[scripts] Add missing iced-x86 dependency for print_got.py

Update init to use the new xoroshiro RNG to create a random load address for dynamic executables.
Also pass an address past the end of the loader in the loader_arg to use when loading dynamic
dependencies.

.clang-format

@@ -1,21 +0,0 @@
----
-Language:        Cpp
-BasedOnStyle:    LLVM
-
-ColumnLimit:     100
-IndentWidth:     4
-TabWidth:        4
-UseTab:          Always
-
-AccessModifierOffset: -4
-
-AlignEscapedNewlinesLeft: true
-
-AllowShortIfStatementsOnASingleLine: true
-AllowShortLoopsOnASingleLine: true
-
-AlwaysBreakAfterReturnType: AllDefinitions
-AlwaysBreakBeforeMultilineStrings: true
-AlwaysBreakTemplateDeclarations: true
-
-BreakBeforeBraces: Linux

.gitignore

@@ -1,8 +1,9 @@
+.cache
 .lock*
 /build*
 *.bak
 tags
-jsix.log
+*.log
 *.out
 *.o
 *.a
@@ -10,3 +11,8 @@ sysroot
 .gdb_history
 .peru
 __pycache__
+/venv
+compile_commands.json
+buddy_allocs.txt
+frame_allocs.txt
+heap_allocs.txt

.vscode/c_cpp_properties.json

@@ -0,0 +1,19 @@
+{
+    "configurations": [
+        {
+            "name": "Linux",
+            "includePath": [
+                "${workspaceFolder}/src/libraries/**",
+                "${workspaceFolder}/build/**",
+                "${workspaceFolder}/sysroot/include"
+            ],
+            "defines": [],
+            "compilerPath": "/usr/bin/clang",
+            "cStandard": "c17",
+            "cppStandard": "c++17",
+            "intelliSenseMode": "linux-clang-x64",
+            "compileCommands": "${workspaceFolder}/compile_commands.json"
+        }
+    ],
+    "version": 4
+}

.vscode/launch.json

@@ -0,0 +1,32 @@
+{
+    // Use IntelliSense to learn about possible attributes.
+    // Hover to view descriptions of existing attributes.
+    // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
+    "version": "0.2.0",
+    "configurations": [
+        {
+            "name": "QEMU Debug Server",
+            "type": "cppdbg",
+            "request": "launch",
+            "program": "${workspaceFolder}/build/jsix.elf",
+            "args": [],
+            "cwd": "${workspaceFolder}",
+            "logging": {
+                //"engineLogging": true,
+                "programOutput": true
+            },  
+            "stopAtConnect": true,
+            "stopAtEntry": false,
+
+            "setupCommands": [
+                {"text": "dashboard -enabled off", "ignoreFailures": true}
+            ],
+
+            "MIMode": "gdb",
+            "miDebuggerServerAddress": "localhost:1234",
+            "debugServerPath": "${workspaceFolder}/qemu.sh",
+            "debugServerArgs": "--debug --no-build",
+            "serverLaunchTimeout": 5000,
+        }
+    ]
+}

.vscode/tasks.json

@@ -1,28 +1,41 @@
 {
     // See https://go.microsoft.com/fwlink/?LinkId=733558
     // for the documentation about the tasks.json format
-    "version": "0.1.0",
+    "version": "2.0.0",
     "tasks": [
         {
-            "taskName": "make",
+            "label": "Ninja",
-            "command": "make.bat",
+            "type": "shell",
-            "isBuildCommand": true
+            "command": "source ${workspaceFolder}/venv/bin/activate.fish; ninja",
+            "detail": "Build the project",
+            "options": {
+                "cwd": "${workspaceFolder}/build"
+            },
+            "group": {
+                "kind": "build",
+                "isDefault": true
+            },
+            "problemMatcher": []
         },
         {
-            "taskName": "clean",
+            "label": "Run QEMU",
-            "command": "make.bat",
+            "command": "./qemu.sh",
-            "args": [ "clean" ],
+            "args": [ "--no-build", "--kvm" ],
-            "isShellCommand": true
+            "dependsOn": ["Ninja"],
+            "options": {
+                "cwd": "${workspaceFolder}"
+            },
         },
         {
-            "taskName": "qemu (windowed)",
+            "label": "clean",
-            "command": "qemu-win.bat",
+            "command": "${workspaceFolder}/venv/bin/ninja",
-            "showOutput": "never",
+            "options": {
-            "isTestCommand": true
+                "cwd": "${workspaceFolder}/build"
-        },
+            },
-        {
+            "args": [
-            "taskName": "qemu",
+                "-t",
-            "command": "qemu.bat"
+                "clean"
+            ]
         }
     ]
 }

NOTES.md

@@ -1,38 +0,0 @@
-# Design / WIP notes
-
-## TODO
-
-- Paging manager
-  - Copy-on-write pages
-  - Better page-allocation model?
-- Allow for more than one IOAPIC in ACPI module
-  - The objects get created, but GSI lookup only uses the one at index 0
-- mark kernel memory pages global
-- Serial out based on circular/bip biffer and interrupts, not spinning on
-  `write_ready()`
-- Split out more code into kutil for testing
-- AHCI / MSI interrupts on Vbox break?
-- FXSAVE to save XMM registers.
-  - optimization using #NM (0x7) to detect SSE usage
-- Clean up of process memory maps
-- Better stack tracer
-- Bootloader rewrite
-  - C++ and sharing library code for ELF, initrd, etc
-  - Parse initrd and pre-load certain ELF images, eg the process loader process?
-  - Do initial memory bootstrap?
-- Calling global ctors
-- Device Tree
-  - Actual serial driver
-  - Disk driver
-    - File system
-- Multiprocessing
-  - Fast syscalls using syscall/sysret
-
-### Build
-
-- Clean up build generator and its templates
-  - More robust objects to represent modules & targets to pass to templates
-  - Read project setup from a simple JSON/TOML/etc
-  - Better compartmentalizing when doing template inheritance
-- Move to LLD as sysroot linker
-

README.md

@@ -3,13 +3,14 @@
 # The jsix operating system
 
 **jsix** is a custom multi-core x64 operating system that I am building from
-scratch. It's far from finished, or even being usable - see the *Status and
+scratch. It's far from finished, or even being usable (see the *Status and
-Roadmap* section, below.
+Roadmap* section, below) but all currently-planned major kernel features are
+now implemented to at least a passable level.
 
 The design goals of the project are:
 
 * Modernity - I'm not interested in designing for legacy systems, or running on
-  all hardware out there. My target is only 64 bit architecutres, and modern
+  all hardware out there. My target is only 64 bit architectures, and modern
   commodity hardware. Currently that means x64 systems with Nehalem or newer
   CPUs and UEFI firmware. (See [this list][cpu_features] for the currently
   required CPU features.) Eventually I'd like to work on an AArch64 port,
@@ -58,12 +59,18 @@ Remaining to do:
 
 - TLB shootdowns
 - Page swapping
+- Large / huge page support
 
 _Physical page allocation: Sufficient._ The current physical page allocator
-implementation suses a group of block representing up-to-1GiB areas of usable
+implementation uses a group of blocks representing up-to-1GiB areas of usable
 memory as defined by the bootloader. Each block has a three-level bitmap
 denoting free/used pages.
 
+Future work:
+
+- Align blocks so that their first page is 1GiB-aligned, making finding free
+  large/huge pages easier.
+
 #### Multitasking
 
 _Sufficient._ The global scheduler object keeps separate ready/blocked lists
@@ -71,72 +78,104 @@ per core. Cores periodically attempt to balance load via work stealing.
 
 User-space tasks are able to create threads as well as other processes.
 
-Several kernel-only tasks exist, though I'm trying to reduce that. Eventually
-only the timekeeping task should be a separate kernel-only thread.
-
 #### API
 
-_In progress._ User-space tasks are able to make syscalls to the kernel via
+_Syscalls: Sufficient._ User-space tasks are able to make syscalls to the
-fast SYSCALL/SYSRET instructions.
+kernel via fast SYSCALL/SYSRET instructions. Syscalls made via `libj6` look to
+both the callee and the caller like standard SysV ABI function calls. The
+implementations are wrapped in generated wrapper functions which validate the
+request, check capabilities, and find the appropriate kernel objects or handles
+before calling the implementation functions.
 
-Major tasks still to do:
+_IPC: Working, needs optimization._ The current IPC primitives are:
 
-- The process initialization protocol needs to be re-built entirely.
+- _Mailboxes_: endpoints for asynchronously-delivered small messages. Currently
-- Processes' handles to kernel objects need the ability to check capabilities
+  these messages are double-copied - once from the caller into a kernel buffer,
+  and once from the kernel to the receiver. This works and is not a major cause
+  of slowdown, but will need to be optimized in the future.
+- _Channels_: endpoints for asynchronous uni-directional streams of bytes.
+  Currently these also suffer from a double-copy problem, and should probably
+  be replaced eventually by userspace shared memory communication.
+- _Events_: objects that can be signalled to send asynchronous notifications to
+  waiting threads.
 
 #### Hardware Support
 
-  * Framebuffer driver: _In progress._ Currently on machines with a video
+- Framebuffer driver: _In progress._ Currently on machines with a video
-	device accessible by UEFI, jsix starts a user-space framebuffer driver that
+  device accessible by UEFI, jsix starts a user-space framebuffer driver that
-	only prints out kernel logs.
+  only prints out kernel logs.
-  * Serial driver: _To do._ Machines without a video device should have a
+- Serial driver: _In progress._ The current UART currently only exposes COM1
-	user-space log output task like the framebuffer driver, but currently this
+  as an output channel, but no input or other serial ports are exposed.
-	is done inside the kernel.
+- USB driver: _To do_
-  * USB driver: _To do_
+- AHCI (SATA) driver: _To do_
-  * AHCI (SATA) driver: _To do_
 
 ## Building
 
-jsix uses the [Ninja][] build tool, and generates the build files for it with a
+jsix uses the [Ninja][] build tool, and generates the build files for it with
-custom tool called [Bonnibel][]. Bonnibel can be installed with [Cargo][], or
+the `configure` script. The build also relies on a custom toolchain sysroot, which can be
-downloaded as a prebuilt binary from its Github repository.
+downloaded or built using the scripts in [jsix-os/toolchain][].
 
-[Ninja]:    https://ninja-build.org
+[Ninja]:             https://ninja-build.org
-[Bonnibel]: https://github.com/justinian/bonnibel_rs
+[jsix-os/toolchain]: https://github.com/jsix-os/toolchain
-[Cargo]:    https://crates.io/crates/bonnibel
 
-Requrirements:
+Other build dependencies:
 
-* bonnibel
+* [clang][]: the C/C++ compiler
-* ninja
+* [nasm][]: the assembler
-* clang
+* [lld][]: the linker
-* nasm
+* [mtools][]: for creating the FAT image
-* mtools
-* curl for downloading the toolchain
 
-### Setting up the cross toolchain
+[clang]:    https://clang.llvm.org
+[nasm]:     https://www.nasm.us
+[lld]:      https://lld.llvm.org
+[mtools]:   https://www.gnu.org/software/mtools/
 
-Running `pb sync` will download and unpack the toolchain into `sysroot`. 
+The `configure` script has some Python dependencies - these can be installed via
+`pip`, though doing so in a python virtual environment is recommended.
+Installing via `pip` will also install `ninja`.
 
