToDecrease]--;
-+					if (huffNode[rankLast[nBitsToDecrease]].nbBits != maxNbBits - nBitsToDecrease)
-+						rankLast[nBitsToDecrease] = noSymbol; /* this rank is now empty */
-+				}
-+			} /* while (totalCost > 0) */
-+
-+			while (totalCost < 0) {		       /* Sometimes, cost correction overshoot */
-+				if (rankLast[1] == noSymbol) { /* special case : no rank 1 symbol (using maxNbBits-1); let's create one from largest rank 0
-+								  (using maxNbBits) */
-+					while (huffNode[n].nbBits == maxNbBits)
-+						n--;
-+					huffNode[n + 1].nbBits--;
-+					rankLast[1] = n + 1;
-+					totalCost++;
-+					continue;
-+				}
-+				huffNode[rankLast[1] + 1].nbBits--;
-+				rankLast[1]++;
-+				totalCost++;
-+			}
-+		}
-+	} /* there are several too large elements (at least >= 2) */
-+
-+	return maxNbBits;
-+}
-+
-+typedef struct {
-+	U32 base;
-+	U32 curr;
-+} rankPos;
-+
-+static void HUF_sort(nodeElt *huffNode, const U32 *count, U32 maxSymbolValue)
-+{
-+	rankPos rank[32];
-+	U32 n;
-+
-+	memset(rank, 0, sizeof(rank));
-+	for (n = 0; n <= maxSymbolValue; n++) {
-+		U32 r = BIT_highbit32(count[n] + 1);
-+		rank[r].base++;
-+	}
-+	for (n = 30; n > 0; n--)
-+		rank[n - 1].base += rank[n].base;
-+	for (n = 0; n < 32; n++)
-+		rank[n].curr = rank[n].base;
-+	for (n = 0; n <= maxSymbolValue; n++) {
-+		U32 const c = count[n];
-+		U32 const r = BIT_highbit32(c + 1) + 1;
-+		U32 pos = rank[r].curr++;
-+		while ((pos > rank[r].base) && (c > huffNode[pos - 1].count))
-+			huffNode[pos] = huffNode[pos - 1], pos--;
-+		huffNode[pos].count = c;
-+		huffNode[pos].byte = (BYTE)n;
-+	}
-+}
-+
-+/** HUF_buildCTable_wksp() :
-+ *  Same as HUF_buildCTable(), but using externally allocated scratch buffer.
-+ *  `workSpace` must be aligned on 4-bytes boundaries, and be at least as large as a table of 1024 unsigned.
-+ */
-+#define STARTNODE (HUF_SYMBOLVALUE_MAX + 1)
-+typedef nodeElt huffNodeTable[2 * HUF_SYMBOLVALUE_MAX + 1 + 1];
-+size_t HUF_buildCTable_wksp(HUF_CElt *tree, const U32 *count, U32 maxSymbolValue, U32 maxNbBits, void *workSpace, size_t wkspSize)
-+{
-+	nodeElt *const huffNode0 = (nodeElt *)workSpace;
-+	nodeElt *const huffNode = huffNode0 + 1;
-+	U32 n, nonNullRank;
-+	int lowS, lowN;
-+	U16 nodeNb = STARTNODE;
-+	U32 nodeRoot;
-+
-+	/* safety checks */
-+	if (wkspSize < sizeof(huffNodeTable))
-+		return ERROR(GENERIC); /* workSpace is not large enough */
-+	if (maxNbBits == 0)
-+		maxNbBits = HUF_TABLELOG_DEFAULT;
-+	if (maxSymbolValue > HUF_SYMBOLVALUE_MAX)
-+		return ERROR(GENERIC);
-+	memset(huffNode0, 0, sizeof(huffNodeTable));
-+
-+	/* sort, decreasing order */
-+	HUF_sort(huffNode, count, maxSymbolValue);
-+
-+	/* init for parents */
-+	nonNullRank = maxSymbolValue;
-+	while (huffNode[nonNullRank].count == 0)
-+		nonNullRank--;
-+	lowS = nonNullRank;
-+	nodeRoot = nodeNb + lowS - 1;
-+	lowN = nodeNb;
-+	huffNode[nodeNb].count = huffNode[lowS].count + huffNode[lowS - 1].count;
-+	huffNode[lowS].parent = huffNode[lowS - 1].parent = nodeNb;
-+	nodeNb++;
-+	lowS -= 2;
-+	for (n = nodeNb; n <= nodeRoot; n++)
-+		huffNode[n].count = (U32)(1U << 30);
-+	huffNode0[0].count = (U32)(1U << 31); /* fake entry, strong barrier */
-+
-+	/* create parents */
-+	while (nodeNb <= nodeRoot) {
-+		U32 n1 = (huffNode[lowS].count < huffNode[lowN].count) ? lowS-- : lowN++;
-+		U32 n2 = (huffNode[lowS].count < huffNode[lowN].count) ? lowS-- : lowN++;
-+		huffNode[nodeNb].count = huffNode[n1].count + huffNode[n2].count;
-+		huffNode[n1].parent = huffNode[n2].parent = nodeNb;
-+		nodeNb++;
-+	}
-+
-+	/* distribute weights (unlimited tree height) */
-+	huffNode[nodeRoot].nbBits = 0;
-+	for (n = nodeRoot - 1; n >= STARTNODE; n--)
-+		huffNode[n].nbBits = huffNode[huffNode[n].parent].nbBits + 1;
-+	for (n = 0; n <= nonNullRank; n++)
-+		huffNode[n].nbBits = huffNode[huffNode[n].parent].nbBits + 1;
-+
-+	/* enforce maxTableLog */
-+	maxNbBits = HUF_setMaxHeight(huffNode, nonNullRank, maxNbBits);
-+
-+	/* fill result into tree (val, nbBits) */
-+	{
-+		U16 nbPerRank[HUF_TABLELOG_MAX + 1] = {0};
-+		U16 valPerRank[HUF_TABLELOG_MAX + 1] = {0};
-+		if (maxNbBits > HUF_TABLELOG_MAX)
-+			return ERROR(GENERIC); /* check fit into table */
-+		for (n = 0; n <= nonNullRank; n++)
-+			nbPerRank[huffNode[n].nbBits]++;
-+		/* determine stating value per rank */
-+		{
-+			U16 min = 0;
-+			for (n = maxNbBits; n > 0; n--) {
-+				valPerRank[n] = min; /* get starting value within each rank */
-+				min += nbPerRank[n];
-+				min >>= 1;
-+			}
-+		}
-+		for (n = 0; n <= maxSymbolValue; n++)
-+			tree[huffNode[n].byte].nbBits = huffNode[n].nbBits; /* push nbBits per symbol, symbol order */
-+		for (n = 0; n <= maxSymbolValue; n++)
-+			tree[n].val = valPerRank[tree[n].nbBits]++; /* assign value within rank, symbol order */
-+	}
-+
-+	return maxNbBits;
-+}
-+
-+static size_t HUF_estimateCompressedSize(HUF_CElt *CTable, const unsigned *count, unsigned maxSymbolValue)
-+{
-+	size_t nbBits = 0;
-+	int s;
-+	for (s = 0; s <= (int)maxSymbolValue; ++s) {
-+		nbBits += CTable[s].nbBits * count[s];
-+	}
-+	return nbBits >> 3;
-+}
-+
-+static int HUF_validateCTable(const HUF_CElt *CTable, const unsigned *count, unsigned maxSymbolValue)
-+{
-+	int bad = 0;
-+	int s;
-+	for (s = 0; s <= (int)maxSymbolValue; ++s) {
-+		bad |= (count[s] != 0) & (CTable[s].nbBits == 0);
-+	}
-+	return !bad;
-+}
-+
-+static void HUF_encodeSymbol(BIT_CStream_t *bitCPtr, U32 symbol, const HUF_CElt *CTable)
-+{
-+	BIT_addBitsFast(bitCPtr, CTable[symbol].val, CTable[symbol].nbBits);
-+}
-+
-+size_t HUF_compressBound(size_t size) { return HUF_COMPRESSBOUND(size); }
-+
-+#define HUF_FLUSHBITS(s)  BIT_flushBits(s)
-+
-+#define HUF_FLUSHBITS_1(stream)                                            \
-+	if (sizeof((stream)->bitContainer) * 8 < HUF_TABLELOG_MAX * 2 + 7) \
-+	HUF_FLUSHBITS(stream)
-+
-+#define HUF_FLUSHBITS_2(stream)                                            \
-+	if (sizeof((stream)->bitContainer) * 8 < HUF_TABLELOG_MAX * 4 + 7) \
-+	HUF_FLUSHBITS(stream)
-+
-+size_t HUF_compress1X_usingCTable(void *dst, size_t dstSize, const void *src, size_t srcSize, const HUF_CElt *CTable)
-+{
-+	const BYTE *ip = (const BYTE *)src;
-+	BYTE *const ostart = (BYTE *)dst;
-+	BYTE *const oend = ostart + dstSize;
-+	BYTE *op = ostart;
-+	size_t n;
-+	BIT_CStream_t bitC;
-+
-+	/* init */
-+	if (dstSize < 8)
-+		return 0; /* not enough space to compress */
-+	{
-+		size_t const initErr = BIT_initCStream(&bitC, op, oend - op);
-+		if (HUF_isError(initErr))
-+			return 0;
-+	}
-+
-+	n = srcSize & ~3; /* join to mod 4 */
-+	switch (srcSize & 3) {
-+	case 3: HUF_encodeSymbol(&bitC, ip[n + 2], CTable); HUF_FLUSHBITS_2(&bitC);
-+	case 2: HUF_encodeSymbol(&bitC, ip[n + 1], CTable); HUF_FLUSHBITS_1(&bitC);
-+	case 1: HUF_encodeSymbol(&bitC, ip[n + 0], CTable); HUF_FLUSHBITS(&bitC);
-+	case 0:
-+	default:;
-+	}
-+
-+	for (; n > 0; n -= 4) { /* note : n&3==0 at this stage */
-+		HUF_encodeSymbol(&bitC, ip[n - 1], CTable);
-+		HUF_FLUSHBITS_1(&bitC);
-+		HUF_encodeSymbol(&bitC, ip[n - 2], CTable);
-+		HUF_FLUSHBITS_2(&bitC);
-+		HUF_encodeSymbol(&bitC, ip[n - 3], CTable);
-+		HUF_FLUSHBITS_1(&bitC);
-+		HUF_encodeSymbol(&bitC, ip[n - 4], CTable);
-+		HUF_FLUSHBITS(&bitC);
-+	}
-+
-+	return BIT_closeCStream(&bitC);
-+}
-+
-+size_t HUF_compress4X_usingCTable(void *dst, size_t dstSize, const void *src, size_t srcSize, const HUF_CElt *CTable)
-+{
-+	size_t const segmentSize = (srcSize + 3) / 4; /* first 3 segments */
-+	const BYTE *ip = (const BYTE *)src;
-+	const BYTE *const iend = ip + srcSize;
-+	BYTE *const ostart = (BYTE *)dst;
-+	BYTE *const oend = ostart + dstSize;
-+	BYTE *op = ostart;
-+
-+	if (dstSize < 6 + 1 + 1 + 1 + 8)
-+		return 0; /* minimum space to compress successfully */
-+	if (srcSize < 12)
-+		return 0; /* no saving possible : too small input */
-+	op += 6;	  /* jumpTable */
-+
-+	{
-+		CHECK_V_F(cSize, HUF_compress1X_usingCTable(op, oend - op, ip, segmentSize, CTable));
-+		if (cSize == 0)
-+			return 0;
-+		ZSTD_writeLE16(ostart, (U16)cSize);
-+		op += cSize;
-+	}
-+
-+	ip += segmentSize;
-+	{
-+		CHECK_V_F(cSize, HUF_compress1X_usingCTable(op, oend - op, ip, segmentSize, CTable));
-+		if (cSize == 0)
-+			return 0;
-+		ZSTD_writeLE16(ostart + 2, (U16)cSize);
-+		op += cSize;
-+	}
-+
-+	ip += segmentSize;
-+	{
-+		CHECK_V_F(cSize, HUF_compress1X_usingCTable(op, oend - op, ip, segmentSize, CTable));
-+		if (cSize == 0)
-+			return 0;
-+		ZSTD_writeLE16(ostart + 4, (U16)cSize);
-+		op += cSize;
-+	}
-+
-+	ip += segmentSize;
-+	{
-+		CHECK_V_F(cSize, HUF_compress1X_usingCTable(op, oend - op, ip, iend - ip, CTable));
-+		if (cSize == 0)
-+			return 0;
-+		op += cSize;
-+	}
-+
-+	return op - ostart;
-+}
-+
-+static size_t HUF_compressCTable_internal(BYTE *const ostart, BYTE *op, BYTE *const oend, const void *src, size_t srcSize, unsigned singleStream,
-+					  const HUF_CElt *CTable)
-+{
-+	size_t const cSize =
-+	    singleStream ? HUF_compress1X_usingCTable(op, oend - op, src, srcSize, CTable) : HUF_compress4X_usingCTable(op, oend - op, src, srcSize, CTable);
-+	if (HUF_isError(cSize)) {
-+		return cSize;
-+	}
-+	if (cSize == 0) {
-+		return 0;
-+	} /* uncompressible */
-+	op += cSize;
-+	/* check compressibility */
-+	if ((size_t)(op - ostart) >= srcSize - 1) {
-+		return 0;
-+	}
-+	return op - ostart;
-+}
-+
-+/* `workSpace` must a table of at least 1024 unsigned */
-+static size_t HUF_compress_internal(void *dst, size_t dstSize, const void *src, size_t srcSize, unsigned maxSymbolValue, unsigned huffLog,
-+				    unsigned singleStream, void *workSpace, size_t wkspSize, HUF_CElt *oldHufTable, HUF_repeat *repeat, int preferRepeat)
-+{
-+	BYTE *const ostart = (BYTE *)dst;
-+	BYTE *const oend = ostart + dstSize;
-+	BYTE *op = ostart;
-+
-+	U32 *count;
-+	size_t const countSize = sizeof(U32) * (HUF_SYMBOLVALUE_MAX + 1);
-+	HUF_CElt *CTable;
-+	size_t const CTableSize = sizeof(HUF_CElt) * (HUF_SYMBOLVALUE_MAX + 1);
-+
-+	/* checks & inits */
-+	if (wkspSize < sizeof(huffNodeTable) + countSize + CTableSize)
-+		return ERROR(GENERIC);
-+	if (!srcSize)
-+		return 0; /* Uncompressed (note : 1 means rle, so first byte must be correct) */
-+	if (!dstSize)
-+		return 0; /* cannot fit within dst budget */
-+	if (srcSize > HUF_BLOCKSIZE_MAX)
-+		return ERROR(srcSize_wrong); /* curr block size limit */
-+	if (huffLog > HUF_TABLELOG_MAX)
-+		return ERROR(tableLog_tooLarge);
-+	if (!maxSymbolValue)
-+		maxSymbolValue = HUF_SYMBOLVALUE_MAX;
-+	if (!huffLog)
-+		huffLog = HUF_TABLELOG_DEFAULT;
-+
-+	count = (U32 *)workSpace;
-+	workSpace = (BYTE *)workSpace + countSize;
-+	wkspSize -= countSize;
-+	CTable = (HUF_CElt *)workSpace;
-+	workSpace = (BYTE *)workSpace + CTableSize;
-+	wkspSize -= CTableSize;
-+
-+	/* Heuristic : If we don't need to check the validity of the old table use the old table for small inputs */
-+	if (preferRepeat && repeat && *repeat == HUF_repeat_valid) {
-+		return HUF_compressCTable_internal(ostart, op, oend, src, srcSize, singleStream, oldHufTable);
-+	}
-+
-+	/* Scan input and build symbol stats */
-+	{
-+		CHECK_V_F(largest, FSE_count_wksp(count, &maxSymbolValue, (const BYTE *)src, srcSize, (U32 *)workSpace));
-+		if (largest == srcSize) {
-+			*ostart = ((const BYTE *)src)[0];
-+			return 1;
-+		} /* single symbol, rle */
-+		if (largest <= (srcSize >> 7) + 1)
-+			return 0; /* Fast heuristic : not compressible enough */
-+	}
-+
-+	/* Check validity of previous table */
-+	if (repeat && *repeat == HUF_repeat_check && !HUF_validateCTable(oldHufTable, count, maxSymbolValue)) {
-+		*repeat = HUF_repeat_none;
-+	}
-+	/* Heuristic : use existing table for small inputs */
-+	if (preferRepeat && repeat && *repeat != HUF_repeat_none) {
-+		return HUF_compressCTable_internal(ostart, op, oend, src, srcSize, singleStream, oldHufTable);
-+	}
-+
-+	/* Build Huffman Tree */
-+	huffLog = HUF_optimalTableLog(huffLog, srcSize, maxSymbolValue);
-+	{
-+		CHECK_V_F(maxBits, HUF_buildCTable_wksp(CTable, count, maxSymbolValue, huffLog, workSpace, wkspSize));
-+		huffLog = (U32)maxBits;
-+		/* Zero the unused symbols so we can check it for validity */
-+		memset(CTable + maxSymbolValue + 1, 0, CTableSize - (maxSymbolValue + 1) * sizeof(HUF_CElt));
-+	}
-+
-+	/* Write table description header */
-+	{
-+		CHECK_V_F(hSize, HUF_writeCTable_wksp(op, dstSize, CTable, maxSymbolValue, huffLog, workSpace, wkspSize));
-+		/* Check if using the previous table will be beneficial */
-+		if (repeat && *repeat != HUF_repeat_none) {
-+			size_t const oldSize = HUF_estimateCompressedSize(oldHufTable, count, maxSymbolValue);
-+			size_t const newSize = HUF_estimateCompressedSize(CTable, count, maxSymbolValue);
-+			if (oldSize <= hSize + newSize || hSize + 12 >= srcSize) {
-+				return HUF_compressCTable_internal(ostart, op, oend, src, srcSize, singleStream, oldHufTable);
-+			}
-+		}
-+		/* Use the new table */
-+		if (hSize + 12ul >= srcSize) {
-+			return 0;
-+		}
-+		op += hSize;
-+		if (repeat) {
-+			*repeat = HUF_repeat_none;
-+		}
-+		if (oldHufTable) {
-+			memcpy(oldHufTable, CTable, CTableSize);
-+		} /* Save the new table */
-+	}
-+	return HUF_compressCTable_internal(ostart, op, oend, src, srcSize, singleStream, CTable);
-+}
-+
-+size_t HUF_compress1X_wksp(void *dst, size_t dstSize, const void *src, size_t srcSize, unsigned maxSymbolValue, unsigned huffLog, void *workSpace,
-+			   size_t wkspSize)
-+{
-+	return HUF_compress_internal(dst, dstSize, src, srcSize, maxSymbolValue, huffLog, 1 /* single stream */, workSpace, wkspSize, NULL, NULL, 0);
-+}
-+
-+size_t HUF_compress1X_repeat(void *dst, size_t dstSize, const void *src, size_t srcSize, unsigned maxSymbolValue, unsigned huffLog, void *workSpace,
-+			     size_t wkspSize, HUF_CElt *hufTable, HUF_repeat *repeat, int preferRepeat)
-+{
-+	return HUF_compress_internal(dst, dstSize, src, srcSize, maxSymbolValue, huffLog, 1 /* single stream */, workSpace, wkspSize, hufTable, repeat,
-+				     preferRepeat);
-+}
-+
-+size_t HUF_compress4X_wksp(void *dst, size_t dstSize, const void *src, size_t srcSize, unsigned maxSymbolValue, unsigned huffLog, void *workSpace,
-+			   size_t wkspSize)
-+{
-+	return HUF_compress_internal(dst, dstSize, src, srcSize, maxSymbolValue, huffLog, 0 /* 4 streams */, workSpace, wkspSize, NULL, NULL, 0);
-+}
-+
-+size_t HUF_compress4X_repeat(void *dst, size_t dstSize, const void *src, size_t srcSize, unsigned maxSymbolValue, unsigned huffLog, void *workSpace,
-+			     size_t wkspSize, HUF_CElt *hufTable, HUF_repeat *repeat, int preferRepeat)
-+{
-+	return HUF_compress_internal(dst, dstSize, src, srcSize, maxSymbolValue, huffLog, 0 /* 4 streams */, workSpace, wkspSize, hufTable, repeat,
-+				     preferRepeat);
-+}
-diff --git a/lib/zstd/huf_decompress.c b/lib/zstd/huf_decompress.c
-new file mode 100644
-index 0000000..6526482
---- /dev/null
-+++ b/lib/zstd/huf_decompress.c
-@@ -0,0 +1,960 @@
-+/*
-+ * Huffman decoder, part of New Generation Entropy library
-+ * Copyright (C) 2013-2016, Yann Collet.
-+ *
-+ * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
-+ *
-+ * Redistribution and use in source and binary forms, with or without
-+ * modification, are permitted provided that the following conditions are
-+ * met:
-+ *
-+ *   * Redistributions of source code must retain the above copyright
-+ * notice, this list of conditions and the following disclaimer.
-+ *   * Redistributions in binary form must reproduce the above
-+ * copyright notice, this list of conditions and the following disclaimer
-+ * in the documentation and/or other materials provided with the
-+ * distribution.
-+ *
-+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
-+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
-+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
-+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
-+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
-+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
-+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
-+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
-+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
-+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
-+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-+ *
-+ * This program is free software; you can redistribute it and/or modify it under
-+ * the terms of the GNU General Public License version 2 as published by the
-+ * Free Software Foundation. This program is dual-licensed; you may select
-+ * either version 2 of the GNU General Public License ("GPL") or BSD license
-+ * ("BSD").
-+ *
-+ * You can contact the author at :
-+ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
-+ */
-+
-+/* **************************************************************
-+*  Compiler specifics
-+****************************************************************/
-+#define FORCE_INLINE static __always_inline
-+
-+/* **************************************************************
-+*  Dependencies
-+****************************************************************/
-+#include "bitstream.h" /* BIT_* */
-+#include "fse.h"       /* header compression */
-+#include "huf.h"
-+#include <linux/compiler.h>
-+#include <linux/kernel.h>
-+#include <linux/string.h> /* memcpy, memset */
-+
-+/* **************************************************************
-+*  Error Management
-+****************************************************************/
-+#define HUF_STATIC_ASSERT(c)                                   \
-+	{                                                      \
-+		enum { HUF_static_assert = 1 / (int)(!!(c)) }; \
-+	} /* use only *after* variable declarations */
-+
-+/*-***************************/
-+/*  generic DTableDesc       */
-+/*-***************************/
-+
-+typedef struct {
-+	BYTE maxTableLog;
-+	BYTE tableType;
-+	BYTE tableLog;
-+	BYTE reserved;
-+} DTableDesc;
-+
-+static DTableDesc HUF_getDTableDesc(const HUF_DTable *table)
-+{
-+	DTableDesc dtd;
-+	memcpy(&dtd, table, sizeof(dtd));
-+	return dtd;
-+}
-+
-+/*-***************************/
-+/*  single-symbol decoding   */
-+/*-***************************/
-+
-+typedef struct {
-+	BYTE byte;
-+	BYTE nbBits;
-+} HUF_DEltX2; /* single-symbol decoding */
-+
-+size_t HUF_readDTableX2_wksp(HUF_DTable *DTable, const void *src, size_t srcSize, void *workspace, size_t workspaceSize)
-+{
-+	U32 tableLog = 0;
-+	U32 nbSymbols = 0;
-+	size_t iSize;
-+	void *const dtPtr = DTable + 1;
-+	HUF_DEltX2 *const dt = (HUF_DEltX2 *)dtPtr;
-+
-+	U32 *rankVal;
-+	BYTE *huffWeight;
-+	size_t spaceUsed32 = 0;
-+
-+	rankVal = (U32 *)workspace + spaceUsed32;
-+	spaceUsed32 += HUF_TABLELOG_ABSOLUTEMAX + 1;
-+	huffWeight = (BYTE *)((U32 *)workspace + spaceUsed32);
-+	spaceUsed32 += ALIGN(HUF_SYMBOLVALUE_MAX + 1, sizeof(U32)) >> 2;
-+
-+	if ((spaceUsed32 << 2) > workspaceSize)
-+		return ERROR(tableLog_tooLarge);
-+	workspace = (U32 *)workspace + spaceUsed32;
-+	workspaceSize -= (spaceUsed32 << 2);
-+
-+	HUF_STATIC_ASSERT(sizeof(DTableDesc) == sizeof(HUF_DTable));
-+	/* memset(huffWeight, 0, sizeof(huffWeight)); */ /* is not necessary, even though some analyzer complain ... */
-+
-+	iSize = HUF_readStats_wksp(huffWeight, HUF_SYMBOLVALUE_MAX + 1, rankVal, &nbSymbols, &tableLog, src, srcSize, workspace, workspaceSize);
-+	if (HUF_isError(iSize))
-+		return iSize;
-+
-+	/* Table header */
-+	{
-+		DTableDesc dtd = HUF_getDTableDesc(DTable);
-+		if (tableLog > (U32)(dtd.maxTableLog + 1))
-+			return ERROR(tableLog_tooLarge); /* DTable too small, Huffman tree cannot fit in */
-+		dtd.tableType = 0;
-+		dtd.tableLog = (BYTE)tableLog;
-+		memcpy(DTable, &dtd, sizeof(dtd));
-+	}
-+
-+	/* Calculate starting value for each rank */
-+	{
-+		U32 n, nextRankStart = 0;
-+		for (n = 1; n < tableLog + 1; n++) {
-+			U32 const curr = nextRankStart;
-+			nextRankStart += (rankVal[n] << (n - 1));
-+			rankVal[n] = curr;
-+		}
-+	}
-+
-+	/* fill DTable */
-+	{
-+		U32 n;
-+		for (n = 0; n < nbSymbols; n++) {
-+			U32 const w = huffWeight[n];
-+			U32 const length = (1 << w) >> 1;
-+			U32 u;
-+			HUF_DEltX2 D;
-+			D.byte = (BYTE)n;
-+			D.nbBits = (BYTE)(tableLog + 1 - w);
-+			for (u = rankVal[w]; u < rankVal[w] + length; u++)
-+				dt[u] = D;
-+			rankVal[w] += length;
-+		}
-+	}
-+
-+	return iSize;
-+}
-+
-+static BYTE HUF_decodeSymbolX2(BIT_DStream_t *Dstream, const HUF_DEltX2 *dt, const U32 dtLog)
-+{
-+	size_t const val = BIT_lookBitsFast(Dstream, dtLog); /* note : dtLog >= 1 */
-+	BYTE const c = dt[val].byte;
-+	BIT_skipBits(Dstream, dt[val].nbBits);
-+	return c;
-+}
-+
-+#define HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr) *ptr++ = HUF_decodeSymbolX2(DStreamPtr, dt, dtLog)
-+
-+#define HUF_DECODE_SYMBOLX2_1(ptr, DStreamPtr)         \
-+	if (ZSTD_64bits() || (HUF_TABLELOG_MAX <= 12)) \
-+	HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr)
-+
-+#define HUF_DECODE_SYMBOLX2_2(ptr, DStreamPtr) \
-+	if (ZSTD_64bits())                     \
-+	HUF_DECODE_SYMBOLX2_0(ptr, DStreamPtr)
-+
-+FORCE_INLINE size_t HUF_decodeStreamX2(BYTE *p, BIT_DStream_t *const bitDPtr, BYTE *const pEnd, const HUF_DEltX2 *const dt, const U32 dtLog)
-+{
-+	BYTE *const pStart = p;
-+
-+	/* up to 4 symbols at a time */
-+	while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) && (p <= pEnd - 4)) {
-+		HUF_DECODE_SYMBOLX2_2(p, bitDPtr);
-+		HUF_DECODE_SYMBOLX2_1(p, bitDPtr);
-+		HUF_DECODE_SYMBOLX2_2(p, bitDPtr);
-+		HUF_DECODE_SYMBOLX2_0(p, bitDPtr);
-+	}
-+
-+	/* closer to the end */
-+	while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) && (p < pEnd))
-+		HUF_DECODE_SYMBOLX2_0(p, bitDPtr);
-+
-+	/* no more data to retrieve from bitstream, hence no need to reload */
-+	while (p < pEnd)
-+		HUF_DECODE_SYMBOLX2_0(p, bitDPtr);
-+
-+	return pEnd - pStart;
-+}
-+
-+static size_t HUF_decompress1X2_usingDTable_internal(void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable)
-+{
-+	BYTE *op = (BYTE *)dst;
-+	BYTE *const oend = op + dstSize;
-+	const void *dtPtr = DTable + 1;
-+	const HUF_DEltX2 *const dt = (const HUF_DEltX2 *)dtPtr;
-+	BIT_DStream_t bitD;
-+	DTableDesc const dtd = HUF_getDTableDesc(DTable);
-+	U32 const dtLog = dtd.tableLog;
-+
-+	{
-+		size_t const errorCode = BIT_initDStream(&bitD, cSrc, cSrcSize);
-+		if (HUF_isError(errorCode))
-+			return errorCode;
-+	}
-+
-+	HUF_decodeStreamX2(op, &bitD, oend, dt, dtLog);
-+
-+	/* check */
-+	if (!BIT_endOfDStream(&bitD))
-+		return ERROR(corruption_detected);
-+
-+	return dstSize;
-+}
-+
-+size_t HUF_decompress1X2_usingDTable(void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable)
-+{
-+	DTableDesc dtd = HUF_getDTableDesc(DTable);
-+	if (dtd.tableType != 0)
-+		return ERROR(GENERIC);
-+	return HUF_decompress1X2_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable);
-+}
-+
-+size_t HUF_decompress1X2_DCtx_wksp(HUF_DTable *DCtx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, size_t workspaceSize)
-+{
-+	const BYTE *ip = (const BYTE *)cSrc;
-+
-+	size_t const hSize = HUF_readDTableX2_wksp(DCtx, cSrc, cSrcSize, workspace, workspaceSize);
-+	if (HUF_isError(hSize))
-+		return hSize;
-+	if (hSize >= cSrcSize)
-+		return ERROR(srcSize_wrong);
-+	ip += hSize;
-+	cSrcSize -= hSize;
-+
-+	return HUF_decompress1X2_usingDTable_internal(dst, dstSize, ip, cSrcSize, DCtx);
-+}
-+
-+static size_t HUF_decompress4X2_usingDTable_internal(void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable)
-+{
-+	/* Check */
-+	if (cSrcSize < 10)
-+		return ERROR(corruption_detected); /* strict minimum : jump table + 1 byte per stream */
-+
-+	{
-+		const BYTE *const istart = (const BYTE *)cSrc;
-+		BYTE *const ostart = (BYTE *)dst;
-+		BYTE *const oend = ostart + dstSize;
-+		const void *const dtPtr = DTable + 1;
-+		const HUF_DEltX2 *const dt = (const HUF_DEltX2 *)dtPtr;
-+
-+		/* Init */
-+		BIT_DStream_t bitD1;
-+		BIT_DStream_t bitD2;
-+		BIT_DStream_t bitD3;
-+		BIT_DStream_t bitD4;
-+		size_t const length1 = ZSTD_readLE16(istart);
-+		size_t const length2 = ZSTD_readLE16(istart + 2);
-+		size_t const length3 = ZSTD_readLE16(istart + 4);
-+		size_t const length4 = cSrcSize - (length1 + length2 + length3 + 6);
-+		const BYTE *const istart1 = istart + 6; /* jumpTable */
-+		const BYTE *const istart2 = istart1 + length1;
-+		const BYTE *const istart3 = istart2 + length2;
-+		const BYTE *const istart4 = istart3 + length3;
-+		const size_t segmentSize = (dstSize + 3) / 4;
-+		BYTE *const opStart2 = ostart + segmentSize;
-+		BYTE *const opStart3 = opStart2 + segmentSize;
-+		BYTE *const opStart4 = opStart3 + segmentSize;
-+		BYTE *op1 = ostart;
-+		BYTE *op2 = opStart2;
-+		BYTE *op3 = opStart3;
-+		BYTE *op4 = opStart4;
-+		U32 endSignal;
-+		DTableDesc const dtd = HUF_getDTableDesc(DTable);
-+		U32 const dtLog = dtd.tableLog;
-+
-+		if (length4 > cSrcSize)
-+			return ERROR(corruption_detected); /* overflow */
-+		{
-+			size_t const errorCode = BIT_initDStream(&bitD1, istart1, length1);
-+			if (HUF_isError(errorCode))
-+				return errorCode;
-+		}
-+		{
-+			size_t const errorCode = BIT_initDStream(&bitD2, istart2, length2);
-+			if (HUF_isError(errorCode))
-+				return errorCode;
-+		}
-+		{
-+			size_t const errorCode = BIT_initDStream(&bitD3, istart3, length3);
-+			if (HUF_isError(errorCode))
-+				return errorCode;
-+		}
-+		{
-+			size_t const errorCode = BIT_initDStream(&bitD4, istart4, length4);
-+			if (HUF_isError(errorCode))
-+				return errorCode;
-+		}
-+
-+		/* 16-32 symbols per loop (4-8 symbols per stream) */
-+		endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4);
-+		for (; (endSignal == BIT_DStream_unfinished) && (op4 < (oend - 7));) {
-+			HUF_DECODE_SYMBOLX2_2(op1, &bitD1);
-+			HUF_DECODE_SYMBOLX2_2(op2, &bitD2);
-+			HUF_DECODE_SYMBOLX2_2(op3, &bitD3);
-+			HUF_DECODE_SYMBOLX2_2(op4, &bitD4);
-+			HUF_DECODE_SYMBOLX2_1(op1, &bitD1);
-+			HUF_DECODE_SYMBOLX2_1(op2, &bitD2);
-+			HUF_DECODE_SYMBOLX2_1(op3, &bitD3);
-+			HUF_DECODE_SYMBOLX2_1(op4, &bitD4);
-+			HUF_DECODE_SYMBOLX2_2(op1, &bitD1);
-+			HUF_DECODE_SYMBOLX2_2(op2, &bitD2);
-+			HUF_DECODE_SYMBOLX2_2(op3, &bitD3);
-+			HUF_DECODE_SYMBOLX2_2(op4, &bitD4);
-+			HUF_DECODE_SYMBOLX2_0(op1, &bitD1);
-+			HUF_DECODE_SYMBOLX2_0(op2, &bitD2);
-+			HUF_DECODE_SYMBOLX2_0(op3, &bitD3);
-+			HUF_DECODE_SYMBOLX2_0(op4, &bitD4);
-+			endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4);
-+		}
-+
-+		/* check corruption */
-+		if (op1 > opStart2)
-+			return ERROR(corruption_detected);
-+		if (op2 > opStart3)
-+			return ERROR(corruption_detected);
-+		if (op3 > opStart4)
-+			return ERROR(corruption_detected);
-+		/* note : op4 supposed already verified within main loop */
-+
-+		/* finish bitStreams one by one */
-+		HUF_decodeStreamX2(op1, &bitD1, opStart2, dt, dtLog);
-+		HUF_decodeStreamX2(op2, &bitD2, opStart3, dt, dtLog);
-+		HUF_decodeStreamX2(op3, &bitD3, opStart4, dt, dtLog);
-+		HUF_decodeStreamX2(op4, &bitD4, oend, dt, dtLog);
-+
-+		/* check */
-+		endSignal = BIT_endOfDStream(&bitD1) & BIT_endOfDStream(&bitD2) & BIT_endOfDStream(&bitD3) & BIT_endOfDStream(&bitD4);
-+		if (!endSignal)
-+			return ERROR(corruption_detected);
-+
-+		/* decoded size */
-+		return dstSize;
-+	}
-+}
-+
-+size_t HUF_decompress4X2_usingDTable(void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable)
-+{
-+	DTableDesc dtd = HUF_getDTableDesc(DTable);
-+	if (dtd.tableType != 0)
-+		return ERROR(GENERIC);
-+	return HUF_decompress4X2_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable);
-+}
-+
-+size_t HUF_decompress4X2_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, size_t workspaceSize)
-+{
-+	const BYTE *ip = (const BYTE *)cSrc;
-+
-+	size_t const hSize = HUF_readDTableX2_wksp(dctx, cSrc, cSrcSize, workspace, workspaceSize);
-+	if (HUF_isError(hSize))
-+		return hSize;
-+	if (hSize >= cSrcSize)
-+		return ERROR(srcSize_wrong);
-+	ip += hSize;
-+	cSrcSize -= hSize;
-+
-+	return HUF_decompress4X2_usingDTable_internal(dst, dstSize, ip, cSrcSize, dctx);
-+}
-+
-+/* *************************/
-+/* double-symbols decoding */
-+/* *************************/
-+typedef struct {
-+	U16 sequence;
-+	BYTE nbBits;
-+	BYTE length;
-+} HUF_DEltX4; /* double-symbols decoding */
-+
-+typedef struct {
-+	BYTE symbol;
-+	BYTE weight;
-+} sortedSymbol_t;
-+
-+/* HUF_fillDTableX4Level2() :
-+ * `rankValOrigin` must be a table of at least (HUF_TABLELOG_MAX + 1) U32 */
-+static void HUF_fillDTableX4Level2(HUF_DEltX4 *DTable, U32 sizeLog, const U32 consumed, const U32 *rankValOrigin, const int minWeight,
-+				   const sortedSymbol_t *sortedSymbols, const U32 sortedListSize, U32 nbBitsBaseline, U16 baseSeq)
-+{
-+	HUF_DEltX4 DElt;
-+	U32 rankVal[HUF_TABLELOG_MAX + 1];
-+
-+	/* get pre-calculated rankVal */
-+	memcpy(rankVal, rankValOrigin, sizeof(rankVal));
-+
-+	/* fill skipped values */
-+	if (minWeight > 1) {
-+		U32 i, skipSize = rankVal[minWeight];
-+		ZSTD_writeLE16(&(DElt.sequence), baseSeq);
-+		DElt.nbBits = (BYTE)(consumed);
-+		DElt.length = 1;
-+		for (i = 0; i < skipSize; i++)
-+			DTable[i] = DElt;
-+	}
-+
-+	/* fill DTable */
-+	{
-+		U32 s;
-+		for (s = 0; s < sortedListSize; s++) { /* note : sortedSymbols already skipped */
-+			const U32 symbol = sortedSymbols[s].symbol;
-+			const U32 weight = sortedSymbols[s].weight;
-+			const U32 nbBits = nbBitsBaseline - weight;
-+			const U32 length = 1 << (sizeLog - nbBits);
-+			const U32 start = rankVal[weight];
-+			U32 i = start;
-+			const U32 end = start + length;
-+
-+			ZSTD_writeLE16(&(DElt.sequence), (U16)(baseSeq + (symbol << 8)));
-+			DElt.nbBits = (BYTE)(nbBits + consumed);
-+			DElt.length = 2;
-+			do {
-+				DTable[i++] = DElt;
-+			} while (i < end); /* since length >= 1 */
-+
-+			rankVal[weight] += length;
-+		}
-+	}
-+}
-+
-+typedef U32 rankVal_t[HUF_TABLELOG_MAX][HUF_TABLELOG_MAX + 1];
-+typedef U32 rankValCol_t[HUF_TABLELOG_MAX + 1];
-+
-+static void HUF_fillDTableX4(HUF_DEltX4 *DTable, const U32 targetLog, const sortedSymbol_t *sortedList, const U32 sortedListSize, const U32 *rankStart,
-+			     rankVal_t rankValOrigin, const U32 maxWeight, const U32 nbBitsBaseline)
-+{
-+	U32 rankVal[HUF_TABLELOG_MAX + 1];
-+	const int scaleLog = nbBitsBaseline - targetLog; /* note : targetLog >= srcLog, hence scaleLog <= 1 */
-+	const U32 minBits = nbBitsBaseline - maxWeight;
-+	U32 s;
-+
-+	memcpy(rankVal, rankValOrigin, sizeof(rankVal));
-+
-+	/* fill DTable */
-+	for (s = 0; s < sortedListSize; s++) {
-+		const U16 symbol = sortedList[s].symbol;
-+		const U32 weight = sortedList[s].weight;
-+		const U32 nbBits = nbBitsBaseline - weight;
-+		const U32 start = rankVal[weight];
-+		const U32 length = 1 << (targetLog - nbBits);
-+
-+		if (targetLog - nbBits >= minBits) { /* enough room for a second symbol */
-+			U32 sortedRank;
-+			int minWeight = nbBits + scaleLog;
-+			if (minWeight < 1)
-+				minWeight = 1;
-+			sortedRank = rankStart[minWeight];
-+			HUF_fillDTableX4Level2(DTable + start, targetLog - nbBits, nbBits, rankValOrigin[nbBits], minWeight, sortedList + sortedRank,
-+					       sortedListSize - sortedRank, nbBitsBaseline, symbol);
-+		} else {
-+			HUF_DEltX4 DElt;
-+			ZSTD_writeLE16(&(DElt.sequence), symbol);
-+			DElt.nbBits = (BYTE)(nbBits);
-+			DElt.length = 1;
-+			{
-+				U32 const end = start + length;
-+				U32 u;
-+				for (u = start; u < end; u++)
-+					DTable[u] = DElt;
-+			}
-+		}
-+		rankVal[weight] += length;
-+	}
-+}
-+
-+size_t HUF_readDTableX4_wksp(HUF_DTable *DTable, const void *src, size_t srcSize, void *workspace, size_t workspaceSize)
-+{
-+	U32 tableLog, maxW, sizeOfSort, nbSymbols;
-+	DTableDesc dtd = HUF_getDTableDesc(DTable);
-+	U32 const maxTableLog = dtd.maxTableLog;
-+	size_t iSize;
-+	void *dtPtr = DTable + 1; /* force compiler to avoid strict-aliasing */
-+	HUF_DEltX4 *const dt = (HUF_DEltX4 *)dtPtr;
-+	U32 *rankStart;
-+
-+	rankValCol_t *rankVal;
-+	U32 *rankStats;
-+	U32 *rankStart0;
-+	sortedSymbol_t *sortedSymbol;
-+	BYTE *weightList;
-+	size_t spaceUsed32 = 0;
-+
-+	HUF_STATIC_ASSERT((sizeof(rankValCol_t) & 3) == 0);
-+
-+	rankVal = (rankValCol_t *)((U32 *)workspace + spaceUsed32);
-+	spaceUsed32 += (sizeof(rankValCol_t) * HUF_TABLELOG_MAX) >> 2;
-+	rankStats = (U32 *)workspace + spaceUsed32;
-+	spaceUsed32 += HUF_TABLELOG_MAX + 1;
-+	rankStart0 = (U32 *)workspace + spaceUsed32;
-+	spaceUsed32 += HUF_TABLELOG_MAX + 2;
-+	sortedSymbol = (sortedSymbol_t *)((U32 *)workspace + spaceUsed32);
-+	spaceUsed32 += ALIGN(sizeof(sortedSymbol_t) * (HUF_SYMBOLVALUE_MAX + 1), sizeof(U32)) >> 2;
-+	weightList = (BYTE *)((U32 *)workspace + spaceUsed32);
-+	spaceUsed32 += ALIGN(HUF_SYMBOLVALUE_MAX + 1, sizeof(U32)) >> 2;
-+
-+	if ((spaceUsed32 << 2) > workspaceSize)
-+		return ERROR(tableLog_tooLarge);
-+	workspace = (U32 *)workspace + spaceUsed32;
-+	workspaceSize -= (spaceUsed32 << 2);
-+
-+	rankStart = rankStart0 + 1;
-+	memset(rankStats, 0, sizeof(U32) * (2 * HUF_TABLELOG_MAX + 2 + 1));
-+
-+	HUF_STATIC_ASSERT(sizeof(HUF_DEltX4) == sizeof(HUF_DTable)); /* if compiler fails here, assertion is wrong */
-+	if (maxTableLog > HUF_TABLELOG_MAX)
-+		return ERROR(tableLog_tooLarge);
-+	/* memset(weightList, 0, sizeof(weightList)); */ /* is not necessary, even though some analyzer complain ... */
-+
-+	iSize = HUF_readStats_wksp(weightList, HUF_SYMBOLVALUE_MAX + 1, rankStats, &nbSymbols, &tableLog, src, srcSize, workspace, workspaceSize);
-+	if (HUF_isError(iSize))
-+		return iSize;
-+
-+	/* check result */
-+	if (tableLog > maxTableLog)
-+		return ERROR(tableLog_tooLarge); /* DTable can't fit code depth */
-+
-+	/* find maxWeight */
-+	for (maxW = tableLog; rankStats[maxW] == 0; maxW--) {
-+	} /* necessarily finds a solution before 0 */
-+
-+	/* Get start index of each weight */
-+	{
-+		U32 w, nextRankStart = 0;
-+		for (w = 1; w < maxW + 1; w++) {
-+			U32 curr = nextRankStart;
-+			nextRankStart += rankStats[w];
-+			rankStart[w] = curr;
-+		}
-+		rankStart[0] = nextRankStart; /* put all 0w symbols at the end of sorted list*/
-+		sizeOfSort = nextRankStart;
-+	}
-+
-+	/* sort symbols by weight */
-+	{
-+		U32 s;
-+		for (s = 0; s < nbSymbols; s++) {
-+			U32 const w = weightList[s];
-+			U32 const r = rankStart[w]++;
-+			sortedSymbol[r].symbol = (BYTE)s;
-+			sortedSymbol[r].weight = (BYTE)w;
-+		}
-+		rankStart[0] = 0; /* forget 0w symbols; this is beginning of weight(1) */
-+	}
-+
-+	/* Build rankVal */
-+	{
-+		U32 *const rankVal0 = rankVal[0];
-+		{
-+			int const rescale = (maxTableLog - tableLog) - 1; /* tableLog <= maxTableLog */
-+			U32 nextRankVal = 0;
-+			U32 w;
-+			for (w = 1; w < maxW + 1; w++) {
-+				U32 curr = nextRankVal;
-+				nextRankVal += rankStats[w] << (w + rescale);
-+				rankVal0[w] = curr;
-+			}
-+		}
-+		{
-+			U32 const minBits = tableLog + 1 - maxW;
-+			U32 consumed;
-+			for (consumed = minBits; consumed < maxTableLog - minBits + 1; consumed++) {
-+				U32 *const rankValPtr = rankVal[consumed];
-+				U32 w;
-+				for (w = 1; w < maxW + 1; w++) {
-+					rankValPtr[w] = rankVal0[w] >> consumed;
-+				}
-+			}
-+		}
-+	}
-+
-+	HUF_fillDTableX4(dt, maxTableLog, sortedSymbol, sizeOfSort, rankStart0, rankVal, maxW, tableLog + 1);
-+
-+	dtd.tableLog = (BYTE)maxTableLog;
-+	dtd.tableType = 1;
-+	memcpy(DTable, &dtd, sizeof(dtd));
-+	return iSize;
-+}
-+
-+static U32 HUF_decodeSymbolX4(void *op, BIT_DStream_t *DStream, const HUF_DEltX4 *dt, const U32 dtLog)
-+{
-+	size_t const val = BIT_lookBitsFast(DStream, dtLog); /* note : dtLog >= 1 */
-+	memcpy(op, dt + val, 2);
-+	BIT_skipBits(DStream, dt[val].nbBits);
-+	return dt[val].length;
-+}
-+
-+static U32 HUF_decodeLastSymbolX4(void *op, BIT_DStream_t *DStream, const HUF_DEltX4 *dt, const U32 dtLog)
-+{
-+	size_t const val = BIT_lookBitsFast(DStream, dtLog); /* note : dtLog >= 1 */
-+	memcpy(op, dt + val, 1);
-+	if (dt[val].length == 1)
-+		BIT_skipBits(DStream, dt[val].nbBits);
-+	else {
-+		if (DStream->bitsConsumed < (sizeof(DStream->bitContainer) * 8)) {
-+			BIT_skipBits(DStream, dt[val].nbBits);
-+			if (DStream->bitsConsumed > (sizeof(DStream->bitContainer) * 8))
-+				/* ugly hack; works only because it's the last symbol. Note : can't easily extract nbBits from just this symbol */
-+				DStream->bitsConsumed = (sizeof(DStream->bitContainer) * 8);
-+		}
-+	}
-+	return 1;
-+}
-+
-+#define HUF_DECODE_SYMBOLX4_0(ptr, DStreamPtr) ptr += HUF_decodeSymbolX4(ptr, DStreamPtr, dt, dtLog)
-+
-+#define HUF_DECODE_SYMBOLX4_1(ptr, DStreamPtr)         \
-+	if (ZSTD_64bits() || (HUF_TABLELOG_MAX <= 12)) \
-+	ptr += HUF_decodeSymbolX4(ptr, DStreamPtr, dt, dtLog)
-+
-+#define HUF_DECODE_SYMBOLX4_2(ptr, DStreamPtr) \
-+	if (ZSTD_64bits())                     \
-+	ptr += HUF_decodeSymbolX4(ptr, DStreamPtr, dt, dtLog)
-+
-+FORCE_INLINE size_t HUF_decodeStreamX4(BYTE *p, BIT_DStream_t *bitDPtr, BYTE *const pEnd, const HUF_DEltX4 *const dt, const U32 dtLog)
-+{
-+	BYTE *const pStart = p;
-+
-+	/* up to 8 symbols at a time */
-+	while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) & (p < pEnd - (sizeof(bitDPtr->bitContainer) - 1))) {
-+		HUF_DECODE_SYMBOLX4_2(p, bitDPtr);
-+		HUF_DECODE_SYMBOLX4_1(p, bitDPtr);
-+		HUF_DECODE_SYMBOLX4_2(p, bitDPtr);
-+		HUF_DECODE_SYMBOLX4_0(p, bitDPtr);
-+	}
-+
-+	/* closer to end : up to 2 symbols at a time */
-+	while ((BIT_reloadDStream(bitDPtr) == BIT_DStream_unfinished) & (p <= pEnd - 2))
-+		HUF_DECODE_SYMBOLX4_0(p, bitDPtr);
-+
-+	while (p <= pEnd - 2)
-+		HUF_DECODE_SYMBOLX4_0(p, bitDPtr); /* no need to reload : reached the end of DStream */
-+
-+	if (p < pEnd)
-+		p += HUF_decodeLastSymbolX4(p, bitDPtr, dt, dtLog);
-+
-+	return p - pStart;
-+}
-+
-+static size_t HUF_decompress1X4_usingDTable_internal(void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable)
-+{
-+	BIT_DStream_t bitD;
-+
-+	/* Init */
-+	{
-+		size_t const errorCode = BIT_initDStream(&bitD, cSrc, cSrcSize);
-+		if (HUF_isError(errorCode))
-+			return errorCode;
-+	}
-+
-+	/* decode */
-+	{
-+		BYTE *const ostart = (BYTE *)dst;
-+		BYTE *const oend = ostart + dstSize;
-+		const void *const dtPtr = DTable + 1; /* force compiler to not use strict-aliasing */
-+		const HUF_DEltX4 *const dt = (const HUF_DEltX4 *)dtPtr;
-+		DTableDesc const dtd = HUF_getDTableDesc(DTable);
-+		HUF_decodeStreamX4(ostart, &bitD, oend, dt, dtd.tableLog);
-+	}
-+
-+	/* check */
-+	if (!BIT_endOfDStream(&bitD))
-+		return ERROR(corruption_detected);
-+
-+	/* decoded size */
-+	return dstSize;
-+}
-+
-+size_t HUF_decompress1X4_usingDTable(void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable)
-+{
-+	DTableDesc dtd = HUF_getDTableDesc(DTable);
-+	if (dtd.tableType != 1)
-+		return ERROR(GENERIC);
-+	return HUF_decompress1X4_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable);
-+}
-+
-+size_t HUF_decompress1X4_DCtx_wksp(HUF_DTable *DCtx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, size_t workspaceSize)
-+{
-+	const BYTE *ip = (const BYTE *)cSrc;
-+
-+	size_t const hSize = HUF_readDTableX4_wksp(DCtx, cSrc, cSrcSize, workspace, workspaceSize);
-+	if (HUF_isError(hSize))
-+		return hSize;
-+	if (hSize >= cSrcSize)
-+		return ERROR(srcSize_wrong);
-+	ip += hSize;
-+	cSrcSize -= hSize;
-+
-+	return HUF_decompress1X4_usingDTable_internal(dst, dstSize, ip, cSrcSize, DCtx);
-+}
-+
-+static size_t HUF_decompress4X4_usingDTable_internal(void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable)
-+{
-+	if (cSrcSize < 10)
-+		return ERROR(corruption_detected); /* strict minimum : jump table + 1 byte per stream */
-+
-+	{
-+		const BYTE *const istart = (const BYTE *)cSrc;
-+		BYTE *const ostart = (BYTE *)dst;
-+		BYTE *const oend = ostart + dstSize;
-+		const void *const dtPtr = DTable + 1;
-+		const HUF_DEltX4 *const dt = (const HUF_DEltX4 *)dtPtr;
-+
-+		/* Init */
-+		BIT_DStream_t bitD1;
-+		BIT_DStream_t bitD2;
-+		BIT_DStream_t bitD3;
-+		BIT_DStream_t bitD4;
-+		size_t const length1 = ZSTD_readLE16(istart);
-+		size_t const length2 = ZSTD_readLE16(istart + 2);
-+		size_t const length3 = ZSTD_readLE16(istart + 4);
-+		size_t const length4 = cSrcSize - (length1 + length2 + length3 + 6);
-+		const BYTE *const istart1 = istart + 6; /* jumpTable */
-+		const BYTE *const istart2 = istart1 + length1;
-+		const BYTE *const istart3 = istart2 + length2;
-+		const BYTE *const istart4 = istart3 + length3;
-+		size_t const segmentSize = (dstSize + 3) / 4;
-+		BYTE *const opStart2 = ostart + segmentSize;
-+		BYTE *const opStart3 = opStart2 + segmentSize;
-+		BYTE *const opStart4 = opStart3 + segmentSize;
-+		BYTE *op1 = ostart;
-+		BYTE *op2 = opStart2;
-+		BYTE *op3 = opStart3;
-+		BYTE *op4 = opStart4;
-+		U32 endSignal;
-+		DTableDesc const dtd = HUF_getDTableDesc(DTable);
-+		U32 const dtLog = dtd.tableLog;
-+
-+		if (length4 > cSrcSize)
-+			return ERROR(corruption_detected); /* overflow */
-+		{
-+			size_t const errorCode = BIT_initDStream(&bitD1, istart1, length1);
-+			if (HUF_isError(errorCode))
-+				return errorCode;
-+		}
-+		{
-+			size_t const errorCode = BIT_initDStream(&bitD2, istart2, length2);
-+			if (HUF_isError(errorCode))
-+				return errorCode;
-+		}
-+		{
-+			size_t const errorCode = BIT_initDStream(&bitD3, istart3, length3);
-+			if (HUF_isError(errorCode))
-+				return errorCode;
-+		}
-+		{
-+			size_t const errorCode = BIT_initDStream(&bitD4, istart4, length4);
-+			if (HUF_isError(errorCode))
-+				return errorCode;
-+		}
-+
-+		/* 16-32 symbols per loop (4-8 symbols per stream) */
-+		endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4);
-+		for (; (endSignal == BIT_DStream_unfinished) & (op4 < (oend - (sizeof(bitD4.bitContainer) - 1)));) {
-+			HUF_DECODE_SYMBOLX4_2(op1, &bitD1);
-+			HUF_DECODE_SYMBOLX4_2(op2, &bitD2);
-+			HUF_DECODE_SYMBOLX4_2(op3, &bitD3);
-+			HUF_DECODE_SYMBOLX4_2(op4, &bitD4);
-+			HUF_DECODE_SYMBOLX4_1(op1, &bitD1);
-+			HUF_DECODE_SYMBOLX4_1(op2, &bitD2);
-+			HUF_DECODE_SYMBOLX4_1(op3, &bitD3);
-+			HUF_DECODE_SYMBOLX4_1(op4, &bitD4);
-+			HUF_DECODE_SYMBOLX4_2(op1, &bitD1);
-+			HUF_DECODE_SYMBOLX4_2(op2, &bitD2);
-+			HUF_DECODE_SYMBOLX4_2(op3, &bitD3);
-+			HUF_DECODE_SYMBOLX4_2(op4, &bitD4);
-+			HUF_DECODE_SYMBOLX4_0(op1, &bitD1);
-+			HUF_DECODE_SYMBOLX4_0(op2, &bitD2);
-+			HUF_DECODE_SYMBOLX4_0(op3, &bitD3);
-+			HUF_DECODE_SYMBOLX4_0(op4, &bitD4);
-+
-+			endSignal = BIT_reloadDStream(&bitD1) | BIT_reloadDStream(&bitD2) | BIT_reloadDStream(&bitD3) | BIT_reloadDStream(&bitD4);
-+		}
-+
-+		/* check corruption */
-+		if (op1 > opStart2)
-+			return ERROR(corruption_detected);
-+		if (op2 > opStart3)
-+			return ERROR(corruption_detected);
-+		if (op3 > opStart4)
-+			return ERROR(corruption_detected);
-+		/* note : op4 already verified within main loop */
-+
-+		/* finish bitStreams one by one */
-+		HUF_decodeStreamX4(op1, &bitD1, opStart2, dt, dtLog);
-+		HUF_decodeStreamX4(op2, &bitD2, opStart3, dt, dtLog);
-+		HUF_decodeStreamX4(op3, &bitD3, opStart4, dt, dtLog);
-+		HUF_decodeStreamX4(op4, &bitD4, oend, dt, dtLog);
-+
-+		/* check */
-+		{
-+			U32 const endCheck = BIT_endOfDStream(&bitD1) & BIT_endOfDStream(&bitD2) & BIT_endOfDStream(&bitD3) & BIT_endOfDStream(&bitD4);
-+			if (!endCheck)
-+				return ERROR(corruption_detected);
-+		}
-+
-+		/* decoded size */
-+		return dstSize;
-+	}
-+}
-+
-+size_t HUF_decompress4X4_usingDTable(void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable)
-+{
-+	DTableDesc dtd = HUF_getDTableDesc(DTable);
-+	if (dtd.tableType != 1)
-+		return ERROR(GENERIC);
-+	return HUF_decompress4X4_usingDTable_internal(dst, dstSize, cSrc, cSrcSize, DTable);
-+}
-+
-+size_t HUF_decompress4X4_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, size_t workspaceSize)
-+{
-+	const BYTE *ip = (const BYTE *)cSrc;
-+
-+	size_t hSize = HUF_readDTableX4_wksp(dctx, cSrc, cSrcSize, workspace, workspaceSize);
-+	if (HUF_isError(hSize))
-+		return hSize;
-+	if (hSize >= cSrcSize)
-+		return ERROR(srcSize_wrong);
-+	ip += hSize;
-+	cSrcSize -= hSize;
-+
-+	return HUF_decompress4X4_usingDTable_internal(dst, dstSize, ip, cSrcSize, dctx);
-+}
-+
-+/* ********************************/
-+/* Generic decompression selector */
-+/* ********************************/
-+
-+size_t HUF_decompress1X_usingDTable(void *dst, size_t maxDstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable)
-+{
-+	DTableDesc const dtd = HUF_getDTableDesc(DTable);
-+	return dtd.tableType ? HUF_decompress1X4_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable)
-+			     : HUF_decompress1X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable);
-+}
-+
-+size_t HUF_decompress4X_usingDTable(void *dst, size_t maxDstSize, const void *cSrc, size_t cSrcSize, const HUF_DTable *DTable)
-+{
-+	DTableDesc const dtd = HUF_getDTableDesc(DTable);
-+	return dtd.tableType ? HUF_decompress4X4_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable)
-+			     : HUF_decompress4X2_usingDTable_internal(dst, maxDstSize, cSrc, cSrcSize, DTable);
-+}
-+
-+typedef struct {
-+	U32 tableTime;
-+	U32 decode256Time;
-+} algo_time_t;
-+static const algo_time_t algoTime[16 /* Quantization */][3 /* single, double, quad */] = {
-+    /* single, double, quad */
-+    {{0, 0}, {1, 1}, {2, 2}},		     /* Q==0 : impossible */
-+    {{0, 0}, {1, 1}, {2, 2}},		     /* Q==1 : impossible */
-+    {{38, 130}, {1313, 74}, {2151, 38}},     /* Q == 2 : 12-18% */
-+    {{448, 128}, {1353, 74}, {2238, 41}},    /* Q == 3 : 18-25% */
-+    {{556, 128}, {1353, 74}, {2238, 47}},    /* Q == 4 : 25-32% */
-+    {{714, 128}, {1418, 74}, {2436, 53}},    /* Q == 5 : 32-38% */
-+    {{883, 128}, {1437, 74}, {2464, 61}},    /* Q == 6 : 38-44% */
-+    {{897, 128}, {1515, 75}, {2622, 68}},    /* Q == 7 : 44-50% */
-+    {{926, 128}, {1613, 75}, {2730, 75}},    /* Q == 8 : 50-56% */
-+    {{947, 128}, {1729, 77}, {3359, 77}},    /* Q == 9 : 56-62% */
-+    {{1107, 128}, {2083, 81}, {4006, 84}},   /* Q ==10 : 62-69% */
-+    {{1177, 128}, {2379, 87}, {4785, 88}},   /* Q ==11 : 69-75% */
-+    {{1242, 128}, {2415, 93}, {5155, 84}},   /* Q ==12 : 75-81% */
-+    {{1349, 128}, {2644, 106}, {5260, 106}}, /* Q ==13 : 81-87% */
-+    {{1455, 128}, {2422, 124}, {4174, 124}}, /* Q ==14 : 87-93% */
-+    {{722, 128}, {1891, 145}, {1936, 146}},  /* Q ==15 : 93-99% */
-+};
-+
-+/** HUF_selectDecoder() :
-+*   Tells which decoder is likely to decode faster,
-+*   based on a set of pre-determined metrics.
-+*   @return : 0==HUF_decompress4X2, 1==HUF_decompress4X4 .
-+*   Assumption : 0 < cSrcSize < dstSize <= 128 KB */
-+U32 HUF_selectDecoder(size_t dstSize, size_t cSrcSize)
-+{
-+	/* decoder timing evaluation */
-+	U32 const Q = (U32)(cSrcSize * 16 / dstSize); /* Q < 16 since dstSize > cSrcSize */
-+	U32 const D256 = (U32)(dstSize >> 8);
-+	U32 const DTime0 = algoTime[Q][0].tableTime + (algoTime[Q][0].decode256Time * D256);
-+	U32 DTime1 = algoTime[Q][1].tableTime + (algoTime[Q][1].decode256Time * D256);
-+	DTime1 += DTime1 >> 3; /* advantage to algorithm using less memory, for cache eviction */
-+
-+	return DTime1 < DTime0;
-+}
-+
-+typedef size_t (*decompressionAlgo)(void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize);
-+
-+size_t HUF_decompress4X_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, size_t workspaceSize)
-+{
-+	/* validation checks */
-+	if (dstSize == 0)
-+		return ERROR(dstSize_tooSmall);
-+	if (cSrcSize > dstSize)
-+		return ERROR(corruption_detected); /* invalid */
-+	if (cSrcSize == dstSize) {
-+		memcpy(dst, cSrc, dstSize);
-+		return dstSize;
-+	} /* not compressed */
-+	if (cSrcSize == 1) {
-+		memset(dst, *(const BYTE *)cSrc, dstSize);
-+		return dstSize;
-+	} /* RLE */
-+
-+	{
-+		U32 const algoNb = HUF_selectDecoder(dstSize, cSrcSize);
-+		return algoNb ? HUF_decompress4X4_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize, workspace, workspaceSize)
-+			      : HUF_decompress4X2_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize, workspace, workspaceSize);
-+	}
-+}
-+
-+size_t HUF_decompress4X_hufOnly_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, size_t workspaceSize)
-+{
-+	/* validation checks */
-+	if (dstSize == 0)
-+		return ERROR(dstSize_tooSmall);
-+	if ((cSrcSize >= dstSize) || (cSrcSize <= 1))
-+		return ERROR(corruption_detected); /* invalid */
-+
-+	{
-+		U32 const algoNb = HUF_selectDecoder(dstSize, cSrcSize);
-+		return algoNb ? HUF_decompress4X4_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize, workspace, workspaceSize)
-+			      : HUF_decompress4X2_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize, workspace, workspaceSize);
-+	}
-+}
-+
-+size_t HUF_decompress1X_DCtx_wksp(HUF_DTable *dctx, void *dst, size_t dstSize, const void *cSrc, size_t cSrcSize, void *workspace, size_t workspaceSize)
-+{
-+	/* validation checks */
-+	if (dstSize == 0)
-+		return ERROR(dstSize_tooSmall);
-+	if (cSrcSize > dstSize)
-+		return ERROR(corruption_detected); /* invalid */
-+	if (cSrcSize == dstSize) {
-+		memcpy(dst, cSrc, dstSize);
-+		return dstSize;
-+	} /* not compressed */
-+	if (cSrcSize == 1) {
-+		memset(dst, *(const BYTE *)cSrc, dstSize);
-+		return dstSize;
-+	} /* RLE */
-+
-+	{
-+		U32 const algoNb = HUF_selectDecoder(dstSize, cSrcSize);
-+		return algoNb ? HUF_decompress1X4_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize, workspace, workspaceSize)
-+			      : HUF_decompress1X2_DCtx_wksp(dctx, dst, dstSize, cSrc, cSrcSize, workspace, workspaceSize);
-+	}
-+}
-diff --git a/lib/zstd/mem.h b/lib/zstd/mem.h
-new file mode 100644
-index 0000000..3a0f34c
---- /dev/null
-+++ b/lib/zstd/mem.h
-@@ -0,0 +1,151 @@
-+/**
-+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
-+ * All rights reserved.
-+ *
-+ * This source code is licensed under the BSD-style license found in the
-+ * LICENSE file in the root directory of https://github.com/facebook/zstd.
-+ * An additional grant of patent rights can be found in the PATENTS file in the
-+ * same directory.
-+ *
-+ * This program is free software; you can redistribute it and/or modify it under
-+ * the terms of the GNU General Public License version 2 as published by the
-+ * Free Software Foundation. This program is dual-licensed; you may select
-+ * either version 2 of the GNU General Public License ("GPL") or BSD license
-+ * ("BSD").
-+ */
-+
-+#ifndef MEM_H_MODULE
-+#define MEM_H_MODULE
-+
-+/*-****************************************
-+*  Dependencies
-+******************************************/
-+#include <asm/unaligned.h>
-+#include <linux/string.h> /* memcpy */
-+#include <linux/types.h>  /* size_t, ptrdiff_t */
-+
-+/*-****************************************
-+*  Compiler specifics
-+******************************************/
-+#define ZSTD_STATIC static __inline __attribute__((unused))
-+
-+/*-**************************************************************
-+*  Basic Types
-+*****************************************************************/
-+typedef uint8_t BYTE;
-+typedef uint16_t U16;
-+typedef int16_t S16;
-+typedef uint32_t U32;
-+typedef int32_t S32;
-+typedef uint64_t U64;
-+typedef int64_t S64;
-+typedef ptrdiff_t iPtrDiff;
-+typedef uintptr_t uPtrDiff;
-+
-+/*-**************************************************************
-+*  Memory I/O
-+*****************************************************************/
-+ZSTD_STATIC unsigned ZSTD_32bits(void) { return sizeof(size_t) == 4; }
-+ZSTD_STATIC unsigned ZSTD_64bits(void) { return sizeof(size_t) == 8; }
-+
-+#if defined(__LITTLE_ENDIAN)
-+#define ZSTD_LITTLE_ENDIAN 1
-+#else
-+#define ZSTD_LITTLE_ENDIAN 0
-+#endif
-+
-+ZSTD_STATIC unsigned ZSTD_isLittleEndian(void) { return ZSTD_LITTLE_ENDIAN; }
-+
-+ZSTD_STATIC U16 ZSTD_read16(const void *memPtr) { return get_unaligned((const U16 *)memPtr); }
-+
-+ZSTD_STATIC U32 ZSTD_read32(const void *memPtr) { return get_unaligned((const U32 *)memPtr); }
-+
-+ZSTD_STATIC U64 ZSTD_read64(const void *memPtr) { return get_unaligned((const U64 *)memPtr); }
-+
-+ZSTD_STATIC size_t ZSTD_readST(const void *memPtr) { return get_unaligned((const size_t *)memPtr); }
-+
-+ZSTD_STATIC void ZSTD_write16(void *memPtr, U16 value) { put_unaligned(value, (U16 *)memPtr); }
-+
-+ZSTD_STATIC void ZSTD_write32(void *memPtr, U32 value) { put_unaligned(value, (U32 *)memPtr); }
-+
-+ZSTD_STATIC void ZSTD_write64(void *memPtr, U64 value) { put_unaligned(value, (U64 *)memPtr); }
-+
-+/*=== Little endian r/w ===*/
-+
-+ZSTD_STATIC U16 ZSTD_readLE16(const void *memPtr) { return get_unaligned_le16(memPtr); }
-+
-+ZSTD_STATIC void ZSTD_writeLE16(void *memPtr, U16 val) { put_unaligned_le16(val, memPtr); }
-+
-+ZSTD_STATIC U32 ZSTD_readLE24(const void *memPtr) { return ZSTD_readLE16(memPtr) + (((const BYTE *)memPtr)[2] << 16); }
-+
-+ZSTD_STATIC void ZSTD_writeLE24(void *memPtr, U32 val)
-+{
-+	ZSTD_writeLE16(memPtr, (U16)val);
-+	((BYTE *)memPtr)[2] = (BYTE)(val >> 16);
-+}
-+
-+ZSTD_STATIC U32 ZSTD_readLE32(const void *memPtr) { return get_unaligned_le32(memPtr); }
-+
-+ZSTD_STATIC void ZSTD_writeLE32(void *memPtr, U32 val32) { put_unaligned_le32(val32, memPtr); }
-+
-+ZSTD_STATIC U64 ZSTD_readLE64(const void *memPtr) { return get_unaligned_le64(memPtr); }
-+
-+ZSTD_STATIC void ZSTD_writeLE64(void *memPtr, U64 val64) { put_unaligned_le64(val64, memPtr); }
-+
-+ZSTD_STATIC size_t ZSTD_readLEST(const void *memPtr)
-+{
-+	if (ZSTD_32bits())
-+		return (size_t)ZSTD_readLE32(memPtr);
-+	else
-+		return (size_t)ZSTD_readLE64(memPtr);
-+}
-+
-+ZSTD_STATIC void ZSTD_writeLEST(void *memPtr, size_t val)
-+{
-+	if (ZSTD_32bits())
-+		ZSTD_writeLE32(memPtr, (U32)val);
-+	else
-+		ZSTD_writeLE64(memPtr, (U64)val);
-+}
-+
-+/*=== Big endian r/w ===*/
-+
-+ZSTD_STATIC U32 ZSTD_readBE32(const void *memPtr) { return get_unaligned_be32(memPtr); }
-+
-+ZSTD_STATIC void ZSTD_writeBE32(void *memPtr, U32 val32) { put_unaligned_be32(val32, memPtr); }
-+
-+ZSTD_STATIC U64 ZSTD_readBE64(const void *memPtr) { return get_unaligned_be64(memPtr); }
-+
-+ZSTD_STATIC void ZSTD_writeBE64(void *memPtr, U64 val64) { put_unaligned_be64(val64, memPtr); }
-+
-+ZSTD_STATIC size_t ZSTD_readBEST(const void *memPtr)
-+{
-+	if (ZSTD_32bits())
-+		return (size_t)ZSTD_readBE32(memPtr);
-+	else
-+		return (size_t)ZSTD_readBE64(memPtr);
-+}
-+
-+ZSTD_STATIC void ZSTD_writeBEST(void *memPtr, size_t val)
-+{
-+	if (ZSTD_32bits())
-+		ZSTD_writeBE32(memPtr, (U32)val);
-+	else
-+		ZSTD_writeBE64(memPtr, (U64)val);
-+}
-+
-+/* function safe only for comparisons */
-+ZSTD_STATIC U32 ZSTD_readMINMATCH(const void *memPtr, U32 length)
-+{
-+	switch (length) {
-+	default:
-+	case 4: return ZSTD_read32(memPtr);
-+	case 3:
-+		if (ZSTD_isLittleEndian())
-+			return ZSTD_read32(memPtr) << 8;
-+		else
-+			return ZSTD_read32(memPtr) >> 8;
-+	}
-+}
-+
-+#endif /* MEM_H_MODULE */
-diff --git a/lib/zstd/zstd_common.c b/lib/zstd/zstd_common.c
-new file mode 100644
-index 0000000..a282624
---- /dev/null
-+++ b/lib/zstd/zstd_common.c
-@@ -0,0 +1,75 @@
-+/**
-+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
-+ * All rights reserved.
-+ *
-+ * This source code is licensed under the BSD-style license found in the
-+ * LICENSE file in the root directory of https://github.com/facebook/zstd.
-+ * An additional grant of patent rights can be found in the PATENTS file in the
-+ * same directory.
-+ *
-+ * This program is free software; you can redistribute it and/or modify it under
-+ * the terms of the GNU General Public License version 2 as published by the
-+ * Free Software Foundation. This program is dual-licensed; you may select
-+ * either version 2 of the GNU General Public License ("GPL") or BSD license
-+ * ("BSD").
-+ */
-+
-+/*-*************************************
-+*  Dependencies
-+***************************************/
-+#include "error_private.h"
-+#include "zstd_internal.h" /* declaration of ZSTD_isError, ZSTD_getErrorName, ZSTD_getErrorCode, ZSTD_getErrorString, ZSTD_versionNumber */
-+#include <linux/kernel.h>
-+
-+/*=**************************************************************
-+*  Custom allocator
-+****************************************************************/
-+
-+#define stack_push(stack, size)                                 \
-+	({                                                      \
-+		void *const ptr = ZSTD_PTR_ALIGN((stack)->ptr); \
-+		(stack)->ptr = (char *)ptr + (size);            \
-+		(stack)->ptr <= (stack)->end ? ptr : NULL;      \
-+	})
-+
-+ZSTD_customMem ZSTD_initStack(void *workspace, size_t workspaceSize)
-+{
-+	ZSTD_customMem stackMem = {ZSTD_stackAlloc, ZSTD_stackFree, workspace};
-+	ZSTD_stack *stack = (ZSTD_stack *)workspace;
-+	/* Verify preconditions */
-+	if (!workspace || workspaceSize < sizeof(ZSTD_stack) || workspace != ZSTD_PTR_ALIGN(workspace)) {
-+		ZSTD_customMem error = {NULL, NULL, NULL};
-+		return error;
-+	}
-+	/* Initialize the stack */
-+	stack->ptr = workspace;
-+	stack->end = (char *)workspace + workspaceSize;
-+	stack_push(stack, sizeof(ZSTD_stack));
-+	return stackMem;
-+}
-+
-+void *ZSTD_stackAllocAll(void *opaque, size_t *size)
-+{
-+	ZSTD_stack *stack = (ZSTD_stack *)opaque;
-+	*size = (BYTE const *)stack->end - (BYTE *)ZSTD_PTR_ALIGN(stack->ptr);
-+	return stack_push(stack, *size);
-+}
-+
-+void *ZSTD_stackAlloc(void *opaque, size_t size)
-+{
-+	ZSTD_stack *stack = (ZSTD_stack *)opaque;
-+	return stack_push(stack, size);
-+}
-+void ZSTD_stackFree(void *opaque, void *address)
-+{
-+	(void)opaque;
-+	(void)address;
-+}
-+
-+void *ZSTD_malloc(size_t size, ZSTD_customMem customMem) { return customMem.customAlloc(customMem.opaque, size); }
-+
-+void ZSTD_free(void *ptr, ZSTD_customMem customMem)
-+{
-+	if (ptr != NULL)
-+		customMem.customFree(customMem.opaque, ptr);
-+}
-diff --git a/lib/zstd/zstd_internal.h b/lib/zstd/zstd_internal.h
-new file mode 100644
-index 0000000..1a79fab
---- /dev/null
-+++ b/lib/zstd/zstd_internal.h
-@@ -0,0 +1,263 @@
-+/**
-+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
-+ * All rights reserved.
-+ *
-+ * This source code is licensed under the BSD-style license found in the
-+ * LICENSE file in the root directory of https://github.com/facebook/zstd.
-+ * An additional grant of patent rights can be found in the PATENTS file in the
-+ * same directory.
-+ *
-+ * This program is free software; you can redistribute it and/or modify it under
-+ * the terms of the GNU General Public License version 2 as published by the
-+ * Free Software Foundation. This program is dual-licensed; you may select
-+ * either version 2 of the GNU General Public License ("GPL") or BSD license
-+ * ("BSD").
-+ */
-+
-+#ifndef ZSTD_CCOMMON_H_MODULE
-+#define ZSTD_CCOMMON_H_MODULE
-+
-+/*-*******************************************************
-+*  Compiler specifics
-+*********************************************************/
-+#define FORCE_INLINE static __always_inline
-+#define FORCE_NOINLINE static noinline
-+
-+/*-*************************************
-+*  Dependencies
-+***************************************/
-+#include "error_private.h"
-+#include "mem.h"
-+#include <linux/compiler.h>
-+#include <linux/kernel.h>
-+#include <linux/xxhash.h>
-+#include <linux/zstd.h>
-+
-+/*-*************************************
-+*  shared macros
-+***************************************/
-+#define MIN(a, b) ((a) < (b) ? (a) : (b))
-+#define MAX(a, b) ((a) > (b) ? (a) : (b))
-+#define CHECK_F(f)                       \
-+	{                                \
-+		size_t const errcod = f; \
-+		if (ERR_isError(errcod)) \
-+			return errcod;   \
-+	} /* check and Forward error code */
-+#define CHECK_E(f, e)                    \
-+	{                                \
-+		size_t const errcod = f; \
-+		if (ERR_isError(errcod)) \
-+			return ERROR(e); \
-+	} /* check and send Error code */
-+#define ZSTD_STATIC_ASSERT(c)                                   \
-+	{                                                       \
-+		enum { ZSTD_static_assert = 1 / (int)(!!(c)) }; \
-+	}
-+
-+/*-*************************************
-+*  Common constants
-+***************************************/
-+#define ZSTD_OPT_NUM (1 << 12)
-+#define ZSTD_DICT_MAGIC 0xEC30A437 /* v0.7+ */
-+
-+#define ZSTD_REP_NUM 3		      /* number of repcodes */
-+#define ZSTD_REP_CHECK (ZSTD_REP_NUM) /* number of repcodes to check by the optimal parser */
-+#define ZSTD_REP_MOVE (ZSTD_REP_NUM - 1)
-+#define ZSTD_REP_MOVE_OPT (ZSTD_REP_NUM)
-+static const U32 repStartValue[ZSTD_REP_NUM] = {1, 4, 8};
-+
-+#define KB *(1 << 10)
-+#define MB *(1 << 20)
-+#define GB *(1U << 30)
-+
-+#define BIT7 128
-+#define BIT6 64
-+#define BIT5 32
-+#define BIT4 16
-+#define BIT1 2
-+#define BIT0 1
-+
-+#define ZSTD_WINDOWLOG_ABSOLUTEMIN 10
-+static const size_t ZSTD_fcs_fieldSize[4] = {0, 2, 4, 8};
-+static const size_t ZSTD_did_fieldSize[4] = {0, 1, 2, 4};
-+
-+#define ZSTD_BLOCKHEADERSIZE 3 /* C standard doesn't allow `static const` variable to be init using another `static const` variable */
-+static const size_t ZSTD_blockHeaderSize = ZSTD_BLOCKHEADERSIZE;
-+typedef enum { bt_raw, bt_rle, bt_compressed, bt_reserved } blockType_e;
-+
-+#define MIN_SEQUENCES_SIZE 1									  /* nbSeq==0 */
-+#define MIN_CBLOCK_SIZE (1 /*litCSize*/ + 1 /* RLE or RAW */ + MIN_SEQUENCES_SIZE /* nbSeq==0 */) /* for a non-null block */
-+
-+#define HufLog 12
-+typedef enum { set_basic, set_rle, set_compressed, set_repeat } symbolEncodingType_e;
-+
-+#define LONGNBSEQ 0x7F00
-+
-+#define MINMATCH 3
-+#define EQUAL_READ32 4
-+
-+#define Litbits 8
-+#define MaxLit ((1 << Litbits) - 1)
-+#define MaxML 52
-+#define MaxLL 35
-+#define MaxOff 28
-+#define MaxSeq MAX(MaxLL, MaxML) /* Assumption : MaxOff < MaxLL,MaxML */
-+#define MLFSELog 9
-+#define LLFSELog 9
-+#define OffFSELog 8
-+
-+static const U32 LL_bits[MaxLL + 1] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 3, 3, 4, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16};
-+static const S16 LL_defaultNorm[MaxLL + 1] = {4, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 2, 1, 1, 1, 1, 1, -1, -1, -1, -1};
-+#define LL_DEFAULTNORMLOG 6 /* for static allocation */
-+static const U32 LL_defaultNormLog = LL_DEFAULTNORMLOG;
-+
-+static const U32 ML_bits[MaxML + 1] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,  0,  0,  0,  0,  0,  0, 0,
-+				       0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 3, 3, 4, 4, 5, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16};
-+static const S16 ML_defaultNorm[MaxML + 1] = {1, 4, 3, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  1,  1,  1,  1,  1,  1, 1,
-+					      1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1};
-+#define ML_DEFAULTNORMLOG 6 /* for static allocation */
-+static const U32 ML_defaultNormLog = ML_DEFAULTNORMLOG;
-+
-+static const S16 OF_defaultNorm[MaxOff + 1] = {1, 1, 1, 1, 1, 1, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1};
-+#define OF_DEFAULTNORMLOG 5 /* for static allocation */
-+static const U32 OF_defaultNormLog = OF_DEFAULTNORMLOG;
-+
-+/*-*******************************************
-+*  Shared functions to include for inlining
-+*********************************************/
-+ZSTD_STATIC void ZSTD_copy8(void *dst, const void *src) {
-+	memcpy(dst, src, 8);
-+}
-+/*! ZSTD_wildcopy() :
-+*   custom version of memcpy(), can copy up to 7 bytes too many (8 bytes if length==0) */
-+#define WILDCOPY_OVERLENGTH 8
-+ZSTD_STATIC void ZSTD_wildcopy(void *dst, const void *src, ptrdiff_t length)
-+{
-+	const BYTE* ip = (const BYTE*)src;
-+	BYTE* op = (BYTE*)dst;
-+	BYTE* const oend = op + length;
-+	/* Work around https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81388.
-+	 * Avoid the bad case where the loop only runs once by handling the
-+	 * special case separately. This doesn't trigger the bug because it
-+	 * doesn't involve pointer/integer overflow.
-+	 */
-+	if (length <= 8)
-+		return ZSTD_copy8(dst, src);
-+	do {
-+		ZSTD_copy8(op, ip);
-+		op += 8;
-+		ip += 8;
-+	} while (op < oend);
-+}
-+
-+/*-*******************************************
-+*  Private interfaces
-+*********************************************/
-+typedef struct ZSTD_stats_s ZSTD_stats_t;
-+
-+typedef struct {
-+	U32 off;
-+	U32 len;
-+} ZSTD_match_t;
-+
-+typedef struct {
-+	U32 price;
-+	U32 off;
-+	U32 mlen;
-+	U32 litlen;
-+	U32 rep[ZSTD_REP_NUM];
-+} ZSTD_optimal_t;
-+
-+typedef struct seqDef_s {
-+	U32 offset;
-+	U16 litLength;
-+	U16 matchLength;
-+} seqDef;
-+
-+typedef struct {
-+	seqDef *sequencesStart;
-+	seqDef *sequences;
-+	BYTE *litStart;
-+	BYTE *lit;
-+	BYTE *llCode;
-+	BYTE *mlCode;
-+	BYTE *ofCode;
-+	U32 longLengthID; /* 0 == no longLength; 1 == Lit.longLength; 2 == Match.longLength; */
-+	U32 longLengthPos;
-+	/* opt */
-+	ZSTD_optimal_t *priceTable;
-+	ZSTD_match_t *matchTable;
-+	U32 *matchLengthFreq;
-+	U32 *litLengthFreq;
-+	U32 *litFreq;
-+	U32 *offCodeFreq;
-+	U32 matchLengthSum;
-+	U32 matchSum;
-+	U32 litLengthSum;
-+	U32 litSum;
-+	U32 offCodeSum;
-+	U32 log2matchLengthSum;
-+	U32 log2matchSum;
-+	U32 log2litLengthSum;
-+	U32 log2litSum;
-+	U32 log2offCodeSum;
-+	U32 factor;
-+	U32 staticPrices;
-+	U32 cachedPrice;
-+	U32 cachedLitLength;
-+	const BYTE *cachedLiterals;
-+} seqStore_t;
-+
-+const seqStore_t *ZSTD_getSeqStore(const ZSTD_CCtx *ctx);
-+void ZSTD_seqToCodes(const seqStore_t *seqStorePtr);
-+int ZSTD_isSkipFrame(ZSTD_DCtx *dctx);
-+
-+/*= Custom memory allocation functions */
-+typedef void *(*ZSTD_allocFunction)(void *opaque, size_t size);
-+typedef void (*ZSTD_freeFunction)(void *opaque, void *address);
-+typedef struct {
-+	ZSTD_allocFunction customAlloc;
-+	ZSTD_freeFunction customFree;
-+	void *opaque;
-+} ZSTD_customMem;
-+
-+void *ZSTD_malloc(size_t size, ZSTD_customMem customMem);
-+void ZSTD_free(void *ptr, ZSTD_customMem customMem);
-+
-+/*====== stack allocation  ======*/
-+
-+typedef struct {
-+	void *ptr;
-+	const void *end;
-+} ZSTD_stack;
-+
-+#define ZSTD_ALIGN(x) ALIGN(x, sizeof(size_t))
-+#define ZSTD_PTR_ALIGN(p) PTR_ALIGN(p, sizeof(size_t))
-+
-+ZSTD_customMem ZSTD_initStack(void *workspace, size_t workspaceSize);
-+
-+void *ZSTD_stackAllocAll(void *opaque, size_t *size);
-+void *ZSTD_stackAlloc(void *opaque, size_t size);
-+void ZSTD_stackFree(void *opaque, void *address);
-+
-+/*======  common function  ======*/
-+
-+ZSTD_STATIC U32 ZSTD_highbit32(U32 val) { return 31 - __builtin_clz(val); }
-+
-+/* hidden functions */
-+
-+/* ZSTD_invalidateRepCodes() :
-+ * ensures next compression will not use repcodes from previous block.
-+ * Note : only works with regular variant;
-+ *        do not use with extDict variant ! */
-+void ZSTD_invalidateRepCodes(ZSTD_CCtx *cctx);
-+
-+size_t ZSTD_freeCCtx(ZSTD_CCtx *cctx);
-+size_t ZSTD_freeDCtx(ZSTD_DCtx *dctx);
-+size_t ZSTD_freeCDict(ZSTD_CDict *cdict);
-+size_t ZSTD_freeDDict(ZSTD_DDict *cdict);
-+size_t ZSTD_freeCStream(ZSTD_CStream *zcs);
-+size_t ZSTD_freeDStream(ZSTD_DStream *zds);
-+
-+#endif /* ZSTD_CCOMMON_H_MODULE */
-diff --git a/lib/zstd/zstd_opt.h b/lib/zstd/zstd_opt.h
-new file mode 100644
-index 0000000..55e1b4c
---- /dev/null
-+++ b/lib/zstd/zstd_opt.h
-@@ -0,0 +1,1014 @@
-+/**
-+ * Copyright (c) 2016-present, Przemyslaw Skibinski, Yann Collet, Facebook, Inc.
-+ * All rights reserved.
-+ *
-+ * This source code is licensed under the BSD-style license found in the
-+ * LICENSE file in the root directory of https://github.com/facebook/zstd.
-+ * An additional grant of patent rights can be found in the PATENTS file in the
-+ * same directory.
-+ *
-+ * This program is free software; you can redistribute it and/or modify it under
-+ * the terms of the GNU General Public License version 2 as published by the
-+ * Free Software Foundation. This program is dual-licensed; you may select
-+ * either version 2 of the GNU General Public License ("GPL") or BSD license
-+ * ("BSD").
-+ */
-+
-+/* Note : this file is intended to be included within zstd_compress.c */
-+
-+#ifndef ZSTD_OPT_H_91842398743
-+#define ZSTD_OPT_H_91842398743
-+
-+#define ZSTD_LITFREQ_ADD 2
-+#define ZSTD_FREQ_DIV 4
-+#define ZSTD_MAX_PRICE (1 << 30)
-+
-+/*-*************************************
-+*  Price functions for optimal parser
-+***************************************/
-+FORCE_INLINE void ZSTD_setLog2Prices(seqStore_t *ssPtr)
-+{
-+	ssPtr->log2matchLengthSum = ZSTD_highbit32(ssPtr->matchLengthSum + 1);
-+	ssPtr->log2litLengthSum = ZSTD_highbit32(ssPtr->litLengthSum + 1);
-+	ssPtr->log2litSum = ZSTD_highbit32(ssPtr->litSum + 1);
-+	ssPtr->log2offCodeSum = ZSTD_highbit32(ssPtr->offCodeSum + 1);
-+	ssPtr->factor = 1 + ((ssPtr->litSum >> 5) / ssPtr->litLengthSum) + ((ssPtr->litSum << 1) / (ssPtr->litSum + ssPtr->matchSum));
-+}
-+
-+ZSTD_STATIC void ZSTD_rescaleFreqs(seqStore_t *ssPtr, const BYTE *src, size_t srcSize)
-+{
-+	unsigned u;
-+
-+	ssPtr->cachedLiterals = NULL;
-+	ssPtr->cachedPrice = ssPtr->cachedLitLength = 0;
-+	ssPtr->staticPrices = 0;
-+
-+	if (ssPtr->litLengthSum == 0) {
-+		if (srcSize <= 1024)
-+			ssPtr->staticPrices = 1;
-+
-+		for (u = 0; u <= MaxLit; u++)
-+			ssPtr->litFreq[u] = 0;
-+		for (u = 0; u < srcSize; u++)
-+			ssPtr->litFreq[src[u]]++;
-+
-+		ssPtr->litSum = 0;
-+		ssPtr->litLengthSum = MaxLL + 1;
-+		ssPtr->matchLengthSum = MaxML + 1;
-+		ssPtr->offCodeSum = (MaxOff + 1);
-+		ssPtr->matchSum = (ZSTD_LITFREQ_ADD << Litbits);
-+
-+		for (u = 0; u <= MaxLit; u++) {
-+			ssPtr->litFreq[u] = 1 + (ssPtr->litFreq[u] >> ZSTD_FREQ_DIV);
-+			ssPtr->litSum += ssPtr->litFreq[u];
-+		}
-+		for (u = 0; u <= MaxLL; u++)
-+			ssPtr->litLengthFreq[u] = 1;
-+		for (u = 0; u <= MaxML; u++)
-+			ssPtr->matchLengthFreq[u] = 1;
-+		for (u = 0; u <= MaxOff; u++)
-+			ssPtr->offCodeFreq[u] = 1;
-+	} else {
-+		ssPtr->matchLengthSum = 0;
-+		ssPtr->litLengthSum = 0;
-+		ssPtr->offCodeSum = 0;
-+		ssPtr->matchSum = 0;
-+		ssPtr->litSum = 0;
-+
-+		for (u = 0; u <= MaxLit; u++) {
-+			ssPtr->litFreq[u] = 1 + (ssPtr->litFreq[u] >> (ZSTD_FREQ_DIV + 1));
-+			ssPtr->litSum += ssPtr->litFreq[u];
-+		}
-+		for (u = 0; u <= MaxLL; u++) {
-+			ssPtr->litLengthFreq[u] = 1 + (ssPtr->litLengthFreq[u] >> (ZSTD_FREQ_DIV + 1));
-+			ssPtr->litLengthSum += ssPtr->litLengthFreq[u];
-+		}
-+		for (u = 0; u <= MaxML; u++) {
-+			ssPtr->matchLengthFreq[u] = 1 + (ssPtr->matchLengthFreq[u] >> ZSTD_FREQ_DIV);
-+			ssPtr->matchLengthSum += ssPtr->matchLengthFreq[u];
-+			ssPtr->matchSum += ssPtr->matchLengthFreq[u] * (u + 3);
-+		}
-+		ssPtr->matchSum *= ZSTD_LITFREQ_ADD;
-+		for (u = 0; u <= MaxOff; u++) {
-+			ssPtr->offCodeFreq[u] = 1 + (ssPtr->offCodeFreq[u] >> ZSTD_FREQ_DIV);
-+			ssPtr->offCodeSum += ssPtr->offCodeFreq[u];
-+		}
-+	}
-+
-+	ZSTD_setLog2Prices(ssPtr);
-+}
-+
-+FORCE_INLINE U32 ZSTD_getLiteralPrice(seqStore_t *ssPtr, U32 litLength, const BYTE *literals)
-+{
-+	U32 price, u;
-+
-+	if (ssPtr->staticPrices)
-+		return ZSTD_highbit32((U32)litLength + 1) + (litLength * 6);
-+
-+	if (litLength == 0)
-+		return ssPtr->log2litLengthSum - ZSTD_highbit32(ssPtr->litLengthFreq[0] + 1);
-+
-+	/* literals */
-+	if (ssPtr->cachedLiterals == literals) {
-+		U32 const additional = litLength - ssPtr->cachedLitLength;
-+		const BYTE *literals2 = ssPtr->cachedLiterals + ssPtr->cachedLitLength;
-+		price = ssPtr->cachedPrice + additional * ssPtr->log2litSum;
-+		for (u = 0; u < additional; u++)
-+			price -= ZSTD_highbit32(ssPtr->litFreq[literals2[u]] + 1);
-+		ssPtr->cachedPrice = price;
-+		ssPtr->cachedLitLength = litLength;
-+	} else {
-+		price = litLength * ssPtr->log2litSum;
-+		for (u = 0; u < litLength; u++)
-+			price -= ZSTD_highbit32(ssPtr->litFreq[literals[u]] + 1);
-+
-+		if (litLength >= 12) {
-+			ssPtr->cachedLiterals = literals;
-+			ssPtr->cachedPrice = price;
-+			ssPtr->cachedLitLength = litLength;
-+		}
-+	}
-+
-+	/* literal Length */
-+	{
-+		const BYTE LL_deltaCode = 19;
-+		const BYTE llCode = (litLength > 63) ? (BYTE)ZSTD_highbit32(litLength) + LL_deltaCode : LL_Code[litLength];
-+		price += LL_bits[llCode] + ssPtr->log2litLengthSum - ZSTD_highbit32(ssPtr->litLengthFreq[llCode] + 1);
-+	}
-+
-+	return price;
-+}
-+
-+FORCE_INLINE U32 ZSTD_getPrice(seqStore_t *seqStorePtr, U32 litLength, const BYTE *literals, U32 offset, U32 matchLength, const int ultra)
-+{
-+	/* offset */
-+	U32 price;
-+	BYTE const offCode = (BYTE)ZSTD_highbit32(offset + 1);
-+
-+	if (seqStorePtr->staticPrices)
-+		return ZSTD_getLiteralPrice(seqStorePtr, litLength, literals) + ZSTD_highbit32((U32)matchLength + 1) + 16 + offCode;
-+
-+	price = offCode + seqStorePtr->log2offCodeSum - ZSTD_highbit32(seqStorePtr->offCodeFreq[offCode] + 1);
-+	if (!ultra && offCode >= 20)
-+		price += (offCode - 19) * 2;
-+
-+	/* match Length */
-+	{
-+		const BYTE ML_deltaCode = 36;
-+		const BYTE mlCode = (matchLength > 127) ? (BYTE)ZSTD_highbit32(matchLength) + ML_deltaCode : ML_Code[matchLength];
-+		price += ML_bits[mlCode] + seqStorePtr->log2matchLengthSum - ZSTD_highbit32(seqStorePtr->matchLengthFreq[mlCode] + 1);
-+	}
-+
-+	return price + ZSTD_getLiteralPrice(seqStorePtr, litLength, literals) + seqStorePtr->factor;
-+}
-+
-+ZSTD_STATIC void ZSTD_updatePrice(seqStore_t *seqStorePtr, U32 litLength, const BYTE *literals, U32 offset, U32 matchLength)
-+{
-+	U32 u;
-+
-+	/* literals */
-+	seqStorePtr->litSum += litLength * ZSTD_LITFREQ_ADD;
-+	for (u = 0; u < litLength; u++)
-+		seqStorePtr->litFreq[literals[u]] += ZSTD_LITFREQ_ADD;
-+
-+	/* literal Length */
-+	{
-+		const BYTE LL_deltaCode = 19;
-+		const BYTE llCode = (litLength > 63) ? (BYTE)ZSTD_highbit32(litLength) + LL_deltaCode : LL_Code[litLength];
-+		seqStorePtr->litLengthFreq[llCode]++;
-+		seqStorePtr->litLengthSum++;
-+	}
-+
-+	/* match offset */
-+	{
-+		BYTE const offCode = (BYTE)ZSTD_highbit32(offset + 1);
-+		seqStorePtr->offCodeSum++;
-+		seqStorePtr->offCodeFreq[offCode]++;
-+	}
-+
-+	/* match Length */
-+	{
-+		const BYTE ML_deltaCode = 36;
-+		const BYTE mlCode = (matchLength > 127) ? (BYTE)ZSTD_highbit32(matchLength) + ML_deltaCode : ML_Code[matchLength];
-+		seqStorePtr->matchLengthFreq[mlCode]++;
-+		seqStorePtr->matchLengthSum++;
-+	}
-+
-+	ZSTD_setLog2Prices(seqStorePtr);
-+}
-+
-+#define SET_PRICE(pos, mlen_, offset_, litlen_, price_)           \
-+	{                                                         \
-+		while (last_pos < pos) {                          \
-+			opt[last_pos + 1].price = ZSTD_MAX_PRICE; \
-+			last_pos++;                               \
-+		}                                                 \
-+		opt[pos].mlen = mlen_;                            \
-+		opt[pos].off = offset_;                           \
-+		opt[pos].litlen = litlen_;                        \
-+		opt[pos].price = price_;                          \
-+	}
-+
-+/* Update hashTable3 up to ip (excluded)
-+   Assumption : always within prefix (i.e. not within extDict) */
-+FORCE_INLINE
-+U32 ZSTD_insertAndFindFirstIndexHash3(ZSTD_CCtx *zc, const BYTE *ip)
-+{
-+	U32 *const hashTable3 = zc->hashTable3;
-+	U32 const hashLog3 = zc->hashLog3;
-+	const BYTE *const base = zc->base;
-+	U32 idx = zc->nextToUpdate3;
-+	const U32 target = zc->nextToUpdate3 = (U32)(ip - base);
-+	const size_t hash3 = ZSTD_hash3Ptr(ip, hashLog3);
-+
-+	while (idx < target) {
-+		hashTable3[ZSTD_hash3Ptr(base + idx, hashLog3)] = idx;
-+		idx++;
-+	}
-+
-+	return hashTable3[hash3];
-+}
-+
-+/*-*************************************
-+*  Binary Tree search
-+***************************************/
-+static U32 ZSTD_insertBtAndGetAllMatches(ZSTD_CCtx *zc, const BYTE *const ip, const BYTE *const iLimit, U32 nbCompares, const U32 mls, U32 extDict,
-+					 ZSTD_match_t *matches, const U32 minMatchLen)
-+{
-+	const BYTE *const base = zc->base;
-+	const U32 curr = (U32)(ip - base);
-+	const U32 hashLog = zc->params.cParams.hashLog;
-+	const size_t h = ZSTD_hashPtr(ip, hashLog, mls);
-+	U32 *const hashTable = zc->hashTable;
-+	U32 matchIndex = hashTable[h];
-+	U32 *const bt = zc->chainTable;
-+	const U32 btLog = zc->params.cParams.chainLog - 1;
-+	const U32 btMask = (1U << btLog) - 1;
-+	size_t commonLengthSmaller = 0, commonLengthLarger = 0;
-+	const BYTE *const dictBase = zc->dictBase;
-+	const U32 dictLimit = zc->dictLimit;
-+	const BYTE *const dictEnd = dictBase + dictLimit;
-+	const BYTE *const prefixStart = base + dictLimit;
-+	const U32 btLow = btMask >= curr ? 0 : curr - btMask;
-+	const U32 windowLow = zc->lowLimit;
-+	U32 *smallerPtr = bt + 2 * (curr & btMask);
-+	U32 *largerPtr = bt + 2 * (curr & btMask) + 1;
-+	U32 matchEndIdx = curr + 8;
-+	U32 dummy32; /* to be nullified at the end */
-+	U32 mnum = 0;
-+
-+	const U32 minMatch = (mls == 3) ? 3 : 4;
-+	size_t bestLength = minMatchLen - 1;
-+
-+	if (minMatch == 3) { /* HC3 match finder */
-+		U32 const matchIndex3 = ZSTD_insertAndFindFirstIndexHash3(zc, ip);
-+		if (matchIndex3 > windowLow && (curr - matchIndex3 < (1 << 18))) {
-+			const BYTE *match;
-+			size_t currMl = 0;
-+			if ((!extDict) || matchIndex3 >= dictLimit) {
-+				match = base + matchIndex3;
-+				if (match[bestLength] == ip[bestLength])
-+					currMl = ZSTD_count(ip, match, iLimit);
-+			} else {
-+				match = dictBase + matchIndex3;
-+				if (ZSTD_readMINMATCH(match, MINMATCH) ==
-+				    ZSTD_readMINMATCH(ip, MINMATCH)) /* assumption : matchIndex3 <= dictLimit-4 (by table construction) */
-+					currMl = ZSTD_count_2segments(ip + MINMATCH, match + MINMATCH, iLimit, dictEnd, prefixStart) + MINMATCH;
-+			}
-+
-+			/* save best solution */
-+			if (currMl > bestLength) {
-+				bestLength = currMl;
-+				matches[mnum].off = ZSTD_REP_MOVE_OPT + curr - matchIndex3;
-+				matches[mnum].len = (U32)currMl;
-+				mnum++;
-+				if (currMl > ZSTD_OPT_NUM)
-+					goto update;
-+				if (ip + currMl == iLimit)
-+					goto update; /* best possible, and avoid read overflow*/
-+			}
-+		}
-+	}
-+
-+	hashTable[h] = curr; /* Update Hash Table */
-+
-+	while (nbCompares-- && (matchIndex > windowLow)) {
-+		U32 *nextPtr = bt + 2 * (matchIndex & btMask);
-+		size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */
-+		const BYTE *match;
-+
-+		if ((!extDict) || (matchIndex + matchLength >= dictLimit)) {
-+			match = base + matchIndex;
-+			if (match[matchLength] == ip[matchLength]) {
-+				matchLength += ZSTD_count(ip + matchLength + 1, match + matchLength + 1, iLimit) + 1;
-+			}
-+		} else {
-+			match = dictBase + matchIndex;
-+			matchLength += ZSTD_count_2segments(ip + matchLength, match + matchLength, iLimit, dictEnd, prefixStart);
-+			if (matchIndex + matchLength >= dictLimit)
-+				match = base + matchIndex; /* to prepare for next usage of match[matchLength] */
-+		}
-+
-+		if (matchLength > bestLength) {
-+			if (matchLength > matchEndIdx - matchIndex)
-+				matchEndIdx = matchIndex + (U32)matchLength;
-+			bestLength = matchLength;
-+			matches[mnum].off = ZSTD_REP_MOVE_OPT + curr - matchIndex;
-+			matches[mnum].len = (U32)matchLength;
-+			mnum++;
-+			if (matchLength > ZSTD_OPT_NUM)
-+				break;
-+			if (ip + matchLength == iLimit) /* equal : no way to know if inf or sup */
-+				break;			/* drop, to guarantee consistency (miss a little bit of compression) */
-+		}
-+
-+		if (match[matchLength] < ip[matchLength]) {
-+			/* match is smaller than curr */
-+			*smallerPtr = matchIndex;	  /* update smaller idx */
-+			commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */
-+			if (matchIndex <= btLow) {
-+				smallerPtr = &dummy32;
-+				break;
-+			}			  /* beyond tree size, stop the search */
-+			smallerPtr = nextPtr + 1; /* new "smaller" => larger of match */
-+			matchIndex = nextPtr[1];  /* new matchIndex larger than previous (closer to curr) */
-+		} else {
-+			/* match is larger than curr */
-+			*largerPtr = matchIndex;
-+			commonLengthLarger = matchLength;
-+			if (matchIndex <= btLow) {
-+				largerPtr = &dummy32;
-+				break;
-+			} /* beyond tree size, stop the search */
-+			largerPtr = nextPtr;
-+			matchIndex = nextPtr[0];
-+		}
-+	}
-+
-+	*smallerPtr = *largerPtr = 0;
-+
-+update:
-+	zc->nextToUpdate = (matchEndIdx > curr + 8) ? matchEndIdx - 8 : curr + 1;
-+	return mnum;
-+}
-+
-+/** Tree updater, providing best match */
-+static U32 ZSTD_BtGetAllMatches(ZSTD_CCtx *zc, const BYTE *const ip, const BYTE *const iLimit, const U32 maxNbAttempts, const U32 mls, ZSTD_match_t *matches,
-+				const U32 minMatchLen)
-+{
-+	if (ip < zc->base + zc->nextToUpdate)
-+		return 0; /* skipped area */
-+	ZSTD_updateTree(zc, ip, iLimit, maxNbAttempts, mls);
-+	return ZSTD_insertBtAndGetAllMatches(zc, ip, iLimit, maxNbAttempts, mls, 0, matches, minMatchLen);
-+}
-+
-+static U32 ZSTD_BtGetAllMatches_selectMLS(ZSTD_CCtx *zc, /* Index table will be updated */
-+					  const BYTE *ip, const BYTE *const iHighLimit, const U32 maxNbAttempts, const U32 matchLengthSearch,
-+					  ZSTD_match_t *matches, const U32 minMatchLen)
-+{
-+	switch (matchLengthSearch) {
-+	case 3: return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 3, matches, minMatchLen);
-+	default:
-+	case 4: return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 4, matches, minMatchLen);
-+	case 5: return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 5, matches, minMatchLen);
-+	case 7:
-+	case 6: return ZSTD_BtGetAllMatches(zc, ip, iHighLimit, maxNbAttempts, 6, matches, minMatchLen);
-+	}
-+}
-+
-+/** Tree updater, providing best match */
-+static U32 ZSTD_BtGetAllMatches_extDict(ZSTD_CCtx *zc, const BYTE *const ip, const BYTE *const iLimit, const U32 maxNbAttempts, const U32 mls,
-+					ZSTD_match_t *matches, const U32 minMatchLen)
-+{
-+	if (ip < zc->base + zc->nextToUpdate)
-+		return 0; /* skipped area */
-+	ZSTD_updateTree_extDict(zc, ip, iLimit, maxNbAttempts, mls);
-+	return ZSTD_insertBtAndGetAllMatches(zc, ip, iLimit, maxNbAttempts, mls, 1, matches, minMatchLen);
-+}
-+
-+static U32 ZSTD_BtGetAllMatches_selectMLS_extDict(ZSTD_CCtx *zc, /* Index table will be updated */
-+						  const BYTE *ip, const BYTE *const iHighLimit, const U32 maxNbAttempts, const U32 matchLengthSearch,
-+						  ZSTD_match_t *matches, const U32 minMatchLen)
-+{
-+	switch (matchLengthSearch) {
-+	case 3: return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 3, matches, minMatchLen);
-+	default:
-+	case 4: return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 4, matches, minMatchLen);
-+	case 5: return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 5, matches, minMatchLen);
-+	case 7:
-+	case 6: return ZSTD_BtGetAllMatches_extDict(zc, ip, iHighLimit, maxNbAttempts, 6, matches, minMatchLen);
-+	}
-+}
-+
-+/*-*******************************
-+*  Optimal parser
-+*********************************/
-+FORCE_INLINE
-+void ZSTD_compressBlock_opt_generic(ZSTD_CCtx *ctx, const void *src, size_t srcSize, const int ultra)
-+{
-+	seqStore_t *seqStorePtr = &(ctx->seqStore);
-+	const BYTE *const istart = (const BYTE *)src;
-+	const BYTE *ip = istart;
-+	const BYTE *anchor = istart;
-+	const BYTE *const iend = istart + srcSize;
-+	const BYTE *const ilimit = iend - 8;
-+	const BYTE *const base = ctx->base;
-+	const BYTE *const prefixStart = base + ctx->dictLimit;
-+
-+	const U32 maxSearches = 1U << ctx->params.cParams.searchLog;
-+	const U32 sufficient_len = ctx->params.cParams.targetLength;
-+	const U32 mls = ctx->params.cParams.searchLength;
-+	const U32 minMatch = (ctx->params.cParams.searchLength == 3) ? 3 : 4;
-+
-+	ZSTD_optimal_t *opt = seqStorePtr->priceTable;
-+	ZSTD_match_t *matches = seqStorePtr->matchTable;
-+	const BYTE *inr;
-+	U32 offset, rep[ZSTD_REP_NUM];
-+
-+	/* init */
-+	ctx->nextToUpdate3 = ctx->nextToUpdate;
-+	ZSTD_rescaleFreqs(seqStorePtr, (const BYTE *)src, srcSize);
-+	ip += (ip == prefixStart);
-+	{
-+		U32 i;
-+		for (i = 0; i < ZSTD_REP_NUM; i++)
-+			rep[i] = ctx->rep[i];
-+	}
-+
-+	/* Match Loop */
-+	while (ip < ilimit) {
-+		U32 cur, match_num, last_pos, litlen, price;
-+		U32 u, mlen, best_mlen, best_off, litLength;
-+		memset(opt, 0, sizeof(ZSTD_optimal_t));
-+		last_pos = 0;
-+		litlen = (U32)(ip - anchor);
-+
-+		/* check repCode */
-+		{
-+			U32 i, last_i = ZSTD_REP_CHECK + (ip == anchor);
-+			for (i = (ip == anchor); i < last_i; i++) {
-+				const S32 repCur = (i == ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : rep[i];
-+				if ((repCur > 0) && (repCur < (S32)(ip - prefixStart)) &&
-+				    (ZSTD_readMINMATCH(ip, minMatch) == ZSTD_readMINMATCH(ip - repCur, minMatch))) {
-+					mlen = (U32)ZSTD_count(ip + minMatch, ip + minMatch - repCur, iend) + minMatch;
-+					if (mlen > sufficient_len || mlen >= ZSTD_OPT_NUM) {
-+						best_mlen = mlen;
-+						best_off = i;
-+						cur = 0;
-+						last_pos = 1;
-+						goto _storeSequence;
-+					}
-+					best_off = i - (ip == anchor);
-+					do {
-+						price = ZSTD_getPrice(seqStorePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra);
-+						if (mlen > last_pos || price < opt[mlen].price)
-+							SET_PRICE(mlen, mlen, i, litlen, price); /* note : macro modifies last_pos */
-+						mlen--;
-+					} while (mlen >= minMatch);
-+				}
-+			}
-+		}
-+
-+		match_num = ZSTD_BtGetAllMatches_selectMLS(ctx, ip, iend, maxSearches, mls, matches, minMatch);
-+
-+		if (!last_pos && !match_num) {
-+			ip++;
-+			continue;
-+		}
-+
-+		if (match_num && (matches[match_num - 1].len > sufficient_len || matches[match_num - 1].len >= ZSTD_OPT_NUM)) {
-+			best_mlen = matches[match_num - 1].len;
-+			best_off = matches[match_num - 1].off;
-+			cur = 0;
-+			last_pos = 1;
-+			goto _storeSequence;
-+		}
-+
-+		/* set prices using matches at position = 0 */
-+		best_mlen = (last_pos) ? last_pos : minMatch;
-+		for (u = 0; u < match_num; u++) {
-+			mlen = (u > 0) ? matches[u - 1].len + 1 : best_mlen;
-+			best_mlen = matches[u].len;
-+			while (mlen <= best_mlen) {
-+				price = ZSTD_getPrice(seqStorePtr, litlen, anchor, matches[u].off - 1, mlen - MINMATCH, ultra);
-+				if (mlen > last_pos || price < opt[mlen].price)
-+					SET_PRICE(mlen, mlen, matches[u].off, litlen, price); /* note : macro modifies last_pos */
-+				mlen++;
-+			}
-+		}
-+
-+		if (last_pos < minMatch) {
-+			ip++;
-+			continue;
-+		}
-+
-+		/* initialize opt[0] */
-+		{
-+			U32 i;
-+			for (i = 0; i < ZSTD_REP_NUM; i++)
-+				opt[0].rep[i] = rep[i];
-+		}
-+		opt[0].mlen = 1;
-+		opt[0].litlen = litlen;
-+
-+		/* check further positions */
-+		for (cur = 1; cur <= last_pos; cur++) {
-+			inr = ip + cur;
-+
-+			if (opt[cur - 1].mlen == 1) {
-+				litlen = opt[cur - 1].litlen + 1;
-+				if (cur > litlen) {
-+					price = opt[cur - litlen].price + ZSTD_getLiteralPrice(seqStorePtr, litlen, inr - litlen);
-+				} else
-+					price = ZSTD_getLiteralPrice(seqStorePtr, litlen, anchor);
-+			} else {
-+				litlen = 1;
-+				price = opt[cur - 1].price + ZSTD_getLiteralPrice(seqStorePtr, litlen, inr - 1);
-+			}
-+
-+			if (cur > last_pos || price <= opt[cur].price)
-+				SET_PRICE(cur, 1, 0, litlen, price);
-+
-+			if (cur == last_pos)
-+				break;
-+
-+			if (inr > ilimit) /* last match must start at a minimum distance of 8 from oend */
-+				continue;
-+
-+			mlen = opt[cur].mlen;
-+			if (opt[cur].off > ZSTD_REP_MOVE_OPT) {
-+				opt[cur].rep[2] = opt[cur - mlen].rep[1];
-+				opt[cur].rep[1] = opt[cur - mlen].rep[0];
-+				opt[cur].rep[0] = opt[cur].off - ZSTD_REP_MOVE_OPT;
-+			} else {
-+				opt[cur].rep[2] = (opt[cur].off > 1) ? opt[cur - mlen].rep[1] : opt[cur - mlen].rep[2];
-+				opt[cur].rep[1] = (opt[cur].off > 0) ? opt[cur - mlen].rep[0] : opt[cur - mlen].rep[1];
-+				opt[cur].rep[0] =
-+				    ((opt[cur].off == ZSTD_REP_MOVE_OPT) && (mlen != 1)) ? (opt[cur - mlen].rep[0] - 1) : (opt[cur - mlen].rep[opt[cur].off]);
-+			}
-+
-+			best_mlen = minMatch;
-+			{
-+				U32 i, last_i = ZSTD_REP_CHECK + (mlen != 1);
-+				for (i = (opt[cur].mlen != 1); i < last_i; i++) { /* check rep */
-+					const S32 repCur = (i == ZSTD_REP_MOVE_OPT) ? (opt[cur].rep[0] - 1) : opt[cur].rep[i];
-+					if ((repCur > 0) && (repCur < (S32)(inr - prefixStart)) &&
-+					    (ZSTD_readMINMATCH(inr, minMatch) == ZSTD_readMINMATCH(inr - repCur, minMatch))) {
-+						mlen = (U32)ZSTD_count(inr + minMatch, inr + minMatch - repCur, iend) + minMatch;
-+
-+						if (mlen > sufficient_len || cur + mlen >= ZSTD_OPT_NUM) {
-+							best_mlen = mlen;
-+							best_off = i;
-+							last_pos = cur + 1;
-+							goto _storeSequence;
-+						}
-+
-+						best_off = i - (opt[cur].mlen != 1);
-+						if (mlen > best_mlen)
-+							best_mlen = mlen;
-+
-+						do {
-+							if (opt[cur].mlen == 1) {
-+								litlen = opt[cur].litlen;
-+								if (cur > litlen) {
-+									price = opt[cur - litlen].price + ZSTD_getPrice(seqStorePtr, litlen, inr - litlen,
-+															best_off, mlen - MINMATCH, ultra);
-+								} else
-+									price = ZSTD_getPrice(seqStorePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra);
-+							} else {
-+								litlen = 0;
-+								price = opt[cur].price + ZSTD_getPrice(seqStorePtr, 0, NULL, best_off, mlen - MINMATCH, ultra);
-+							}
-+
-+							if (cur + mlen > last_pos || price <= opt[cur + mlen].price)
-+								SET_PRICE(cur + mlen, mlen, i, litlen, price);
-+							mlen--;
-+						} while (mlen >= minMatch);
-+					}
-+				}
-+			}
-+
-+			match_num = ZSTD_BtGetAllMatches_selectMLS(ctx, inr, iend, maxSearches, mls, matches, best_mlen);
-+
-+			if (match_num > 0 && (matches[match_num - 1].len > sufficient_len || cur + matches[match_num - 1].len >= ZSTD_OPT_NUM)) {
-+				best_mlen = matches[match_num - 1].len;
-+				best_off = matches[match_num - 1].off;
-+				last_pos = cur + 1;
-+				goto _storeSequence;
-+			}
-+
-+			/* set prices using matches at position = cur */
-+			for (u = 0; u < match_num; u++) {
-+				mlen = (u > 0) ? matches[u - 1].len + 1 : best_mlen;
-+				best_mlen = matches[u].len;
-+
-+				while (mlen <= best_mlen) {
-+					if (opt[cur].mlen == 1) {
-+						litlen = opt[cur].litlen;
-+						if (cur > litlen)
-+							price = opt[cur - litlen].price + ZSTD_getPrice(seqStorePtr, litlen, ip + cur - litlen,
-+													matches[u].off - 1, mlen - MINMATCH, ultra);
-+						else
-+							price = ZSTD_getPrice(seqStorePtr, litlen, anchor, matches[u].off - 1, mlen - MINMATCH, ultra);
-+					} else {
-+						litlen = 0;
-+						price = opt[cur].price + ZSTD_getPrice(seqStorePtr, 0, NULL, matches[u].off - 1, mlen - MINMATCH, ultra);
-+					}
-+
-+					if (cur + mlen > last_pos || (price < opt[cur + mlen].price))
-+						SET_PRICE(cur + mlen, mlen, matches[u].off, litlen, price);
-+
-+					mlen++;
-+				}
-+			}
-+		}
-+
-+		best_mlen = opt[last_pos].mlen;
-+		best_off = opt[last_pos].off;
-+		cur = last_pos - best_mlen;
-+
-+	/* store sequence */
-+_storeSequence: /* cur, last_pos, best_mlen, best_off have to be set */
-+		opt[0].mlen = 1;
-+
-+		while (1) {
-+			mlen = opt[cur].mlen;
-+			offset = opt[cur].off;
-+			opt[cur].mlen = best_mlen;
-+			opt[cur].off = best_off;
-+			best_mlen = mlen;
-+			best_off = offset;
-+			if (mlen > cur)
-+				break;
-+			cur -= mlen;
-+		}
-+
-+		for (u = 0; u <= last_pos;) {
-+			u += opt[u].mlen;
-+		}
-+
-+		for (cur = 0; cur < last_pos;) {
-+			mlen = opt[cur].mlen;
-+			if (mlen == 1) {
-+				ip++;
-+				cur++;
-+				continue;
-+			}
-+			offset = opt[cur].off;
-+			cur += mlen;
-+			litLength = (U32)(ip - anchor);
-+
-+			if (offset > ZSTD_REP_MOVE_OPT) {
-+				rep[2] = rep[1];
-+				rep[1] = rep[0];
-+				rep[0] = offset - ZSTD_REP_MOVE_OPT;
-+				offset--;
-+			} else {
-+				if (offset != 0) {
-+					best_off = (offset == ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : (rep[offset]);
-+					if (offset != 1)
-+						rep[2] = rep[1];
-+					rep[1] = rep[0];
-+					rep[0] = best_off;
-+				}
-+				if (litLength == 0)
-+					offset--;
-+			}
-+
-+			ZSTD_updatePrice(seqStorePtr, litLength, anchor, offset, mlen - MINMATCH);
-+			ZSTD_storeSeq(seqStorePtr, litLength, anchor, offset, mlen - MINMATCH);
-+			anchor = ip = ip + mlen;
-+		}
-+	} /* for (cur=0; cur < last_pos; ) */
-+
-+	/* Save reps for next block */
-+	{
-+		int i;
-+		for (i = 0; i < ZSTD_REP_NUM; i++)
-+			ctx->repToConfirm[i] = rep[i];
-+	}
-+
-+	/* Last Literals */
-+	{
-+		size_t const lastLLSize = iend - anchor;
-+		memcpy(seqStorePtr->lit, anchor, lastLLSize);
-+		seqStorePtr->lit += lastLLSize;
-+	}
-+}
-+
-+FORCE_INLINE
-+void ZSTD_compressBlock_opt_extDict_generic(ZSTD_CCtx *ctx, const void *src, size_t srcSize, const int ultra)
-+{
-+	seqStore_t *seqStorePtr = &(ctx->seqStore);
-+	const BYTE *const istart = (const BYTE *)src;
-+	const BYTE *ip = istart;
-+	const BYTE *anchor = istart;
-+	const BYTE *const iend = istart + srcSize;
-+	const BYTE *const ilimit = iend - 8;
-+	const BYTE *const base = ctx->base;
-+	const U32 lowestIndex = ctx->lowLimit;
-+	const U32 dictLimit = ctx->dictLimit;
-+	const BYTE *const prefixStart = base + dictLimit;
-+	const BYTE *const dictBase = ctx->dictBase;
-+	const BYTE *const dictEnd = dictBase + dictLimit;
-+
-+	const U32 maxSearches = 1U << ctx->params.cParams.searchLog;
-+	const U32 sufficient_len = ctx->params.cParams.targetLength;
-+	const U32 mls = ctx->params.cParams.searchLength;
-+	const U32 minMatch = (ctx->params.cParams.searchLength == 3) ? 3 : 4;
-+
-+	ZSTD_optimal_t *opt = seqStorePtr->priceTable;
-+	ZSTD_match_t *matches = seqStorePtr->matchTable;
-+	const BYTE *inr;
-+
-+	/* init */
-+	U32 offset, rep[ZSTD_REP_NUM];
-+	{
-+		U32 i;
-+		for (i = 0; i < ZSTD_REP_NUM; i++)
-+			rep[i] = ctx->rep[i];
-+	}
-+
-+	ctx->nextToUpdate3 = ctx->nextToUpdate;
-+	ZSTD_rescaleFreqs(seqStorePtr, (const BYTE *)src, srcSize);
-+	ip += (ip == prefixStart);
-+
-+	/* Match Loop */
-+	while (ip < ilimit) {
-+		U32 cur, match_num, last_pos, litlen, price;
-+		U32 u, mlen, best_mlen, best_off, litLength;
-+		U32 curr = (U32)(ip - base);
-+		memset(opt, 0, sizeof(ZSTD_optimal_t));
-+		last_pos = 0;
-+		opt[0].litlen = (U32)(ip - anchor);
-+
-+		/* check repCode */
-+		{
-+			U32 i, last_i = ZSTD_REP_CHECK + (ip == anchor);
-+			for (i = (ip == anchor); i < last_i; i++) {
-+				const S32 repCur = (i == ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : rep[i];
-+				const U32 repIndex = (U32)(curr - repCur);
-+				const BYTE *const repBase = repIndex < dictLimit ? dictBase : base;
-+				const BYTE *const repMatch = repBase + repIndex;
-+				if ((repCur > 0 && repCur <= (S32)curr) &&
-+				    (((U32)((dictLimit - 1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */
-+				    && (ZSTD_readMINMATCH(ip, minMatch) == ZSTD_readMINMATCH(repMatch, minMatch))) {
-+					/* repcode detected we should take it */
-+					const BYTE *const repEnd = repIndex < dictLimit ? dictEnd : iend;
-+					mlen = (U32)ZSTD_count_2segments(ip + minMatch, repMatch + minMatch, iend, repEnd, prefixStart) + minMatch;
-+
-+					if (mlen > sufficient_len || mlen >= ZSTD_OPT_NUM) {
-+						best_mlen = mlen;
-+						best_off = i;
-+						cur = 0;
-+						last_pos = 1;
-+						goto _storeSequence;
-+					}
-+
-+					best_off = i - (ip == anchor);
-+					litlen = opt[0].litlen;
-+					do {
-+						price = ZSTD_getPrice(seqStorePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra);
-+						if (mlen > last_pos || price < opt[mlen].price)
-+							SET_PRICE(mlen, mlen, i, litlen, price); /* note : macro modifies last_pos */
-+						mlen--;
-+					} while (mlen >= minMatch);
-+				}
-+			}
-+		}
-+
-+		match_num = ZSTD_BtGetAllMatches_selectMLS_extDict(ctx, ip, iend, maxSearches, mls, matches, minMatch); /* first search (depth 0) */
-+
-+		if (!last_pos && !match_num) {
-+			ip++;
-+			continue;
-+		}
-+
-+		{
-+			U32 i;
-+			for (i = 0; i < ZSTD_REP_NUM; i++)
-+				opt[0].rep[i] = rep[i];
-+		}
-+		opt[0].mlen = 1;
-+
-+		if (match_num && (matches[match_num - 1].len > sufficient_len || matches[match_num - 1].len >= ZSTD_OPT_NUM)) {
-+			best_mlen = matches[match_num - 1].len;
-+			best_off = matches[match_num - 1].off;
-+			cur = 0;
-+			last_pos = 1;
-+			goto _storeSequence;
-+		}
-+
-+		best_mlen = (last_pos) ? last_pos : minMatch;
-+
-+		/* set prices using matches at position = 0 */
-+		for (u = 0; u < match_num; u++) {
-+			mlen = (u > 0) ? matches[u - 1].len + 1 : best_mlen;
-+			best_mlen = matches[u].len;
-+			litlen = opt[0].litlen;
-+			while (mlen <= best_mlen) {
-+				price = ZSTD_getPrice(seqStorePtr, litlen, anchor, matches[u].off - 1, mlen - MINMATCH, ultra);
-+				if (mlen > last_pos || price < opt[mlen].price)
-+					SET_PRICE(mlen, mlen, matches[u].off, litlen, price);
-+				mlen++;
-+			}
-+		}
-+
-+		if (last_pos < minMatch) {
-+			ip++;
-+			continue;
-+		}
-+
-+		/* check further positions */
-+		for (cur = 1; cur <= last_pos; cur++) {
-+			inr = ip + cur;
-+
-+			if (opt[cur - 1].mlen == 1) {
-+				litlen = opt[cur - 1].litlen + 1;
-+				if (cur > litlen) {
-+					price = opt[cur - litlen].price + ZSTD_getLiteralPrice(seqStorePtr, litlen, inr - litlen);
-+				} else
-+					price = ZSTD_getLiteralPrice(seqStorePtr, litlen, anchor);
-+			} else {
-+				litlen = 1;
-+				price = opt[cur - 1].price + ZSTD_getLiteralPrice(seqStorePtr, litlen, inr - 1);
-+			}
-+
-+			if (cur > last_pos || price <= opt[cur].price)
-+				SET_PRICE(cur, 1, 0, litlen, price);
-+
-+			if (cur == last_pos)
-+				break;
-+
-+			if (inr > ilimit) /* last match must start at a minimum distance of 8 from oend */
-+				continue;
-+
-+			mlen = opt[cur].mlen;
-+			if (opt[cur].off > ZSTD_REP_MOVE_OPT) {
-+				opt[cur].rep[2] = opt[cur - mlen].rep[1];
-+				opt[cur].rep[1] = opt[cur - mlen].rep[0];
-+				opt[cur].rep[0] = opt[cur].off - ZSTD_REP_MOVE_OPT;
-+			} else {
-+				opt[cur].rep[2] = (opt[cur].off > 1) ? opt[cur - mlen].rep[1] : opt[cur - mlen].rep[2];
-+				opt[cur].rep[1] = (opt[cur].off > 0) ? opt[cur - mlen].rep[0] : opt[cur - mlen].rep[1];
-+				opt[cur].rep[0] =
-+				    ((opt[cur].off == ZSTD_REP_MOVE_OPT) && (mlen != 1)) ? (opt[cur - mlen].rep[0] - 1) : (opt[cur - mlen].rep[opt[cur].off]);
-+			}
-+
-+			best_mlen = minMatch;
-+			{
-+				U32 i, last_i = ZSTD_REP_CHECK + (mlen != 1);
-+				for (i = (mlen != 1); i < last_i; i++) {
-+					const S32 repCur = (i == ZSTD_REP_MOVE_OPT) ? (opt[cur].rep[0] - 1) : opt[cur].rep[i];
-+					const U32 repIndex = (U32)(curr + cur - repCur);
-+					const BYTE *const repBase = repIndex < dictLimit ? dictBase : base;
-+					const BYTE *const repMatch = repBase + repIndex;
-+					if ((repCur > 0 && repCur <= (S32)(curr + cur)) &&
-+					    (((U32)((dictLimit - 1) - repIndex) >= 3) & (repIndex > lowestIndex)) /* intentional overflow */
-+					    && (ZSTD_readMINMATCH(inr, minMatch) == ZSTD_readMINMATCH(repMatch, minMatch))) {
-+						/* repcode detected */
-+						const BYTE *const repEnd = repIndex < dictLimit ? dictEnd : iend;
-+						mlen = (U32)ZSTD_count_2segments(inr + minMatch, repMatch + minMatch, iend, repEnd, prefixStart) + minMatch;
-+
-+						if (mlen > sufficient_len || cur + mlen >= ZSTD_OPT_NUM) {
-+							best_mlen = mlen;
-+							best_off = i;
-+							last_pos = cur + 1;
-+							goto _storeSequence;
-+						}
-+
-+						best_off = i - (opt[cur].mlen != 1);
-+						if (mlen > best_mlen)
-+							best_mlen = mlen;
-+
-+						do {
-+							if (opt[cur].mlen == 1) {
-+								litlen = opt[cur].litlen;
-+								if (cur > litlen) {
-+									price = opt[cur - litlen].price + ZSTD_getPrice(seqStorePtr, litlen, inr - litlen,
-+															best_off, mlen - MINMATCH, ultra);
-+								} else
-+									price = ZSTD_getPrice(seqStorePtr, litlen, anchor, best_off, mlen - MINMATCH, ultra);
-+							} else {
-+								litlen = 0;
-+								price = opt[cur].price + ZSTD_getPrice(seqStorePtr, 0, NULL, best_off, mlen - MINMATCH, ultra);
-+							}
-+
-+							if (cur + mlen > last_pos || price <= opt[cur + mlen].price)
-+								SET_PRICE(cur + mlen, mlen, i, litlen, price);
-+							mlen--;
-+						} while (mlen >= minMatch);
-+					}
-+				}
-+			}
-+
-+			match_num = ZSTD_BtGetAllMatches_selectMLS_extDict(ctx, inr, iend, maxSearches, mls, matches, minMatch);
-+
-+			if (match_num > 0 && (matches[match_num - 1].len > sufficient_len || cur + matches[match_num - 1].len >= ZSTD_OPT_NUM)) {
-+				best_mlen = matches[match_num - 1].len;
-+				best_off = matches[match_num - 1].off;
-+				last_pos = cur + 1;
-+				goto _storeSequence;
-+			}
-+
-+			/* set prices using matches at position = cur */
-+			for (u = 0; u < match_num; u++) {
-+				mlen = (u > 0) ? matches[u - 1].len + 1 : best_mlen;
-+				best_mlen = matches[u].len;
-+
-+				while (mlen <= best_mlen) {
-+					if (opt[cur].mlen == 1) {
-+						litlen = opt[cur].litlen;
-+						if (cur > litlen)
-+							price = opt[cur - litlen].price + ZSTD_getPrice(seqStorePtr, litlen, ip + cur - litlen,
-+													matches[u].off - 1, mlen - MINMATCH, ultra);
-+						else
-+							price = ZSTD_getPrice(seqStorePtr, litlen, anchor, matches[u].off - 1, mlen - MINMATCH, ultra);
-+					} else {
-+						litlen = 0;
-+						price = opt[cur].price + ZSTD_getPrice(seqStorePtr, 0, NULL, matches[u].off - 1, mlen - MINMATCH, ultra);
-+					}
-+
-+					if (cur + mlen > last_pos || (price < opt[cur + mlen].price))
-+						SET_PRICE(cur + mlen, mlen, matches[u].off, litlen, price);
-+
-+					mlen++;
-+				}
-+			}
-+		} /* for (cur = 1; cur <= last_pos; cur++) */
-+
-+		best_mlen = opt[last_pos].mlen;
-+		best_off = opt[last_pos].off;
-+		cur = last_pos - best_mlen;
-+
-+	/* store sequence */
-+_storeSequence: /* cur, last_pos, best_mlen, best_off have to be set */
-+		opt[0].mlen = 1;
-+
-+		while (1) {
-+			mlen = opt[cur].mlen;
-+			offset = opt[cur].off;
-+			opt[cur].mlen = best_mlen;
-+			opt[cur].off = best_off;
-+			best_mlen = mlen;
-+			best_off = offset;
-+			if (mlen > cur)
-+				break;
-+			cur -= mlen;
-+		}
-+
-+		for (u = 0; u <= last_pos;) {
-+			u += opt[u].mlen;
-+		}
-+
-+		for (cur = 0; cur < last_pos;) {
-+			mlen = opt[cur].mlen;
-+			if (mlen == 1) {
-+				ip++;
-+				cur++;
-+				continue;
-+			}
-+			offset = opt[cur].off;
-+			cur += mlen;
-+			litLength = (U32)(ip - anchor);
-+
-+			if (offset > ZSTD_REP_MOVE_OPT) {
-+				rep[2] = rep[1];
-+				rep[1] = rep[0];
-+				rep[0] = offset - ZSTD_REP_MOVE_OPT;
-+				offset--;
-+			} else {
-+				if (offset != 0) {
-+					best_off = (offset == ZSTD_REP_MOVE_OPT) ? (rep[0] - 1) : (rep[offset]);
-+					if (offset != 1)
-+						rep[2] = rep[1];
-+					rep[1] = rep[0];
-+					rep[0] = best_off;
-+				}
-+
-+				if (litLength == 0)
-+					offset--;
-+			}
-+
-+			ZSTD_updatePrice(seqStorePtr, litLength, anchor, offset, mlen - MINMATCH);
-+			ZSTD_storeSeq(seqStorePtr, litLength, anchor, offset, mlen - MINMATCH);
-+			anchor = ip = ip + mlen;
-+		}
-+	} /* for (cur=0; cur < last_pos; ) */
-+
-+	/* Save reps for next block */
-+	{
-+		int i;
-+		for (i = 0; i < ZSTD_REP_NUM; i++)
-+			ctx->repToConfirm[i] = rep[i];
-+	}
-+
-+	/* Last Literals */
-+	{
-+		size_t lastLLSize = iend - anchor;
-+		memcpy(seqStorePtr->lit, anchor, lastLLSize);
-+		seqStorePtr->lit += lastLLSize;
-+	}
-+}
-+
-+#endif /* ZSTD_OPT_H_91842398743 */
---
-2.9.3
diff --git a/0003-btrfs-Add-zstd-support.patch b/0003-btrfs-Add-zstd-support.patch
deleted file mode 100644
index edc7839..0000000
--- a/0003-btrfs-Add-zstd-support.patch
+++ /dev/null
@@ -1,740 +0,0 @@
-From 8a9dddfbf6551afea73911e367dd4be64d62b9fd Mon Sep 17 00:00:00 2001
-From: Nick Terrell <terrelln@fb.com>
-Date: Mon, 17 Jul 2017 17:08:39 -0700
-Subject: [PATCH v5 3/5] btrfs: Add zstd support
-
-Add zstd compression and decompression support to BtrFS. zstd at its
-fastest level compresses almost as well as zlib, while offering much
-faster compression and decompression, approaching lzo speeds.
-
-I benchmarked btrfs with zstd compression against no compression, lzo
-compression, and zlib compression. I benchmarked two scenarios. Copying
-a set of files to btrfs, and then reading the files. Copying a tarball
-to btrfs, extracting it to btrfs, and then reading the extracted files.
-After every operation, I call `sync` and include the sync time.
-Between every pair of operations I unmount and remount the filesystem
-to avoid caching. The benchmark files can be found in the upstream
-zstd source repository under
-`contrib/linux-kernel/{btrfs-benchmark.sh,btrfs-extract-benchmark.sh}`
-[1] [2].
-
-I ran the benchmarks on a Ubuntu 14.04 VM with 2 cores and 4 GiB of RAM.
-The VM is running on a MacBook Pro with a 3.1 GHz Intel Core i7 processor,
-16 GB of RAM, and a SSD.
-
-The first compression benchmark is copying 10 copies of the unzipped
-Silesia corpus [3] into a BtrFS filesystem mounted with
-`-o compress-force=Method`. The decompression benchmark times how long
-it takes to `tar` all 10 copies into `/dev/null`. The compression ratio is
-measured by comparing the output of `df` and `du`. See the benchmark file
-[1] for details. I benchmarked multiple zstd compression levels, although
-the patch uses zstd level 1.
-
-| Method  | Ratio | Compression MB/s | Decompression speed |
-|---------|-------|------------------|---------------------|
-| None    |  0.99 |              504 |                 686 |
-| lzo     |  1.66 |              398 |                 442 |
-| zlib    |  2.58 |               65 |                 241 |
-| zstd 1  |  2.57 |              260 |                 383 |
-| zstd 3  |  2.71 |              174 |                 408 |
-| zstd 6  |  2.87 |               70 |                 398 |
-| zstd 9  |  2.92 |               43 |                 406 |
-| zstd 12 |  2.93 |               21 |                 408 |
-| zstd 15 |  3.01 |               11 |                 354 |
-
-The next benchmark first copies `linux-4.11.6.tar` [4] to btrfs. Then it
-measures the compression ratio, extracts the tar, and deletes the tar.
-Then it measures the compression ratio again, and `tar`s the extracted
-files into `/dev/null`. See the benchmark file [2] for details.
-
-| Method | Tar Ratio | Extract Ratio | Copy (s) | Extract (s)| Read (s) |
-|--------|-----------|---------------|----------|------------|----------|
-| None   |      0.97 |          0.78 |    0.981 |      5.501 |    8.807 |
-| lzo    |      2.06 |          1.38 |    1.631 |      8.458 |    8.585 |
-| zlib   |      3.40 |          1.86 |    7.750 |     21.544 |   11.744 |
-| zstd 1 |      3.57 |          1.85 |    2.579 |     11.479 |    9.389 |
-
-[1] https://github.com/facebook/zstd/blob/dev/contrib/linux-kernel/btrfs-benchmark.sh
-[2] https://github.com/facebook/zstd/blob/dev/contrib/linux-kernel/btrfs-extract-benchmark.sh
-[3] http://sun.aei.polsl.pl/~sdeor/index.php?page=silesia
-[4] https://cdn.kernel.org/pub/linux/kernel/v4.x/linux-4.11.6.tar.xz
-
-zstd source repository: https://github.com/facebook/zstd
-
-Signed-off-by: Nick Terrell <terrelln@fb.com>
----
-v2 -> v3:
-- Port upstream BtrFS commits e1ddce71d6, 389a6cfc2a, and 6acafd1eff
-- Change default compression level for BtrFS to 3
-
-v3 -> v4:
-- Add missing includes, which fixes the aarch64 build
-- Fix minor linter warnings
-
- fs/btrfs/Kconfig           |   2 +
- fs/btrfs/Makefile          |   2 +-
- fs/btrfs/compression.c     |   1 +
- fs/btrfs/compression.h     |   6 +-
- fs/btrfs/ctree.h           |   1 +
- fs/btrfs/disk-io.c         |   2 +
- fs/btrfs/ioctl.c           |   6 +-
- fs/btrfs/props.c           |   6 +
- fs/btrfs/super.c           |  12 +-
- fs/btrfs/sysfs.c           |   2 +
- fs/btrfs/zstd.c            | 432 +++++++++++++++++++++++++++++++++++++++++++++
- include/uapi/linux/btrfs.h |   8 +-
- 12 files changed, 468 insertions(+), 12 deletions(-)
- create mode 100644 fs/btrfs/zstd.c
-
-diff --git a/fs/btrfs/Kconfig b/fs/btrfs/Kconfig
-index 80e9c18..a26c63b 100644
---- a/fs/btrfs/Kconfig
-+++ b/fs/btrfs/Kconfig
-@@ -6,6 +6,8 @@ config BTRFS_FS
- 	select ZLIB_DEFLATE
- 	select LZO_COMPRESS
- 	select LZO_DECOMPRESS
-+	select ZSTD_COMPRESS
-+	select ZSTD_DECOMPRESS
- 	select RAID6_PQ
- 	select XOR_BLOCKS
- 	select SRCU
-diff --git a/fs/btrfs/Makefile b/fs/btrfs/Makefile
-index 128ce17..962a95a 100644
---- a/fs/btrfs/Makefile
-+++ b/fs/btrfs/Makefile
-@@ -6,7 +6,7 @@ btrfs-y += super.o ctree.o extent-tree.o print-tree.o root-tree.o dir-item.o \
- 	   transaction.o inode.o file.o tree-defrag.o \
- 	   extent_map.o sysfs.o struct-funcs.o xattr.o ordered-data.o \
- 	   extent_io.o volumes.o async-thread.o ioctl.o locking.o orphan.o \
--	   export.o tree-log.o free-space-cache.o zlib.o lzo.o \
-+	   export.o tree-log.o free-space-cache.o zlib.o lzo.o zstd.o \
- 	   compression.o delayed-ref.o relocation.o delayed-inode.o scrub.o \
- 	   reada.o backref.o ulist.o qgroup.o send.o dev-replace.o raid56.o \
- 	   uuid-tree.o props.o hash.o free-space-tree.o
-diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c
-index d2ef9ac..4ff42d1 100644
---- a/fs/btrfs/compression.c
-+++ b/fs/btrfs/compression.c
-@@ -704,6 +704,7 @@ static struct {
- static const struct btrfs_compress_op * const btrfs_compress_op[] = {
- 	&btrfs_zlib_compress,
- 	&btrfs_lzo_compress,
-+	&btrfs_zstd_compress,
- };
-
- void __init btrfs_init_compress(void)
-diff --git a/fs/btrfs/compression.h b/fs/btrfs/compression.h
-index 87f6d33..2269e00 100644
---- a/fs/btrfs/compression.h
-+++ b/fs/btrfs/compression.h
-@@ -99,8 +99,9 @@ enum btrfs_compression_type {
- 	BTRFS_COMPRESS_NONE  = 0,
- 	BTRFS_COMPRESS_ZLIB  = 1,
- 	BTRFS_COMPRESS_LZO   = 2,
--	BTRFS_COMPRESS_TYPES = 2,
--	BTRFS_COMPRESS_LAST  = 3,
-+	BTRFS_COMPRESS_ZSTD  = 3,
-+	BTRFS_COMPRESS_TYPES = 3,
-+	BTRFS_COMPRESS_LAST  = 4,
- };
-
- struct btrfs_compress_op {
-@@ -128,5 +129,6 @@ struct btrfs_compress_op {
-
- extern const struct btrfs_compress_op btrfs_zlib_compress;
- extern const struct btrfs_compress_op btrfs_lzo_compress;
-+extern const struct btrfs_compress_op btrfs_zstd_compress;
-
- #endif
-diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
-index 3f3eb7b..845d77c 100644
---- a/fs/btrfs/ctree.h
-+++ b/fs/btrfs/ctree.h
-@@ -270,6 +270,7 @@ struct btrfs_super_block {
- 	 BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS |		\
- 	 BTRFS_FEATURE_INCOMPAT_BIG_METADATA |		\
- 	 BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO |		\
-+	 BTRFS_FEATURE_INCOMPAT_COMPRESS_ZSTD |		\
- 	 BTRFS_FEATURE_INCOMPAT_RAID56 |		\
- 	 BTRFS_FEATURE_INCOMPAT_EXTENDED_IREF |		\
- 	 BTRFS_FEATURE_INCOMPAT_SKINNY_METADATA |	\
-diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
-index 080e2eb..04632f4 100644
---- a/fs/btrfs/disk-io.c
-+++ b/fs/btrfs/disk-io.c
-@@ -2828,6 +2828,8 @@ int open_ctree(struct super_block *sb,
- 	features |= BTRFS_FEATURE_INCOMPAT_MIXED_BACKREF;
- 	if (fs_info->compress_type == BTRFS_COMPRESS_LZO)
- 		features |= BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO;
-+	else if (fs_info->compress_type == BTRFS_COMPRESS_ZSTD)
-+		features |= BTRFS_FEATURE_INCOMPAT_COMPRESS_ZSTD;
-
- 	if (features & BTRFS_FEATURE_INCOMPAT_SKINNY_METADATA)
- 		btrfs_info(fs_info, "has skinny extents");
-diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
-index fa1b78c..b9963d9 100644
---- a/fs/btrfs/ioctl.c
-+++ b/fs/btrfs/ioctl.c
-@@ -327,8 +327,10 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg)
-
- 		if (fs_info->compress_type == BTRFS_COMPRESS_LZO)
- 			comp = "lzo";
--		else
-+		else if (fs_info->compress_type == BTRFS_COMPRESS_ZLIB)
- 			comp = "zlib";
-+		else
-+			comp = "zstd";
- 		ret = btrfs_set_prop(inode, "btrfs.compression",
- 				     comp, strlen(comp), 0);
- 		if (ret)
-@@ -1466,6 +1468,8 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
-
- 	if (range->compress_type == BTRFS_COMPRESS_LZO) {
- 		btrfs_set_fs_incompat(fs_info, COMPRESS_LZO);
-+	} else if (range->compress_type == BTRFS_COMPRESS_ZSTD) {
-+		btrfs_set_fs_incompat(fs_info, COMPRESS_ZSTD);
- 	}
-
- 	ret = defrag_count;
-diff --git a/fs/btrfs/props.c b/fs/btrfs/props.c
-index 4b23ae5..20631e9 100644
---- a/fs/btrfs/props.c
-+++ b/fs/btrfs/props.c
-@@ -390,6 +390,8 @@ static int prop_compression_validate(const char *value, size_t len)
- 		return 0;
- 	else if (!strncmp("zlib", value, len))
- 		return 0;
-+	else if (!strncmp("zstd", value, len))
-+		return 0;
-
- 	return -EINVAL;
- }
-@@ -412,6 +414,8 @@ static int prop_compression_apply(struct inode *inode,
- 		type = BTRFS_COMPRESS_LZO;
- 	else if (!strncmp("zlib", value, len))
- 		type = BTRFS_COMPRESS_ZLIB;
-+	else if (!strncmp("zstd", value, len))
-+		type = BTRFS_COMPRESS_ZSTD;
- 	else
- 		return -EINVAL;
-
-@@ -429,6 +433,8 @@ static const char *prop_compression_extract(struct inode *inode)
- 		return "zlib";
- 	case BTRFS_COMPRESS_LZO:
- 		return "lzo";
-+	case BTRFS_COMPRESS_ZSTD:
-+		return "zstd";
- 	}
-
- 	return NULL;
-diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
-index 12540b6..c370dea 100644
---- a/fs/btrfs/super.c
-+++ b/fs/btrfs/super.c
-@@ -513,6 +513,14 @@ int btrfs_parse_options(struct btrfs_fs_info *info, char *options,
- 				btrfs_clear_opt(info->mount_opt, NODATASUM);
- 				btrfs_set_fs_incompat(info, COMPRESS_LZO);
- 				no_compress = 0;
-+			} else if (strcmp(args[0].from, "zstd") == 0) {
-+				compress_type = "zstd";
-+				info->compress_type = BTRFS_COMPRESS_ZSTD;
-+				btrfs_set_opt(info->mount_opt, COMPRESS);
-+				btrfs_clear_opt(info->mount_opt, NODATACOW);
-+				btrfs_clear_opt(info->mount_opt, NODATASUM);
-+				btrfs_set_fs_incompat(info, COMPRESS_ZSTD);
-+				no_compress = 0;
- 			} else if (strncmp(args[0].from, "no", 2) == 0) {
- 				compress_type = "no";
- 				btrfs_clear_opt(info->mount_opt, COMPRESS);
-@@ -1227,8 +1235,10 @@ static int btrfs_show_options(struct seq_file *seq, struct dentry *dentry)
- 	if (btrfs_test_opt(info, COMPRESS)) {
- 		if (info->compress_type == BTRFS_COMPRESS_ZLIB)
- 			compress_type = "zlib";
--		else
-+		else if (info->compress_type == BTRFS_COMPRESS_LZO)
- 			compress_type = "lzo";
-+		else
-+			compress_type = "zstd";
- 		if (btrfs_test_opt(info, FORCE_COMPRESS))
- 			seq_printf(seq, ",compress-force=%s", compress_type);
- 		else
-diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c
-index c2d5f35..2b6d37c 100644
---- a/fs/btrfs/sysfs.c
-+++ b/fs/btrfs/sysfs.c
-@@ -200,6 +200,7 @@ BTRFS_FEAT_ATTR_INCOMPAT(mixed_backref, MIXED_BACKREF);
- BTRFS_FEAT_ATTR_INCOMPAT(default_subvol, DEFAULT_SUBVOL);
- BTRFS_FEAT_ATTR_INCOMPAT(mixed_groups, MIXED_GROUPS);
- BTRFS_FEAT_ATTR_INCOMPAT(compress_lzo, COMPRESS_LZO);
-+BTRFS_FEAT_ATTR_INCOMPAT(compress_zstd, COMPRESS_ZSTD);
- BTRFS_FEAT_ATTR_INCOMPAT(big_metadata, BIG_METADATA);
- BTRFS_FEAT_ATTR_INCOMPAT(extended_iref, EXTENDED_IREF);
- BTRFS_FEAT_ATTR_INCOMPAT(raid56, RAID56);
-@@ -212,6 +213,7 @@ static struct attribute *btrfs_supported_feature_attrs[] = {
- 	BTRFS_FEAT_ATTR_PTR(default_subvol),
- 	BTRFS_FEAT_ATTR_PTR(mixed_groups),
- 	BTRFS_FEAT_ATTR_PTR(compress_lzo),
-+	BTRFS_FEAT_ATTR_PTR(compress_zstd),
- 	BTRFS_FEAT_ATTR_PTR(big_metadata),
- 	BTRFS_FEAT_ATTR_PTR(extended_iref),
- 	BTRFS_FEAT_ATTR_PTR(raid56),
-diff --git a/fs/btrfs/zstd.c b/fs/btrfs/zstd.c
-new file mode 100644
-index 0000000..607ce47
---- /dev/null
-+++ b/fs/btrfs/zstd.c
-@@ -0,0 +1,432 @@
-+/*
-+ * Copyright (c) 2016-present, Facebook, Inc.
-+ * All rights reserved.
-+ *
-+ * This program is free software; you can redistribute it and/or
-+ * modify it under the terms of the GNU General Public
-+ * License v2 as published by the Free Software Foundation.
-+ *
-+ * This program is distributed in the hope that it will be useful,
-+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
-+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
-+ * General Public License for more details.
-+ */
-+#include <linux/bio.h>
-+#include <linux/err.h>
-+#include <linux/init.h>
-+#include <linux/kernel.h>
-+#include <linux/mm.h>
-+#include <linux/pagemap.h>
-+#include <linux/refcount.h>
-+#include <linux/sched.h>
-+#include <linux/slab.h>
-+#include <linux/zstd.h>
-+#include "compression.h"
-+
-+#define ZSTD_BTRFS_MAX_WINDOWLOG 17
-+#define ZSTD_BTRFS_MAX_INPUT (1 << ZSTD_BTRFS_MAX_WINDOWLOG)
-+#define ZSTD_BTRFS_DEFAULT_LEVEL 3
-+
-+static ZSTD_parameters zstd_get_btrfs_parameters(size_t src_len)
-+{
-+	ZSTD_parameters params = ZSTD_getParams(ZSTD_BTRFS_DEFAULT_LEVEL,
-+						src_len, 0);
-+
-+	if (params.cParams.windowLog > ZSTD_BTRFS_MAX_WINDOWLOG)
-+		params.cParams.windowLog = ZSTD_BTRFS_MAX_WINDOWLOG;
-+	WARN_ON(src_len > ZSTD_BTRFS_MAX_INPUT);
-+	return params;
-+}
-+
-+struct workspace {
-+	void *mem;
-+	size_t size;
-+	char *buf;
-+	struct list_head list;
-+};
-+
-+static void zstd_free_workspace(struct list_head *ws)
-+{
-+	struct workspace *workspace = list_entry(ws, struct workspace, list);
-+
-+	kvfree(workspace->mem);
-+	kfree(workspace->buf);
-+	kfree(workspace);
-+}
-+
-+static struct list_head *zstd_alloc_workspace(void)
-+{
-+	ZSTD_parameters params =
-+			zstd_get_btrfs_parameters(ZSTD_BTRFS_MAX_INPUT);
-+	struct workspace *workspace;
-+
-+	workspace = kzalloc(sizeof(*workspace), GFP_KERNEL);
-+	if (!workspace)
-+		return ERR_PTR(-ENOMEM);
-+
-+	workspace->size = max_t(size_t,
-+			ZSTD_CStreamWorkspaceBound(params.cParams),
-+			ZSTD_DStreamWorkspaceBound(ZSTD_BTRFS_MAX_INPUT));
-+	workspace->mem = kvmalloc(workspace->size, GFP_KERNEL);
-+	workspace->buf = kmalloc(PAGE_SIZE, GFP_KERNEL);
-+	if (!workspace->mem || !workspace->buf)
-+		goto fail;
-+
-+	INIT_LIST_HEAD(&workspace->list);
-+
-+	return &workspace->list;
-+fail:
-+	zstd_free_workspace(&workspace->list);
-+	return ERR_PTR(-ENOMEM);
-+}
-+
-+static int zstd_compress_pages(struct list_head *ws,
-+		struct address_space *mapping,
-+		u64 start,
-+		struct page **pages,
-+		unsigned long *out_pages,
-+		unsigned long *total_in,
-+		unsigned long *total_out)
-+{
-+	struct workspace *workspace = list_entry(ws, struct workspace, list);
-+	ZSTD_CStream *stream;
-+	int ret = 0;
-+	int nr_pages = 0;
-+	struct page *in_page = NULL;  /* The current page to read */
-+	struct page *out_page = NULL; /* The current page to write to */
-+	ZSTD_inBuffer in_buf = { NULL, 0, 0 };
-+	ZSTD_outBuffer out_buf = { NULL, 0, 0 };
-+	unsigned long tot_in = 0;
-+	unsigned long tot_out = 0;
-+	unsigned long len = *total_out;
-+	const unsigned long nr_dest_pages = *out_pages;
-+	unsigned long max_out = nr_dest_pages * PAGE_SIZE;
-+	ZSTD_parameters params = zstd_get_btrfs_parameters(len);
-+
-+	*out_pages = 0;
-+	*total_out = 0;
-+	*total_in = 0;
-+
-+	/* Initialize the stream */
-+	stream = ZSTD_initCStream(params, len, workspace->mem,
-+			workspace->size);
-+	if (!stream) {
-+		pr_warn("BTRFS: ZSTD_initCStream failed\n");
-+		ret = -EIO;
-+		goto out;
-+	}
-+
-+	/* map in the first page of input data */
-+	in_page = find_get_page(mapping, start >> PAGE_SHIFT);
-+	in_buf.src = kmap(in_page);
-+	in_buf.pos = 0;
-+	in_buf.size = min_t(size_t, len, PAGE_SIZE);
-+
-+
-+	/* Allocate and map in the output buffer */
-+	out_page = alloc_page(GFP_NOFS | __GFP_HIGHMEM);
-+	if (out_page == NULL) {
-+		ret = -ENOMEM;
-+		goto out;
-+	}
-+	pages[nr_pages++] = out_page;
-+	out_buf.dst = kmap(out_page);
-+	out_buf.pos = 0;
-+	out_buf.size = min_t(size_t, max_out, PAGE_SIZE);
-+
-+	while (1) {
-+		size_t ret2;
-+
-+		ret2 = ZSTD_compressStream(stream, &out_buf, &in_buf);
-+		if (ZSTD_isError(ret2)) {
-+			pr_debug("BTRFS: ZSTD_compressStream returned %d\n",
-+					ZSTD_getErrorCode(ret2));
-+			ret = -EIO;
-+			goto out;
-+		}
-+
-+		/* Check to see if we are making it bigger */
-+		if (tot_in + in_buf.pos > 8192 &&
-+				tot_in + in_buf.pos <
-+				tot_out + out_buf.pos) {
-+			ret = -E2BIG;
-+			goto out;
-+		}
-+
-+		/* We've reached the end of our output range */
-+		if (out_buf.pos >= max_out) {
-+			tot_out += out_buf.pos;
-+			ret = -E2BIG;
-+			goto out;
-+		}
-+
-+		/* Check if we need more output space */
-+		if (out_buf.pos == out_buf.size) {
-+			tot_out += PAGE_SIZE;
-+			max_out -= PAGE_SIZE;
-+			kunmap(out_page);
-+			if (nr_pages == nr_dest_pages) {
-+				out_page = NULL;
-+				ret = -E2BIG;
-+				goto out;
-+			}
-+			out_page = alloc_page(GFP_NOFS | __GFP_HIGHMEM);
-+			if (out_page == NULL) {
-+				ret = -ENOMEM;
-+				goto out;
-+			}
-+			pages[nr_pages++] = out_page;
-+			out_buf.dst = kmap(out_page);
-+			out_buf.pos = 0;
-+			out_buf.size = min_t(size_t, max_out, PAGE_SIZE);
-+		}
-+
-+		/* We've reached the end of the input */
-+		if (in_buf.pos >= len) {
-+			tot_in += in_buf.pos;
-+			break;
-+		}
-+
-+		/* Check if we need more input */
-+		if (in_buf.pos == in_buf.size) {
-+			tot_in += PAGE_SIZE;
-+			kunmap(in_page);
-+			put_page(in_page);
-+
-+			start += PAGE_SIZE;
-+			len -= PAGE_SIZE;
-+			in_page = find_get_page(mapping, start >> PAGE_SHIFT);
-+			in_buf.src = kmap(in_page);
-+			in_buf.pos = 0;
-+			in_buf.size = min_t(size_t, len, PAGE_SIZE);
-+		}
-+	}
-+	while (1) {
-+		size_t ret2;
-+
-+		ret2 = ZSTD_endStream(stream, &out_buf);
-+		if (ZSTD_isError(ret2)) {
-+			pr_debug("BTRFS: ZSTD_endStream returned %d\n",
-+					ZSTD_getErrorCode(ret2));
-+			ret = -EIO;
-+			goto out;
-+		}
-+		if (ret2 == 0) {
-+			tot_out += out_buf.pos;
-+			break;
-+		}
-+		if (out_buf.pos >= max_out) {
-+			tot_out += out_buf.pos;
-+			ret = -E2BIG;
-+			goto out;
-+		}
-+
-+		tot_out += PAGE_SIZE;
-+		max_out -= PAGE_SIZE;
-+		kunmap(out_page);
-+		if (nr_pages == nr_dest_pages) {
-+			out_page = NULL;
-+			ret = -E2BIG;
-+			goto out;
-+		}
-+		out_page = alloc_page(GFP_NOFS | __GFP_HIGHMEM);
-+		if (out_page == NULL) {
-+			ret = -ENOMEM;
-+			goto out;
-+		}
-+		pages[nr_pages++] = out_page;
-+		out_buf.dst = kmap(out_page);
-+		out_buf.pos = 0;
-+		out_buf.size = min_t(size_t, max_out, PAGE_SIZE);
-+	}
-+
-+	if (tot_out >= tot_in) {
-+		ret = -E2BIG;
-+		goto out;
-+	}
-+
-+	ret = 0;
-+	*total_in = tot_in;
-+	*total_out = tot_out;
-+out:
-+	*out_pages = nr_pages;
-+	/* Cleanup */
-+	if (in_page) {
-+		kunmap(in_page);
-+		put_page(in_page);
-+	}
-+	if (out_page)
-+		kunmap(out_page);
-+	return ret;
-+}
-+
-+static int zstd_decompress_bio(struct list_head *ws, struct compressed_bio *cb)
-+{
-+	struct workspace *workspace = list_entry(ws, struct workspace, list);
-+	struct page **pages_in = cb->compressed_pages;
-+	u64 disk_start = cb->start;
-+	struct bio *orig_bio = cb->orig_bio;
-+	size_t srclen = cb->compressed_len;
-+	ZSTD_DStream *stream;
-+	int ret = 0;
-+	unsigned long page_in_index = 0;
-+	unsigned long total_pages_in = DIV_ROUND_UP(srclen, PAGE_SIZE);
-+	unsigned long buf_start;
-+	unsigned long total_out = 0;
-+	ZSTD_inBuffer in_buf = { NULL, 0, 0 };
-+	ZSTD_outBuffer out_buf = { NULL, 0, 0 };
-+
-+	stream = ZSTD_initDStream(
-+			ZSTD_BTRFS_MAX_INPUT, workspace->mem, workspace->size);
-+	if (!stream) {
-+		pr_debug("BTRFS: ZSTD_initDStream failed\n");
-+		ret = -EIO;
-+		goto done;
-+	}
-+
-+	in_buf.src = kmap(pages_in[page_in_index]);
-+	in_buf.pos = 0;
-+	in_buf.size = min_t(size_t, srclen, PAGE_SIZE);
-+
-+	out_buf.dst = workspace->buf;
-+	out_buf.pos = 0;
-+	out_buf.size = PAGE_SIZE;
-+
-+	while (1) {
-+		size_t ret2;
-+
-+		ret2 = ZSTD_decompressStream(stream, &out_buf, &in_buf);
-+		if (ZSTD_isError(ret2)) {
-+			pr_debug("BTRFS: ZSTD_decompressStream returned %d\n",
-+					ZSTD_getErrorCode(ret2));
-+			ret = -EIO;
-+			goto done;
-+		}
-+		buf_start = total_out;
-+		total_out += out_buf.pos;
-+		out_buf.pos = 0;
-+
-+		ret = btrfs_decompress_buf2page(out_buf.dst, buf_start,
-+				total_out, disk_start, orig_bio);
-+		if (ret == 0)
-+			break;
-+
-+		if (in_buf.pos >= srclen)
-+			break;
-+
-+		/* Check if we've hit the end of a frame */
-+		if (ret2 == 0)
-+			break;
-+
-+		if (in_buf.pos == in_buf.size) {
-+			kunmap(pages_in[page_in_index++]);
-+			if (page_in_index >= total_pages_in) {
-+				in_buf.src = NULL;
-+				ret = -EIO;
-+				goto done;
-+			}
-+			srclen -= PAGE_SIZE;
-+			in_buf.src = kmap(pages_in[page_in_index]);
-+			in_buf.pos = 0;
-+			in_buf.size = min_t(size_t, srclen, PAGE_SIZE);
-+		}
-+	}
-+	ret = 0;
-+	zero_fill_bio(orig_bio);
-+done:
-+	if (in_buf.src)
-+		kunmap(pages_in[page_in_index]);
-+	return ret;
-+}
-+
-+static int zstd_decompress(struct list_head *ws, unsigned char *data_in,
-+		struct page *dest_page,
-+		unsigned long start_byte,
-+		size_t srclen, size_t destlen)
-+{
-+	struct workspace *workspace = list_entry(ws, struct workspace, list);
-+	ZSTD_DStream *stream;
-+	int ret = 0;
-+	size_t ret2;
-+	ZSTD_inBuffer in_buf = { NULL, 0, 0 };
-+	ZSTD_outBuffer out_buf = { NULL, 0, 0 };
-+	unsigned long total_out = 0;
-+	unsigned long pg_offset = 0;
-+	char *kaddr;
-+
-+	stream = ZSTD_initDStream(
-+			ZSTD_BTRFS_MAX_INPUT, workspace->mem, workspace->size);
-+	if (!stream) {
-+		pr_warn("BTRFS: ZSTD_initDStream failed\n");
-+		ret = -EIO;
-+		goto finish;
-+	}
-+
-+	destlen = min_t(size_t, destlen, PAGE_SIZE);
-+
-+	in_buf.src = data_in;
-+	in_buf.pos = 0;
-+	in_buf.size = srclen;
-+
-+	out_buf.dst = workspace->buf;
-+	out_buf.pos = 0;
-+	out_buf.size = PAGE_SIZE;
-+
-+	ret2 = 1;
-+	while (pg_offset < destlen && in_buf.pos < in_buf.size) {
-+		unsigned long buf_start;
-+		unsigned long buf_offset;
-+		unsigned long bytes;
-+
-+		/* Check if the frame is over and we still need more input */
-+		if (ret2 == 0) {
-+			pr_debug("BTRFS: ZSTD_decompressStream ended early\n");
-+			ret = -EIO;
-+			goto finish;
-+		}
-+		ret2 = ZSTD_decompressStream(stream, &out_buf, &in_buf);
-+		if (ZSTD_isError(ret2)) {
-+			pr_debug("BTRFS: ZSTD_decompressStream returned %d\n",
-+					ZSTD_getErrorCode(ret2));
-+			ret = -EIO;
-+			goto finish;
-+		}
-+
-+		buf_start = total_out;
-+		total_out += out_buf.pos;
-+		out_buf.pos = 0;
-+
-+		if (total_out <= start_byte)
-+			continue;
-+
-+		if (total_out > start_byte && buf_start < start_byte)
-+			buf_offset = start_byte - buf_start;
-+		else
-+			buf_offset = 0;
-+
-+		bytes = min_t(unsigned long, destlen - pg_offset,
-+				out_buf.size - buf_offset);
-+
-+		kaddr = kmap_atomic(dest_page);
-+		memcpy(kaddr + pg_offset, out_buf.dst + buf_offset, bytes);
-+		kunmap_atomic(kaddr);
-+
-+		pg_offset += bytes;
-+	}
-+	ret = 0;
-+finish:
-+	if (pg_offset < destlen) {
-+		kaddr = kmap_atomic(dest_page);
-+		memset(kaddr + pg_offset, 0, destlen - pg_offset);
-+		kunmap_atomic(kaddr);
-+	}
-+	return ret;
-+}
-+
-+const struct btrfs_compress_op btrfs_zstd_compress = {
-+	.alloc_workspace = zstd_alloc_workspace,
-+	.free_workspace = zstd_free_workspace,
-+	.compress_pages = zstd_compress_pages,
-+	.decompress_bio = zstd_decompress_bio,
-+	.decompress = zstd_decompress,
-+};
-diff --git a/include/uapi/linux/btrfs.h b/include/uapi/linux/btrfs.h
-index 9aa74f3..378230c 100644
---- a/include/uapi/linux/btrfs.h
-+++ b/include/uapi/linux/btrfs.h
-@@ -255,13 +255,7 @@ struct btrfs_ioctl_fs_info_args {
- #define BTRFS_FEATURE_INCOMPAT_DEFAULT_SUBVOL	(1ULL << 1)
- #define BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS	(1ULL << 2)
- #define BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO	(1ULL << 3)
--/*
-- * some patches floated around with a second compression method
-- * lets save that incompat here for when they do get in
-- * Note we don't actually support it, we're just reserving the
-- * number
-- */
--#define BTRFS_FEATURE_INCOMPAT_COMPRESS_LZOv2	(1ULL << 4)
-+#define BTRFS_FEATURE_INCOMPAT_COMPRESS_ZSTD	(1ULL << 4)
-
- /*
-  * older kernels tried to do bigger metadata blocks, but the
---
-2.9.3
diff --git a/0003-kbuild-LLVMLinux-Add-support-for-generating-LLVM-bit.patch b/0003-kbuild-LLVMLinux-Add-support-for-generating-LLVM-bit.patch
deleted file mode 100644
index f84f1f2..0000000
--- a/0003-kbuild-LLVMLinux-Add-support-for-generating-LLVM-bit.patch
+++ /dev/null
@@ -1,94 +0,0 @@
-From 282c44dbde5c83338bd94bffe8a17390cc22b6fa Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?= <bero@linaro.org>
-Date: Thu, 23 Feb 2017 00:08:21 +0100
-Subject: [PATCH 03/30] kbuild, LLVMLinux: Add support for generating LLVM
- bitcode files
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-Add rules to kbuild in order to generate LLVM bitcode files with the .ll
-extension when using clang.
-
-  # from c code
-  CC=clang make kernel/pid.ll
-
-  # from asm code
-  CC=clang make arch/x86/kernel/preempt.ll
-
-Author: Vinicius Tinti <viniciustinti@gmail.com>
-Forward-ported-by: Bernhard Rosenkränzer <bero@linaro.org>
-Signed-off-by: Bernhard Rosenkränzer <bero@linaro.org>
-Signed-off-by: Behan Webster <behanw@converseincode.com>
----
- .gitignore             |  1 +
- Makefile               |  8 +++++++-
- scripts/Makefile.build | 14 ++++++++++++++
- 3 files changed, 22 insertions(+), 1 deletion(-)
-
-diff --git a/.gitignore b/.gitignore
-index c2ed4ecb0acd..0c39aa20b6ba 100644
---- a/.gitignore
-+++ b/.gitignore
-@@ -33,6 +33,7 @@
- *.lzo
- *.patch
- *.gcno
-+*.ll
- modules.builtin
- Module.symvers
- *.dwo
-diff --git a/Makefile b/Makefile
-index e3e8522f0322..0e76a1ea9cae 100644
---- a/Makefile
-+++ b/Makefile
-@@ -1377,7 +1377,9 @@ help:
- 	@echo  '  firmware_install- Install all firmware to INSTALL_FW_PATH'
- 	@echo  '                    (default: $$(INSTALL_MOD_PATH)/lib/firmware)'
- 	@echo  '  dir/            - Build all files in dir and below'
--	@echo  '  dir/file.[ois]  - Build specified target only'
-+	@echo  '  dir/file.[oisS] - Build specified target only'
-+	@echo  '  dir/file.ll     - Build the LLVM bitcode file'
-+	@echo  '                    (requires compiler support for LLVM bitcode generation)'
- 	@echo  '  dir/file.lst    - Build specified mixed source/assembly target only'
- 	@echo  '                    (requires a recent binutils and recent build (System.map))'
- 	@echo  '  dir/file.ko     - Build module including final link'
-@@ -1665,6 +1667,10 @@ endif
- 	$(Q)$(MAKE) $(build)=$(build-dir) $(target-dir)$(notdir $@)
- %.symtypes: %.c prepare scripts FORCE
- 	$(Q)$(MAKE) $(build)=$(build-dir) $(target-dir)$(notdir $@)
-+%.ll: %.c prepare scripts FORCE
-+	$(Q)$(MAKE) $(build)=$(build-dir) $(target-dir)$(notdir $@)
-+%.ll: %.S prepare scripts FORCE
-+	$(Q)$(MAKE) $(build)=$(build-dir) $(target-dir)$(notdir $@)
- 
- # Modules
- /: prepare scripts FORCE
-diff --git a/scripts/Makefile.build b/scripts/Makefile.build
-index d883116ebaa4..e5a28da2e6fa 100644
---- a/scripts/Makefile.build
-+++ b/scripts/Makefile.build
-@@ -177,6 +177,20 @@ cmd_cc_symtypes_c =                                                         \
- $(obj)/%.symtypes : $(src)/%.c FORCE
- 	$(call cmd,cc_symtypes_c)
- 
-+# LLVM bitcode
-+# Generate .ll files from .s and .c
-+quiet_cmd_cc_ll_c = CC $(quiet_modtag)  $@
-+      cmd_cc_ll_c = $(CC) $(c_flags) -emit-llvm -S -o $@ $<
-+
-+$(obj)/%.ll: $(src)/%.c FORCE
-+	$(call if_changed_dep,cc_ll_c)
-+
-+quiet_cmd_as_ll_S = CPP $(quiet_modtag) $@
-+      cmd_as_ll_S = $(CPP) $(a_flags)   -o $@ $<
-+
-+$(obj)/%.ll: $(src)/%.S FORCE
-+	$(call if_changed_dep,as_ll_S)
-+
- # C (.c) files
- # The C file is compiled and updated dependency information is generated.
- # (See cmd_cc_o_c + relevant part of rule_cc_o_c)
--- 
-2.11.0
-
diff --git a/0004-kbuild-LLVMLinux-Make-asm-offset-generation-work-wit.patch b/0004-kbuild-LLVMLinux-Make-asm-offset-generation-work-wit.patch
deleted file mode 100644
index ac4b7b1..0000000
--- a/0004-kbuild-LLVMLinux-Make-asm-offset-generation-work-wit.patch
+++ /dev/null
@@ -1,121 +0,0 @@
-From d2699a2c0021c7837bc871a34b82419a0e178fa2 Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?= <bero@linaro.org>
-Date: Thu, 23 Feb 2017 00:18:44 +0100
-Subject: [PATCH 04/30] kbuild, LLVMLinux: Make asm-offset generation work with
- clang
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-When using clang with -fno-integerated-as, clang will use the gnu assembler
-instead of the integrated assembler. However clang will still perform asm
-error checking before sending the inline assembly language to gas.
-
-The generation of asm-offsets from within C code is dependent on gcc's blind
-passing of whatever is in asm() through to gas. Arbirary text is
-passed through which is then modified by a sed script into the appropriate .h
-and .S code. Since the arbitrary text is not valid assembly language, clang fails.
-
-This can be fixed by making the arbitrary text into an ASM comment and then
-updating the sed scripts accordingly to work as expected.
-
-This solution works for both gcc and clang.
-
-Signed-off-by: Behan Webster <behanw@converseincode.com>
-Forward-ported-by: Bernhard Rosenkränzer <bero@linaro.org>
-Signed-off-by: Bernhard Rosenkränzer <bero@linaro.org>
-Reviewed-by: Mark Charlebois <charlebm@gmail.com>
-Reviewed-by: Jan-Simon Möller <dl9pf@gmx.de>
-Cc: Jan Moskyto Matejka <mq@suse.cz>
-Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
-Cc: "H. Peter Anvin" <hpa@linux.intel.com>
-Cc: Tom Gundersen <teg@jklm.no>
-Cc: Masahiro Yamada <yamada.m@jp.panasonic.com>
-Cc: Arnd Bergmann <arnd@arndb.de>
----
- Kbuild                 | 25 +++++++++++++++++++++++++
- include/linux/kbuild.h |  6 +++---
- scripts/mod/Makefile   |  8 ++++----
- 3 files changed, 32 insertions(+), 7 deletions(-)
-
-diff --git a/Kbuild b/Kbuild
-index 3d0ae152af7c..0a4022da6b6e 100644
---- a/Kbuild
-+++ b/Kbuild
-@@ -75,6 +75,31 @@ offsets-file := include/generated/asm-offsets.h
- always  += $(offsets-file)
- targets += arch/$(SRCARCH)/kernel/asm-offsets.s
- 
-+# Default sed regexp - multiline due to syntax constraints
-+define sed-y
-+	"/^@->/{s:@->#\(.*\):/* \1 */:; \
-+	s:^@->\([^ ]*\) [\$$#]*\([-0-9]*\) \(.*\):#define \1 \2 /* \3 */:; \
-+	s:^@->\([^ ]*\) [\$$#]*\([^ ]*\) \(.*\):#define \1 \2 /* \3 */:; \
-+	s:@->::; p;}"
-+endef
-+
-+quiet_cmd_offsets = GEN     $@
-+define cmd_offsets
-+	(set -e; \
-+	 echo "#ifndef __ASM_OFFSETS_H__"; \
-+	 echo "#define __ASM_OFFSETS_H__"; \
-+	 echo "/*"; \
-+	 echo " * DO NOT MODIFY."; \
-+	 echo " *"; \
-+	 echo " * This file was generated by Kbuild"; \
-+	 echo " *"; \
-+	 echo " */"; \
-+	 echo ""; \
-+	 sed -ne $(sed-y) $<; \
-+	 echo ""; \
-+	 echo "#endif" ) > $@
-+endef
-+
- # We use internal kbuild rules to avoid the "is up to date" message from make
- arch/$(SRCARCH)/kernel/asm-offsets.s: arch/$(SRCARCH)/kernel/asm-offsets.c \
-                                       $(obj)/$(timeconst-file) $(obj)/$(bounds-file) FORCE
-diff --git a/include/linux/kbuild.h b/include/linux/kbuild.h
-index 22a72198c14b..75fa2c3e0e1d 100644
---- a/include/linux/kbuild.h
-+++ b/include/linux/kbuild.h
-@@ -2,14 +2,14 @@
- #define __LINUX_KBUILD_H
- 
- #define DEFINE(sym, val) \
--        asm volatile("\n->" #sym " %0 " #val : : "i" (val))
-+	asm volatile("\n@->" #sym " %0 " #val : : "i" (val))
- 
--#define BLANK() asm volatile("\n->" : : )
-+#define BLANK() asm volatile("\n@->" : : )
- 
- #define OFFSET(sym, str, mem) \
- 	DEFINE(sym, offsetof(struct str, mem))
- 
- #define COMMENT(x) \
--	asm volatile("\n->#" x)
-+	asm volatile("\n@->#" x)
- 
- #endif
-diff --git a/scripts/mod/Makefile b/scripts/mod/Makefile
-index 19d9bcadc0cc..c3db2fa31052 100644
---- a/scripts/mod/Makefile
-+++ b/scripts/mod/Makefile
-@@ -8,10 +8,10 @@ modpost-objs	:= modpost.o file2alias.o sumversion.o
- devicetable-offsets-file := devicetable-offsets.h
- 
- define sed-y
--	"/^->/{s:->#\(.*\):/* \1 */:; \
--	s:^->\([^ ]*\) [\$$#]*\([-0-9]*\) \(.*\):#define \1 \2 /* \3 */:; \
--	s:^->\([^ ]*\) [\$$#]*\([^ ]*\) \(.*\):#define \1 \2 /* \3 */:; \
--	s:->::; p;}"
-+	"/^@->/{s:@->#\(.*\):/* \1 */:; \
-+	s:^@->\([^ ]*\) [\$$#]*\([-0-9]*\) \(.*\):#define \1 \2 /* \3 */:; \
-+	s:^@->\([^ ]*\) [\$$#]*\([^ ]*\) \(.*\):#define \1 \2 /* \3 */:; \
-+	s:@->::; p;}"
- endef
- 
- quiet_cmd_offsets = GEN     $@
--- 
-2.11.0
-
diff --git a/0004-squashfs-Add-zstd-support.patch b/0004-squashfs-Add-zstd-support.patch
deleted file mode 100644
index 36cdf71..0000000
--- a/0004-squashfs-Add-zstd-support.patch
+++ /dev/null
@@ -1,306 +0,0 @@
-From 46bf8f6d30d6ddf2446c110f122482b5e5e16933 Mon Sep 17 00:00:00 2001
-From: Sean Purcell <me@seanp.xyz>
-Date: Mon, 17 Jul 2017 17:08:59 -0700
-Subject: [PATCH v5 4/5] squashfs: Add zstd support
-
-Add zstd compression and decompression support to SquashFS. zstd is a
-great fit for SquashFS because it can compress at ratios approaching xz,
-while decompressing twice as fast as zlib. For SquashFS in particular,
-it can decompress as fast as lzo and lz4. It also has the flexibility
-to turn down the compression ratio for faster compression times.
-
-The compression benchmark is run on the file tree from the SquashFS archive
-found in ubuntu-16.10-desktop-amd64.iso [1]. It uses `mksquashfs` with the
-default block size (128 KB) and and various compression algorithms/levels.
-xz and zstd are also benchmarked with 256 KB blocks. The decompression
-benchmark times how long it takes to `tar` the file tree into `/dev/null`.
-See the benchmark file in the upstream zstd source repository located under
-`contrib/linux-kernel/squashfs-benchmark.sh` [2] for details.
-
-I ran the benchmarks on a Ubuntu 14.04 VM with 2 cores and 4 GiB of RAM.
-The VM is running on a MacBook Pro with a 3.1 GHz Intel Core i7 processor,
-16 GB of RAM, and a SSD.
-
-| Method         | Ratio | Compression MB/s | Decompression MB/s |
-|----------------|-------|------------------|--------------------|
-| gzip           |  2.92 |               15 |                128 |
-| lzo            |  2.64 |              9.5 |                217 |
-| lz4            |  2.12 |               94 |                218 |
-| xz             |  3.43 |              5.5 |                 35 |
-| xz 256 KB      |  3.53 |              5.4 |                 40 |
-| zstd 1         |  2.71 |               96 |                210 |
-| zstd 5         |  2.93 |               69 |                198 |
-| zstd 10        |  3.01 |               41 |                225 |
-| zstd 15        |  3.13 |             11.4 |                224 |
-| zstd 16 256 KB |  3.24 |              8.1 |                210 |
-
-This patch was written by Sean Purcell <me@seanp.xyz>, but I will be
-taking over the submission process.
-
-[1] http://releases.ubuntu.com/16.10/
-[2] https://github.com/facebook/zstd/blob/dev/contrib/linux-kernel/squashfs-benchmark.sh
-
-zstd source repository: https://github.com/facebook/zstd
-
-Signed-off-by: Sean Purcell <me@seanp.xyz>
-Signed-off-by: Nick Terrell <terrelln@fb.com>
----
-v3 -> v4:
-- Fix minor linter warnings
-
-v4 -> v5:
-- Fix ZSTD_DStream initialization code in squashfs
-- Fix patch documentation to reflect that Sean Purcell is the author
-
- fs/squashfs/Kconfig        |  14 +++++
- fs/squashfs/Makefile       |   1 +
- fs/squashfs/decompressor.c |   7 +++
- fs/squashfs/decompressor.h |   4 ++
- fs/squashfs/squashfs_fs.h  |   1 +
- fs/squashfs/zstd_wrapper.c | 151 +++++++++++++++++++++++++++++++++++++++++++++
- 6 files changed, 178 insertions(+)
- create mode 100644 fs/squashfs/zstd_wrapper.c
-
-diff --git a/fs/squashfs/Kconfig b/fs/squashfs/Kconfig
-index ffb093e..1adb334 100644
---- a/fs/squashfs/Kconfig
-+++ b/fs/squashfs/Kconfig
-@@ -165,6 +165,20 @@ config SQUASHFS_XZ
-
- 	  If unsure, say N.
-
-+config SQUASHFS_ZSTD
-+	bool "Include support for ZSTD compressed file systems"
-+	depends on SQUASHFS
-+	select ZSTD_DECOMPRESS
-+	help
-+	  Saying Y here includes support for reading Squashfs file systems
-+	  compressed with ZSTD compression.  ZSTD gives better compression than
-+	  the default ZLIB compression, while using less CPU.
-+
-+	  ZSTD is not the standard compression used in Squashfs and so most
-+	  file systems will be readable without selecting this option.
-+
-+	  If unsure, say N.
-+
- config SQUASHFS_4K_DEVBLK_SIZE
- 	bool "Use 4K device block size?"
- 	depends on SQUASHFS
-diff --git a/fs/squashfs/Makefile b/fs/squashfs/Makefile
-index 246a6f3..6655631 100644
---- a/fs/squashfs/Makefile
-+++ b/fs/squashfs/Makefile
-@@ -15,3 +15,4 @@ squashfs-$(CONFIG_SQUASHFS_LZ4) += lz4_wrapper.o
- squashfs-$(CONFIG_SQUASHFS_LZO) += lzo_wrapper.o
- squashfs-$(CONFIG_SQUASHFS_XZ) += xz_wrapper.o
- squashfs-$(CONFIG_SQUASHFS_ZLIB) += zlib_wrapper.o
-+squashfs-$(CONFIG_SQUASHFS_ZSTD) += zstd_wrapper.o
-diff --git a/fs/squashfs/decompressor.c b/fs/squashfs/decompressor.c
-index d2bc136..8366398 100644
---- a/fs/squashfs/decompressor.c
-+++ b/fs/squashfs/decompressor.c
-@@ -65,6 +65,12 @@ static const struct squashfs_decompressor squashfs_zlib_comp_ops = {
- };
- #endif
-
-+#ifndef CONFIG_SQUASHFS_ZSTD
-+static const struct squashfs_decompressor squashfs_zstd_comp_ops = {
-+	NULL, NULL, NULL, NULL, ZSTD_COMPRESSION, "zstd", 0
-+};
-+#endif
-+
- static const struct squashfs_decompressor squashfs_unknown_comp_ops = {
- 	NULL, NULL, NULL, NULL, 0, "unknown", 0
- };
-@@ -75,6 +81,7 @@ static const struct squashfs_decompressor *decompressor[] = {
- 	&squashfs_lzo_comp_ops,
- 	&squashfs_xz_comp_ops,
- 	&squashfs_lzma_unsupported_comp_ops,
-+	&squashfs_zstd_comp_ops,
- 	&squashfs_unknown_comp_ops
- };
-
-diff --git a/fs/squashfs/decompressor.h b/fs/squashfs/decompressor.h
-index a25713c..0f5a8e4 100644
---- a/fs/squashfs/decompressor.h
-+++ b/fs/squashfs/decompressor.h
-@@ -58,4 +58,8 @@ extern const struct squashfs_decompressor squashfs_lzo_comp_ops;
- extern const struct squashfs_decompressor squashfs_zlib_comp_ops;
- #endif
-
-+#ifdef CONFIG_SQUASHFS_ZSTD
-+extern const struct squashfs_decompressor squashfs_zstd_comp_ops;
-+#endif
-+
- #endif
-diff --git a/fs/squashfs/squashfs_fs.h b/fs/squashfs/squashfs_fs.h
-index 506f4ba..24d12fd 100644
---- a/fs/squashfs/squashfs_fs.h
-+++ b/fs/squashfs/squashfs_fs.h
-@@ -241,6 +241,7 @@ struct meta_index {
- #define LZO_COMPRESSION		3
- #define XZ_COMPRESSION		4
- #define LZ4_COMPRESSION		5
-+#define ZSTD_COMPRESSION	6
-
- struct squashfs_super_block {
- 	__le32			s_magic;
-diff --git a/fs/squashfs/zstd_wrapper.c b/fs/squashfs/zstd_wrapper.c
-new file mode 100644
-index 0000000..eeaabf8
---- /dev/null
-+++ b/fs/squashfs/zstd_wrapper.c
-@@ -0,0 +1,151 @@
-+/*
-+ * Squashfs - a compressed read only filesystem for Linux
-+ *
-+ * Copyright (c) 2016-present, Facebook, Inc.
-+ * All rights reserved.
-+ *
-+ * This program is free software; you can redistribute it and/or
-+ * modify it under the terms of the GNU General Public License
-+ * as published by the Free Software Foundation; either version 2,
-+ * or (at your option) any later version.
-+ *
-+ * This program is distributed in the hope that it will be useful,
-+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
-+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
-+ * GNU General Public License for more details.
-+ *
-+ * zstd_wrapper.c
-+ */
-+
-+#include <linux/mutex.h>
-+#include <linux/buffer_head.h>
-+#include <linux/slab.h>
-+#include <linux/zstd.h>
-+#include <linux/vmalloc.h>
-+
-+#include "squashfs_fs.h"
-+#include "squashfs_fs_sb.h"
-+#include "squashfs.h"
-+#include "decompressor.h"
-+#include "page_actor.h"
-+
-+struct workspace {
-+	void *mem;
-+	size_t mem_size;
-+	size_t window_size;
-+};
-+
-+static void *zstd_init(struct squashfs_sb_info *msblk, void *buff)
-+{
-+	struct workspace *wksp = kmalloc(sizeof(*wksp), GFP_KERNEL);
-+
-+	if (wksp == NULL)
-+		goto failed;
-+	wksp->window_size = max_t(size_t,
-+			msblk->block_size, SQUASHFS_METADATA_SIZE);
-+	wksp->mem_size = ZSTD_DStreamWorkspaceBound(wksp->window_size);
-+	wksp->mem = vmalloc(wksp->mem_size);
-+	if (wksp->mem == NULL)
-+		goto failed;
-+
-+	return wksp;
-+
-+failed:
-+	ERROR("Failed to allocate zstd workspace\n");
-+	kfree(wksp);
-+	return ERR_PTR(-ENOMEM);
-+}
-+
-+
-+static void zstd_free(void *strm)
-+{
-+	struct workspace *wksp = strm;
-+
-+	if (wksp)
-+		vfree(wksp->mem);
-+	kfree(wksp);
-+}
-+
-+
-+static int zstd_uncompress(struct squashfs_sb_info *msblk, void *strm,
-+	struct buffer_head **bh, int b, int offset, int length,
-+	struct squashfs_page_actor *output)
-+{
-+	struct workspace *wksp = strm;
-+	ZSTD_DStream *stream;
-+	size_t total_out = 0;
-+	size_t zstd_err;
-+	int k = 0;
-+	ZSTD_inBuffer in_buf = { NULL, 0, 0 };
-+	ZSTD_outBuffer out_buf = { NULL, 0, 0 };
-+
-+	stream = ZSTD_initDStream(wksp->window_size, wksp->mem, wksp->mem_size);
-+
-+	if (!stream) {
-+		ERROR("Failed to initialize zstd decompressor\n");
-+		goto out;
-+	}
-+
-+	out_buf.size = PAGE_SIZE;
-+	out_buf.dst = squashfs_first_page(output);
-+
-+	do {
-+		if (in_buf.pos == in_buf.size && k < b) {
-+			int avail = min(length, msblk->devblksize - offset);
-+
-+			length -= avail;
-+			in_buf.src = bh[k]->b_data + offset;
-+			in_buf.size = avail;
-+			in_buf.pos = 0;
-+			offset = 0;
-+		}
-+
-+		if (out_buf.pos == out_buf.size) {
-+			out_buf.dst = squashfs_next_page(output);
-+			if (out_buf.dst == NULL) {
-+				/* Shouldn't run out of pages
-+				 * before stream is done.
-+				 */
-+				squashfs_finish_page(output);
-+				goto out;
-+			}
-+			out_buf.pos = 0;
-+			out_buf.size = PAGE_SIZE;
-+		}
-+
-+		total_out -= out_buf.pos;
-+		zstd_err = ZSTD_decompressStream(stream, &out_buf, &in_buf);
-+		total_out += out_buf.pos; /* add the additional data produced */
-+
-+		if (in_buf.pos == in_buf.size && k < b)
-+			put_bh(bh[k++]);
-+	} while (zstd_err != 0 && !ZSTD_isError(zstd_err));
-+
-+	squashfs_finish_page(output);
-+
-+	if (ZSTD_isError(zstd_err)) {
-+		ERROR("zstd decompression error: %d\n",
-+				(int)ZSTD_getErrorCode(zstd_err));
-+		goto out;
-+	}
-+
-+	if (k < b)
-+		goto out;
-+
-+	return (int)total_out;
-+
-+out:
-+	for (; k < b; k++)
-+		put_bh(bh[k]);
-+
-+	return -EIO;
-+}
-+
-+const struct squashfs_decompressor squashfs_zstd_comp_ops = {
-+	.init = zstd_init,
-+	.free = zstd_free,
-+	.decompress = zstd_uncompress,
-+	.id = ZSTD_COMPRESSION,
-+	.name = "zstd",
-+	.supported = 1
-+};
---
-2.9.3
diff --git a/0005-crypto-Add-zstd-support.patch b/0005-crypto-Add-zstd-support.patch
deleted file mode 100644
index 9c29f0b..0000000
--- a/0005-crypto-Add-zstd-support.patch
+++ /dev/null
@@ -1,393 +0,0 @@
---- linux-4.13/crypto/Kconfig.0134~	2017-09-03 22:56:17.000000000 +0200
-+++ linux-4.13/crypto/Kconfig	2017-09-04 14:35:24.681180391 +0200
-@@ -1662,6 +1662,15 @@ config CRYPTO_LZ4HC
- 	help
- 	  This is the LZ4 high compression mode algorithm.
- 
-+config CRYPTO_ZSTD
-+	tristate "Zstd compression algorithm"
-+	select CRYPTO_ALGAPI
-+	select CRYPTO_ACOMP2
-+	select ZSTD_COMPRESS
-+	select ZSTD_DECOMPRESS
-+	help
-+	  This is the zstd algorithm.
-+
- comment "Random Number Generation"
- 
- config CRYPTO_ANSI_CPRNG
---- linux-4.13/crypto/Makefile.0134~	2017-09-04 14:35:24.681180391 +0200
-+++ linux-4.13/crypto/Makefile	2017-09-04 14:38:40.919589876 +0200
-@@ -138,6 +138,7 @@ ecdh_generic-y := ecc.o
- ecdh_generic-y += ecdh.o
- ecdh_generic-y += ecdh_helper.o
- obj-$(CONFIG_CRYPTO_ECDH) += ecdh_generic.o
-+obj-$(CONFIG_CRYPTO_ZSTD) += zstd.o
- 
- #
- # generic algorithms and the async_tx api
---- linux-4.13/crypto/testmgr.c.0134~	2017-09-03 22:56:17.000000000 +0200
-+++ linux-4.13/crypto/testmgr.c	2017-09-04 14:35:24.682180389 +0200
-@@ -3603,6 +3603,16 @@ static const struct alg_test_desc alg_te
- 				.decomp = __VECS(zlib_deflate_decomp_tv_template)
- 			}
- 		}
-+	}, {
-+		.alg = "zstd",
-+		.test = alg_test_comp,
-+		.fips_allowed = 1,
-+		.suite = {
-+			.comp = {
-+				.comp = __VECS(zstd_comp_tv_template),
-+				.decomp = __VECS(zstd_decomp_tv_template)
-+			}
-+		}
- 	}
- };
- 
---- linux-4.13/crypto/testmgr.h.0134~	2017-09-03 22:56:17.000000000 +0200
-+++ linux-4.13/crypto/testmgr.h	2017-09-04 14:35:24.684180383 +0200
-@@ -34638,4 +34638,75 @@ static const struct comp_testvec lz4hc_d
- 	},
- };
- 
-+static const struct comp_testvec zstd_comp_tv_template[] = {
-+	{
-+		.inlen	= 68,
-+		.outlen	= 39,
-+		.input	= "The algorithm is zstd. "
-+			  "The algorithm is zstd. "
-+			  "The algorithm is zstd.",
-+		.output	= "\x28\xb5\x2f\xfd\x00\x50\xf5\x00\x00\xb8\x54\x68\x65"
-+			  "\x20\x61\x6c\x67\x6f\x72\x69\x74\x68\x6d\x20\x69\x73"
-+			  "\x20\x7a\x73\x74\x64\x2e\x20\x01\x00\x55\x73\x36\x01"
-+			  ,
-+	},
-+	{
-+		.inlen	= 244,
-+		.outlen	= 151,
-+		.input	= "zstd, short for Zstandard, is a fast lossless "
-+			  "compression algorithm, targeting real-time "
-+			  "compression scenarios at zlib-level and better "
-+			  "compression ratios. The zstd compression library "
-+			  "provides in-memory compression and decompression "
-+			  "functions.",
-+		.output	= "\x28\xb5\x2f\xfd\x00\x50\x75\x04\x00\x42\x4b\x1e\x17"
-+			  "\x90\x81\x31\x00\xf2\x2f\xe4\x36\xc9\xef\x92\x88\x32"
-+			  "\xc9\xf2\x24\x94\xd8\x68\x9a\x0f\x00\x0c\xc4\x31\x6f"
-+			  "\x0d\x0c\x38\xac\x5c\x48\x03\xcd\x63\x67\xc0\xf3\xad"
-+			  "\x4e\x90\xaa\x78\xa0\xa4\xc5\x99\xda\x2f\xb6\x24\x60"
-+			  "\xe2\x79\x4b\xaa\xb6\x6b\x85\x0b\xc9\xc6\x04\x66\x86"
-+			  "\xe2\xcc\xe2\x25\x3f\x4f\x09\xcd\xb8\x9d\xdb\xc1\x90"
-+			  "\xa9\x11\xbc\x35\x44\x69\x2d\x9c\x64\x4f\x13\x31\x64"
-+			  "\xcc\xfb\x4d\x95\x93\x86\x7f\x33\x7f\x1a\xef\xe9\x30"
-+			  "\xf9\x67\xa1\x94\x0a\x69\x0f\x60\xcd\xc3\xab\x99\xdc"
-+			  "\x42\xed\x97\x05\x00\x33\xc3\x15\x95\x3a\x06\xa0\x0e"
-+			  "\x20\xa9\x0e\x82\xb9\x43\x45\x01",
-+	},
-+};
-+
-+static const struct comp_testvec zstd_decomp_tv_template[] = {
-+	{
-+		.inlen	= 43,
-+		.outlen	= 68,
-+		.input	= "\x28\xb5\x2f\xfd\x04\x50\xf5\x00\x00\xb8\x54\x68\x65"
-+			  "\x20\x61\x6c\x67\x6f\x72\x69\x74\x68\x6d\x20\x69\x73"
-+			  "\x20\x7a\x73\x74\x64\x2e\x20\x01\x00\x55\x73\x36\x01"
-+			  "\x6b\xf4\x13\x35",
-+		.output	= "The algorithm is zstd. "
-+			  "The algorithm is zstd. "
-+			  "The algorithm is zstd.",
-+	},
-+	{
-+		.inlen	= 155,
-+		.outlen	= 244,
-+		.input	= "\x28\xb5\x2f\xfd\x04\x50\x75\x04\x00\x42\x4b\x1e\x17"
-+			  "\x90\x81\x31\x00\xf2\x2f\xe4\x36\xc9\xef\x92\x88\x32"
-+			  "\xc9\xf2\x24\x94\xd8\x68\x9a\x0f\x00\x0c\xc4\x31\x6f"
-+			  "\x0d\x0c\x38\xac\x5c\x48\x03\xcd\x63\x67\xc0\xf3\xad"
-+			  "\x4e\x90\xaa\x78\xa0\xa4\xc5\x99\xda\x2f\xb6\x24\x60"
-+			  "\xe2\x79\x4b\xaa\xb6\x6b\x85\x0b\xc9\xc6\x04\x66\x86"
-+			  "\xe2\xcc\xe2\x25\x3f\x4f\x09\xcd\xb8\x9d\xdb\xc1\x90"
-+			  "\xa9\x11\xbc\x35\x44\x69\x2d\x9c\x64\x4f\x13\x31\x64"
-+			  "\xcc\xfb\x4d\x95\x93\x86\x7f\x33\x7f\x1a\xef\xe9\x30"
-+			  "\xf9\x67\xa1\x94\x0a\x69\x0f\x60\xcd\xc3\xab\x99\xdc"
-+			  "\x42\xed\x97\x05\x00\x33\xc3\x15\x95\x3a\x06\xa0\x0e"
-+			  "\x20\xa9\x0e\x82\xb9\x43\x45\x01\xaa\x6d\xda\x0d",
-+		.output	= "zstd, short for Zstandard, is a fast lossless "
-+			  "compression algorithm, targeting real-time "
-+			  "compression scenarios at zlib-level and better "
-+			  "compression ratios. The zstd compression library "
-+			  "provides in-memory compression and decompression "
-+			  "functions.",
-+	},
-+};
- #endif	/* _CRYPTO_TESTMGR_H */
---- linux-4.13/crypto/zstd.c.0134~	2017-09-04 14:35:24.685180380 +0200
-+++ linux-4.13/crypto/zstd.c	2017-09-04 14:35:24.685180380 +0200
-@@ -0,0 +1,265 @@
-+/*
-+ * Cryptographic API.
-+ *
-+ * Copyright (c) 2017-present, Facebook, Inc.
-+ *
-+ * This program is free software; you can redistribute it and/or modify it
-+ * under the terms of the GNU General Public License version 2 as published by
-+ * the Free Software Foundation.
-+ *
-+ * This program is distributed in the hope that it will be useful, but WITHOUT
-+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
-+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
-+ * more details.
-+ */
-+#include <linux/crypto.h>
-+#include <linux/init.h>
-+#include <linux/interrupt.h>
-+#include <linux/mm.h>
-+#include <linux/module.h>
-+#include <linux/net.h>
-+#include <linux/vmalloc.h>
-+#include <linux/zstd.h>
-+#include <crypto/internal/scompress.h>
-+
-+
-+#define ZSTD_DEF_LEVEL	3
-+
-+struct zstd_ctx {
-+	ZSTD_CCtx *cctx;
-+	ZSTD_DCtx *dctx;
-+	void *cwksp;
-+	void *dwksp;
-+};
-+
-+static ZSTD_parameters zstd_params(void)
-+{
-+	return ZSTD_getParams(ZSTD_DEF_LEVEL, 0, 0);
-+}
-+
-+static int zstd_comp_init(struct zstd_ctx *ctx)
-+{
-+	int ret = 0;
-+	const ZSTD_parameters params = zstd_params();
-+	const size_t wksp_size = ZSTD_CCtxWorkspaceBound(params.cParams);
-+
-+	ctx->cwksp = vzalloc(wksp_size);
-+	if (!ctx->cwksp) {
-+		ret = -ENOMEM;
-+		goto out;
-+	}
-+
-+	ctx->cctx = ZSTD_initCCtx(ctx->cwksp, wksp_size);
-+	if (!ctx->cctx) {
-+		ret = -EINVAL;
-+		goto out_free;
-+	}
-+out:
-+	return ret;
-+out_free:
-+	vfree(ctx->cwksp);
-+	goto out;
-+}
-+
-+static int zstd_decomp_init(struct zstd_ctx *ctx)
-+{
-+	int ret = 0;
-+	const size_t wksp_size = ZSTD_DCtxWorkspaceBound();
-+
-+	ctx->dwksp = vzalloc(wksp_size);
-+	if (!ctx->dwksp) {
-+		ret = -ENOMEM;
-+		goto out;
-+	}
-+
-+	ctx->dctx = ZSTD_initDCtx(ctx->dwksp, wksp_size);
-+	if (!ctx->dctx) {
-+		ret = -EINVAL;
-+		goto out_free;
-+	}
-+out:
-+	return ret;
-+out_free:
-+	vfree(ctx->dwksp);
-+	goto out;
-+}
-+
-+static void zstd_comp_exit(struct zstd_ctx *ctx)
-+{
-+	vfree(ctx->cwksp);
-+	ctx->cwksp = NULL;
-+	ctx->cctx = NULL;
-+}
-+
-+static void zstd_decomp_exit(struct zstd_ctx *ctx)
-+{
-+	vfree(ctx->dwksp);
-+	ctx->dwksp = NULL;
-+	ctx->dctx = NULL;
-+}
-+
-+static int __zstd_init(void *ctx)
-+{
-+	int ret;
-+
-+	ret = zstd_comp_init(ctx);
-+	if (ret)
-+		return ret;
-+	ret = zstd_decomp_init(ctx);
-+	if (ret)
-+		zstd_comp_exit(ctx);
-+	return ret;
-+}
-+
-+static void *zstd_alloc_ctx(struct crypto_scomp *tfm)
-+{
-+	int ret;
-+	struct zstd_ctx *ctx;
-+
-+	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
-+	if (!ctx)
-+		return ERR_PTR(-ENOMEM);
-+
-+	ret = __zstd_init(ctx);
-+	if (ret) {
-+		kfree(ctx);
-+		return ERR_PTR(ret);
-+	}
-+
-+	return ctx;
-+}
-+
-+static int zstd_init(struct crypto_tfm *tfm)
-+{
-+	struct zstd_ctx *ctx = crypto_tfm_ctx(tfm);
-+
-+	return __zstd_init(ctx);
-+}
-+
-+static void __zstd_exit(void *ctx)
-+{
-+	zstd_comp_exit(ctx);
-+	zstd_decomp_exit(ctx);
-+}
-+
-+static void zstd_free_ctx(struct crypto_scomp *tfm, void *ctx)
-+{
-+	__zstd_exit(ctx);
-+	kzfree(ctx);
-+}
-+
-+static void zstd_exit(struct crypto_tfm *tfm)
-+{
-+	struct zstd_ctx *ctx = crypto_tfm_ctx(tfm);
-+
-+	__zstd_exit(ctx);
-+}
-+
-+static int __zstd_compress(const u8 *src, unsigned int slen,
-+			   u8 *dst, unsigned int *dlen, void *ctx)
-+{
-+	size_t out_len;
-+	struct zstd_ctx *zctx = ctx;
-+	const ZSTD_parameters params = zstd_params();
-+
-+	out_len = ZSTD_compressCCtx(zctx->cctx, dst, *dlen, src, slen, params);
-+	if (ZSTD_isError(out_len))
-+		return -EINVAL;
-+	*dlen = out_len;
-+	return 0;
-+}
-+
-+static int zstd_compress(struct crypto_tfm *tfm, const u8 *src,
-+			 unsigned int slen, u8 *dst, unsigned int *dlen)
-+{
-+	struct zstd_ctx *ctx = crypto_tfm_ctx(tfm);
-+
-+	return __zstd_compress(src, slen, dst, dlen, ctx);
-+}
-+
-+static int zstd_scompress(struct crypto_scomp *tfm, const u8 *src,
-+			  unsigned int slen, u8 *dst, unsigned int *dlen,
-+			  void *ctx)
-+{
-+	return __zstd_compress(src, slen, dst, dlen, ctx);
-+}
-+
-+static int __zstd_decompress(const u8 *src, unsigned int slen,
-+			     u8 *dst, unsigned int *dlen, void *ctx)
-+{
-+	size_t out_len;
-+	struct zstd_ctx *zctx = ctx;
-+
-+	out_len = ZSTD_decompressDCtx(zctx->dctx, dst, *dlen, src, slen);
-+	if (ZSTD_isError(out_len))
-+		return -EINVAL;
-+	*dlen = out_len;
-+	return 0;
-+}
-+
-+static int zstd_decompress(struct crypto_tfm *tfm, const u8 *src,
-+			   unsigned int slen, u8 *dst, unsigned int *dlen)
-+{
-+	struct zstd_ctx *ctx = crypto_tfm_ctx(tfm);
-+
-+	return __zstd_decompress(src, slen, dst, dlen, ctx);
-+}
-+
-+static int zstd_sdecompress(struct crypto_scomp *tfm, const u8 *src,
-+			    unsigned int slen, u8 *dst, unsigned int *dlen,
-+			    void *ctx)
-+{
-+	return __zstd_decompress(src, slen, dst, dlen, ctx);
-+}
-+
-+static struct crypto_alg alg = {
-+	.cra_name		= "zstd",
-+	.cra_flags		= CRYPTO_ALG_TYPE_COMPRESS,
-+	.cra_ctxsize		= sizeof(struct zstd_ctx),
-+	.cra_module		= THIS_MODULE,
-+	.cra_init		= zstd_init,
-+	.cra_exit		= zstd_exit,
-+	.cra_u			= { .compress = {
-+	.coa_compress		= zstd_compress,
-+	.coa_decompress		= zstd_decompress } }
-+};
-+
-+static struct scomp_alg scomp = {
-+	.alloc_ctx		= zstd_alloc_ctx,
-+	.free_ctx		= zstd_free_ctx,
-+	.compress		= zstd_scompress,
-+	.decompress		= zstd_sdecompress,
-+	.base			= {
-+		.cra_name	= "zstd",
-+		.cra_driver_name = "zstd-scomp",
-+		.cra_module	 = THIS_MODULE,
-+	}
-+};
-+
-+static int __init zstd_mod_init(void)
-+{
-+	int ret;
-+
-+	ret = crypto_register_alg(&alg);
-+	if (ret)
-+		return ret;
-+
-+	ret = crypto_register_scomp(&scomp);
-+	if (ret)
-+		crypto_unregister_alg(&alg);
-+
-+	return ret;
-+}
-+
-+static void __exit zstd_mod_fini(void)
-+{
-+	crypto_unregister_alg(&alg);
-+	crypto_unregister_scomp(&scomp);
-+}
-+
-+module_init(zstd_mod_init);
-+module_exit(zstd_mod_fini);
-+
-+MODULE_LICENSE("GPL");
-+MODULE_DESCRIPTION("Zstd Compression Algorithm");
-+MODULE_ALIAS_CRYPTO("zstd");
diff --git a/0005-md-sysfs-LLVMLinux-Remove-nested-function-from-bcach.patch b/0005-md-sysfs-LLVMLinux-Remove-nested-function-from-bcach.patch
deleted file mode 100644
index a9b3e20..0000000
--- a/0005-md-sysfs-LLVMLinux-Remove-nested-function-from-bcach.patch
+++ /dev/null
@@ -1,65 +0,0 @@
-From 36def6a1b04411fb3b0ae97871c4308a59d6350a Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?= <bero@linaro.org>
-Date: Thu, 23 Feb 2017 00:20:49 +0100
-Subject: [PATCH 05/30] md, sysfs, LLVMLinux: Remove nested function from
- bcache sysfs
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-Replace the use of nested functions where a normal function will suffice.
-
-Nested functions are not liked by upstream kernel developers in general. Their
-use breaks the use of clang as a compiler, and doesn't make the code any
-better.
-
-This code now works for both gcc and clang.
-
-Signed-off-by: Behan Webster <behanw@converseincode.com>
-Reviewed-by: Mark Charlebois <charlebm@gmail.com>
-Suggested-by: Arnd Bergmann <arnd@arndb.de>
-Cc: Arnd Bergmann <arnd@arndb.de>
-Forward-ported-by: Bernhard Rosenkränzer <bero@linaro.org>
-Signed-off-by: Bernhard Rosenkränzer <bero@linaro.org>
----
- drivers/md/bcache/sysfs.c | 10 ++++++----
- 1 file changed, 6 insertions(+), 4 deletions(-)
-
-diff --git a/drivers/md/bcache/sysfs.c b/drivers/md/bcache/sysfs.c
-index b3ff57d61dde..53d8baa741fb 100644
---- a/drivers/md/bcache/sysfs.c
-+++ b/drivers/md/bcache/sysfs.c
-@@ -731,6 +731,11 @@ static struct attribute *bch_cache_set_internal_files[] = {
- };
- KTYPE(bch_cache_set_internal);
- 
-+static int __bch_cache_cmp(const void *l, const void *r)
-+{
-+	return *((uint16_t *) r) - *((uint16_t *) l);
-+}
-+
- SHOW(__bch_cache)
- {
- 	struct cache *ca = container_of(kobj, struct cache, kobj);
-@@ -755,9 +760,6 @@ SHOW(__bch_cache)
- 					       CACHE_REPLACEMENT(&ca->sb));
- 
- 	if (attr == &sysfs_priority_stats) {
--		int cmp(const void *l, const void *r)
--		{	return *((uint16_t *) r) - *((uint16_t *) l); }
--
- 		struct bucket *b;
- 		size_t n = ca->sb.nbuckets, i;
- 		size_t unused = 0, available = 0, dirty = 0, meta = 0;
-@@ -786,7 +788,7 @@ SHOW(__bch_cache)
- 			p[i] = ca->buckets[i].prio;
- 		mutex_unlock(&ca->set->bucket_lock);
- 
--		sort(p, n, sizeof(uint16_t), cmp, NULL);
-+		sort(p, n, sizeof(uint16_t), __bch_cache_cmp, NULL);
- 
- 		while (n &&
- 		       !cached[n - 1])
--- 
-2.11.0
-
diff --git a/0006-apparmor-LLVMLinux-Remove-VLAIS.patch b/0006-apparmor-LLVMLinux-Remove-VLAIS.patch
deleted file mode 100644
index 9fde18d..0000000
--- a/0006-apparmor-LLVMLinux-Remove-VLAIS.patch
+++ /dev/null
@@ -1,108 +0,0 @@
-From de33aa01ff91fb63a9a22f362135b9aeb56b5ccc Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?= <bero@linaro.org>
-Date: Thu, 23 Feb 2017 00:22:50 +0100
-Subject: [PATCH 06/30] apparmor, LLVMLinux: Remove VLAIS
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-Replaced the use of a Variable Length Array In Struct (VLAIS) with a C99
-compliant equivalent.  This patch instead allocates the appropriate amount of
-memory using a char array using the SHASH_DESC_ON_STACK macro.
-
-The new code can be compiled with both gcc and clang.
-
-Based on an older version by Vinicius Tinti <viniciustinti@gmail.com>
-
-Signed-off-by: Bernhard Rosenkränzer <bero@linaro.org>
-Cc: Vinícius Tinti <viniciustinti@gmail.com>
-Cc: Jan-Simon Möller <dl9pf@gmx.de>
-Cc: Mark Charlebois <charlebm@gmail.com>
-Cc: Behan Webster <behanw@converseincode.com>
----
- security/apparmor/crypto.c | 36 +++++++++++++++++-------------------
- 1 file changed, 17 insertions(+), 19 deletions(-)
-
-diff --git a/security/apparmor/crypto.c b/security/apparmor/crypto.c
-index de8dc78b6144..198c10b2f46c 100644
---- a/security/apparmor/crypto.c
-+++ b/security/apparmor/crypto.c
-@@ -31,10 +31,6 @@ unsigned int aa_hash_size(void)
- 
- char *aa_calc_hash(void *data, size_t len)
- {
--	struct {
--		struct shash_desc shash;
--		char ctx[crypto_shash_descsize(apparmor_tfm)];
--	} desc;
- 	char *hash = NULL;
- 	int error = -ENOMEM;
- 
-@@ -45,16 +41,19 @@ char *aa_calc_hash(void *data, size_t len)
- 	if (!hash)
- 		goto fail;
- 
--	desc.shash.tfm = apparmor_tfm;
--	desc.shash.flags = 0;
-+	SHASH_DESC_ON_STACK(shash, apparmor_tfm);
-+	if (!shash)
-+		goto fail;
-+	shash->tfm = apparmor_tfm;
-+	shash->flags = 0;
- 
--	error = crypto_shash_init(&desc.shash);
-+	error = crypto_shash_init(shash);
- 	if (error)
- 		goto fail;
--	error = crypto_shash_update(&desc.shash, (u8 *) data, len);
-+	error = crypto_shash_update(shash, (u8 *) data, len);
- 	if (error)
- 		goto fail;
--	error = crypto_shash_final(&desc.shash, hash);
-+	error = crypto_shash_final(shash, hash);
- 	if (error)
- 		goto fail;
- 
-@@ -69,10 +68,6 @@ char *aa_calc_hash(void *data, size_t len)
- int aa_calc_profile_hash(struct aa_profile *profile, u32 version, void *start,
- 			 size_t len)
- {
--	struct {
--		struct shash_desc shash;
--		char ctx[crypto_shash_descsize(apparmor_tfm)];
--	} desc;
- 	int error = -ENOMEM;
- 	__le32 le32_version = cpu_to_le32(version);
- 
-@@ -86,19 +81,22 @@ int aa_calc_profile_hash(struct aa_profile *profile, u32 version, void *start,
- 	if (!profile->hash)
- 		goto fail;
- 
--	desc.shash.tfm = apparmor_tfm;
--	desc.shash.flags = 0;
-+	SHASH_DESC_ON_STACK(shash, apparmor_tfm);
-+	if (!shash)
-+		goto fail;
-+	shash->tfm = apparmor_tfm;
-+	shash->flags = 0;
- 
--	error = crypto_shash_init(&desc.shash);
-+	error = crypto_shash_init(shash);
- 	if (error)
- 		goto fail;
--	error = crypto_shash_update(&desc.shash, (u8 *) &le32_version, 4);
-+	error = crypto_shash_update(shash, (u8 *) &le32_version, 4);
- 	if (error)
- 		goto fail;
--	error = crypto_shash_update(&desc.shash, (u8 *) start, len);
-+	error = crypto_shash_update(shash, (u8 *) start, len);
- 	if (error)
- 		goto fail;
--	error = crypto_shash_final(&desc.shash, profile->hash);
-+	error = crypto_shash_final(shash, profile->hash);
- 	if (error)
- 		goto fail;
- 
--- 
-2.11.0
-
diff --git a/0007-exofs-LLVMLinux-Remove-VLAIS-from-exofs-FIXME-Check-.patch b/0007-exofs-LLVMLinux-Remove-VLAIS-from-exofs-FIXME-Check-.patch
deleted file mode 100644
index 9f8fffd..0000000
--- a/0007-exofs-LLVMLinux-Remove-VLAIS-from-exofs-FIXME-Check-.patch
+++ /dev/null
@@ -1,91 +0,0 @@
-From 1cbc1d31554f7e80ab76442c8e5d72007ba4d16b Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?= <bero@linaro.org>
-Date: Thu, 23 Feb 2017 00:23:43 +0100
-Subject: [PATCH 07/30] exofs, LLVMLinux: Remove VLAIS from exofs FIXME Check
- __weak and BROKEN fragment
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-This makes it possible to build exofs with clang.
-
-Signed-off-by: Behan Webster <behanw@converseincode.com>
-Suggested-by: Arnd Bergmann <arnd@arndb.de>
-Cc: Arnd Bergmann <arnd@arndb.de>
-Forward-ported-by: Bernhard Rosenkränzer <bero@linaro.org>
-Signed-off-by: Bernhard Rosenkränzer <bero@linaro.org>
----
- fs/exofs/Kconfig  |  2 +-
- fs/exofs/common.h |  2 +-
- fs/exofs/super.c  | 21 +++++++++------------
- 3 files changed, 11 insertions(+), 14 deletions(-)
-
-diff --git a/fs/exofs/Kconfig b/fs/exofs/Kconfig
-index 86194b2f799d..492746f9e61b 100644
---- a/fs/exofs/Kconfig
-+++ b/fs/exofs/Kconfig
-@@ -1,6 +1,6 @@
- config EXOFS_FS
- 	tristate "exofs: OSD based file system support"
--	depends on SCSI_OSD_ULD
-+	depends on SCSI_OSD_ULD && BROKEN
- 	help
- 	  EXOFS is a file system that uses an OSD storage device,
- 	  as its backing storage.
-diff --git a/fs/exofs/common.h b/fs/exofs/common.h
-index 7d88ef566213..9fdf02333c20 100644
---- a/fs/exofs/common.h
-+++ b/fs/exofs/common.h
-@@ -183,7 +183,7 @@ struct exofs_fcb {
- #define EXOFS_INO_ATTR_SIZE	sizeof(struct exofs_fcb)
- 
- /* This is the Attribute the fcb is stored in */
--static const struct __weak osd_attr g_attr_inode_data = ATTR_DEF(
-+static const struct osd_attr g_attr_inode_data __maybe_unused = ATTR_DEF(
- 	EXOFS_APAGE_FS_DATA,
- 	EXOFS_ATTR_INODE_DATA,
- 	EXOFS_INO_ATTR_SIZE);
-diff --git a/fs/exofs/super.c b/fs/exofs/super.c
-index 1076a4233b39..410572994dad 100644
---- a/fs/exofs/super.c
-+++ b/fs/exofs/super.c
-@@ -546,27 +546,24 @@ static int exofs_devs_2_odi(struct exofs_dt_device_info *dt_dev,
- static int __alloc_dev_table(struct exofs_sb_info *sbi, unsigned numdevs,
- 		      struct exofs_dev **peds)
- {
--	struct __alloc_ore_devs_and_exofs_devs {
--		/* Twice bigger table: See exofs_init_comps() and comment at
--		 * exofs_read_lookup_dev_table()
--		 */
--		struct ore_dev *oreds[numdevs * 2 - 1];
--		struct exofs_dev eds[numdevs];
--	} *aoded;
- 	struct exofs_dev *eds;
- 	unsigned i;
- 
--	aoded = kzalloc(sizeof(*aoded), GFP_KERNEL);
--	if (unlikely(!aoded)) {
-+	/* Twice bigger table: See exofs_init_comps() and comment at
-+	 * exofs_read_lookup_dev_table()
-+	 * XXX: why -1?
-+	 */
-+	sbi->oc.ods = kzalloc(sizeof(struct ore_dev) * (numdevs * 2 - 1) +
-+			      sizeof(struct exofs_dev) * numdevs, GFP_KERNEL);
-+	if (unlikely(!sbi->oc.ods)) {
- 		EXOFS_ERR("ERROR: failed allocating Device array[%d]\n",
- 			  numdevs);
- 		return -ENOMEM;
- 	}
- 
--	sbi->oc.ods = aoded->oreds;
--	*peds = eds = aoded->eds;
-+	*peds = eds = (void *)sbi->oc.ods[numdevs * 2 - 1];
- 	for (i = 0; i < numdevs; ++i)
--		aoded->oreds[i] = &eds[i].ored;
-+		sbi->oc.ods[i] = &eds[i].ored;
- 	return 0;
- }
- 
--- 
-2.11.0
-
diff --git a/0011-Kbuild-LLVMLinux-Use-Oz-instead-of-Os-when-using-cla.patch b/0011-Kbuild-LLVMLinux-Use-Oz-instead-of-Os-when-using-cla.patch
deleted file mode 100644
index 62b4b90..0000000
--- a/0011-Kbuild-LLVMLinux-Use-Oz-instead-of-Os-when-using-cla.patch
+++ /dev/null
@@ -1,35 +0,0 @@
-From 638df6b609af6cfacc3134137c7b513113a1676a Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?= <bero@linaro.org>
-Date: Thu, 23 Feb 2017 00:32:12 +0100
-Subject: [PATCH 11/30] Kbuild, LLVMLinux: Use -Oz instead of -Os when using
- clang
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-This generates smaller resulting object code when compiled with clang.
-
-Signed-off-by: Behan Webster <behanw@converseincode.com>
-Forward-ported-by: Bernhard Rosenkränzer <bero@linaro.org>
-Signed-off-by: Bernhard Rosenkränzer <bero@linaro.org>
----
- Makefile | 3 ++-
- 1 file changed, 2 insertions(+), 1 deletion(-)
-
-diff --git a/Makefile b/Makefile
-index 0e76a1ea9cae..a15d91aebe52 100644
---- a/Makefile
-+++ b/Makefile
-@@ -651,7 +651,8 @@ KBUILD_CFLAGS	+= $(call cc-option,-fdata-sections,)
- endif
- 
- ifdef CONFIG_CC_OPTIMIZE_FOR_SIZE
--KBUILD_CFLAGS	+= -Os $(call cc-disable-warning,maybe-uninitialized,)
-+KBUILD_CFLAGS	+= $(call cc-option,-Oz,-Os)
-+KBUILD_CFLAGS	+= $(call cc-disable-warning,maybe-uninitialized,)
- else
- ifdef CONFIG_PROFILE_ALL_BRANCHES
- KBUILD_CFLAGS	+= -O2 $(call cc-disable-warning,maybe-uninitialized,)
--- 
-2.11.0
-
diff --git a/0012-WORKAROUND-x86-boot-LLVMLinux-Work-around-clang-PR39.patch b/0012-WORKAROUND-x86-boot-LLVMLinux-Work-around-clang-PR39.patch
deleted file mode 100644
index 769ff06..0000000
--- a/0012-WORKAROUND-x86-boot-LLVMLinux-Work-around-clang-PR39.patch
+++ /dev/null
@@ -1,71 +0,0 @@
-From 0b2388ba1b93f5fed9746a6f24a83e02c5548a62 Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?= <bero@linaro.org>
-Date: Thu, 23 Feb 2017 00:33:18 +0100
-Subject: [PATCH 12/30] WORKAROUND x86, boot, LLVMLinux: Work around clang
- PR3997
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-Clang's intrinsics ignore -mregparm=3 when they fall back to calling the
-out-of-line implementations. Putting the args on the stack when memcpy()
-expects them in registers is not a recipe for a happy kernel.
-
-This bites with -m32 too, so clang is presumably catastrophically
-broken for the i386 kernel until this is fixed, unless I'm missing
-something.
-
-For information/testing only; do not apply. With this, I can use
-'clang -m16' to build all the kernel's 16-bit code and get a successful
-boot.
-
-Not-signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
-Forward-ported-by: Jan-Simon Möller <dl9pf@gmx.de>
-Forward-ported-by: Bernhard Rosenkränzer <bero@linaro.org>
----
- arch/x86/boot/memory.c | 7 ++++++-
- arch/x86/boot/string.h | 2 ++
- 2 files changed, 8 insertions(+), 1 deletion(-)
-
-diff --git a/arch/x86/boot/memory.c b/arch/x86/boot/memory.c
-index db75d07c3645..0a61ad084927 100644
---- a/arch/x86/boot/memory.c
-+++ b/arch/x86/boot/memory.c
-@@ -63,8 +63,13 @@ static int detect_memory_e820(void)
- 			count = 0;
- 			break;
- 		}
--
-+#ifdef __clang__
-+		/* PR18415, PR3997 */
-+		memcpy(desc, &buf, sizeof(*desc));
-+		desc++;
-+#else
- 		*desc++ = buf;
-+#endif
- 		count++;
- 	} while (ireg.ebx && count < ARRAY_SIZE(boot_params.e820_map));
- 
-diff --git a/arch/x86/boot/string.h b/arch/x86/boot/string.h
-index 113588ddb43f..d03f4e21a7f3 100644
---- a/arch/x86/boot/string.h
-+++ b/arch/x86/boot/string.h
-@@ -10,6 +10,7 @@ void *memcpy(void *dst, const void *src, size_t len);
- void *memset(void *dst, int c, size_t len);
- int memcmp(const void *s1, const void *s2, size_t len);
- 
-+#ifndef __clang__ /* PR3997, PR18415 */
- /*
-  * Access builtin version by default. If one needs to use optimized version,
-  * do "undef memcpy" in .c file and link against right string.c
-@@ -17,6 +18,7 @@ int memcmp(const void *s1, const void *s2, size_t len);
- #define memcpy(d,s,l) __builtin_memcpy(d,s,l)
- #define memset(d,c,l) __builtin_memset(d,c,l)
- #define memcmp	__builtin_memcmp
-+#endif
- 
- extern int strcmp(const char *str1, const char *str2);
- extern int strncmp(const char *cs, const char *ct, size_t count);
--- 
-2.11.0
-
diff --git a/0013-DO-NOT-UPSTREAM-xen-LLVMLinux-Remove-VLAIS-from-xen-.patch b/0013-DO-NOT-UPSTREAM-xen-LLVMLinux-Remove-VLAIS-from-xen-.patch
deleted file mode 100644
index c78c351..0000000
--- a/0013-DO-NOT-UPSTREAM-xen-LLVMLinux-Remove-VLAIS-from-xen-.patch
+++ /dev/null
@@ -1,76 +0,0 @@
-From e1be0e294634148b3b538d303bc6291b2d6cf801 Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Vin=C3=ADcius=20Tinti?= <viniciustinti@gmail.com>
-Date: Sun, 14 Dec 2014 00:19:00 -0200
-Subject: [PATCH 13/30] DO-NOT-UPSTREAM xen, LLVMLinux: Remove VLAIS from xen
- mmu
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-Signed-off-by: Vinícius Tinti <viniciustinti@gmail.com>
----
- arch/x86/xen/mmu.c | 35 ++++++++++++++++++-----------------
- 1 file changed, 18 insertions(+), 17 deletions(-)
-
-diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c
-index f6740b5b1738..5adf2b65f1fb 100644
---- a/arch/x86/xen/mmu.c
-+++ b/arch/x86/xen/mmu.c
-@@ -1375,36 +1375,37 @@ static void xen_flush_tlb_others(const struct cpumask *cpus,
- 				 struct mm_struct *mm, unsigned long start,
- 				 unsigned long end)
- {
--	struct {
--		struct mmuext_op op;
--#ifdef CONFIG_SMP
--		DECLARE_BITMAP(mask, num_processors);
--#else
--		DECLARE_BITMAP(mask, NR_CPUS);
--#endif
--	} *args;
- 	struct multicall_space mcs;
-+	struct mmuext_op *op;
-+	struct cpumask *mask;
- 
- 	trace_xen_mmu_flush_tlb_others(cpus, mm, start, end);
- 
- 	if (cpumask_empty(cpus))
- 		return;		/* nothing to do */
- 
--	mcs = xen_mc_entry(sizeof(*args));
--	args = mcs.args;
--	args->op.arg2.vcpumask = to_cpumask(args->mask);
-+#ifdef CONFIG_SMP
-+	mcs = xen_mc_entry(sizeof(struct mmuext_op) + BITS_TO_LONGS(num_processors)*sizeof(unsigned long));
-+#else
-+	mcs = xen_mc_entry(sizeof(struct mmuext_op) + BITS_TO_LONGS(NR_CPUS)*sizeof(unsigned long));
-+#endif
-+	/* Extract fields */
-+	op = mcs.args;
-+	mask = to_cpumask(mcs.args + sizeof(struct mmuext_op));
-+
-+	op->arg2.vcpumask = mask;
- 
- 	/* Remove us, and any offline CPUS. */
--	cpumask_and(to_cpumask(args->mask), cpus, cpu_online_mask);
--	cpumask_clear_cpu(smp_processor_id(), to_cpumask(args->mask));
-+	cpumask_and(mask, cpus, cpu_online_mask);
-+	cpumask_clear_cpu(smp_processor_id(), mask);
- 
--	args->op.cmd = MMUEXT_TLB_FLUSH_MULTI;
-+	op->cmd = MMUEXT_TLB_FLUSH_MULTI;
- 	if (end != TLB_FLUSH_ALL && (end - start) <= PAGE_SIZE) {
--		args->op.cmd = MMUEXT_INVLPG_MULTI;
--		args->op.arg1.linear_addr = start;
-+		op->cmd = MMUEXT_INVLPG_MULTI;
-+		op->arg1.linear_addr = start;
- 	}
- 
--	MULTI_mmuext_op(mcs.mc, &args->op, 1, NULL, DOMID_SELF);
-+	MULTI_mmuext_op(mcs.mc, op, 1, NULL, DOMID_SELF);
- 
- 	xen_mc_issue(PARAVIRT_LAZY_MMU);
- }
--- 
-2.11.0
-
diff --git a/0014-DO-NOT-UPSTREAM-arm-LLVMLinux-Provide-__aeabi_-symbo.patch b/0014-DO-NOT-UPSTREAM-arm-LLVMLinux-Provide-__aeabi_-symbo.patch
deleted file mode 100644
index 912d448..0000000
--- a/0014-DO-NOT-UPSTREAM-arm-LLVMLinux-Provide-__aeabi_-symbo.patch
+++ /dev/null
@@ -1,72 +0,0 @@
-From a4e5575ace4bd519270b1ff89a3e0ac069798966 Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?=
- <Bernhard.Rosenkranzer@linaro.org>
-Date: Wed, 30 Mar 2016 00:57:50 +0200
-Subject: [PATCH 14/30] DO-NOT-UPSTREAM arm: LLVMLinux: Provide __aeabi_*
- symbols which are needed for clang
-
-These symbols are required when compiling the Linux kernel for arch ARM with
-clang.
-
-KernelVersion:
-Author: Mark Charlebois <charlebm@gmail.com>
-Signed-off-by: Mark Charlebois <charlebm@gmail.com>
-Not-signed-off-by: Behan Webster <behanw@converseincode.com>
----
- arch/arm/lib/Makefile |  4 ++++
- arch/arm/lib/eabi.c   | 28 ++++++++++++++++++++++++++++
- 2 files changed, 32 insertions(+)
- create mode 100644 arch/arm/lib/eabi.c
-
-diff --git a/arch/arm/lib/Makefile b/arch/arm/lib/Makefile
-index 27f4d96258a2..53f1c85178a8 100644
---- a/arch/arm/lib/Makefile
-+++ b/arch/arm/lib/Makefile
-@@ -18,6 +18,10 @@ lib-y		:= backtrace.o changebit.o csumipv6.o csumpartial.o   \
- mmu-y		:= clear_user.o copy_page.o getuser.o putuser.o       \
- 		   copy_from_user.o copy_to_user.o
- 
-+ifeq ($(COMPILER),clang)
-+lib-y += eabi.o
-+endif
-+
- # using lib_ here won't override already available weak symbols
- obj-$(CONFIG_UACCESS_WITH_MEMCPY) += uaccess_with_memcpy.o
- 
-diff --git a/arch/arm/lib/eabi.c b/arch/arm/lib/eabi.c
-new file mode 100644
-index 000000000000..636a403c9579
---- /dev/null
-+++ b/arch/arm/lib/eabi.c
-@@ -0,0 +1,28 @@
-+/*
-+ * EABI routines
-+ *
-+ *  Copyright (C) 2012  Mark Charlebois
-+ */
-+
-+#include <linux/types.h>
-+#include <linux/string.h>
-+#include <linux/ctype.h>
-+#include <linux/export.h>
-+
-+void __aeabi_memcpy(void *dest, const void *src, size_t n)
-+{
-+	(void)memcpy(dest, src, n);
-+}
-+EXPORT_SYMBOL(__aeabi_memcpy);
-+
-+void __aeabi_memmove(void *dest, const void *src, size_t n)
-+{
-+	(void)memmove(dest, src, n);
-+}
-+EXPORT_SYMBOL(__aeabi_memmove);
-+
-+void __aeabi_memset(void *s, size_t n, int c)
-+{
-+	(void)memset(s, c, n);
-+}
-+EXPORT_SYMBOL(__aeabi_memset);
--- 
-2.11.0
-
diff --git a/0015-DO-NOT-UPSTREAM-arm-firmware-LLVMLinux-replace-naked.patch b/0015-DO-NOT-UPSTREAM-arm-firmware-LLVMLinux-replace-naked.patch
deleted file mode 100644
index ab71d53..0000000
--- a/0015-DO-NOT-UPSTREAM-arm-firmware-LLVMLinux-replace-naked.patch
+++ /dev/null
@@ -1,47 +0,0 @@
-From 5068a19bab6aff3cc63d2cde3e3ffff26e0fe9c5 Mon Sep 17 00:00:00 2001
-From: Behan Webster <behanw@converseincode.com>
-Date: Fri, 30 May 2014 11:35:50 -0700
-Subject: [PATCH 15/30] DO-NOT-UPSTREAM arm, firmware, LLVMLinux: replace naked
- function with hard coded parameters
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-As documented in GCC naked functions should only use Basic asm syntax. The
-Extended asm or mixture of Basic asm and "C" code is not guarantee. Currently
-this works because it was hard coded to follow and check GCC behavior for
-arguments and register placement.
-
-By replacing the naked function with an inline one it is possible to have the
-same result in a more portable and reliable way. Register placement checks
-are not needed since there is no hard code.
-
-Not-signed-off-by: Behan Webster <behanw@converseincode.com>
-Not-signed-off-by: Vinícius Tinti <viniciustinti@gmail.com>
----
- arch/arm/firmware/trusted_foundations.c | 5 +----
- 1 file changed, 1 insertion(+), 4 deletions(-)
-
-diff --git a/arch/arm/firmware/trusted_foundations.c b/arch/arm/firmware/trusted_foundations.c
-index 3fb1b5a1dce9..f319c3fa0fe5 100644
---- a/arch/arm/firmware/trusted_foundations.c
-+++ b/arch/arm/firmware/trusted_foundations.c
-@@ -31,14 +31,11 @@
- 
- static unsigned long cpu_boot_addr;
- 
--static void __naked tf_generic_smc(u32 type, u32 arg1, u32 arg2)
-+static inline void tf_generic_smc(u32 type, u32 arg1, u32 arg2)
- {
- 	asm volatile(
- 		".arch_extension	sec\n\t"
- 		"stmfd	sp!, {r4 - r11, lr}\n\t"
--		__asmeq("%0", "r0")
--		__asmeq("%1", "r1")
--		__asmeq("%2", "r2")
- 		"mov	r3, #0\n\t"
- 		"mov	r4, #0\n\t"
- 		"smc	#0\n\t"
--- 
-2.11.0
-
diff --git a/0016-arm-LLVMLinux-Remove-unreachable-from-naked-function.patch b/0016-arm-LLVMLinux-Remove-unreachable-from-naked-function.patch
deleted file mode 100644
index 93f7a75..0000000
--- a/0016-arm-LLVMLinux-Remove-unreachable-from-naked-function.patch
+++ /dev/null
@@ -1,28 +0,0 @@
-From 21b07dfa5426577f604743948a45481794e7a241 Mon Sep 17 00:00:00 2001
-From: Behan Webster <behanw@converseincode.com>
-Date: Tue, 23 Sep 2014 22:43:30 -0700
-Subject: [PATCH 16/30] arm, LLVMLinux: Remove unreachable from naked function
-
-Signed-off-by: Behan Webster <behanw@converseincode.com>
-Suggested-by: Arnd Bergmann <arnd@arndb.de>
-Cc: Arnd Bergmann <arnd@arndb.de>
----
- drivers/bus/arm-cci.c | 2 +-
- 1 file changed, 1 insertion(+), 1 deletion(-)
-
-diff --git a/drivers/bus/arm-cci.c b/drivers/bus/arm-cci.c
-index c49da15d9790..49c257d572f0 100644
---- a/drivers/bus/arm-cci.c
-+++ b/drivers/bus/arm-cci.c
-@@ -2101,7 +2101,7 @@ asmlinkage void __naked cci_enable_port_for_self(void)
- 	[sizeof_struct_ace_port] "i" (sizeof(struct cci_ace_port)),
- 	[offsetof_port_phys] "i" (offsetof(struct cci_ace_port, phys)) );
- 
--	unreachable();
-+	// unreachable(); // not allowed in naked function
- }
- 
- /**
--- 
-2.11.0
-
diff --git a/0017-MIPS-LLVMLinux-Fix-a-cast-to-type-not-present-in-uni.patch b/0017-MIPS-LLVMLinux-Fix-a-cast-to-type-not-present-in-uni.patch
deleted file mode 100644
index 577cf49..0000000
--- a/0017-MIPS-LLVMLinux-Fix-a-cast-to-type-not-present-in-uni.patch
+++ /dev/null
@@ -1,36 +0,0 @@
-From 0ad5200769177645876d71c2375f4659e49a1bcf Mon Sep 17 00:00:00 2001
-From: Toma Tabacu <toma.tabacu@imgtec.com>
-Date: Mon, 8 Dec 2014 14:55:09 +0000
-Subject: [PATCH 17/30] MIPS: LLVMLinux: Fix a 'cast to type not present in
- union' error.
-
-Replace a 16-bit unsigned int member bitfield with an unsigned short member
-variable in the 'mips16e_instruction' union.
-
-This follows the example set by the 'mips_instruction' union, which doesn't use
-bitfields for its generic union members.
-
-This error gets reported by clang, but not by gcc.
-The changed code can be compiled successfully by both gcc and clang.
-
-Signed-off-by: Toma Tabacu <toma.tabacu@imgtec.com>
----
- arch/mips/include/uapi/asm/inst.h | 2 +-
- 1 file changed, 1 insertion(+), 1 deletion(-)
-
-diff --git a/arch/mips/include/uapi/asm/inst.h b/arch/mips/include/uapi/asm/inst.h
-index 77429d1622b3..248e718180a0 100644
---- a/arch/mips/include/uapi/asm/inst.h
-+++ b/arch/mips/include/uapi/asm/inst.h
-@@ -1071,7 +1071,7 @@ union mips_instruction {
- };
- 
- union mips16e_instruction {
--	unsigned int full : 16;
-+	unsigned short full;
- 	struct m16e_rr rr;
- 	struct m16e_jal jal;
- 	struct m16e_i64 i64;
--- 
-2.11.0
-
diff --git a/0018-MIPS-LLVMLinux-Fix-an-inline-asm-input-output-type-m.patch b/0018-MIPS-LLVMLinux-Fix-an-inline-asm-input-output-type-m.patch
deleted file mode 100644
index 3183812..0000000
--- a/0018-MIPS-LLVMLinux-Fix-an-inline-asm-input-output-type-m.patch
+++ /dev/null
@@ -1,41 +0,0 @@
-From a4bf1a8f23047a1c91ca2d3db64892af92c2436d Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?= <bero@lindev.ch>
-Date: Wed, 1 Feb 2017 16:13:16 +0100
-Subject: [PATCH 18/30] MIPS: LLVMLinux: Fix an 'inline asm input/output type
- mismatch' error.
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-Change the type of csum_ipv6_magic's 'proto' argument from unsigned short to
-__u32.
-
-This fixes a type mismatch between the 'htonl(proto)' inline asm input, which
-is __u32, and the 'proto' output, which is unsigned short.
-
-This error gets reported by clang, but not by gcc.
-The changed code can be compiled successfully by both gcc and clang.
-
-Signed-off-by: Toma Tabacu <toma.tabacu@imgtec.com>
-Forward-ported-by: Bernhard Rosenkränzer <bero@linaro.org>
-Signed-off-by: Bernhard Rosenkränzer <bero@linaro.org>
----
- arch/mips/include/asm/checksum.h | 2 +-
- 1 file changed, 1 insertion(+), 1 deletion(-)
-
-diff --git a/arch/mips/include/asm/checksum.h b/arch/mips/include/asm/checksum.h
-index c8b574f7e0cc..f963fff3be29 100644
---- a/arch/mips/include/asm/checksum.h
-+++ b/arch/mips/include/asm/checksum.h
-@@ -217,7 +217,7 @@ static inline __sum16 ip_compute_csum(const void *buff, int len)
- #define _HAVE_ARCH_IPV6_CSUM
- static __inline__ __sum16 csum_ipv6_magic(const struct in6_addr *saddr,
- 					  const struct in6_addr *daddr,
--					  __u32 len, __u8 proto,
-+					  __u32 len, __u32 proto,
- 					  __wsum sum)
- {
- 	__wsum tmp;
--- 
-2.11.0
-
diff --git a/0019-MIPS-LLVMLinux-Silence-variable-self-assignment-warn.patch b/0019-MIPS-LLVMLinux-Silence-variable-self-assignment-warn.patch
deleted file mode 100644
index 89b8430..0000000
--- a/0019-MIPS-LLVMLinux-Silence-variable-self-assignment-warn.patch
+++ /dev/null
@@ -1,123 +0,0 @@
-From 8bd652c08ac8662f3ce461c99517fab21b75a1cd Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?=
- <Bernhard.Rosenkranzer@linaro.org>
-Date: Wed, 30 Mar 2016 01:02:19 +0200
-Subject: [PATCH 19/30] MIPS: LLVMLinux: Silence variable self-assignment
- warnings.
-
-Replace variable self-assignments with explanatory comments.
-
-This silences a bunch of -Wself-assign warnings reported by clang.
-The changed code can be compiled without warnings by both gcc and clang.
-
-Signed-off-by: Toma Tabacu <toma.tabacu@imgtec.com>
----
- arch/mips/math-emu/dp_add.c | 5 +++++
- arch/mips/math-emu/dp_sub.c | 5 +++++
- arch/mips/math-emu/sp_add.c | 5 +++++
- arch/mips/math-emu/sp_sub.c | 5 +++++
- 4 files changed, 20 insertions(+)
-
-diff --git a/arch/mips/math-emu/dp_add.c b/arch/mips/math-emu/dp_add.c
-index 8954ef031f84..34081484db6c 100644
---- a/arch/mips/math-emu/dp_add.c
-+++ b/arch/mips/math-emu/dp_add.c
-@@ -151,6 +151,8 @@ union ieee754dp ieee754dp_add(union ieee754dp x, union ieee754dp y)
- 		 * leaving result in xm, xs and xe.
- 		 */
- 		xm = xm + ym;
-+		/* xe does not change */
-+		/* xs does not change */
- 
- 		if (xm >> (DP_FBITS + 1 + 3)) { /* carry out */
- 			xm = XDPSRS1(xm);
-@@ -159,8 +161,11 @@ union ieee754dp ieee754dp_add(union ieee754dp x, union ieee754dp y)
- 	} else {
- 		if (xm >= ym) {
- 			xm = xm - ym;
-+			/* xe does not change */
-+			/* xs does not change */
- 		} else {
- 			xm = ym - xm;
-+			/* xe does not change */
- 			xs = ys;
- 		}
- 		if (xm == 0)
-diff --git a/arch/mips/math-emu/dp_sub.c b/arch/mips/math-emu/dp_sub.c
-index fc17a781b9ae..22fac9097e7f 100644
---- a/arch/mips/math-emu/dp_sub.c
-+++ b/arch/mips/math-emu/dp_sub.c
-@@ -154,6 +154,8 @@ union ieee754dp ieee754dp_sub(union ieee754dp x, union ieee754dp y)
- 		/* generate 28 bit result of adding two 27 bit numbers
- 		 */
- 		xm = xm + ym;
-+		/* xe does not change */
-+		/* xs does not change */
- 
- 		if (xm >> (DP_FBITS + 1 + 3)) { /* carry out */
- 			xm = XDPSRS1(xm);	/* shift preserving sticky */
-@@ -162,8 +164,11 @@ union ieee754dp ieee754dp_sub(union ieee754dp x, union ieee754dp y)
- 	} else {
- 		if (xm >= ym) {
- 			xm = xm - ym;
-+			/* xe does not change */
-+			/* xs does not change */
- 		} else {
- 			xm = ym - xm;
-+			/* xe does not change */
- 			xs = ys;
- 		}
- 		if (xm == 0) {
-diff --git a/arch/mips/math-emu/sp_add.c b/arch/mips/math-emu/sp_add.c
-index c55c0c00bca8..91566a8bb8f6 100644
---- a/arch/mips/math-emu/sp_add.c
-+++ b/arch/mips/math-emu/sp_add.c
-@@ -151,6 +151,8 @@ union ieee754sp ieee754sp_add(union ieee754sp x, union ieee754sp y)
- 		 * leaving result in xm, xs and xe.
- 		 */
- 		xm = xm + ym;
-+		/* xe does not change */
-+		/* xs does not change */
- 
- 		if (xm >> (SP_FBITS + 1 + 3)) { /* carry out */
- 			SPXSRSX1();
-@@ -158,8 +160,11 @@ union ieee754sp ieee754sp_add(union ieee754sp x, union ieee754sp y)
- 	} else {
- 		if (xm >= ym) {
- 			xm = xm - ym;
-+			/* xe does not change */
-+			/* xs does not change */
- 		} else {
- 			xm = ym - xm;
-+			/* xe does not change */
- 			xs = ys;
- 		}
- 		if (xm == 0)
-diff --git a/arch/mips/math-emu/sp_sub.c b/arch/mips/math-emu/sp_sub.c
-index dc998ed47295..0bacd49b9815 100644
---- a/arch/mips/math-emu/sp_sub.c
-+++ b/arch/mips/math-emu/sp_sub.c
-@@ -151,6 +151,8 @@ union ieee754sp ieee754sp_sub(union ieee754sp x, union ieee754sp y)
- 		/* generate 28 bit result of adding two 27 bit numbers
- 		 */
- 		xm = xm + ym;
-+		/* xe does not change */
-+		/* xs does not change */
- 
- 		if (xm >> (SP_FBITS + 1 + 3)) { /* carry out */
- 			SPXSRSX1();	/* shift preserving sticky */
-@@ -158,8 +160,11 @@ union ieee754sp ieee754sp_sub(union ieee754sp x, union ieee754sp y)
- 	} else {
- 		if (xm >= ym) {
- 			xm = xm - ym;
-+			/* xe does not change */
-+			/* xs does not change */
- 		} else {
- 			xm = ym - xm;
-+			/* xe does not change */
- 			xs = ys;
- 		}
- 		if (xm == 0) {
--- 
-2.11.0
-
diff --git a/0020-MIPS-LLVMLinux-Silence-unicode-warnings-when-preproc.patch b/0020-MIPS-LLVMLinux-Silence-unicode-warnings-when-preproc.patch
deleted file mode 100644
index 785878d..0000000
--- a/0020-MIPS-LLVMLinux-Silence-unicode-warnings-when-preproc.patch
+++ /dev/null
@@ -1,44 +0,0 @@
-From 706cb8d4c962002068f26b0acc69642d25c74f76 Mon Sep 17 00:00:00 2001
-From: Toma Tabacu <toma.tabacu@imgtec.com>
-Date: Mon, 8 Dec 2014 15:05:17 +0000
-Subject: [PATCH 20/30] MIPS: LLVMLinux: Silence unicode warnings when
- preprocessing assembly.
-
-Differentiate the 'u' GAS .macro argument from the '\u' C preprocessor unicode
-escape sequence by renaming it to '_u'.
-
-When the 'u' argument is evaluated, we end up writing '\u', which causes
-clang to emit -Wunicode warnings.
-
-This silences a couple of -Wunicode warnings reported by clang.
-The changed code can be preprocessed without warnings by both gcc and clang.
-
-Signed-off-by: Toma Tabacu <toma.tabacu@imgtec.com>
----
- arch/mips/include/asm/asmmacro.h | 8 ++++----
- 1 file changed, 4 insertions(+), 4 deletions(-)
-
-diff --git a/arch/mips/include/asm/asmmacro.h b/arch/mips/include/asm/asmmacro.h
-index 83054f79f72a..6588350df089 100644
---- a/arch/mips/include/asm/asmmacro.h
-+++ b/arch/mips/include/asm/asmmacro.h
-@@ -225,12 +225,12 @@
- 	.word	0x41600021 | (\reg << 16)
- 	.endm
- 
--	.macro	MFTR	rt=0, rd=0, u=0, sel=0
--	 .word	0x41000000 | (\rt << 16) | (\rd << 11) | (\u << 5) | (\sel)
-+	.macro	MFTR	rt=0, rd=0, _u=0, sel=0
-+	 .word	0x41000000 | (\rt << 16) | (\rd << 11) | (\_u << 5) | (\sel)
- 	.endm
- 
--	.macro	MTTR	rt=0, rd=0, u=0, sel=0
--	 .word	0x41800000 | (\rt << 16) | (\rd << 11) | (\u << 5) | (\sel)
-+	.macro	MTTR	rt=0, rd=0, _u=0, sel=0
-+	 .word	0x41800000 | (\rt << 16) | (\rd << 11) | (\_u << 5) | (\sel)
- 	.endm
- 
- #ifdef TOOLCHAIN_SUPPORTS_MSA
--- 
-2.11.0
-
diff --git a/0021-Don-t-use-attributes-error-and-warning-with-clang.patch b/0021-Don-t-use-attributes-error-and-warning-with-clang.patch
deleted file mode 100644
index 99c7108..0000000
--- a/0021-Don-t-use-attributes-error-and-warning-with-clang.patch
+++ /dev/null
@@ -1,30 +0,0 @@
-From e476573e2f1b83f6a7cdf4059f344a4c89a55be4 Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?=
- <Bernhard.Rosenkranzer@linaro.org>
-Date: Tue, 5 Apr 2016 23:30:17 +0200
-Subject: [PATCH 21/30] Don't use attributes error and warning with clang
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-Signed-off-by: Bernhard Rosenkränzer <Bernhard.Rosenkranzer@linaro.org>
----
- include/linux/compiler-gcc.h | 2 +-
- 1 file changed, 1 insertion(+), 1 deletion(-)
-
-diff --git a/include/linux/compiler-gcc.h b/include/linux/compiler-gcc.h
-index 0444b1336268..85bf55a08b3f 100644
---- a/include/linux/compiler-gcc.h
-+++ b/include/linux/compiler-gcc.h
-@@ -181,7 +181,7 @@
- 
- #define __UNIQUE_ID(prefix) __PASTE(__PASTE(__UNIQUE_ID_, prefix), __COUNTER__)
- 
--#ifndef __CHECKER__
-+#if !defined(__CHECKER__) && !defined(__clang__)
- # define __compiletime_warning(message) __attribute__((warning(message)))
- # define __compiletime_error(message) __attribute__((error(message)))
- #endif /* __CHECKER__ */
--- 
-2.11.0
-
diff --git a/0022-Fix-undefined-references-to-acpi_idle_driver-on-aarc.patch b/0022-Fix-undefined-references-to-acpi_idle_driver-on-aarc.patch
deleted file mode 100644
index f8268c8..0000000
--- a/0022-Fix-undefined-references-to-acpi_idle_driver-on-aarc.patch
+++ /dev/null
@@ -1,51 +0,0 @@
-From 61e4d47a66ba32eed9e83257f36e82d795bf4c9e Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?= <bero@linaro.org>
-Date: Thu, 23 Feb 2017 00:41:15 +0100
-Subject: [PATCH 22/30] Fix undefined references to acpi_idle_driver on aarch64
- with clang
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-gcc removes the reference to acpi_idle_driver in processor_driver.c when
-building without CONFIG_ACPI_PROCESSOR_IDLE - it thinks the code is
-unreachable. clang doesn't, creating an undefined reference.
-
-Signed-off-by: Bernhard Rosenkränzer <bero@linaro.org>
----
- drivers/acpi/processor_driver.c | 2 ++
- include/acpi/processor.h        | 2 +-
- 2 files changed, 3 insertions(+), 1 deletion(-)
-
-diff --git a/drivers/acpi/processor_driver.c b/drivers/acpi/processor_driver.c
-index 9d5f0c7ed3f7..aafe719e6bb8 100644
---- a/drivers/acpi/processor_driver.c
-+++ b/drivers/acpi/processor_driver.c
-@@ -239,8 +239,10 @@ static int __acpi_processor_start(struct acpi_device *device)
- 	if (result && !IS_ENABLED(CONFIG_ACPI_CPU_FREQ_PSS))
- 		dev_warn(&device->dev, "CPPC data invalid or not present\n");
- 
-+#ifdef CONFIG_ACPI_PROCESSOR_IDLE
- 	if (!cpuidle_get_driver() || cpuidle_get_driver() == &acpi_idle_driver)
- 		acpi_processor_power_init(pr);
-+#endif
- 
- 	result = acpi_pss_perf_init(pr, device);
- 	if (result)
-diff --git a/include/acpi/processor.h b/include/acpi/processor.h
-index c1ba00fc4888..06848b600926 100644
---- a/include/acpi/processor.h
-+++ b/include/acpi/processor.h
-@@ -387,8 +387,8 @@ static inline void acpi_processor_throttling_init(void) {}
- #endif	/* CONFIG_ACPI_CPU_FREQ_PSS */
- 
- /* in processor_idle.c */
--extern struct cpuidle_driver acpi_idle_driver;
- #ifdef CONFIG_ACPI_PROCESSOR_IDLE
-+extern struct cpuidle_driver acpi_idle_driver;
- int acpi_processor_power_init(struct acpi_processor *pr);
- int acpi_processor_power_exit(struct acpi_processor *pr);
- int acpi_processor_power_state_has_changed(struct acpi_processor *pr);
--- 
-2.11.0
-
diff --git a/0023-HACK-firmware-LLVMLinux-fix-EFI-libstub-with-clang.patch b/0023-HACK-firmware-LLVMLinux-fix-EFI-libstub-with-clang.patch
deleted file mode 100644
index 327a516..0000000
--- a/0023-HACK-firmware-LLVMLinux-fix-EFI-libstub-with-clang.patch
+++ /dev/null
@@ -1,46 +0,0 @@
-From 746104bee67aca342c77887648efc471215dd6d8 Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?= <bero@linaro.org>
-Date: Thu, 23 Feb 2017 00:42:02 +0100
-Subject: [PATCH 23/30] HACK: firmware, LLVMLinux: fix EFI libstub with clang
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-Without any extra guidance, clang will generate libstub with either
-absolute or relative ELF relocations.  Use the right combination of
--fpic and -fno-pic on different files to avoid this.
-
-Signed-off-by: Greg Hackmann <ghackmann@google.com>
-Forward-ported-by: Bernhard Rosenkränzer <bero@linaro.org>
-Signed-off-by: Bernhard Rosenkränzer <bero@linaro.org>
----
- drivers/firmware/efi/libstub/Makefile | 6 ++++++
- 1 file changed, 6 insertions(+)
-
-diff --git a/drivers/firmware/efi/libstub/Makefile b/drivers/firmware/efi/libstub/Makefile
-index f7425960f6a5..abe4d91bcc89 100644
---- a/drivers/firmware/efi/libstub/Makefile
-+++ b/drivers/firmware/efi/libstub/Makefile
-@@ -11,6 +11,9 @@ cflags-$(CONFIG_X86)		+= -m$(BITS) -D__KERNEL__ -O2 \
- 				   -mno-mmx -mno-sse
- 
- cflags-$(CONFIG_ARM64)		:= $(subst -pg,,$(KBUILD_CFLAGS))
-+ifeq ($(cc-name),clang)
-+cflags-$(CONFIG_ARM64)		+= -fpic
-+endif
- cflags-$(CONFIG_ARM)		:= $(subst -pg,,$(KBUILD_CFLAGS)) \
- 				   -fno-builtin -fpic -mno-single-pic-base
- 
-@@ -42,6 +45,9 @@ lib-$(CONFIG_EFI_ARMSTUB)	+= arm-stub.o fdt.o string.o random.o \
- lib-$(CONFIG_ARM)		+= arm32-stub.o
- lib-$(CONFIG_ARM64)		+= arm64-stub.o
- CFLAGS_arm64-stub.o 		:= -DTEXT_OFFSET=$(TEXT_OFFSET)
-+ifeq ($(cc-name),clang)
-+CFLAGS_arm64-stub.o 		+= -fno-pic
-+endif
- 
- #
- # arm64 puts the stub in the kernel proper, which will unnecessarily retain all
--- 
-2.11.0
-
diff --git a/0024-aarch64-crypto-LLVMLinux-Fix-inline-assembly-for-cla.patch b/0024-aarch64-crypto-LLVMLinux-Fix-inline-assembly-for-cla.patch
deleted file mode 100644
index ba06cf6..0000000
--- a/0024-aarch64-crypto-LLVMLinux-Fix-inline-assembly-for-cla.patch
+++ /dev/null
@@ -1,46 +0,0 @@
-From 87257c567c10ef01b28988ef6caf890949f33593 Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?= <bero@linaro.org>
-Date: Thu, 23 Feb 2017 00:43:25 +0100
-Subject: [PATCH 24/30] aarch64: crypto: LLVMLinux: Fix inline assembly for
- clang
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-Fix inline assembly to be compatible with both gcc and clang
-
-Signed-off-by: Bernhard Rosenkränzer <bero@linaro.org>
----
- arch/arm64/crypto/sha1-ce-glue.c | 2 +-
- arch/arm64/crypto/sha2-ce-glue.c | 2 +-
- 2 files changed, 2 insertions(+), 2 deletions(-)
-
-diff --git a/arch/arm64/crypto/sha1-ce-glue.c b/arch/arm64/crypto/sha1-ce-glue.c
-index aefda9868627..c71e94ba0e43 100644
---- a/arch/arm64/crypto/sha1-ce-glue.c
-+++ b/arch/arm64/crypto/sha1-ce-glue.c
-@@ -18,7 +18,7 @@
- #include <linux/module.h>
- 
- #define ASM_EXPORT(sym, val) \
--	asm(".globl " #sym "; .set " #sym ", %0" :: "I"(val));
-+	asm(".globl " #sym "; .set " #sym ", %c0" :: "I"(val));
- 
- MODULE_DESCRIPTION("SHA1 secure hash using ARMv8 Crypto Extensions");
- MODULE_AUTHOR("Ard Biesheuvel <ard.biesheuvel@linaro.org>");
-diff --git a/arch/arm64/crypto/sha2-ce-glue.c b/arch/arm64/crypto/sha2-ce-glue.c
-index 7cd587564a41..381b5fb2dcb2 100644
---- a/arch/arm64/crypto/sha2-ce-glue.c
-+++ b/arch/arm64/crypto/sha2-ce-glue.c
-@@ -18,7 +18,7 @@
- #include <linux/module.h>
- 
- #define ASM_EXPORT(sym, val) \
--	asm(".globl " #sym "; .set " #sym ", %0" :: "I"(val));
-+	asm(".globl " #sym "; .set " #sym ", %c0" :: "I"(val));
- 
- MODULE_DESCRIPTION("SHA-224/SHA-256 secure hash using ARMv8 Crypto Extensions");
- MODULE_AUTHOR("Ard Biesheuvel <ard.biesheuvel@linaro.org>");
--- 
-2.11.0
-
diff --git a/0025-aarch64-LLVMLinux-Make-spin_lock_prefetch-asm-code-c.patch b/0025-aarch64-LLVMLinux-Make-spin_lock_prefetch-asm-code-c.patch
deleted file mode 100644
index 82c1b2c..0000000
--- a/0025-aarch64-LLVMLinux-Make-spin_lock_prefetch-asm-code-c.patch
+++ /dev/null
@@ -1,34 +0,0 @@
-From f0d4542049a4041f34ec8c66a16d02d23a272599 Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?= <bero@linaro.org>
-Date: Thu, 23 Feb 2017 00:44:18 +0100
-Subject: [PATCH 25/30] aarch64: LLVMLinux: Make spin_lock_prefetch asm code
- compatible with clang
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-This version is accepted by both gcc and clang.
-
-Signed-off-by: Bernhard Rosenkränzer <bero@linaro.org>
----
- arch/arm64/include/asm/processor.h | 4 ++--
- 1 file changed, 2 insertions(+), 2 deletions(-)
-
-diff --git a/arch/arm64/include/asm/processor.h b/arch/arm64/include/asm/processor.h
-index c97b8bd2acba..edc9fe9bb7c6 100644
---- a/arch/arm64/include/asm/processor.h
-+++ b/arch/arm64/include/asm/processor.h
-@@ -178,8 +178,8 @@ static inline void prefetchw(const void *ptr)
- static inline void spin_lock_prefetch(const void *ptr)
- {
- 	asm volatile(ARM64_LSE_ATOMIC_INSN(
--		     "prfm pstl1strm, %a0",
--		     "nop") : : "p" (ptr));
-+		     "prfm pstl1strm, [%x0]",
-+		     "nop") : : "r" (ptr));
- }
- 
- #define HAVE_ARCH_PICK_MMAP_LAYOUT
--- 
-2.11.0
-
diff --git a/0026-LLVMLinux-Don-t-use-attribute-externally_visible-whe.patch b/0026-LLVMLinux-Don-t-use-attribute-externally_visible-whe.patch
deleted file mode 100644
index 43548de..0000000
--- a/0026-LLVMLinux-Don-t-use-attribute-externally_visible-whe.patch
+++ /dev/null
@@ -1,33 +0,0 @@
-From d94e33448ec3cd7807600c83e46db3c96084be76 Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?= <bero@linaro.org>
-Date: Thu, 23 Feb 2017 00:45:31 +0100
-Subject: [PATCH 26/30] LLVMLinux: Don't use attribute externally_visible when
- building with clang
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-As of 4.0, clang doesn't support attribute externally_visible (PR16683).
-Don't use the attribute when building with clang.
-
-Signed-off-by: Bernhard Rosenkränzer <bero@linaro.org>
----
- include/linux/compiler-gcc.h | 2 +-
- 1 file changed, 1 insertion(+), 1 deletion(-)
-
-diff --git a/include/linux/compiler-gcc.h b/include/linux/compiler-gcc.h
-index 85bf55a08b3f..44dc519ca5e2 100644
---- a/include/linux/compiler-gcc.h
-+++ b/include/linux/compiler-gcc.h
-@@ -211,7 +211,7 @@
- 
- #endif /* GCC_VERSION >= 40500 */
- 
--#if GCC_VERSION >= 40600
-+#if GCC_VERSION >= 40600 && !defined(__clang__)
- /*
-  * When used with Link Time Optimization, gcc can optimize away C functions or
-  * variables which are referenced only from assembly code.  __visible tells the
--- 
-2.11.0
-
diff --git a/0027-x86-crypto-LLVMLinux-Fix-building-x86_64-AES-extensi.patch b/0027-x86-crypto-LLVMLinux-Fix-building-x86_64-AES-extensi.patch
deleted file mode 100644
index e301ca0..0000000
--- a/0027-x86-crypto-LLVMLinux-Fix-building-x86_64-AES-extensi.patch
+++ /dev/null
@@ -1,32 +0,0 @@
-From 68c9b1821b70a7c58a66be42b0e18cf0a0ede537 Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?= <bero@linaro.org>
-Date: Thu, 23 Feb 2017 00:46:28 +0100
-Subject: [PATCH 27/30] x86, crypto, LLVMLinux: Fix building x86_64 AES
- extensions with clang
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-This is a workaround for PR24494
-
-Signed-off-by: Bernhard Rosenkränzer <bero@linaro.org>
----
- arch/x86/crypto/aes_ctrby8_avx-x86_64.S | 2 +-
- 1 file changed, 1 insertion(+), 1 deletion(-)
-
-diff --git a/arch/x86/crypto/aes_ctrby8_avx-x86_64.S b/arch/x86/crypto/aes_ctrby8_avx-x86_64.S
-index a916c4a61165..7a71553b23dc 100644
---- a/arch/x86/crypto/aes_ctrby8_avx-x86_64.S
-+++ b/arch/x86/crypto/aes_ctrby8_avx-x86_64.S
-@@ -93,7 +93,7 @@
- 
- #define tmp		%r10
- #define	DDQ(i)		CONCAT(ddq_add_,i)
--#define	XMM(i)		CONCAT(%xmm, i)
-+#define	XMM(i)		CONCAT(%xmm,i)
- #define	DDQ_DATA	0
- #define	XDATA		1
- #define KEY_128		1
--- 
-2.11.0
-
diff --git a/0028-x86-LLVMLinux-Qualify-mul-as-mulq-to-make-clang-happ.patch b/0028-x86-LLVMLinux-Qualify-mul-as-mulq-to-make-clang-happ.patch
deleted file mode 100644
index cacd635..0000000
--- a/0028-x86-LLVMLinux-Qualify-mul-as-mulq-to-make-clang-happ.patch
+++ /dev/null
@@ -1,33 +0,0 @@
-From d582ae8c61d52a3a7c3841c87ac48942c2b556ef Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?= <bero@linaro.org>
-Date: Thu, 23 Feb 2017 00:47:09 +0100
-Subject: [PATCH 28/30] x86, LLVMLinux: Qualify mul as mulq to make clang happy
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-Signed-off-by: Bernhard Rosenkränzer <bero@linaro.org>
----
- arch/x86/lib/kaslr.c | 2 +-
- 1 file changed, 1 insertion(+), 1 deletion(-)
-
-diff --git a/arch/x86/lib/kaslr.c b/arch/x86/lib/kaslr.c
-index 121f59c6ee54..9915ff924a5e 100644
---- a/arch/x86/lib/kaslr.c
-+++ b/arch/x86/lib/kaslr.c
-@@ -79,7 +79,11 @@ unsigned long kaslr_get_random_long(const char *purpose)
- 	}
- 
- 	/* Circular multiply for better bit diffusion */
--	asm("mul %3"
-+#ifdef CONFIG_X86_64
-+	asm("mulq %3"
-+#else
-+	asm("mull %3"
-+#endif
- 	    : "=a" (random), "=d" (raw)
- 	    : "a" (random), "rm" (mix_const));
- 	random += raw;
--- 
-2.11.0
-
diff --git a/0029-kbuild-LLVMLinux-Add-Werror-to-cc-option-in-order-to.patch b/0029-kbuild-LLVMLinux-Add-Werror-to-cc-option-in-order-to.patch
deleted file mode 100644
index fc62b87..0000000
--- a/0029-kbuild-LLVMLinux-Add-Werror-to-cc-option-in-order-to.patch
+++ /dev/null
@@ -1,55 +0,0 @@
-From 8de733e8e653892cae10581b816e58f8ad29846e Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?= <bero@linaro.org>
-Date: Thu, 23 Feb 2017 05:09:42 +0100
-Subject: [PATCH 29/30] kbuild, LLVMLinux: Add -Werror to cc-option in order to
- support clang
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-Clang will warn about unknown warnings but will not return false
-unless -Werror is set. GCC will return false if an unknown
-warning is passed.
-
-Adding -Werror make both compiler behave the same.
-
-Signed-off-by: Mark Charlebois <charlebm@gmail.com>
-Signed-off-by: Behan Webster <behanw@converseincode.com>
-Forward-ported-by: Bernhard Rosenkränzer <bero@linaro.org>
-Signed-off-by: Bernhard Rosenkränzer <bero@linaro.org>
-Reviewed-by: Jan-Simon Möller <dl9pf@gmx.de>
----
- scripts/Kbuild.include | 6 +++---
- 1 file changed, 3 insertions(+), 3 deletions(-)
-
-diff --git a/scripts/Kbuild.include b/scripts/Kbuild.include
-index d6ca649cb0e9..a70fd26204de 100644
---- a/scripts/Kbuild.include
-+++ b/scripts/Kbuild.include
-@@ -116,12 +116,12 @@ CC_OPTION_CFLAGS = $(filter-out $(GCC_PLUGINS_CFLAGS),$(KBUILD_CFLAGS))
- # Usage: cflags-y += $(call cc-option,-march=winchip-c6,-march=i586)
- 
- cc-option = $(call try-run,\
--	$(CC) $(KBUILD_CPPFLAGS) $(CC_OPTION_CFLAGS) $(1) -c -x c /dev/null -o "$$TMP",$(1),$(2))
-+	$(CC) -Werror $(KBUILD_CPPFLAGS) $(CC_OPTION_CFLAGS) $(1) -c -x c /dev/null -o "$$TMP",$(1),$(2))
- 
- # cc-option-yn
- # Usage: flag := $(call cc-option-yn,-march=winchip-c6)
- cc-option-yn = $(call try-run,\
--	$(CC) $(KBUILD_CPPFLAGS) $(CC_OPTION_CFLAGS) $(1) -c -x c /dev/null -o "$$TMP",y,n)
-+	$(CC) -Werror $(KBUILD_CPPFLAGS) $(CC_OPTION_CFLAGS) $(1) -c -x c /dev/null -o "$$TMP",y,n)
- 
- # cc-option-align
- # Prefix align with either -falign or -malign
-@@ -131,7 +131,7 @@ cc-option-align = $(subst -functions=0,,\
- # cc-disable-warning
- # Usage: cflags-y += $(call cc-disable-warning,unused-but-set-variable)
- cc-disable-warning = $(call try-run,\
--	$(CC) $(KBUILD_CPPFLAGS) $(CC_OPTION_CFLAGS) -W$(strip $(1)) -c -x c /dev/null -o "$$TMP",-Wno-$(strip $(1)))
-+	$(CC) -Werror $(KBUILD_CPPFLAGS) $(CC_OPTION_CFLAGS) -W$(strip $(1)) -c -x c /dev/null -o "$$TMP",-Wno-$(strip $(1)))
- 
- # cc-name
- # Expands to either gcc or clang
--- 
-2.11.0
-
diff --git a/0030-x86-kbuild-LLVMLinux-Check-for-compiler-support-of-f.patch b/0030-x86-kbuild-LLVMLinux-Check-for-compiler-support-of-f.patch
deleted file mode 100644
index 79fc95a..0000000
--- a/0030-x86-kbuild-LLVMLinux-Check-for-compiler-support-of-f.patch
+++ /dev/null
@@ -1,37 +0,0 @@
-From 5747a302540b1b3b1fe66caf7aa724b8bdc8cc6f Mon Sep 17 00:00:00 2001
-From: =?UTF-8?q?Bernhard=20Rosenkr=C3=A4nzer?= <bero@linaro.org>
-Date: Thu, 23 Feb 2017 15:03:10 +0100
-Subject: [PATCH 30/30] x86, kbuild, LLVMLinux: Check for compiler support of
- -falign-*=1
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-As of 4.0, clang doesn't support -falign-loops=1 -falign-jumps=1.
-Don't add those flags unconditionally.
-
-Signed-off-by: Bernhard Rosenkränzer <bero@linaro.org>
----
- arch/x86/Makefile | 4 ++--
- 1 file changed, 2 insertions(+), 2 deletions(-)
-
-diff --git a/arch/x86/Makefile b/arch/x86/Makefile
-index 2d449337a360..0a62ebada700 100644
---- a/arch/x86/Makefile
-+++ b/arch/x86/Makefile
-@@ -88,10 +88,10 @@ else
-         KBUILD_CFLAGS += -m64
- 
-         # Align jump targets to 1 byte, not the default 16 bytes:
--        KBUILD_CFLAGS += -falign-jumps=1
-+        KBUILD_CFLAGS += $(call cc-option,-falign-jumps=1)
- 
-         # Pack loops tightly as well:
--        KBUILD_CFLAGS += -falign-loops=1
-+        KBUILD_CFLAGS += $(call cc-option,-falign-loops=1)
- 
-         # Don't autogenerate traditional x87 instructions
-         KBUILD_CFLAGS += $(call cc-option,-mno-80387)
--- 
-2.11.0
-
diff --git a/01-revert-6602f080cb28745259e2fab1a4cf55eeb5894f93.patch b/01-revert-6602f080cb28745259e2fab1a4cf55eeb5894f93.patch
new file mode 100644
index 0000000..d527b7c
--- /dev/null
+++ b/01-revert-6602f080cb28745259e2fab1a4cf55eeb5894f93.patch
@@ -0,0 +1,166 @@
+diff --git a/drivers/net/wireless/ath/ath9k/hif_usb.c b/drivers/net/wireless/ath/ath9k/hif_usb.c
+index 4ed21dad6a8e..6049d3766c64 100644
+--- a/drivers/net/wireless/ath/ath9k/hif_usb.c
++++ b/drivers/net/wireless/ath/ath9k/hif_usb.c
+@@ -643,9 +643,9 @@ static void ath9k_hif_usb_rx_stream(struct hif_device_usb *hif_dev,
+ 
+ static void ath9k_hif_usb_rx_cb(struct urb *urb)
+ {
+-	struct rx_buf *rx_buf = (struct rx_buf *)urb->context;
+-	struct hif_device_usb *hif_dev = rx_buf->hif_dev;
+-	struct sk_buff *skb = rx_buf->skb;
++	struct sk_buff *skb = (struct sk_buff *) urb->context;
++	struct hif_device_usb *hif_dev =
++		usb_get_intfdata(usb_ifnum_to_if(urb->dev, 0));
+ 	int ret;
+ 
+ 	if (!skb)
+@@ -685,15 +685,14 @@ static void ath9k_hif_usb_rx_cb(struct urb *urb)
+ 	return;
+ free:
+ 	kfree_skb(skb);
+-	kfree(rx_buf);
+ }
+ 
+ static void ath9k_hif_usb_reg_in_cb(struct urb *urb)
+ {
+-	struct rx_buf *rx_buf = (struct rx_buf *)urb->context;
+-	struct hif_device_usb *hif_dev = rx_buf->hif_dev;
+-	struct sk_buff *skb = rx_buf->skb;
++	struct sk_buff *skb = (struct sk_buff *) urb->context;
+ 	struct sk_buff *nskb;
++	struct hif_device_usb *hif_dev =
++		usb_get_intfdata(usb_ifnum_to_if(urb->dev, 0));
+ 	int ret;
+ 
+ 	if (!skb)
+@@ -751,7 +750,6 @@ static void ath9k_hif_usb_reg_in_cb(struct urb *urb)
+ 	return;
+ free:
+ 	kfree_skb(skb);
+-	kfree(rx_buf);
+ 	urb->context = NULL;
+ }
+ 
+@@ -797,7 +795,7 @@ static int ath9k_hif_usb_alloc_tx_urbs(struct hif_device_usb *hif_dev)
+ 	init_usb_anchor(&hif_dev->mgmt_submitted);
+ 
+ 	for (i = 0; i < MAX_TX_URB_NUM; i++) {
+-		tx_buf = kzalloc(sizeof(*tx_buf), GFP_KERNEL);
++		tx_buf = kzalloc(sizeof(struct tx_buf), GFP_KERNEL);
+ 		if (!tx_buf)
+ 			goto err;
+ 
+@@ -834,9 +832,8 @@ static void ath9k_hif_usb_dealloc_rx_urbs(struct hif_device_usb *hif_dev)
+ 
+ static int ath9k_hif_usb_alloc_rx_urbs(struct hif_device_usb *hif_dev)
+ {
+-	struct rx_buf *rx_buf = NULL;
+-	struct sk_buff *skb = NULL;
+ 	struct urb *urb = NULL;
++	struct sk_buff *skb = NULL;
+ 	int i, ret;
+ 
+ 	init_usb_anchor(&hif_dev->rx_submitted);
+@@ -844,12 +841,6 @@ static int ath9k_hif_usb_alloc_rx_urbs(struct hif_device_usb *hif_dev)
+ 
+ 	for (i = 0; i < MAX_RX_URB_NUM; i++) {
+ 
+-		rx_buf = kzalloc(sizeof(*rx_buf), GFP_KERNEL);
+-		if (!rx_buf) {
+-			ret = -ENOMEM;
+-			goto err_rxb;
+-		}
+-
+ 		/* Allocate URB */
+ 		urb = usb_alloc_urb(0, GFP_KERNEL);
+ 		if (urb == NULL) {
+@@ -864,14 +855,11 @@ static int ath9k_hif_usb_alloc_rx_urbs(struct hif_device_usb *hif_dev)
+ 			goto err_skb;
+ 		}
+ 
+-		rx_buf->hif_dev = hif_dev;
+-		rx_buf->skb = skb;
+-
+ 		usb_fill_bulk_urb(urb, hif_dev->udev,
+ 				  usb_rcvbulkpipe(hif_dev->udev,
+ 						  USB_WLAN_RX_PIPE),
+ 				  skb->data, MAX_RX_BUF_SIZE,
+-				  ath9k_hif_usb_rx_cb, rx_buf);
++				  ath9k_hif_usb_rx_cb, skb);
+ 
+ 		/* Anchor URB */
+ 		usb_anchor_urb(urb, &hif_dev->rx_submitted);
+@@ -897,8 +885,6 @@ static int ath9k_hif_usb_alloc_rx_urbs(struct hif_device_usb *hif_dev)
+ err_skb:
+ 	usb_free_urb(urb);
+ err_urb:
+-	kfree(rx_buf);
+-err_rxb:
+ 	ath9k_hif_usb_dealloc_rx_urbs(hif_dev);
+ 	return ret;
+ }
+@@ -910,21 +896,14 @@ static void ath9k_hif_usb_dealloc_reg_in_urbs(struct hif_device_usb *hif_dev)
+ 
+ static int ath9k_hif_usb_alloc_reg_in_urbs(struct hif_device_usb *hif_dev)
+ {
+-	struct rx_buf *rx_buf = NULL;
+-	struct sk_buff *skb = NULL;
+ 	struct urb *urb = NULL;
++	struct sk_buff *skb = NULL;
+ 	int i, ret;
+ 
+ 	init_usb_anchor(&hif_dev->reg_in_submitted);
+ 
+ 	for (i = 0; i < MAX_REG_IN_URB_NUM; i++) {
+ 
+-		rx_buf = kzalloc(sizeof(*rx_buf), GFP_KERNEL);
+-		if (!rx_buf) {
+-			ret = -ENOMEM;
+-			goto err_rxb;
+-		}
+-
+ 		/* Allocate URB */
+ 		urb = usb_alloc_urb(0, GFP_KERNEL);
+ 		if (urb == NULL) {
+@@ -939,14 +918,11 @@ static int ath9k_hif_usb_alloc_reg_in_urbs(struct hif_device_usb *hif_dev)
+ 			goto err_skb;
+ 		}
+ 
+-		rx_buf->hif_dev = hif_dev;
+-		rx_buf->skb = skb;
+-
+ 		usb_fill_int_urb(urb, hif_dev->udev,
+ 				  usb_rcvintpipe(hif_dev->udev,
+ 						  USB_REG_IN_PIPE),
+ 				  skb->data, MAX_REG_IN_BUF_SIZE,
+-				  ath9k_hif_usb_reg_in_cb, rx_buf, 1);
++				  ath9k_hif_usb_reg_in_cb, skb, 1);
+ 
+ 		/* Anchor URB */
+ 		usb_anchor_urb(urb, &hif_dev->reg_in_submitted);
+@@ -972,8 +948,6 @@ static int ath9k_hif_usb_alloc_reg_in_urbs(struct hif_device_usb *hif_dev)
+ err_skb:
+ 	usb_free_urb(urb);
+ err_urb:
+-	kfree(rx_buf);
+-err_rxb:
+ 	ath9k_hif_usb_dealloc_reg_in_urbs(hif_dev);
+ 	return ret;
+ }
+diff --git a/drivers/net/wireless/ath/ath9k/hif_usb.h b/drivers/net/wireless/ath/ath9k/hif_usb.h
+index 5985aa15ca93..a94e7e1c86e9 100644
+--- a/drivers/net/wireless/ath/ath9k/hif_usb.h
++++ b/drivers/net/wireless/ath/ath9k/hif_usb.h
+@@ -86,11 +86,6 @@ struct tx_buf {
+ 	struct list_head list;
+ };
+ 
+-struct rx_buf {
+-	struct sk_buff *skb;
+-	struct hif_device_usb *hif_dev;
+-};
+-
+ #define HIF_USB_TX_STOP  BIT(0)
+ #define HIF_USB_TX_FLUSH BIT(1)
+ 
diff --git a/0107-intel_idle-tweak-cpuidle-cstates.patch b/0107-intel_idle-tweak-cpuidle-cstates.patch
index 984d4f6..4804923 100644
--- a/0107-intel_idle-tweak-cpuidle-cstates.patch
+++ b/0107-intel_idle-tweak-cpuidle-cstates.patch
@@ -1,32 +1,16 @@
-From 790db86f51a23533d457b361bb61e2845b6de6b8 Mon Sep 17 00:00:00 2001
-From: Arjan van de Ven <arjan@linux.intel.com>
-Date: Sat, 19 Mar 2016 21:32:19 -0400
-Subject: [PATCH 107/126] intel_idle: tweak cpuidle cstates
-
-Increase target_residency in cpuidle cstate
-
-Tune intel_idle to be a bit less agressive;
-Clear linux is cleaner in hygiene (wakupes) than the average linux,
-so we can afford changing these in a way that increases
-performance while keeping power efficiency
----
- drivers/idle/intel_idle.c | 46 +++++++++++++++++++++++-----------------------
- 1 file changed, 23 insertions(+), 23 deletions(-)
-
-diff --git a/drivers/idle/intel_idle.c b/drivers/idle/intel_idle.c
-index f0b06b14e782..24609fbb1010 100644
---- a/drivers/idle/intel_idle.c
-+++ b/drivers/idle/intel_idle.c
-@@ -466,7 +466,7 @@ static struct cpuidle_state hsw_cstates[] = {
+diff -up linux-5.6/drivers/idle/intel_idle.c.23~ linux-5.6/drivers/idle/intel_idle.c
+--- linux-5.6/drivers/idle/intel_idle.c.23~	2020-03-30 00:25:41.000000000 +0200
++++ linux-5.6/drivers/idle/intel_idle.c	2020-04-01 01:18:28.274248595 +0200
+@@ -462,7 +462,7 @@ static struct cpuidle_state hsw_cstates[
  		.desc = "MWAIT 0x01",
- 		.flags = MWAIT2flg(0x01),
+ 		.flags = MWAIT2flg(0x01) | CPUIDLE_FLAG_ALWAYS_ENABLE,
  		.exit_latency = 10,
 -		.target_residency = 20,
 +		.target_residency = 120,
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -474,7 +474,7 @@ static struct cpuidle_state hsw_cstates[] = {
+@@ -470,7 +470,7 @@ static struct cpuidle_state hsw_cstates[
  		.desc = "MWAIT 0x10",
  		.flags = MWAIT2flg(0x10) | CPUIDLE_FLAG_TLB_FLUSHED,
  		.exit_latency = 33,
@@ -35,7 +19,7 @@ index f0b06b14e782..24609fbb1010 100644
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -482,7 +482,7 @@ static struct cpuidle_state hsw_cstates[] = {
+@@ -478,7 +478,7 @@ static struct cpuidle_state hsw_cstates[
  		.desc = "MWAIT 0x20",
  		.flags = MWAIT2flg(0x20) | CPUIDLE_FLAG_TLB_FLUSHED,
  		.exit_latency = 133,
@@ -44,7 +28,7 @@ index f0b06b14e782..24609fbb1010 100644
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -490,7 +490,7 @@ static struct cpuidle_state hsw_cstates[] = {
+@@ -486,7 +486,7 @@ static struct cpuidle_state hsw_cstates[
  		.desc = "MWAIT 0x32",
  		.flags = MWAIT2flg(0x32) | CPUIDLE_FLAG_TLB_FLUSHED,
  		.exit_latency = 166,
@@ -53,7 +37,7 @@ index f0b06b14e782..24609fbb1010 100644
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -498,7 +498,7 @@ static struct cpuidle_state hsw_cstates[] = {
+@@ -494,7 +494,7 @@ static struct cpuidle_state hsw_cstates[
  		.desc = "MWAIT 0x40",
  		.flags = MWAIT2flg(0x40) | CPUIDLE_FLAG_TLB_FLUSHED,
  		.exit_latency = 300,
@@ -62,7 +46,7 @@ index f0b06b14e782..24609fbb1010 100644
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -506,7 +506,7 @@ static struct cpuidle_state hsw_cstates[] = {
+@@ -502,7 +502,7 @@ static struct cpuidle_state hsw_cstates[
  		.desc = "MWAIT 0x50",
  		.flags = MWAIT2flg(0x50) | CPUIDLE_FLAG_TLB_FLUSHED,
  		.exit_latency = 600,
@@ -71,7 +55,7 @@ index f0b06b14e782..24609fbb1010 100644
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -514,7 +514,7 @@ static struct cpuidle_state hsw_cstates[] = {
+@@ -510,7 +510,7 @@ static struct cpuidle_state hsw_cstates[
  		.desc = "MWAIT 0x60",
  		.flags = MWAIT2flg(0x60) | CPUIDLE_FLAG_TLB_FLUSHED,
  		.exit_latency = 2600,
@@ -80,16 +64,16 @@ index f0b06b14e782..24609fbb1010 100644
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -534,7 +534,7 @@ static struct cpuidle_state bdw_cstates[] = {
+@@ -530,7 +530,7 @@ static struct cpuidle_state bdw_cstates[
  		.desc = "MWAIT 0x01",
- 		.flags = MWAIT2flg(0x01),
+ 		.flags = MWAIT2flg(0x01) | CPUIDLE_FLAG_ALWAYS_ENABLE,
  		.exit_latency = 10,
 -		.target_residency = 20,
 +		.target_residency = 120,
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -542,7 +542,7 @@ static struct cpuidle_state bdw_cstates[] = {
+@@ -538,7 +538,7 @@ static struct cpuidle_state bdw_cstates[
  		.desc = "MWAIT 0x10",
  		.flags = MWAIT2flg(0x10) | CPUIDLE_FLAG_TLB_FLUSHED,
  		.exit_latency = 40,
@@ -98,7 +82,7 @@ index f0b06b14e782..24609fbb1010 100644
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -550,7 +550,7 @@ static struct cpuidle_state bdw_cstates[] = {
+@@ -546,7 +546,7 @@ static struct cpuidle_state bdw_cstates[
  		.desc = "MWAIT 0x20",
  		.flags = MWAIT2flg(0x20) | CPUIDLE_FLAG_TLB_FLUSHED,
  		.exit_latency = 133,
@@ -107,7 +91,7 @@ index f0b06b14e782..24609fbb1010 100644
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -558,7 +558,7 @@ static struct cpuidle_state bdw_cstates[] = {
+@@ -554,7 +554,7 @@ static struct cpuidle_state bdw_cstates[
  		.desc = "MWAIT 0x32",
  		.flags = MWAIT2flg(0x32) | CPUIDLE_FLAG_TLB_FLUSHED,
  		.exit_latency = 166,
@@ -116,7 +100,7 @@ index f0b06b14e782..24609fbb1010 100644
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -566,7 +566,7 @@ static struct cpuidle_state bdw_cstates[] = {
+@@ -562,7 +562,7 @@ static struct cpuidle_state bdw_cstates[
  		.desc = "MWAIT 0x40",
  		.flags = MWAIT2flg(0x40) | CPUIDLE_FLAG_TLB_FLUSHED,
  		.exit_latency = 300,
@@ -125,7 +109,7 @@ index f0b06b14e782..24609fbb1010 100644
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -574,7 +574,7 @@ static struct cpuidle_state bdw_cstates[] = {
+@@ -570,7 +570,7 @@ static struct cpuidle_state bdw_cstates[
  		.desc = "MWAIT 0x50",
  		.flags = MWAIT2flg(0x50) | CPUIDLE_FLAG_TLB_FLUSHED,
  		.exit_latency = 600,
@@ -134,7 +118,7 @@ index f0b06b14e782..24609fbb1010 100644
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -582,7 +582,7 @@ static struct cpuidle_state bdw_cstates[] = {
+@@ -578,7 +578,7 @@ static struct cpuidle_state bdw_cstates[
  		.desc = "MWAIT 0x60",
  		.flags = MWAIT2flg(0x60) | CPUIDLE_FLAG_TLB_FLUSHED,
  		.exit_latency = 2600,
@@ -143,16 +127,16 @@ index f0b06b14e782..24609fbb1010 100644
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -603,7 +603,7 @@ static struct cpuidle_state skl_cstates[] = {
+@@ -599,7 +599,7 @@ static struct cpuidle_state skl_cstates[
  		.desc = "MWAIT 0x01",
- 		.flags = MWAIT2flg(0x01),
+ 		.flags = MWAIT2flg(0x01) | CPUIDLE_FLAG_ALWAYS_ENABLE,
  		.exit_latency = 10,
 -		.target_residency = 20,
 +		.target_residency = 120,
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -611,7 +611,7 @@ static struct cpuidle_state skl_cstates[] = {
+@@ -607,7 +607,7 @@ static struct cpuidle_state skl_cstates[
  		.desc = "MWAIT 0x10",
  		.flags = MWAIT2flg(0x10) | CPUIDLE_FLAG_TLB_FLUSHED,
  		.exit_latency = 70,
@@ -161,7 +145,7 @@ index f0b06b14e782..24609fbb1010 100644
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -619,7 +619,7 @@ static struct cpuidle_state skl_cstates[] = {
+@@ -615,7 +615,7 @@ static struct cpuidle_state skl_cstates[
  		.desc = "MWAIT 0x20",
  		.flags = MWAIT2flg(0x20) | CPUIDLE_FLAG_TLB_FLUSHED,
  		.exit_latency = 85,
@@ -170,7 +154,7 @@ index f0b06b14e782..24609fbb1010 100644
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -627,7 +627,7 @@ static struct cpuidle_state skl_cstates[] = {
+@@ -623,7 +623,7 @@ static struct cpuidle_state skl_cstates[
  		.desc = "MWAIT 0x33",
  		.flags = MWAIT2flg(0x33) | CPUIDLE_FLAG_TLB_FLUSHED,
  		.exit_latency = 124,
@@ -179,7 +163,7 @@ index f0b06b14e782..24609fbb1010 100644
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -635,7 +635,7 @@ static struct cpuidle_state skl_cstates[] = {
+@@ -631,7 +631,7 @@ static struct cpuidle_state skl_cstates[
  		.desc = "MWAIT 0x40",
  		.flags = MWAIT2flg(0x40) | CPUIDLE_FLAG_TLB_FLUSHED,
  		.exit_latency = 200,
@@ -188,7 +172,7 @@ index f0b06b14e782..24609fbb1010 100644
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -643,7 +643,7 @@ static struct cpuidle_state skl_cstates[] = {
+@@ -639,7 +639,7 @@ static struct cpuidle_state skl_cstates[
  		.desc = "MWAIT 0x50",
  		.flags = MWAIT2flg(0x50) | CPUIDLE_FLAG_TLB_FLUSHED,
  		.exit_latency = 480,
@@ -197,7 +181,7 @@ index f0b06b14e782..24609fbb1010 100644
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -651,7 +651,7 @@ static struct cpuidle_state skl_cstates[] = {
+@@ -647,7 +647,7 @@ static struct cpuidle_state skl_cstates[
  		.desc = "MWAIT 0x60",
  		.flags = MWAIT2flg(0x60) | CPUIDLE_FLAG_TLB_FLUSHED,
  		.exit_latency = 890,
@@ -206,16 +190,16 @@ index f0b06b14e782..24609fbb1010 100644
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -672,7 +672,7 @@ static struct cpuidle_state skx_cstates[] = {
+@@ -668,7 +668,7 @@ static struct cpuidle_state skx_cstates[
  		.desc = "MWAIT 0x01",
- 		.flags = MWAIT2flg(0x01),
+ 		.flags = MWAIT2flg(0x01) | CPUIDLE_FLAG_ALWAYS_ENABLE,
  		.exit_latency = 10,
 -		.target_residency = 20,
 +		.target_residency = 1000,
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
-@@ -680,7 +680,7 @@ static struct cpuidle_state skx_cstates[] = {
+@@ -676,7 +676,7 @@ static struct cpuidle_state skx_cstates[
  		.desc = "MWAIT 0x20",
  		.flags = MWAIT2flg(0x20) | CPUIDLE_FLAG_TLB_FLUSHED,
  		.exit_latency = 133,
@@ -224,5 +208,3 @@ index f0b06b14e782..24609fbb1010 100644
  		.enter = &intel_idle,
  		.enter_s2idle = intel_idle_s2idle, },
  	{
--- 
-2.15.0
diff --git a/2-0001-x86-Don-t-declare-__force_order-in-kaslr_64.c.patch b/2-0001-x86-Don-t-declare-__force_order-in-kaslr_64.c.patch
new file mode 100644
index 0000000..b809432
--- /dev/null
+++ b/2-0001-x86-Don-t-declare-__force_order-in-kaslr_64.c.patch
@@ -0,0 +1,37 @@
+From c8c26194cf5a344cd53763eaaf16c3ab609736f4 Mon Sep 17 00:00:00 2001
+From: "H.J. Lu" <hjl.tools@gmail.com>
+Date: Thu, 16 Jan 2020 12:46:51 -0800
+Subject: [PATCH] x86: Don't declare __force_order in kaslr_64.c
+
+GCC 10 changed the default to -fno-common, which leads to
+
+  LD      arch/x86/boot/compressed/vmlinux
+ld: arch/x86/boot/compressed/pgtable_64.o:(.bss+0x0): multiple definition of `__force_order'; arch/x86/boot/compressed/kaslr_64.o:(.bss+0x0): first defined here
+make[2]: *** [arch/x86/boot/compressed/Makefile:119: arch/x86/boot/compressed/vmlinux] Error 1
+
+Since __force_order is already provided in pgtable_64.c, there is no
+need to declare __force_order in kaslr_64.c.
+
+Signed-off-by: H.J. Lu <hjl.tools@gmail.com>
+Signed-off-by: Yu-cheng Yu <yu-cheng.yu@intel.com>
+---
+ arch/x86/boot/compressed/kaslr_64.c | 3 ---
+ 1 file changed, 3 deletions(-)
+
+diff --git a/arch/x86/boot/compressed/kaslr_64.c b/arch/x86/boot/compressed/kaslr_64.c
+index 748456c365f4..9557c5a15b91 100644
+--- a/arch/x86/boot/compressed/kaslr_64.c
++++ b/arch/x86/boot/compressed/kaslr_64.c
+@@ -29,9 +29,6 @@
+ #define __PAGE_OFFSET __PAGE_OFFSET_BASE
+ #include "../../mm/ident_map.c"
+ 
+-/* Used by pgtable.h asm code to force instruction serialization. */
+-unsigned long __force_order;
+-
+ /* Used to track our page table allocation area. */
+ struct alloc_pgt_data {
+ 	unsigned char *pgt_buf;
+-- 
+2.24.1
+
diff --git a/7a8b64d17e35810dc3176fe61208b45c15d25402.patch b/7a8b64d17e35810dc3176fe61208b45c15d25402.patch
new file mode 100644
index 0000000..efa63b9
--- /dev/null
+++ b/7a8b64d17e35810dc3176fe61208b45c15d25402.patch
@@ -0,0 +1,77 @@
+commit 7a8b64d17e35810dc3176fe61208b45c15d25402
+Author: Rob Herring <robh@kernel.org>
+Date:   Thu Feb 6 14:02:30 2020 +0000
+
+    of/address: use range parser for of_dma_get_range
+    
+    of_dma_get_range() does the same ranges parsing as
+    of_pci_range_parser_one(), so let's refactor of_dma_get_range() to use
+    it instead.
+    
+    This commit is no functional change. Subsequent commits will parse more
+    than the 1st dma-ranges entry.
+    
+    Signed-off-by: Rob Herring <robh@kernel.org>
+
+diff --git a/drivers/of/address.c b/drivers/of/address.c
+index 6d33f849f11440..a2c45812a50e52 100644
+--- a/drivers/of/address.c
++++ b/drivers/of/address.c
+@@ -939,10 +939,11 @@ int of_dma_get_range(struct device_node *np, u64 *dma_addr, u64 *paddr, u64 *siz
+ {
+ 	struct device_node *node = of_node_get(np);
+ 	const __be32 *ranges = NULL;
+-	int len, naddr, nsize, pna;
++	int len;
+ 	int ret = 0;
+ 	bool found_dma_ranges = false;
+-	u64 dmaaddr;
++	struct of_range_parser parser;
++	struct of_range range;
+ 
+ 	while (node) {
+ 		ranges = of_get_property(node, "dma-ranges", &len);
+@@ -967,33 +968,20 @@ int of_dma_get_range(struct device_node *np, u64 *dma_addr, u64 *paddr, u64 *siz
+ 		goto out;
+ 	}
+ 
+-	naddr = of_bus_n_addr_cells(node);
+-	nsize = of_bus_n_size_cells(node);
+-	pna = of_n_addr_cells(node);
+-	if ((len / sizeof(__be32)) % (pna + naddr + nsize)) {
+-		ret = -EINVAL;
+-		goto out;
+-	}
++	of_dma_range_parser_init(&parser, node);
++
++	for_each_of_range(&parser, &range) {
++		pr_debug("dma_addr(%llx) cpu_addr(%llx) size(%llx)\n",
++			 range.bus_addr, range.cpu_addr, range.size);
++
++		*dma_addr = range.bus_addr;
++		*paddr = range.cpu_addr;
++		*size = range.size;
+ 
+-	/* dma-ranges format:
+-	 * DMA addr	: naddr cells
+-	 * CPU addr	: pna cells
+-	 * size		: nsize cells
+-	 */
+-	dmaaddr = of_read_number(ranges, naddr);
+-	*paddr = of_translate_dma_address(node, ranges + naddr);
+-	if (*paddr == OF_BAD_ADDR) {
+-		pr_err("translation of DMA address(%llx) to CPU address failed node(%pOF)\n",
+-		       dmaaddr, np);
+-		ret = -EINVAL;
+ 		goto out;
+ 	}
+-	*dma_addr = dmaaddr;
+-
+-	*size = of_read_number(ranges + naddr + pna, nsize);
+ 
+-	pr_debug("dma_addr(%llx) cpu_addr(%llx) size(%llx)\n",
+-		 *dma_addr, *paddr, *size);
++	pr_err("translation of DMA ranges failed on node(%pOF)\n", np);
+ 
+ out:
+ 	of_node_put(node);
diff --git a/9d55bebd9816903b821a403a69a94190442ac043.patch b/9d55bebd9816903b821a403a69a94190442ac043.patch
new file mode 100644
index 0000000..b3ead93
--- /dev/null
+++ b/9d55bebd9816903b821a403a69a94190442ac043.patch
@@ -0,0 +1,68 @@
+commit 9d55bebd9816903b821a403a69a94190442ac043
+Author: Rob Herring <robh@kernel.org>
+Date:   Thu Feb 6 14:26:40 2020 +0000
+
+    of/address: Support multiple 'dma-ranges' entries
+    
+    Currently, the DMA offset and mask for a device are set based only on the
+    first 'dma-ranges' entry. We should really be using all the entries. The
+    kernel doesn't yet support multiple offsets and sizes, so the best we can
+    do is to find the biggest size for a single offset. The algorithm is
+    copied from acpi_dma_get_range().
+    
+    If there's different offsets from the first entry, then we warn and
+    continue. It really should be an error, but this will likely break
+    existing DTs.
+    
+    Signed-off-by: Rob Herring <robh@kernel.org>
+
+diff --git a/drivers/of/address.c b/drivers/of/address.c
+index a2c45812a50e52..8eea3f6e29a441 100644
+--- a/drivers/of/address.c
++++ b/drivers/of/address.c
+@@ -944,6 +944,7 @@ int of_dma_get_range(struct device_node *np, u64 *dma_addr, u64 *paddr, u64 *siz
+ 	bool found_dma_ranges = false;
+ 	struct of_range_parser parser;
+ 	struct of_range range;
++	u64 dma_start = U64_MAX, dma_end = 0, dma_offset = 0;
+ 
+ 	while (node) {
+ 		ranges = of_get_property(node, "dma-ranges", &len);
+@@ -974,14 +975,33 @@ int of_dma_get_range(struct device_node *np, u64 *dma_addr, u64 *paddr, u64 *siz
+ 		pr_debug("dma_addr(%llx) cpu_addr(%llx) size(%llx)\n",
+ 			 range.bus_addr, range.cpu_addr, range.size);
+ 
+-		*dma_addr = range.bus_addr;
+-		*paddr = range.cpu_addr;
+-		*size = range.size;
++		if (dma_offset && range.cpu_addr - range.bus_addr != dma_offset) {
++			pr_warn("Can't handle multiple dma-ranges with different offsets on node(%pOF)\n", node);
++			/* Don't error out as we'd break some existing DTs */
++			continue;
++		}
++		dma_offset = range.cpu_addr - range.bus_addr;
++
++		/* Take lower and upper limits */
++		if (range.bus_addr < dma_start)
++			dma_start = range.bus_addr;
++		if (range.bus_addr + range.size > dma_end)
++			dma_end = range.bus_addr + range.size;
++	}
+ 
++	if (dma_start >= dma_end) {
++		ret = -EINVAL;
++		pr_debug("Invalid DMA ranges configuration on node(%pOF)\n",
++			 node);
+ 		goto out;
+ 	}
+ 
+-	pr_err("translation of DMA ranges failed on node(%pOF)\n", np);
++	*dma_addr = dma_start;
++	*size = dma_end - dma_start;
++	*paddr = dma_start + dma_offset;
++
++	pr_debug("final: dma_addr(%llx) cpu_addr(%llx) size(%llx)\n",
++		 *dma_addr, *paddr, *size);
+ 
+ out:
+ 	of_node_put(node);
diff --git a/android-enable-building-ashmem-and-binder-as-modules.patch b/android-enable-building-ashmem-and-binder-as-modules.patch
index 115ec80..05ad8f8 100644
--- a/android-enable-building-ashmem-and-binder-as-modules.patch
+++ b/android-enable-building-ashmem-and-binder-as-modules.patch
@@ -1,30 +1,18 @@
-From: Ben Hutchings <ben@decadent.org.uk>
-Date: Fri, 22 Jun 2018 17:27:00 +0100
-Subject: android: Enable building ashmem and binder as modules
-Bug-Debian: https://bugs.debian.org/901492
-
-We want to enable use of the Android ashmem and binder drivers to
-support Anbox, but they should not be built-in as that would waste
-resources and increase security attack surface on systems that don't
-need them.
-
-- Add a MODULE_LICENSE declaration to ashmem
-- Change the Makefiles to build each driver as an object with the
-  "_linux" suffix (which is what Anbox expects)
-- Change config symbol types to tristate
----
- drivers/android/Kconfig          | 2 +-
- drivers/android/Makefile         | 7 ++++---
- drivers/android/binder_alloc.c   | 2 +-
- drivers/staging/android/Kconfig  | 2 +-
- drivers/staging/android/Makefile | 3 ++-
- drivers/staging/android/ashmem.c | 3 +++
- 6 files changed, 12 insertions(+), 7 deletions(-)
-
-diff --git a/drivers/android/Kconfig b/drivers/android/Kconfig
-index 6fdf2abe4598..329ec9941fbf 100644
---- a/drivers/android/Kconfig
-+++ b/drivers/android/Kconfig
+diff -up linux-5.6-rc3/drivers/android/binder_alloc.c.22~ linux-5.6-rc3/drivers/android/binder_alloc.c
+--- linux-5.6-rc3/drivers/android/binder_alloc.c.22~	2020-02-24 01:17:42.000000000 +0100
++++ linux-5.6-rc3/drivers/android/binder_alloc.c	2020-02-24 23:26:47.054727484 +0100
+@@ -38,7 +38,7 @@ enum {
+ };
+ static uint32_t binder_alloc_debug_mask = BINDER_DEBUG_USER_ERROR;
+ 
+-module_param_named(debug_mask, binder_alloc_debug_mask,
++module_param_named(alloc_debug_mask, binder_alloc_debug_mask,
+ 		   uint, 0644);
+ 
+ #define binder_alloc_debug(mask, x...) \
+diff -up linux-5.6-rc3/drivers/android/Kconfig.22~ linux-5.6-rc3/drivers/android/Kconfig
+--- linux-5.6-rc3/drivers/android/Kconfig.22~	2020-02-24 01:17:42.000000000 +0100
++++ linux-5.6-rc3/drivers/android/Kconfig	2020-02-24 23:26:47.054727484 +0100
 @@ -9,7 +9,7 @@ config ANDROID
  if ANDROID
  