-#### Compiling the toolchain yourself
+A Debian 11 (Bullseye) system can be configured with the necessary build
+dependencies by running the following commands from the jsix repository root:
 
-If you have `clang` and `curl` installed, runing the `scripts/build_sysroot.sh`
+```bash
-script will download and build a LLVM toolchain configured for building jsix
+sudo apt install clang lld nasm mtools python3-pip python3-venv
-host binaries.
+python3 -m venv ./venv
+source venv/bin/activate
+pip install -r requirements.txt
+peru sync
+```
+
+### Setting up the sysroot
+
+Build or download the toolchain sysroot as mentioned above with
+[jsix-os/toolchain][], and symlink the built toolchain directory as `sysroot`
+at the root of this project.
+
+```bash
+# Example if both the toolchain and this project are cloned under ~/src
+ln -s ~/src/toolchain/toolchains/llvm-13 ~/src/jsix/sysroot
+```
 
 ### Building and running jsix
 
-Once the toolchain has been set up, running Bonnibel's `pb init` command will
+Once the toolchain has been set up, running the `./configure` script (see
-set up the build configuration, and `pb build` will actually run the build.  If
+`./configure --help` for available options) will set up the build configuration,
-you have `qemu-system-x86_64` installed, the `qemu.sh` script will to run jsix
+and `ninja -C build` (or wherever you put the build directory) will actually run
-in QEMU `-nographic` mode.
+the build. If you have `qemu-system-x86_64` installed, the `qemu.sh` script will
+to run jsix in QEMU `-nographic` mode.
 
-I personally run this either from a real debian amd64 testing/buster machine or
+I personally run this either from a real debian amd64 bullseye machine or
-a windows WSL debian testing/buster installation. The following should be
+a windows WSL debian bullseye installation. Your mileage may vary with other
-enough to set up such a system to build the kernel:
+setups and distros.
 
-    sudo apt install qemu-system-x86 nasm clang-10 mtools curl ninja-build
+### Running the test suite
-    sudo update-alternatives /usr/bin/clang clang /usr/bin/clang-10 1000
-    sudo update-alternatives /usr/bin/clang++ clang++ /usr/bin/clang++-10 1000
-	curl -L -o pb https://github.com/justinian/bonnibel_rs/releases/download/v2.3.0/pb-linux-amd64 && chmod a+x pb
 
+jsix now has the `test_runner` userspace program that runs various automated
+tests. It is not included in the default build, but if you use the `test.yml`
+manifest it will be built, and can be run with the `test.sh` script or the
+`qemu.sh` script.
+
+```bash
+./configure --manifest=assets/manifests/test.yml
+if ./test.sh; then echo "All tests passed!"; else echo "Failed."; fi
+```

assets/build/config.debug.yaml

@@ -0,0 +1,9 @@
+---
+ccflags: [
+  "-g3",
+  "-ggdb",
+]
+
+ldflags: [
+  "-g",
+]

assets/build/global.yaml

@@ -0,0 +1,39 @@
+---
+cc: "${source_root}/sysroot/bin/clang"
+cxx: "${source_root}/sysroot/bin/clang++"
+ld: "${source_root}/sysroot/bin/ld.lld"
+ar: ar
+nasm: nasm
+objcopy: objcopy
+
+ccflags: [
+  "-I${source_root}/src/include",
+  "-fcolor-diagnostics",
+  "-U__STDCPP_THREADS__",
+  "-D_LIBCPP_HAS_NO_THREADS",
+  "-D__jsix_config=${build_config}",
+  "-D__jsix_config_${build_config}",
+  "-DVERSION_MAJOR=${version_major}",
+  "-DVERSION_MINOR=${version_minor}",
+  "-DVERSION_PATCH=${version_patch}",
+  "-DVERSION_GITSHA=0x${version_sha}",
+  '-DGIT_VERSION=\"${version_major}.${version_minor}.${version_patch}+${version_sha}\"',
+  '-DGIT_VERSION_WIDE=L\"${version_major}.${version_minor}.${version_patch}+${version_sha}\"',
+
+  "-Wformat=2", "-Winit-self", "-Winline", "-Wmissing-format-attribute",
+  "-Wmissing-include-dirs", "-Wswitch", "-Wundef", "-Wdisabled-optimization",
+  "-Wpointer-arith", "-Wno-attributes", "-Wno-sign-compare", "-Wno-multichar",
+  "-Wno-div-by-zero", "-Wno-endif-labels", "-Wno-pragmas", "-Wno-format-extra-args",
+  "-Wno-unused-result", "-Wno-deprecated-declarations", "-Wno-unused-function",
+  "-Wno-address-of-packed-member", "-Wno-invalid-offsetof", "-Wno-format-nonliteral",
+  "-Werror" ]
+
+asflags: [
+  "-DVERSION_MAJOR=${version_major}",
+  "-DVERSION_MINOR=${version_minor}",
+  "-DVERSION_PATCH=${version_patch}",
+  "-DVERSION_GITSHA=0x${version_sha}",
+  "-I${source_root}/src/include" ]
+
+cflags: [ "-std=c11" ]
+cxxflags: [ "-std=c++17" ]

assets/build/rules.ninja

@@ -0,0 +1,91 @@
+rule compile.c
+  command = $cc -MMD -MF $out.d $cflags $ccflags -o $out -c $in
+  description = Compiling [$target]:$name
+  depfile = $out.d
+  deps = gcc
+
+rule dump_c_defs
+  command = echo | $cc $ccflags $cflags -dM -E - > $out
+  description = Dumping C defines for $target
+
+rule dump_c_run
+  command = echo '#!/bin/bash' > $out; echo '$cc $ccflags $cflags $$*' >> $
+      $out; chmod a+x $out
+  description = Dumping C arguments for $target
+
+rule compile.cpp
+  command = $cxx -MMD -MF $out.d $cxxflags $ccflags -o $out -c $in
+  description = Compiling [$target]:$name
+  depfile = $out.d
+  deps = gcc
+
+rule dump_cpp_defs
+  command = echo | $cxx -x c++ $ccflags $cxxflags -dM -E - > $out
+  description = Dumping C++ defines for $target
+
+rule dump_cpp_run
+  command = echo '#!/bin/bash' > $out; echo '$cxx $ccflags $cxxflags $$*' $
+      >> $out; chmod a+x $out
+  description = Dumping C++ arguments for $target
+
+rule compile.s
+  command = $nasm -o $out -felf64 -MD $out.d $asflags $in
+  description = Assembling [$target]:$name
+  depfile = $out.d
+  deps = gcc
+
+rule parse.cog
+  command = cog -o $out -d -D target=$target $cogflags $in
+  description = Parsing [$target]:$name
+
+rule exe
+  command = $ld $ldflags -o $out $in $libs
+  description = Linking exe [$target]:$name
+
+rule driver
+  command = $ld $ldflags -o $out $in $libs
+  description = Linking driver [$target]:$name
+
+rule lib
+  command = $ld -shared -soname $soname $ldflags -o $out $in $libs
+  description = Linking [$target]:$name
+
+rule lib_static
+  command = $ar qcs $out $in
+  description = Archiving [$target]:$name
+
+rule cp
+  command = cp $in $out
+  description = Copying [$target]:$name
+
+rule dump
+  command = objdump -DSC -M intel $in > $out
+  description = Dumping decompiled $name
+
+rule makest
+    description = Making symbol table
+    command = nm -n -S --demangle $in | ${source_root}/scripts/build_symbol_table.py $out
+
+rule makeinitrd
+    description = Creating $name
+    command = ${source_root}/scripts/mkj6romfs.py -c $format $in $out
+
+rule makefat
+    description = Creating $name
+    command = $
+        cp $in $out; $
+        mcopy -s -D o -i $out@@1M ${build_root}/fatroot/* ::/
+
+rule strip
+    description = Stripping $name
+    command = $
+        cp $in $out; $
+        objcopy --only-keep-debug $out $debug; $
+        strip --discard-all -g $out; $
+        objcopy --add-gnu-debuglink=$debug $out
+
+rule touch
+    command = touch $out
+
+rule compdb
+    command = ninja -t compdb > $out

assets/build/target.boot.yaml

@@ -0,0 +1,26 @@
+---
+ld: clang++
+
+ccflags: [
+  "-nostdlib",
+  "-nodefaultlibs",
+  "-fno-builtin",
+
+  "-I${source_root}/external",
+  "--target=x86_64-unknown-windows",
+  "-ffreestanding",
+  "-mno-red-zone",
+  "-fshort-wchar",
+  "-fno-omit-frame-pointer",
+]
+
+cxxflags: [ "-fno-exceptions", "-fno-rtti" ]
+
+ldflags: [
+  "--target=x86_64-unknown-windows",
+  "-nostdlib",
+  "-Wl,-entry:efi_main",
+  "-Wl,-subsystem:efi_application",
+  "-fuse-ld=lld-link",
+]
+

assets/build/target.init.yaml

@@ -0,0 +1,39 @@
+---
+ccflags: [
+  "--target=x86_64-jsix-elf",
+  "-fno-omit-frame-pointer",
+  "-fno-stack-protector",
+
+  "-fvisibility=hidden",
+  "-fvisibility-inlines-hidden",
+
+  "-D__ELF__",
+  "-D__jsix__",
+  "-U__linux",
+  "-U__linux__",
+
+  "-DMSPACES",
+
+  "--sysroot='${source_root}/sysroot'"
+]
+
+
+cxxflags: [
+  "-fno-exceptions",
+  "-fno-rtti",
+]
+
+ldflags: [
+  "-Bstatic",
+  "-m", "elf_x86_64",
+  "--sysroot='${source_root}/sysroot'",
+  "--no-eh-frame-hdr",
+  "-L", "${source_root}/sysroot/lib",
+  "-z", "separate-code",
+  "-lc++", "-lc++abi", "-lunwind",
+  "--no-dependent-libraries",
+]
+
+libs: [
+  "${target_dir}/crt0.o",
+]

assets/build/target.kernel.yaml

@@ -0,0 +1,52 @@
+---
+asflags: [ "-I${source_root}/src/kernel/" ]
+
+ccflags: [
+  "--target=x86_64-jsix-elf",
+  "-fno-stack-protector",
+
+  "-I${source_root}/external",
+
+  "-nostdinc",
+  "-nostdlib",
+  "-ffreestanding",
+  "-nodefaultlibs",
+  "-fno-builtin",
+  "-fno-plt",
+
+  "-mno-sse",
+  "-fno-omit-frame-pointer",
+  "-mno-red-zone",
+  "-mcmodel=kernel",
+  "-fvisibility=hidden",
+  "-fvisibility-inlines-hidden",
+
+  "-D__ELF__",
+  "-D__jsix__",
+  "-D__j6kernel",
+  "-U__linux",
+  "-U__linux__",
+  "-DPRINTF_ALIAS_STANDARD_FUNCTION_NAMES=1",
+  "-DPRINTF_INCLUDE_CONFIG_H=1",
+
+  "--sysroot='${source_root}/sysroot'"
+]
+
+
+cflags: [ '-nostdinc' ]
+
+cxxflags: [
+  "-fno-exceptions",
+  "-fno-rtti",
+  "-nostdinc",
+]
+
+ldflags: [
+  "-m", "elf_x86_64",
+  "-nostdlib",
+  "-Bstatic",
+  "--no-eh-frame-hdr",
+  "-z", "norelro",
+  "-z", "separate-code"
+]
+

assets/build/target.user.exe.yaml

@@ -0,0 +1,15 @@
+---
+
+ccflags: [
+  "-fpie"
+]
+
+ldflags: [
+  "-pie",
+  "--dynamic-linker", "/jsix/lib/ld.so",
+  "--push-state", "--as-needed", "-Bstatic", "-lc++", "-lc++abi", "-lunwind", "--pop-state",
+]
+
+libs: [
+  "${target_dir}/crt0.o",
+]

assets/build/target.user.shared.yaml

@@ -0,0 +1,7 @@
+---
+ccflags: [
+]
+
+ldflags: [
+  "-shared",
+]

assets/build/target.user.yaml

@@ -0,0 +1,33 @@
+---
+asflags: []
+
+ccflags: [
+  "--target=x86_64-jsix-elf",
+  "-fno-omit-frame-pointer",
+  "-fno-stack-protector",
+
+  "-fvisibility=hidden",
+  "-fvisibility-inlines-hidden",
+
+  "-D__ELF__",
+  "-D__jsix__",
+  "-U__linux",
+  "-U__linux__",
+
+  "--sysroot='${source_root}/sysroot'",
+  "-fpic",
+]
+
+cxxflags: [
+  "-fno-exceptions",
+  "-fno-rtti",
+  ]
+
+ldflags: [
+  "-m", "elf_x86_64",
+  "--sysroot='${source_root}/sysroot'",
+  "--no-eh-frame-hdr",
+  "-L", "${source_root}/sysroot/lib",
+  "-z", "separate-code",
+  "--no-dependent-libraries",
+]

assets/debugging/jsix.elf-gdb.py

@@ -1,4 +1,14 @@
 import gdb
+import gdb.printing
+
+import sys
+sys.path.append('./scripts')
+
+import re
+
+from collections import namedtuple
+Capability = namedtuple("Capability", ["id", "parent", "refcount", "caps", "type", "koid"])
+LogEntry = namedtuple("LogHeader", ["id", "bytes", "severity", "area", "message"])
 
 class PrintStackCommand(gdb.Command):
     def __init__(self):
@@ -10,6 +20,7 @@ class PrintStackCommand(gdb.Command):
         base = "$rsp"
         if len(args) > 0:
             base = args[0]
+        base = int(gdb.parse_and_eval(base))
 
         depth = 22
         if len(args) > 1:
@@ -18,14 +29,35 @@ class PrintStackCommand(gdb.Command):
         for i in range(depth-1, -1, -1):
             try:
                 offset = i * 8
-                base_addr = gdb.parse_and_eval(base)
+                value = gdb.parse_and_eval(f"*(uint64_t*)({base:#x} + {offset:#x})")
-                value = gdb.parse_and_eval(f"*(uint64_t*)({base} + {offset})")
+                print("{:016x} (+{:04x}): {:016x}".format(base + offset, offset, int(value)))
-                print("{:016x} (+{:04x}): {:016x}".format(int(base_addr) + offset, offset, int(value)))
             except Exception as e:
                 print(e)
                 continue
 
 
+def stack_walk(frame, depth):
+    for i in range(depth-1, -1, -1):
+        ret = gdb.parse_and_eval(f"*(uint64_t*)({frame:#x} + 0x8)")
+
+        name = ""
+        try:
+            block = gdb.block_for_pc(int(ret))
+            if block:
+                name = block.function or ""
+        except RuntimeError:
+            pass
+
+        try:
+            print("{:016x}: {:016x} {}".format(int(frame), int(ret), name))
+            frame = int(gdb.parse_and_eval(f"*(uint64_t*)({frame:#x})"))
+        except gdb.MemoryError:
+            return
+
+        if frame == 0 or ret == 0:
+            return
+
+
 class PrintBacktraceCommand(gdb.Command):
     def __init__(self):
         super().__init__("j6bt", gdb.COMMAND_DATA)
@@ -33,27 +65,17 @@ class PrintBacktraceCommand(gdb.Command):
     def invoke(self, arg, from_tty):
         args = gdb.string_to_argv(arg)
 
-        depth = 30
-        if len(args) > 0:
-            depth = int(args[0])
-
         frame = "$rbp"
+        if len(args) > 0:
+            frame = args[0]
+
+        frame = int(gdb.parse_and_eval(f"{frame}"))
+
+        depth = 30
         if len(args) > 1:
-            frame = args[1]
+            depth = int(gdb.parse_and_eval(args[1]))
 
-        for i in range(depth-1, -1, -1):
+        stack_walk(frame, depth)
-            ret = gdb.parse_and_eval(f"*(uint64_t*)({frame} + 8)")
-            frame = gdb.parse_and_eval(f"*(uint64_t*)({frame})")
-
-            name = ""
-            block = gdb.block_for_pc(int(ret))
-            if block:
-                name = block.function or ""
-
-            print("{:016x} {}".format(int(ret), name))
-
-            if frame == 0 or ret == 0:
-                return
 
 
 class TableWalkCommand(gdb.Command):
@@ -109,7 +131,7 @@ class TableWalkCommand(gdb.Command):
         table = pml4
         entry = 0
         for i in range(len(indices)):
-            entry = int(gdb.parse_and_eval(f'((uint64_t*){table})[{indices[i]}]'))
+            entry = int(gdb.parse_and_eval(f'((uint64_t*)0x{table:x})[0x{indices[i]:x}]'))
             flagset = flagsets[i]
             flag_names = " | ".join([f[1] for f in flagset if (entry & f[0]) == f[0]])
 
@@ -123,10 +145,384 @@ class TableWalkCommand(gdb.Command):
             table = (entry & 0x7ffffffffffffe00) | 0xffffc00000000000
 
 
+class GetThreadsCommand(gdb.Command):
+    FLAGS = {
+        "ready":    0x01,
+        "loading":  0x02,
+        "exited":   0x04,
+        "constant": 0x80,
+    }
+
+    def __init__(self):
+        super().__init__("j6threads", gdb.COMMAND_DATA)
+
+    def get_flags(self, bitset):
+        flags = []
+        for k, v in GetThreadsCommand.FLAGS.items():
+            if bitset & v:
+                flags.append(k)
+        return " ".join(flags)
+
+    def print_thread(self, addr):
+        if addr == 0:
+            print("  <no thread>\n")
+            return
+
+        tcb = f"((TCB*){addr:#x})"
+        thread = f"({tcb}->thread)"
+        stack = int(gdb.parse_and_eval(f"{tcb}->kernel_stack"))
+        rsp = int(gdb.parse_and_eval(f"{tcb}->rsp"))
+        pri = int(gdb.parse_and_eval(f"{tcb}->priority"))
+        flags = int(gdb.parse_and_eval(f"{thread}->m_state"))
+        koid = int(gdb.parse_and_eval(f"{thread}->m_obj_id"))
+        proc = int(gdb.parse_and_eval(f"{thread}->m_parent.m_obj_id"))
+
+        creator = int(gdb.parse_and_eval(f"{thread}->m_creator"))
+        if creator != 0:
+            creator_koid = int(gdb.parse_and_eval(f"{thread}->m_creator->m_obj_id"))
+            creator = f"{creator_koid:x}"
+        else:
+            creator = "<no thread>"
+
+        print(f"  Thread {proc:x}:{koid:x}")
+        print(f"         creator: {creator}")
+        print(f"        priority: {pri}")
+        print(f"           flags: {self.get_flags(flags)}")
+        print(f"          kstack: {stack:#x}")
+        print(f"             rsp: {rsp:#x}")
+
+        if stack == 0: return 0
+        return int(gdb.parse_and_eval(f"{tcb}->rsp"))
+
+
+    def print_thread_list(self, addr, name):
+        if addr == 0:
+            return
+
+        print(f"=== {name} ===")
+
+        while addr != 0:
+            rsp = self.print_thread(addr)
+            if rsp != 0:
+                print("------------------------------------")
+                stack_walk(rsp + 5*8, 5)
+
+            addr = int(gdb.parse_and_eval(f"((tcb_node*){addr:#x})->m_next"))
+            print()
+
+    def print_cpudata(self, index):
+        cpu = f"(g_cpu_data[{index}])"
+        tss = f"{cpu}->tss"
+        tss_rsp0 = int(gdb.parse_and_eval(f"{tss}->m_rsp[0]"))
+        tss_ist = [int(gdb.parse_and_eval(f"{tss}->m_ist[{i}]")) for i in range(8)]
+        print(f"        tss rsp0: {tss_rsp0:#x}")
+        for i in range(1,8):
+            if tss_ist[i] == 0: continue
+            print(f"        tss ist{i}: {tss_ist[i]:#x}")
+
+    def invoke(self, arg, from_tty):
+        args = gdb.string_to_argv(arg)
+        if len(args) > 1:
+            raise RuntimeError("Usage: j6threads [cpu]")
+
+        ncpus = int(gdb.parse_and_eval("g_num_cpus"))
+        cpus = list(range(ncpus))
+        if len(args) == 1:
+            cpus = [int(args[0])]
+
+        for cpu in cpus:
+            runlist = f"scheduler::s_instance->m_run_queues.m_elements[{cpu:#x}]"
+
+            print(f"=== CPU {cpu:2}: CURRENT ===")
+            current = int(gdb.parse_and_eval(f"{runlist}.current"))
+            self.print_thread(current)
+            self.print_cpudata(cpu)
+
+            previous = int(gdb.parse_and_eval(f"{runlist}.prev"))
+            print(f"            prev: {previous:x}")
+            print()
+
+            for pri in range(8):
+                ready = int(gdb.parse_and_eval(f"{runlist}.ready[{pri:#x}].m_head"))
+                self.print_thread_list(ready, f"CPU {cpu:2}: PRIORITY {pri}")
+
+            blocked = int(gdb.parse_and_eval(f"{runlist}.blocked.m_head"))
+            self.print_thread_list(blocked, f"CPU {cpu:2}: BLOCKED")
+            print()
+
+
+class PrintProfilesCommand(gdb.Command):
+    def __init__(self):
+        super().__init__("j6prof", gdb.COMMAND_DATA)
+
+    def invoke(self, arg, from_tty):
+        args = gdb.string_to_argv(arg)
+        if len(args) != 1:
+            raise RuntimeError("Usage: j6prof <profiler class>")
+
+        profclass = args[0]
+        root_type = f"profile_class<{profclass}>"
+
+        try:
+            baseclass = gdb.lookup_type(root_type)
+        except Exception as e:
+            print(e)
+            return
+
+        results = {}
+        max_len = 0
+        count = gdb.parse_and_eval(f"{profclass}::count")
+        for i in range(count):
+            name = gdb.parse_and_eval(f"{root_type}::function_names[{i:#x}]")
+            if name == 0: continue
+
+            call_counts = gdb.parse_and_eval(f"{root_type}::call_counts[{i:#x}]")
+            call_durations = gdb.parse_and_eval(f"{root_type}::call_durations[{i:#x}]")
+            results[name.string()] = float(call_durations) / float(call_counts)
+            max_len = max(max_len, len(name.string()))
+
+        for name, avg in results.items():
+            print(f"{name:>{max_len}}: {avg:15.3f}")
+
+
+class DumpLogCommand(gdb.Command):
+    level_names = ["", "fatal", "error", "warn", "info", "verbose", "spam"]
+
+    def __init__(self):
+        super().__init__("j6log", gdb.COMMAND_DATA)
+
+        from memory import Layout
+        layout = Layout("definitions/memory_layout.yaml")
+        for region in layout.regions:
+            if region.name == "logs":
+                self.base_addr = region.start
+                break
+
+        self.areas = []
+        area_re = re.compile(r"LOG\(\s*(\w+).*")
+        with open("src/libraries/j6/include/j6/tables/log_areas.inc", 'r') as areas_inc:
+            for line in areas_inc:
+                m = area_re.match(line)
+                if m:
+                    self.areas.append(m.group(1))
+
+    def get_entry(self, addr):
+        addr = int(addr)
+        size = int(gdb.parse_and_eval(f"((j6_log_entry*){addr:#x})->bytes"))
+        mlen = size - 8
+
+        return LogEntry(
+                int(gdb.parse_and_eval(f"((j6_log_entry*){addr:#x})->id")),
+                size,
+                int(gdb.parse_and_eval(f"((j6_log_entry*){addr:#x})->severity")),
+                int(gdb.parse_and_eval(f"((j6_log_entry*){addr:#x})->area")),
+                gdb.parse_and_eval(f"((j6_log_entry*){addr:#x})->message").string(length=mlen))
+
+    def invoke(self, arg, from_tty):
+        start = gdb.parse_and_eval("g_logger.m_start & (g_logger.m_buffer.count - 1)")
+        end = gdb.parse_and_eval("g_logger.m_end & (g_logger.m_buffer.count - 1)")
+        if end < start:
+            end += gdb.parse_and_eval("g_logger.m_buffer.count")
+
+        print(f"Logs are {start} -> {end}")
+
+        addr = self.base_addr + start
+        end += self.base_addr
+        while addr < end:
+            entry = self.get_entry(addr)
+            if entry.bytes < 8:
+                print(f"Bad log header size: {entry.bytes}")
+                break
+            addr += entry.bytes
+            area = "??"
+            if entry.area < len(self.areas):
+                area = self.areas[entry.area]
+            level = self.level_names[entry.severity]
+            print(f"{area:>7}:{level:7}  {entry.message}")
+
+
+class ShowCurrentProcessCommand(gdb.Command):
+    def __init__(self):
+        super().__init__("j6current", gdb.COMMAND_DATA)
+
+    def invoke(self, arg, from_tty):
+        def get_obj_and_id(name):
+            obj = int(gdb.parse_and_eval(f"((cpu_data*)$gs_base)->{name}"))
+            oid = -1
+            if obj != 0:
+                oid = int(gdb.parse_and_eval(f"((obj::kobject*){obj:#x})->m_obj_id"))
+            return obj, oid
+
+        process, pid = get_obj_and_id("process")
+        thread, tid = get_obj_and_id("thread")
+        print(f"{pid:02x}/{tid:02x}  [ {process:x} / {thread:x} ]")
+
+
+class CapTablePrinter:
+    def __init__(self, val):
+        node_map = val["m_caps"]
+        self.nodes = node_map["m_nodes"]
+        self.count = int(node_map["m_count"])
+        self.capacity = int(node_map["m_capacity"])
+
+    class _iterator:
+        def __init__(self, nodes, capacity):
+            self.nodes = []
+
+            for i in range(capacity):
+                node = nodes[i]
+
+                node_id = int(node["id"])
+                if node_id == 0:
+                    continue
+
+                self.nodes.append(Capability(
+                    id = node_id,
+                    parent = int(node["parent"]),
+                    refcount = int(node["holders"]),
+                    caps = int(node["caps"]),
+                    type = str(node["type"])[14:],
+                    koid = node['object']['m_obj_id']))
+
+            self.nodes.sort(key=lambda n: n.id, reverse=True)
+
+        def __iter__(self):
+            return self
+
+        def __next__(self):
+            if not self.nodes:
+                raise StopIteration
+
+            node = self.nodes.pop()
+
+            desc = f'p:{node.parent:016x}  refs:{node.refcount:2}  caps:{node.caps:04x}  {node.type:14} {node.koid}'
+            return (f"{node.id:016x}", desc)
+
+    def to_string(self):
+        return f"Cap table with {self.count}/{self.capacity} nodes.\n"
+
+    def children(self):
+        return self._iterator(self.nodes, self.capacity)
+
+
+class VectorPrinter:
+    def __init__(self, vector):
+        self.name = vector.type.tag
+        self.count = vector["m_size"]
+        self.array = vector["m_elements"]
+
+    class _iterator:
+        def __init__(self, array, count, deref):
+            self.array = array
+            self.count = count
+            self.deref = deref
+            self.index = 0
+
+        def __iter__(self):
+            return self
+
+        def __next__(self):
+            if self.index >= self.count:
+                raise StopIteration
+
+            item = self.array[self.index]
+            if self.deref:
+                item = item.dereference()
+
+            result = (f"{self.index:3}", item)
+            self.index += 1
+            return result
+
+    def to_string(self):
+        return f"{self.name} [{self.count}]"
+
+    def children(self):
+        deref = self.count > 0 and self.array[0].type.code == gdb.TYPE_CODE_PTR
+        return self._iterator(self.array, int(self.count), deref)
+
+
+class HandleSetPrinter:
+    def __init__(self, nodeset):
+        self.node_map = nodeset['m_map']
+        self.count = self.node_map['m_count']
+        self.capacity = self.node_map['m_capacity']
+
+    class _iterator:
+        def __init__(self, nodes, capacity):
+            self.items = []
+            self.index = 0
+            for i in range(capacity):
+                item = nodes[i]
+                if int(item) != 0:
+                    self.items.append(int(item))
+            self.items.sort()
+
+        def __iter__(self):
+            return self
+
+        def __next__(self):
+            if self.index >= len(self.items):
+                raise StopIteration
+            result = (f"{self.index}", f"{self.items[self.index]:016x}")
+            self.index += 1
+            return result
+
+    def to_string(self):
+        return f"node_set[{self.count} / {self.capacity}]:"
+
+    def children(self):
+        return self._iterator(self.node_map['m_nodes'], self.capacity)
+
+
+class LinkedListPrinter:
+    def __init__(self, llist):
+        self.name = llist.type.tag
+        self.head = llist['m_head']
+        self.tail = llist['m_tail']
+
+        self.items = []
+        current = self.head
+        while current:
+            item = current.dereference()
+            self.items.append((str(len(self.items)), item))
+            current = item['m_next']
+
+    class _iterator:
+        def __iter__(self):
+            return self
+
+        def __next__(self):
+            raise StopIteration
+
+    def to_string(self):
+        return f"{self.name}[{len(self.items)}]"
+
+    def children(self):
+        return self.items
+
+
+def build_pretty_printers():
+    pp = gdb.printing.RegexpCollectionPrettyPrinter("jsix")
+    pp.add_printer("cap table", '^cap_table$', CapTablePrinter)
+    pp.add_printer("handle set", '^util::node_set<unsigned long, 0, heap_allocated>$', HandleSetPrinter)
+    pp.add_printer("vector", '^util::vector<.*>$', VectorPrinter)
+    pp.add_printer("linked list", '^util::linked_list<.*>$', LinkedListPrinter)
+    return pp
+
+gdb.printing.register_pretty_printer(
+    gdb.current_objfile(),
+    build_pretty_printers())
 