@@ -34,10 +22,9 @@ index 6fdf2abe4598..329ec9941fbf 100644
  	depends on MMU
  	default n
  	---help---
-diff --git a/drivers/android/Makefile b/drivers/android/Makefile
-index c9d3d0c99c25..55411d9a9c2a 100644
---- a/drivers/android/Makefile
-+++ b/drivers/android/Makefile
+diff -up linux-5.6-rc3/drivers/android/Makefile.22~ linux-5.6-rc3/drivers/android/Makefile
+--- linux-5.6-rc3/drivers/android/Makefile.22~	2020-02-24 01:17:42.000000000 +0100
++++ linux-5.6-rc3/drivers/android/Makefile	2020-02-24 23:26:47.054727484 +0100
 @@ -1,6 +1,7 @@
  # SPDX-License-Identifier: GPL-2.0-only
  ccflags-y += -I$(src)			# needed for trace events
@@ -49,23 +36,26 @@ index c9d3d0c99c25..55411d9a9c2a 100644
 +binder_linux-y := binder.o binder_alloc.o
 +binder_linux-$(CONFIG_ANDROID_BINDERFS)	+= binderfs.o
 +binder_linux-$(CONFIG_ANDROID_BINDER_IPC_SELFTEST) += binder_alloc_selftest.o
-diff --git a/drivers/android/binder_alloc.c b/drivers/android/binder_alloc.c
-index ce5603c2291c..099fef731148 100644
---- a/drivers/android/binder_alloc.c
-+++ b/drivers/android/binder_alloc.c
-@@ -37,7 +37,7 @@ enum {
- };
- static uint32_t binder_alloc_debug_mask = BINDER_DEBUG_USER_ERROR;
- 
--module_param_named(debug_mask, binder_alloc_debug_mask,
-+module_param_named(alloc_debug_mask, binder_alloc_debug_mask,
- 		   uint, 0644);
+diff -up linux-5.6-rc3/drivers/staging/android/ashmem.c.22~ linux-5.6-rc3/drivers/staging/android/ashmem.c
+--- linux-5.6-rc3/drivers/staging/android/ashmem.c.22~	2020-02-24 01:17:42.000000000 +0100
++++ linux-5.6-rc3/drivers/staging/android/ashmem.c	2020-02-24 23:26:47.054727484 +0100
+@@ -24,6 +24,7 @@
+ #include <linux/bitops.h>
+ #include <linux/mutex.h>
+ #include <linux/shmem_fs.h>
++#include <linux/module.h>
+ #include "ashmem.h"
  
- #define binder_alloc_debug(mask, x...) \
-diff --git a/drivers/staging/android/Kconfig b/drivers/staging/android/Kconfig
-index d6d605d5cbde..afe9085369a3 100644
---- a/drivers/staging/android/Kconfig
-+++ b/drivers/staging/android/Kconfig
+ #define ASHMEM_NAME_PREFIX "dev/ashmem/"
+@@ -953,3 +954,5 @@ out:
+ 	return ret;
+ }
+ device_initcall(ashmem_init);
++
++MODULE_LICENSE("GPL v2");
+diff -up linux-5.6-rc3/drivers/staging/android/Kconfig.22~ linux-5.6-rc3/drivers/staging/android/Kconfig
+--- linux-5.6-rc3/drivers/staging/android/Kconfig.22~	2020-02-24 01:17:42.000000000 +0100
++++ linux-5.6-rc3/drivers/staging/android/Kconfig	2020-02-24 23:26:47.054727484 +0100
 @@ -4,7 +4,7 @@ menu "Android"
  if ANDROID
  
@@ -75,36 +65,14 @@ index d6d605d5cbde..afe9085369a3 100644
  	depends on SHMEM
  	help
  	  The ashmem subsystem is a new shared memory allocator, similar to
-diff --git a/drivers/staging/android/Makefile b/drivers/staging/android/Makefile
-index 14bd9c6ce10d..c202aa4f8323 100644
---- a/drivers/staging/android/Makefile
-+++ b/drivers/staging/android/Makefile
-@@ -3,5 +3,6 @@ ccflags-y += -I$(src)			# needed for trace events
+diff -up linux-5.6-rc3/drivers/staging/android/Makefile.22~ linux-5.6-rc3/drivers/staging/android/Makefile
+--- linux-5.6-rc3/drivers/staging/android/Makefile.22~	2020-02-24 23:26:47.054727484 +0100
++++ linux-5.6-rc3/drivers/staging/android/Makefile	2020-02-24 23:27:56.058734039 +0100
+@@ -3,4 +3,6 @@ ccflags-y += -I$(src)			# needed for tra
  
  obj-y					+= ion/
  
 -obj-$(CONFIG_ASHMEM)			+= ashmem.o
 +obj-$(CONFIG_ASHMEM)			+= ashmem_linux.o
 +ashmem_linux-y				+= ashmem.o
- obj-$(CONFIG_ANDROID_VSOC)		+= vsoc.o
-diff --git a/drivers/staging/android/ashmem.c b/drivers/staging/android/ashmem.c
-index 74d497d39c5a..fbbe467b0ba8 100644
---- a/drivers/staging/android/ashmem.c
-+++ b/drivers/staging/android/ashmem.c
-@@ -24,6 +24,7 @@
- #include <linux/bitops.h>
- #include <linux/mutex.h>
- #include <linux/shmem_fs.h>
-+#include <linux/module.h>
- #include "ashmem.h"
- 
- #define ASHMEM_NAME_PREFIX "dev/ashmem/"
-@@ -925,3 +926,5 @@ static int __init ashmem_init(void)
- 	return ret;
- }
- device_initcall(ashmem_init);
 +
-+MODULE_LICENSE("GPL v2");
--- 
-2.20.1
-
diff --git a/arm64-common.config b/arm64-common.config
index c444a48..3a6539a 100644
--- a/arm64-common.config
+++ b/arm64-common.config
@@ -1029,3 +1029,9 @@ CONFIG_OF_RECONFIG_NOTIFIER_ERROR_INJECT=m
 CONFIG_ARM64_PTDUMP_CORE=y
 CONFIG_ARM64_PTDUMP_DEBUGFS=y
 CONFIG_ARM64_RELOC_TEST=m
+CONFIG_GOOGLE_FIRMWARE=y
+CONFIG_GOOGLE_COREBOOT_TABLE=m
+CONFIG_GOOGLE_FRAMEBUFFER_COREBOOT=m
+CONFIG_GOOGLE_MEMCONSOLE_COREBOOT=m
+CONFIG_GOOGLE_VPD=m
+CONFIG_IMX_DSP=m
diff --git a/btrfs-zstd-backport-to-4.12-api.patch b/btrfs-zstd-backport-to-4.12-api.patch
deleted file mode 100644
index e52717e..0000000
--- a/btrfs-zstd-backport-to-4.12-api.patch
+++ /dev/null
@@ -1,17 +0,0 @@
---- linux-4.12/fs/btrfs/zstd.c.omv~	2017-08-17 21:12:35.674612608 +0200
-+++ linux-4.12/fs/btrfs/zstd.c	2017-08-17 21:13:29.883512195 +0200
-@@ -260,13 +260,9 @@ out:
- 	return ret;
- }
- 
--static int zstd_decompress_bio(struct list_head *ws, struct compressed_bio *cb)
-+static int zstd_decompress_bio(struct list_head *ws, struct page **pages_in, u64 disk_start, struct bio *orig_bio, size_t srclen)
- {
- 	struct workspace *workspace = list_entry(ws, struct workspace, list);
--	struct page **pages_in = cb->compressed_pages;
--	u64 disk_start = cb->start;
--	struct bio *orig_bio = cb->orig_bio;
--	size_t srclen = cb->compressed_len;
- 	ZSTD_DStream *stream;
- 	int ret = 0;
- 	unsigned long page_in_index = 0;
diff --git a/common-desktop.config b/common-desktop.config
index e1d0ebf..185ad4c 100644
--- a/common-desktop.config
+++ b/common-desktop.config
@@ -30,3 +30,5 @@ CONFIG_UKSM=y
 CONFIG_PREEMPT_VOLUNTARY=y
 # CONFIG_PREEMPT is not set
 CONFIG_PREEMPT_COUNT=y
+CONFIG_HZ_1000=y
+CONFIG_HZ=1000
diff --git a/common-server.config b/common-server.config
index a6d6012..7f739c0 100644
--- a/common-server.config
+++ b/common-server.config
@@ -1,5 +1,3 @@
-# CONFIG_SCHED_MUQSS is not set
-
 #
 # RCU Subsystem
 #
@@ -23,6 +21,7 @@ CONFIG_INLINE_READ_UNLOCK_IRQ=y
 CONFIG_INLINE_WRITE_UNLOCK=y
 CONFIG_INLINE_WRITE_UNLOCK_IRQ=y
 