 PrintStackCommand()
 PrintBacktraceCommand()
 TableWalkCommand()
+GetThreadsCommand()
+PrintProfilesCommand()
+DumpLogCommand()
+ShowCurrentProcessCommand()
 
-gdb.execute("target remote :1234")
 gdb.execute("display/i $rip")
+gdb.execute("define hook-quit\nkill\nend")
+if not gdb.selected_inferior().was_attached:
+    gdb.execute("add-symbol-file build/panic.serial.elf")
+    gdb.execute("target remote :1234")

assets/grammars/definitions.g

@@ -0,0 +1,49 @@
+start: import_statement* (object|interface)+
+
+import_statement: "import" PATH
+
+object: description? "object" name options? super? "{" uid cname? capabilities? method* "}"
+
+interface: description? "interface" name options? "{" uid interface_param* "}"
+
+?interface_param: expose | function
+
+expose: "expose" type
+
+uid: "uid" UID
+
+cname: "cname" IDENTIFIER
+
+capabilities: "capabilities" "[" IDENTIFIER+ "]"
+
+super: ":" name
+
+function: description? "function" name options? ("{" param* "}")?
+
+method: description? "method" name options? ("{" param* "}")?
+
+param: "param" name type options? description?
+
+?type: PRIMITIVE | object_name | struct_name
+
+object_name: "ref" name
+
+struct_name: "struct" name
+
+id: NUMBER
+name: IDENTIFIER
+options: "[" ( OPTION | IDENTIFIER )+ "]"
+description: COMMENT+
+
+PRIMITIVE: INT_TYPE "*"? | "size" | "string" | "buffer" | "address"
+INT_TYPE: /u?int(8|16|32|64)?/
+NUMBER: /(0x)?[0-9a-fA-F]+/
+UID: /[0-9a-fA-F]{16}/
+OPTION.2: IDENTIFIER ":" IDENTIFIER
+COMMENT: /#.*/
+PATH: /"[^"]*"/
+
+%import common.LETTER
+%import common.CNAME -> IDENTIFIER
+%import common.WS
+%ignore WS

assets/initrd.toml

@@ -1,28 +0,0 @@
-# This is the manifest for the initial ramdisk, read by the `makerd` tool.
-# The contents should be a table array of files to add to the ramdistk:
-#
-#  [[files]]
-#  dest = "foo.bar"             # Name of the file in the ramdisk
-#  source = "build/foo/foo.bar" # Location of the file from the project root
-#  executable = true            # Optional, default false. Whether this is an
-#                               # initial application for the kernel to execute
-#                               # on startup
-
-[[files]]
-dest = "screenfont.psf"
-source = "../assets/fonts/tamsyn8x16r.psf"
-
-[[files]]
-dest = "symbol_table.dat"
-source = "symbol_table.dat"
-symbols = true
-
-[[files]]
-dest = "nulldrv1"
-source = "user/nulldrv"
-executable = true
-
-[[files]]
-dest = "nulldrv2"
-source = "user/nulldrv"
-executable = true

assets/manifests/default.yaml

@@ -0,0 +1,16 @@
+---
+location: jsix
+init: srv.init
+initrd:
+  name: initrd.dat
+  format: zstd
+panic:
+  - panic.serial
+services:
+  - srv.logger
+  - testapp
+drivers:
+  - drv.uart
+  - drv.uefi_fb
+libs:
+  - ld.so

assets/manifests/test.yaml

@@ -0,0 +1,11 @@
+---
+location: jsix
+init: srv.init
+initrd:
+  name: initrd.dat
+  format: zstd
+flags: ["test"]
+panic:
+  - panic.serial
+services:
+  - test_runner

configure

@@ -0,0 +1,76 @@
+#!/usr/bin/env python3
+
+def generate(output, config, manifest):
+    from os import makedirs
+    from glob import iglob
+    from pathlib import Path
+    from bonnibel.module import Module
+    from bonnibel.project import Project
+
+    root = Path(__file__).parent.resolve()
+    project = Project(root)
+
+    output = root / output
+    manifest = root / manifest
+
+    sources = [
+        str(root / "src/**/*.module"),
+        str(root / "external/*.module"),
+    ]
+
+    modules = {}
+    for source in sources:
+        for modfile in iglob(source, recursive=True):
+            path = Path(modfile).parent
+
+            def module_init(name, **kwargs):
+                if not "root" in kwargs:
+                    kwargs["root"] = path
+                m = Module(name, modfile, **kwargs)
+                modules[m.name] = m
+                return m
+
+            glo = {
+                "module": module_init,
+                "source_root": root,
+                "build_root": output,
+                "module_root": path,
+                "config": config,
+            }
+            code = compile(open(modfile, 'r').read(), modfile, "exec")
+
+            loc = {}
+            exec(code, glo, loc)
+
+    Module.update(modules)
+
+    makedirs(output.resolve(), exist_ok=True)
+    project.generate(root, output, modules, config, manifest)
+    for mod in modules.values():
+        mod.generate(output)
+
+if __name__ == "__main__":
+    import sys
+    from pathlib import Path
+    sys.path.insert(0, str(Path(__file__).parent / "scripts"))
+
+    from argparse import ArgumentParser
+    from bonnibel import BonnibelError
+
+    p = ArgumentParser(description="Generate jsix build files")
+    p.add_argument("--manifest", "-m", metavar="FILE", default="assets/manifests/default.yaml",
+            help="File to use as the system manifest")
+    p.add_argument("--config", "-c", metavar="NAME", default="debug",
+            help="Configuration to build (eg, 'debug' or 'release')")
+    p.add_argument("output", metavar="DIR", default="build", nargs='?',
+            help="Where to create the build root")
+
+    args = p.parse_args()
+
+    try:
+        generate(args.output, args.config, args.manifest)
+
+    except BonnibelError as be:
+        import sys
+        print(f"Error: {be}", file=sys.stderr)
+        sys.exit(1)