+
 #
 # Processor type and features
 #
@@ -31,3 +30,7 @@ CONFIG_PREEMPT_NONE=y
 # CONFIG_PREEMPT is not set
 # CONFIG_PREEMPT_COUNT is not set
 CONFIG_SLAB=y
+CONFIG_HZ_300=y
+CONFIG_HZ=300
+# CONFIG_HZ_250 is not set
+# CONFIG_HZ_100 is not set
diff --git a/common.config b/common.config
index 57e65c7..28c1ebe 100644
--- a/common.config
+++ b/common.config
@@ -1,7 +1,7 @@
 # CONFIG_KERNEL_GZIP is not set
 CONFIG_KERNEL_ZSTD=y
 CONFIG_DEFAULT_HOSTNAME="omv"
-CONFIG_CC_OPTIMIZE_FOR_PERFORMANCE_O3=y
+# CONFIG_CC_OPTIMIZE_FOR_PERFORMANCE_O3 is not set
 # CONFIG_USELIB is not set
 CONFIG_GENERIC_IRQ_CHIP=y
 CONFIG_IRQ_SIM=y
@@ -31,13 +31,14 @@ CONFIG_CGROUP_HUGETLB=y
 CONFIG_CGROUP_DEVICE=y
 CONFIG_CGROUP_PERF=y
 CONFIG_CGROUP_BPF=y
+CONFIG_CGROUP_FREEZER=y
 CONFIG_HAVE_GCC_PLUGINS=y
 CONFIG_GCC_PLUGINS=y
 CONFIG_SOCK_CGROUP_DATA=y
 CONFIG_USER_NS=y
 CONFIG_CHECKPOINT_RESTORE=y
 CONFIG_SCHED_AUTOGROUP=y
-CONFIG_EXPERT=y
+# CONFIG_EXPERT is not set
 # CONFIG_SYSFS_SYSCALL is not set
 # CONFIG_SYSFS_DEPRECATED is not set
 CONFIG_SYSCTL_SYSCALL=y
@@ -131,6 +132,7 @@ CONFIG_ACPI_PCI_SLOT=y
 CONFIG_ACPI_HOTPLUG_MEMORY=y
 CONFIG_ACPI_CUSTOM_METHOD=m
 CONFIG_ACPI_NFIT=m
+CONFIG_ACPI_HMAT=m
 CONFIG_DPTF_POWER=m
 CONFIG_ACPI_WATCHDOG=y
 CONFIG_ACPI_EXTLOG=m
@@ -220,7 +222,7 @@ CONFIG_FW_CFG_SYSFS=m
 CONFIG_FW_CFG_SYSFS_CMDLINE=y
 CONFIG_FRAMEBUFFER_CONSOLE_DEFERRED_TAKEOVER=y
 CONFIG_FRAMEBUFFER_CONSOLE_ROTATION=y
-# CONFIG_EFI_VARS is not set
+CONFIG_EFI_VARS=m
 CONFIG_EFI_CAPSULE_LOADER=m
 CONFIG_EFI_TEST=m
 CONFIG_APPLE_PROPERTIES=y
@@ -365,6 +367,8 @@ CONFIG_NET_IPVTI=m
 CONFIG_NET_UDP_TUNNEL=m
 CONFIG_NET_FOU=m
 CONFIG_NET_FOU_IP_TUNNELS=y
+CONFIG_WIREGUARD=m
+# CONFIG_WIREGUARD_DEBUG is not set
 CONFIG_INET_AH=m
 CONFIG_INET_ESP=m
 CONFIG_INET_ESP_OFFLOAD=m
@@ -393,6 +397,7 @@ CONFIG_TCP_CONG_YEAH=m
 CONFIG_TCP_CONG_ILLINOIS=m
 CONFIG_TCP_CONG_DCTCP=m
 CONFIG_TCP_CONG_BBR=m
+CONFIG_MPTCP=y
 CONFIG_IPV6=m
 CONFIG_IPV6_ROUTER_PREF=y
 CONFIG_IPV6_ROUTE_INFO=y
@@ -829,6 +834,7 @@ CONFIG_IEEE802154_SOCKET=m
 CONFIG_IEEE802154_6LOWPAN=m
 CONFIG_MAC802154=m
 CONFIG_NET_SCH_CBQ=m
+CONFIG_NET_SCH_ETS=m
 CONFIG_NET_SCH_CAKE=m
 CONFIG_NET_SCH_HTB=m
 CONFIG_NET_SCH_HFSC=m
@@ -851,6 +857,7 @@ CONFIG_NET_SCH_QFQ=m
 CONFIG_NET_SCH_CODEL=m
 CONFIG_NET_SCH_FQ_CODEL=y
 CONFIG_NET_SCH_FQ=m
+CONFIG_NET_SCH_FQ_PIE=m
 CONFIG_NET_SCH_HHF=m
 CONFIG_NET_SCH_PIE=m
 CONFIG_NET_SCH_INGRESS=m
@@ -1421,7 +1428,7 @@ CONFIG_IDE_GD_ATAPI=y
 CONFIG_BLK_DEV_IDECS=m
 CONFIG_BLK_DEV_DELKIN=m
 CONFIG_BLK_DEV_IDECD=m
-CONFIG_BLK_DEV_IDECD_VERBOSE_ERRORS=y
+# CONFIG_BLK_DEV_IDECD_VERBOSE_ERRORS is not set
 CONFIG_BLK_DEV_IDETAPE=m
 CONFIG_BLK_DEV_IDEACPI=y
 CONFIG_IDE_TASK_IOCTL=y
@@ -1892,6 +1899,7 @@ CONFIG_BE2NET_SKYHAWK=y
 CONFIG_PCMCIA_FMVJ18X=m
 CONFIG_HP100=m
 CONFIG_HINIC=m
+CONFIG_IGC=m
 CONFIG_E100=m
 CONFIG_E1000=m
 CONFIG_E1000E=m
@@ -2219,6 +2227,8 @@ CONFIG_LIBIPW=m
 CONFIG_IWLEGACY=m
 CONFIG_IWL4965=m
 CONFIG_IWL3945=m
+CONFIG_RTL8723DE=m
+CONFIG_RTL8821CE=m
 
 #
 # iwl3945 / iwl4965 Debugging Options
@@ -4333,8 +4343,8 @@ CONFIG_CHASH=m
 # CONFIG_CHASH_STATS is not set
 # CONFIG_CHASH_SELFTEST is not set
 CONFIG_DRM_NOUVEAU=m
-CONFIG_NOUVEAU_DEBUG=5
-CONFIG_NOUVEAU_DEBUG_DEFAULT=3
+CONFIG_NOUVEAU_DEBUG=3
+CONFIG_NOUVEAU_DEBUG_DEFAULT=1
 # CONFIG_NOUVEAU_DEBUG_MMU is not set
 CONFIG_DRM_NOUVEAU_BACKLIGHT=y
 CONFIG_DRM_I915=m
@@ -4519,7 +4529,7 @@ CONFIG_SOUND=m
 CONFIG_SOUND_OSS_CORE=y
 # CONFIG_SOUND_OSS_CORE_PRECLAIM is not set
 CONFIG_SND=m
-CONFIG_SND_DEBUG_VERBOSE=y
+# CONFIG_SND_DEBUG_VERBOSE is not set
 CONFIG_SND_TIMER=m
 CONFIG_SND_PCM=m
 CONFIG_SND_PCM_ELD=y
@@ -4914,6 +4924,7 @@ CONFIG_HID_BELKIN=m
 CONFIG_HID_BETOP_FF=m
 CONFIG_HID_CHERRY=m
 CONFIG_HID_CHICONY=m
+CONFIG_HID_CORSAIR=m
 CONFIG_HID_PRODIKEYS=m
 CONFIG_HID_CMEDIA=m
 CONFIG_HID_CP2112=m
@@ -6955,6 +6966,9 @@ CONFIG_842_DECOMPRESS=y
 CONFIG_LZ4_COMPRESS=m
 CONFIG_LZ4HC_COMPRESS=m
 CONFIG_ZSTD_COMPRESS=y
+CONFIG_RD_ZSTD=y
+CONFIG_DECOMPRESS_ZSTD=y
+CONFIG_INITRAMFS_COMPRESSION_ZSTD=y
 # CONFIG_XZ_DEC_POWERPC is not set
 # CONFIG_XZ_DEC_IA64 is not set
 # CONFIG_XZ_DEC_ARM is not set
@@ -7037,8 +7051,10 @@ CONFIG_TRACING_EVENTS_GPIO=y
 # CONFIG_RUNTIME_TESTING_MENU is not set
 CONFIG_MEMTEST=y
 CONFIG_IO_STRICT_DEVMEM=y
+# CONFIG_EARLY_PRINTK is not set
 # CONFIG_EARLY_PRINTK_DBGP is not set
-CONFIG_EARLY_PRINTK_USB_XDBC=y
+# CONFIG_EARLY_PRINTK_USB_XDBC is not set
+# CONFIG_X86_VERBOSE_BOOTUP is not set
 # CONFIG_IOMMU_DEBUG is not set
 # CONFIG_DEBUG_BOOT_PARAMS is not set
 # CONFIG_UNWINDER_GUESS is not set
diff --git a/export-symbols-needed-by-android-drivers.patch b/export-symbols-needed-by-android-drivers.patch
index ab7ab0f..437a37e 100644
--- a/export-symbols-needed-by-android-drivers.patch
+++ b/export-symbols-needed-by-android-drivers.patch
@@ -22,9 +22,10 @@ Export the currently un-exported symbols they depend on.
  security/security.c | 4 ++++
  9 files changed, 17 insertions(+)
 
---- a/fs/file.c
-+++ b/fs/file.c
-@@ -409,6 +409,7 @@ struct files_struct *get_files_struct(st
+diff -uraN linux-5.7.8/fs/file.c omv-linux-5.7.8/fs/file.c
+--- linux-5.7.8/fs/file.c	2020-07-09 09:39:40.000000000 +0200
++++ omv-linux-5.7.8/fs/file.c	2020-07-13 15:14:49.188263989 +0200
+@@ -409,6 +409,7 @@
  
  	return files;
  }
@@ -32,7 +33,7 @@ Export the currently un-exported symbols they depend on.
  
  void put_files_struct(struct files_struct *files)
  {
-@@ -421,6 +422,7 @@ void put_files_struct(struct files_struc
+@@ -421,6 +422,7 @@
  		kmem_cache_free(files_cachep, files);
  	}
  }
@@ -40,7 +41,7 @@ Export the currently un-exported symbols they depend on.
  
  void reset_files_struct(struct files_struct *files)
  {
-@@ -533,6 +535,7 @@ out:
+@@ -534,6 +536,7 @@
  	spin_unlock(&files->file_lock);
  	return error;
  }
@@ -48,7 +49,7 @@ Export the currently un-exported symbols they depend on.
  
  static int alloc_fd(unsigned start, unsigned flags)
  {
-@@ -606,6 +609,7 @@ void __fd_install(struct files_struct *f
+@@ -612,6 +615,7 @@
  	rcu_assign_pointer(fdt->fd[fd], file);
  	rcu_read_unlock_sched();
  }
@@ -56,7 +57,7 @@ Export the currently un-exported symbols they depend on.
  
  void fd_install(unsigned int fd, struct file *file)
  {
-@@ -668,6 +672,7 @@ out_unlock:
+@@ -676,6 +680,7 @@
  	*res = NULL;
  	return -ENOENT;
  }
@@ -64,9 +65,10 @@ Export the currently un-exported symbols they depend on.
  
  void do_close_on_exec(struct files_struct *files)
  {
---- a/kernel/fork.c
-+++ b/kernel/fork.c
-@@ -1085,6 +1085,7 @@ void mmput_async(struct mm_struct *mm)
+diff -uraN linux-5.7.8/kernel/fork.c omv-linux-5.7.8/kernel/fork.c
+--- linux-5.7.8/kernel/fork.c	2020-07-09 09:39:40.000000000 +0200
++++ omv-linux-5.7.8/kernel/fork.c	2020-07-13 15:15:57.177994389 +0200
+@@ -1123,6 +1123,7 @@
  		schedule_work(&mm->async_put_work);
  	}
  }
@@ -74,9 +76,10 @@ Export the currently un-exported symbols they depend on.
  #endif
  
  /**
---- a/kernel/sched/core.c
-+++ b/kernel/sched/core.c
-@@ -3943,6 +3943,7 @@ int can_nice(const struct task_struct *p
+diff -uraN linux-5.7.8/kernel/sched/core.c omv-linux-5.7.8/kernel/sched/core.c
+--- linux-5.7.8/kernel/sched/core.c	2020-07-09 09:39:40.000000000 +0200
++++ omv-linux-5.7.8/kernel/sched/core.c	2020-07-13 15:16:48.193591744 +0200
+@@ -4572,6 +4572,7 @@
  	return (nice_rlim <= task_rlimit(p, RLIMIT_NICE) ||
  		capable(CAP_SYS_NICE));
  }
@@ -84,9 +87,10 @@ Export the currently un-exported symbols they depend on.
  
  #ifdef __ARCH_WANT_SYS_NICE
  
---- a/kernel/signal.c
-+++ b/kernel/signal.c
-@@ -1353,6 +1353,7 @@ struct sighand_struct *__lock_task_sigha
+diff -uraN linux-5.7.8/kernel/signal.c omv-linux-5.7.8/kernel/signal.c
+--- linux-5.7.8/kernel/signal.c	2020-07-09 09:39:40.000000000 +0200
++++ omv-linux-5.7.8/kernel/signal.c	2020-07-13 15:18:04.854196194 +0200
+@@ -1396,6 +1396,7 @@
  
  	return sighand;
  }
@@ -94,19 +98,21 @@ Export the currently un-exported symbols they depend on.
  
  /*
   * send signal info to all the members of a group
---- a/kernel/task_work.c
-+++ b/kernel/task_work.c
-@@ -40,6 +40,7 @@ task_work_add(struct task_struct *task,
- 		set_notify_resume(task);
+diff -uraN linux-5.7.8/kernel/task_work.c omv-linux-5.7.8/kernel/task_work.c
+--- linux-5.7.8/kernel/task_work.c	2020-07-09 09:39:40.000000000 +0200
++++ omv-linux-5.7.8/kernel/task_work.c	2020-07-13 15:18:57.298331300 +0200
+@@ -52,6 +52,7 @@
+ 
  	return 0;
  }
 +EXPORT_SYMBOL(task_work_add);
  
  /**
   * task_work_cancel - cancel a pending work added by task_work_add()
---- a/mm/memory.c
-+++ b/mm/memory.c
-@@ -1364,6 +1364,7 @@ void zap_page_range(struct vm_area_struc
+diff -uraN linux-5.7.8/mm/memory.c omv-linux-5.7.8/mm/memory.c
+--- linux-5.7.8/mm/memory.c	2020-07-09 09:39:40.000000000 +0200
++++ omv-linux-5.7.8/mm/memory.c	2020-07-13 15:19:48.003475257 +0200
+@@ -1370,6 +1370,7 @@
  	mmu_notifier_invalidate_range_end(&range);
  	tlb_finish_mmu(&tlb, start, range.end);
  }
@@ -114,9 +120,10 @@ Export the currently un-exported symbols they depend on.
  
  /**
   * zap_page_range_single - remove user pages in a given range
---- a/mm/shmem.c
-+++ b/mm/shmem.c
-@@ -3992,6 +3992,7 @@ int shmem_zero_setup(struct vm_area_stru
+diff -uraN linux-5.7.8/mm/shmem.c omv-linux-5.7.8/mm/shmem.c
+--- linux-5.7.8/mm/shmem.c	2020-07-09 09:39:40.000000000 +0200
++++ omv-linux-5.7.8/mm/shmem.c	2020-07-13 15:20:27.829371644 +0200
+@@ -4185,6 +4185,7 @@
  
  	return 0;
  }
@@ -124,9 +131,10 @@ Export the currently un-exported symbols they depend on.
  
  /**
   * shmem_read_mapping_page_gfp - read into page cache, using specified page allocation flags.
---- a/mm/vmalloc.c
-+++ b/mm/vmalloc.c
-@@ -1295,6 +1295,7 @@ int map_kernel_range_noflush(unsigned lo
+diff -uraN linux-5.7.8/mm/vmalloc.c omv-linux-5.7.8/mm/vmalloc.c
+--- linux-5.7.8/mm/vmalloc.c	2020-07-09 09:39:40.000000000 +0200
++++ omv-linux-5.7.8/mm/vmalloc.c	2020-07-13 15:22:00.544919570 +0200
+@@ -2011,6 +2011,7 @@
  {
  	return vmap_page_range_noflush(addr, addr + size, prot, pages);
  }
@@ -134,7 +142,7 @@ Export the currently un-exported symbols they depend on.
  
  /**
   * unmap_kernel_range_noflush - unmap kernel VM area
-@@ -1435,6 +1436,7 @@ struct vm_struct *get_vm_area(unsigned l
+@@ -2161,6 +2162,7 @@
  				  NUMA_NO_NODE, GFP_KERNEL,
  				  __builtin_return_address(0));
  }
@@ -142,9 +150,10 @@ Export the currently un-exported symbols they depend on.
  
  struct vm_struct *get_vm_area_caller(unsigned long size, unsigned long flags,
  				const void *caller)
---- a/security/security.c
-+++ b/security/security.c
-@@ -247,24 +247,28 @@ int security_binder_set_context_mgr(stru
+diff -uraN linux-5.7.8/security/security.c omv-linux-5.7.8/security/security.c
+--- linux-5.7.8/security/security.c	2020-07-09 09:39:40.000000000 +0200
++++ omv-linux-5.7.8/security/security.c	2020-07-13 15:23:59.601954504 +0200
+@@ -725,24 +725,28 @@
  {
  	return call_int_hook(binder_set_context_mgr, 0, mgr);
  }
diff --git a/extra-wifi-drivers-compile.patch b/extra-wifi-drivers-compile.patch
new file mode 100644
index 0000000..e948733
--- /dev/null
+++ b/extra-wifi-drivers-compile.patch
@@ -0,0 +1,50 @@
+diff -up linux-5.5/drivers/net/wireless/rtl8723de/include/ieee80211.h.omv~ linux-5.5/drivers/net/wireless/rtl8723de/include/ieee80211.h
+--- linux-5.5/drivers/net/wireless/rtl8723de/include/ieee80211.h.omv~	2020-03-01 04:31:59.060616017 +0100
++++ linux-5.5/drivers/net/wireless/rtl8723de/include/ieee80211.h	2020-03-01 04:32:21.404617002 +0100
+@@ -1481,18 +1481,18 @@ enum ieee80211_state {
+ 	(((Addr[2]) & 0xff) == 0xff) && (((Addr[3]) & 0xff) == 0xff) && (((Addr[4]) & 0xff) == 0xff) && \
+ 				     (((Addr[5]) & 0xff) == 0xff))
+ #else
+-extern __inline int is_multicast_mac_addr(const u8 *addr)
++static inline int is_multicast_mac_addr(const u8 *addr)
+ {
+ 	return (addr[0] != 0xff) && (0x01 & addr[0]);
+ }
+ 
+-extern __inline int is_broadcast_mac_addr(const u8 *addr)
++static inline int is_broadcast_mac_addr(const u8 *addr)
+ {
+ 	return ((addr[0] == 0xff) && (addr[1] == 0xff) && (addr[2] == 0xff) &&   \
+ 		(addr[3] == 0xff) && (addr[4] == 0xff) && (addr[5] == 0xff));
+ }
+ 
+-extern __inline int is_zero_mac_addr(const u8 *addr)
++static inline int is_zero_mac_addr(const u8 *addr)
+ {
+ 	return ((addr[0] == 0x00) && (addr[1] == 0x00) && (addr[2] == 0x00) &&   \
+ 		(addr[3] == 0x00) && (addr[4] == 0x00) && (addr[5] == 0x00));
+diff -up linux-5.5/drivers/net/wireless/rtl8821ce/include/ieee80211.h.omv~ linux-5.5/drivers/net/wireless/rtl8821ce/include/ieee80211.h
+--- linux-5.5/drivers/net/wireless/rtl8821ce/include/ieee80211.h.omv~	2020-03-01 04:41:48.093641990 +0100
++++ linux-5.5/drivers/net/wireless/rtl8821ce/include/ieee80211.h	2020-03-01 04:42:05.069642739 +0100
+@@ -1581,18 +1581,18 @@ enum ieee80211_state {
+ 	(((Addr[2]) & 0xff) == 0xff) && (((Addr[3]) & 0xff) == 0xff) && (((Addr[4]) & 0xff) == 0xff) && \
+ 				     (((Addr[5]) & 0xff) == 0xff))
+ #else
+-extern __inline int is_multicast_mac_addr(const u8 *addr)
++static inline int is_multicast_mac_addr(const u8 *addr)
+ {
+ 	return (addr[0] != 0xff) && (0x01 & addr[0]);
+ }
+ 
+-extern __inline int is_broadcast_mac_addr(const u8 *addr)
++static inline int is_broadcast_mac_addr(const u8 *addr)
+ {
+ 	return ((addr[0] == 0xff) && (addr[1] == 0xff) && (addr[2] == 0xff) &&   \
+ 		(addr[3] == 0xff) && (addr[4] == 0xff) && (addr[5] == 0xff));
+ }
+ 
+-extern __inline int is_zero_mac_addr(const u8 *addr)
++static inline int is_zero_mac_addr(const u8 *addr)
+ {
+ 	return ((addr[0] == 0x00) && (addr[1] == 0x00) && (addr[2] == 0x00) &&   \
+ 		(addr[3] == 0x00) && (addr[4] == 0x00) && (addr[5] == 0x00));
diff --git a/extra-wifi-drivers-port-to-5.6.patch b/extra-wifi-drivers-port-to-5.6.patch
new file mode 100644
index 0000000..adcce3b
--- /dev/null
+++ b/extra-wifi-drivers-port-to-5.6.patch
@@ -0,0 +1,412 @@
+diff -up linux-5.6-rc4/drivers/net/wireless/rtl8723de/os_dep/linux/rtw_proc.c.omv~ linux-5.6-rc4/drivers/net/wireless/rtl8723de/os_dep/linux/rtw_proc.c
+--- linux-5.6-rc4/drivers/net/wireless/rtl8723de/os_dep/linux/rtw_proc.c.omv~	2020-03-02 19:01:13.164067081 +0100
++++ linux-5.6-rc4/drivers/net/wireless/rtl8723de/os_dep/linux/rtw_proc.c	2020-03-02 23:08:55.704439445 +0100
+@@ -69,12 +69,12 @@ inline struct proc_dir_entry *rtw_proc_c
+ }
+ 
+ inline struct proc_dir_entry *rtw_proc_create_entry(const char *name, struct proc_dir_entry *parent,
+-	const struct file_operations *fops, void * data)
++	const struct proc_ops *ops, void * data)
+ {
+ 	struct proc_dir_entry *entry;
+ 
+ #if (LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 26))
+-	entry = proc_create_data(name,  S_IFREG | S_IRUGO | S_IWUGO, parent, fops, data);
++	entry = proc_create_data(name,  S_IFREG | S_IRUGO | S_IWUGO, parent, ops, data);
+ #else
+ 	entry = create_proc_entry(name, S_IFREG | S_IRUGO | S_IWUGO, parent);
+ 	if (entry) {
+@@ -219,22 +219,20 @@ static ssize_t rtw_drv_proc_write(struct
+ 	return -EROFS;
+ }
+ 
+-static const struct file_operations rtw_drv_proc_seq_fops = {
+-	.owner = THIS_MODULE,
+-	.open = rtw_drv_proc_open,
+-	.read = seq_read,
+-	.llseek = seq_lseek,
+-	.release = seq_release,
+-	.write = rtw_drv_proc_write,
++static const struct proc_ops rtw_drv_proc_seq_ops = {
++	.proc_open = rtw_drv_proc_open,
++	.proc_read = seq_read,
++	.proc_lseek = seq_lseek,
++	.proc_release = seq_release,
++	.proc_write = rtw_drv_proc_write,
+ };
+ 
+-static const struct file_operations rtw_drv_proc_sseq_fops = {
+-	.owner = THIS_MODULE,
+-	.open = rtw_drv_proc_open,
+-	.read = seq_read,
+-	.llseek = seq_lseek,
+-	.release = single_release,
+-	.write = rtw_drv_proc_write,
++static const struct proc_ops rtw_drv_proc_sseq_ops = {
++	.proc_open = rtw_drv_proc_open,
++	.proc_read = seq_read,
++	.proc_lseek = seq_lseek,
++	.proc_release = single_release,
++	.proc_write = rtw_drv_proc_write,
+ };
+ 
+ int rtw_drv_proc_init(void)
+@@ -257,9 +255,9 @@ int rtw_drv_proc_init(void)
+ 
+ 	for (i = 0; i < drv_proc_hdls_num; i++) {
+ 		if (drv_proc_hdls[i].type == RTW_PROC_HDL_TYPE_SEQ)
+-			entry = rtw_proc_create_entry(drv_proc_hdls[i].name, rtw_proc, &rtw_drv_proc_seq_fops, (void *)i);
++			entry = rtw_proc_create_entry(drv_proc_hdls[i].name, rtw_proc, &rtw_drv_proc_seq_ops, (void *)i);
+ 		else if (drv_proc_hdls[i].type == RTW_PROC_HDL_TYPE_SSEQ)
+-			entry = rtw_proc_create_entry(drv_proc_hdls[i].name, rtw_proc, &rtw_drv_proc_sseq_fops, (void *)i);
++			entry = rtw_proc_create_entry(drv_proc_hdls[i].name, rtw_proc, &rtw_drv_proc_sseq_ops, (void *)i);
+ 		else
+ 			entry = NULL;
+ 
+@@ -2335,22 +2333,20 @@ static ssize_t rtw_adapter_proc_write(st
+ 	return -EROFS;
+ }
+ 
+-static const struct file_operations rtw_adapter_proc_seq_fops = {
+-	.owner = THIS_MODULE,
+-	.open = rtw_adapter_proc_open,
+-	.read = seq_read,
+-	.llseek = seq_lseek,
+-	.release = seq_release,
+-	.write = rtw_adapter_proc_write,
++static const struct proc_ops rtw_adapter_proc_seq_ops = {
++	.proc_open = rtw_adapter_proc_open,
++	.proc_read = seq_read,
++	.proc_lseek = seq_lseek,
++	.proc_release = seq_release,
++	.proc_write = rtw_adapter_proc_write,
+ };
+ 
+-static const struct file_operations rtw_adapter_proc_sseq_fops = {
+-	.owner = THIS_MODULE,
+-	.open = rtw_adapter_proc_open,
+-	.read = seq_read,
+-	.llseek = seq_lseek,
+-	.release = single_release,
+-	.write = rtw_adapter_proc_write,
++static const struct proc_ops rtw_adapter_proc_sseq_ops = {
++	.proc_open = rtw_adapter_proc_open,
++	.proc_read = seq_read,
++	.proc_lseek = seq_lseek,
++	.proc_release = single_release,
++	.proc_write = rtw_adapter_proc_write,
+ };
+ 
+ int proc_get_odm_dbg_comp(struct seq_file *m, void *v)
+@@ -2670,22 +2666,20 @@ static ssize_t rtw_odm_proc_write(struct
+ 	return -EROFS;
+ }
+ 
+-static const struct file_operations rtw_odm_proc_seq_fops = {
+-	.owner = THIS_MODULE,
+-	.open = rtw_odm_proc_open,
+-	.read = seq_read,
+-	.llseek = seq_lseek,
+-	.release = seq_release,
+-	.write = rtw_odm_proc_write,
++static const struct proc_ops rtw_odm_proc_seq_ops = {
++	.proc_open = rtw_odm_proc_open,
++	.proc_read = seq_read,
++	.proc_lseek = seq_lseek,
++	.proc_release = seq_release,
++	.proc_write = rtw_odm_proc_write,
+ };
+ 
+-static const struct file_operations rtw_odm_proc_sseq_fops = {
+-	.owner = THIS_MODULE,
+-	.open = rtw_odm_proc_open,
+-	.read = seq_read,
+-	.llseek = seq_lseek,
+-	.release = single_release,
+-	.write = rtw_odm_proc_write,
++static const struct proc_ops rtw_odm_proc_sseq_ops = {
++	.proc_open = rtw_odm_proc_open,
++	.proc_read = seq_read,
++	.proc_lseek = seq_lseek,
++	.proc_release = single_release,
++	.proc_write = rtw_odm_proc_write,
+ };
+ 
+ struct proc_dir_entry *rtw_odm_proc_init(struct net_device *dev)
+@@ -2715,9 +2709,9 @@ struct proc_dir_entry *rtw_odm_proc_init
+ 
+ 	for (i = 0; i < odm_proc_hdls_num; i++) {
+ 		if (odm_proc_hdls[i].type == RTW_PROC_HDL_TYPE_SEQ)
+-			entry = rtw_proc_create_entry(odm_proc_hdls[i].name, dir_odm, &rtw_odm_proc_seq_fops, (void *)i);
++			entry = rtw_proc_create_entry(odm_proc_hdls[i].name, dir_odm, &rtw_odm_proc_seq_ops, (void *)i);
+ 		else if (odm_proc_hdls[i].type == RTW_PROC_HDL_TYPE_SSEQ)
+-			entry = rtw_proc_create_entry(odm_proc_hdls[i].name, dir_odm, &rtw_odm_proc_sseq_fops, (void *)i);
++			entry = rtw_proc_create_entry(odm_proc_hdls[i].name, dir_odm, &rtw_odm_proc_sseq_ops, (void *)i);
+ 		else
+ 			entry = NULL;
+ 
+@@ -2809,22 +2803,20 @@ static ssize_t rtw_mcc_proc_write(struct
+ 	return -EROFS;
+ }
+ 
+-static const struct file_operations rtw_mcc_proc_seq_fops = {
+-	.owner = THIS_MODULE,
+-	.open = rtw_mcc_proc_open,
+-	.read = seq_read,
+-	.llseek = seq_lseek,
+-	.release = seq_release,
+-	.write = rtw_mcc_proc_write,
++static const struct file_operations rtw_mcc_proc_seq_ops = {
++	.proc_open = rtw_mcc_proc_open,
++	.proc_read = seq_read,
++	.proc_lseek = seq_lseek,
++	.proc_release = seq_release,
++	.proc_write = rtw_mcc_proc_write,
+ };
+ 
+-static const struct file_operations rtw_mcc_proc_sseq_fops = {
+-	.owner = THIS_MODULE,
+-	.open = rtw_mcc_proc_open,
+-	.read = seq_read,
+-	.llseek = seq_lseek,
+-	.release = single_release,
+-	.write = rtw_mcc_proc_write,
++static const struct file_operations rtw_mcc_proc_sseq_ops = {
++	.proc_open = rtw_mcc_proc_open,
++	.proc_read = seq_read,
++	.proc_lseek = seq_lseek,
++	.proc_release = single_release,
++	.proc_write = rtw_mcc_proc_write,
+ };
+ 
+ struct proc_dir_entry *rtw_mcc_proc_init(struct net_device *dev)
+@@ -2854,9 +2846,9 @@ struct proc_dir_entry *rtw_mcc_proc_init
+ 
+ 	for (i = 0; i < mcc_proc_hdls_num; i++) {
+ 		if (mcc_proc_hdls[i].type == RTW_PROC_HDL_TYPE_SEQ)
+-			entry = rtw_proc_create_entry(mcc_proc_hdls[i].name, dir_mcc, &rtw_mcc_proc_seq_fops, (void *)i);
++			entry = rtw_proc_create_entry(mcc_proc_hdls[i].name, dir_mcc, &rtw_mcc_proc_seq_ops, (void *)i);
+ 		else if (mcc_proc_hdls[i].type == RTW_PROC_HDL_TYPE_SSEQ)
+-			entry = rtw_proc_create_entry(mcc_proc_hdls[i].name, dir_mcc, &rtw_mcc_proc_sseq_fops, (void *)i);
++			entry = rtw_proc_create_entry(mcc_proc_hdls[i].name, dir_mcc, &rtw_mcc_proc_sseq_ops, (void *)i);
+ 		else
+ 			entry = NULL;
+ 
+@@ -2920,9 +2912,9 @@ struct proc_dir_entry *rtw_adapter_proc_
+ 
+ 	for (i = 0; i < adapter_proc_hdls_num; i++) {
+ 		if (adapter_proc_hdls[i].type == RTW_PROC_HDL_TYPE_SEQ)
+-			entry = rtw_proc_create_entry(adapter_proc_hdls[i].name, dir_dev, &rtw_adapter_proc_seq_fops, (void *)i);
++			entry = rtw_proc_create_entry(adapter_proc_hdls[i].name, dir_dev, &rtw_adapter_proc_seq_ops, (void *)i);
+ 		else if (adapter_proc_hdls[i].type == RTW_PROC_HDL_TYPE_SSEQ)
+-			entry = rtw_proc_create_entry(adapter_proc_hdls[i].name, dir_dev, &rtw_adapter_proc_sseq_fops, (void *)i);
++			entry = rtw_proc_create_entry(adapter_proc_hdls[i].name, dir_dev, &rtw_adapter_proc_sseq_ops, (void *)i);
+ 		else
+ 			entry = NULL;
+ 
+diff -up linux-5.6-rc4/drivers/net/wireless/rtl8821ce/os_dep/linux/rtw_proc.c.omv~ linux-5.6-rc4/drivers/net/wireless/rtl8821ce/os_dep/linux/rtw_proc.c
+--- linux-5.6-rc4/drivers/net/wireless/rtl8821ce/os_dep/linux/rtw_proc.c.omv~	2020-03-02 23:04:39.749450254 +0100
++++ linux-5.6-rc4/drivers/net/wireless/rtl8821ce/os_dep/linux/rtw_proc.c	2020-03-02 23:08:18.209441028 +0100
+@@ -62,12 +62,12 @@ inline struct proc_dir_entry *rtw_proc_c
+ }
+ 
+ inline struct proc_dir_entry *rtw_proc_create_entry(const char *name, struct proc_dir_entry *parent,
+-	const struct file_operations *fops, void * data)
++	const struct proc_ops *ops, void * data)
+ {
+ 	struct proc_dir_entry *entry;
+ 
+ #if (LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 26))
+-	entry = proc_create_data(name,  S_IFREG | S_IRUGO | S_IWUGO, parent, fops, data);
++	entry = proc_create_data(name,  S_IFREG | S_IRUGO | S_IWUGO, parent, ops, data);
+ #else
+ 	entry = create_proc_entry(name, S_IFREG | S_IRUGO | S_IWUGO, parent);
+ 	if (entry) {
+@@ -238,22 +238,20 @@ static ssize_t rtw_drv_proc_write(struct
+ 	return -EROFS;
+ }
+ 
+-static const struct file_operations rtw_drv_proc_seq_fops = {
+-	.owner = THIS_MODULE,
+-	.open = rtw_drv_proc_open,
+-	.read = seq_read,
+-	.llseek = seq_lseek,
+-	.release = seq_release,
+-	.write = rtw_drv_proc_write,
++static const struct proc_ops rtw_drv_proc_seq_ops = {
++	.proc_open = rtw_drv_proc_open,
++	.proc_read = seq_read,
++	.proc_lseek = seq_lseek,
++	.proc_release = seq_release,
++	.proc_write = rtw_drv_proc_write,
+ };
+ 
+-static const struct file_operations rtw_drv_proc_sseq_fops = {
+-	.owner = THIS_MODULE,
+-	.open = rtw_drv_proc_open,
+-	.read = seq_read,
+-	.llseek = seq_lseek,
+-	.release = single_release,
+-	.write = rtw_drv_proc_write,
++static const struct proc_ops rtw_drv_proc_sseq_ops = {
++	.proc_open = rtw_drv_proc_open,
++	.proc_read = seq_read,
++	.proc_lseek = seq_lseek,
++	.proc_release = single_release,
++	.proc_write = rtw_drv_proc_write,
+ };
+ 
+ int rtw_drv_proc_init(void)
+@@ -276,9 +274,9 @@ int rtw_drv_proc_init(void)
+ 
+ 	for (i = 0; i < drv_proc_hdls_num; i++) {
+ 		if (drv_proc_hdls[i].type == RTW_PROC_HDL_TYPE_SEQ)
+-			entry = rtw_proc_create_entry(drv_proc_hdls[i].name, rtw_proc, &rtw_drv_proc_seq_fops, (void *)i);
++			entry = rtw_proc_create_entry(drv_proc_hdls[i].name, rtw_proc, &rtw_drv_proc_seq_ops, (void *)i);
+ 		else if (drv_proc_hdls[i].type == RTW_PROC_HDL_TYPE_SSEQ)
+-			entry = rtw_proc_create_entry(drv_proc_hdls[i].name, rtw_proc, &rtw_drv_proc_sseq_fops, (void *)i);
++			entry = rtw_proc_create_entry(drv_proc_hdls[i].name, rtw_proc, &rtw_drv_proc_sseq_ops, (void *)i);
+ 		else
+ 			entry = NULL;
+ 
+@@ -3899,22 +3897,20 @@ static ssize_t rtw_adapter_proc_write(st
+ 	return -EROFS;
+ }
+ 
+-static const struct file_operations rtw_adapter_proc_seq_fops = {
+-	.owner = THIS_MODULE,
+-	.open = rtw_adapter_proc_open,
+-	.read = seq_read,
+-	.llseek = seq_lseek,
+-	.release = seq_release,
+-	.write = rtw_adapter_proc_write,
++static const struct proc_ops rtw_adapter_proc_seq_ops = {
++	.proc_open = rtw_adapter_proc_open,
++	.proc_read = seq_read,
++	.proc_lseek = seq_lseek,
++	.proc_release = seq_release,
++	.proc_write = rtw_adapter_proc_write,
+ };
+ 
+-static const struct file_operations rtw_adapter_proc_sseq_fops = {
+-	.owner = THIS_MODULE,
+-	.open = rtw_adapter_proc_open,
+-	.read = seq_read,
+-	.llseek = seq_lseek,
+-	.release = single_release,
+-	.write = rtw_adapter_proc_write,
++static const struct proc_ops rtw_adapter_proc_sseq_ops = {
++	.proc_open = rtw_adapter_proc_open,
++	.proc_read = seq_read,
++	.proc_lseek = seq_lseek,
++	.proc_release = single_release,
++	.proc_write = rtw_adapter_proc_write,
+ };
+ 
+ int proc_get_odm_adaptivity(struct seq_file *m, void *v)
+@@ -4068,22 +4064,20 @@ static ssize_t rtw_odm_proc_write(struct
+ 	return -EROFS;
+ }
+ 
+-static const struct file_operations rtw_odm_proc_seq_fops = {
+-	.owner = THIS_MODULE,
+-	.open = rtw_odm_proc_open,
+-	.read = seq_read,
+-	.llseek = seq_lseek,
+-	.release = seq_release,
+-	.write = rtw_odm_proc_write,
++static const struct proc_ops rtw_odm_proc_seq_ops = {
++	.proc_open = rtw_odm_proc_open,
++	.proc_read = seq_read,
++	.proc_lseek = seq_lseek,
++	.proc_release = seq_release,
++	.proc_write = rtw_odm_proc_write,
+ };
+ 
+-static const struct file_operations rtw_odm_proc_sseq_fops = {
+-	.owner = THIS_MODULE,
+-	.open = rtw_odm_proc_open,
+-	.read = seq_read,
+-	.llseek = seq_lseek,
+-	.release = single_release,
+-	.write = rtw_odm_proc_write,
++static const struct proc_ops rtw_odm_proc_sseq_ops = {
++	.proc_open = rtw_odm_proc_open,
++	.proc_read = seq_read,
++	.proc_lseek = seq_lseek,
++	.proc_release = single_release,
++	.proc_write = rtw_odm_proc_write,
+ };
+ 
+ struct proc_dir_entry *rtw_odm_proc_init(struct net_device *dev)
+@@ -4113,9 +4107,9 @@ struct proc_dir_entry *rtw_odm_proc_init
+ 
+ 	for (i = 0; i < odm_proc_hdls_num; i++) {
+ 		if (odm_proc_hdls[i].type == RTW_PROC_HDL_TYPE_SEQ)
+-			entry = rtw_proc_create_entry(odm_proc_hdls[i].name, dir_odm, &rtw_odm_proc_seq_fops, (void *)i);
++			entry = rtw_proc_create_entry(odm_proc_hdls[i].name, dir_odm, &rtw_odm_proc_seq_ops, (void *)i);
+ 		else if (odm_proc_hdls[i].type == RTW_PROC_HDL_TYPE_SSEQ)
+-			entry = rtw_proc_create_entry(odm_proc_hdls[i].name, dir_odm, &rtw_odm_proc_sseq_fops, (void *)i);
++			entry = rtw_proc_create_entry(odm_proc_hdls[i].name, dir_odm, &rtw_odm_proc_sseq_ops, (void *)i);
+ 		else
+ 			entry = NULL;
+ 
+@@ -4209,22 +4203,20 @@ static ssize_t rtw_mcc_proc_write(struct
+ 	return -EROFS;
+ }
+ 
+-static const struct file_operations rtw_mcc_proc_seq_fops = {
+-	.owner = THIS_MODULE,
+-	.open = rtw_mcc_proc_open,
+-	.read = seq_read,
+-	.llseek = seq_lseek,
+-	.release = seq_release,
+-	.write = rtw_mcc_proc_write,
++static const struct proc_ops rtw_mcc_proc_seq_ops = {
++	.proc_open = rtw_mcc_proc_open,
++	.proc_read = seq_read,
++	.proc_lseek = seq_lseek,
++	.proc_release = seq_release,
++	.proc_write = rtw_mcc_proc_write,
+ };
+ 
+-static const struct file_operations rtw_mcc_proc_sseq_fops = {
+-	.owner = THIS_MODULE,
+-	.open = rtw_mcc_proc_open,
+-	.read = seq_read,
+-	.llseek = seq_lseek,
+-	.release = single_release,
+-	.write = rtw_mcc_proc_write,
++static const struct proc_ops rtw_mcc_proc_sseq_ops = {
++	.proc_open = rtw_mcc_proc_open,
++	.proc_read = seq_read,
++	.proc_lseek = seq_lseek,
++	.proc_release = single_release,
++	.proc_write = rtw_mcc_proc_write,
+ };
+ 
+ struct proc_dir_entry *rtw_mcc_proc_init(struct net_device *dev)
+@@ -4254,9 +4246,9 @@ struct proc_dir_entry *rtw_mcc_proc_init
+ 
+ 	for (i = 0; i < mcc_proc_hdls_num; i++) {
+ 		if (mcc_proc_hdls[i].type == RTW_PROC_HDL_TYPE_SEQ)
+-			entry = rtw_proc_create_entry(mcc_proc_hdls[i].name, dir_mcc, &rtw_mcc_proc_seq_fops, (void *)i);
++			entry = rtw_proc_create_entry(mcc_proc_hdls[i].name, dir_mcc, &rtw_mcc_proc_seq_ops, (void *)i);
+ 		else if (mcc_proc_hdls[i].type == RTW_PROC_HDL_TYPE_SSEQ)
+-			entry = rtw_proc_create_entry(mcc_proc_hdls[i].name, dir_mcc, &rtw_mcc_proc_sseq_fops, (void *)i);
++			entry = rtw_proc_create_entry(mcc_proc_hdls[i].name, dir_mcc, &rtw_mcc_proc_sseq_ops, (void *)i);
+ 		else
+ 			entry = NULL;
+ 
+@@ -4319,9 +4311,9 @@ struct proc_dir_entry *rtw_adapter_proc_
+ 
+ 	for (i = 0; i < adapter_proc_hdls_num; i++) {
+ 		if (adapter_proc_hdls[i].type == RTW_PROC_HDL_TYPE_SEQ)
+-			entry = rtw_proc_create_entry(adapter_proc_hdls[i].name, dir_dev, &rtw_adapter_proc_seq_fops, (void *)i);
++			entry = rtw_proc_create_entry(adapter_proc_hdls[i].name, dir_dev, &rtw_adapter_proc_seq_ops, (void *)i);
+ 		else if (adapter_proc_hdls[i].type == RTW_PROC_HDL_TYPE_SSEQ)
+-			entry = rtw_proc_create_entry(adapter_proc_hdls[i].name, dir_dev, &rtw_adapter_proc_sseq_fops, (void *)i);
++			entry = rtw_proc_create_entry(adapter_proc_hdls[i].name, dir_dev, &rtw_adapter_proc_sseq_ops, (void *)i);
+ 		else
+ 			entry = NULL;
+ 
diff --git a/kernel-5.6-kvm-gcc10.patch b/kernel-5.6-kvm-gcc10.patch
new file mode 100644
index 0000000..bb29126
--- /dev/null
+++ b/kernel-5.6-kvm-gcc10.patch
@@ -0,0 +1,12 @@
+diff -up linux-5.6/virt/kvm/kvm_main.c.omv~ linux-5.6/virt/kvm/kvm_main.c
+--- linux-5.6/virt/kvm/kvm_main.c.omv~	2020-04-17 21:04:40.240584475 +0200
++++ linux-5.6/virt/kvm/kvm_main.c	2020-04-17 21:04:43.150676715 +0200
+@@ -2219,7 +2219,7 @@ static int __kvm_gfn_to_hva_cache_init(s
+ 	gfn_t start_gfn = gpa >> PAGE_SHIFT;
+ 	gfn_t end_gfn = (gpa + len - 1) >> PAGE_SHIFT;
+ 	gfn_t nr_pages_needed = end_gfn - start_gfn + 1;
+-	gfn_t nr_pages_avail;
++	gfn_t nr_pages_avail = 0;
+ 
+ 	/* Update ghc->generation before performing any error checks. */
+ 	ghc->generation = slots->generation;
diff --git a/kernel-5.7-fewer-conditions-for-ARM64_PTR_AUTH.patch b/kernel-5.7-fewer-conditions-for-ARM64_PTR_AUTH.patch
new file mode 100644
index 0000000..63ed69a
--- /dev/null
+++ b/kernel-5.7-fewer-conditions-for-ARM64_PTR_AUTH.patch
@@ -0,0 +1,27 @@
+This is wrong and should never go upstream.
+
+It is needed nevertheless: Without this patch, the config builder doesn't emit
+any value for ARM64_PTR_AUTH because the preconditions aren't met.
+Then on "real" make, it knows more about the toolchain, sees it can be
+enabled, and "make syncconfig" hangs prompting for the value.
+
+Given our toolchain is known good (we'll always have binutils >= 2.33.1 and
+current versions of gcc and clang), inside our package disabling those
+preconditions doesn't do any harm.
+
+diff -up linux-5.7/arch/arm64/Kconfig.omv~ linux-5.7/arch/arm64/Kconfig
+--- linux-5.7/arch/arm64/Kconfig.omv~	2020-07-25 23:31:02.229462957 +0000
++++ linux-5.7/arch/arm64/Kconfig	2020-07-25 23:31:12.161907200 +0000
+@@ -1502,12 +1502,6 @@ config ARM64_PTR_AUTH
+ 	bool "Enable support for pointer authentication"
+ 	default y
+ 	depends on !KVM || ARM64_VHE
+-	depends on (CC_HAS_SIGN_RETURN_ADDRESS || CC_HAS_BRANCH_PROT_PAC_RET) && AS_HAS_PAC
+-	# GCC 9.1 and later inserts a .note.gnu.property section note for PAC
+-	# which is only understood by binutils starting with version 2.33.1.
+-	depends on !CC_IS_GCC || GCC_VERSION < 90100 || LD_VERSION >= 233010000
+-	depends on !CC_IS_CLANG || AS_HAS_CFI_NEGATE_RA_STATE
+-	depends on (!FUNCTION_GRAPH_TRACER || DYNAMIC_FTRACE_WITH_REGS)
+ 	help
+ 	  Pointer authentication (part of the ARMv8.3 Extensions) provides
+ 	  instructions for signing and authenticating pointers against secret
diff --git a/kernel-5.8-nouveau-write-combining-only-on-x86.patch b/kernel-5.8-nouveau-write-combining-only-on-x86.patch
new file mode 100644
index 0000000..41aa007
--- /dev/null
+++ b/kernel-5.8-nouveau-write-combining-only-on-x86.patch
@@ -0,0 +1,72 @@
+commit bb14834cef399ae99771f0f33fe6b57a68165370
+Author: Bernhard Rosenkränzer <bero@lindev.ch>
+Date:   Tue Jul 21 15:45:49 2020 +0200
+
+    nouveau: avoid write-combine mappings for VRAM and IO space
+    
+    NOUPSTREAM
+    
+    Author: Ard Biesheuvel <ard.biesheuvel@linaro.org>
+    Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
+    Signed-off-by: Graeme Gregory <graeme.gregory@linaro.org>
+    Signed-off-by: Bernhard Rosenkränzer <bero@lindev.ch>
+    [daniel.thompson@linaro.org:
+      Updated to undo changes to the bar2 mapping in v4.15-rc1, see
+      dfcbd5506817 ("drm/nouveau/imem/nv50: map bar2 write-combined")
+      and b807270cbd87 ("drm/nouveau/imem/nv40: map bar2
+      write-combined") for more details]
+    Signed-off-by: Daniel Thompson <daniel.thompson@linaro.org>
+
+diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
+index c40f127de3d0b2..525b2dc1ed91af 100644
+--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
++++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
+@@ -662,9 +662,14 @@ nouveau_bo_init_mem_type(struct ttm_bo_device *bdev, uint32_t type,
+ 	case TTM_PL_VRAM:
+ 		man->flags = TTM_MEMTYPE_FLAG_FIXED |
+ 			     TTM_MEMTYPE_FLAG_MAPPABLE;
++#ifdef CONFIG_X86
+ 		man->available_caching = TTM_PL_FLAG_UNCACHED |
+ 					 TTM_PL_FLAG_WC;
+ 		man->default_caching = TTM_PL_FLAG_WC;
++#else
++		man->available_caching = TTM_PL_FLAG_UNCACHED;
++		man->default_caching = TTM_PL_FLAG_UNCACHED;
++#endif
+ 
+ 		if (drm->client.device.info.family >= NV_DEVICE_INFO_V0_TESLA) {
+ 			/* Some BARs do not support being ioremapped WC */
+diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv40.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv40.c
+index 086c118488ef5f..d7e4c0e325afb1 100644
+--- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv40.c
++++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv40.c
+@@ -253,8 +253,13 @@ nv40_instmem_new(struct nvkm_device *device, int index,
+ 	else
+ 		bar = 3;
+ 
++#ifdef CONFIG_X86
+	imem->iomem = ioremap_wc(device->func->resource_addr(device, bar),
+				 device->func->resource_size(device, bar));
++#else
++	imem->iomem = ioremap(device->func->resource_addr(device, bar),
++			      device->func->resource_size(device, bar));
++#endif
+ 	if (!imem->iomem) {
+ 		nvkm_error(&imem->base.subdev, "unable to map PRAMIN BAR\n");
+ 		return -EFAULT;
+diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c
+index 02c4eb28cef44d..80a28164c15041 100644
+--- a/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c
++++ b/drivers/gpu/drm/nouveau/nvkm/subdev/instmem/nv50.c
+@@ -171,7 +171,11 @@ nv50_instobj_kmap(struct nv50_instobj *iobj, struct nvkm_vmm *vmm)
+ 
+ 	/* Make the mapping visible to the host. */
+ 	iobj->bar = bar;
++#ifdef CONFIG_X86
+ 	iobj->map = ioremap_wc(device->func->resource_addr(device, 3) +
++#else
++	iobj->map = ioremap(device->func->resource_addr(device, 3) +
++#endif
+ 			       (u32)iobj->bar->addr, size);
+ 	if (!iobj->map) {
+ 		nvkm_warn(subdev, "PRAMIN ioremap failed\n");
diff --git a/kernel-release.spec b/kernel-release.spec
index c1cba9e..fe33307 100644
--- a/kernel-release.spec
+++ b/kernel-release.spec
@@ -4,8 +4,8 @@
 #end
 %define _disable_ld_no_undefined 1
 
-# (tpg) try to speed up things
-%global optflags %{optflags} -O3
+## STOP: Adding weird and unsupported upstream kernel C/LD flags of any sort
+## yes , including ftlo . O3 and whatever else
 
 # (crazy) , well that new way of doing buil-id symlinks
 # does not seems to work, see:
@@ -17,22 +17,22 @@
 # This is the place where you set kernel version i.e 4.5.0
 # compose tar.xz name and release
 %define kernelversion	5
-%define patchlevel	5
-%define sublevel	2
+%define patchlevel	7
+%define sublevel	11
 %define relc		%{nil}
 # Only ever wrong on x.0 releases...
 %define previous	%{kernelversion}.%(echo $((%{patchlevel}-1)))
 
 %define buildrel	%{kversion}-%{buildrpmrel}
-%define rpmtag	%{disttag}
+%define rpmtag		%{disttag}
 
 # IMPORTANT
 # This is the place where you set release version %{version}-1omv2015
 %if 0%{relc}
 %define rpmrel		0.rc%{relc}.1
-%define tar_ver   	%{kernelversion}.%{patchlevel}-rc%{relc}
+%define tar_ver		%{kernelversion}.%{patchlevel}-rc%{relc}
 %else
-%define rpmrel		1
+%define rpmrel		2
 %define tar_ver		%{kernelversion}.%{patchlevel}
 %endif
 %define buildrpmrel	%{rpmrel}%{rpmtag}
@@ -65,7 +65,7 @@
 # Build defines
 %bcond_with build_doc
 %ifarch %{ix86} %{x86_64}
-%bcond_with uksm
+%bcond_without uksm
 %else
 %bcond_with uksm
 %endif
@@ -73,7 +73,12 @@
 %bcond_without build_devel
 %bcond_with build_debug
 %bcond_with clang
-%bcond_with bootsplash
+## enabled it runs dracut -f --regenerate-all
+## we *should* enable that, is bc we keep or can keep lots
+## kernel around and the initrd is created using sys libs, sys configs,
+## *systemd* service & apps etc. IOW, a old initrd may have old files, libs, etc
+## changed since last rebuild and may result in either broken boot, or very hard to debug bugs.
+%bcond_with dracut_all_initrd
 # (tpg) enable patches from ClearLinux
 %bcond_without clr
 %if %mdvver > 3000000
@@ -82,7 +87,7 @@
 %bcond_with cross_headers
 %endif
 
-%global cross_header_archs	aarch64-linux armv7hnl-linux i686-linux x86_64-linux x32-linux riscv32-linux riscv64-linux aarch64-linuxmusl armv7hnl-linuxmusl i686-linuxmusl x86_64-linuxmusl x32-linuxmusl riscv32-linuxmusl riscv64-linuxmusl aarch64-android armv7l-android armv8l-android
+%global cross_header_archs	aarch64-linux armv7hnl-linux i686-linux x86_64-linux x32-linux riscv32-linux riscv64-linux aarch64-linuxmusl armv7hnl-linuxmusl i686-linuxmusl x86_64-linuxmusl x32-linuxmusl riscv32-linuxmusl riscv64-linuxmusl aarch64-android armv7l-android armv8l-android x86_64-android aarch64-linuxuclibc armv7hnl-linuxuclibc i686-linuxuclibc x86_64-linuxuclibc x32-linuxuclibc riscv32-linuxuclibc riscv64-linuxuclibc
 %global long_cross_header_archs %(
 	for i in %{cross_header_archs}; do
 		CPU=$(echo $i |cut -d- -f1)
@@ -120,7 +125,6 @@
 # unfortunately kmod does not support Zstandard for now, so kernel modules
 # compressed with zstd will not bo loaded and system will fail
 # https://github.com/facebook/zstd/issues/1121
-# Currently only supported on x86
 %ifarch %{ix86} %{x86_64}
 %bcond_without build_modzstd
 # compress modules with XZ
@@ -222,84 +226,29 @@ Source51:	cpupower.config
 %if 0%{sublevel}
 Source90:	https://cdn.kernel.org/pub/linux/kernel/v%(echo %{version}|cut -d. -f1).x/patch-%{version}.xz
 %endif
+Patch1:		linux-5.6-fix-disassembler-4args-detection.patch
 Patch2:		die-floppy-die.patch
 Patch3:		0001-Add-support-for-Acer-Predator-macro-keys.patch
 Patch4:		linux-4.7-intel-dvi-duallink.patch
-Patch5:		linux-4.8.1-buildfix.patch
+Patch5:		kernel-5.6-kvm-gcc10.patch
 Patch6:		linux-5.2.9-riscv-compile.patch
 # Work around rpm dependency generator screaming about
 # error: Illegal char ']' (0x5d) in: 1.2.1[50983]_custom
 # caused by aacraid versioning ("1.2.1[50983]-custom")
 Patch7:		aacraid-dont-freak-out-dependency-generator.patch
-
-%if %{with clang}
-# Patches to make it build with clang
-Patch1000:	0001-kbuild-LLVMLinux-Set-compiler-flags-for-clang.patch
-Patch1001:	0002-fs-LLVMLinux-Remove-warning-from-COMPATIBLE_IOCTL.patch
-Patch1002:	0003-kbuild-LLVMLinux-Add-support-for-generating-LLVM-bit.patch
-Patch1003:	0004-kbuild-LLVMLinux-Make-asm-offset-generation-work-wit.patch
-Patch1004:	0005-md-sysfs-LLVMLinux-Remove-nested-function-from-bcach.patch
-Patch1005:	0006-apparmor-LLVMLinux-Remove-VLAIS.patch
-Patch1006:	0007-exofs-LLVMLinux-Remove-VLAIS-from-exofs-FIXME-Check-.patch
-Patch1007:	0008-md-raid10-LLVMLinux-Remove-VLAIS-from-raid10-driver.patch
-Patch1008:	0009-fs-nfs-LLVMLinux-Remove-VLAIS-from-nfs.patch
-Patch1009:	0010-net-wimax-i2400-LLVMLinux-Remove-VLAIS-from-wimax-i2.patch
-Patch1010:	0011-Kbuild-LLVMLinux-Use-Oz-instead-of-Os-when-using-cla.patch
-Patch1011:	0012-WORKAROUND-x86-boot-LLVMLinux-Work-around-clang-PR39.patch
-Patch1012:	0013-DO-NOT-UPSTREAM-xen-LLVMLinux-Remove-VLAIS-from-xen-.patch
-Patch1013:	0014-DO-NOT-UPSTREAM-arm-LLVMLinux-Provide-__aeabi_-symbo.patch
-Patch1014:	0015-DO-NOT-UPSTREAM-arm-firmware-LLVMLinux-replace-naked.patch
-Patch1015:	0016-arm-LLVMLinux-Remove-unreachable-from-naked-function.patch
-Patch1016:	0017-MIPS-LLVMLinux-Fix-a-cast-to-type-not-present-in-uni.patch
-Patch1017:	0018-MIPS-LLVMLinux-Fix-an-inline-asm-input-output-type-m.patch
-Patch1018:	0019-MIPS-LLVMLinux-Silence-variable-self-assignment-warn.patch
-Patch1019:	0020-MIPS-LLVMLinux-Silence-unicode-warnings-when-preproc.patch
-Patch1020:	0021-Don-t-use-attributes-error-and-warning-with-clang.patch
-Patch1021:	0022-Fix-undefined-references-to-acpi_idle_driver-on-aarc.patch
-Patch1022:	0023-HACK-firmware-LLVMLinux-fix-EFI-libstub-with-clang.patch
-Patch1023:	0024-aarch64-crypto-LLVMLinux-Fix-inline-assembly-for-cla.patch
-Patch1024:	0025-aarch64-LLVMLinux-Make-spin_lock_prefetch-asm-code-c.patch
-Patch1025:	0026-LLVMLinux-Don-t-use-attribute-externally_visible-whe.patch
-Patch1026:	0027-x86-crypto-LLVMLinux-Fix-building-x86_64-AES-extensi.patch
-Patch1027:	0028-x86-LLVMLinux-Qualify-mul-as-mulq-to-make-clang-happ.patch
-Patch1028:	0029-kbuild-LLVMLinux-Add-Werror-to-cc-option-in-order-to.patch
-Patch1029:	0030-x86-kbuild-LLVMLinux-Check-for-compiler-support-of-f.patch
-#Patch1030:	0031-x86-cmpxchg-break.patch
-Patch1031:	0001-Fix-for-compilation-with-clang.patch
-%endif
-
-# Bootsplash system
-# (tpg) disable it for now 2018-11-07
-%if %{with bootsplash}
-# https://lkml.org/lkml/2017/10/25/346
-# https://patchwork.kernel.org/patch/10172665/, rebased
-Patch100:	RFC-v3-01-13-bootsplash-Initial-implementation-showing-black-screen.patch
-# https://patchwork.kernel.org/patch/10172669/
-Patch101:	RFC-v3-02-13-bootsplash-Add-file-reading-and-picture-rendering.patch
-# https://patchwork.kernel.org/patch/10172715/
-Patch102:	RFC-v3-03-13-bootsplash-Flush-framebuffer-after-drawing.patch
-# https://patchwork.kernel.org/patch/10172699/
-Patch103:	RFC-v3-04-13-bootsplash-Add-corner-positioning.patch
-# https://patchwork.kernel.org/patch/10172667/
-Patch104:	RFC-v3-05-13-bootsplash-Add-animation-support.patch
-# https://patchwork.kernel.org/patch/10172605/, rebased
-Patch105:	RFC-v3-06-13-vt-Redraw-bootsplash-fully-on-console_unblank.patch
-# https://patchwork.kernel.org/patch/10172599/
-Patch106:	RFC-v3-07-13-vt-Add-keyboard-hook-to-disable-bootsplash.patch
-# https://patchwork.kernel.org/patch/10172603/
-Patch107:	RFC-v3-08-13-sysrq-Disable-bootsplash-on-SAK.patch
-# https://patchwork.kernel.org/patch/10172601/
-Patch108:	RFC-v3-09-13-fbcon-Disable-bootsplash-on-oops.patch
-# https://patchwork.kernel.org/patch/10172663/
-Patch109:	RFC-v3-10-13-Documentation-Add-bootsplash-main-documentation.patch
-# https://patchwork.kernel.org/patch/10172685/
-Patch110:	RFC-v3-11-13-bootsplash-sysfs-entries-to-load-and-unload-files.patch
-# https://patchwork.kernel.org/patch/10172597/
-Patch111:	RFC-v3-12-13-tools-bootsplash-Add-a-basic-splash-file-creation-tool.patch
-# https://patchwork.kernel.org/patch/10172661/
-# Contains git binary patch -- needs to be applied with git apply instead of autopatch -p1
-Source112:	RFC-v3-13-13-tools-bootsplash-Add-script-and-data-to-create-sample-file.patch
-%endif
+# Make uClibc-ng happy
+Patch8:		socket.h-include-bitsperlong.h.patch
+# Make Nouveau work on SynQuacer (and probably all other non-x86 boards)
+Patch9:		kernel-5.8-nouveau-write-combining-only-on-x86.patch
+Patch10:	kernel-5.7-fewer-conditions-for-ARM64_PTR_AUTH.patch
+
+# FIXME git bisect shows upstream commit
+# 7a8b64d17e35810dc3176fe61208b45c15d25402 breaks
+# booting SynQuacer from USB flash drives.
+# 9d55bebd9816903b821a403a69a94190442ac043 builds on
+# 7a8b64d17e35810dc3176fe61208b45c15d25402.
+Source100:      7a8b64d17e35810dc3176fe61208b45c15d25402.patch
+Source101:      9d55bebd9816903b821a403a69a94190442ac043.patch
 
 # Patches to VirtualBox and other external modules are
 # pulled in as Source: rather than Patch: because it's arch specific
@@ -308,20 +257,16 @@ Source112:	RFC-v3-13-13-tools-bootsplash-Add-script-and-data-to-create-sample-fi
 # (tpg) The Ultra Kernel Same Page Deduplication
 # (tpg) http://kerneldedup.org/en/projects/uksm/download/
 # (tpg) sources can be found here https://github.com/dolohow/uksm
+# (crazy) each new patch has to be checked for GPL violation on ksm.h
+#  it cannot be re-licensed to GPL3 by random patches.
 %if %{with uksm}
 # brokes armx builds
-Patch120:	https://raw.githubusercontent.com/dolohow/uksm/master/v5.x/uksm-5.4.patch
-# Sometimes other people are ahead of upstream porting to new releases...
-# No UKSM for 5.5 yet... :/
-#Patch120:	https://github.com/sirlucjan/kernel-patches/raw/master/5.1/uksm-pf/0001-uksm-5.1-initial-submission.patch
-#Patch121:	https://github.com/sirlucjan/kernel-patches/raw/master/5.1/uksm-pf-fix/0001-uksm-5.1-apply-52d1e606ee733.patch
+Patch120:	https://raw.githubusercontent.com/dolohow/uksm/master/v5.x/uksm-5.7.patch
 %endif
 
 %if %{with build_modzstd}
-# https://patchwork.kernel.org/patch/10003007/
-Patch126:	v2-1-2-lib-Add-support-for-ZSTD-compressed-kernel.patch
-# https://patchwork.kernel.org/patch/10003011/
-Patch127:	v2-2-2-x86-Add-support-for-ZSTD-compressed-kernel.patch
+# v4 -> https://lkml.org/lkml/2020/4/1/29
+Patch126: https://gitweb.frugalware.org/frugalware-current/raw/master/source/base/kernel/support-kernel-and-ramfs-comp-and-decomp-with-zstd.patch
 %endif
 
 ### Additional hardware support
@@ -342,20 +287,23 @@ Patch146:	saa716x-4.15.patch
 Patch147:	saa716x-linux-4.19.patch
 Patch148:	saa716x-5.4.patch
 
+# Additional WiFi drivers taken from the Endless kernel
+# git clone https://github.com/endlessm/linux.git
+# cd linux
+# tar cf extra-wifi-drivers-`date +%Y%m%d`.tar drivers/net/wireless/rtl8*
+# zstd -19 extra-wifi-drivers*.tar
+Source200:	extra-wifi-drivers-20200301.tar.zst
+Patch201:	extra-wifi-drivers-compile.patch
+Patch202:	extra-wifi-drivers-port-to-5.6.patch
+
 # Lima driver for ARM Mali graphics chips
 # Generated from https://gitlab.freedesktop.org/lima/linux.git
 # using git diff v5.1..lima/lima-5.1
 # Currently no patch necessary
 
-# NOT YET
-#Patch250:	4.14-C11.patch
-
-# VirtualBox shared folders support
-# https://patchwork.kernel.org/patch/10906949/
-# For newer versions, check
-# https://patchwork.kernel.org/project/linux-fsdevel/list/?submitter=582
-Patch300:	v19-fs-Add-VirtualBox-guest-shared-folder-vboxsf-support.diff
-Source300:	virtualbox-kernel-5.3.patch
+# VirtualBox patches -- added as Source: rather than Patch:
+# because they need to be applied after stuff from the
+# virtualbox-kernel-module-sources package is copied around
 Source301:	vbox-6.1-fix-build-on-znver1-hosts.patch
 
 # Better support for newer x86 processors
@@ -405,11 +353,6 @@ Patch800:	Unknow-SSD-HFM128GDHTNG-8310B-QUIRK_NO_APST.patch
 # Restore ACPI loglevels to sane values
 Patch801:	https://gitweb.frugalware.org/wip_kernel/raw/86234abea5e625043153f6b8295642fd9f42bff0/source/base/kernel/acpi-use-kern_warning_even_when_error.patch
 Patch802:	https://gitweb.frugalware.org/wip_kernel/raw/23f5e50042768b823e18613151cc81b4c0cf6e22/source/base/kernel/fix-acpi_dbg_level.patch
-# (tpg) enable MuQSS CPU scheduler
-# FIXME re-enable when ported to 5.3
-Patch803:	http://ck.kolivas.org/patches/muqss/5.0/5.4/0001-MultiQueue-Skiplist-Scheduler-v0.196.patch
-# (bero) And make it compatible with modular binder
-Patch804:	MuQSS-export-can_nice-for-binder.patch
 # (crazy) need to know what function() breaks on nvme failures
 Patch809:	nvme-pci-more-info.patch
 # ( crazy ) this one is adding be_silent mod parameter to acer-wmi
@@ -419,6 +362,8 @@ Patch809:	nvme-pci-more-info.patch
 # until is implemented / fixed.
 #Patch810:  acer-wmi-silence-unknow-functions-messages.patch
 Patch810:	linux-5.4.5-fix-build.patch
+Patch812:	linux-5.5-corsair-strafe-quirks.patch
+Patch814:	http://crazy.dev.frugalware.org/smpboot-no-stack-protector-for-gcc10.patch
 
 # Defines for the things that are needed for all the kernels
 #
@@ -465,9 +410,9 @@ BuildRequires:	flex
 BuildRequires:	bison
 BuildRequires:	binutils
 BuildRequires:	hostname
-BuildRequires:	gcc >= 7.2.1_2017.11-3
-BuildRequires:	gcc-plugin-devel >= 7.2.1_2017.11-3
-BuildRequires:	gcc-c++ >= 7.2.1_2017.11-3
+BuildRequires:	gcc
+BuildRequires:	gcc-c++
+BuildRequires:  pkgconfig(libcap)
 BuildRequires:	pkgconfig(libssl)
 BuildRequires:	diffutils
 # For git apply
@@ -485,6 +430,10 @@ BuildRequires:	numa-devel
 BuildRequires:	pkgconfig(libpci)
 %endif
 
+%if %{with build_turbostat}
+BuildRequires:  pkgconfig(libpcap)
+%endif
+
 # for docs
 %if %{with build_doc}
 BuildRequires:	xmlto
@@ -499,17 +448,15 @@ BuildRequires:	asciidoc
 BuildRequires:	pkgconfig(audit)
 BuildRequires:	binutils-devel
 BuildRequires:	bison
-# BuildRequires:	docbook-style-xsl
 BuildRequires:	flex
-# BuildRequires:	gettext
-# BuildRequires:	gtk2-devel
 BuildRequires:	pkgconfig(libunwind)
 BuildRequires:	pkgconfig(libnewt)
 BuildRequires:	perl-devel
-# BuildRequires:	perl(ExtUtils::Embed)
 BuildRequires:	pkgconfig(gtk+-2.0)
 BuildRequires:	pkgconfig(python)
 BuildRequires:	pkgconfig(zlib)
+# (tpg) needed for bfd
+BuildRequires:	binutils-devel
 %endif
 
 %ifarch %{arm}
@@ -524,8 +471,8 @@ Suggests:	microcode-intel
 # get compiler error messages on failures)
 %if %mdvver >= 3000000
 %ifarch %{x86_64}
-BuildRequires:	virtualbox-kernel-module-sources
-BuildRequires:	virtualbox-guest-kernel-module-sources
+BuildRequires:	virtualbox-kernel-module-sources >= 6.1.10
+BuildRequires:	virtualbox-guest-kernel-module-sources >= 6.1.10
 %endif
 %endif
 
@@ -552,7 +499,7 @@ Requires(pre):	%requires3 %requires4			\
 Requires:	%requires5				\
 Obsoletes:	%kobsoletes1 %kobsoletes2 %kobsoletes3	\
 Conflicts:	%kconflicts1 %kconflicts2 %kconflicts3	\
-Conflicts:	%kconflicts4 %kconflicts5		\
+Conflicts:	%kconflicts4 %kconflicts5 \
 Conflicts:	%{kname}-%{1}-latest <= %{kversion}-%{rpmrel}	\
 Obsoletes:	%{kname}-%{1}-latest <= %{kversion}-%{rpmrel}	\
 Provides:	installonlypkg(kernel)			\
@@ -636,7 +583,6 @@ needs debugging info from the kernel, this package may help. \
 							\
 %post -n %{kname}-%{1} -f kernel_files.%{1}-post 	\
 %posttrans -n %{kname}-%{1} -f kernel_files.%{1}-posttrans \
-%preun -n %{kname}-%{1} -f kernel_files.%{1}-preun 	\
 %postun -n %{kname}-%{1} -f kernel_files.%{1}-postun 	\
 							\
 %if %{with build_devel}					\
@@ -678,7 +624,7 @@ voluntary preempt, CFS cpu scheduler and BFQ i/o scheduler, ONDEMAND governor.
 %define summary_server Linux Kernel for server use with i686 & 64GB RAM
 %define info_server This kernel is compiled for server use, single or \
 multiple i686 processor(s)/core(s) and up to 64GB RAM using PAE, using \
-no preempt, HZ_100, CFS cpu scheduler and BFQ i/o scheduler, PERFORMANCE governor.
+no preempt, HZ_300, CFS cpu scheduler and BFQ i/o scheduler, PERFORMANCE governor.
 %else
 %define summary_server Linux Kernel for server use with %{_arch}
 %define info_server This kernel is compiled for server use, single or \
@@ -780,15 +726,6 @@ Conflicts:	%{_lib}cpufreq-devel
 This package contains the development files for cpupower.
 %endif
 
-%package -n bootsplash-packer
-Summary:	Tool for packing bootsplash images
-Group:		System/Kernel and hardware
-Version:	%{kversion}
-Release:	%{rpmrel}
-
-%description -n bootsplash-packer
-Tool for packing bootsplash images.
-
 %if %{with build_x86_energy_perf_policy}
 %package -n x86_energy_perf_policy
 Version:	%{kversion}
@@ -811,6 +748,34 @@ Group:		System/Kernel and hardware
 Tool to report processor frequency and idle statistics.
 %endif
 
+%define bpf_major 0
+%define libbpf %mklibname bpf %{bpf_major}
+%define libbpfdevel %mklibname bpf -d
+
+%package -n bpftool
+Summary:	Inspection and simple manipulation of eBPF programs and maps
+Group:		System/Kernel and hardware
+
+%description -n bpftool
+This package contains the bpftool, which allows inspection and simple
+manipulation of eBPF programs and maps.
+
+%package -n %{libbpf}
+Summary:	The bpf library from kernel source
+Group:		System/Libraries
+
+%description -n %{libbpf}
+This package contains the kernel source bpf library.
+
+%package -n %{libbpfdevel}
+Summary:	Developement files for the bpf library from kernel source
+Group:		Development/Kernel
+Requires:	%{libbpf} = %{EVRD}
+
+%description -n %{libbpfdevel}
+This package includes libraries and header files needed for development
+of applications which use bpf library from kernel sour
+
 %package headers
 Version:	%{kversion}
 Release:	%{rpmrel}
@@ -871,20 +836,19 @@ done
 # End packages - here begins build stage
 #
 %prep
-%setup -q -n linux-%{tar_ver} -a 140
+%setup -q -n linux-%{tar_ver} -a 140 -a 200
 cp %{S:6} %{S:7} %{S:8} %{S:9} %{S:10} %{S:11} %{S:12} %{S:13} kernel/configs/
 %if 0%{sublevel}
 [ -e .git ] || git init
 xzcat %{SOURCE90} |git apply - || git apply %{SOURCE90}
 rm -rf .git
 %endif
-%if %mdvver > 3000000
 %autopatch -p1
-%else
-%autopatch -p1
-%endif
-%if %{with bootsplash}
-git apply %{SOURCE112}
+
+%ifarch %{aarch64}
+# FIXME SynQuacer workaround
+patch -p1 -R <%{S:101}
+patch -p1 -R <%{S:100}
 %endif
 
 # merge SAA716x DVB driver from extra tarball
@@ -892,6 +856,14 @@ sed -i -e '/saa7164/isource "drivers/media/pci/saa716x/Kconfig"' drivers/media/p
 sed -i -e '/saa7164/iobj-$(CONFIG_SAA716X_CORE) += saa716x/' drivers/media/pci/Makefile
 find drivers/media/tuners drivers/media/dvb-frontends -name "*.c" -o -name "*.h" |xargs sed -i -e 's,"dvb_frontend.h",<media/dvb_frontend.h>,g'
 
+# Merge RTL8723DE and RTL8821CE drivers
+cd drivers/net/wireless
+sed -i -e '/quantenna\/Kconfig/asource "drivers/net/wireless/rtl8821ce/Kconfig' Kconfig
+sed -i -e '/quantenna\/Kconfig/asource "drivers/net/wireless/rtl8723de/Kconfig' Kconfig
+sed -i -e '/QUANTENNA/aobj-$(CONFIG_RTL8821CE) += rtl8821ce/' Makefile
+sed -i -e '/QUANTENNA/aobj-$(CONFIG_RTL8723DE) += rtl8723de/' Makefile
+cd -
+
 %if %{with build_debug}
 %define debug --debug
 %else
@@ -910,9 +882,9 @@ LC_ALL=C sed -i -e "s/^SUBLEVEL.*/SUBLEVEL = %{sublevel}/" Makefile
 # There is an in-kernel version of vboxvideo -- unfortunately
 # it doesn't seem to work properly with vbox just yet
 # Let's replace it with the one that comes with VB for now
-rm -rf drivers/staging/vboxvideo
-cp -a $(ls --sort=time -1d /usr/src/vboxadditions-*|head -n1)/vboxvideo drivers/staging
-cat >drivers/staging/vboxvideo/Kconfig <<'EOF'
+rm -rf drivers/gpu/drm/vboxvideo
+cp -a $(ls --sort=time -1d /usr/src/vboxadditions-*|head -n1)/vboxvideo drivers/gpu/drm/
+cat >drivers/gpu/drm/vboxvideo/Kconfig <<'EOF'
 config DRM_VBOXVIDEO
 	tristate "Virtual Box Graphics Card"
 	depends on DRM && X86 && PCI
@@ -929,14 +901,14 @@ config DRM_VBOXVIDEO
 	  driver as a module and add support for these devices via drm/kms
 	  interfaces.
 EOF
-sed -i -e 's,\$(KBUILD_EXTMOD),drivers/gpu/drm/vboxvideo,g' drivers/staging/vboxvideo/Makefile*
-sed -i -e "s,^KERN_DIR.*,KERN_DIR := $(pwd)," drivers/staging/vboxvideo/Makefile*
+sed -i -e 's,\$(KBUILD_EXTMOD),drivers/gpu/drm/vboxvideo,g' drivers/gpu/drm/vboxvideo/Makefile*
+sed -i -e "s,^KERN_DIR.*,KERN_DIR := $(pwd)," drivers/gpu/drm/vboxvideo/Makefile*
 %endif
 