definitions/memory_layout.yaml

@@ -0,0 +1,29 @@
+---
+- name: linear
+  size: 64T
+  shared: true
+
+- name: bitmap
+  size: 1T
+  shared: true
+
+- name: heapmap
+  size: 32G
+
+- name: heap
+  size: 32G
+
+- name: capsmap
+  size: 32G
+
+- name: caps
+  size: 32G
+
+- name: stacks
+  size: 64G
+
+- name: buffers
+  size: 64G
+
+- name: logs
+  size: 2G

definitions/objects/event.def

@@ -0,0 +1,22 @@
+object event : object {
+    uid f441e03da5516b1a
+
+    capabilities [
+        signal
+        wait
+    ]
+
+    method create [constructor]
+
+    # Signal events on this object
+    method signal [cap:signal] {
+        param signals uint64        # A bitset of which events to signal
+    }
+
+    # Wait for signaled events on this object
+    method wait [cap:wait] {
+        param signals uint64 [out]  # A bitset of which events were signaled
+        param timeout uint64        # Wait timeout in nanoseconds
+    }
+}
+

definitions/objects/mailbox.def

@@ -0,0 +1,39 @@
+# Mailboxes are objects that enable synchronous IPC via arbitrary
+# message-passing of tagged data and/or handles. Not as efficient
+# as shared memory channels, but more flexible.
+
+object mailbox : object {
+    uid 99934ad04ece1e07
+
+    capabilities [
+        send
+        receive
+        close
+    ]
+
+    method create [constructor]
+    method close [destructor cap:close]
+
+    # Send a message to the reciever, and block until a response is
+    # sent. Note that getting this response does not require the
+    # receive capability.
+    method call [cap:send] {
+        param tag uint64 [inout]
+        param data buffer [optional inout]
+        param data_in_len size  # number of bytes in data used for input
+        param handles ref object [optional inout handle list]
+    }
+
+    # Respond to a message sent using call, and wait for another
+    # message to arrive. Note that this does not require the send
+    # capability. A reply tag of 0 skips the reply and goes directly
+    # to waiting for a new message.
+    method respond [cap:receive] {
+        param tag uint64 [inout]
+        param data buffer [optional inout]
+        param data_in_len size  # number of bytes in data used for input
+        param handles ref object [optional inout handle list]
+        param reply_tag uint64 [inout]
+        param flags uint64
+    }
+}

definitions/objects/object.def

@@ -0,0 +1,14 @@
+# The base type of all kernel-exposed objects
+object object [virtual] {
+    uid 667f61fb2cd57bb4
+    cname kobject
+
+    capabilities [
+        clone
+    ]
+
+    # Get the internal kernel object id of an object
+    method koid {
+        param koid uint64 [out]
+    }
+}

definitions/objects/process.def

@@ -0,0 +1,30 @@
+import "objects/object.def"
+
+# Processes are a collection of handles and a virtual memory
+# space inside which threads are run.
+
+object process : object {
+    uid 0c69ee0b7502ba31
+
+    capabilities [
+        kill
+        create_thread
+    ]
+
+    # Create a new empty process
+    method create [constructor]
+
+    # Stop all threads and exit the given process
+    method kill [destructor cap:kill]
+
+    # Stop all threads and exit the current process
+    method exit [static noreturn] {
+        param result int32    # The result to retrun to the parent process
+    }
+
+    # Give the given process a handle that points to the same
+    # object as the specified handle.
+    method give_handle {
+        param target ref object [handle]          # A handle in the caller process to send
+    }
+}

definitions/objects/system.def

@@ -0,0 +1,41 @@
+# The system object represents a handle to kernel functionality
+# needed by drivers and other priviledged services
+object system : object {
+    uid fa72506a2cf71a30
+
+    capabilities [
+        get_log
+        bind_irq
+        map_phys
+        change_iopl
+    ]
+
+    # Get the next log line from the kernel log
+    method get_log [cap:get_log] {
+        param seen uint64                 # Last seen log id
+        param buffer buffer [out zero_ok] # Buffer for the log message data structure
+    }
+
+    # Ask the kernel to send this process messages whenever
+    # the given IRQ fires
+    method bind_irq [cap:bind_irq] {
+        param dest ref event     # Event object that will receive messages
+        param irq uint           # IRQ number to bind
+        param signal uint        # Signal number on the event to bind to
+    }
+
+    # Create a VMA and map an area of physical memory into it,
+    # also mapping that VMA into the current process
+    method map_phys [cap:map_phys] {
+        param area ref vma [out] # Receives a handle to the VMA created
+        param phys address       # The physical address of the area
+        param size size          # Size of the area, in bytes
+        param flags uint32       # Flags to apply to the created VMA
+    }
+
+    # Request the kernel change the IOPL for this process. The only values
+    # that make sense are 0 and 3.
+    method request_iopl [cap:change_iopl] {
+        param iopl uint             # The IOPL to set for this process
+    }
+}

definitions/objects/thread.def

@@ -0,0 +1,26 @@
+object thread : object {
+    uid 11f23e593d5761bd
+
+    capabilities [
+        kill
+        join
+    ]
+
+    method create [constructor] {
+        param process ref process [optional cap:create_thread]
+        param stack_top address
+        param entrypoint address
+        param arg0 uint64
+        param arg1 uint64
+    }
+
+    method kill [destructor cap:kill]
+
+    method join [cap:join]
+
+    method exit [static]
+
+    method sleep [static] {
+        param duration uint64
+    }
+}

definitions/objects/vma.def

@@ -0,0 +1,37 @@
+import "objects/process.def"
+
+object vma : object {
+    uid d6a12b63b3ed3937
+    cname vm_area
+
+    capabilities [
+        map
+        unmap
+        resize
+    ]
+
+    method create [constructor] {
+        param size size
+        param flags uint32
+    }
+
+    method create_map [constructor cap:map] {
+        param size size
+        param address address [inout]
+        param flags uint32
+    }
+
+    method map [cap:map] {
+        param process ref process [optional]
+        param address address [inout]
+        param flags uint32
+    }
+
+    method unmap [cap:unmap] {
+        param process ref process [optional]
+    }
+
+    method resize [cap:resize] {
+        param size size [inout]  # New size for the VMA, or 0 to query the current size without changing
+    }
+}

definitions/syscalls.def

@@ -0,0 +1,62 @@
+import "objects/object.def"
+
+import "objects/event.def"
+import "objects/mailbox.def"
+import "objects/process.def"
+import "objects/system.def"
+import "objects/thread.def"
+import "objects/vma.def"
+
+interface syscalls [syscall] {
+    uid 01d9b6a948961097
+
+    expose ref object
+
+    expose ref event
+    expose ref mailbox
+    expose ref process
+    expose ref system
+    expose ref thread
+    expose ref vma
+
+    # Simple no-op syscall for testing
+    function noop
+
+    # Write a message to the kernel log
+    function log {
+        param message string
+    }
+
+    # Get a list of handles owned by this process. If the
+    # supplied list is not big enough, will set the size
+    # needed in `size` and return j6_err_insufficient
+    function handle_list {
+        param handles struct handle_descriptor [list inout zero_ok] # A list of handles to be filled
+    }
+
+    # Create a clone of an existing handle, possibly with
+    # some capabilities masked out.
+    function handle_clone {
+        param orig ref object [handle cap:clone] # The handle to clone
+        param clone ref object [out]             # The new handle
+        param mask uint32                        # The capability bitmask
+    }
+
+    # Block waiting on a futex
+    function futex_wait [static] {
+        param address uint32*  # Address of the futex value
+        param current uint32   # Current value of the futex
+        param timeout uint64   # Wait timeout in nanoseconds
+    }
+
+    # Wake threads waiting on a futex
+    function futex_wake [static] {
+        param address uint32*  # Address of the futex value
+        param count uint64     # Number of threads to wake, or 0 for all
+    }
+
+    # Testing mode only: Have the kernel finish and exit QEMU with the given exit code
+    function test_finish [test] {
+        param exit_code uint32
+    }
+}

definitions/sysconf.yaml

@@ -0,0 +1,34 @@
+---
+address: 0x6a360000
+vars:
+  - name: version_major
+    section: kernel
+    type: uint8_t
+
+  - name: version_minor
+    section: kernel
+    type: uint8_t
+
+  - name: version_patch
+    section: kernel
+    type: uint16_t
+
+  - name: version_gitsha
+    section: kernel
+    type: uint32_t
+
+  - name: page_size
+    section: sys
+    type: size_t
+
+  - name: large_page_size
+    section: sys
+    type: size_t
+
+  - name: huge_page_size
+    section: sys
+    type: size_t
+
+  - name: num_cpus
+    section: sys
+    type: uint32_t

external/uefi/boot_services.h

@@ -22,7 +22,7 @@ namespace bs_impl {
 	using create_event = status (*)(evt, tpl, event_notify, void*, event*);
 	using exit_boot_services = status (*)(handle, size_t);
 	using locate_protocol = status (*)(const guid*, void*, void**);
-	using copy_mem = void (*)(void*, void*, size_t);
+	using copy_mem = void (*)(void*, const void*, size_t);
 	using set_mem = void (*)(void*, uint64_t, uint8_t);
 }
 

external/zstd.module

@@ -0,0 +1,33 @@
+# vim: ft=python
+
+zstd = module("zstd",
+    root = "${source_root}/external/zstd",
+    kind = "lib",
+    copy_headers = True,
+    deps = [ "libc" ],
+    sources = [
+        "decompress/zstd_decompress.c",
+        "decompress/zstd_ddict.c",
+        "decompress/huf_decompress.c",
+        "decompress/zstd_decompress_block.c",
+        "common/threading.c",
+        "common/fse_decompress.c",
+        "common/debug.c",
+        "common/xxhash.c",
+        "common/pool.c",
+        "common/error_private.c",
+        "common/entropy_common.c",
+        "common/zstd_common.c",
+    ],
+    public_headers = [
+        "zdict.h",
+        "zstd.h",
+        "zstd_errors.h",
+    ])
+
+zstd.variables['ccflags'] = [
+    "${ccflags}",
+    "-DXXH_NAMESPACE=ZSTD_",
+    "-DDEBUGLEVEL=0",
+    "-DZSTD_DISABLE_ASM",
+]