 # 800x600 is too small to be useful -- even calamares doesn't
 # fit into that anymore (this fix is needed for both the in-kernel
 # version and the vbox version of the driver)
-sed -i -e 's|800, 600|1024, 768|g' drivers/staging/vboxvideo/vbox_mode.c
+sed -i -e 's|800, 600|1024, 768|g' drivers/gpu/drm/vboxvideo/vbox_mode.c
 # VirtualBox shared folders now come in through patch 300
 
 # === VirtualBox host modules ===
@@ -960,7 +932,6 @@ cp -a $(ls --sort=time -1d /usr/src/virtualbox-*|head -n1)/vboxpci drivers/pci/
 sed -i -e 's,\$(KBUILD_EXTMOD),drivers/pci/vboxpci,g' drivers/pci/vboxpci/Makefile*
 sed -i -e "s,^KERN_DIR.*,KERN_DIR := $(pwd)," drivers/pci/vboxpci/Makefile*
 echo 'obj-m += vboxpci/' >>drivers/pci/Makefile
-#patch -p1 -z .300a~ -b <%{S:300}
 patch -p1 -z .301a~ -b <%{S:301}
 %endif
 %endif
@@ -1001,12 +972,6 @@ CreateConfig() {
 	type="$2"
 	rm -f .config
 
-%if %{with clang}
-	CLANG_EXTRAS=clang-workarounds
-%else
-	CLANG_EXTRAS=""
-%endif
-
 %if %{with build_modxz}
 sed -i -e "s/^# CONFIG_KERNEL_XZ is not set/CONFIG_KERNEL_XZ=y/g" kernel/configs/common.config
 %endif
@@ -1028,7 +993,7 @@ sed -i -e "s/^# CONFIG_RD_ZSTD is not set/CONFIG_RD_ZSTD=y/g" kernel/configs/com
 		;;
 	esac
 
-	for i in common common-${type} ${arch}-common ${arch}-${type} $CLANG_EXTRAS; do
+	for i in common common-${type} ${arch}-common ${arch}-${type}; do
 		[ -e kernel/configs/$i.config ] && CONFIGS="$CONFIGS $i.config"
 	done
 	if [ "$arch" = "znver1" ]; then
@@ -1064,9 +1029,9 @@ BuildKernel() {
 # (tpg) build with gcc, as kernel is not yet ready for LLVM/clang
 %ifarch %{x86_64}
 %if %{with clang}
-    %kmake all CC=clang CXX=clang++ CFLAGS="$CFLAGS -flto"
+    %kmake all CC=clang CXX=clang++ CFLAGS="$CFLAGS"
 %else
-    %kmake all CC=gcc CXX=g++ CFLAGS="$CFLAGS -flto"
+    %kmake all CC=gcc CXX=g++ CFLAGS="$CFLAGS"
 %endif
 %else
 %if %{with clang}
@@ -1170,9 +1135,6 @@ SaveDevel() {
 # Needed for external dvb tree (#41418)
     cp -fR drivers/media/dvb-frontends/lgdt330x.h $TempDevelRoot/drivers/media/dvb-frontends/
 
-# add acpica header files, needed for fglrx build
-    cp -fR drivers/acpi/acpica/*.h $TempDevelRoot/drivers/acpi/acpica/
-
 # orc unwinder needs theese
     cp -fR tools/build/Build{,.include} $TempDevelRoot/tools/build
     cp -fR tools/build/fixdep.c $TempDevelRoot/tools/build
@@ -1181,6 +1143,10 @@ SaveDevel() {
     cp -fR tools/objtool/* $TempDevelRoot/tools/objtool
     cp -fR tools/scripts/utilities.mak $TempDevelRoot/tools/scripts
 
+# Make clean fails on the include statements in the Makefiles - and the drivers aren't relevant for -devel
+    rm -rf $TempDevelRoot/drivers/net/wireless/rtl8*
+    sed -i -e '/rtl8.*/d' $TempDevelRoot/drivers/net/wireless/{Makefile,Kconfig}
+
     for i in alpha arc avr32 blackfin c6x cris csky frv h8300 hexagon ia64 m32r m68k m68knommu metag microblaze \
 		 mips mn10300 nds32 nios2 openrisc parisc powerpc s390 score sh sparc tile unicore32 xtensa; do
 	rm -rf $TempDevelRoot/arch/$i
@@ -1337,54 +1303,19 @@ cat kernel_exclude_debug_files.$kernel_flavour >> $kernel_files
 ### Create kernel Post script
 cat > $kernel_files-post <<EOF
 
-# create initrd/grub.cfg for installed kernel first.
+%if %{with dracut_all_initrd}
+[ -x /sbin/dracut ] && /sbin/dracut -f --regenerate-all
+%endif
 
 /sbin/depmod -a %{kversion}-$kernel_flavour-%{buildrpmrel}
-/usr/bin/dracut -f --kver %{kversion}-$kernel_flavour-%{buildrpmrel}
+[ -x /sbin/dracut ] && /sbin/dracut -f --kver %{kversion}-$kernel_flavour-%{buildrpmrel}
 
-# try rebuild all other initrd's , however that may take a while with lots
-# kernels installed
-cd /boot > /dev/null
-
-for i in $(ls vmlinuz-[0-9]*| sed 's/.*vmlinuz-//g')
-do
-	if [[ vmlinuz-$i =~ vmlinuz-%{kversion}-$kernel_flavour-%{buildrpmrel} ]]; then
-		# we just create this
-		continue
-	fi
-	if [[ -e "initrd-$i.img" ]]; then
-		## if exist ignore
-		continue
-	fi
-	/sbin/depmod -a "$i"
-	/usr/bin/dracut -f --kver "$i"
-done
 
 ## cleanup some werid symlinks we never used anyway
 rm -rf vmlinuz-{server,desktop} initrd0.img initrd-{server,desktop}
 
 # run update-grub2
-/usr/sbin/update-grub2
-
-# (crazy) only half the story , need grub patches , OM scripts ( including ARM ) removed suport for systemd-boot
-# and so on .. we hit a limit here with lots kernels installed.
-# also half of that is not used bc missing grub part support. Also we produce ofc broken symlinks and ducplicate
-# 'wath-should-be-machine-id' too. I cannot see why we need that anyway.
-
-#/usr/bin/kernel-install add %{kversion}-$kernel_flavour-%{buildrpmrel} /boot/vmlinuz-%{kversion}-$kernel_flavour-%{buildrpmrel}
-#cd /boot > /dev/null
-#if [ -L vmlinuz-$kernel_flavour ]; then
-#    rm -f vmlinuz-$kernel_flavour
-#fi
-#ln -sf vmlinuz-%{kversion}-$kernel_flavour-%{buildrpmrel} vmlinuz-$kernel_flavour
-#if [ -L initrd-$kernel_flavour.img ]; then
-#    rm -f initrd-$kernel_flavour.img
-#fi
-#ln -sf initrd-%{kversion}-$kernel_flavour-%{buildrpmrel}.img initrd-$kernel_flavour.img
-#if [ -e initrd-%{kversion}-$kernel_flavour-%{buildrpmrel}.img ]; then
-#    ln -sf vmlinuz-%{kversion}-$kernel_flavour-%{buildrpmrel} vmlinuz
-#    ln -sf initrd-%{kversion}-$kernel_flavour-%{buildrpmrel}.img initrd.img
-#fi
+[ -x /usr/sbin/update-grub2 ] && /usr/sbin/update-grub2
 
 cd - > /dev/null
 %if %{with build_devel}
@@ -1399,11 +1330,11 @@ EOF
 
 ### Create kernel Posttrans script
 cat > $kernel_files-posttrans <<EOF
-if [ -x /usr/sbin/dkms_autoinstaller -a -d /usr/src/linux-%{kversion}-$kernel_flavour-%{buildrpmrel} ]; then
+if [ -x /usr/sbin/dkms_autoinstaller ] && [ -d /usr/src/linux-%{kversion}-$kernel_flavour-%{buildrpmrel} ]; then
     /usr/sbin/dkms_autoinstaller start %{kversion}-$kernel_flavour-%{buildrpmrel}
 fi
 
-if [ -x %{_sbindir}/dkms -a -e %{_unitdir}/dkms.service -a -d /usr/src/linux-%{kversion}-$kernel_flavour-%{buildrpmrel} ]; then
+if [ -x %{_sbindir}/dkms ] && [ -e %{_unitdir}/dkms.service ] && [ -d /usr/src/linux-%{kversion}-$kernel_flavour-%{buildrpmrel} ]; then
     /bin/systemctl --quiet restart dkms.service
     /bin/systemctl --quiet try-restart fedora-loadmodules.service
     %{_sbindir}/dkms autoinstall --verbose --kernelver %{kversion}-$kernel_flavour-%{buildrpmrel}
@@ -1411,10 +1342,10 @@ fi
 
 EOF
 
-### Create kernel Preun script on the fly
-cat > $kernel_files-preun <<EOF
+### Create kernel Postun script on the fly
+cat > $kernel_files-postun <<EOF
 
-rm -rf /lib/modules/%{kversion}-$kernel_flavour-%{buildrpmrel}/modules.{alias{,.bin},builtin.bin,dep{,.bin},devname,softdep,symbols{,.bin}}
+rm -rf /lib/modules/%{kversion}-$kernel_flavour-%{buildrpmrel}/modules.{alias{,.bin},builtin.bin,dep{,.bin},devname,softdep,symbols{,.bin}} ||:
 cd /boot > /dev/null
 
 if [ -e vmlinuz-%{kversion}-$kernel_flavour-%{buildrpmrel} ]; then
@@ -1422,26 +1353,20 @@ if [ -e vmlinuz-%{kversion}-$kernel_flavour-%{buildrpmrel} ]; then
 fi
 
 if [ -e initrd-%{kversion}-$kernel_flavour-%{buildrpmrel}.img ]; then
-        rm -rf initrd-%{kversion}-$kernel_flavour-%{buildrpmrel}.img
+	rm -rf initrd-%{kversion}-$kernel_flavour-%{buildrpmrel}.img
 fi
 
 
-#/usr/bin/kernel-install remove %{kversion}-$kernel_flavour-%{buildrpmrel}
-#cd /boot > /dev/null
-## (crazy) we dont use ( nor have support in grub to look ) for initrd-fooname or vmlinuz-fooname
-## so that never worked anyway.
-#if [ -L vmlinuz-$kernel_flavour ]; then
-#    if [ "$(readlink vmlinuz-$kernel_flavour)" = "vmlinuz-%{kversion}-$kernel_flavour-%{buildrpmrel}" ]; then
-#	rm -f vmlinuz-$kernel_flavour
-#    fi
-#fi
-#if [ -L initrd-$kernel_flavour.img ]; then
-#    if [ "$(readlink initrd-$kernel_flavour.img)" = "initrd-%{kversion}-$kernel_flavour-%{buildrpmrel}.img" ]; then
-#	rm -f initrd-$kernel_flavour.img
-#    fi
-#fi
-
 cd - > /dev/null
+
+rm -rf /lib/modules/%{kversion}-$kernel_flavour-%{buildrpmrel} >/dev/null
+if [ -d /var/lib/dkms ]; then
+    rm -f /var/lib/dkms/*/kernel-%{kversion}-$devel_flavour-%{buildrpmrel}-%{_target_cpu} >/dev/null
+    rm -rf /var/lib/dkms/*/*/%{kversion}-$devel_flavour-%{buildrpmrel} >/dev/null
+    rm -f /var/lib/dkms-binary/*/kernel-%{kversion}-$devel_flavour-%{buildrpmrel}-%{_target_cpu} >/dev/null
+    rm -rf /var/lib/dkms-binary/*/*/%{kversion}-$devel_flavour-%{buildrpmrel} >/dev/null
+fi
+
 %if %{with build_devel}
 if [ -L /lib/modules/%{kversion}-$kernel_flavour-%{buildrpmrel}/build ]; then
     rm -f /lib/modules/%{kversion}-$kernel_flavour-%{buildrpmrel}/build
@@ -1452,18 +1377,6 @@ fi
 %endif
 exit 0
 EOF
-
-### Create kernel Postun script on the fly
-cat > $kernel_files-postun <<EOF
-rm -f /boot/initrd-%{kversion}-$kernel_flavour-%{buildrpmrel}.img
-rm -rf /lib/modules/%{kversion}-$kernel_flavour-%{buildrpmrel} >/dev/null
-if [ -d /var/lib/dkms ]; then
-    rm -f /var/lib/dkms/*/kernel-%{kversion}-$devel_flavour-%{buildrpmrel}-%{_target_cpu} >/dev/null
-    rm -rf /var/lib/dkms/*/*/%{kversion}-$devel_flavour-%{buildrpmrel} >/dev/null
-    rm -f /var/lib/dkms-binary/*/kernel-%{kversion}-$devel_flavour-%{buildrpmrel}-%{_target_cpu} >/dev/null
-    rm -rf /var/lib/dkms-binary/*/*/%{kversion}-$devel_flavour-%{buildrpmrel} >/dev/null
-fi
-EOF
 }
 
 CreateKernel() {
@@ -1567,16 +1480,12 @@ sed -ri "s|^(EXTRAVERSION =).*|\1 -%{rpmrel}|" Makefile
 %if %{with build_cpupower}
 # make sure version-gen.sh is executable.
 chmod +x tools/power/cpupower/utils/version-gen.sh
-%kmake -C tools/power/cpupower CPUFREQ_BENCH=false LDFLAGS="%{optflags}"
-%endif
-
-%if %{with bootsplash}
-%kmake -C tools/bootsplash LDFLAGS="%{optflags}"
+%kmake -C tools/power/cpupower CPUFREQ_BENCH=false
 %endif
 
 %ifarch %{ix86} %{x86_64}
 %if %{with build_x86_energy_perf_policy}
-%kmake -C tools/power/x86/x86_energy_perf_policy CC=clang LDFLAGS="%{optflags} -Wl,--build-id=none"
+%kmake -C tools/power/x86/x86_energy_perf_policy CC=clang LDFLAGS="-Wl,--build-id=none"
 %endif
 
 %if %{with build_turbostat}
@@ -1584,6 +1493,11 @@ chmod +x tools/power/cpupower/utils/version-gen.sh
 %endif
 %endif
 
+%kmake -C tools/lib/bpf CC=clang libbpf.a libbpf.pc libbpf.so.0.0.8
+cd tools/bpf/bpftool
+%kmake CC=clang bpftool
+cd -
+
 ############################################################
 ###  Linker end3 > Check point to build for omv or rosa  ###
 ############################################################
@@ -1649,17 +1563,17 @@ sed -ri "s|^(EXTRAVERSION =).*|\1 -%{rpmrel}|" Makefile
 %if %{with build_perf}
 
 # perf tool binary and supporting scripts/binaries
-make -C tools/perf -s CC=%{__cc} V=1 DESTDIR=%{buildroot} WERROR=0 HAVE_CPLUS_DEMANGLE=1 prefix=%{_prefix} install
+make -C tools/perf -s CC=%{__cc} DESTDIR=%{buildroot} WERROR=0 HAVE_CPLUS_DEMANGLE=1 prefix=%{_prefix} install
 
 # perf man pages (note: implicit rpm magic compresses them later)
-make -C tools/perf  -s CC=%{__cc} V=1 DESTDIR=%{buildroot} WERROR=0 HAVE_CPLUS_DEMANGLE=1 prefix=%{_prefix} install-man
+make -C tools/perf  -s CC=%{__cc} DESTDIR=%{buildroot} WERROR=0 HAVE_CPLUS_DEMANGLE=1 prefix=%{_prefix} install-man
 %endif
 
 ############################################################
 ### Linker start4 > Check point to build for omv or rosa ###
 ############################################################
 %if %{with build_cpupower}
-%{make_build} -C tools/power/cpupower DESTDIR=%{buildroot} libdir=%{_libdir} mandir=%{_mandir} CPUFREQ_BENCH=false CC=%{__cc} LDFLAGS="%{optflags}" install
+%make_install -C tools/power/cpupower DESTDIR=%{buildroot} libdir=%{_libdir} mandir=%{_mandir} CPUFREQ_BENCH=false CC=%{__cc} install
 
 rm -f %{buildroot}%{_libdir}/*.{a,la}
 %find_lang cpupower
@@ -1669,22 +1583,21 @@ install -m644 %{SOURCE50} %{buildroot}%{_unitdir}/cpupower.service
 install -m644 %{SOURCE51} %{buildroot}%{_sysconfdir}/sysconfig/cpupower
 %endif
 
-%if %{with bootsplash}
-mkdir -p %{buildroot}%{_bindir}
-install -m755 tools/bootsplash/bootsplash-packer %{buildroot}%{_bindir}/
-%endif
-
 %ifarch %{ix86} %{x86_64}
 %if %{with build_x86_energy_perf_policy}
 mkdir -p %{buildroot}%{_bindir} %{buildroot}%{_mandir}/man8
-%kmake -C tools/power/x86/x86_energy_perf_policy install DESTDIR="%{buildroot}"
+%make_install -C tools/power/x86/x86_energy_perf_policy install DESTDIR="%{buildroot}"
 %endif
 %if %{with build_turbostat}
 mkdir -p %{buildroot}%{_bindir} %{buildroot}%{_mandir}/man8
-%kmake -C tools/power/x86/turbostat install DESTDIR="%{buildroot}"
+%make_install -C tools/power/x86/turbostat install DESTDIR="%{buildroot}"
 %endif
 %endif
 
+# install bpftool and libbpf
+%make_install -C tools/lib/bpf install install_headers DESTDIR=%{buildroot} prefix=%{_prefix} libdir=%{_libdir}
+%make_install -C tools/bpf/bpftool install DESTDIR=%{buildroot} prefix=%{_prefix} bash_compdir=%{_sysconfdir}/bash_completion.d/ mandir=%{_mandir}
+
 # Create directories infastructure
 %if %{with build_source}
 install -d %{target_source}
@@ -1842,11 +1755,6 @@ cd -
 %{_includedir}/cpufreq.h
 %endif
 
-%if %{with bootsplash}
-%files -n bootsplash-packer
-%{_bindir}/bootsplash-packer
-%endif
-
 %ifarch %{ix86} %{x86_64}
 %if %{with build_x86_energy_perf_policy}
 %files -n x86_energy_perf_policy
@@ -1860,3 +1768,17 @@ cd -
 %{_mandir}/man8/turbostat.8*
 %endif
 %endif
+
+%files -n bpftool
+%{_sbindir}/bpftool
+%{_sysconfdir}/bash_completion.d/bpftool
+
+%files -n %{libbpf}
+%{_libdir}/libbpf.so.%{bpf_major}*
+
+%files -n %{libbpfdevel}
+%{_libdir}/libbpf.a
+%{_libdir}/libbpf.so
+%{_libdir}/pkgconfig/*.pc
+%dir %{_includedir}/bpf
+%{_includedir}/bpf/*.h
diff --git a/linux-4.7-intel-dvi-duallink.patch b/linux-4.7-intel-dvi-duallink.patch
index 99a0d13..a9951f8 100644
--- a/linux-4.7-intel-dvi-duallink.patch
+++ b/linux-4.7-intel-dvi-duallink.patch
@@ -1,19 +1,19 @@
-diff -up linux-5.1/drivers/gpu/drm/i915/display/intel_hdmi.c.3~ linux-5.1/drivers/gpu/drm/i915/display/intel_hdmi.c
---- linux-5.1/drivers/gpu/drm/i915/display/intel_hdmi.c.3~	2019-05-07 17:17:24.996151319 +0200
-+++ linux-5.1/drivers/gpu/drm/i915/display/intel_hdmi.c	2019-05-07 17:19:03.873570987 +0200
-@@ -1584,13 +1584,13 @@ intel_hdmi_mode_valid(struct drm_connect
+diff -up linux-5.7-rc1/drivers/gpu/drm/i915/display/intel_hdmi.c.4~ linux-5.7-rc1/drivers/gpu/drm/i915/display/intel_hdmi.c
+--- linux-5.7-rc1/drivers/gpu/drm/i915/display/intel_hdmi.c.4~	2020-04-15 00:04:04.207753898 +0200
++++ linux-5.7-rc1/drivers/gpu/drm/i915/display/intel_hdmi.c	2020-04-15 00:08:26.075745335 +0200
+@@ -2200,13 +2200,13 @@ intel_hdmi_mode_valid(struct drm_connect
  		clock /= 2;
  
  	/* check if we can do 8bpc */
--	status = hdmi_port_clock_valid(hdmi, clock, true, force_dvi);
-+	status = hdmi_port_clock_valid(hdmi, clock, false, force_dvi);
+-	status = hdmi_port_clock_valid(hdmi, clock, true, has_hdmi_sink);
++	status = hdmi_port_clock_valid(hdmi, clock, false, has_hdmi_sink);
  
- 	if (hdmi->has_hdmi_sink && !force_dvi) {
+ 	if (has_hdmi_sink) {
  		/* if we can't do 8bpc we may still be able to do 12bpc */
  		if (status != MODE_OK && !HAS_GMCH(dev_priv))
  			status = hdmi_port_clock_valid(hdmi, clock * 3 / 2,
--						       true, force_dvi);
-+						       false, force_dvi);
+-						       true, has_hdmi_sink);
++						       false, has_hdmi_sink);
  
  		/* if we can't do 8,12bpc we may still be able to do 10bpc */
  		if (status != MODE_OK && INTEL_GEN(dev_priv) >= 11)
diff --git a/linux-4.8.1-buildfix.patch b/linux-4.8.1-buildfix.patch
deleted file mode 100644
index 4c1632b..0000000
--- a/linux-4.8.1-buildfix.patch
+++ /dev/null
@@ -1,33 +0,0 @@
-diff -up linux-5.3-rc1/arch/arm64/kernel/vdso/Makefile.4~ linux-5.3-rc1/arch/arm64/kernel/vdso/Makefile
---- linux-5.3-rc1/arch/arm64/kernel/vdso/Makefile.4~	2019-07-24 00:26:39.079420412 +0200
-+++ linux-5.3-rc1/arch/arm64/kernel/vdso/Makefile	2019-07-24 00:32:03.196644465 +0200
-@@ -17,7 +17,7 @@ obj-vdso := vgettimeofday.o note.o sigre
- targets := $(obj-vdso) vdso.so vdso.so.dbg
- obj-vdso := $(addprefix $(obj)/, $(obj-vdso))
- 
--ldflags-y := -shared -nostdlib -soname=linux-vdso.so.1 --hash-style=sysv \
-+ldflags-y := -shared -nostdlib -soname=linux-vdso.so.1 --hash-style=sysv -fuse-ld=bfd \
- 		--build-id -n -T
- 
- ccflags-y := -fno-common -fno-builtin -fno-stack-protector -ffixed-x18
-diff -up linux-5.3-rc1/arch/x86/entry/vdso/Makefile.4~ linux-5.3-rc1/arch/x86/entry/vdso/Makefile
---- linux-5.3-rc1/arch/x86/entry/vdso/Makefile.4~	2019-07-21 23:05:38.000000000 +0200
-+++ linux-5.3-rc1/arch/x86/entry/vdso/Makefile	2019-07-24 00:32:22.289564153 +0200
-@@ -53,7 +53,7 @@ targets += $(vdso_img_sodbg) $(vdso_img-
- CPPFLAGS_vdso.lds += -P -C
- 
- VDSO_LDFLAGS_vdso.lds = -m elf_x86_64 -soname linux-vdso.so.1 --no-undefined \
--			-z max-page-size=4096
-+			-z max-page-size=4096 -fuse-ld=bfd
- 
- $(obj)/vdso64.so.dbg: $(obj)/vdso.lds $(vobjs) FORCE
- 	$(call if_changed,vdso_and_check)
-@@ -104,7 +104,7 @@ CFLAGS_REMOVE_vvar.o = -pg
- 
- CPPFLAGS_vdsox32.lds = $(CPPFLAGS_vdso.lds)
- VDSO_LDFLAGS_vdsox32.lds = -m elf32_x86_64 -soname linux-vdso.so.1 \
--			   -z max-page-size=4096
-+			   -z max-page-size=4096 -fuse-ld=bfd
- 
- # x32-rebranded versions
- vobjx32s-y := $(vobjs-y:.o=-x32.o)
diff --git a/linux-5.5-corsair-strafe-quirks.patch b/linux-5.5-corsair-strafe-quirks.patch
new file mode 100644
index 0000000..56d9ffe
--- /dev/null
+++ b/linux-5.5-corsair-strafe-quirks.patch
@@ -0,0 +1,22 @@
+diff -up linux-5.5/drivers/hid/hid-ids.h.omv~ linux-5.5/drivers/hid/hid-ids.h
+--- linux-5.5/drivers/hid/hid-ids.h.omv~	2020-02-13 15:44:42.401817937 +0100
++++ linux-5.5/drivers/hid/hid-ids.h	2020-02-13 15:45:29.512814852 +0100
+@@ -310,6 +310,7 @@
+ #define USB_DEVICE_ID_CORSAIR_K70RGB_RAPIDFIRE  0x1b38
+ #define USB_DEVICE_ID_CORSAIR_K65RGB_RAPIDFIRE  0x1b39
+ #define USB_DEVICE_ID_CORSAIR_SCIMITAR_PRO_RGB  0x1b3e
++#define USB_DEVICE_ID_CORSAIR_STRAFE_2  0x1b44
+ 
+ #define USB_VENDOR_ID_CREATIVELABS	0x041e
+ #define USB_DEVICE_ID_CREATIVE_SB_OMNI_SURROUND_51	0x322c
+diff -up linux-5.5/drivers/hid/hid-quirks.c.omv~ linux-5.5/drivers/hid/hid-quirks.c
+--- linux-5.5/drivers/hid/hid-quirks.c.omv~	2020-02-13 15:46:19.977811548 +0100
++++ linux-5.5/drivers/hid/hid-quirks.c	2020-02-13 15:46:22.903811356 +0100
+@@ -64,6 +64,7 @@ static const struct hid_device_id hid_qu
+ 	{ HID_USB_DEVICE(USB_VENDOR_ID_CORSAIR, USB_DEVICE_ID_CORSAIR_GLAIVE_RGB), HID_QUIRK_NO_INIT_REPORTS | HID_QUIRK_ALWAYS_POLL },
+ 	{ HID_USB_DEVICE(USB_VENDOR_ID_CORSAIR, USB_DEVICE_ID_CORSAIR_SCIMITAR_PRO_RGB), HID_QUIRK_NO_INIT_REPORTS | HID_QUIRK_ALWAYS_POLL },
+ 	{ HID_USB_DEVICE(USB_VENDOR_ID_CORSAIR, USB_DEVICE_ID_CORSAIR_STRAFE), HID_QUIRK_NO_INIT_REPORTS | HID_QUIRK_ALWAYS_POLL },
++	{ HID_USB_DEVICE(USB_VENDOR_ID_CORSAIR, USB_DEVICE_ID_CORSAIR_STRAFE_2), HID_QUIRK_NO_INIT_REPORTS | HID_QUIRK_ALWAYS_POLL },
+ 	{ HID_USB_DEVICE(USB_VENDOR_ID_CREATIVELABS, USB_DEVICE_ID_CREATIVE_SB_OMNI_SURROUND_51), HID_QUIRK_NOGET },
+ 	{ HID_USB_DEVICE(USB_VENDOR_ID_DELL, USB_DEVICE_ID_DELL_PIXART_USB_OPTICAL_MOUSE), HID_QUIRK_ALWAYS_POLL },
+ 	{ HID_USB_DEVICE(USB_VENDOR_ID_DMI, USB_DEVICE_ID_DMI_ENC), HID_QUIRK_NOGET },
diff --git a/linux-5.5.tar.sign b/linux-5.5.tar.sign
deleted file mode 100644
index 360859c..0000000
--- a/linux-5.5.tar.sign
+++ /dev/null
@@ -1,19 +0,0 @@
------BEGIN PGP SIGNATURE-----
-Comment: This signature is for the .tar version of the archive
-Comment: git archive --format tar --prefix=linux-5.5/ v5.5
-Comment: git version 2.25.0
-
-iQIzBAABCAAdFiEEZH8oZUiU471FcZm+ONu9yGCSaT4FAl4uhxEACgkQONu9yGCS
-aT7Mig//e4oxjnoKWzrGpC3EtiaOG/ueY0NJWvXWU66xMMzqRnNT8BOL0HdKkZWB
-xS60lR7vKXKTK2w1yz92Tmeo0bbF77J5+uthCQ0F5pvxHvIXGMLhsS5uRwSsjO49
-T0rRea0n+oyjmZQpECyVDLtgjjtaYLSBPSEEylix+m7yetRONcmD49YoizzU9Wtc
-YdpK0bWES/hwcDQlb9owi6giv5qeCvs69CQ2/dzL9ByeSzFuL5TTYw/9CoyCzvMi
-XLCabg5IVdxwGWePLnKamEQRrLbSXanjBacNZXOb9CxRmsESy1NFPIy+yXwSIKIS
-/rMe4zwTj6UK5Sdn5J1onUtk064+TFRnG6QEv7503sml+QMjUcWeTXpSRruoXHov
-H/vnLQBo+my4qatejzdLQLnLKnZB1SSbNA8LeVwxOFKsMxfuIZ5SdztUKUljAaeV
-ppohOLsVoC8XeECL74ZT3KzxOTXy1KWg6Qq5uUIJi40NKTvXNQoi+8q1URcLAk6+
-3s6w/m4nDcSA5x1wkOS7P22qX2oHOwbBOq14xoaR9y/UdfNiNdRCTpMVZcplFvFl
-hNoVhjArabK9M7ufib29mEAJZhqNjp8Q0kXlHVP9scHXIJTRAWCAlqSonmYrNew9
-jocFQoSaD0DdDTlRSA0RFn/YVRfXhvarwKeSTFtv1sVax9yfLd4=
-=kpiR
------END PGP SIGNATURE-----
diff --git a/linux-5.6-fix-disassembler-4args-detection.patch b/linux-5.6-fix-disassembler-4args-detection.patch
new file mode 100644
index 0000000..36aa382
--- /dev/null
+++ b/linux-5.6-fix-disassembler-4args-detection.patch
@@ -0,0 +1,13 @@
+diff -up linux-5.6/tools/bpf/bpftool/jit_disasm.c.omv~ linux-5.6/tools/bpf/bpftool/jit_disasm.c
+diff -up linux-5.6/tools/build/feature/Makefile.omv~ linux-5.6/tools/build/feature/Makefile
+--- linux-5.6/tools/build/feature/Makefile.omv~	2020-04-02 06:07:08.488570621 +0200
++++ linux-5.6/tools/build/feature/Makefile	2020-04-02 06:07:19.983631490 +0200
+@@ -224,7 +224,7 @@ $(OUTPUT)test-libbfd.bin:
+ 	$(BUILD) -DPACKAGE='"perf"' -lbfd -ldl
+ 
+ $(OUTPUT)test-disassembler-four-args.bin:
+-	$(BUILD) -DPACKAGE='"perf"' -lbfd -lopcodes
++	$(BUILD) -DPACKAGE='"perf"' -lbfd -ldl -lopcodes
+ 
+ $(OUTPUT)test-reallocarray.bin:
+ 	$(BUILD)
diff --git a/smpboot-no-stack-protector-for-gcc10.patch b/smpboot-no-stack-protector-for-gcc10.patch
new file mode 100644
index 0000000..d7285e5
--- /dev/null
+++ b/smpboot-no-stack-protector-for-gcc10.patch
@@ -0,0 +1,13 @@
+diff -Naur linux-5.5.9/arch/x86/kernel/Makefile linux-5.5.9-p/arch/x86/kernel/Makefile
+--- linux-5.5.9/arch/x86/kernel/Makefile	2020-03-12 07:18:49.000000000 +0100
++++ linux-5.5.9-p/arch/x86/kernel/Makefile	2020-03-16 15:06:33.354637851 +0100
+@@ -11,6 +11,9 @@
+ 
+ CPPFLAGS_vmlinux.lds += -U$(UTS_MACHINE)
+ 
++dropstackprotector := $(call cc-option, -fno-stack-protector)
++CFLAGS_smpboot.o := $(dropstackprotector)
++
+ ifdef CONFIG_FUNCTION_TRACER
+ # Do not profile debug and lowlevel utilities
+ CFLAGS_REMOVE_tsc.o = -pg
diff --git a/socket.h-include-bitsperlong.h.patch b/socket.h-include-bitsperlong.h.patch
new file mode 100644
index 0000000..d021976
--- /dev/null
+++ b/socket.h-include-bitsperlong.h.patch
@@ -0,0 +1,11 @@
+diff -up linux-5.7/include/uapi/asm-generic/socket.h.omv~ linux-5.7/include/uapi/asm-generic/socket.h
+--- linux-5.7/include/uapi/asm-generic/socket.h.omv~	2020-06-21 14:40:19.478327294 +0200
++++ linux-5.7/include/uapi/asm-generic/socket.h	2020-06-21 14:41:14.638889566 +0200
+@@ -3,6 +3,7 @@
+ #define __ASM_GENERIC_SOCKET_H
+ 
+ #include <linux/posix_types.h>
++#include <asm/bitsperlong.h>
+ #include <asm/sockios.h>
+ 
+ /* For setsockopt(2) */
diff --git a/v2-1-2-lib-Add-support-for-ZSTD-compressed-kernel.patch b/support-kernel-and-ramfs-comp-and-decomp-with-zstd.patch
similarity index 59%
rename from v2-1-2-lib-Add-support-for-ZSTD-compressed-kernel.patch
rename to support-kernel-and-ramfs-comp-and-decomp-with-zstd.patch
index 8acb356..28141f1 100644
--- a/v2-1-2-lib-Add-support-for-ZSTD-compressed-kernel.patch
+++ b/support-kernel-and-ramfs-comp-and-decomp-with-zstd.patch
@@ -1,23 +1,125 @@
-diff -up linux-5.1/include/linux/decompress/unzstd.h.19~ linux-5.1/include/linux/decompress/unzstd.h
---- linux-5.1/include/linux/decompress/unzstd.h.19~	2019-05-07 17:19:25.014249610 +0200
-+++ linux-5.1/include/linux/decompress/unzstd.h	2019-05-07 17:19:25.014249610 +0200
-@@ -0,0 +1,26 @@
-+/*
-+ * Copyright (C) 2017 Facebook
-+ *
-+ * This program is free software; you can redistribute it and/or
-+ * modify it under the terms of the GNU General Public
-+ * License v2 as published by the Free Software Foundation.
-+ *
-+ * This program is distributed in the hope that it will be useful,
-+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
-+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
-+ * General Public License for more details.
-+ *
-+ * You should have received a copy of the GNU General Public License
-+ * along with this program.  If not, see <https://www.gnu.org/licenses/>.
-+ */
+diff -Naur linux-5.6.1/arch/x86/boot/compressed/Makefile linux-5.6.1-p/arch/x86/boot/compressed/Makefile
+--- linux-5.6.1/arch/x86/boot/compressed/Makefile	2020-04-01 10:58:19.000000000 +0200
++++ linux-5.6.1-p/arch/x86/boot/compressed/Makefile	2020-04-01 13:29:07.958649103 +0200
+@@ -24,7 +24,7 @@
+ KCOV_INSTRUMENT		:= n
+ 
+ targets := vmlinux vmlinux.bin vmlinux.bin.gz vmlinux.bin.bz2 vmlinux.bin.lzma \
+-	vmlinux.bin.xz vmlinux.bin.lzo vmlinux.bin.lz4
++	vmlinux.bin.xz vmlinux.bin.lzo vmlinux.bin.lz4 vmlinux.bin.zst
+ 
+ KBUILD_CFLAGS := -m$(BITS) -O2
+ KBUILD_CFLAGS += -fno-strict-aliasing $(call cc-option, -fPIE, -fPIC)
+@@ -145,6 +145,8 @@
+ 	$(call if_changed,lzo)
+ $(obj)/vmlinux.bin.lz4: $(vmlinux.bin.all-y) FORCE
+ 	$(call if_changed,lz4)
++$(obj)/vmlinux.bin.zst: $(vmlinux.bin.all-y) FORCE
++	$(call if_changed,zstd)
+ 
+ suffix-$(CONFIG_KERNEL_GZIP)	:= gz
+ suffix-$(CONFIG_KERNEL_BZIP2)	:= bz2
+@@ -152,6 +154,7 @@
+ suffix-$(CONFIG_KERNEL_XZ)	:= xz
+ suffix-$(CONFIG_KERNEL_LZO) 	:= lzo
+ suffix-$(CONFIG_KERNEL_LZ4) 	:= lz4
++suffix-$(CONFIG_KERNEL_ZSTD)	:= zst
+ 
+ quiet_cmd_mkpiggy = MKPIGGY $@
+       cmd_mkpiggy = $(obj)/mkpiggy $< > $@
+diff -Naur linux-5.6.1/arch/x86/boot/compressed/misc.c linux-5.6.1-p/arch/x86/boot/compressed/misc.c
+--- linux-5.6.1/arch/x86/boot/compressed/misc.c	2020-04-01 10:58:19.000000000 +0200
++++ linux-5.6.1-p/arch/x86/boot/compressed/misc.c	2020-04-01 13:29:07.958649103 +0200
+@@ -77,6 +77,10 @@
+ #ifdef CONFIG_KERNEL_LZ4
+ #include "../../../../lib/decompress_unlz4.c"
+ #endif
 +
++#ifdef CONFIG_KERNEL_ZSTD
++#include "../../../../lib/decompress_unzstd.c"
++#endif
+ /*
+  * NOTE: When adding a new decompressor, please update the analysis in
+  * ../header.S.
+diff -Naur linux-5.6.1/arch/x86/boot/header.S linux-5.6.1-p/arch/x86/boot/header.S
+--- linux-5.6.1/arch/x86/boot/header.S	2020-04-01 10:58:19.000000000 +0200
++++ linux-5.6.1-p/arch/x86/boot/header.S	2020-04-01 13:29:03.958623285 +0200
+@@ -536,8 +536,14 @@
+ # the size-dependent part now grows so fast.
+ #
+ # extra_bytes = (uncompressed_size >> 8) + 65536
++#
++# ZSTD compressed data grows by at most 3 bytes per 128K, and only has a 22
++# byte fixed overhead but has a maximum block size of 128K, so it needs a
++# larger margin.
++#
++# extra_bytes = (uncompressed_size >> 8) + 131072
+ 
+-#define ZO_z_extra_bytes	((ZO_z_output_len >> 8) + 65536)
++#define ZO_z_extra_bytes	((ZO_z_output_len >> 8) + 131072)
+ #if ZO_z_output_len > ZO_z_input_len
+ # define ZO_z_extract_offset	(ZO_z_output_len + ZO_z_extra_bytes - \
+ 				 ZO_z_input_len)
+diff -Naur linux-5.6.1/arch/x86/include/asm/boot.h linux-5.6.1-p/arch/x86/include/asm/boot.h
+--- linux-5.6.1/arch/x86/include/asm/boot.h	2020-04-01 10:58:19.000000000 +0200
++++ linux-5.6.1-p/arch/x86/include/asm/boot.h	2020-04-01 13:29:07.958649103 +0200
+@@ -24,9 +24,11 @@
+ # error "Invalid value for CONFIG_PHYSICAL_ALIGN"
+ #endif
+ 
+-#ifdef CONFIG_KERNEL_BZIP2
++#if defined(CONFIG_KERNEL_BZIP2)
+ # define BOOT_HEAP_SIZE		0x400000
+-#else /* !CONFIG_KERNEL_BZIP2 */
++#elif defined(CONFIG_KERNEL_ZSTD)
++# define BOOT_HEAP_SIZE		 0x30000
++#else
+ # define BOOT_HEAP_SIZE		 0x10000
+ #endif
+ 
+diff -Naur linux-5.6.1/arch/x86/Kconfig linux-5.6.1-p/arch/x86/Kconfig
+--- linux-5.6.1/arch/x86/Kconfig	2020-04-01 10:58:19.000000000 +0200
++++ linux-5.6.1-p/arch/x86/Kconfig	2020-04-01 13:29:07.958649103 +0200
+@@ -183,6 +183,7 @@
+ 	select HAVE_KERNEL_LZMA
+ 	select HAVE_KERNEL_LZO
+ 	select HAVE_KERNEL_XZ
++	select HAVE_KERNEL_ZSTD
+ 	select HAVE_KPROBES
+ 	select HAVE_KPROBES_ON_FTRACE
+ 	select HAVE_FUNCTION_ERROR_INJECTION
+diff -Naur linux-5.6.1/Documentation/x86/boot.rst linux-5.6.1-p/Documentation/x86/boot.rst
+--- linux-5.6.1/Documentation/x86/boot.rst	2020-04-01 10:58:19.000000000 +0200
++++ linux-5.6.1-p/Documentation/x86/boot.rst	2020-04-01 13:29:07.958649103 +0200
+@@ -786,9 +786,9 @@
+   uncompressed data should be determined using the standard magic
+   numbers.  The currently supported compression formats are gzip
+   (magic numbers 1F 8B or 1F 9E), bzip2 (magic number 42 5A), LZMA
+-  (magic number 5D 00), XZ (magic number FD 37), and LZ4 (magic number
+-  02 21).  The uncompressed payload is currently always ELF (magic
+-  number 7F 45 4C 46).
++  (magic number 5D 00), XZ (magic number FD 37), LZ4 (magic number
++  02 21) and ZSTD (magic number 28 B5). The uncompressed payload is
++  currently always ELF (magic number 7F 45 4C 46).
+ 
+ ============	==============
+ Field name:	payload_length
+diff -Naur linux-5.6.1/.gitignore linux-5.6.1-p/.gitignore
+--- linux-5.6.1/.gitignore	2020-04-01 10:58:19.000000000 +0200
++++ linux-5.6.1-p/.gitignore	2020-04-01 13:29:11.158669756 +0200
+@@ -43,6 +43,7 @@
+ *.tab.[ch]
+ *.tar
+ *.xz
++*.zst
+ Module.symvers
+ modules.builtin
+ modules.order
+diff -Naur linux-5.6.1/include/linux/decompress/unzstd.h linux-5.6.1-p/include/linux/decompress/unzstd.h
+--- linux-5.6.1/include/linux/decompress/unzstd.h	1970-01-01 01:00:00.000000000 +0100
++++ linux-5.6.1-p/include/linux/decompress/unzstd.h	2020-04-01 13:28:53.998558979 +0200
+@@ -0,0 +1,11 @@
++/* SPDX-License-Identifier: GPL-2.0 */
 +#ifndef LINUX_DECOMPRESS_UNZSTD_H
 +#define LINUX_DECOMPRESS_UNZSTD_H
 +
@@ -28,10 +130,10 @@ diff -up linux-5.1/include/linux/decompress/unzstd.h.19~ linux-5.1/include/linux
 +	   long *pos,
 +	   void (*error_fn)(char *x));
 +#endif
-diff -up linux-5.1/init/Kconfig.19~ linux-5.1/init/Kconfig
---- linux-5.1/init/Kconfig.19~	2019-05-06 02:42:58.000000000 +0200
-+++ linux-5.1/init/Kconfig	2019-05-07 17:19:25.015249595 +0200
-@@ -153,13 +153,16 @@ config HAVE_KERNEL_LZO
+diff -Naur linux-5.6.1/init/Kconfig linux-5.6.1-p/init/Kconfig
+--- linux-5.6.1/init/Kconfig	2020-04-01 10:58:19.000000000 +0200
++++ linux-5.6.1-p/init/Kconfig	2020-04-01 13:28:58.038585064 +0200
+@@ -173,13 +173,16 @@
  config HAVE_KERNEL_LZ4
  	bool
  
@@ -49,7 +151,7 @@ diff -up linux-5.1/init/Kconfig.19~ linux-5.1/init/Kconfig
  	help
  	  The linux kernel is a kind of self-extracting executable.
  	  Several compression algorithms are available, which differ
-@@ -238,6 +241,15 @@ config KERNEL_LZ4
+@@ -258,6 +261,16 @@
  	  is about 8% bigger than LZO. But the decompression speed is
  	  faster than LZO.
  
@@ -60,14 +162,15 @@ diff -up linux-5.1/init/Kconfig.19~ linux-5.1/init/Kconfig
 +	  ZSTD is a compression algorithm targeting intermediate compression
 +	  with fast decompression speed. It will compress better than GZIP and
 +	  decompress around the same speed as LZO, but slower than LZ4. You
-+	  will need at least 192 KB RAM or more for booting.
++	  will need at least 192 KB RAM or more for booting. The zstd command
++	  line tools is required for compression.
 +
  config KERNEL_UNCOMPRESSED
  	bool "None"
  	depends on HAVE_KERNEL_UNCOMPRESSED
-diff -up linux-5.1/lib/decompress.c.19~ linux-5.1/lib/decompress.c
---- linux-5.1/lib/decompress.c.19~	2019-05-06 02:42:58.000000000 +0200
-+++ linux-5.1/lib/decompress.c	2019-05-07 17:19:25.015249595 +0200
+diff -Naur linux-5.6.1/lib/decompress.c linux-5.6.1-p/lib/decompress.c
+--- linux-5.6.1/lib/decompress.c	2020-04-01 10:58:19.000000000 +0200
++++ linux-5.6.1-p/lib/decompress.c	2020-04-01 13:28:53.998558979 +0200
 @@ -13,6 +13,7 @@
  #include <linux/decompress/inflate.h>
  #include <linux/decompress/unlzo.h>
@@ -86,7 +189,7 @@ diff -up linux-5.1/lib/decompress.c.19~ linux-5.1/lib/decompress.c
  
  struct compress_format {
  	unsigned char magic[2];
-@@ -52,6 +56,7 @@ static const struct compress_format comp
+@@ -52,6 +56,7 @@
  	{ {0xfd, 0x37}, "xz", unxz },
  	{ {0x89, 0x4c}, "lzo", unlzo },
  	{ {0x02, 0x21}, "lz4", unlz4 },
@@ -94,25 +197,11 @@ diff -up linux-5.1/lib/decompress.c.19~ linux-5.1/lib/decompress.c
  	{ {0, 0}, NULL, NULL }
  };
  
-diff -up linux-5.1/lib/decompress_unzstd.c.19~ linux-5.1/lib/decompress_unzstd.c
---- linux-5.1/lib/decompress_unzstd.c.19~	2019-05-07 17:19:25.015249595 +0200
-+++ linux-5.1/lib/decompress_unzstd.c	2019-05-07 17:19:25.015249595 +0200
-@@ -0,0 +1,341 @@
-+/*
-+ * Copyright (C) 2017 Facebook
-+ *
-+ * This program is free software; you can redistribute it and/or
-+ * modify it under the terms of the GNU General Public
-+ * License v2 as published by the Free Software Foundation.
-+ *
-+ * This program is distributed in the hope that it will be useful,
-+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
-+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
-+ * General Public License for more details.
-+ *
-+ * You should have received a copy of the GNU General Public License
-+ * along with this program.  If not, see <https://www.gnu.org/licenses/>.
-+ */
+diff -Naur linux-5.6.1/lib/decompress_unzstd.c linux-5.6.1-p/lib/decompress_unzstd.c
+--- linux-5.6.1/lib/decompress_unzstd.c	1970-01-01 01:00:00.000000000 +0100
++++ linux-5.6.1-p/lib/decompress_unzstd.c	2020-04-01 13:28:53.998558979 +0200
+@@ -0,0 +1,342 @@
++// SPDX-License-Identifier: GPL-2.0
 +
 +/*
 + * Important notes about in-place decompression
@@ -167,32 +256,38 @@ diff -up linux-5.1/lib/decompress_unzstd.c.19~ linux-5.1/lib/decompress_unzstd.c
 + *                 <= 22 + (uncompressed_size >> 15) + 131072
 + */
 +
++/*
++ * Preboot environments #include "path/to/decompress_unzstd.c".
++ * All of the source files we depend on must be #included.
++ * zstd's only source dependeny is xxhash, which has no source
++ * dependencies.
++ *
++ * zstd and xxhash avoid declaring themselves as modules
++ * when ZSTD_PREBOOT and XXH_PREBOOT are defined.
++ */
 +#ifdef STATIC
-+	/* Preboot environments #include "path/to/decompress_unzstd.c".
-+	 * All of the source files we depend on must be #included.
-+	 * zstd's only source dependeny is xxhash, which has no source
-+	 * dependencies.
-+	 *
-+	 * zstd and xxhash both avoid declaring themselves as modules
-+	 * when PREBOOT is defined.
-+	 */
-+#	define PREBOOT
-+#	include "xxhash.c"
-+#	include "zstd/entropy_common.c"
-+#	include "zstd/fse_decompress.c"
-+#	include "zstd/huf_decompress.c"
-+#	include "zstd/zstd_common.c"
-+#	include "zstd/decompress.c"
++# define ZSTD_PREBOOT
++# define XXH_PREBOOT
++# include "xxhash.c"
++# include "zstd/entropy_common.c"
++# include "zstd/fse_decompress.c"
++# include "zstd/huf_decompress.c"
++# include "zstd/zstd_common.c"
++# include "zstd/decompress.c"
 +#endif
 +
 +#include <linux/decompress/mm.h>
 +#include <linux/kernel.h>
 +#include <linux/zstd.h>
 +
-+/* 8 MB maximum window size */
-+#define ZSTD_WINDOWSIZE_MAX	(1 << 23)
-+/* Size of the input and output buffers in multi-call mdoe */
-+#define ZSTD_IOBUF_SIZE		4096
++/* 128MB is the maximum window size supported by zstd. */
++#define ZSTD_WINDOWSIZE_MAX	(1 << ZSTD_WINDOWLOG_MAX)
++/* Size of the input and output buffers in multi-call mode.
++ * Pick a larger size because it isn't used during kernel decompression,
++ * since that is single pass, and we have to allocate a large buffer for
++ * zstd's window anyways. The larger size speeds up initramfs decompression.
++ */
++#define ZSTD_IOBUF_SIZE		(1 << 17)
 +
 +static int INIT handle_zstd_error(size_t ret, void (*error)(char *x))
 +{
@@ -220,7 +315,8 @@ diff -up linux-5.1/lib/decompress_unzstd.c.19~ linux-5.1/lib/decompress_unzstd.c
 +	return -1;
 +}
 +
-+/* Handle the case where we have the entire input and output in one segment.
++/*
++ * Handle the case where we have the entire input and output in one segment.
 + * We can allocate less memory (no circular buffer for the sliding window),
 + * and avoid some memcpy() calls.
 + */
@@ -239,7 +335,8 @@ diff -up linux-5.1/lib/decompress_unzstd.c.19~ linux-5.1/lib/decompress_unzstd.c
 +		err = -1;
 +		goto out;
 +	}
-+	/* Find out how large the frame actually is, there may be junk at
++	/*
++	 * Find out how large the frame actually is, there may be junk at
 +	 * the end of the frame that ZSTD_decompressDCtx() can't handle.
 +	 */
 +	ret = ZSTD_findFrameCompressedSize(in_buf, in_len);
@@ -285,18 +382,20 @@ diff -up linux-5.1/lib/decompress_unzstd.c.19~ linux-5.1/lib/decompress_unzstd.c
 +		out_len = LONG_MAX; /* no limit */
 +
 +	if (fill == NULL && flush == NULL)
-+		/* We can decompress faster and with less memory when we have a
++		/*
++		 * We can decompress faster and with less memory when we have a
 +		 * single chunk.
 +		 */
 +		return decompress_single(in_buf, in_len, out_buf, out_len,
 +					 in_pos, error);
 +
-+	/* If in_buf is not provided, we must be using fill(), so allocate
++	/*
++	 * If in_buf is not provided, we must be using fill(), so allocate
 +	 * a large enough buffer. If it is provided, it must be at least
 +	 * ZSTD_IOBUF_SIZE large.
 +	 */
 +	if (in_buf == NULL) {
-+		in_allocated = malloc(ZSTD_IOBUF_SIZE);
++		in_allocated = large_malloc(ZSTD_IOBUF_SIZE);
 +		if (in_allocated == NULL) {
 +			error("Out of memory while allocating input buffer");
 +			err = -1;
@@ -305,7 +404,7 @@ diff -up linux-5.1/lib/decompress_unzstd.c.19~ linux-5.1/lib/decompress_unzstd.c
 +		in_buf = in_allocated;
 +		in_len = 0;
 +	}
-+	/* Read the first chunk, since we need to decode the frame header */
++	/* Read the first chunk, since we need to decode the frame header. */
 +	if (fill != NULL)
 +		in_len = fill(in_buf, ZSTD_IOBUF_SIZE);
 +	if (in_len < 0) {
@@ -313,13 +412,13 @@ diff -up linux-5.1/lib/decompress_unzstd.c.19~ linux-5.1/lib/decompress_unzstd.c
 +		err = -1;
 +		goto out;
 +	}
-+	/* Set the first non-empty input buffer */
++	/* Set the first non-empty input buffer. */
 +	in.src = in_buf;
 +	in.pos = 0;
 +	in.size = in_len;
 +	/* Allocate the output buffer if we are using flush(). */
 +	if (flush != NULL) {
-+		out_allocated = malloc(ZSTD_IOBUF_SIZE);
++		out_allocated = large_malloc(ZSTD_IOBUF_SIZE);
 +		if (out_allocated == NULL) {
 +			error("Out of memory while allocating output buffer");
 +			err = -1;
@@ -328,12 +427,13 @@ diff -up linux-5.1/lib/decompress_unzstd.c.19~ linux-5.1/lib/decompress_unzstd.c
 +		out_buf = out_allocated;
 +		out_len = ZSTD_IOBUF_SIZE;
 +	}
-+	/* Set the output buffer */
++	/* Set the output buffer. */
 +	out.dst = out_buf;
 +	out.pos = 0;
 +	out.size = out_len;
 +
-+	/* We need to know the window size to allocate the ZSTD_DStream.
++	/*
++	 * We need to know the window size to allocate the ZSTD_DStream.
 +	 * Since we are streaming, we need to allocate a buffer for the sliding
 +	 * window. The window size varies from 1 KB to ZSTD_WINDOWSIZE_MAX
 +	 * (8 MB), so it is important to use the actual value so as not to
@@ -354,7 +454,8 @@ diff -up linux-5.1/lib/decompress_unzstd.c.19~ linux-5.1/lib/decompress_unzstd.c
 +		goto out;
 +	}
 +
-+	/* Allocate the ZSTD_DStream now that we know how much memory is
++	/*
++	 * Allocate the ZSTD_DStream now that we know how much memory is
 +	 * required.
 +	 */
 +	wksp_size = ZSTD_DStreamWorkspaceBound(params.windowSize);
@@ -365,7 +466,9 @@ diff -up linux-5.1/lib/decompress_unzstd.c.19~ linux-5.1/lib/decompress_unzstd.c
 +		err = -1;
 +		goto out;
 +	}
-+	/* Decompression loop:
++
++	/*
++	 * Decompression loop:
 +	 * Read more data if necessary (error if no more data can be read).
 +	 * Call the decompression function, which returns 0 when finished.
 +	 * Flush any data produced if using flush().
@@ -373,7 +476,8 @@ diff -up linux-5.1/lib/decompress_unzstd.c.19~ linux-5.1/lib/decompress_unzstd.c
 +	if (in_pos != NULL)
 +		*in_pos = 0;
 +	do {
-+		/* If we need to reload data, either we have fill() and can
++		/*
++		 * If we need to reload data, either we have fill() and can
 +		 * try to get more data, or we don't and the input is truncated.
 +		 */
 +		if (in.pos == in.size) {
@@ -388,12 +492,12 @@ diff -up linux-5.1/lib/decompress_unzstd.c.19~ linux-5.1/lib/decompress_unzstd.c
 +			in.pos = 0;
 +			in.size = in_len;
 +		}
-+		/* Returns zero when the frame is complete */
++		/* Returns zero when the frame is complete. */
 +		ret = ZSTD_decompressStream(dstream, &out, &in);
 +		err = handle_zstd_error(ret, error);
 +		if (err)
 +			goto out;
-+		/* Flush all of the data produced if using flush() */
++		/* Flush all of the data produced if using flush(). */
 +		if (flush != NULL && out.pos > 0) {
 +			if (out.pos != flush(out.dst, out.pos)) {
 +				error("Failed to flush()");
@@ -410,15 +514,15 @@ diff -up linux-5.1/lib/decompress_unzstd.c.19~ linux-5.1/lib/decompress_unzstd.c
 +	err = 0;
 +out:
 +	if (in_allocated != NULL)
-+		free(in_allocated);
++		large_free(in_allocated);
 +	if (out_allocated != NULL)
-+		free(out_allocated);
++		large_free(out_allocated);
 +	if (wksp != NULL)
 +		large_free(wksp);
 +	return err;
 +}
 +
-+#ifndef PREBOOT
++#ifndef ZSTD_PREBOOT
 +STATIC int INIT unzstd(unsigned char *buf, long len,
 +		       long (*fill)(void*, unsigned long),
 +		       long (*flush)(void*, unsigned long),
@@ -439,10 +543,10 @@ diff -up linux-5.1/lib/decompress_unzstd.c.19~ linux-5.1/lib/decompress_unzstd.c
 +	return __unzstd(buf, len, fill, flush, out_buf, out_len, pos, error);
 +}
 +#endif
-diff -up linux-5.1/lib/Kconfig.19~ linux-5.1/lib/Kconfig
---- linux-5.1/lib/Kconfig.19~	2019-05-06 02:42:58.000000000 +0200
-+++ linux-5.1/lib/Kconfig	2019-05-07 17:19:25.015249595 +0200
-@@ -312,6 +312,10 @@ config DECOMPRESS_LZ4
+diff -Naur linux-5.6.1/lib/Kconfig linux-5.6.1-p/lib/Kconfig
+--- linux-5.6.1/lib/Kconfig	2020-04-01 10:58:19.000000000 +0200
++++ linux-5.6.1-p/lib/Kconfig	2020-04-01 13:28:53.998558979 +0200
+@@ -336,6 +336,10 @@
  	select LZ4_DECOMPRESS
  	tristate
  
@@ -453,10 +557,10 @@ diff -up linux-5.1/lib/Kconfig.19~ linux-5.1/lib/Kconfig
  #
  # Generic allocator support is selected if needed
  #
-diff -up linux-5.1/lib/Makefile.19~ linux-5.1/lib/Makefile
---- linux-5.1/lib/Makefile.19~	2019-05-07 17:19:25.005249746 +0200
-+++ linux-5.1/lib/Makefile	2019-05-07 17:19:25.016249580 +0200
-@@ -156,6 +156,7 @@ lib-$(CONFIG_DECOMPRESS_LZMA) += decompr
+diff -Naur linux-5.6.1/lib/Makefile linux-5.6.1-p/lib/Makefile
+--- linux-5.6.1/lib/Makefile	2020-04-01 10:58:19.000000000 +0200
++++ linux-5.6.1-p/lib/Makefile	2020-04-01 13:28:53.998558979 +0200
+@@ -160,6 +160,7 @@
  lib-$(CONFIG_DECOMPRESS_XZ) += decompress_unxz.o
  lib-$(CONFIG_DECOMPRESS_LZO) += decompress_unlzo.o
  lib-$(CONFIG_DECOMPRESS_LZ4) += decompress_unlz4.o
@@ -464,10 +568,10 @@ diff -up linux-5.1/lib/Makefile.19~ linux-5.1/lib/Makefile
  
  obj-$(CONFIG_TEXTSEARCH) += textsearch.o
  obj-$(CONFIG_TEXTSEARCH_KMP) += ts_kmp.o
-diff -up linux-5.1/lib/xxhash.c.19~ linux-5.1/lib/xxhash.c
---- linux-5.1/lib/xxhash.c.19~	2019-05-06 02:42:58.000000000 +0200
-+++ linux-5.1/lib/xxhash.c	2019-05-07 17:19:25.016249580 +0200
-@@ -80,13 +80,11 @@ void xxh32_copy_state(struct xxh32_state
+diff -Naur linux-5.6.1/lib/xxhash.c linux-5.6.1-p/lib/xxhash.c
+--- linux-5.6.1/lib/xxhash.c	2020-04-01 10:58:19.000000000 +0200
++++ linux-5.6.1-p/lib/xxhash.c	2020-04-01 13:28:51.078540121 +0200
+@@ -80,13 +80,11 @@
  {
  	memcpy(dst, src, sizeof(*dst));
  }
@@ -481,7 +585,7 @@ diff -up linux-5.1/lib/xxhash.c.19~ linux-5.1/lib/xxhash.c
  
  /*-***************************
   * Simple Hash Functions
-@@ -151,7 +149,6 @@ uint32_t xxh32(const void *input, const
+@@ -151,7 +149,6 @@
  
  	return h32;
  }
@@ -489,7 +593,7 @@ diff -up linux-5.1/lib/xxhash.c.19~ linux-5.1/lib/xxhash.c
  
  static uint64_t xxh64_round(uint64_t acc, const uint64_t input)
  {
-@@ -234,7 +231,6 @@ uint64_t xxh64(const void *input, const
+@@ -234,7 +231,6 @@
  
  	return h64;
  }
@@ -497,7 +601,7 @@ diff -up linux-5.1/lib/xxhash.c.19~ linux-5.1/lib/xxhash.c
  
  /*-**************************************************
   * Advanced Hash Functions
-@@ -251,7 +247,6 @@ void xxh32_reset(struct xxh32_state *sta
+@@ -251,7 +247,6 @@
  	state.v4 = seed - PRIME32_1;
  	memcpy(statePtr, &state, sizeof(state));
  }
@@ -505,7 +609,7 @@ diff -up linux-5.1/lib/xxhash.c.19~ linux-5.1/lib/xxhash.c
  
  void xxh64_reset(struct xxh64_state *statePtr, const uint64_t seed)
  {
-@@ -265,7 +260,6 @@ void xxh64_reset(struct xxh64_state *sta
+@@ -265,7 +260,6 @@
  	state.v4 = seed - PRIME64_1;
  	memcpy(statePtr, &state, sizeof(state));
  }
@@ -513,7 +617,7 @@ diff -up linux-5.1/lib/xxhash.c.19~ linux-5.1/lib/xxhash.c
  
  int xxh32_update(struct xxh32_state *state, const void *input, const size_t len)
  {
-@@ -334,7 +328,6 @@ int xxh32_update(struct xxh32_state *sta
+@@ -334,7 +328,6 @@
  
  	return 0;
  }
@@ -521,7 +625,7 @@ diff -up linux-5.1/lib/xxhash.c.19~ linux-5.1/lib/xxhash.c
  
  uint32_t xxh32_digest(const struct xxh32_state *state)
  {
-@@ -372,7 +365,6 @@ uint32_t xxh32_digest(const struct xxh32
+@@ -372,7 +365,6 @@
  
  	return h32;
  }
@@ -529,7 +633,7 @@ diff -up linux-5.1/lib/xxhash.c.19~ linux-5.1/lib/xxhash.c
  
  int xxh64_update(struct xxh64_state *state, const void *input, const size_t len)
  {
-@@ -439,7 +431,6 @@ int xxh64_update(struct xxh64_state *sta
+@@ -439,7 +431,6 @@
  
  	return 0;
  }
@@ -537,12 +641,12 @@ diff -up linux-5.1/lib/xxhash.c.19~ linux-5.1/lib/xxhash.c
  
  uint64_t xxh64_digest(const struct xxh64_state *state)
  {
-@@ -494,7 +485,19 @@ uint64_t xxh64_digest(const struct xxh64
+@@ -494,7 +485,19 @@
  
  	return h64;
  }
 +
-+#ifndef PREBOOT
++#ifndef XXH_PREBOOT
 +EXPORT_SYMBOL(xxh32_copy_state);
 +EXPORT_SYMBOL(xxh64_copy_state);
 +EXPORT_SYMBOL(xxh32);
@@ -557,25 +661,25 @@ diff -up linux-5.1/lib/xxhash.c.19~ linux-5.1/lib/xxhash.c
  MODULE_LICENSE("Dual BSD/GPL");
  MODULE_DESCRIPTION("xxHash");
 +#endif
-diff -up linux-5.1/lib/zstd/decompress.c.19~ linux-5.1/lib/zstd/decompress.c
---- linux-5.1/lib/zstd/decompress.c.19~	2019-05-06 02:42:58.000000000 +0200
-+++ linux-5.1/lib/zstd/decompress.c	2019-05-07 17:19:25.016249580 +0200
-@@ -2487,6 +2487,7 @@ size_t ZSTD_decompressStream(ZSTD_DStrea
+diff -Naur linux-5.6.1/lib/zstd/decompress.c linux-5.6.1-p/lib/zstd/decompress.c
+--- linux-5.6.1/lib/zstd/decompress.c	2020-04-01 10:58:19.000000000 +0200
++++ linux-5.6.1-p/lib/zstd/decompress.c	2020-04-01 13:28:46.908513191 +0200
+@@ -2490,6 +2490,7 @@
  	}
  }
  
-+#ifndef PREBOOT
++#ifndef ZSTD_PREBOOT
  EXPORT_SYMBOL(ZSTD_DCtxWorkspaceBound);
  EXPORT_SYMBOL(ZSTD_initDCtx);
  EXPORT_SYMBOL(ZSTD_decompressDCtx);
-@@ -2526,3 +2527,4 @@ EXPORT_SYMBOL(ZSTD_insertBlock);
+@@ -2529,3 +2530,4 @@
  
  MODULE_LICENSE("Dual BSD/GPL");
  MODULE_DESCRIPTION("Zstd Decompressor");
 +#endif
-diff -up linux-5.1/lib/zstd/fse_decompress.c.19~ linux-5.1/lib/zstd/fse_decompress.c
---- linux-5.1/lib/zstd/fse_decompress.c.19~	2019-05-06 02:42:58.000000000 +0200
-+++ linux-5.1/lib/zstd/fse_decompress.c	2019-05-07 17:19:25.016249580 +0200
+diff -Naur linux-5.6.1/lib/zstd/fse_decompress.c linux-5.6.1-p/lib/zstd/fse_decompress.c
+--- linux-5.6.1/lib/zstd/fse_decompress.c	2020-04-01 10:58:19.000000000 +0200
++++ linux-5.6.1-p/lib/zstd/fse_decompress.c	2020-04-01 13:28:46.908513191 +0200
 @@ -47,6 +47,7 @@
  ****************************************************************/
  #include "bitstream.h"
@@ -599,15 +703,16 @@ diff -up linux-5.1/lib/zstd/fse_decompress.c.19~ linux-5.1/lib/zstd/fse_decompre
  /* **************************************************************
  *  Templates
  ****************************************************************/
-diff -up linux-5.1/lib/zstd/zstd_internal.h.19~ linux-5.1/lib/zstd/zstd_internal.h
---- linux-5.1/lib/zstd/zstd_internal.h.19~	2019-05-06 02:42:58.000000000 +0200
-+++ linux-5.1/lib/zstd/zstd_internal.h	2019-05-07 17:19:25.016249580 +0200
-@@ -127,7 +127,13 @@ static const U32 OF_defaultNormLog = OF_
+diff -Naur linux-5.6.1/lib/zstd/zstd_internal.h linux-5.6.1-p/lib/zstd/zstd_internal.h
+--- linux-5.6.1/lib/zstd/zstd_internal.h	2020-04-01 10:58:19.000000000 +0200
++++ linux-5.6.1-p/lib/zstd/zstd_internal.h	2020-04-01 13:28:46.908513191 +0200
+@@ -127,7 +127,14 @@
  *  Shared functions to include for inlining
  *********************************************/
  ZSTD_STATIC void ZSTD_copy8(void *dst, const void *src) {
 -	memcpy(dst, src, 8);
-+	/* zstd relies heavily on gcc being able to analyze and inline this
++	/*
++	 * zstd relies heavily on gcc being able to analyze and inline this
 +	 * memcpy() call, since it is called in a tight loop. Preboot mode
 +	 * is compiled in freestanding mode, which stops gcc from analyzing
 +	 * memcpy(). Use __builtin_memcpy() to tell gcc to analyze this as a
@@ -617,15 +722,17 @@ diff -up linux-5.1/lib/zstd/zstd_internal.h.19~ linux-5.1/lib/zstd/zstd_internal
  }
  /*! ZSTD_wildcopy() :
  *   custom version of memcpy(), can copy up to 7 bytes too many (8 bytes if length==0) */
-@@ -137,6 +143,7 @@ ZSTD_STATIC void ZSTD_wildcopy(void *dst
+@@ -137,13 +144,16 @@
  	const BYTE* ip = (const BYTE*)src;
  	BYTE* op = (BYTE*)dst;
  	BYTE* const oend = op + length;
-+#if GCC_VERSION >= 70000 && GCC_VERSION < 70200
- 	/* Work around https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81388.
+-	/* Work around https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81388.
++#if defined(GCC_VERSION) && GCC_VERSION >= 70000 && GCC_VERSION < 70200
++	/*
++	 * Work around https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81388.
  	 * Avoid the bad case where the loop only runs once by handling the
  	 * special case separately. This doesn't trigger the bug because it
-@@ -144,6 +151,7 @@ ZSTD_STATIC void ZSTD_wildcopy(void *dst
+ 	 * doesn't involve pointer/integer overflow.
  	 */
  	if (length <= 8)
  		return ZSTD_copy8(dst, src);
@@ -633,10 +740,10 @@ diff -up linux-5.1/lib/zstd/zstd_internal.h.19~ linux-5.1/lib/zstd/zstd_internal
  	do {
  		ZSTD_copy8(op, ip);
  		op += 8;
-diff -up linux-5.1/scripts/Makefile.lib.19~ linux-5.1/scripts/Makefile.lib
---- linux-5.1/scripts/Makefile.lib.19~	2019-05-07 17:19:25.017249565 +0200
-+++ linux-5.1/scripts/Makefile.lib	2019-05-07 17:19:58.549750595 +0200
-@@ -402,6 +402,21 @@ quiet_cmd_xzkern = XZKERN  $@
+diff -Naur linux-5.6.1/scripts/Makefile.lib linux-5.6.1-p/scripts/Makefile.lib
+--- linux-5.6.1/scripts/Makefile.lib	2020-04-01 10:58:19.000000000 +0200
++++ linux-5.6.1-p/scripts/Makefile.lib	2020-04-01 13:28:58.038585064 +0200
+@@ -394,6 +394,21 @@
  quiet_cmd_xzmisc = XZMISC  $@
        cmd_xzmisc = cat $(real-prereqs) | xz --check=crc32 --lzma2=dict=1MiB > $@
  
@@ -651,17 +758,17 @@ diff -up linux-5.1/scripts/Makefile.lib.19~ linux-5.1/scripts/Makefile.lib
 +
 +quiet_cmd_zstd = ZSTD    $@
 +cmd_zstd = (cat $(filter-out FORCE,$^) | \
-+        zstd -19 && \
++	zstd -19 && \
 +        $(call size_append, $(filter-out FORCE,$^))) > $@ || \
-+        (rm -f $@ ; false)
++	(rm -f $@ ; false)
 +
  # ASM offsets
  # ---------------------------------------------------------------------------
  
-diff -up linux-5.1/usr/Kconfig.19~ linux-5.1/usr/Kconfig
---- linux-5.1/usr/Kconfig.19~	2019-05-06 02:42:58.000000000 +0200
-+++ linux-5.1/usr/Kconfig	2019-05-07 17:19:25.017249565 +0200
-@@ -106,6 +106,15 @@ config RD_LZ4
+diff -Naur linux-5.6.1/usr/Kconfig linux-5.6.1-p/usr/Kconfig
+--- linux-5.6.1/usr/Kconfig	2020-04-01 10:58:19.000000000 +0200
++++ linux-5.6.1-p/usr/Kconfig	2020-04-01 13:29:00.978604046 +0200
+@@ -100,6 +100,15 @@
  	  Support loading of a LZ4 encoded initial ramdisk or cpio buffer
  	  If unsure, say N.
  
@@ -676,10 +783,10 @@ diff -up linux-5.1/usr/Kconfig.19~ linux-5.1/usr/Kconfig
 +
  choice
  	prompt "Built-in initramfs compression mode"
- 	depends on INITRAMFS_SOURCE!=""
-@@ -214,6 +223,17 @@ config INITRAMFS_COMPRESSION_LZ4
- 	  If you choose this, keep in mind that most distros don't provide lz4
- 	  by default which could cause a build failure.
+ 	depends on INITRAMFS_SOURCE != ""
+@@ -207,4 +216,15 @@
+ 	  short time at boot, while both the cpio image and the unpacked
+ 	  filesystem image will be present in memory simultaneously
  
 +config INITRAMFS_COMPRESSION_ZSTD
 +	bool "ZSTD"
@@ -693,18 +800,14 @@ diff -up linux-5.1/usr/Kconfig.19~ linux-5.1/usr/Kconfig
 +	  tool to be able to compress the initram.
 +
  endchoice
+diff -Naur linux-5.6.1/usr/Makefile linux-5.6.1-p/usr/Makefile
+--- linux-5.6.1/usr/Makefile	2020-04-01 10:58:19.000000000 +0200
++++ linux-5.6.1-p/usr/Makefile	2020-04-01 13:29:00.978604046 +0200
+@@ -15,6 +15,7 @@
+ compress-$(CONFIG_INITRAMFS_COMPRESSION_XZ)	:= xzmisc
+ compress-$(CONFIG_INITRAMFS_COMPRESSION_LZO)	:= lzo
+ compress-$(CONFIG_INITRAMFS_COMPRESSION_LZ4)	:= lz4
++compress-$(CONFIG_INITRAMFS_COMPRESSION_ZSTD)	:= zstd
+ 
+ obj-$(CONFIG_BLK_DEV_INITRD) := initramfs_data.o
  
- config INITRAMFS_COMPRESSION
-@@ -226,10 +246,12 @@ config INITRAMFS_COMPRESSION
- 	default ".xz"   if INITRAMFS_COMPRESSION_XZ
- 	default ".lzo"  if INITRAMFS_COMPRESSION_LZO
- 	default ".lz4"  if INITRAMFS_COMPRESSION_LZ4
-+	default ".zst"  if INITRAMFS_COMPRESSION_ZSTD
- 	default ".gz"   if RD_GZIP
- 	default ".lz4"  if RD_LZ4
- 	default ".lzo"  if RD_LZO
- 	default ".xz"   if RD_XZ
- 	default ".lzma" if RD_LZMA
- 	default ".bz2"  if RD_BZIP2
-+	default ".zst"  if RD_ZSTD
- 	default ""
diff --git a/uksm-5.4.patch b/uksm-5.6.patch
similarity index 98%
rename from uksm-5.4.patch
rename to uksm-5.6.patch
index 31fc641..0fbc88d 100644
--- a/uksm-5.4.patch
+++ b/uksm-5.6.patch
@@ -66,10 +66,10 @@ index 000000000000..be19a3127001
 +2016-09-10 UKSM 0.1.2.5 Fix a bug in dedup ratio calculation.
 +2017-02-26 UKSM 0.1.2.6 Fix a bug in hugetlbpage handling and a race bug with page migration.
 diff --git a/fs/exec.c b/fs/exec.c
-index 555e93c7dec8..7e1cfd568e6c 100644
+index db17be51b112..0ce4c6303f53 100644
 --- a/fs/exec.c
 +++ b/fs/exec.c
-@@ -63,6 +63,7 @@
+@@ -62,6 +62,7 @@
  #include <linux/oom.h>
  #include <linux/compat.h>
  #include <linux/vmalloc.h>
@@ -77,7 +77,7 @@ index 555e93c7dec8..7e1cfd568e6c 100644
  
  #include <linux/uaccess.h>
  #include <asm/mmu_context.h>
-@@ -1384,6 +1385,7 @@ void setup_new_exec(struct linux_binprm * bprm)
+@@ -1387,6 +1388,7 @@ void setup_new_exec(struct linux_binprm * bprm)
  	/* An exec changes our domain. We are no longer part of the thread
  	   group */
  	current->self_exec_id++;
@@ -102,10 +102,10 @@ index 8c1f1bb1a5ce..62e28cf10bbf 100644
  		    global_node_page_state(NR_UNSTABLE_NFS));
  	show_val_kb(m, "Bounce:         ",
 diff --git a/include/asm-generic/pgtable.h b/include/asm-generic/pgtable.h
-index 818691846c90..b2169f37c696 100644
+index e2e2bef07dd2..59675a348878 100644
 --- a/include/asm-generic/pgtable.h
 +++ b/include/asm-generic/pgtable.h
-@@ -855,12 +855,25 @@ extern void untrack_pfn(struct vm_area_struct *vma, unsigned long pfn,
+@@ -866,12 +866,25 @@ extern void untrack_pfn(struct vm_area_struct *vma, unsigned long pfn,
  extern void untrack_pfn_moved(struct vm_area_struct *vma);
  #endif
  
@@ -132,7 +132,7 @@ index 818691846c90..b2169f37c696 100644
  }
  
  #define my_zero_pfn(addr)	page_to_pfn(ZERO_PAGE(addr))
-@@ -869,7 +882,7 @@ static inline int is_zero_pfn(unsigned long pfn)
+@@ -880,7 +893,7 @@ static inline int is_zero_pfn(unsigned long pfn)
  static inline int is_zero_pfn(unsigned long pfn)
  {
  	extern unsigned long zero_pfn;
@@ -147,7 +147,7 @@ index e48b1e453ff5..095d59310ce0 100644
 +++ b/include/linux/ksm.h
 @@ -1,4 +1,4 @@
 -/* SPDX-License-Identifier: GPL-2.0 */
-+/* SPDX-License-Identifier: GPL-3.0 */
++/* SPDX-License-Identifier: GPL-2.0 */
  #ifndef __LINUX_KSM_H
  #define __LINUX_KSM_H
  /*
@@ -219,10 +219,10 @@ index e48b1e453ff5..095d59310ce0 100644
 +
  #endif /* __LINUX_KSM_H */
 diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h
-index 270aa8fd2800..8e1eaf6a2de3 100644
+index c28911c3afa8..d14689c20e56 100644
 --- a/include/linux/mm_types.h
 +++ b/include/linux/mm_types.h
-@@ -353,6 +353,9 @@ struct vm_area_struct {
+@@ -358,6 +358,9 @@ struct vm_area_struct {
  	struct mempolicy *vm_policy;	/* NUMA policy for the VMA */
  #endif
  	struct vm_userfaultfd_ctx vm_userfaultfd_ctx;
@@ -233,7 +233,7 @@ index 270aa8fd2800..8e1eaf6a2de3 100644
  
  struct core_thread {
 diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
-index bda20282746b..dab82e7cc3df 100644
+index 462f6873905a..43edd793d4e7 100644
 --- a/include/linux/mmzone.h
 +++ b/include/linux/mmzone.h
 @@ -206,6 +206,9 @@ enum zone_stat_item {
@@ -485,10 +485,10 @@ index 000000000000..bb8651f534f2
 +#endif /* !CONFIG_UKSM */
 +#endif /* __LINUX_UKSM_H */
 diff --git a/kernel/fork.c b/kernel/fork.c
-index 13b38794efb5..acb562745f1c 100644
+index d90af13431c7..b9d771b1acfb 100644
 --- a/kernel/fork.c
 +++ b/kernel/fork.c
-@@ -595,7 +595,7 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm,
+@@ -598,7 +598,7 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm,
  		__vma_link_rb(mm, tmp, rb_link, rb_parent);
  		rb_link = &tmp->vm_rb.rb_right;
  		rb_parent = &tmp->vm_rb;
@@ -498,18 +498,18 @@ index 13b38794efb5..acb562745f1c 100644
  		if (!(tmp->vm_flags & VM_WIPEONFORK))
  			retval = copy_page_range(mm, oldmm, mpnt);
 diff --git a/lib/Makefile b/lib/Makefile
-index c5892807e06f..4b0d37da9fa0 100644
+index 611872c06926..1e3d86def26c 100644
 --- a/lib/Makefile
 +++ b/lib/Makefile
-@@ -25,7 +25,7 @@ CFLAGS_string.o := $(call cc-option, -fno-stack-protector)
+@@ -26,7 +26,7 @@ CFLAGS_string.o := $(call cc-option, -fno-stack-protector)
  endif
  
  lib-y := ctype.o string.o vsprintf.o cmdline.o \
 -	 rbtree.o radix-tree.o timerqueue.o xarray.o \
 +	 rbtree.o radix-tree.o sradix-tree.o timerqueue.o xarray.o \
- 	 idr.o extable.o \
- 	 sha1.o chacha.o irq_regs.o argv_split.o \
+ 	 idr.o extable.o sha1.o irq_regs.o argv_split.o \
  	 flex_proportions.o ratelimit.o show_mem.o \
+ 	 is_single_threaded.o plist.o decompress.o kobject_uevent.o \
 diff --git a/lib/sradix-tree.c b/lib/sradix-tree.c
 new file mode 100644
 index 000000000000..ab21e6309b93
@@ -993,10 +993,10 @@ index 000000000000..ab21e6309b93
 +	return 0;
 +}
 diff --git a/mm/Kconfig b/mm/Kconfig
-index a5dae9a7eb51..eda3a78226b0 100644
+index ab80933be65f..04796660748f 100644
 --- a/mm/Kconfig
 +++ b/mm/Kconfig
-@@ -299,6 +299,32 @@ config KSM
+@@ -300,6 +300,32 @@ config KSM
  	  See Documentation/vm/ksm.rst for more information: KSM is inactive
  	  until a program has madvised that an area is MADV_MERGEABLE, and
  	  root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set).
@@ -1028,12 +1028,12 @@ index a5dae9a7eb51..eda3a78226b0 100644
 +endchoice
  
  config DEFAULT_MMAP_MIN_ADDR
-         int "Low address space to protect from user allocation"
+ 	int "Low address space to protect from user allocation"
 diff --git a/mm/Makefile b/mm/Makefile
-index d996846697ef..786cb9fad4b9 100644
+index 272e66039e70..20ec701eda60 100644
 --- a/mm/Makefile
 +++ b/mm/Makefile
-@@ -66,7 +66,8 @@ obj-$(CONFIG_SPARSEMEM)	+= sparse.o
+@@ -67,7 +67,8 @@ obj-$(CONFIG_SPARSEMEM)	+= sparse.o
  obj-$(CONFIG_SPARSEMEM_VMEMMAP) += sparse-vmemmap.o
  obj-$(CONFIG_SLOB) += slob.o
  obj-$(CONFIG_MMU_NOTIFIER) += mmu_notifier.o
@@ -1044,7 +1044,7 @@ index d996846697ef..786cb9fad4b9 100644
  obj-$(CONFIG_SLAB) += slab.o
  obj-$(CONFIG_SLUB) += slub.o
 diff --git a/mm/ksm.c b/mm/ksm.c
-index 7905934cd3ad..3e9d0064eaf2 100644
+index d17c7d57d0d8..d9942db0a21e 100644
 --- a/mm/ksm.c
 +++ b/mm/ksm.c
 @@ -857,17 +857,6 @@ static int unmerge_ksm_pages(struct vm_area_struct *vma,
@@ -1066,10 +1066,10 @@ index 7905934cd3ad..3e9d0064eaf2 100644
  /*
   * Only called through the sysfs control interface:
 diff --git a/mm/memory.c b/mm/memory.c
-index b1ca51a079f2..98a57e508b9b 100644
+index e8bfdf0d9d1d..3507bfef1c5e 100644
 --- a/mm/memory.c
 +++ b/mm/memory.c
-@@ -130,6 +130,25 @@ EXPORT_SYMBOL(zero_pfn);
+@@ -144,6 +144,25 @@ EXPORT_SYMBOL(zero_pfn);
  
  unsigned long highest_memmap_pfn __read_mostly;
  
@@ -1095,15 +1095,15 @@ index b1ca51a079f2..98a57e508b9b 100644
  /*
   * CONFIG_MMU architectures set up ZERO_PAGE in their paging_init()
   */
-@@ -141,6 +160,7 @@ static int __init init_zero_pfn(void)
- core_initcall(init_zero_pfn);
- 
+@@ -159,6 +178,7 @@ void mm_trace_rss_stat(struct mm_struct *mm, int member, long count)
+ 	trace_rss_stat(mm, member, count);
+ }
  
 +
  #if defined(SPLIT_RSS_COUNTING)
  
  void sync_mm_rss(struct mm_struct *mm)
-@@ -772,8 +792,13 @@ copy_one_pte(struct mm_struct *dst_mm, struct mm_struct *src_mm,
+@@ -790,8 +810,13 @@ copy_one_pte(struct mm_struct *dst_mm, struct mm_struct *src_mm,
  		get_page(page);
  		page_dup_rmap(page, false);
  		rss[mm_counter(page)]++;
@@ -1117,7 +1117,7 @@ index b1ca51a079f2..98a57e508b9b 100644
  	}
  
  out_set_pte:
-@@ -1046,8 +1071,10 @@ static unsigned long zap_pte_range(struct mmu_gather *tlb,
+@@ -1064,8 +1089,10 @@ static unsigned long zap_pte_range(struct mmu_gather *tlb,
  			ptent = ptep_get_and_clear_full(mm, addr, pte,
  							tlb->fullmm);
  			tlb_remove_tlb_entry(tlb, pte, addr);
@@ -1129,19 +1129,15 @@ index b1ca51a079f2..98a57e508b9b 100644
  
  			if (!PageAnon(page)) {
  				if (pte_dirty(ptent)) {
-@@ -2169,8 +2196,10 @@ static inline void cow_user_page(struct page *dst, struct page *src, unsigned lo
- 			clear_page(kaddr);
- 		kunmap_atomic(kaddr);
- 		flush_dcache_page(dst);
--	} else
-+	} else {
- 		copy_user_highpage(dst, src, va, vma);
+@@ -2266,6 +2293,7 @@ static inline bool cow_user_page(struct page *dst, struct page *src,
+ 
+ 	if (likely(src)) {
+ 		copy_user_highpage(dst, src, addr, vma);
 +		uksm_cow_page(vma, src);
-+	}
- }
+ 		return true;
+ 	}
  
- static gfp_t __get_fault_gfp_mask(struct vm_area_struct *vma)
-@@ -2322,6 +2351,7 @@ static vm_fault_t wp_page_copy(struct vm_fault *vmf)
+@@ -2512,6 +2540,7 @@ static vm_fault_t wp_page_copy(struct vm_fault *vmf)
  							      vmf->address);
  		if (!new_page)
  			goto oom;
@@ -1149,7 +1145,7 @@ index b1ca51a079f2..98a57e508b9b 100644
  	} else {
  		new_page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma,
  				vmf->address);
-@@ -2351,7 +2381,9 @@ static vm_fault_t wp_page_copy(struct vm_fault *vmf)
+@@ -2553,7 +2582,9 @@ static vm_fault_t wp_page_copy(struct vm_fault *vmf)
  						mm_counter_file(old_page));
  				inc_mm_counter_fast(mm, MM_ANONPAGES);
  			}
@@ -1160,7 +1156,7 @@ index b1ca51a079f2..98a57e508b9b 100644
  		}
  		flush_cache_page(vma, vmf->address, pte_pfn(vmf->orig_pte));
 diff --git a/mm/mmap.c b/mm/mmap.c
-index a7d8c84d19b7..797582240047 100644
+index d681a20eb4ea..acf07b12677c 100644
 --- a/mm/mmap.c
 +++ b/mm/mmap.c
 @@ -46,6 +46,7 @@
@@ -1171,7 +1167,7 @@ index a7d8c84d19b7..797582240047 100644
  #include <linux/sched/mm.h>
  
  #include <linux/uaccess.h>
-@@ -184,6 +185,7 @@ static struct vm_area_struct *remove_vma(struct vm_area_struct *vma)
+@@ -178,6 +179,7 @@ static struct vm_area_struct *remove_vma(struct vm_area_struct *vma)
  	if (vma->vm_file)
  		fput(vma->vm_file);
  	mpol_put(vma_policy(vma));
@@ -1179,7 +1175,7 @@ index a7d8c84d19b7..797582240047 100644
  	vm_area_free(vma);
  	return next;
  }
-@@ -736,9 +738,16 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
+@@ -705,9 +707,16 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
  	long adjust_next = 0;
  	int remove_next = 0;
  
@@ -1196,7 +1192,7 @@ index a7d8c84d19b7..797582240047 100644
  		if (end >= next->vm_end) {
  			/*
  			 * vma expands, overlapping all the next, and
-@@ -871,6 +880,7 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
+@@ -838,6 +847,7 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
  		end_changed = true;
  	}
  	vma->vm_pgoff = pgoff;
@@ -1204,7 +1200,7 @@ index a7d8c84d19b7..797582240047 100644
  	if (adjust_next) {
  		next->vm_start += adjust_next << PAGE_SHIFT;
  		next->vm_pgoff += adjust_next;
-@@ -976,6 +986,7 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
+@@ -943,6 +953,7 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
  		if (remove_next == 2) {
  			remove_next = 1;
  			end = next->vm_end;
@@ -1212,7 +1208,7 @@ index a7d8c84d19b7..797582240047 100644
  			goto again;
  		}
  		else if (next)
-@@ -1002,10 +1013,14 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
+@@ -969,10 +980,14 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
  			 */
  			VM_WARN_ON(mm->highest_vm_end != vm_end_gap(vma));
  		}
@@ -1227,7 +1223,7 @@ index a7d8c84d19b7..797582240047 100644
  	validate_mm(mm);
  
  	return 0;
-@@ -1465,6 +1480,9 @@ unsigned long do_mmap(struct file *file, unsigned long addr,
+@@ -1431,6 +1446,9 @@ unsigned long do_mmap(struct file *file, unsigned long addr,
  	vm_flags |= calc_vm_prot_bits(prot, pkey) | calc_vm_flag_bits(flags) |
  			mm->def_flags | VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC;
  
@@ -1237,7 +1233,7 @@ index a7d8c84d19b7..797582240047 100644
  	if (flags & MAP_LOCKED)
  		if (!can_do_mlock())
  			return -EPERM;
-@@ -1835,6 +1853,7 @@ unsigned long mmap_region(struct file *file, unsigned long addr,
+@@ -1799,6 +1817,7 @@ unsigned long mmap_region(struct file *file, unsigned long addr,
  			allow_write_access(file);
  	}
  	file = vma->vm_file;
@@ -1245,7 +1241,7 @@ index a7d8c84d19b7..797582240047 100644
  out:
  	perf_event_mmap(vma);
  
-@@ -1877,6 +1896,7 @@ unsigned long mmap_region(struct file *file, unsigned long addr,
+@@ -1841,6 +1860,7 @@ unsigned long mmap_region(struct file *file, unsigned long addr,
  	if (vm_flags & VM_DENYWRITE)
  		allow_write_access(file);
  free_vma:
@@ -1253,7 +1249,7 @@ index a7d8c84d19b7..797582240047 100644
  	vm_area_free(vma);
  unacct_error:
  	if (charged)
-@@ -2706,6 +2726,8 @@ int __split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
+@@ -2670,6 +2690,8 @@ int __split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
  	else
  		err = vma_adjust(vma, vma->vm_start, addr, vma->vm_pgoff, new);
  
@@ -1262,15 +1258,15 @@ index a7d8c84d19b7..797582240047 100644
  	/* Success. */
  	if (!err)
  		return 0;
-@@ -3011,6 +3033,7 @@ static int do_brk_flags(unsigned long addr, unsigned long len, unsigned long fla
+@@ -2976,6 +2998,7 @@ static int do_brk_flags(unsigned long addr, unsigned long len, unsigned long fla
  	if ((flags & (~VM_EXEC)) != 0)
  		return -EINVAL;
  	flags |= VM_DATA_DEFAULT_FLAGS | VM_ACCOUNT | mm->def_flags;
 +	uksm_vm_flags_mod(&flags);
  
- 	error = get_unmapped_area(NULL, addr, len, 0, MAP_FIXED);
- 	if (offset_in_page(error))
-@@ -3061,6 +3084,7 @@ static int do_brk_flags(unsigned long addr, unsigned long len, unsigned long fla
+ 	mapped_addr = get_unmapped_area(NULL, addr, len, 0, MAP_FIXED);
+ 	if (IS_ERR_VALUE(mapped_addr))
+@@ -3026,6 +3049,7 @@ static int do_brk_flags(unsigned long addr, unsigned long len, unsigned long fla
  	vma->vm_flags = flags;
  	vma->vm_page_prot = vm_get_page_prot(flags);
  	vma_link(mm, vma, prev, rb_link, rb_parent);
@@ -1278,7 +1274,7 @@ index a7d8c84d19b7..797582240047 100644
  out:
  	perf_event_mmap(vma);
  	mm->total_vm += len >> PAGE_SHIFT;
-@@ -3138,6 +3162,12 @@ void exit_mmap(struct mm_struct *mm)
+@@ -3103,6 +3127,12 @@ void exit_mmap(struct mm_struct *mm)
  		up_write(&mm->mmap_sem);
  	}
  
@@ -1291,7 +1287,7 @@ index a7d8c84d19b7..797582240047 100644
  	if (mm->locked_vm) {
  		vma = mm->mmap;
  		while (vma) {
-@@ -3172,6 +3202,11 @@ void exit_mmap(struct mm_struct *mm)
+@@ -3137,6 +3167,11 @@ void exit_mmap(struct mm_struct *mm)
  		vma = remove_vma(vma);
  	}
  	vm_unacct_memory(nr_accounted);
@@ -1303,7 +1299,7 @@ index a7d8c84d19b7..797582240047 100644
  }
  
  /* Insert vm structure into process list sorted by address
-@@ -3279,6 +3314,7 @@ struct vm_area_struct *copy_vma(struct vm_area_struct **vmap,
+@@ -3244,6 +3279,7 @@ struct vm_area_struct *copy_vma(struct vm_area_struct **vmap,
  			new_vma->vm_ops->open(new_vma);
  		vma_link(mm, new_vma, prev, rb_link, rb_parent);
  		*need_rmap_locks = false;
@@ -1311,7 +1307,7 @@ index a7d8c84d19b7..797582240047 100644
  	}
  	return new_vma;
  
-@@ -3429,6 +3465,7 @@ static struct vm_area_struct *__install_special_mapping(
+@@ -3396,6 +3432,7 @@ static struct vm_area_struct *__install_special_mapping(
  	vm_stat_account(mm, vma->vm_flags, len >> PAGE_SHIFT);
  
  	perf_event_mmap(vma);
@@ -6939,10 +6935,10 @@ index 000000000000..ef068d5dc307
 +#endif
 +
 diff --git a/mm/vmstat.c b/mm/vmstat.c
-index a8222041bd44..7058e8322cbd 100644
+index 78d53378db99..fbb16b46b41b 100644
 --- a/mm/vmstat.c
 +++ b/mm/vmstat.c
-@@ -1168,6 +1168,9 @@ const char * const vmstat_text[] = {
+@@ -1169,6 +1169,9 @@ const char * const vmstat_text[] = {
  	"nr_written",
  	"nr_kernel_misc_reclaimable",
  