external/zstd/.gitignore

@@ -0,0 +1,3 @@
+# make install artefact
+libzstd.pc
+libzstd-nomt

external/zstd/BUCK

@@ -0,0 +1,232 @@
+cxx_library(
+    name='zstd',
+    header_namespace='',
+    exported_headers=['zstd.h'],
+    visibility=['PUBLIC'],
+    deps=[
+        ':common',
+        ':compress',
+        ':decompress',
+        ':deprecated',
+    ],
+)
+
+cxx_library(
+    name='compress',
+    header_namespace='',
+    visibility=['PUBLIC'],
+    exported_headers=subdir_glob([
+        ('compress', 'zstd*.h'),
+    ]),
+    srcs=glob(['compress/zstd*.c', 'compress/hist.c']),
+    deps=[':common'],
+)
+
+cxx_library(
+    name='decompress',
+    header_namespace='',
+    visibility=['PUBLIC'],
+    headers=subdir_glob([
+        ('decompress', '*_impl.h'),
+    ]),
+    srcs=glob(['decompress/zstd*.c']),
+    deps=[
+        ':common',
+        ':legacy',
+    ],
+)
+
+cxx_library(
+    name='deprecated',
+    header_namespace='',
+    visibility=['PUBLIC'],
+    exported_headers=subdir_glob([
+        ('deprecated', '*.h'),
+    ]),
+    srcs=glob(['deprecated/*.c']),
+    deps=[':common'],
+)
+
+cxx_library(
+    name='legacy',
+    header_namespace='',
+    visibility=['PUBLIC'],
+    exported_headers=subdir_glob([
+        ('legacy', '*.h'),
+    ]),
+    srcs=glob(['legacy/*.c']),
+    deps=[':common'],
+    exported_preprocessor_flags=[
+        '-DZSTD_LEGACY_SUPPORT=4',
+    ],
+)
+
+cxx_library(
+    name='zdict',
+    header_namespace='',
+    visibility=['PUBLIC'],
+    exported_headers=['zdict.h'],
+    headers=subdir_glob([
+        ('dictBuilder', 'divsufsort.h'),
+        ('dictBuilder', 'cover.h'),
+    ]),
+    srcs=glob(['dictBuilder/*.c']),
+    deps=[':common'],
+)
+
+cxx_library(
+    name='compiler',
+    header_namespace='',
+    visibility=['PUBLIC'],
+    exported_headers=subdir_glob([
+        ('common', 'compiler.h'),
+    ]),
+)
+
+cxx_library(
+    name='cpu',
+    header_namespace='',
+    visibility=['PUBLIC'],
+    exported_headers=subdir_glob([
+        ('common', 'cpu.h'),
+    ]),
+)
+
+cxx_library(
+    name='bitstream',
+    header_namespace='',
+    visibility=['PUBLIC'],
+    exported_headers=subdir_glob([
+        ('common', 'bitstream.h'),
+    ]),
+)
+
+cxx_library(
+    name='entropy',
+    header_namespace='',
+    visibility=['PUBLIC'],
+    exported_headers=subdir_glob([
+        ('common', 'fse.h'),
+        ('common', 'huf.h'),
+    ]),
+    srcs=[
+        'common/entropy_common.c',
+        'common/fse_decompress.c',
+        'compress/fse_compress.c',
+        'compress/huf_compress.c',
+        'decompress/huf_decompress.c',
+    ],
+    deps=[
+        ':debug',
+        ':bitstream',
+        ':compiler',
+        ':errors',
+        ':mem',
+    ],
+)
+
+cxx_library(
+    name='errors',
+    header_namespace='',
+    visibility=['PUBLIC'],
+    exported_headers=[
+        'zstd_errors.h',
+        'common/error_private.h',
+    ]
+    srcs=['common/error_private.c'],
+)
+
+cxx_library(
+    name='mem',
+    header_namespace='',
+    visibility=['PUBLIC'],
+    exported_headers=subdir_glob([
+        ('common', 'mem.h'),
+    ]),
+)
+
+cxx_library(
+    name='pool',
+    header_namespace='',
+    visibility=['PUBLIC'],
+    exported_headers=subdir_glob([
+        ('common', 'pool.h'),
+    ]),
+    srcs=['common/pool.c'],
+    deps=[
+        ':threading',
+        ':zstd_common',
+    ],
+)
+
+cxx_library(
+    name='threading',
+    header_namespace='',
+    visibility=['PUBLIC'],
+    exported_headers=subdir_glob([
+        ('common', 'threading.h'),
+    ]),
+    srcs=['common/threading.c'],
+    exported_preprocessor_flags=[
+        '-DZSTD_MULTITHREAD',
+    ],
+    exported_linker_flags=[
+        '-pthread',
+    ],
+)
+
+cxx_library(
+    name='xxhash',
+    header_namespace='',
+    visibility=['PUBLIC'],
+    exported_headers=subdir_glob([
+        ('common', 'xxhash.h'),
+    ]),
+    srcs=['common/xxhash.c'],
+    exported_preprocessor_flags=[
+        '-DXXH_NAMESPACE=ZSTD_',
+    ],
+)
+
+cxx_library(
+    name='zstd_common',
+    header_namespace='',
+    visibility=['PUBLIC'],
+    exported_headers=subdir_glob([
+        ('', 'zstd.h'),
+        ('common', 'zstd_internal.h'),
+    ]),
+    srcs=['common/zstd_common.c'],
+    deps=[
+        ':compiler',
+        ':errors',
+        ':mem',
+    ],
+)
+
+cxx_library(
+    name='debug',
+    header_namespace='',
+    visibility=['PUBLIC'],
+    exported_headers=subdir_glob([
+        ('common', 'debug.h'),
+    ]),
+    srcs=['common/debug.c'],
+)
+
+cxx_library(
+    name='common',
+    deps=[
+        ':debug',
+        ':bitstream',
+        ':compiler',
+        ':cpu',
+        ':entropy',
+        ':errors',
+        ':mem',
+        ':pool',
+        ':threading',
+        ':xxhash',
+        ':zstd_common',
+    ]
+)

external/zstd/LICENSE

@@ -0,0 +1,30 @@
+BSD License
+
+For Zstandard software
+
+Copyright (c) Meta Platforms, Inc. and affiliates. All rights reserved.
+
+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.
+
+ * Neither the name Facebook, nor Meta, nor the names of its contributors may
+   be used to endorse or promote products derived from this software without
+   specific prior written permission.
+
+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 HOLDER 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.