diff --git a/uksm-5.7.patch b/uksm-5.7.patch
new file mode 100644
index 0000000..471f357
--- /dev/null
+++ b/uksm-5.7.patch
@@ -0,0 +1,6936 @@
+diff --git a/Documentation/vm/uksm.txt b/Documentation/vm/uksm.txt
+new file mode 100644
+index 000000000000..be19a3127001
+--- /dev/null
++++ b/Documentation/vm/uksm.txt
+@@ -0,0 +1,61 @@
++The Ultra Kernel Samepage Merging feature
++----------------------------------------------
++/*
++ * Ultra KSM. Copyright (C) 2011-2012 Nai Xia
++ *
++ * This is an improvement upon KSM. Some basic data structures and routines
++ * are borrowed from ksm.c .
++ *
++ * Its new features:
++ * 1. Full system scan:
++ *      It automatically scans all user processes' anonymous VMAs. Kernel-user
++ *      interaction to submit a memory area to KSM is no longer needed.
++ *
++ * 2. Rich area detection:
++ *      It automatically detects rich areas containing abundant duplicated
++ *      pages based. Rich areas are given a full scan speed. Poor areas are
++ *      sampled at a reasonable speed with very low CPU consumption.
++ *
++ * 3. Ultra Per-page scan speed improvement:
++ *      A new hash algorithm is proposed. As a result, on a machine with
++ *      Core(TM)2 Quad Q9300 CPU in 32-bit mode and 800MHZ DDR2 main memory, it
++ *      can scan memory areas that does not contain duplicated pages at speed of
++ *      627MB/sec ~ 2445MB/sec and can merge duplicated areas at speed of
++ *      477MB/sec ~ 923MB/sec.
++ *
++ * 4. Thrashing area avoidance:
++ *      Thrashing area(an VMA that has frequent Ksm page break-out) can be
++ *      filtered out. My benchmark shows it's more efficient than KSM's per-page
++ *      hash value based volatile page detection.
++ *
++ *
++ * 5. Misc changes upon KSM:
++ *      * It has a fully x86-opitmized memcmp dedicated for 4-byte-aligned page
++ *        comparison. It's much faster than default C version on x86.
++ *      * rmap_item now has an struct *page member to loosely cache a
++ *        address-->page mapping, which reduces too much time-costly
++ *        follow_page().
++ *      * The VMA creation/exit procedures are hooked to let the Ultra KSM know.
++ *      * try_to_merge_two_pages() now can revert a pte if it fails. No break_
++ *        ksm is needed for this case.
++ *
++ * 6. Full Zero Page consideration(contributed by Figo Zhang)
++ *    Now uksmd consider full zero pages as special pages and merge them to an
++ *    special unswappable uksm zero page.
++ */
++
++ChangeLog:
++
++2012-05-05 The creation of this Doc
++2012-05-08 UKSM 0.1.1.1 libc crash bug fix, api clean up, doc clean up.
++2012-05-28 UKSM 0.1.1.2 bug fix release
++2012-06-26 UKSM 0.1.2-beta1 first beta release for 0.1.2
++2012-07-2  UKSM 0.1.2-beta2
++2012-07-10 UKSM 0.1.2-beta3
++2012-07-26 UKSM 0.1.2 Fine grained speed control, more scan optimization.
++2012-10-13 UKSM 0.1.2.1 Bug fixes.
++2012-12-31 UKSM 0.1.2.2 Minor bug fixes.
++2014-07-02 UKSM 0.1.2.3 Fix a " __this_cpu_read() in preemptible bug".
++2015-04-22 UKSM 0.1.2.4 Fix a race condition that can sometimes trigger anonying warnings.
++2016-09-10 UKSM 0.1.2.5 Fix a bug in dedup ratio calculation.
++2017-02-26 UKSM 0.1.2.6 Fix a bug in hugetlbpage handling and a race bug with page migration.
+diff --git a/fs/exec.c b/fs/exec.c
+index 2c465119affc..50890203108a 100644
+--- a/fs/exec.c
++++ b/fs/exec.c
+@@ -62,6 +62,7 @@
+ #include <linux/oom.h>
+ #include <linux/compat.h>
+ #include <linux/vmalloc.h>
++#include <linux/ksm.h>
+ 
+ #include <linux/uaccess.h>
+ #include <asm/mmu_context.h>
+diff --git a/fs/proc/meminfo.c b/fs/proc/meminfo.c
+index 8c1f1bb1a5ce..62e28cf10bbf 100644
+--- a/fs/proc/meminfo.c
++++ b/fs/proc/meminfo.c
+@@ -105,7 +105,10 @@ static int meminfo_proc_show(struct seq_file *m, void *v)
+ 		   global_zone_page_state(NR_KERNEL_STACK_KB));
+ 	show_val_kb(m, "PageTables:     ",
+ 		    global_zone_page_state(NR_PAGETABLE));
+-
++#ifdef CONFIG_UKSM
++	show_val_kb(m, "KsmZeroPages:     ",
++		    global_zone_page_state(NR_UKSM_ZERO_PAGES));
++#endif
+ 	show_val_kb(m, "NFS_Unstable:   ",
+ 		    global_node_page_state(NR_UNSTABLE_NFS));
+ 	show_val_kb(m, "Bounce:         ",
+diff --git a/include/asm-generic/pgtable.h b/include/asm-generic/pgtable.h
+index 329b8c8ca703..22fb5a592b98 100644
+--- a/include/asm-generic/pgtable.h
++++ b/include/asm-generic/pgtable.h
+@@ -867,12 +867,25 @@ extern void untrack_pfn(struct vm_area_struct *vma, unsigned long pfn,
+ extern void untrack_pfn_moved(struct vm_area_struct *vma);
+ #endif
+ 
++#ifdef CONFIG_UKSM
++static inline int is_uksm_zero_pfn(unsigned long pfn)
++{
++	extern unsigned long uksm_zero_pfn;
++	return pfn == uksm_zero_pfn;
++}
++#else
++static inline int is_uksm_zero_pfn(unsigned long pfn)
++{
++	return 0;
++}
++#endif
++
+ #ifdef __HAVE_COLOR_ZERO_PAGE
+ static inline int is_zero_pfn(unsigned long pfn)
+ {
+ 	extern unsigned long zero_pfn;
+ 	unsigned long offset_from_zero_pfn = pfn - zero_pfn;
+-	return offset_from_zero_pfn <= (zero_page_mask >> PAGE_SHIFT);
++	return offset_from_zero_pfn <= (zero_page_mask >> PAGE_SHIFT) || is_uksm_zero_pfn(pfn);
+ }
+ 
+ #define my_zero_pfn(addr)	page_to_pfn(ZERO_PAGE(addr))
+@@ -881,7 +894,7 @@ static inline int is_zero_pfn(unsigned long pfn)
+ static inline int is_zero_pfn(unsigned long pfn)
+ {
+ 	extern unsigned long zero_pfn;
+-	return pfn == zero_pfn;
++	return (pfn == zero_pfn) || (is_uksm_zero_pfn(pfn));
+ }
+ 
+ static inline unsigned long my_zero_pfn(unsigned long addr)
+diff --git a/include/linux/ksm.h b/include/linux/ksm.h
+index e48b1e453ff5..8cc8077cebf4 100644
+--- a/include/linux/ksm.h
++++ b/include/linux/ksm.h
+@@ -21,20 +21,16 @@ struct mem_cgroup;
+ #ifdef CONFIG_KSM
+ int ksm_madvise(struct vm_area_struct *vma, unsigned long start,
+ 		unsigned long end, int advice, unsigned long *vm_flags);
+-int __ksm_enter(struct mm_struct *mm);
+-void __ksm_exit(struct mm_struct *mm);
+ 
+-static inline int ksm_fork(struct mm_struct *mm, struct mm_struct *oldmm)
++static inline struct stable_node *page_stable_node(struct page *page)
+ {
+-	if (test_bit(MMF_VM_MERGEABLE, &oldmm->flags))
+-		return __ksm_enter(mm);
+-	return 0;
++	return PageKsm(page) ? page_rmapping(page) : NULL;
+ }
+ 
+-static inline void ksm_exit(struct mm_struct *mm)
++static inline void set_page_stable_node(struct page *page,
++					struct stable_node *stable_node)
+ {
+-	if (test_bit(MMF_VM_MERGEABLE, &mm->flags))
+-		__ksm_exit(mm);
++	page->mapping = (void *)((unsigned long)stable_node | PAGE_MAPPING_KSM);
+ }
+ 
+ /*
+@@ -56,6 +52,33 @@ void ksm_migrate_page(struct page *newpage, struct page *oldpage);
+ bool reuse_ksm_page(struct page *page,
+ 			struct vm_area_struct *vma, unsigned long address);
+ 
++#ifdef CONFIG_KSM_LEGACY
++int __ksm_enter(struct mm_struct *mm);
++void __ksm_exit(struct mm_struct *mm);
++static inline int ksm_fork(struct mm_struct *mm, struct mm_struct *oldmm)
++{
++	if (test_bit(MMF_VM_MERGEABLE, &oldmm->flags))
++		return __ksm_enter(mm);
++	return 0;
++}
++
++static inline void ksm_exit(struct mm_struct *mm)
++{
++	if (test_bit(MMF_VM_MERGEABLE, &mm->flags))
++		__ksm_exit(mm);
++}
++
++#elif defined(CONFIG_UKSM)
++static inline int ksm_fork(struct mm_struct *mm, struct mm_struct *oldmm)
++{
++	return 0;
++}
++
++static inline void ksm_exit(struct mm_struct *mm)
++{
++}
++#endif /* !CONFIG_UKSM */
++
+ #else  /* !CONFIG_KSM */
+ 
+ static inline int ksm_fork(struct mm_struct *mm, struct mm_struct *oldmm)
+@@ -96,4 +119,6 @@ static inline bool reuse_ksm_page(struct page *page,
+ #endif /* CONFIG_MMU */
+ #endif /* !CONFIG_KSM */
+ 
++#include <linux/uksm.h>
++
+ #endif /* __LINUX_KSM_H */
+diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h
+index 4aba6c0c2ba8..d4a58fce7f42 100644
+--- a/include/linux/mm_types.h
++++ b/include/linux/mm_types.h
+@@ -363,6 +363,9 @@ struct vm_area_struct {
+ 	struct mempolicy *vm_policy;	/* NUMA policy for the VMA */
+ #endif
+ 	struct vm_userfaultfd_ctx vm_userfaultfd_ctx;
++#ifdef CONFIG_UKSM
++	struct vma_slot *uksm_vma_slot;
++#endif
+ } __randomize_layout;
+ 
+ struct core_thread {
+diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
+index 1b9de7d220fb..84519d66a253 100644
+--- a/include/linux/mmzone.h
++++ b/include/linux/mmzone.h
+@@ -162,6 +162,9 @@ enum zone_stat_item {
+ 	NR_ZSPAGES,		/* allocated in zsmalloc */
+ #endif
+ 	NR_FREE_CMA_PAGES,
++#ifdef CONFIG_UKSM
++	NR_UKSM_ZERO_PAGES,
++#endif
+ 	NR_VM_ZONE_STAT_ITEMS };
+ 
+ enum node_stat_item {
+diff --git a/include/linux/sradix-tree.h b/include/linux/sradix-tree.h
+new file mode 100644
+index 000000000000..d71edba6b63f
+--- /dev/null
++++ b/include/linux/sradix-tree.h
+@@ -0,0 +1,77 @@
++#ifndef _LINUX_SRADIX_TREE_H
++#define _LINUX_SRADIX_TREE_H
++
++
++#define INIT_SRADIX_TREE(root, mask)					\
++do {									\
++	(root)->height = 0;						\
++	(root)->gfp_mask = (mask);					\
++	(root)->rnode = NULL;						\
++} while (0)
++
++#define ULONG_BITS	(sizeof(unsigned long) * 8)
++#define SRADIX_TREE_INDEX_BITS  (8 /* CHAR_BIT */ * sizeof(unsigned long))
++//#define SRADIX_TREE_MAP_SHIFT	6
++//#define SRADIX_TREE_MAP_SIZE	(1UL << SRADIX_TREE_MAP_SHIFT)
++//#define SRADIX_TREE_MAP_MASK	(SRADIX_TREE_MAP_SIZE-1)
++
++struct sradix_tree_node {
++	unsigned int	height;		/* Height from the bottom */
++	unsigned int	count;
++	unsigned int	fulls;		/* Number of full sublevel trees */
++	struct sradix_tree_node *parent;
++	void *stores[0];
++};
++
++/* A simple radix tree implementation */
++struct sradix_tree_root {
++	unsigned int            height;
++	struct sradix_tree_node *rnode;
++
++	/* Where found to have available empty stores in its sublevels */
++	struct sradix_tree_node *enter_node;
++	unsigned int shift;
++	unsigned int stores_size;
++	unsigned int mask;
++	unsigned long min;	/* The first hole index */
++	unsigned long num;
++	//unsigned long *height_to_maxindex;
++
++	/* How the node is allocated and freed. */
++	struct sradix_tree_node *(*alloc)(void);
++	void (*free)(struct sradix_tree_node *node);
++
++	/* When a new node is added and removed */
++	void (*extend)(struct sradix_tree_node *parent, struct sradix_tree_node *child);
++	void (*assign)(struct sradix_tree_node *node, unsigned int index, void *item);
++	void (*rm)(struct sradix_tree_node *node, unsigned int offset);
++};
++
++struct sradix_tree_path {
++	struct sradix_tree_node *node;
++	int offset;
++};
++
++static inline
++void init_sradix_tree_root(struct sradix_tree_root *root, unsigned long shift)
++{
++	root->height = 0;
++	root->rnode = NULL;
++	root->shift = shift;
++	root->stores_size = 1UL << shift;
++	root->mask = root->stores_size - 1;
++}
++
++
++extern void *sradix_tree_next(struct sradix_tree_root *root,
++		       struct sradix_tree_node *node, unsigned long index,
++		       int (*iter)(void *, unsigned long));
++
++extern int sradix_tree_enter(struct sradix_tree_root *root, void **item, int num);
++
++extern void sradix_tree_delete_from_leaf(struct sradix_tree_root *root,
++			struct sradix_tree_node *node, unsigned long index);
++
++extern void *sradix_tree_lookup(struct sradix_tree_root *root, unsigned long index);
++
++#endif /* _LINUX_SRADIX_TREE_H */
+diff --git a/include/linux/uksm.h b/include/linux/uksm.h
+new file mode 100644
+index 000000000000..bb8651f534f2
+--- /dev/null
++++ b/include/linux/uksm.h
+@@ -0,0 +1,149 @@
++#ifndef __LINUX_UKSM_H
++#define __LINUX_UKSM_H
++/*
++ * Memory merging support.
++ *
++ * This code enables dynamic sharing of identical pages found in different
++ * memory areas, even if they are not shared by fork().
++ */
++
++/* if !CONFIG_UKSM this file should not be compiled at all. */
++#ifdef CONFIG_UKSM
++
++#include <linux/bitops.h>
++#include <linux/mm.h>
++#include <linux/pagemap.h>
++#include <linux/rmap.h>
++#include <linux/sched.h>
++
++extern unsigned long zero_pfn __read_mostly;
++extern unsigned long uksm_zero_pfn __read_mostly;
++extern struct page *empty_uksm_zero_page;
++
++/* must be done before linked to mm */
++extern void uksm_vma_add_new(struct vm_area_struct *vma);
++extern void uksm_remove_vma(struct vm_area_struct *vma);
++
++#define UKSM_SLOT_NEED_SORT	(1 << 0)
++#define UKSM_SLOT_NEED_RERAND	(1 << 1)
++#define UKSM_SLOT_SCANNED	(1 << 2) /* It's scanned in this round */
++#define UKSM_SLOT_FUL_SCANNED	(1 << 3)
++#define UKSM_SLOT_IN_UKSM	(1 << 4)
++
++struct vma_slot {
++	struct sradix_tree_node *snode;
++	unsigned long sindex;
++
++	struct list_head slot_list;
++	unsigned long fully_scanned_round;
++	unsigned long dedup_num;
++	unsigned long pages_scanned;
++	unsigned long this_sampled;
++	unsigned long last_scanned;
++	unsigned long pages_to_scan;
++	struct scan_rung *rung;
++	struct page **rmap_list_pool;
++	unsigned int *pool_counts;
++	unsigned long pool_size;
++	struct vm_area_struct *vma;
++	struct mm_struct *mm;
++	unsigned long ctime_j;
++	unsigned long pages;
++	unsigned long flags;
++	unsigned long pages_cowed; /* pages cowed this round */
++	unsigned long pages_merged; /* pages merged this round */
++	unsigned long pages_bemerged;
++
++	/* when it has page merged in this eval round */
++	struct list_head dedup_list;
++};
++
++static inline void uksm_unmap_zero_page(pte_t pte)
++{
++	if (pte_pfn(pte) == uksm_zero_pfn)
++		__dec_zone_page_state(empty_uksm_zero_page, NR_UKSM_ZERO_PAGES);
++}
++
++static inline void uksm_map_zero_page(pte_t pte)
++{
++	if (pte_pfn(pte) == uksm_zero_pfn)
++		__inc_zone_page_state(empty_uksm_zero_page, NR_UKSM_ZERO_PAGES);
++}
++
++static inline void uksm_cow_page(struct vm_area_struct *vma, struct page *page)
++{
++	if (vma->uksm_vma_slot && PageKsm(page))
++		vma->uksm_vma_slot->pages_cowed++;
++}
++
++static inline void uksm_cow_pte(struct vm_area_struct *vma, pte_t pte)
++{
++	if (vma->uksm_vma_slot && pte_pfn(pte) == uksm_zero_pfn)
++		vma->uksm_vma_slot->pages_cowed++;
++}
++
++static inline int uksm_flags_can_scan(unsigned long vm_flags)
++{
++#ifdef VM_SAO
++		if (vm_flags & VM_SAO)
++			return 0;
++#endif
++
++	return !(vm_flags & (VM_PFNMAP | VM_IO  | VM_DONTEXPAND |
++			     VM_HUGETLB | VM_MIXEDMAP | VM_SHARED
++			     | VM_MAYSHARE | VM_GROWSUP | VM_GROWSDOWN));
++}
++
++static inline void uksm_vm_flags_mod(unsigned long *vm_flags_p)
++{
++	if (uksm_flags_can_scan(*vm_flags_p))
++		*vm_flags_p |= VM_MERGEABLE;
++}
++
++/*
++ * Just a wrapper for BUG_ON for where ksm_zeropage must not be. TODO: it will
++ * be removed when uksm zero page patch is stable enough.
++ */
++static inline void uksm_bugon_zeropage(pte_t pte)
++{
++	BUG_ON(pte_pfn(pte) == uksm_zero_pfn);
++}
++#else
++static inline void uksm_vma_add_new(struct vm_area_struct *vma)
++{
++}
++
++static inline void uksm_remove_vma(struct vm_area_struct *vma)
++{
++}
++
++static inline void uksm_unmap_zero_page(pte_t pte)
++{
++}
++
++static inline void uksm_map_zero_page(pte_t pte)
++{
++}
++
++static inline void uksm_cow_page(struct vm_area_struct *vma, struct page *page)
++{
++}
++
++static inline void uksm_cow_pte(struct vm_area_struct *vma, pte_t pte)
++{
++}
++
++static inline int uksm_flags_can_scan(unsigned long vm_flags)
++{
++	return 0;
++}
++
++static inline void uksm_vm_flags_mod(unsigned long *vm_flags_p)
++{
++}
++
++static inline void uksm_bugon_zeropage(pte_t pte)
++{
++}
++#endif /* !CONFIG_UKSM */
++#endif /* __LINUX_UKSM_H */
+diff --git a/kernel/fork.c b/kernel/fork.c
+index 48ed22774efa..9b5b1ff9babf 100644
+--- a/kernel/fork.c
++++ b/kernel/fork.c
+@@ -601,7 +601,7 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm,
+ 		__vma_link_rb(mm, tmp, rb_link, rb_parent);
+ 		rb_link = &tmp->vm_rb.rb_right;
+ 		rb_parent = &tmp->vm_rb;
+-
++		uksm_vma_add_new(tmp);
+ 		mm->map_count++;
+ 		if (!(tmp->vm_flags & VM_WIPEONFORK))
+ 			retval = copy_page_range(mm, oldmm, mpnt);
+diff --git a/lib/Makefile b/lib/Makefile
+index 685aee60de1d..e9584575b51c 100644
+--- a/lib/Makefile
++++ b/lib/Makefile
+@@ -26,7 +26,7 @@ CFLAGS_string.o := $(call cc-option, -fno-stack-protector)
+ endif
+ 
+ lib-y := ctype.o string.o vsprintf.o cmdline.o \
+-	 rbtree.o radix-tree.o timerqueue.o xarray.o \
++	 rbtree.o radix-tree.o sradix-tree.o timerqueue.o xarray.o \
+ 	 idr.o extable.o sha1.o irq_regs.o argv_split.o \
+ 	 flex_proportions.o ratelimit.o show_mem.o \
+ 	 is_single_threaded.o plist.o decompress.o kobject_uevent.o \
+diff --git a/lib/sradix-tree.c b/lib/sradix-tree.c
+new file mode 100644
+index 000000000000..ab21e6309b93
+--- /dev/null
++++ b/lib/sradix-tree.c
+@@ -0,0 +1,476 @@
++#include <linux/errno.h>
++#include <linux/mm.h>
++#include <linux/mman.h>
++#include <linux/spinlock.h>
++#include <linux/slab.h>
++#include <linux/gcd.h>
++#include <linux/sradix-tree.h>
++
++static inline int sradix_node_full(struct sradix_tree_root *root, struct sradix_tree_node *node)
++{
++	return node->fulls == root->stores_size ||
++		(node->height == 1 && node->count == root->stores_size);
++}
++
++/*
++ *	Extend a sradix tree so it can store key @index.
++ */
++static int sradix_tree_extend(struct sradix_tree_root *root, unsigned long index)
++{
++	struct sradix_tree_node *node;
++	unsigned int height;
++
++	if (unlikely(root->rnode == NULL)) {
++		if (!(node = root->alloc()))
++			return -ENOMEM;
++
++		node->height = 1;
++		root->rnode = node;
++		root->height = 1;
++	}
++
++	/* Figure out what the height should be.  */
++	height = root->height;
++	index >>= root->shift * height;
++
++	while (index) {
++		index >>= root->shift;
++		height++;
++	}
++
++	while (height > root->height) {
++		unsigned int newheight;
++
++		if (!(node = root->alloc()))
++			return -ENOMEM;
++
++		/* Increase the height.  */
++		node->stores[0] = root->rnode;
++		root->rnode->parent = node;
++		if (root->extend)
++			root->extend(node, root->rnode);
++
++		newheight = root->height + 1;
++		node->height = newheight;
++		node->count = 1;
++		if (sradix_node_full(root, root->rnode))
++			node->fulls = 1;
++
++		root->rnode = node;
++		root->height = newheight;
++	}
++
++	return 0;
++}
++
++/*
++ * Search the next item from the current node, that is not NULL
++ * and can satify root->iter().
++ */
++void *sradix_tree_next(struct sradix_tree_root *root,
++		       struct sradix_tree_node *node, unsigned long index,
++		       int (*iter)(void *item, unsigned long height))
++{
++	unsigned long offset;
++	void *item;
++
++	if (unlikely(node == NULL)) {
++		node = root->rnode;
++		for (offset = 0; offset < root->stores_size; offset++) {
++			item = node->stores[offset];
++			if (item && (!iter || iter(item, node->height)))
++				break;
++		}
++
++		if (unlikely(offset >= root->stores_size))
++			return NULL;
++
++		if (node->height == 1)
++			return item;
++		else
++			goto go_down;
++	}
++
++	while (node) {
++		offset = (index & root->mask) + 1;
++		for (; offset < root->stores_size; offset++) {
++			item = node->stores[offset];
++			if (item && (!iter || iter(item, node->height)))
++				break;
++		}
++
++		if (offset < root->stores_size)
++			break;
++
++		node = node->parent;
++		index >>= root->shift;
++	}
++
++	if (!node)
++		return NULL;
++
++	while (node->height > 1) {
++go_down:
++		node = item;
++		for (offset = 0; offset < root->stores_size; offset++) {
++			item = node->stores[offset];
++			if (item && (!iter || iter(item, node->height)))
++				break;
++		}
++
++		if (unlikely(offset >= root->stores_size))
++			return NULL;
++	}
++
++	BUG_ON(offset > root->stores_size);
++
++	return item;
++}
++
++/*
++ * Blindly insert the item to the tree. Typically, we reuse the
++ * first empty store item.
++ */
++int sradix_tree_enter(struct sradix_tree_root *root, void **item, int num)
++{
++	unsigned long index;
++	unsigned int height;
++	struct sradix_tree_node *node, *tmp = NULL;
++	int offset, offset_saved;
++	void **store = NULL;
++	int error, i, j, shift;
++
++go_on:
++	index = root->min;
++
++	if (root->enter_node && !sradix_node_full(root, root->enter_node)) {
++		node = root->enter_node;
++		BUG_ON((index >> (root->shift * root->height)));
++	} else {
++		node = root->rnode;
++		if (node == NULL || (index >> (root->shift * root->height))
++		    || sradix_node_full(root, node)) {
++			error = sradix_tree_extend(root, index);
++			if (error)
++				return error;
++
++			node = root->rnode;
++		}
++	}
++
++
++	height = node->height;
++	shift = (height - 1) * root->shift;
++	offset = (index >> shift) & root->mask;
++	while (shift > 0) {
++		offset_saved = offset;
++		for (; offset < root->stores_size; offset++) {
++			store = &node->stores[offset];
++			tmp = *store;
++
++			if (!tmp || !sradix_node_full(root, tmp))
++				break;
++		}
++		BUG_ON(offset >= root->stores_size);
++
++		if (offset != offset_saved) {
++			index += (offset - offset_saved) << shift;
++			index &= ~((1UL << shift) - 1);
++		}
++
++		if (!tmp) {
++			if (!(tmp = root->alloc()))
++				return -ENOMEM;
++
++			tmp->height = shift / root->shift;
++			*store = tmp;
++			tmp->parent = node;
++			node->count++;
++//			if (root->extend)
++//				root->extend(node, tmp);
++		}
++
++		node = tmp;
++		shift -= root->shift;
++		offset = (index >> shift) & root->mask;
++	}
++
++	BUG_ON(node->height != 1);
++
++
++	store = &node->stores[offset];
++	for (i = 0, j = 0;
++	      j < root->stores_size - node->count &&
++	      i < root->stores_size - offset && j < num; i++) {
++		if (!store[i]) {
++			store[i] = item[j];
++			if (root->assign)
++				root->assign(node, index + i, item[j]);
++			j++;
++		}
++	}
++
++	node->count += j;
++	root->num += j;
++	num -= j;
++
++	while (sradix_node_full(root, node)) {
++		node = node->parent;
++		if (!node)
++			break;
++
++		node->fulls++;
++	}
++
++	if (unlikely(!node)) {
++		/* All nodes are full */
++		root->min = 1 << (root->height * root->shift);
++		root->enter_node = NULL;
++	} else {
++		root->min = index + i - 1;
++		root->min |= (1UL << (node->height - 1)) - 1;
++		root->min++;
++		root->enter_node = node;
++	}
++
++	if (num) {
++		item += j;
++		goto go_on;
++	}
++
++	return 0;
++}
++
++
++/**
++ *	sradix_tree_shrink    -    shrink height of a sradix tree to minimal
++ *      @root		sradix tree root
++ *
++ */
++static inline void sradix_tree_shrink(struct sradix_tree_root *root)
++{
++	/* try to shrink tree height */
++	while (root->height > 1) {
++		struct sradix_tree_node *to_free = root->rnode;
++
++		/*
++		 * The candidate node has more than one child, or its child
++		 * is not at the leftmost store, we cannot shrink.
++		 */
++		if (to_free->count != 1 || !to_free->stores[0])
++			break;
++
++		root->rnode = to_free->stores[0];
++		root->rnode->parent = NULL;
++		root->height--;
++		if (unlikely(root->enter_node == to_free))
++			root->enter_node = NULL;
++		root->free(to_free);
++	}
++}
++
++/*
++ * Del the item on the known leaf node and index
++ */
++void sradix_tree_delete_from_leaf(struct sradix_tree_root *root,
++				  struct sradix_tree_node *node, unsigned long index)
++{
++	unsigned int offset;
++	struct sradix_tree_node *start, *end;
++
++	BUG_ON(node->height != 1);
++
++	start = node;
++	while (node && !(--node->count))
++		node = node->parent;
++
++	end = node;
++	if (!node) {
++		root->rnode = NULL;
++		root->height = 0;
++		root->min = 0;
++		root->num = 0;
++		root->enter_node = NULL;
++	} else {
++		offset = (index >> (root->shift * (node->height - 1))) & root->mask;
++		if (root->rm)
++			root->rm(node, offset);
++		node->stores[offset] = NULL;
++		root->num--;
++		if (root->min > index) {
++			root->min = index;
++			root->enter_node = node;
++		}
++	}
++
++	if (start != end) {
++		do {
++			node = start;
++			start = start->parent;
++			if (unlikely(root->enter_node == node))
++				root->enter_node = end;
++			root->free(node);
++		} while (start != end);
++
++		/*
++		 * Note that shrink may free "end", so enter_node still need to
++		 * be checked inside.
++		 */
++		sradix_tree_shrink(root);
++	} else if (node->count == root->stores_size - 1) {
++		/* It WAS a full leaf node. Update the ancestors */
++		node = node->parent;
++		while (node) {
++			node->fulls--;
++			if (node->fulls != root->stores_size - 1)
++				break;
++
++			node = node->parent;
++		}
++	}
++}
++
++void *sradix_tree_lookup(struct sradix_tree_root *root, unsigned long index)
++{
++	unsigned int height, offset;
++	struct sradix_tree_node *node;
++	int shift;
++
++	node = root->rnode;
++	if (node == NULL || (index >> (root->shift * root->height)))
++		return NULL;
++
++	height = root->height;
++	shift = (height - 1) * root->shift;
++
++	do {
++		offset = (index >> shift) & root->mask;
++		node = node->stores[offset];
++		if (!node)
++			return NULL;
++
++		shift -= root->shift;
++	} while (shift >= 0);
++
++	return node;
++}
++
++/*
++ * Return the item if it exists, otherwise create it in place
++ * and return the created item.
++ */
++void *sradix_tree_lookup_create(struct sradix_tree_root *root,
++			unsigned long index, void *(*item_alloc)(void))
++{
++	unsigned int height, offset;
++	struct sradix_tree_node *node, *tmp;
++	void *item;
++	int shift, error;
++
++	if (root->rnode == NULL || (index >> (root->shift * root->height))) {
++		if (item_alloc) {
++			error = sradix_tree_extend(root, index);
++			if (error)
++				return NULL;
++		} else {
++			return NULL;
++		}
++	}
++
++	node = root->rnode;
++	height = root->height;
++	shift = (height - 1) * root->shift;
++
++	do {
++		offset = (index >> shift) & root->mask;
++		if (!node->stores[offset]) {
++			if (!(tmp = root->alloc()))
++				return NULL;
++
++			tmp->height = shift / root->shift;
++			node->stores[offset] = tmp;
++			tmp->parent = node;
++			node->count++;
++			node = tmp;
++		} else {
++			node = node->stores[offset];
++		}
++
++		shift -= root->shift;
++	} while (shift > 0);
++
++	BUG_ON(node->height != 1);
++	offset = index & root->mask;
++	if (node->stores[offset]) {
++		return node->stores[offset];
++	} else if (item_alloc) {
++		if (!(item = item_alloc()))
++			return NULL;
++
++		node->stores[offset] = item;
++
++		/*
++		 * NOTE: we do NOT call root->assign here, since this item is
++		 * newly created by us having no meaning. Caller can call this
++		 * if it's necessary to do so.
++		 */
++
++		node->count++;
++		root->num++;
++
++		while (sradix_node_full(root, node)) {
++			node = node->parent;
++			if (!node)
++				break;
++
++			node->fulls++;
++		}
++
++		if (unlikely(!node)) {
++			/* All nodes are full */
++			root->min = 1 << (root->height * root->shift);
++		} else {
++			if (root->min == index) {
++				root->min |= (1UL << (node->height - 1)) - 1;
++				root->min++;
++				root->enter_node = node;
++			}
++		}
++
++		return item;
++	} else {
++		return NULL;
++	}
++
++}
++
++int sradix_tree_delete(struct sradix_tree_root *root, unsigned long index)
++{
++	unsigned int height, offset;
++	struct sradix_tree_node *node;
++	int shift;
++
++	node = root->rnode;
++	if (node == NULL || (index >> (root->shift * root->height)))
++		return -ENOENT;
++
++	height = root->height;
++	shift = (height - 1) * root->shift;
++
++	do {
++		offset = (index >> shift) & root->mask;
++		node = node->stores[offset];
++		if (!node)
++			return -ENOENT;
++
++		shift -= root->shift;
++	} while (shift > 0);
++
++	offset = index & root->mask;
++	if (!node->stores[offset])
++		return -ENOENT;
++
++	sradix_tree_delete_from_leaf(root, node, index);
++
++	return 0;
++}
+diff --git a/mm/Kconfig b/mm/Kconfig
+index c1acc34c1c35..ecdcd88f5275 100644
+--- a/mm/Kconfig
++++ b/mm/Kconfig
+@@ -316,6 +316,32 @@ config KSM
+ 	  See Documentation/vm/ksm.rst for more information: KSM is inactive
+ 	  until a program has madvised that an area is MADV_MERGEABLE, and
+ 	  root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set).
++choice
++	prompt "Choose UKSM/KSM strategy"
++	default UKSM
++	depends on KSM
++	help
++	  This option allows to select a UKSM/KSM stragety.
++
++config UKSM
++	bool "Ultra-KSM for page merging"
++	depends on KSM
++	help
++	UKSM is inspired by the Linux kernel project \u2014 KSM(Kernel Same
++	page Merging), but with a fundamentally rewritten core algorithm. With
++	an advanced algorithm, UKSM now can transparently scans all anonymously
++	mapped user space applications with an significantly improved scan speed
++	and CPU efficiency. Since KVM is friendly to KSM, KVM can also benefit from
++	UKSM. Now UKSM has its first stable release and first real world enterprise user.
++	For more information, please goto its project page.
++	(github.com/dolohow/uksm)
++
++config KSM_LEGACY
++	bool "Legacy KSM implementation"
++	depends on KSM
++	help
++	The legacy KSM implementation from Red Hat.
++endchoice
+ 
+ config DEFAULT_MMAP_MIN_ADDR
+ 	int "Low address space to protect from user allocation"
+diff --git a/mm/Makefile b/mm/Makefile
+index fccd3756b25f..c157e08bd87a 100644
+--- a/mm/Makefile
++++ b/mm/Makefile
+@@ -68,7 +68,8 @@ obj-$(CONFIG_SPARSEMEM)	+= sparse.o
+ obj-$(CONFIG_SPARSEMEM_VMEMMAP) += sparse-vmemmap.o
+ obj-$(CONFIG_SLOB) += slob.o
+ obj-$(CONFIG_MMU_NOTIFIER) += mmu_notifier.o
+-obj-$(CONFIG_KSM) += ksm.o
++obj-$(CONFIG_KSM_LEGACY) += ksm.o
++obj-$(CONFIG_UKSM) += uksm.o
+ obj-$(CONFIG_PAGE_POISONING) += page_poison.o
+ obj-$(CONFIG_SLAB) += slab.o
+ obj-$(CONFIG_SLUB) += slub.o
+diff --git a/mm/ksm.c b/mm/ksm.c
+index 281c00129a2e..48d9ac624335 100644
+--- a/mm/ksm.c
++++ b/mm/ksm.c
+@@ -857,17 +857,6 @@ static int unmerge_ksm_pages(struct vm_area_struct *vma,
+ 	return err;
+ }
+ 
+-static inline struct stable_node *page_stable_node(struct page *page)
+-{
+-	return PageKsm(page) ? page_rmapping(page) : NULL;
+-}
+-
+-static inline void set_page_stable_node(struct page *page,
+-					struct stable_node *stable_node)
+-{
+-	page->mapping = (void *)((unsigned long)stable_node | PAGE_MAPPING_KSM);
+-}
+-
+ #ifdef CONFIG_SYSFS
+ /*
+  * Only called through the sysfs control interface:
+diff --git a/mm/memory.c b/mm/memory.c
+index f703fe8c8346..0bedec37098f 100644
+--- a/mm/memory.c
++++ b/mm/memory.c
+@@ -144,6 +144,25 @@ EXPORT_SYMBOL(zero_pfn);
+ 
+ unsigned long highest_memmap_pfn __read_mostly;
+ 
++#ifdef CONFIG_UKSM
++unsigned long uksm_zero_pfn __read_mostly;
++EXPORT_SYMBOL_GPL(uksm_zero_pfn);
++struct page *empty_uksm_zero_page;
++
++static int __init setup_uksm_zero_page(void)
++{
++	empty_uksm_zero_page = alloc_pages(__GFP_ZERO & ~__GFP_MOVABLE, 0);
++	if (!empty_uksm_zero_page)
++		panic("Oh boy, that early out of memory?");
++
++	SetPageReserved(empty_uksm_zero_page);
++	uksm_zero_pfn = page_to_pfn(empty_uksm_zero_page);
++
++	return 0;
++}
++core_initcall(setup_uksm_zero_page);
++#endif
++
+ /*
+  * CONFIG_MMU architectures set up ZERO_PAGE in their paging_init()
+  */
+@@ -159,6 +178,7 @@ void mm_trace_rss_stat(struct mm_struct *mm, int member, long count)
+ 	trace_rss_stat(mm, member, count);
+ }
+ 
++
+ #if defined(SPLIT_RSS_COUNTING)
+ 
+ void sync_mm_rss(struct mm_struct *mm)
+@@ -802,8 +822,13 @@ copy_one_pte(struct mm_struct *dst_mm, struct mm_struct *src_mm,
+ 		get_page(page);
+ 		page_dup_rmap(page, false);
+ 		rss[mm_counter(page)]++;
++
++		/* Should return NULL in vm_normal_page() */
++		uksm_bugon_zeropage(pte);
+ 	} else if (pte_devmap(pte)) {
+ 		page = pte_page(pte);
++	} else {
++		uksm_map_zero_page(pte);
+ 	}
+ 
+ out_set_pte:
+@@ -1076,8 +1101,10 @@ static unsigned long zap_pte_range(struct mmu_gather *tlb,
+ 			ptent = ptep_get_and_clear_full(mm, addr, pte,
+ 							tlb->fullmm);
+ 			tlb_remove_tlb_entry(tlb, pte, addr);
+-			if (unlikely(!page))
++			if (unlikely(!page)) {
++				uksm_unmap_zero_page(ptent);
+ 				continue;
++			}
+ 
+ 			if (!PageAnon(page)) {
+ 				if (pte_dirty(ptent)) {
+@@ -2412,6 +2439,7 @@ static inline bool cow_user_page(struct page *dst, struct page *src,
+ 
+ 	if (likely(src)) {
+ 		copy_user_highpage(dst, src, addr, vma);
++		uksm_cow_page(vma, src);
+ 		return true;
+ 	}
+ 
+@@ -2658,6 +2686,7 @@ static vm_fault_t wp_page_copy(struct vm_fault *vmf)
+ 							      vmf->address);
+ 		if (!new_page)
+ 			goto oom;
++		uksm_cow_pte(vma, vmf->orig_pte);
+ 	} else {
+ 		new_page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma,
+ 				vmf->address);
+@@ -2699,7 +2728,9 @@ static vm_fault_t wp_page_copy(struct vm_fault *vmf)
+ 						mm_counter_file(old_page));
+ 				inc_mm_counter_fast(mm, MM_ANONPAGES);
+ 			}
++			uksm_bugon_zeropage(vmf->orig_pte);
+ 		} else {
++			uksm_unmap_zero_page(vmf->orig_pte);
+ 			inc_mm_counter_fast(mm, MM_ANONPAGES);
+ 		}
+ 		flush_cache_page(vma, vmf->address, pte_pfn(vmf->orig_pte));
+diff --git a/mm/mmap.c b/mm/mmap.c
+index f609e9ec4a25..464058e2f637 100644
+--- a/mm/mmap.c
++++ b/mm/mmap.c
+@@ -46,6 +46,7 @@
+ #include <linux/moduleparam.h>
+ #include <linux/pkeys.h>
+ #include <linux/oom.h>
++#include <linux/ksm.h>
+ #include <linux/sched/mm.h>
+ 
+ #include <linux/uaccess.h>
+@@ -181,6 +182,7 @@ static struct vm_area_struct *remove_vma(struct vm_area_struct *vma)
+ 	if (vma->vm_file)
+ 		fput(vma->vm_file);
+ 	mpol_put(vma_policy(vma));
++       uksm_remove_vma(vma);
+ 	vm_area_free(vma);
+ 	return next;
+ }
+@@ -708,9 +710,16 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
+ 	long adjust_next = 0;
+ 	int remove_next = 0;
+ 
++/*
++ * to avoid deadlock, ksm_remove_vma must be done before any spin_lock is
++ * acquired
++ */
++	uksm_remove_vma(vma);
++
+ 	if (next && !insert) {
+ 		struct vm_area_struct *exporter = NULL, *importer = NULL;
+ 
++		uksm_remove_vma(next);
+ 		if (end >= next->vm_end) {
+ 			/*
+ 			 * vma expands, overlapping all the next, and
+@@ -841,6 +850,7 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
+ 		end_changed = true;
+ 	}
+ 	vma->vm_pgoff = pgoff;
++
+ 	if (adjust_next) {
+ 		next->vm_start += adjust_next << PAGE_SHIFT;
+ 		next->vm_pgoff += adjust_next;
+@@ -946,6 +956,7 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
+ 		if (remove_next == 2) {
+ 			remove_next = 1;
+ 			end = next->vm_end;
++			uksm_remove_vma(next);
+ 			goto again;
+ 		}
+ 		else if (next)
+@@ -972,10 +983,14 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start,
+ 			 */
+ 			VM_WARN_ON(mm->highest_vm_end != vm_end_gap(vma));
+ 		}
++	} else {
++		if (next && !insert)
++			uksm_vma_add_new(next);
+ 	}
+ 	if (insert && file)
+ 		uprobe_mmap(insert);
+ 
++	uksm_vma_add_new(vma);
+ 	validate_mm(mm);
+ 
+ 	return 0;
+@@ -1434,6 +1449,9 @@ unsigned long do_mmap(struct file *file, unsigned long addr,
+ 	vm_flags |= calc_vm_prot_bits(prot, pkey) | calc_vm_flag_bits(flags) |
+ 			mm->def_flags | VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC;
+ 
++	/* If uksm is enabled, we add VM_MERGEABLE to new VMAs. */
++	uksm_vm_flags_mod(&vm_flags);
++
+ 	if (flags & MAP_LOCKED)
+ 		if (!can_do_mlock())
+ 			return -EPERM;
+@@ -1801,6 +1819,7 @@ unsigned long mmap_region(struct file *file, unsigned long addr,
+ 			allow_write_access(file);
+ 	}
+ 	file = vma->vm_file;
++	uksm_vma_add_new(vma);
+ out:
+ 	perf_event_mmap(vma);
+ 
+@@ -1843,6 +1862,7 @@ unsigned long mmap_region(struct file *file, unsigned long addr,
+ 	if (vm_flags & VM_DENYWRITE)
+ 		allow_write_access(file);
+ free_vma:
++	uksm_remove_vma(vma);
+ 	vm_area_free(vma);
+ unacct_error:
+ 	if (charged)
+@@ -2694,6 +2714,8 @@ int __split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
+ 	else
+ 		err = vma_adjust(vma, vma->vm_start, addr, vma->vm_pgoff, new);
+ 
++	uksm_vma_add_new(new);
++
+ 	/* Success. */
+ 	if (!err)
+ 		return 0;
+@@ -3000,6 +3022,7 @@ static int do_brk_flags(unsigned long addr, unsigned long len, unsigned long fla
+ 	if ((flags & (~VM_EXEC)) != 0)
+ 		return -EINVAL;
+ 	flags |= VM_DATA_DEFAULT_FLAGS | VM_ACCOUNT | mm->def_flags;
++	uksm_vm_flags_mod(&flags);
+ 
+ 	mapped_addr = get_unmapped_area(NULL, addr, len, 0, MAP_FIXED);
+ 	if (IS_ERR_VALUE(mapped_addr))
+@@ -3050,6 +3073,7 @@ static int do_brk_flags(unsigned long addr, unsigned long len, unsigned long fla
+ 	vma->vm_flags = flags;
+ 	vma->vm_page_prot = vm_get_page_prot(flags);
+ 	vma_link(mm, vma, prev, rb_link, rb_parent);
++	uksm_vma_add_new(vma);
+ out:
+ 	perf_event_mmap(vma);
+ 	mm->total_vm += len >> PAGE_SHIFT;
+@@ -3127,6 +3151,12 @@ void exit_mmap(struct mm_struct *mm)
+ 		up_write(&mm->mmap_sem);
+ 	}
+ 
++	/*
++	 * Taking write lock on mmap_sem does not harm others,
++	 * but it's crucial for uksm to avoid races.
++	 */
++	down_write(&mm->mmap_sem);
++
+ 	if (mm->locked_vm) {
+ 		vma = mm->mmap;
+ 		while (vma) {
+@@ -3161,6 +3191,11 @@ void exit_mmap(struct mm_struct *mm)
+ 		vma = remove_vma(vma);
+ 	}
+ 	vm_unacct_memory(nr_accounted);
++
++	mm->mmap = NULL;
++	mm->mm_rb = RB_ROOT;
++	vmacache_invalidate(mm);
++	up_write(&mm->mmap_sem);
+ }
+ 
+ /* Insert vm structure into process list sorted by address
+@@ -3268,6 +3303,7 @@ struct vm_area_struct *copy_vma(struct vm_area_struct **vmap,
+ 			new_vma->vm_ops->open(new_vma);
+ 		vma_link(mm, new_vma, prev, rb_link, rb_parent);
+ 		*need_rmap_locks = false;
++		uksm_vma_add_new(new_vma);
+ 	}
+ 	return new_vma;
+ 
+@@ -3420,6 +3456,7 @@ static struct vm_area_struct *__install_special_mapping(
+ 	vm_stat_account(mm, vma->vm_flags, len >> PAGE_SHIFT);
+ 
+ 	perf_event_mmap(vma);
++	uksm_vma_add_new(vma);
+ 
+ 	return vma;
+ 
+diff --git a/mm/uksm.c b/mm/uksm.c
+new file mode 100644
+index 000000000000..ef068d5dc307
+--- /dev/null
++++ b/mm/uksm.c
+@@ -0,0 +1,5613 @@
++/*
++ * Ultra KSM. Copyright (C) 2011-2012 Nai Xia
++ *
++ * This is an improvement upon KSM. Some basic data structures and routines
++ * are borrowed from ksm.c .
++ *
++ * Its new features:
++ * 1. Full system scan:
++ *      It automatically scans all user processes' anonymous VMAs. Kernel-user
++ *      interaction to submit a memory area to KSM is no longer needed.
++ *
++ * 2. Rich area detection:
++ *      It automatically detects rich areas containing abundant duplicated
++ *      pages based. Rich areas are given a full scan speed. Poor areas are
++ *      sampled at a reasonable speed with very low CPU consumption.
++ *
++ * 3. Ultra Per-page scan speed improvement:
++ *      A new hash algorithm is proposed. As a result, on a machine with
++ *      Core(TM)2 Quad Q9300 CPU in 32-bit mode and 800MHZ DDR2 main memory, it
++ *      can scan memory areas that does not contain duplicated pages at speed of
++ *      627MB/sec ~ 2445MB/sec and can merge duplicated areas at speed of
++ *      477MB/sec ~ 923MB/sec.
++ *
++ * 4. Thrashing area avoidance:
++ *      Thrashing area(an VMA that has frequent Ksm page break-out) can be
++ *      filtered out. My benchmark shows it's more efficient than KSM's per-page
++ *      hash value based volatile page detection.
++ *
++ *
++ * 5. Misc changes upon KSM:
++ *      * It has a fully x86-opitmized memcmp dedicated for 4-byte-aligned page
++ *        comparison. It's much faster than default C version on x86.
++ *      * rmap_item now has an struct *page member to loosely cache a
++ *        address-->page mapping, which reduces too much time-costly
++ *        follow_page().
++ *      * The VMA creation/exit procedures are hooked to let the Ultra KSM know.
++ *      * try_to_merge_two_pages() now can revert a pte if it fails. No break_
++ *        ksm is needed for this case.
++ *
++ * 6. Full Zero Page consideration(contributed by Figo Zhang)
++ *    Now uksmd consider full zero pages as special pages and merge them to an
++ *    special unswappable uksm zero page.
++ */
++
++#include <linux/errno.h>
++#include <linux/mm.h>
++#include <linux/fs.h>
++#include <linux/mman.h>
++#include <linux/sched.h>
++#include <linux/sched/mm.h>
++#include <linux/sched/coredump.h>
++#include <linux/sched/cputime.h>
++#include <linux/rwsem.h>
++#include <linux/pagemap.h>
++#include <linux/rmap.h>
++#include <linux/spinlock.h>
++#include <linux/jhash.h>
++#include <linux/delay.h>
++#include <linux/kthread.h>
++#include <linux/wait.h>
++#include <linux/slab.h>
++#include <linux/rbtree.h>
++#include <linux/memory.h>
++#include <linux/mmu_notifier.h>
++#include <linux/swap.h>
++#include <linux/ksm.h>
++#include <linux/crypto.h>
++#include <linux/scatterlist.h>
++#include <crypto/hash.h>
++#include <linux/random.h>
++#include <linux/math64.h>
++#include <linux/gcd.h>
++#include <linux/freezer.h>
++#include <linux/oom.h>
++#include <linux/numa.h>
++#include <linux/sradix-tree.h>
++
++#include <asm/tlbflush.h>
++#include "internal.h"
++
++#ifdef CONFIG_X86
++#undef memcmp
++
++#ifdef CONFIG_X86_32
++#define memcmp memcmpx86_32
++/*
++ * Compare 4-byte-aligned address s1 and s2, with length n
++ */
++int memcmpx86_32(void *s1, void *s2, size_t n)
++{
++	size_t num = n / 4;
++	register int res;
++
++	__asm__ __volatile__
++	(
++	 "testl %3,%3\n\t"
++	 "repe; cmpsd\n\t"
++	 "je        1f\n\t"
++	 "sbbl      %0,%0\n\t"
++	 "orl       $1,%0\n"
++	 "1:"
++	 : "=&a" (res), "+&S" (s1), "+&D" (s2), "+&c" (num)
++	 : "0" (0)
++	 : "cc");
++
++	return res;
++}
++
++/*
++ * Check the page is all zero ?
++ */
++static int is_full_zero(const void *s1, size_t len)
++{
++	unsigned char same;
++
++	len /= 4;
++
++	__asm__ __volatile__
++	("repe; scasl;"
++	 "sete %0"
++	 : "=qm" (same), "+D" (s1), "+c" (len)
++	 : "a" (0)
++	 : "cc");
++
++	return same;
++}
++
++
++#elif defined(CONFIG_X86_64)
++#define memcmp memcmpx86_64
++/*
++ * Compare 8-byte-aligned address s1 and s2, with length n
++ */
++int memcmpx86_64(void *s1, void *s2, size_t n)
++{
++	size_t num = n / 8;
++	register int res;
++
++	__asm__ __volatile__
++	(
++	 "testq %q3,%q3\n\t"
++	 "repe; cmpsq\n\t"
++	 "je        1f\n\t"
++	 "sbbq      %q0,%q0\n\t"
++	 "orq       $1,%q0\n"
++	 "1:"
++	 : "=&a" (res), "+&S" (s1), "+&D" (s2), "+&c" (num)
++	 : "0" (0)
++	 : "cc");
++
++	return res;
++}
++
++static int is_full_zero(const void *s1, size_t len)
++{
++	unsigned char same;
++
++	len /= 8;
++
++	__asm__ __volatile__
++	("repe; scasq;"
++	 "sete %0"
++	 : "=qm" (same), "+D" (s1), "+c" (len)
++	 : "a" (0)
++	 : "cc");
++
++	return same;
++}
++
++#endif
++#else
++static int is_full_zero(const void *s1, size_t len)
++{
++	unsigned long *src = s1;
++	int i;
++
++	len /= sizeof(*src);
++
++	for (i = 0; i < len; i++) {
++		if (src[i])
++			return 0;
++	}
++
++	return 1;
++}
++#endif
++
++#define UKSM_RUNG_ROUND_FINISHED  (1 << 0)
++#define TIME_RATIO_SCALE	10000
++
++#define SLOT_TREE_NODE_SHIFT	8
++#define SLOT_TREE_NODE_STORE_SIZE	(1UL << SLOT_TREE_NODE_SHIFT)
++struct slot_tree_node {
++	unsigned long size;
++	struct sradix_tree_node snode;
++	void *stores[SLOT_TREE_NODE_STORE_SIZE];
++};
++
++static struct kmem_cache *slot_tree_node_cachep;
++
++static struct sradix_tree_node *slot_tree_node_alloc(void)
++{
++	struct slot_tree_node *p;
++
++	p = kmem_cache_zalloc(slot_tree_node_cachep, GFP_KERNEL |
++			      __GFP_NORETRY | __GFP_NOWARN);
++	if (!p)
++		return NULL;
++
++	return &p->snode;
++}
++
++static void slot_tree_node_free(struct sradix_tree_node *node)
++{
++	struct slot_tree_node *p;
++
++	p = container_of(node, struct slot_tree_node, snode);
++	kmem_cache_free(slot_tree_node_cachep, p);
++}
++
++static void slot_tree_node_extend(struct sradix_tree_node *parent,
++				  struct sradix_tree_node *child)
++{
++	struct slot_tree_node *p, *c;
++
++	p = container_of(parent, struct slot_tree_node, snode);
++	c = container_of(child, struct slot_tree_node, snode);
++
++	p->size += c->size;
++}
++
++void slot_tree_node_assign(struct sradix_tree_node *node,
++			   unsigned int index, void *item)
++{
++	struct vma_slot *slot = item;
++	struct slot_tree_node *cur;
++
++	slot->snode = node;
++	slot->sindex = index;
++
++	while (node) {
++		cur = container_of(node, struct slot_tree_node, snode);
++		cur->size += slot->pages;
++		node = node->parent;
++	}
++}
++
++void slot_tree_node_rm(struct sradix_tree_node *node, unsigned int offset)
++{
++	struct vma_slot *slot;
++	struct slot_tree_node *cur;
++	unsigned long pages;
++
++	if (node->height == 1) {
++		slot = node->stores[offset];
++		pages = slot->pages;
++	} else {
++		cur = container_of(node->stores[offset],
++				   struct slot_tree_node, snode);
++		pages = cur->size;
++	}
++
++	while (node) {
++		cur = container_of(node, struct slot_tree_node, snode);
++		cur->size -= pages;
++		node = node->parent;
++	}
++}
++
++unsigned long slot_iter_index;
++int slot_iter(void *item,  unsigned long height)
++{
++	struct slot_tree_node *node;
++	struct vma_slot *slot;
++
++	if (height == 1) {
++		slot = item;
++		if (slot_iter_index < slot->pages) {
++			/*in this one*/
++			return 1;
++		} else {
++			slot_iter_index -= slot->pages;
++			return 0;
++		}
++
++	} else {
++		node = container_of(item, struct slot_tree_node, snode);
++		if (slot_iter_index < node->size) {
++			/*in this one*/
++			return 1;
++		} else {
++			slot_iter_index -= node->size;
++			return 0;
++		}
++	}
++}
++
++
++static inline void slot_tree_init_root(struct sradix_tree_root *root)
++{
++	init_sradix_tree_root(root, SLOT_TREE_NODE_SHIFT);
++	root->alloc = slot_tree_node_alloc;
++	root->free = slot_tree_node_free;
++	root->extend = slot_tree_node_extend;
++	root->assign = slot_tree_node_assign;
++	root->rm = slot_tree_node_rm;
++}
++
++void slot_tree_init(void)
++{
++	slot_tree_node_cachep = kmem_cache_create("slot_tree_node",
++				sizeof(struct slot_tree_node), 0,
++				SLAB_PANIC | SLAB_RECLAIM_ACCOUNT,
++				NULL);
++}
++
++
++/* Each rung of this ladder is a list of VMAs having a same scan ratio */
++struct scan_rung {
++	//struct list_head scanned_list;
++	struct sradix_tree_root vma_root;
++	struct sradix_tree_root vma_root2;
++
++	struct vma_slot *current_scan;
++	unsigned long current_offset;
++
++	/*
++	 * The initial value for current_offset, it should loop over
++	 * [0~ step - 1] to let all slot have its chance to be scanned.
++	 */
++	unsigned long offset_init;
++	unsigned long step; /* dynamic step for current_offset */
++	unsigned int flags;
++	unsigned long pages_to_scan;
++	//unsigned long fully_scanned_slots;
++	/*
++	 * a little bit tricky - if cpu_time_ratio > 0, then the value is the
++	 * the cpu time ratio it can spend in rung_i for every scan
++	 * period. if < 0, then it is the cpu time ratio relative to the
++	 * max cpu percentage user specified. Both in unit of
++	 * 1/TIME_RATIO_SCALE
++	 */
++	int cpu_ratio;
++
++	/*
++	 * How long it will take for all slots in this rung to be fully
++	 * scanned? If it's zero, we don't care about the cover time:
++	 * it's fully scanned.
++	 */
++	unsigned int cover_msecs;
++	//unsigned long vma_num;
++	//unsigned long pages; /* Sum of all slot's pages in rung */
++};
++
++/**
++ * node of either the stable or unstale rbtree
++ *
++ */
++struct tree_node {
++	struct rb_node node; /* link in the main (un)stable rbtree */
++	struct rb_root sub_root; /* rb_root for sublevel collision rbtree */
++	u32 hash;
++	unsigned long count; /* TODO: merged with sub_root */
++	struct list_head all_list; /* all tree nodes in stable/unstable tree */
++};
++
++/**
++ * struct stable_node - node of the stable rbtree
++ * @node: rb node of this ksm page in the stable tree
++ * @hlist: hlist head of rmap_items using this ksm page
++ * @kpfn: page frame number of this ksm page
++ */
++struct stable_node {
++	struct rb_node node; /* link in sub-rbtree */
++	struct tree_node *tree_node; /* it's tree node root in stable tree, NULL if it's in hell list */
++	struct hlist_head hlist;
++	unsigned long kpfn;
++	u32 hash_max; /* if ==0 then it's not been calculated yet */
++	struct list_head all_list; /* in a list for all stable nodes */
++};
++
++/**
++ * struct node_vma - group rmap_items linked in a same stable
++ * node together.
++ */
++struct node_vma {
++	union {
++		struct vma_slot *slot;
++		unsigned long key;  /* slot is used as key sorted on hlist */
++	};
++	struct hlist_node hlist;
++	struct hlist_head rmap_hlist;
++	struct stable_node *head;
++};
++
++/**
++ * struct rmap_item - reverse mapping item for virtual addresses
++ * @rmap_list: next rmap_item in mm_slot's singly-linked rmap_list
++ * @anon_vma: pointer to anon_vma for this mm,address, when in stable tree
++ * @mm: the memory structure this rmap_item is pointing into
++ * @address: the virtual address this rmap_item tracks (+ flags in low bits)
++ * @node: rb node of this rmap_item in the unstable tree
++ * @head: pointer to stable_node heading this list in the stable tree
++ * @hlist: link into hlist of rmap_items hanging off that stable_node
++ */
++struct rmap_item {
++	struct vma_slot *slot;
++	struct page *page;
++	unsigned long address;	/* + low bits used for flags below */
++	unsigned long hash_round;
++	unsigned long entry_index;
++	union {
++		struct {/* when in unstable tree */
++			struct rb_node node;
++			struct tree_node *tree_node;
++			u32 hash_max;
++		};
++		struct { /* when in stable tree */
++			struct node_vma *head;
++			struct hlist_node hlist;
++			struct anon_vma *anon_vma;
++		};
++	};
++} __aligned(4);
++
++struct rmap_list_entry {
++	union {
++		struct rmap_item *item;
++		unsigned long addr;
++	};
++	/* lowest bit is used for is_addr tag */
++} __aligned(4); /* 4 aligned to fit in to pages*/
++
++
++/* Basic data structure definition ends */
++
++
++/*
++ * Flags for rmap_item to judge if it's listed in the stable/unstable tree.
++ * The flags use the low bits of rmap_item.address
++ */
++#define UNSTABLE_FLAG	0x1
++#define STABLE_FLAG	0x2
++#define get_rmap_addr(x)	((x)->address & PAGE_MASK)
++
++/*
++ * rmap_list_entry helpers
++ */
++#define IS_ADDR_FLAG	1
++#define is_addr(ptr)		((unsigned long)(ptr) & IS_ADDR_FLAG)
++#define set_is_addr(ptr)	((ptr) |= IS_ADDR_FLAG)
++#define get_clean_addr(ptr)	(((ptr) & ~(__typeof__(ptr))IS_ADDR_FLAG))
++
++
++/*
++ * High speed caches for frequently allocated and freed structs
++ */
++static struct kmem_cache *rmap_item_cache;
++static struct kmem_cache *stable_node_cache;
++static struct kmem_cache *node_vma_cache;
++static struct kmem_cache *vma_slot_cache;
++static struct kmem_cache *tree_node_cache;
++#define UKSM_KMEM_CACHE(__struct, __flags) kmem_cache_create("uksm_"#__struct,\
++		sizeof(struct __struct), __alignof__(struct __struct),\
++		(__flags), NULL)
++
++/* Array of all scan_rung, uksm_scan_ladder[0] having the minimum scan ratio */
++#define SCAN_LADDER_SIZE 4
++static struct scan_rung uksm_scan_ladder[SCAN_LADDER_SIZE];
++
++/* The evaluation rounds uksmd has finished */
++static unsigned long long uksm_eval_round = 1;
++
++/*
++ * we add 1 to this var when we consider we should rebuild the whole
++ * unstable tree.
++ */
++static unsigned long uksm_hash_round = 1;
++
++/*
++ * How many times the whole memory is scanned.
++ */
++static unsigned long long fully_scanned_round = 1;
++
++/* The total number of virtual pages of all vma slots */
++static u64 uksm_pages_total;
++
++/* The number of pages has been scanned since the start up */
++static u64 uksm_pages_scanned;
++
++static u64 scanned_virtual_pages;
++
++/* The number of pages has been scanned since last encode_benefit call */
++static u64 uksm_pages_scanned_last;
++
++/* If the scanned number is tooo large, we encode it here */
++static u64 pages_scanned_stored;
++
++static unsigned long pages_scanned_base;
++
++/* The number of nodes in the stable tree */
++static unsigned long uksm_pages_shared;
++
++/* The number of page slots additionally sharing those nodes */
++static unsigned long uksm_pages_sharing;
++
++/* The number of nodes in the unstable tree */
++static unsigned long uksm_pages_unshared;
++
++/*
++ * Milliseconds ksmd should sleep between scans,
++ * >= 100ms to be consistent with
++ * scan_time_to_sleep_msec()
++ */
++static unsigned int uksm_sleep_jiffies;
++
++/* The real value for the uksmd next sleep */
++static unsigned int uksm_sleep_real;
++
++/* Saved value for user input uksm_sleep_jiffies when it's enlarged */
++static unsigned int uksm_sleep_saved;
++
++/* Max percentage of cpu utilization ksmd can take to scan in one batch */
++static unsigned int uksm_max_cpu_percentage;
++
++static int uksm_cpu_governor;
++
++static char *uksm_cpu_governor_str[4] = { "full", "medium", "low", "quiet" };
++
++struct uksm_cpu_preset_s {
++	int cpu_ratio[SCAN_LADDER_SIZE];
++	unsigned int cover_msecs[SCAN_LADDER_SIZE];
++	unsigned int max_cpu; /* percentage */
++};
++
++struct uksm_cpu_preset_s uksm_cpu_preset[4] = {
++	{ {20, 40, -2500, -10000}, {1000, 500, 200, 50}, 95},
++	{ {20, 30, -2500, -10000}, {1000, 500, 400, 100}, 50},
++	{ {10, 20, -5000, -10000}, {1500, 1000, 1000, 250}, 20},
++	{ {10, 20, 40, 75}, {2000, 1000, 1000, 1000}, 1},
++};
++
++/* The default value for uksm_ema_page_time if it's not initialized */
++#define UKSM_PAGE_TIME_DEFAULT	500
++
++/*cost to scan one page by expotional moving average in nsecs */
++static unsigned long uksm_ema_page_time = UKSM_PAGE_TIME_DEFAULT;
++
++/* The expotional moving average alpha weight, in percentage. */
++#define EMA_ALPHA	20
++
++/*
++ * The threshold used to filter out thrashing areas,
++ * If it == 0, filtering is disabled, otherwise it's the percentage up-bound
++ * of the thrashing ratio of all areas. Any area with a bigger thrashing ratio
++ * will be considered as having a zero duplication ratio.
++ */
++static unsigned int uksm_thrash_threshold = 50;
++
++/* How much dedup ratio is considered to be abundant*/
++static unsigned int uksm_abundant_threshold = 10;
++
++/* All slots having merged pages in this eval round. */
++struct list_head vma_slot_dedup = LIST_HEAD_INIT(vma_slot_dedup);
++
++/* How many times the ksmd has slept since startup */
++static unsigned long long uksm_sleep_times;
++
++#define UKSM_RUN_STOP	0
++#define UKSM_RUN_MERGE	1
++static unsigned int uksm_run = 1;
++
++static DECLARE_WAIT_QUEUE_HEAD(uksm_thread_wait);
++static DEFINE_MUTEX(uksm_thread_mutex);
++
++/*
++ * List vma_slot_new is for newly created vma_slot waiting to be added by
++ * ksmd. If one cannot be added(e.g. due to it's too small), it's moved to
++ * vma_slot_noadd. vma_slot_del is the list for vma_slot whose corresponding
++ * VMA has been removed/freed.
++ */
++struct list_head vma_slot_new = LIST_HEAD_INIT(vma_slot_new);
++struct list_head vma_slot_noadd = LIST_HEAD_INIT(vma_slot_noadd);
++struct list_head vma_slot_del = LIST_HEAD_INIT(vma_slot_del);
++static DEFINE_SPINLOCK(vma_slot_list_lock);
++
++/* The unstable tree heads */
++static struct rb_root root_unstable_tree = RB_ROOT;
++
++/*
++ * All tree_nodes are in a list to be freed at once when unstable tree is
++ * freed after each scan round.
++ */
++static struct list_head unstable_tree_node_list =
++				LIST_HEAD_INIT(unstable_tree_node_list);
++
++/* List contains all stable nodes */
++static struct list_head stable_node_list = LIST_HEAD_INIT(stable_node_list);
++
++/*
++ * When the hash strength is changed, the stable tree must be delta_hashed and
++ * re-structured. We use two set of below structs to speed up the
++ * re-structuring of stable tree.
++ */
++static struct list_head
++stable_tree_node_list[2] = {LIST_HEAD_INIT(stable_tree_node_list[0]),
++			    LIST_HEAD_INIT(stable_tree_node_list[1])};
++
++static struct list_head *stable_tree_node_listp = &stable_tree_node_list[0];
++static struct rb_root root_stable_tree[2] = {RB_ROOT, RB_ROOT};
++static struct rb_root *root_stable_treep = &root_stable_tree[0];
++static unsigned long stable_tree_index;
++
++/* The hash strength needed to hash a full page */
++#define HASH_STRENGTH_FULL		(PAGE_SIZE / sizeof(u32))
++
++/* The hash strength needed for loop-back hashing */
++#define HASH_STRENGTH_MAX		(HASH_STRENGTH_FULL + 10)
++
++/* The random offsets in a page */
++static u32 *random_nums;
++
++/* The hash strength */
++static unsigned long hash_strength = HASH_STRENGTH_FULL >> 4;
++
++/* The delta value each time the hash strength increases or decreases */
++static unsigned long hash_strength_delta;
++#define HASH_STRENGTH_DELTA_MAX	5
++
++/* The time we have saved due to random_sample_hash */
++static u64 rshash_pos;
++
++/* The time we have wasted due to hash collision */
++static u64 rshash_neg;
++
++struct uksm_benefit {
++	u64 pos;
++	u64 neg;
++	u64 scanned;
++	unsigned long base;
++} benefit;
++
++/*
++ * The relative cost of memcmp, compared to 1 time unit of random sample
++ * hash, this value is tested when ksm module is initialized
++ */
++static unsigned long memcmp_cost;
++
++static unsigned long  rshash_neg_cont_zero;
++static unsigned long  rshash_cont_obscure;
++
++/* The possible states of hash strength adjustment heuristic */
++enum rshash_states {
++		RSHASH_STILL,
++		RSHASH_TRYUP,
++		RSHASH_TRYDOWN,
++		RSHASH_NEW,
++		RSHASH_PRE_STILL,
++};
++
++/* The possible direction we are about to adjust hash strength */
++enum rshash_direct {
++	GO_UP,
++	GO_DOWN,
++	OBSCURE,
++	STILL,
++};
++
++/* random sampling hash state machine */
++static struct {
++	enum rshash_states state;
++	enum rshash_direct pre_direct;
++	u8 below_count;
++	/* Keep a lookup window of size 5, iff above_count/below_count > 3
++	 * in this window we stop trying.
++	 */
++	u8 lookup_window_index;
++	u64 stable_benefit;
++	unsigned long turn_point_down;
++	unsigned long turn_benefit_down;
++	unsigned long turn_point_up;
++	unsigned long turn_benefit_up;
++	unsigned long stable_point;
++} rshash_state;
++
++/*zero page hash table, hash_strength [0 ~ HASH_STRENGTH_MAX]*/
++static u32 *zero_hash_table;
++
++static inline struct node_vma *alloc_node_vma(void)
++{
++	struct node_vma *node_vma;
++
++	node_vma = kmem_cache_zalloc(node_vma_cache, GFP_KERNEL |
++				     __GFP_NORETRY | __GFP_NOWARN);
++	if (node_vma) {
++		INIT_HLIST_HEAD(&node_vma->rmap_hlist);
++		INIT_HLIST_NODE(&node_vma->hlist);
++	}
++	return node_vma;
++}
++
++static inline void free_node_vma(struct node_vma *node_vma)
++{
++	kmem_cache_free(node_vma_cache, node_vma);
++}
++
++
++static inline struct vma_slot *alloc_vma_slot(void)
++{
++	struct vma_slot *slot;
++
++	/*
++	 * In case ksm is not initialized by now.
++	 * Oops, we need to consider the call site of uksm_init() in the future.
++	 */
++	if (!vma_slot_cache)
++		return NULL;
++
++	slot = kmem_cache_zalloc(vma_slot_cache, GFP_KERNEL |
++				 __GFP_NORETRY | __GFP_NOWARN);
++	if (slot) {
++		INIT_LIST_HEAD(&slot->slot_list);
++		INIT_LIST_HEAD(&slot->dedup_list);
++		slot->flags |= UKSM_SLOT_NEED_RERAND;
++	}
++	return slot;
++}
++
++static inline void free_vma_slot(struct vma_slot *vma_slot)
++{
++	kmem_cache_free(vma_slot_cache, vma_slot);
++}
++
++
++
++static inline struct rmap_item *alloc_rmap_item(void)
++{
++	struct rmap_item *rmap_item;
++
++	rmap_item = kmem_cache_zalloc(rmap_item_cache, GFP_KERNEL |
++				      __GFP_NORETRY | __GFP_NOWARN);
++	if (rmap_item) {
++		/* bug on lowest bit is not clear for flag use */
++		BUG_ON(is_addr(rmap_item));
++	}
++	return rmap_item;
++}
++
++static inline void free_rmap_item(struct rmap_item *rmap_item)
++{
++	rmap_item->slot = NULL;	/* debug safety */
++	kmem_cache_free(rmap_item_cache, rmap_item);
++}
++
++static inline struct stable_node *alloc_stable_node(void)
++{
++	struct stable_node *node;
++
++	node = kmem_cache_alloc(stable_node_cache, GFP_KERNEL |
++				__GFP_NORETRY | __GFP_NOWARN);
++	if (!node)
++		return NULL;
++
++	INIT_HLIST_HEAD(&node->hlist);
++	list_add(&node->all_list, &stable_node_list);
++	return node;
++}
++
++static inline void free_stable_node(struct stable_node *stable_node)
++{
++	list_del(&stable_node->all_list);
++	kmem_cache_free(stable_node_cache, stable_node);
++}
++
++static inline struct tree_node *alloc_tree_node(struct list_head *list)
++{
++	struct tree_node *node;
++
++	node = kmem_cache_zalloc(tree_node_cache, GFP_KERNEL |
++				 __GFP_NORETRY | __GFP_NOWARN);
++	if (!node)
++		return NULL;
++
++	list_add(&node->all_list, list);
++	return node;
++}
++
++static inline void free_tree_node(struct tree_node *node)
++{
++	list_del(&node->all_list);
++	kmem_cache_free(tree_node_cache, node);
++}
++
++static void uksm_drop_anon_vma(struct rmap_item *rmap_item)
++{
++	struct anon_vma *anon_vma = rmap_item->anon_vma;
++
++	put_anon_vma(anon_vma);
++}
++
++
++/**
++ * Remove a stable node from stable_tree, may unlink from its tree_node and
++ * may remove its parent tree_node if no other stable node is pending.
++ *
++ * @stable_node	    The node need to be removed
++ * @unlink_rb	    Will this node be unlinked from the rbtree?
++ * @remove_tree_    node Will its tree_node be removed if empty?
++ */
++static void remove_node_from_stable_tree(struct stable_node *stable_node,
++					 int unlink_rb,  int remove_tree_node)
++{
++	struct node_vma *node_vma;
++	struct rmap_item *rmap_item;
++	struct hlist_node *n;
++
++	if (!hlist_empty(&stable_node->hlist)) {
++		hlist_for_each_entry_safe(node_vma, n,
++					  &stable_node->hlist, hlist) {
++			hlist_for_each_entry(rmap_item, &node_vma->rmap_hlist, hlist) {
++				uksm_pages_sharing--;
++
++				uksm_drop_anon_vma(rmap_item);
++				rmap_item->address &= PAGE_MASK;
++			}
++			free_node_vma(node_vma);
++			cond_resched();
++		}
++
++		/* the last one is counted as shared */
++		uksm_pages_shared--;
++		uksm_pages_sharing++;
++	}
++
++	if (stable_node->tree_node && unlink_rb) {
++		rb_erase(&stable_node->node,
++			 &stable_node->tree_node->sub_root);
++
++		if (RB_EMPTY_ROOT(&stable_node->tree_node->sub_root) &&
++		    remove_tree_node) {
++			rb_erase(&stable_node->tree_node->node,
++				 root_stable_treep);
++			free_tree_node(stable_node->tree_node);
++		} else {
++			stable_node->tree_node->count--;
++		}
++	}
++
++	free_stable_node(stable_node);
++}
++
++
++/*
++ * get_uksm_page: checks if the page indicated by the stable node
++ * is still its ksm page, despite having held no reference to it.
++ * In which case we can trust the content of the page, and it
++ * returns the gotten page; but if the page has now been zapped,
++ * remove the stale node from the stable tree and return NULL.
++ *
++ * You would expect the stable_node to hold a reference to the ksm page.
++ * But if it increments the page's count, swapping out has to wait for
++ * ksmd to come around again before it can free the page, which may take
++ * seconds or even minutes: much too unresponsive.  So instead we use a
++ * "keyhole reference": access to the ksm page from the stable node peeps
++ * out through its keyhole to see if that page still holds the right key,
++ * pointing back to this stable node.  This relies on freeing a PageAnon
++ * page to reset its page->mapping to NULL, and relies on no other use of
++ * a page to put something that might look like our key in page->mapping.
++ *
++ * include/linux/pagemap.h page_cache_get_speculative() is a good reference,
++ * but this is different - made simpler by uksm_thread_mutex being held, but
++ * interesting for assuming that no other use of the struct page could ever
++ * put our expected_mapping into page->mapping (or a field of the union which
++ * coincides with page->mapping).  The RCU calls are not for KSM at all, but
++ * to keep the page_count protocol described with page_cache_get_speculative.
++ *
++ * Note: it is possible that get_uksm_page() will return NULL one moment,
++ * then page the next, if the page is in between page_freeze_refs() and
++ * page_unfreeze_refs(): this shouldn't be a problem anywhere, the page
++ * is on its way to being freed; but it is an anomaly to bear in mind.
++ *
++ * @unlink_rb:			if the removal of this node will firstly unlink from
++ * its rbtree. stable_node_reinsert will prevent this when restructuring the
++ * node from its old tree.
++ *
++ * @remove_tree_node:	if this is the last one of its tree_node, will the
++ * tree_node be freed ? If we are inserting stable node, this tree_node may
++ * be reused, so don't free it.
++ */
++static struct page *get_uksm_page(struct stable_node *stable_node,
++				 int unlink_rb, int remove_tree_node)
++{
++	struct page *page;
++	void *expected_mapping;
++	unsigned long kpfn;
++
++	expected_mapping = (void *)((unsigned long)stable_node |
++				    PAGE_MAPPING_KSM);
++again:
++	kpfn = READ_ONCE(stable_node->kpfn);
++	page = pfn_to_page(kpfn);
++
++	/*
++	 * page is computed from kpfn, so on most architectures reading
++	 * page->mapping is naturally ordered after reading node->kpfn,
++	 * but on Alpha we need to be more careful.
++	 */
++	smp_read_barrier_depends();
++
++	if (READ_ONCE(page->mapping) != expected_mapping)
++		goto stale;
++
++	/*
++	 * We cannot do anything with the page while its refcount is 0.
++	 * Usually 0 means free, or tail of a higher-order page: in which
++	 * case this node is no longer referenced, and should be freed;
++	 * however, it might mean that the page is under page_freeze_refs().
++	 * The __remove_mapping() case is easy, again the node is now stale;
++	 * but if page is swapcache in migrate_page_move_mapping(), it might
++	 * still be our page, in which case it's essential to keep the node.
++	 */
++	while (!get_page_unless_zero(page)) {
++		/*
++		 * Another check for page->mapping != expected_mapping would
++		 * work here too.  We have chosen the !PageSwapCache test to
++		 * optimize the common case, when the page is or is about to
++		 * be freed: PageSwapCache is cleared (under spin_lock_irq)
++		 * in the freeze_refs section of __remove_mapping(); but Anon
++		 * page->mapping reset to NULL later, in free_pages_prepare().
++		 */
++		if (!PageSwapCache(page))
++			goto stale;
++		cpu_relax();
++	}
++
++	if (READ_ONCE(page->mapping) != expected_mapping) {
++		put_page(page);
++		goto stale;
++	}
++
++	lock_page(page);
++	if (READ_ONCE(page->mapping) != expected_mapping) {
++		unlock_page(page);
++		put_page(page);
++		goto stale;
++	}
++	unlock_page(page);
++	return page;
++stale:
++	/*
++	 * We come here from above when page->mapping or !PageSwapCache
++	 * suggests that the node is stale; but it might be under migration.
++	 * We need smp_rmb(), matching the smp_wmb() in ksm_migrate_page(),
++	 * before checking whether node->kpfn has been changed.
++	 */
++	smp_rmb();
++	if (stable_node->kpfn != kpfn)
++		goto again;
++
++	remove_node_from_stable_tree(stable_node, unlink_rb, remove_tree_node);
++
++	return NULL;
++}
++
++/*
++ * Removing rmap_item from stable or unstable tree.
++ * This function will clean the information from the stable/unstable tree.
++ */
++static inline void remove_rmap_item_from_tree(struct rmap_item *rmap_item)
++{
++	if (rmap_item->address & STABLE_FLAG) {
++		struct stable_node *stable_node;
++		struct node_vma *node_vma;
++		struct page *page;
++
++		node_vma = rmap_item->head;
++		stable_node = node_vma->head;
++		page = get_uksm_page(stable_node, 1, 1);
++		if (!page)
++			goto out;
++
++		/*
++		 * page lock is needed because it's racing with
++		 * try_to_unmap_ksm(), etc.
++		 */
++		lock_page(page);
++		hlist_del(&rmap_item->hlist);
++
++		if (hlist_empty(&node_vma->rmap_hlist)) {
++			hlist_del(&node_vma->hlist);
++			free_node_vma(node_vma);
++		}
++		unlock_page(page);
++
++		put_page(page);
++		if (hlist_empty(&stable_node->hlist)) {
++			/* do NOT call remove_node_from_stable_tree() here,
++			 * it's possible for a forked rmap_item not in
++			 * stable tree while the in-tree rmap_items were
++			 * deleted.
++			 */
++			uksm_pages_shared--;
++		} else
++			uksm_pages_sharing--;
++
++
++		uksm_drop_anon_vma(rmap_item);
++	} else if (rmap_item->address & UNSTABLE_FLAG) {
++		if (rmap_item->hash_round == uksm_hash_round) {
++
++			rb_erase(&rmap_item->node,
++				 &rmap_item->tree_node->sub_root);
++			if (RB_EMPTY_ROOT(&rmap_item->tree_node->sub_root)) {
++				rb_erase(&rmap_item->tree_node->node,
++					 &root_unstable_tree);
++
++				free_tree_node(rmap_item->tree_node);
++			} else
++				rmap_item->tree_node->count--;
++		}
++		uksm_pages_unshared--;
++	}
++
++	rmap_item->address &= PAGE_MASK;
++	rmap_item->hash_max = 0;
++
++out:
++	cond_resched();		/* we're called from many long loops */
++}
++
++static inline int slot_in_uksm(struct vma_slot *slot)
++{
++	return list_empty(&slot->slot_list);
++}
++
++/*
++ * Test if the mm is exiting
++ */
++static inline bool uksm_test_exit(struct mm_struct *mm)
++{
++	return atomic_read(&mm->mm_users) == 0;
++}
++
++static inline unsigned long vma_pool_size(struct vma_slot *slot)
++{
++	return round_up(sizeof(struct rmap_list_entry) * slot->pages,
++			PAGE_SIZE) >> PAGE_SHIFT;
++}
++
++#define CAN_OVERFLOW_U64(x, delta) (U64_MAX - (x) < (delta))
++
++/* must be done with sem locked */
++static int slot_pool_alloc(struct vma_slot *slot)
++{
++	unsigned long pool_size;
++
++	if (slot->rmap_list_pool)
++		return 0;
++
++	pool_size = vma_pool_size(slot);
++	slot->rmap_list_pool = kcalloc(pool_size, sizeof(struct page *),
++				       GFP_KERNEL);
++	if (!slot->rmap_list_pool)
++		return -ENOMEM;
++
++	slot->pool_counts = kcalloc(pool_size, sizeof(unsigned int),
++				    GFP_KERNEL);
++	if (!slot->pool_counts) {
++		kfree(slot->rmap_list_pool);
++		return -ENOMEM;
++	}
++
++	slot->pool_size = pool_size;
++	BUG_ON(CAN_OVERFLOW_U64(uksm_pages_total, slot->pages));
++	slot->flags |= UKSM_SLOT_IN_UKSM;
++	uksm_pages_total += slot->pages;
++
++	return 0;
++}
++
++/*
++ * Called after vma is unlinked from its mm
++ */
++void uksm_remove_vma(struct vm_area_struct *vma)
++{
++	struct vma_slot *slot;
++
++	if (!vma->uksm_vma_slot)
++		return;
++
++	spin_lock(&vma_slot_list_lock);
++	slot = vma->uksm_vma_slot;
++	if (!slot)
++		goto out;
++
++	if (slot_in_uksm(slot)) {
++		/**
++		 * This slot has been added by ksmd, so move to the del list
++		 * waiting ksmd to free it.
++		 */
++		list_add_tail(&slot->slot_list, &vma_slot_del);
++	} else {
++		/**
++		 * It's still on new list. It's ok to free slot directly.
++		 */
++		list_del(&slot->slot_list);
++		free_vma_slot(slot);
++	}
++out:
++	vma->uksm_vma_slot = NULL;
++	spin_unlock(&vma_slot_list_lock);
++}
++
++/**
++ * Need to do two things:
++ * 1. check if slot was moved to del list
++ * 2. make sure the mmap_sem is manipulated under valid vma.
++ *
++ * My concern here is that in some cases, this may make
++ * vma_slot_list_lock() waiters to serialized further by some
++ * sem->wait_lock, can this really be expensive?
++ *
++ *
++ * @return
++ * 0: if successfully locked mmap_sem
++ * -ENOENT: this slot was moved to del list
++ * -EBUSY: vma lock failed
++ */
++static int try_down_read_slot_mmap_sem(struct vma_slot *slot)
++{
++	struct vm_area_struct *vma;
++	struct mm_struct *mm;
++	struct rw_semaphore *sem;
++
++	spin_lock(&vma_slot_list_lock);
++
++	/* the slot_list was removed and inited from new list, when it enters
++	 * uksm_list. If now it's not empty, then it must be moved to del list
++	 */
++	if (!slot_in_uksm(slot)) {
++		spin_unlock(&vma_slot_list_lock);
++		return -ENOENT;
++	}
++
++	BUG_ON(slot->pages != vma_pages(slot->vma));
++	/* Ok, vma still valid */
++	vma = slot->vma;
++	mm = vma->vm_mm;
++	sem = &mm->mmap_sem;
++
++	if (uksm_test_exit(mm)) {
++		spin_unlock(&vma_slot_list_lock);
++		return -ENOENT;
++	}
++
++	if (down_read_trylock(sem)) {
++		spin_unlock(&vma_slot_list_lock);
++		if (slot_pool_alloc(slot)) {
++			uksm_remove_vma(vma);
++			up_read(sem);
++			return -ENOENT;
++		}
++		return 0;
++	}
++
++	spin_unlock(&vma_slot_list_lock);
++	return -EBUSY;
++}
++
++static inline unsigned long
++vma_page_address(struct page *page, struct vm_area_struct *vma)
++{
++	pgoff_t pgoff = page->index;
++	unsigned long address;
++
++	address = vma->vm_start + ((pgoff - vma->vm_pgoff) << PAGE_SHIFT);
++	if (unlikely(address < vma->vm_start || address >= vma->vm_end)) {
++		/* page should be within @vma mapping range */
++		return -EFAULT;
++	}
++	return address;
++}
++
++
++/* return 0 on success with the item's mmap_sem locked */
++static inline int get_mergeable_page_lock_mmap(struct rmap_item *item)
++{
++	struct mm_struct *mm;
++	struct vma_slot *slot = item->slot;
++	int err = -EINVAL;
++
++	struct page *page;
++
++	/*
++	 * try_down_read_slot_mmap_sem() returns non-zero if the slot
++	 * has been removed by uksm_remove_vma().
++	 */
++	if (try_down_read_slot_mmap_sem(slot))
++		return -EBUSY;
++
++	mm = slot->vma->vm_mm;
++
++	if (uksm_test_exit(mm))
++		goto failout_up;
++
++	page = item->page;
++	rcu_read_lock();
++	if (!get_page_unless_zero(page)) {
++		rcu_read_unlock();
++		goto failout_up;
++	}
++
++	/* No need to consider huge page here. */
++	if (item->slot->vma->anon_vma != page_anon_vma(page) ||
++	    vma_page_address(page, item->slot->vma) != get_rmap_addr(item)) {
++		/*
++		 * TODO:
++		 * should we release this item becase of its stale page
++		 * mapping?
++		 */
++		put_page(page);
++		rcu_read_unlock();
++		goto failout_up;
++	}
++	rcu_read_unlock();
++	return 0;
++
++failout_up:
++	up_read(&mm->mmap_sem);
++	return err;
++}
++
++/*
++ * What kind of VMA is considered ?
++ */
++static inline int vma_can_enter(struct vm_area_struct *vma)
++{
++	return uksm_flags_can_scan(vma->vm_flags);
++}
++
++/*
++ * Called whenever a fresh new vma is created A new vma_slot.
++ * is created and inserted into a global list Must be called.
++ * after vma is inserted to its mm.
++ */
++void uksm_vma_add_new(struct vm_area_struct *vma)
++{
++	struct vma_slot *slot;
++
++	if (!vma_can_enter(vma)) {
++		vma->uksm_vma_slot = NULL;
++		return;
++	}
++
++	slot = alloc_vma_slot();
++	if (!slot) {
++		vma->uksm_vma_slot = NULL;
++		return;
++	}
++
++	vma->uksm_vma_slot = slot;
++	vma->vm_flags |= VM_MERGEABLE;
++	slot->vma = vma;
++	slot->mm = vma->vm_mm;
++	slot->ctime_j = jiffies;
++	slot->pages = vma_pages(vma);
++	spin_lock(&vma_slot_list_lock);
++	list_add_tail(&slot->slot_list, &vma_slot_new);
++	spin_unlock(&vma_slot_list_lock);
++}
++
++/*   32/3 < they < 32/2 */
++#define shiftl	8
++#define shiftr	12
++
++#define HASH_FROM_TO(from, to)			\
++for (index = from; index < to; index++) {	\
++	pos = random_nums[index];		\
++	hash += key[pos];			\
++	hash += (hash << shiftl);		\
++	hash ^= (hash >> shiftr);		\
++}
++
++
++#define HASH_FROM_DOWN_TO(from, to)		\
++for (index = from - 1; index >= to; index--) {	\
++	hash ^= (hash >> shiftr);		\
++	hash ^= (hash >> (shiftr*2));		\
++	hash -= (hash << shiftl);		\
++	hash += (hash << (shiftl*2));		\
++	pos = random_nums[index];		\
++	hash -= key[pos];			\
++}
++
++/*
++ * The main random sample hash function.
++ */
++static u32 random_sample_hash(void *addr, u32 hash_strength)
++{
++	u32 hash = 0xdeadbeef;
++	int index, pos, loop = hash_strength;
++	u32 *key = (u32 *)addr;
++
++	if (loop > HASH_STRENGTH_FULL)
++		loop = HASH_STRENGTH_FULL;
++
++	HASH_FROM_TO(0, loop);
++
++	if (hash_strength > HASH_STRENGTH_FULL) {
++		loop = hash_strength - HASH_STRENGTH_FULL;
++		HASH_FROM_TO(0, loop);
++	}
++
++	return hash;
++}
++
++
++/**
++ * It's used when hash strength is adjusted
++ *
++ * @addr The page's virtual address
++ * @from The original hash strength
++ * @to   The hash strength changed to
++ * @hash The hash value generated with "from" hash value
++ *
++ * return the hash value
++ */
++static u32 delta_hash(void *addr, int from, int to, u32 hash)
++{
++	u32 *key = (u32 *)addr;
++	int index, pos; /* make sure they are int type */
++
++	if (to > from) {
++		if (from >= HASH_STRENGTH_FULL) {
++			from -= HASH_STRENGTH_FULL;
++			to -= HASH_STRENGTH_FULL;
++			HASH_FROM_TO(from, to);
++		} else if (to <= HASH_STRENGTH_FULL) {
++			HASH_FROM_TO(from, to);
++		} else {
++			HASH_FROM_TO(from, HASH_STRENGTH_FULL);
++			HASH_FROM_TO(0, to - HASH_STRENGTH_FULL);
++		}
++	} else {
++		if (from <= HASH_STRENGTH_FULL) {
++			HASH_FROM_DOWN_TO(from, to);
++		} else if (to >= HASH_STRENGTH_FULL) {
++			from -= HASH_STRENGTH_FULL;
++			to -= HASH_STRENGTH_FULL;
++			HASH_FROM_DOWN_TO(from, to);
++		} else {
++			HASH_FROM_DOWN_TO(from - HASH_STRENGTH_FULL, 0);
++			HASH_FROM_DOWN_TO(HASH_STRENGTH_FULL, to);
++		}
++	}
++
++	return hash;
++}
++
++/**
++ *
++ * Called when: rshash_pos or rshash_neg is about to overflow or a scan round
++ * has finished.
++ *
++ * return 0 if no page has been scanned since last call, 1 otherwise.
++ */
++static inline int encode_benefit(void)
++{
++	u64 scanned_delta, pos_delta, neg_delta;
++	unsigned long base = benefit.base;
++
++	scanned_delta = uksm_pages_scanned - uksm_pages_scanned_last;
++
++	if (!scanned_delta)
++		return 0;
++
++	scanned_delta >>= base;
++	pos_delta = rshash_pos >> base;
++	neg_delta = rshash_neg >> base;
++
++	if (CAN_OVERFLOW_U64(benefit.pos, pos_delta) ||
++	    CAN_OVERFLOW_U64(benefit.neg, neg_delta) ||
++	    CAN_OVERFLOW_U64(benefit.scanned, scanned_delta)) {
++		benefit.scanned >>= 1;
++		benefit.neg >>= 1;
++		benefit.pos >>= 1;
++		benefit.base++;
++		scanned_delta >>= 1;
++		pos_delta >>= 1;
++		neg_delta >>= 1;
++	}
++
++	benefit.pos += pos_delta;
++	benefit.neg += neg_delta;
++	benefit.scanned += scanned_delta;
++
++	BUG_ON(!benefit.scanned);
++
++	rshash_pos = rshash_neg = 0;
++	uksm_pages_scanned_last = uksm_pages_scanned;
++
++	return 1;
++}
++
++static inline void reset_benefit(void)
++{
++	benefit.pos = 0;
++	benefit.neg = 0;
++	benefit.base = 0;
++	benefit.scanned = 0;
++}
++
++static inline void inc_rshash_pos(unsigned long delta)
++{
++	if (CAN_OVERFLOW_U64(rshash_pos, delta))
++		encode_benefit();
++
++	rshash_pos += delta;
++}
++
++static inline void inc_rshash_neg(unsigned long delta)
++{
++	if (CAN_OVERFLOW_U64(rshash_neg, delta))
++		encode_benefit();
++
++	rshash_neg += delta;
++}
++
++
++static inline u32 page_hash(struct page *page, unsigned long hash_strength,
++			    int cost_accounting)
++{
++	u32 val;
++	unsigned long delta;
++
++	void *addr = kmap_atomic(page);
++
++	val = random_sample_hash(addr, hash_strength);
++	kunmap_atomic(addr);
++
++	if (cost_accounting) {
++		if (hash_strength < HASH_STRENGTH_FULL)
++			delta = HASH_STRENGTH_FULL - hash_strength;
++		else
++			delta = 0;
++
++		inc_rshash_pos(delta);
++	}
++
++	return val;
++}
++
++static int memcmp_pages_with_cost(struct page *page1, struct page *page2,
++			int cost_accounting)
++{
++	char *addr1, *addr2;
++	int ret;
++
++	addr1 = kmap_atomic(page1);
++	addr2 = kmap_atomic(page2);
++	ret = memcmp(addr1, addr2, PAGE_SIZE);
++	kunmap_atomic(addr2);
++	kunmap_atomic(addr1);
++
++	if (cost_accounting)
++		inc_rshash_neg(memcmp_cost);
++
++	return ret;
++}
++
++static inline int pages_identical_with_cost(struct page *page1, struct page *page2)
++{
++	return !memcmp_pages_with_cost(page1, page2, 0);
++}
++
++static inline int is_page_full_zero(struct page *page)
++{
++	char *addr;
++	int ret;
++
++	addr = kmap_atomic(page);
++	ret = is_full_zero(addr, PAGE_SIZE);
++	kunmap_atomic(addr);
++
++	return ret;
++}
++
++static int write_protect_page(struct vm_area_struct *vma, struct page *page,
++			      pte_t *orig_pte, pte_t *old_pte)
++{
++	struct mm_struct *mm = vma->vm_mm;
++	struct page_vma_mapped_walk pvmw = {
++		.page = page,
++		.vma = vma,
++	};
++       struct mmu_notifier_range range;
++	int swapped;
++	int err = -EFAULT;
++
++	pvmw.address = page_address_in_vma(page, vma);
++	if (pvmw.address == -EFAULT)
++		goto out;
++
++	BUG_ON(PageTransCompound(page));
++
++        mmu_notifier_range_init(&range, MMU_NOTIFY_CLEAR, 0, vma, mm, pvmw.address,
++                                pvmw.address + PAGE_SIZE);
++	mmu_notifier_invalidate_range_start(&range);
++
++	if (!page_vma_mapped_walk(&pvmw))
++		goto out_mn;
++	if (WARN_ONCE(!pvmw.pte, "Unexpected PMD mapping?"))
++		goto out_unlock;
++
++	if (old_pte)
++		*old_pte = *pvmw.pte;
++
++	if (pte_write(*pvmw.pte) || pte_dirty(*pvmw.pte) ||
++	    (pte_protnone(*pvmw.pte) && pte_savedwrite(*pvmw.pte)) || mm_tlb_flush_pending(mm)) {
++		pte_t entry;
++
++		swapped = PageSwapCache(page);
++		flush_cache_page(vma, pvmw.address, page_to_pfn(page));
++		/*
++		 * Ok this is tricky, when get_user_pages_fast() run it doesn't
++		 * take any lock, therefore the check that we are going to make
++		 * with the pagecount against the mapcount is racey and
++		 * O_DIRECT can happen right after the check.
++		 * So we clear the pte and flush the tlb before the check
++		 * this assure us that no O_DIRECT can happen after the check
++		 * or in the middle of the check.
++		 */
++		entry = ptep_clear_flush_notify(vma, pvmw.address, pvmw.pte);
++		/*
++		 * Check that no O_DIRECT or similar I/O is in progress on the
++		 * page
++		 */
++		if (page_mapcount(page) + 1 + swapped != page_count(page)) {
++			set_pte_at(mm, pvmw.address, pvmw.pte, entry);
++			goto out_unlock;
++		}
++		if (pte_dirty(entry))
++			set_page_dirty(page);
++
++		if (pte_protnone(entry))
++			entry = pte_mkclean(pte_clear_savedwrite(entry));
++		else
++			entry = pte_mkclean(pte_wrprotect(entry));
++
++		set_pte_at_notify(mm, pvmw.address, pvmw.pte, entry);
++	}
++	*orig_pte = *pvmw.pte;
++	err = 0;
++
++out_unlock:
++	page_vma_mapped_walk_done(&pvmw);
++out_mn:
++	mmu_notifier_invalidate_range_end(&range);
++out:
++	return err;
++}
++
++#define MERGE_ERR_PGERR		1 /* the page is invalid cannot continue */
++#define MERGE_ERR_COLLI		2 /* there is a collision */
++#define MERGE_ERR_COLLI_MAX	3 /* collision at the max hash strength */
++#define MERGE_ERR_CHANGED	4 /* the page has changed since last hash */
++
++
++/**
++ * replace_page - replace page in vma by new ksm page
++ * @vma:      vma that holds the pte pointing to page
++ * @page:     the page we are replacing by kpage
++ * @kpage:    the ksm page we replace page by
++ * @orig_pte: the original value of the pte
++ *
++ * Returns 0 on success, MERGE_ERR_PGERR on failure.
++ */
++static int replace_page(struct vm_area_struct *vma, struct page *page,
++			struct page *kpage, pte_t orig_pte)
++{
++	struct mm_struct *mm = vma->vm_mm;
++       struct mmu_notifier_range range;
++	pgd_t *pgd;
++	p4d_t *p4d;
++	pud_t *pud;
++	pmd_t *pmd;
++	pte_t *ptep;
++	spinlock_t *ptl;
++	pte_t entry;
++
++	unsigned long addr;
++	int err = MERGE_ERR_PGERR;
++
++	addr = page_address_in_vma(page, vma);
++	if (addr == -EFAULT)
++		goto out;
++
++	pgd = pgd_offset(mm, addr);
++	if (!pgd_present(*pgd))
++		goto out;
++
++	p4d = p4d_offset(pgd, addr);
++	pud = pud_offset(p4d, addr);
++	if (!pud_present(*pud))
++		goto out;
++
++	pmd = pmd_offset(pud, addr);
++	BUG_ON(pmd_trans_huge(*pmd));
++	if (!pmd_present(*pmd))
++		goto out;
++
++        mmu_notifier_range_init(&range, MMU_NOTIFY_CLEAR, 0, vma, mm, addr,
++                                addr + PAGE_SIZE);
++	mmu_notifier_invalidate_range_start(&range);
++
++	ptep = pte_offset_map_lock(mm, pmd, addr, &ptl);
++	if (!pte_same(*ptep, orig_pte)) {
++		pte_unmap_unlock(ptep, ptl);
++		goto out_mn;
++	}
++
++	flush_cache_page(vma, addr, pte_pfn(*ptep));
++	ptep_clear_flush_notify(vma, addr, ptep);
++	entry = mk_pte(kpage, vma->vm_page_prot);
++
++	/* special treatment is needed for zero_page */
++	if ((page_to_pfn(kpage) == uksm_zero_pfn) ||
++				(page_to_pfn(kpage) == zero_pfn)) {
++		entry = pte_mkspecial(entry);
++		dec_mm_counter(mm, MM_ANONPAGES);
++		inc_zone_page_state(page, NR_UKSM_ZERO_PAGES);
++	} else {
++		get_page(kpage);
++		page_add_anon_rmap(kpage, vma, addr, false);
++	}
++
++	set_pte_at_notify(mm, addr, ptep, entry);
++
++	page_remove_rmap(page, false);
++	if (!page_mapped(page))
++		try_to_free_swap(page);
++	put_page(page);
++
++	pte_unmap_unlock(ptep, ptl);
++	err = 0;
++out_mn:
++	mmu_notifier_invalidate_range_end(&range);
++out:
++	return err;
++}
++
++
++/**
++ *  Fully hash a page with HASH_STRENGTH_MAX return a non-zero hash value. The
++ *  zero hash value at HASH_STRENGTH_MAX is used to indicated that its
++ *  hash_max member has not been calculated.
++ *
++ * @page The page needs to be hashed
++ * @hash_old The hash value calculated with current hash strength
++ *
++ * return the new hash value calculated at HASH_STRENGTH_MAX
++ */
++static inline u32 page_hash_max(struct page *page, u32 hash_old)
++{
++	u32 hash_max = 0;
++	void *addr;
++
++	addr = kmap_atomic(page);
++	hash_max = delta_hash(addr, hash_strength,
++			      HASH_STRENGTH_MAX, hash_old);
++
++	kunmap_atomic(addr);
++
++	if (!hash_max)
++		hash_max = 1;
++
++	inc_rshash_neg(HASH_STRENGTH_MAX - hash_strength);
++	return hash_max;
++}
++
++/*
++ * We compare the hash again, to ensure that it is really a hash collision
++ * instead of being caused by page write.
++ */
++static inline int check_collision(struct rmap_item *rmap_item,
++				  u32 hash)
++{
++	int err;
++	struct page *page = rmap_item->page;
++
++	/* if this rmap_item has already been hash_maxed, then the collision
++	 * must appears in the second-level rbtree search. In this case we check
++	 * if its hash_max value has been changed. Otherwise, the collision
++	 * happens in the first-level rbtree search, so we check against it's
++	 * current hash value.
++	 */
++	if (rmap_item->hash_max) {
++		inc_rshash_neg(memcmp_cost);
++		inc_rshash_neg(HASH_STRENGTH_MAX - hash_strength);
++
++		if (rmap_item->hash_max == page_hash_max(page, hash))
++			err = MERGE_ERR_COLLI;
++		else
++			err = MERGE_ERR_CHANGED;
++	} else {
++		inc_rshash_neg(memcmp_cost + hash_strength);
++
++		if (page_hash(page, hash_strength, 0) == hash)
++			err = MERGE_ERR_COLLI;
++		else
++			err = MERGE_ERR_CHANGED;
++	}
++
++	return err;
++}
++
++/**
++ * Try to merge a rmap_item.page with a kpage in stable node. kpage must
++ * already be a ksm page.
++ *
++ * @return 0 if the pages were merged, -EFAULT otherwise.
++ */
++static int try_to_merge_with_uksm_page(struct rmap_item *rmap_item,
++				      struct page *kpage, u32 hash)
++{
++	struct vm_area_struct *vma = rmap_item->slot->vma;
++	struct mm_struct *mm = vma->vm_mm;
++	pte_t orig_pte = __pte(0);
++	int err = MERGE_ERR_PGERR;
++	struct page *page;
++
++	if (uksm_test_exit(mm))
++		goto out;
++
++	page = rmap_item->page;
++
++	if (page == kpage) { /* ksm page forked */
++		err = 0;
++		goto out;
++	}
++
++	/*
++	 * We need the page lock to read a stable PageSwapCache in
++	 * write_protect_page().  We use trylock_page() instead of
++	 * lock_page() because we don't want to wait here - we
++	 * prefer to continue scanning and merging different pages,
++	 * then come back to this page when it is unlocked.
++	 */
++	if (!trylock_page(page))
++		goto out;
++
++	if (!PageAnon(page) || !PageKsm(kpage))
++		goto out_unlock;
++
++	if (PageTransCompound(page)) {
++		err = split_huge_page(page);
++		if (err)
++			goto out_unlock;
++	}
++
++	/*
++	 * If this anonymous page is mapped only here, its pte may need
++	 * to be write-protected.  If it's mapped elsewhere, all of its
++	 * ptes are necessarily already write-protected.  But in either
++	 * case, we need to lock and check page_count is not raised.
++	 */
++	if (write_protect_page(vma, page, &orig_pte, NULL) == 0) {
++		if (pages_identical_with_cost(page, kpage))
++			err = replace_page(vma, page, kpage, orig_pte);
++		else
++			err = check_collision(rmap_item, hash);
++	}
++
++	if ((vma->vm_flags & VM_LOCKED) && kpage && !err) {
++		munlock_vma_page(page);
++		if (!PageMlocked(kpage)) {
++			unlock_page(page);
++			lock_page(kpage);
++			mlock_vma_page(kpage);
++			page = kpage;		/* for final unlock */
++		}
++	}
++
++out_unlock:
++	unlock_page(page);
++out:
++	return err;
++}
++
++
++
++/**
++ * If two pages fail to merge in try_to_merge_two_pages, then we have a chance
++ * to restore a page mapping that has been changed in try_to_merge_two_pages.
++ *
++ * @return 0 on success.
++ */
++static int restore_uksm_page_pte(struct vm_area_struct *vma, unsigned long addr,
++			     pte_t orig_pte, pte_t wprt_pte)
++{
++	struct mm_struct *mm = vma->vm_mm;
++	pgd_t *pgd;
++	p4d_t *p4d;
++	pud_t *pud;
++	pmd_t *pmd;
++	pte_t *ptep;
++	spinlock_t *ptl;
++
++	int err = -EFAULT;
++
++	pgd = pgd_offset(mm, addr);
++	if (!pgd_present(*pgd))
++		goto out;
++
++	p4d = p4d_offset(pgd, addr);
++	pud = pud_offset(p4d, addr);
++	if (!pud_present(*pud))
++		goto out;
++
++	pmd = pmd_offset(pud, addr);
++	if (!pmd_present(*pmd))
++		goto out;
++
++	ptep = pte_offset_map_lock(mm, pmd, addr, &ptl);
++	if (!pte_same(*ptep, wprt_pte)) {
++		/* already copied, let it be */
++		pte_unmap_unlock(ptep, ptl);
++		goto out;
++	}
++
++	/*
++	 * Good boy, still here. When we still get the ksm page, it does not
++	 * return to the free page pool, there is no way that a pte was changed
++	 * to other page and gets back to this page. And remind that ksm page
++	 * do not reuse in do_wp_page(). So it's safe to restore the original
++	 * pte.
++	 */
++	flush_cache_page(vma, addr, pte_pfn(*ptep));
++	ptep_clear_flush_notify(vma, addr, ptep);
++	set_pte_at_notify(mm, addr, ptep, orig_pte);
++
++	pte_unmap_unlock(ptep, ptl);
++	err = 0;
++out:
++	return err;
++}
++
++/**
++ * try_to_merge_two_pages() - take two identical pages and prepare
++ * them to be merged into one page(rmap_item->page)
++ *
++ * @return 0 if we successfully merged two identical pages into
++ *         one ksm page. MERGE_ERR_COLLI if it's only a hash collision
++ *         search in rbtree. MERGE_ERR_CHANGED if rmap_item has been
++ *         changed since it's hashed. MERGE_ERR_PGERR otherwise.
++ *
++ */
++static int try_to_merge_two_pages(struct rmap_item *rmap_item,
++				  struct rmap_item *tree_rmap_item,
++				  u32 hash)
++{
++	pte_t orig_pte1 = __pte(0), orig_pte2 = __pte(0);
++	pte_t wprt_pte1 = __pte(0), wprt_pte2 = __pte(0);
++	struct vm_area_struct *vma1 = rmap_item->slot->vma;
++	struct vm_area_struct *vma2 = tree_rmap_item->slot->vma;
++	struct page *page = rmap_item->page;
++	struct page *tree_page = tree_rmap_item->page;
++	int err = MERGE_ERR_PGERR;
++	struct address_space *saved_mapping;
++
++
++	if (rmap_item->page == tree_rmap_item->page)
++		goto out;
++
++	if (!trylock_page(page))
++		goto out;
++
++	if (!PageAnon(page))
++		goto out_unlock;
++
++	if (PageTransCompound(page)) {
++		err = split_huge_page(page);
++		if (err)
++			goto out_unlock;
++	}
++
++	if (write_protect_page(vma1, page, &wprt_pte1, &orig_pte1) != 0) {
++		unlock_page(page);
++		goto out;
++	}
++
++	/*
++	 * While we hold page lock, upgrade page from
++	 * PageAnon+anon_vma to PageKsm+NULL stable_node:
++	 * stable_tree_insert() will update stable_node.
++	 */
++	saved_mapping = page->mapping;
++	set_page_stable_node(page, NULL);
++	mark_page_accessed(page);
++	if (!PageDirty(page))
++		SetPageDirty(page);
++
++	unlock_page(page);
++
++	if (!trylock_page(tree_page))
++		goto restore_out;
++
++	if (!PageAnon(tree_page)) {
++		unlock_page(tree_page);
++		goto restore_out;
++	}
++
++	if (PageTransCompound(tree_page)) {
++		err = split_huge_page(tree_page);
++		if (err) {
++			unlock_page(tree_page);
++			goto restore_out;
++		}
++	}
++
++	if (write_protect_page(vma2, tree_page, &wprt_pte2, &orig_pte2) != 0) {
++		unlock_page(tree_page);
++		goto restore_out;
++	}
++
++	if (pages_identical_with_cost(page, tree_page)) {
++		err = replace_page(vma2, tree_page, page, wprt_pte2);
++		if (err) {
++			unlock_page(tree_page);
++			goto restore_out;
++		}
++
++		if ((vma2->vm_flags & VM_LOCKED)) {
++			munlock_vma_page(tree_page);
++			if (!PageMlocked(page)) {
++				unlock_page(tree_page);
++				lock_page(page);
++				mlock_vma_page(page);
++				tree_page = page; /* for final unlock */
++			}
++		}
++
++		unlock_page(tree_page);
++
++		goto out; /* success */
++
++	} else {
++		if (tree_rmap_item->hash_max &&
++		    tree_rmap_item->hash_max == rmap_item->hash_max) {
++			err = MERGE_ERR_COLLI_MAX;
++		} else if (page_hash(page, hash_strength, 0) ==
++		    page_hash(tree_page, hash_strength, 0)) {
++			inc_rshash_neg(memcmp_cost + hash_strength * 2);
++			err = MERGE_ERR_COLLI;
++		} else {
++			err = MERGE_ERR_CHANGED;
++		}
++
++		unlock_page(tree_page);
++	}
++
++restore_out:
++	lock_page(page);
++	if (!restore_uksm_page_pte(vma1, get_rmap_addr(rmap_item),
++				  orig_pte1, wprt_pte1))
++		page->mapping = saved_mapping;
++
++out_unlock:
++	unlock_page(page);
++out:
++	return err;
++}
++
++static inline int hash_cmp(u32 new_val, u32 node_val)
++{
++	if (new_val > node_val)
++		return 1;
++	else if (new_val < node_val)
++		return -1;
++	else
++		return 0;
++}
++
++static inline u32 rmap_item_hash_max(struct rmap_item *item, u32 hash)
++{
++	u32 hash_max = item->hash_max;
++
++	if (!hash_max) {
++		hash_max = page_hash_max(item->page, hash);
++
++		item->hash_max = hash_max;
++	}
++
++	return hash_max;
++}
++
++
++
++/**
++ * stable_tree_search() - search the stable tree for a page
++ *
++ * @item:	the rmap_item we are comparing with
++ * @hash:	the hash value of this item->page already calculated
++ *
++ * @return	the page we have found, NULL otherwise. The page returned has
++ *			been gotten.
++ */
++static struct page *stable_tree_search(struct rmap_item *item, u32 hash)
++{
++	struct rb_node *node = root_stable_treep->rb_node;
++	struct tree_node *tree_node;
++	unsigned long hash_max;
++	struct page *page = item->page;
++	struct stable_node *stable_node;
++
++	stable_node = page_stable_node(page);
++	if (stable_node) {
++		/* ksm page forked, that is
++		 * if (PageKsm(page) && !in_stable_tree(rmap_item))
++		 * it's actually gotten once outside.
++		 */
++		get_page(page);
++		return page;
++	}
++
++	while (node) {
++		int cmp;
++
++		tree_node = rb_entry(node, struct tree_node, node);
++
++		cmp = hash_cmp(hash, tree_node->hash);
++
++		if (cmp < 0)
++			node = node->rb_left;
++		else if (cmp > 0)
++			node = node->rb_right;
++		else
++			break;
++	}
++
++	if (!node)
++		return NULL;
++
++	if (tree_node->count == 1) {
++		stable_node = rb_entry(tree_node->sub_root.rb_node,
++				       struct stable_node, node);
++		BUG_ON(!stable_node);
++
++		goto get_page_out;
++	}
++
++	/*
++	 * ok, we have to search the second
++	 * level subtree, hash the page to a
++	 * full strength.
++	 */
++	node = tree_node->sub_root.rb_node;
++	BUG_ON(!node);
++	hash_max = rmap_item_hash_max(item, hash);
++
++	while (node) {
++		int cmp;
++
++		stable_node = rb_entry(node, struct stable_node, node);
++
++		cmp = hash_cmp(hash_max, stable_node->hash_max);
++
++		if (cmp < 0)
++			node = node->rb_left;
++		else if (cmp > 0)
++			node = node->rb_right;
++		else
++			goto get_page_out;
++	}
++
++	return NULL;
++
++get_page_out:
++	page = get_uksm_page(stable_node, 1, 1);
++	return page;
++}
++
++static int try_merge_rmap_item(struct rmap_item *item,
++			       struct page *kpage,
++			       struct page *tree_page)
++{
++	struct vm_area_struct *vma = item->slot->vma;
++	struct page_vma_mapped_walk pvmw = {
++		.page = kpage,
++		.vma = vma,
++	};
++
++	pvmw.address = get_rmap_addr(item);
++	if (!page_vma_mapped_walk(&pvmw))
++		return 0;
++
++	if (pte_write(*pvmw.pte)) {
++		/* has changed, abort! */
++		page_vma_mapped_walk_done(&pvmw);
++		return 0;
++	}
++
++	get_page(tree_page);
++	page_add_anon_rmap(tree_page, vma, pvmw.address, false);
++
++	flush_cache_page(vma, pvmw.address, page_to_pfn(kpage));
++	ptep_clear_flush_notify(vma, pvmw.address, pvmw.pte);
++	set_pte_at_notify(vma->vm_mm, pvmw.address, pvmw.pte,
++			  mk_pte(tree_page, vma->vm_page_prot));
++
++	page_remove_rmap(kpage, false);
++	put_page(kpage);
++
++	page_vma_mapped_walk_done(&pvmw);
++
++	return 1;
++}
++
++/**
++ * try_to_merge_with_stable_page() - when two rmap_items need to be inserted
++ * into stable tree, the page was found to be identical to a stable ksm page,
++ * this is the last chance we can merge them into one.
++ *
++ * @item1:	the rmap_item holding the page which we wanted to insert
++ *		into stable tree.
++ * @item2:	the other rmap_item we found when unstable tree search
++ * @oldpage:	the page currently mapped by the two rmap_items
++ * @tree_page:	the page we found identical in stable tree node
++ * @success1:	return if item1 is successfully merged
++ * @success2:	return if item2 is successfully merged
++ */
++static void try_merge_with_stable(struct rmap_item *item1,
++				  struct rmap_item *item2,
++				  struct page **kpage,
++				  struct page *tree_page,
++				  int *success1, int *success2)
++{
++	struct vm_area_struct *vma1 = item1->slot->vma;
++	struct vm_area_struct *vma2 = item2->slot->vma;
++	*success1 = 0;
++	*success2 = 0;
++
++	if (unlikely(*kpage == tree_page)) {
++		/* I don't think this can really happen */
++		pr_warn("UKSM: unexpected condition detected in "
++			"%s -- *kpage == tree_page !\n", __func__);
++		*success1 = 1;
++		*success2 = 1;
++		return;
++	}
++
++	if (!PageAnon(*kpage) || !PageKsm(*kpage))
++		goto failed;
++
++	if (!trylock_page(tree_page))
++		goto failed;
++
++	/* If the oldpage is still ksm and still pointed
++	 * to in the right place, and still write protected,
++	 * we are confident it's not changed, no need to
++	 * memcmp anymore.
++	 * be ware, we cannot take nested pte locks,
++	 * deadlock risk.
++	 */
++	if (!try_merge_rmap_item(item1, *kpage, tree_page))
++		goto unlock_failed;
++
++	/* ok, then vma2, remind that pte1 already set */
++	if (!try_merge_rmap_item(item2, *kpage, tree_page))
++		goto success_1;
++
++	*success2 = 1;
++success_1:
++	*success1 = 1;
++
++
++	if ((*success1 && vma1->vm_flags & VM_LOCKED) ||
++	    (*success2 && vma2->vm_flags & VM_LOCKED)) {
++		munlock_vma_page(*kpage);
++		if (!PageMlocked(tree_page))
++			mlock_vma_page(tree_page);
++	}
++
++	/*
++	 * We do not need oldpage any more in the caller, so can break the lock
++	 * now.
++	 */
++	unlock_page(*kpage);
++	*kpage = tree_page; /* Get unlocked outside. */
++	return;
++
++unlock_failed:
++	unlock_page(tree_page);
++failed:
++	return;
++}
++
++static inline void stable_node_hash_max(struct stable_node *node,
++					 struct page *page, u32 hash)
++{
++	u32 hash_max = node->hash_max;
++
++	if (!hash_max) {
++		hash_max = page_hash_max(page, hash);
++		node->hash_max = hash_max;
++	}
++}
++
++static inline
++struct stable_node *new_stable_node(struct tree_node *tree_node,
++				    struct page *kpage, u32 hash_max)
++{
++	struct stable_node *new_stable_node;
++
++	new_stable_node = alloc_stable_node();
++	if (!new_stable_node)
++		return NULL;
++
++	new_stable_node->kpfn = page_to_pfn(kpage);
++	new_stable_node->hash_max = hash_max;
++	new_stable_node->tree_node = tree_node;
++	set_page_stable_node(kpage, new_stable_node);
++
++	return new_stable_node;
++}
++
++static inline
++struct stable_node *first_level_insert(struct tree_node *tree_node,
++				       struct rmap_item *rmap_item,
++				       struct rmap_item *tree_rmap_item,
++				       struct page **kpage, u32 hash,
++				       int *success1, int *success2)
++{
++	int cmp;
++	struct page *tree_page;
++	u32 hash_max = 0;
++	struct stable_node *stable_node, *new_snode;
++	struct rb_node *parent = NULL, **new;
++
++	/* this tree node contains no sub-tree yet */
++	stable_node = rb_entry(tree_node->sub_root.rb_node,
++			       struct stable_node, node);
++
++	tree_page = get_uksm_page(stable_node, 1, 0);
++	if (tree_page) {
++		cmp = memcmp_pages_with_cost(*kpage, tree_page, 1);
++		if (!cmp) {
++			try_merge_with_stable(rmap_item, tree_rmap_item, kpage,
++					      tree_page, success1, success2);
++			put_page(tree_page);
++			if (!*success1 && !*success2)
++				goto failed;
++
++			return stable_node;
++
++		} else {
++			/*
++			 * collision in first level try to create a subtree.
++			 * A new node need to be created.
++			 */
++			put_page(tree_page);
++
++			stable_node_hash_max(stable_node, tree_page,
++					     tree_node->hash);
++			hash_max = rmap_item_hash_max(rmap_item, hash);
++			cmp = hash_cmp(hash_max, stable_node->hash_max);
++
++			parent = &stable_node->node;
++			if (cmp < 0)
++				new = &parent->rb_left;
++			else if (cmp > 0)
++				new = &parent->rb_right;
++			else
++				goto failed;
++		}
++
++	} else {
++		/* the only stable_node deleted, we reuse its tree_node.
++		 */
++		parent = NULL;
++		new = &tree_node->sub_root.rb_node;
++	}
++
++	new_snode = new_stable_node(tree_node, *kpage, hash_max);
++	if (!new_snode)
++		goto failed;
++
++	rb_link_node(&new_snode->node, parent, new);
++	rb_insert_color(&new_snode->node, &tree_node->sub_root);
++	tree_node->count++;
++	*success1 = *success2 = 1;
++
++	return new_snode;
++
++failed:
++	return NULL;
++}
++
++static inline
++struct stable_node *stable_subtree_insert(struct tree_node *tree_node,
++					  struct rmap_item *rmap_item,
++					  struct rmap_item *tree_rmap_item,
++					  struct page **kpage, u32 hash,
++					  int *success1, int *success2)
++{
++	struct page *tree_page;
++	u32 hash_max;
++	struct stable_node *stable_node, *new_snode;
++	struct rb_node *parent, **new;
++
++research:
++	parent = NULL;
++	new = &tree_node->sub_root.rb_node;
++	BUG_ON(!*new);
++	hash_max = rmap_item_hash_max(rmap_item, hash);
++	while (*new) {
++		int cmp;
++
++		stable_node = rb_entry(*new, struct stable_node, node);
++
++		cmp = hash_cmp(hash_max, stable_node->hash_max);
++
++		if (cmp < 0) {
++			parent = *new;
++			new = &parent->rb_left;
++		} else if (cmp > 0) {
++			parent = *new;
++			new = &parent->rb_right;
++		} else {
++			tree_page = get_uksm_page(stable_node, 1, 0);
++			if (tree_page) {
++				cmp = memcmp_pages_with_cost(*kpage, tree_page, 1);
++				if (!cmp) {
++					try_merge_with_stable(rmap_item,
++						tree_rmap_item, kpage,
++						tree_page, success1, success2);
++
++					put_page(tree_page);
++					if (!*success1 && !*success2)
++						goto failed;
++					/*
++					 * successfully merged with a stable
++					 * node
++					 */
++					return stable_node;
++				} else {
++					put_page(tree_page);
++					goto failed;
++				}
++			} else {
++				/*
++				 * stable node may be deleted,
++				 * and subtree maybe
++				 * restructed, cannot
++				 * continue, research it.
++				 */
++				if (tree_node->count) {
++					goto research;
++				} else {
++					/* reuse the tree node*/
++					parent = NULL;
++					new = &tree_node->sub_root.rb_node;
++				}
++			}
++		}
++	}
++
++	new_snode = new_stable_node(tree_node, *kpage, hash_max);
++	if (!new_snode)
++		goto failed;
++
++	rb_link_node(&new_snode->node, parent, new);
++	rb_insert_color(&new_snode->node, &tree_node->sub_root);
++	tree_node->count++;
++	*success1 = *success2 = 1;
++
++	return new_snode;
++
++failed:
++	return NULL;
++}
++
++
++/**
++ * stable_tree_insert() - try to insert a merged page in unstable tree to
++ * the stable tree
++ *
++ * @kpage:		the page need to be inserted
++ * @hash:		the current hash of this page
++ * @rmap_item:		the rmap_item being scanned
++ * @tree_rmap_item:	the rmap_item found on unstable tree
++ * @success1:		return if rmap_item is merged
++ * @success2:		return if tree_rmap_item is merged
++ *
++ * @return		the stable_node on stable tree if at least one
++ *			rmap_item is inserted into stable tree, NULL
++ *			otherwise.
++ */
++static struct stable_node *
++stable_tree_insert(struct page **kpage, u32 hash,
++		   struct rmap_item *rmap_item,
++		   struct rmap_item *tree_rmap_item,
++		   int *success1, int *success2)
++{
++	struct rb_node **new = &root_stable_treep->rb_node;
++	struct rb_node *parent = NULL;
++	struct stable_node *stable_node;
++	struct tree_node *tree_node;
++	u32 hash_max = 0;
++
++	*success1 = *success2 = 0;
++
++	while (*new) {
++		int cmp;
++
++		tree_node = rb_entry(*new, struct tree_node, node);
++
++		cmp = hash_cmp(hash, tree_node->hash);
++
++		if (cmp < 0) {
++			parent = *new;
++			new = &parent->rb_left;
++		} else if (cmp > 0) {
++			parent = *new;
++			new = &parent->rb_right;
++		} else
++			break;
++	}
++
++	if (*new) {
++		if (tree_node->count == 1) {
++			stable_node = first_level_insert(tree_node, rmap_item,
++						tree_rmap_item, kpage,
++						hash, success1, success2);
++		} else {
++			stable_node = stable_subtree_insert(tree_node,
++					rmap_item, tree_rmap_item, kpage,
++					hash, success1, success2);
++		}
++	} else {
++
++		/* no tree node found */
++		tree_node = alloc_tree_node(stable_tree_node_listp);
++		if (!tree_node) {
++			stable_node = NULL;
++			goto out;
++		}
++
++		stable_node = new_stable_node(tree_node, *kpage, hash_max);
++		if (!stable_node) {
++			free_tree_node(tree_node);
++			goto out;
++		}
++
++		tree_node->hash = hash;
++		rb_link_node(&tree_node->node, parent, new);
++		rb_insert_color(&tree_node->node, root_stable_treep);
++		parent = NULL;
++		new = &tree_node->sub_root.rb_node;
++
++		rb_link_node(&stable_node->node, parent, new);
++		rb_insert_color(&stable_node->node, &tree_node->sub_root);
++		tree_node->count++;
++		*success1 = *success2 = 1;
++	}
++
++out:
++	return stable_node;
++}
++
++
++/**
++ * get_tree_rmap_item_page() - try to get the page and lock the mmap_sem
++ *
++ * @return	0 on success, -EBUSY if unable to lock the mmap_sem,
++ *		-EINVAL if the page mapping has been changed.
++ */
++static inline int get_tree_rmap_item_page(struct rmap_item *tree_rmap_item)
++{
++	int err;
++
++	err = get_mergeable_page_lock_mmap(tree_rmap_item);
++
++	if (err == -EINVAL) {
++		/* its page map has been changed, remove it */
++		remove_rmap_item_from_tree(tree_rmap_item);
++	}
++
++	/* The page is gotten and mmap_sem is locked now. */
++	return err;
++}
++
++
++/**
++ * unstable_tree_search_insert() - search an unstable tree rmap_item with the
++ * same hash value. Get its page and trylock the mmap_sem
++ */
++static inline
++struct rmap_item *unstable_tree_search_insert(struct rmap_item *rmap_item,
++					      u32 hash)
++
++{
++	struct rb_node **new = &root_unstable_tree.rb_node;
++	struct rb_node *parent = NULL;
++	struct tree_node *tree_node;
++	u32 hash_max;
++	struct rmap_item *tree_rmap_item;
++
++	while (*new) {
++		int cmp;
++
++		tree_node = rb_entry(*new, struct tree_node, node);
++
++		cmp = hash_cmp(hash, tree_node->hash);
++
++		if (cmp < 0) {
++			parent = *new;
++			new = &parent->rb_left;
++		} else if (cmp > 0) {
++			parent = *new;
++			new = &parent->rb_right;
++		} else
++			break;
++	}
++
++	if (*new) {
++		/* got the tree_node */
++		if (tree_node->count == 1) {
++			tree_rmap_item = rb_entry(tree_node->sub_root.rb_node,
++						  struct rmap_item, node);
++			BUG_ON(!tree_rmap_item);
++
++			goto get_page_out;
++		}
++
++		/* well, search the collision subtree */
++		new = &tree_node->sub_root.rb_node;
++		BUG_ON(!*new);
++		hash_max = rmap_item_hash_max(rmap_item, hash);
++
++		while (*new) {
++			int cmp;
++
++			tree_rmap_item = rb_entry(*new, struct rmap_item,
++						  node);
++
++			cmp = hash_cmp(hash_max, tree_rmap_item->hash_max);
++			parent = *new;
++			if (cmp < 0)
++				new = &parent->rb_left;
++			else if (cmp > 0)
++				new = &parent->rb_right;
++			else
++				goto get_page_out;
++		}
++	} else {
++		/* alloc a new tree_node */
++		tree_node = alloc_tree_node(&unstable_tree_node_list);
++		if (!tree_node)
++			return NULL;
++
++		tree_node->hash = hash;
++		rb_link_node(&tree_node->node, parent, new);
++		rb_insert_color(&tree_node->node, &root_unstable_tree);
++		parent = NULL;
++		new = &tree_node->sub_root.rb_node;
++	}
++
++	/* did not found even in sub-tree */
++	rmap_item->tree_node = tree_node;
++	rmap_item->address |= UNSTABLE_FLAG;
++	rmap_item->hash_round = uksm_hash_round;
++	rb_link_node(&rmap_item->node, parent, new);
++	rb_insert_color(&rmap_item->node, &tree_node->sub_root);
++
++	uksm_pages_unshared++;
++	return NULL;
++
++get_page_out:
++	if (tree_rmap_item->page == rmap_item->page)
++		return NULL;
++
++	if (get_tree_rmap_item_page(tree_rmap_item))
++		return NULL;
++
++	return tree_rmap_item;
++}
++
++static void hold_anon_vma(struct rmap_item *rmap_item,
++			  struct anon_vma *anon_vma)
++{
++	rmap_item->anon_vma = anon_vma;
++	get_anon_vma(anon_vma);
++}
++
++
++/**
++ * stable_tree_append() - append a rmap_item to a stable node. Deduplication
++ * ratio statistics is done in this function.
++ *
++ */
++static void stable_tree_append(struct rmap_item *rmap_item,
++			       struct stable_node *stable_node, int logdedup)
++{
++	struct node_vma *node_vma = NULL, *new_node_vma, *node_vma_cont = NULL;
++	unsigned long key = (unsigned long)rmap_item->slot;
++	unsigned long factor = rmap_item->slot->rung->step;
++
++	BUG_ON(!stable_node);
++	rmap_item->address |= STABLE_FLAG;
++
++	if (hlist_empty(&stable_node->hlist)) {
++		uksm_pages_shared++;
++		goto node_vma_new;
++	} else {
++		uksm_pages_sharing++;
++	}
++
++	hlist_for_each_entry(node_vma, &stable_node->hlist, hlist) {
++		if (node_vma->key >= key)
++			break;
++
++		if (logdedup) {
++			node_vma->slot->pages_bemerged += factor;
++			if (list_empty(&node_vma->slot->dedup_list))
++				list_add(&node_vma->slot->dedup_list,
++					 &vma_slot_dedup);
++		}
++	}
++
++	if (node_vma) {
++		if (node_vma->key == key) {
++			node_vma_cont = hlist_entry_safe(node_vma->hlist.next, struct node_vma, hlist);
++			goto node_vma_ok;
++		} else if (node_vma->key > key) {
++			node_vma_cont = node_vma;
++		}
++	}
++
++node_vma_new:
++	/* no same vma already in node, alloc a new node_vma */
++	new_node_vma = alloc_node_vma();
++	BUG_ON(!new_node_vma);
++	new_node_vma->head = stable_node;
++	new_node_vma->slot = rmap_item->slot;
++
++	if (!node_vma) {
++		hlist_add_head(&new_node_vma->hlist, &stable_node->hlist);
++	} else if (node_vma->key != key) {
++		if (node_vma->key < key)
++			hlist_add_behind(&new_node_vma->hlist, &node_vma->hlist);
++		else {
++			hlist_add_before(&new_node_vma->hlist,
++					 &node_vma->hlist);
++		}
++
++	}
++	node_vma = new_node_vma;
++
++node_vma_ok: /* ok, ready to add to the list */
++	rmap_item->head = node_vma;
++	hlist_add_head(&rmap_item->hlist, &node_vma->rmap_hlist);
++	hold_anon_vma(rmap_item, rmap_item->slot->vma->anon_vma);
++	if (logdedup) {
++		rmap_item->slot->pages_merged++;
++		if (node_vma_cont) {
++			node_vma = node_vma_cont;
++			hlist_for_each_entry_continue(node_vma, hlist) {
++				node_vma->slot->pages_bemerged += factor;
++				if (list_empty(&node_vma->slot->dedup_list))
++					list_add(&node_vma->slot->dedup_list,
++						 &vma_slot_dedup);
++			}
++		}
++	}
++}
++
++/*
++ * We use break_ksm to break COW on a ksm page: it's a stripped down
++ *
++ *	if (get_user_pages(addr, 1, 1, 1, &page, NULL) == 1)
++ *		put_page(page);
++ *
++ * but taking great care only to touch a ksm page, in a VM_MERGEABLE vma,
++ * in case the application has unmapped and remapped mm,addr meanwhile.
++ * Could a ksm page appear anywhere else?  Actually yes, in a VM_PFNMAP
++ * mmap of /dev/mem or /dev/kmem, where we would not want to touch it.
++ */
++static int break_ksm(struct vm_area_struct *vma, unsigned long addr)
++{
++	struct page *page;
++	int ret = 0;
++
++	do {
++		cond_resched();
++		page = follow_page(vma, addr, FOLL_GET | FOLL_MIGRATION | FOLL_REMOTE);
++		if (IS_ERR_OR_NULL(page))
++			break;
++		if (PageKsm(page)) {
++			ret = handle_mm_fault(vma, addr,
++					      FAULT_FLAG_WRITE | FAULT_FLAG_REMOTE);
++		} else
++			ret = VM_FAULT_WRITE;
++		put_page(page);
++	} while (!(ret & (VM_FAULT_WRITE | VM_FAULT_SIGBUS | VM_FAULT_SIGSEGV | VM_FAULT_OOM)));
++	/*
++	 * We must loop because handle_mm_fault() may back out if there's
++	 * any difficulty e.g. if pte accessed bit gets updated concurrently.
++	 *
++	 * VM_FAULT_WRITE is what we have been hoping for: it indicates that
++	 * COW has been broken, even if the vma does not permit VM_WRITE;
++	 * but note that a concurrent fault might break PageKsm for us.
++	 *
++	 * VM_FAULT_SIGBUS could occur if we race with truncation of the
++	 * backing file, which also invalidates anonymous pages: that's
++	 * okay, that truncation will have unmapped the PageKsm for us.
++	 *
++	 * VM_FAULT_OOM: at the time of writing (late July 2009), setting
++	 * aside mem_cgroup limits, VM_FAULT_OOM would only be set if the
++	 * current task has TIF_MEMDIE set, and will be OOM killed on return
++	 * to user; and ksmd, having no mm, would never be chosen for that.
++	 *
++	 * But if the mm is in a limited mem_cgroup, then the fault may fail
++	 * with VM_FAULT_OOM even if the current task is not TIF_MEMDIE; and
++	 * even ksmd can fail in this way - though it's usually breaking ksm
++	 * just to undo a merge it made a moment before, so unlikely to oom.
++	 *
++	 * That's a pity: we might therefore have more kernel pages allocated
++	 * than we're counting as nodes in the stable tree; but uksm_do_scan
++	 * will retry to break_cow on each pass, so should recover the page
++	 * in due course.  The important thing is to not let VM_MERGEABLE
++	 * be cleared while any such pages might remain in the area.
++	 */
++	return (ret & VM_FAULT_OOM) ? -ENOMEM : 0;
++}
++
++static void break_cow(struct rmap_item *rmap_item)
++{
++	struct vm_area_struct *vma = rmap_item->slot->vma;
++	struct mm_struct *mm = vma->vm_mm;
++	unsigned long addr = get_rmap_addr(rmap_item);
++
++	if (uksm_test_exit(mm))
++		goto out;
++
++	break_ksm(vma, addr);
++out:
++	return;
++}
++
++/*
++ * Though it's very tempting to unmerge in_stable_tree(rmap_item)s rather
++ * than check every pte of a given vma, the locking doesn't quite work for
++ * that - an rmap_item is assigned to the stable tree after inserting ksm
++ * page and upping mmap_sem.  Nor does it fit with the way we skip dup'ing
++ * rmap_items from parent to child at fork time (so as not to waste time
++ * if exit comes before the next scan reaches it).
++ *
++ * Similarly, although we'd like to remove rmap_items (so updating counts
++ * and freeing memory) when unmerging an area, it's easier to leave that
++ * to the next pass of ksmd - consider, for example, how ksmd might be
++ * in cmp_and_merge_page on one of the rmap_items we would be removing.
++ */
++inline int unmerge_uksm_pages(struct vm_area_struct *vma,
++		      unsigned long start, unsigned long end)
++{
++	unsigned long addr;
++	int err = 0;
++
++	for (addr = start; addr < end && !err; addr += PAGE_SIZE) {
++		if (uksm_test_exit(vma->vm_mm))
++			break;
++		if (signal_pending(current))
++			err = -ERESTARTSYS;
++		else
++			err = break_ksm(vma, addr);
++	}
++	return err;
++}
++
++static inline void inc_uksm_pages_scanned(void)
++{
++	u64 delta;
++
++
++	if (uksm_pages_scanned == U64_MAX) {
++		encode_benefit();
++
++		delta = uksm_pages_scanned >> pages_scanned_base;
++
++		if (CAN_OVERFLOW_U64(pages_scanned_stored, delta)) {
++			pages_scanned_stored >>= 1;
++			delta >>= 1;
++			pages_scanned_base++;
++		}
++
++		pages_scanned_stored += delta;
++
++		uksm_pages_scanned = uksm_pages_scanned_last = 0;
++	}
++
++	uksm_pages_scanned++;
++}
++
++static inline int find_zero_page_hash(int strength, u32 hash)
++{
++	return (zero_hash_table[strength] == hash);
++}
++
++static
++int cmp_and_merge_zero_page(struct vm_area_struct *vma, struct page *page)
++{
++	struct page *zero_page = empty_uksm_zero_page;
++	struct mm_struct *mm = vma->vm_mm;
++	pte_t orig_pte = __pte(0);
++	int err = -EFAULT;
++
++	if (uksm_test_exit(mm))
++		goto out;
++
++	if (!trylock_page(page))
++		goto out;
++
++	if (!PageAnon(page))
++		goto out_unlock;
++
++	if (PageTransCompound(page)) {
++		err = split_huge_page(page);
++		if (err)
++			goto out_unlock;
++	}
++
++	if (write_protect_page(vma, page, &orig_pte, 0) == 0) {
++		if (is_page_full_zero(page))
++			err = replace_page(vma, page, zero_page, orig_pte);
++	}
++
++out_unlock:
++	unlock_page(page);
++out:
++	return err;
++}
++
++/*
++ * cmp_and_merge_page() - first see if page can be merged into the stable
++ * tree; if not, compare hash to previous and if it's the same, see if page
++ * can be inserted into the unstable tr