external/zstd/Makefile

@@ -0,0 +1,357 @@
+# ################################################################
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under both the BSD-style license (found in the
+# LICENSE file in the root directory of this source tree) and the GPLv2 (found
+# in the COPYING file in the root directory of this source tree).
+# You may select, at your option, one of the above-listed licenses.
+# ################################################################
+
+# Modules
+ZSTD_LIB_COMPRESSION ?= 1
+ZSTD_LIB_DECOMPRESSION ?= 1
+ZSTD_LIB_DICTBUILDER ?= 1
+ZSTD_LIB_DEPRECATED ?= 0
+
+# Input variables for libzstd.mk
+ifeq ($(ZSTD_LIB_COMPRESSION), 0)
+  ZSTD_LIB_DICTBUILDER = 0
+  ZSTD_LIB_DEPRECATED = 0
+endif
+
+ifeq ($(ZSTD_LIB_DECOMPRESSION), 0)
+  ZSTD_LEGACY_SUPPORT = 0
+  ZSTD_LIB_DEPRECATED = 0
+endif
+
+include libzstd.mk
+
+ZSTD_FILES := $(ZSTD_COMMON_FILES) $(ZSTD_LEGACY_FILES)
+
+ifneq ($(ZSTD_LIB_COMPRESSION), 0)
+  ZSTD_FILES += $(ZSTD_COMPRESS_FILES)
+endif
+
+ifneq ($(ZSTD_LIB_DECOMPRESSION), 0)
+  ZSTD_FILES += $(ZSTD_DECOMPRESS_FILES)
+endif
+
+ifneq ($(ZSTD_LIB_DEPRECATED), 0)
+  ZSTD_FILES += $(ZSTD_DEPRECATED_FILES)
+endif
+
+ifneq ($(ZSTD_LIB_DICTBUILDER), 0)
+  ZSTD_FILES += $(ZSTD_DICTBUILDER_FILES)
+endif
+
+ZSTD_LOCAL_SRC := $(notdir $(ZSTD_FILES))
+ZSTD_LOCAL_OBJ0 := $(ZSTD_LOCAL_SRC:.c=.o)
+ZSTD_LOCAL_OBJ := $(ZSTD_LOCAL_OBJ0:.S=.o)
+
+VERSION := $(ZSTD_VERSION)
+
+# Note: by default, the static library is built single-threaded and dynamic library is built
+# multi-threaded. It is possible to force multi or single threaded builds by appending
+# -mt or -nomt to the build target (like lib-mt for multi-threaded, lib-nomt for single-threaded).
+.PHONY: default
+default: lib-release
+
+CPPFLAGS_DYNLIB  += -DZSTD_MULTITHREAD # dynamic library build defaults to multi-threaded
+LDFLAGS_DYNLIB   += -pthread
+CPPFLAGS_STATLIB +=                    # static library build defaults to single-threaded
+
+
+ifeq ($(findstring GCC,$(CCVER)),GCC)
+decompress/zstd_decompress_block.o : CFLAGS+=-fno-tree-vectorize
+endif
+
+
+# macOS linker doesn't support -soname, and use different extension
+# see : https://developer.apple.com/library/mac/documentation/DeveloperTools/Conceptual/DynamicLibraries/100-Articles/DynamicLibraryDesignGuidelines.html
+ifeq ($(UNAME), Darwin)
+  SHARED_EXT = dylib
+  SHARED_EXT_MAJOR = $(LIBVER_MAJOR).$(SHARED_EXT)
+  SHARED_EXT_VER = $(LIBVER).$(SHARED_EXT)
+  SONAME_FLAGS = -install_name $(LIBDIR)/libzstd.$(SHARED_EXT_MAJOR) -compatibility_version $(LIBVER_MAJOR) -current_version $(LIBVER)
+else
+  ifeq ($(UNAME), AIX)
+    SONAME_FLAGS =
+  else
+    SONAME_FLAGS = -Wl,-soname=libzstd.$(SHARED_EXT).$(LIBVER_MAJOR)
+  endif
+  SHARED_EXT = so
+  SHARED_EXT_MAJOR = $(SHARED_EXT).$(LIBVER_MAJOR)
+  SHARED_EXT_VER = $(SHARED_EXT).$(LIBVER)
+endif
+
+
+.PHONY: all
+all: lib
+
+
+.PHONY: libzstd.a  # must be run every time
+libzstd.a: CPPFLAGS += $(CPPFLAGS_STATLIB)
+
+SET_CACHE_DIRECTORY = \
+   +$(MAKE) --no-print-directory $@ \
+    BUILD_DIR=obj/$(HASH_DIR) \
+    CPPFLAGS="$(CPPFLAGS)" \
+    CFLAGS="$(CFLAGS)" \
+    LDFLAGS="$(LDFLAGS)"
+
+ifndef BUILD_DIR
+# determine BUILD_DIR from compilation flags
+
+libzstd.a:
+	$(SET_CACHE_DIRECTORY)
+
+else
+# BUILD_DIR is defined
+
+ZSTD_STATLIB_DIR := $(BUILD_DIR)/static
+ZSTD_STATLIB := $(ZSTD_STATLIB_DIR)/libzstd.a
+ZSTD_STATLIB_OBJ := $(addprefix $(ZSTD_STATLIB_DIR)/,$(ZSTD_LOCAL_OBJ))
+$(ZSTD_STATLIB): ARFLAGS = rcs
+$(ZSTD_STATLIB): | $(ZSTD_STATLIB_DIR)
+$(ZSTD_STATLIB): $(ZSTD_STATLIB_OBJ)
+  # Check for multithread flag at target execution time
+	$(if $(filter -DZSTD_MULTITHREAD,$(CPPFLAGS)),\
+    @echo compiling multi-threaded static library $(LIBVER),\
+    @echo compiling single-threaded static library $(LIBVER))
+	$(AR) $(ARFLAGS) $@ $^
+
+libzstd.a: $(ZSTD_STATLIB)
+	cp -f $< $@
+
+endif
+
+ifneq (,$(filter Windows%,$(TARGET_SYSTEM)))
+
+LIBZSTD = dll/libzstd.dll
+$(LIBZSTD): $(ZSTD_FILES)
+	@echo compiling dynamic library $(LIBVER)
+	$(CC) $(FLAGS) -DZSTD_DLL_EXPORT=1 -Wl,--out-implib,dll/libzstd.dll.a -shared $^ -o $@
+
+else  # not Windows
+
+LIBZSTD = libzstd.$(SHARED_EXT_VER)
+.PHONY: $(LIBZSTD)  # must be run every time
+$(LIBZSTD): CPPFLAGS += $(CPPFLAGS_DYNLIB)
+$(LIBZSTD): CFLAGS   += -fPIC -fvisibility=hidden
+$(LIBZSTD): LDFLAGS  += -shared $(LDFLAGS_DYNLIB)
+
+ifndef BUILD_DIR
+# determine BUILD_DIR from compilation flags
+
+$(LIBZSTD):
+	$(SET_CACHE_DIRECTORY)
+
+else
+# BUILD_DIR is defined
+
+ZSTD_DYNLIB_DIR := $(BUILD_DIR)/dynamic
+ZSTD_DYNLIB := $(ZSTD_DYNLIB_DIR)/$(LIBZSTD)
+ZSTD_DYNLIB_OBJ := $(addprefix $(ZSTD_DYNLIB_DIR)/,$(ZSTD_LOCAL_OBJ))
+
+$(ZSTD_DYNLIB): | $(ZSTD_DYNLIB_DIR)
+$(ZSTD_DYNLIB): $(ZSTD_DYNLIB_OBJ)
+# Check for multithread flag at target execution time
+	$(if $(filter -DZSTD_MULTITHREAD,$(CPPFLAGS)),\
+    @echo compiling multi-threaded dynamic library $(LIBVER),\
+    @echo compiling single-threaded dynamic library $(LIBVER))
+	$(CC) $(FLAGS) $^ $(LDFLAGS) $(SONAME_FLAGS) -o $@
+	@echo creating versioned links
+	ln -sf $@ libzstd.$(SHARED_EXT_MAJOR)
+	ln -sf $@ libzstd.$(SHARED_EXT)
+
+$(LIBZSTD): $(ZSTD_DYNLIB)
+	cp -f $< $@
+
+endif  # ifndef BUILD_DIR
+endif  # if windows
+
+.PHONY: libzstd
+libzstd : $(LIBZSTD)
+
+.PHONY: lib
+lib : libzstd.a libzstd
+
+
+# note : do not define lib-mt or lib-release as .PHONY
+# make does not consider implicit pattern rule for .PHONY target
+
+%-mt : CPPFLAGS_DYNLIB  := -DZSTD_MULTITHREAD
+%-mt : CPPFLAGS_STATLIB := -DZSTD_MULTITHREAD
+%-mt : LDFLAGS_DYNLIB   := -pthread
+%-mt : %
+	@echo multi-threaded build completed
+
+%-nomt : CPPFLAGS_DYNLIB  :=
+%-nomt : LDFLAGS_DYNLIB   :=
+%-nomt : CPPFLAGS_STATLIB :=
+%-nomt : %
+	@echo single-threaded build completed
+
+%-release : DEBUGFLAGS :=
+%-release : %
+	@echo release build completed
+
+
+# Generate .h dependencies automatically
+
+DEPFLAGS = -MT $@ -MMD -MP -MF
+
+$(ZSTD_DYNLIB_DIR)/%.o : %.c $(ZSTD_DYNLIB_DIR)/%.d | $(ZSTD_DYNLIB_DIR)
+	@echo CC $@
+	$(COMPILE.c) $(DEPFLAGS) $(ZSTD_DYNLIB_DIR)/$*.d $(OUTPUT_OPTION) $<
+
+$(ZSTD_STATLIB_DIR)/%.o : %.c $(ZSTD_STATLIB_DIR)/%.d | $(ZSTD_STATLIB_DIR)
+	@echo CC $@
+	$(COMPILE.c) $(DEPFLAGS) $(ZSTD_STATLIB_DIR)/$*.d $(OUTPUT_OPTION) $<
+
+$(ZSTD_DYNLIB_DIR)/%.o : %.S | $(ZSTD_DYNLIB_DIR)
+	@echo AS $@
+	$(COMPILE.S) $(OUTPUT_OPTION) $<
+
+$(ZSTD_STATLIB_DIR)/%.o : %.S | $(ZSTD_STATLIB_DIR)
+	@echo AS $@
+	$(COMPILE.S) $(OUTPUT_OPTION) $<
+
+MKDIR ?= mkdir
+$(BUILD_DIR) $(ZSTD_DYNLIB_DIR) $(ZSTD_STATLIB_DIR):
+	$(MKDIR) -p $@
+
+DEPFILES := $(ZSTD_DYNLIB_OBJ:.o=.d) $(ZSTD_STATLIB_OBJ:.o=.d)
+$(DEPFILES):
+
+include $(wildcard $(DEPFILES))
+
+
+# Special case : building library in single-thread mode _and_ without zstdmt_compress.c
+ZSTDMT_FILES = compress/zstdmt_compress.c
+ZSTD_NOMT_FILES = $(filter-out $(ZSTDMT_FILES),$(ZSTD_FILES))
+libzstd-nomt: CFLAGS += -fPIC -fvisibility=hidden
+libzstd-nomt: LDFLAGS += -shared
+libzstd-nomt: $(ZSTD_NOMT_FILES)
+	@echo compiling single-thread dynamic library $(LIBVER)
+	@echo files : $(ZSTD_NOMT_FILES)
+	$(CC) $(FLAGS) $^ $(LDFLAGS) $(SONAME_FLAGS) -o $@
+
+.PHONY: clean
+clean:
+	$(RM) -r *.dSYM   # macOS-specific
+	$(RM) core *.o *.a *.gcda *.$(SHARED_EXT) *.$(SHARED_EXT).* libzstd.pc
+	$(RM) dll/libzstd.dll dll/libzstd.lib libzstd-nomt*
+	$(RM) -r obj/*
+	@echo Cleaning library completed
+
+#-----------------------------------------------------------------------------
+# make install is validated only for below listed environments
+#-----------------------------------------------------------------------------
+ifneq (,$(filter $(UNAME),Linux Darwin GNU/kFreeBSD GNU OpenBSD FreeBSD NetBSD DragonFly SunOS Haiku AIX))
+
+lib: libzstd.pc
+
+HAS_EXPLICIT_EXEC_PREFIX := $(if $(or $(EXEC_PREFIX),$(exec_prefix)),1,)
+
+DESTDIR     ?=
+# directory variables : GNU conventions prefer lowercase
+# see https://www.gnu.org/prep/standards/html_node/Makefile-Conventions.html
+# support both lower and uppercase (BSD), use uppercase in script
+prefix      ?= /usr/local
+PREFIX      ?= $(prefix)
+exec_prefix ?= $(PREFIX)
+EXEC_PREFIX ?= $(exec_prefix)
+libdir      ?= $(EXEC_PREFIX)/lib
+LIBDIR      ?= $(libdir)
+includedir  ?= $(PREFIX)/include
+INCLUDEDIR  ?= $(includedir)
+
+PCINCDIR := $(patsubst $(PREFIX)%,%,$(INCLUDEDIR))
+PCLIBDIR := $(patsubst $(EXEC_PREFIX)%,%,$(LIBDIR))
+
+# If we successfully stripped off a prefix, we'll add a reference to the
+# relevant pc variable.
+PCINCPREFIX := $(if $(findstring $(INCLUDEDIR),$(PCINCDIR)),,$${prefix})
+PCLIBPREFIX := $(if $(findstring $(LIBDIR),$(PCLIBDIR)),,$${exec_prefix})
+
+# If no explicit EXEC_PREFIX was set by the caller, write it out as a reference
+# to PREFIX, rather than as a resolved value.
+PCEXEC_PREFIX := $(if $(HAS_EXPLICIT_EXEC_PREFIX),$(EXEC_PREFIX),$${prefix})
+
+ifneq (,$(filter $(UNAME),FreeBSD NetBSD DragonFly))
+  PKGCONFIGDIR ?= $(PREFIX)/libdata/pkgconfig
+else
+  PKGCONFIGDIR ?= $(LIBDIR)/pkgconfig
+endif
+
+ifneq (,$(filter $(UNAME),SunOS))
+  INSTALL ?= ginstall
+else
+  INSTALL ?= install
+endif
+
+INSTALL_PROGRAM ?= $(INSTALL)
+INSTALL_DATA    ?= $(INSTALL) -m 644
+
+
+libzstd.pc: libzstd.pc.in
+	@echo creating pkgconfig
+	@sed \
+	        -e 's|@PREFIX@|$(PREFIX)|' \
+	        -e 's|@EXEC_PREFIX@|$(PCEXEC_PREFIX)|' \
+	        -e 's|@INCLUDEDIR@|$(PCINCPREFIX)$(PCINCDIR)|' \
+	        -e 's|@LIBDIR@|$(PCLIBPREFIX)$(PCLIBDIR)|' \
+	        -e 's|@VERSION@|$(VERSION)|' \
+	        -e 's|@LIBS_PRIVATE@|$(LDFLAGS_DYNLIB)|' \
+	        $< >$@
+
+.PHONY: install
+install: install-pc install-static install-shared install-includes
+	@echo zstd static and shared library installed
+
+.PHONY: install-pc
+install-pc: libzstd.pc
+	[ -e $(DESTDIR)$(PKGCONFIGDIR) ] || $(INSTALL) -d -m 755 $(DESTDIR)$(PKGCONFIGDIR)/
+	$(INSTALL_DATA) libzstd.pc $(DESTDIR)$(PKGCONFIGDIR)/
+
+.PHONY: install-static
+install-static:
+	# only generate libzstd.a if it's not already present
+	[ -e libzstd.a ] || $(MAKE) libzstd.a-release
+	[ -e $(DESTDIR)$(LIBDIR) ] || $(INSTALL) -d -m 755 $(DESTDIR)$(LIBDIR)/
+	@echo Installing static library
+	$(INSTALL_DATA) libzstd.a $(DESTDIR)$(LIBDIR)
+
+.PHONY: install-shared
+install-shared:
+	# only generate libzstd.so if it's not already present
+	[ -e $(LIBZSTD) ] || $(MAKE) libzstd-release
+	[ -e $(DESTDIR)$(LIBDIR) ] || $(INSTALL) -d -m 755 $(DESTDIR)$(LIBDIR)/
+	@echo Installing shared library
+	$(INSTALL_PROGRAM) $(LIBZSTD) $(DESTDIR)$(LIBDIR)
+	ln -sf $(LIBZSTD) $(DESTDIR)$(LIBDIR)/libzstd.$(SHARED_EXT_MAJOR)
+	ln -sf $(LIBZSTD) $(DESTDIR)$(LIBDIR)/libzstd.$(SHARED_EXT)
+
+.PHONY: install-includes
+install-includes:
+	[ -e $(DESTDIR)$(INCLUDEDIR) ] || $(INSTALL) -d -m 755 $(DESTDIR)$(INCLUDEDIR)/
+	@echo Installing includes
+	$(INSTALL_DATA) zstd.h $(DESTDIR)$(INCLUDEDIR)
+	$(INSTALL_DATA) zstd_errors.h $(DESTDIR)$(INCLUDEDIR)
+	$(INSTALL_DATA) zdict.h $(DESTDIR)$(INCLUDEDIR)
+
+.PHONY: uninstall
+uninstall:
+	$(RM) $(DESTDIR)$(LIBDIR)/libzstd.a
+	$(RM) $(DESTDIR)$(LIBDIR)/libzstd.$(SHARED_EXT)
+	$(RM) $(DESTDIR)$(LIBDIR)/libzstd.$(SHARED_EXT_MAJOR)
+	$(RM) $(DESTDIR)$(LIBDIR)/$(LIBZSTD)
+	$(RM) $(DESTDIR)$(PKGCONFIGDIR)/libzstd.pc
+	$(RM) $(DESTDIR)$(INCLUDEDIR)/zstd.h
+	$(RM) $(DESTDIR)$(INCLUDEDIR)/zstd_errors.h
+	$(RM) $(DESTDIR)$(INCLUDEDIR)/zdict.h
+	@echo zstd libraries successfully uninstalled
+
+endif

external/zstd/README.md

@@ -0,0 +1,224 @@
+Zstandard library files
+================================
+
+The __lib__ directory is split into several sub-directories,
+in order to make it easier to select or exclude features.
+
+
+#### Building
+
+`Makefile` script is provided, supporting [Makefile conventions](https://www.gnu.org/prep/standards/html_node/Makefile-Conventions.html#Makefile-Conventions),
+including commands variables, staged install, directory variables and standard targets.
+- `make` : generates both static and dynamic libraries
+- `make install` : install libraries and headers in target system directories
+
+`libzstd` default scope is pretty large, including compression, decompression, dictionary builder,
+and support for decoding legacy formats >= v0.5.0.
+The scope can be reduced on demand (see paragraph _modular build_).
+
+
+#### Multithreading support
+
+When building with `make`, by default the dynamic library is multithreaded and static library is single-threaded (for compatibility reasons).
+
+Enabling multithreading requires 2 conditions :
+- set build macro `ZSTD_MULTITHREAD` (`-DZSTD_MULTITHREAD` for `gcc`)
+- for POSIX systems : compile with pthread (`-pthread` compilation flag for `gcc`)
+
+For convenience, we provide a build target to generate multi and single threaded libraries:
+- Force enable multithreading on both dynamic and static libraries by appending `-mt` to the target, e.g. `make lib-mt`.
+- Force disable multithreading on both dynamic and static libraries by appending `-nomt` to the target, e.g. `make lib-nomt`.
+- By default, as mentioned before, dynamic library is multithreaded, and static library is single-threaded, e.g. `make lib`.
+
+When linking a POSIX program with a multithreaded version of `libzstd`,
+note that it's necessary to invoke the `-pthread` flag during link stage.
+
+Multithreading capabilities are exposed
+via the [advanced API defined in `lib/zstd.h`](https://github.com/facebook/zstd/blob/v1.4.3/lib/zstd.h#L351).
+
+
+#### API
+
+Zstandard's stable API is exposed within [lib/zstd.h](zstd.h).
+
+
+#### Advanced API
+
+Optional advanced features are exposed via :
+
+- `lib/zstd_errors.h` : translates `size_t` function results
+                        into a `ZSTD_ErrorCode`, for accurate error handling.
+
+- `ZSTD_STATIC_LINKING_ONLY` : if this macro is defined _before_ including `zstd.h`,
+                          it unlocks access to the experimental API,
+                          exposed in the second part of `zstd.h`.
+                          All definitions in the experimental APIs are unstable,
+                          they may still change in the future, or even be removed.
+                          As a consequence, experimental definitions shall ___never be used with dynamic library___ !
+                          Only static linking is allowed.
+
+
+#### Modular build
+
+It's possible to compile only a limited set of features within `libzstd`.
+The file structure is designed to make this selection manually achievable for any build system :
+
+- Directory `lib/common` is always required, for all variants.
+
+- Compression source code lies in `lib/compress`
+
+- Decompression source code lies in `lib/decompress`
+
+- It's possible to include only `compress` or only `decompress`, they don't depend on each other.
+
+- `lib/dictBuilder` : makes it possible to generate dictionaries from a set of samples.
+        The API is exposed in `lib/dictBuilder/zdict.h`.
+        This module depends on both `lib/common` and `lib/compress` .
+
+- `lib/legacy` : makes it possible to decompress legacy zstd formats, starting from `v0.1.0`.
+        This module depends on `lib/common` and `lib/decompress`.
+        To enable this feature, define `ZSTD_LEGACY_SUPPORT` during compilation.
+        Specifying a number limits versions supported to that version onward.
+        For example, `ZSTD_LEGACY_SUPPORT=2` means : "support legacy formats >= v0.2.0".
+        Conversely, `ZSTD_LEGACY_SUPPORT=0` means "do __not__ support legacy formats".
+        By default, this build macro is set as `ZSTD_LEGACY_SUPPORT=5`.
+        Decoding supported legacy format is a transparent capability triggered within decompression functions.
+        It's also allowed to invoke legacy API directly, exposed in `lib/legacy/zstd_legacy.h`.
+        Each version does also provide its own set of advanced API.
+        For example, advanced API for version `v0.4` is exposed in `lib/legacy/zstd_v04.h` .
+
+- While invoking `make libzstd`, it's possible to define build macros
+        `ZSTD_LIB_COMPRESSION, ZSTD_LIB_DECOMPRESSION`, `ZSTD_LIB_DICTBUILDER`,
+        and `ZSTD_LIB_DEPRECATED` as `0` to forgo compilation of the
+        corresponding features. This will also disable compilation of all
+        dependencies (e.g. `ZSTD_LIB_COMPRESSION=0` will also disable
+        dictBuilder).
+
+- There are a number of options that can help minimize the binary size of
+  `libzstd`.
+
+  The first step is to select the components needed (using the above-described
+  `ZSTD_LIB_COMPRESSION` etc.).
+
+  The next step is to set `ZSTD_LIB_MINIFY` to `1` when invoking `make`. This
+  disables various optional components and changes the compilation flags to
+  prioritize space-saving.
+
+  Detailed options: Zstandard's code and build environment is set up by default
+  to optimize above all else for performance. In pursuit of this goal, Zstandard
+  makes significant trade-offs in code size. For example, Zstandard often has
+  more than one implementation of a particular component, with each
+  implementation optimized for different scenarios. For example, the Huffman
+  decoder has complementary implementations that decode the stream one symbol at
+  a time or two symbols at a time. Zstd normally includes both (and dispatches
+  between them at runtime), but by defining `HUF_FORCE_DECOMPRESS_X1` or
+  `HUF_FORCE_DECOMPRESS_X2`, you can force the use of one or the other, avoiding
+  compilation of the other. Similarly, `ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT`
+  and `ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG` force the compilation and use of
+  only one or the other of two decompression implementations. The smallest
+  binary is achieved by using `HUF_FORCE_DECOMPRESS_X1` and
+  `ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT` (implied by `ZSTD_LIB_MINIFY`).
+
+  For squeezing the last ounce of size out, you can also define
+  `ZSTD_NO_INLINE`, which disables inlining, and `ZSTD_STRIP_ERROR_STRINGS`,
+  which removes the error messages that are otherwise returned by
+  `ZSTD_getErrorName` (implied by `ZSTD_LIB_MINIFY`).
+
+  Finally, when integrating into your application, make sure you're doing link-
+  time optimization and unused symbol garbage collection (via some combination of,
+  e.g., `-flto`, `-ffat-lto-objects`, `-fuse-linker-plugin`,
+  `-ffunction-sections`, `-fdata-sections`, `-fmerge-all-constants`,
+  `-Wl,--gc-sections`, `-Wl,-z,norelro`, and an archiver that understands
+  the compiler's intermediate representation, e.g., `AR=gcc-ar`). Consult your
+  compiler's documentation.
+
+- While invoking `make libzstd`, the build macro `ZSTD_LEGACY_MULTITHREADED_API=1`
+  will expose the deprecated `ZSTDMT` API exposed by `zstdmt_compress.h` in
+  the shared library, which is now hidden by default.
+
+- The build macro `DYNAMIC_BMI2` can be set to 1 or 0 in order to generate binaries
+  which can detect at runtime the presence of BMI2 instructions, and use them only if present.
+  These instructions contribute to better performance, notably on the decoder side.
+  By default, this feature is automatically enabled on detecting
+  the right instruction set (x64) and compiler (clang or gcc >= 5).
+  It's obviously disabled for different cpus,
+  or when BMI2 instruction set is _required_ by the compiler command line
+  (in this case, only the BMI2 code path is generated).
+  Setting this macro will either force to generate the BMI2 dispatcher (1)
+  or prevent it (0). It overrides automatic detection.
+
+- The build macro `ZSTD_NO_UNUSED_FUNCTIONS` can be defined to hide the definitions of functions
+  that zstd does not use. Not all unused functions are hidden, but they can be if needed.
+  Currently, this macro will hide function definitions in FSE and HUF that use an excessive
+  amount of stack space.
+
+- The build macro `ZSTD_NO_INTRINSICS` can be defined to disable all explicit intrinsics.
+  Compiler builtins are still used.
+
+- The build macro `ZSTD_DECODER_INTERNAL_BUFFER` can be set to control
+  the amount of extra memory used during decompression to store literals.
+  This defaults to 64kB.  Reducing this value reduces the memory footprint of
+  `ZSTD_DCtx` decompression contexts,
+  but might also result in a small decompression speed cost.
+
+- The C compiler macros `ZSTDLIB_VISIBLE`, `ZSTDERRORLIB_VISIBLE` and `ZDICTLIB_VISIBLE`
+  can be overridden to control the visibility of zstd's API. Additionally,
+  `ZSTDLIB_STATIC_API` and `ZDICTLIB_STATIC_API` can be overridden to control the visibility
+  of zstd's static API. Specifically, it can be set to `ZSTDLIB_HIDDEN` to hide the symbols
+  from the shared library. These macros default to `ZSTDLIB_VISIBILITY`,
+  `ZSTDERRORLIB_VSIBILITY`, and `ZDICTLIB_VISIBILITY` if unset, for backwards compatibility
+  with the old macro names.
+
+#### Windows : using MinGW+MSYS to create DLL
+
+DLL can be created using MinGW+MSYS with the `make libzstd` command.
+This command creates `dll\libzstd.dll` and the import library `dll\libzstd.lib`.
+The import library is only required with Visual C++.
+The header file `zstd.h` and the dynamic library `dll\libzstd.dll` are required to
+compile a project using gcc/MinGW.
+The dynamic library has to be added to linking options.
+It means that if a project that uses ZSTD consists of a single `test-dll.c`
+file it should be linked with `dll\libzstd.dll`. For example:
+```
+    gcc $(CFLAGS) -Iinclude/ test-dll.c -o test-dll dll\libzstd.dll
+```
+The compiled executable will require ZSTD DLL which is available at `dll\libzstd.dll`.
+
+
+#### Advanced Build options
+
+The build system requires a hash function in order to
+separate object files created with different compilation flags.
+By default, it tries to use `md5sum` or equivalent.
+The hash function can be manually switched by setting the `HASH` variable.
+For example : `make HASH=xxhsum`
+The hash function needs to generate at least 64-bit using hexadecimal format.
+When no hash function is found,
+the Makefile just generates all object files into the same default directory,
+irrespective of compilation flags.
+This functionality only matters if `libzstd` is compiled multiple times
+with different build flags.
+
+The build directory, where object files are stored
+can also be manually controlled using variable `BUILD_DIR`,
+for example `make BUILD_DIR=objectDir/v1`.
+In which case, the hash function doesn't matter.
+
+
+#### Deprecated API
+
+Obsolete API on their way out are stored in directory `lib/deprecated`.
+At this stage, it contains older streaming prototypes, in `lib/deprecated/zbuff.h`.
+These prototypes will be removed in some future version.
+Consider migrating code towards supported streaming API exposed in `zstd.h`.
+
+
+#### Miscellaneous
+
+The other files are not source code. There are :
+
+ - `BUCK` : support for `buck` build system (https://buckbuild.com/)
+ - `Makefile` : `make` script to build and install zstd library (static and dynamic)
+ - `README.md` : this file
+ - `dll/` : resources directory for Windows compilation
+ - `libzstd.pc.in` : script for `pkg-config` (used in `make install`)

external/zstd/common/bits.h

@@ -0,0 +1,175 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_BITS_H
+#define ZSTD_BITS_H
+
+#include "mem.h"
+
+MEM_STATIC unsigned ZSTD_countTrailingZeros32_fallback(U32 val)
+{
+    assert(val != 0);
+    {
+        static const int DeBruijnBytePos[32] = {0, 1, 28, 2, 29, 14, 24, 3,
+                                                30, 22, 20, 15, 25, 17, 4, 8,
+                                                31, 27, 13, 23, 21, 19, 16, 7,
+                                                26, 12, 18, 6, 11, 5, 10, 9};
+        return DeBruijnBytePos[((U32) ((val & -(S32) val) * 0x077CB531U)) >> 27];
+    }
+}
+
+MEM_STATIC unsigned ZSTD_countTrailingZeros32(U32 val)
+{
+    assert(val != 0);
+#   if defined(_MSC_VER)
+#       if STATIC_BMI2 == 1
+            return _tzcnt_u32(val);
+#       else
+            if (val != 0) {
+                unsigned long r;
+                _BitScanForward(&r, val);
+                return (unsigned)r;
+            } else {
+                /* Should not reach this code path */
+                __assume(0);
+            }
+#       endif
+#   elif defined(__GNUC__) && (__GNUC__ >= 4)
+        return (unsigned)__builtin_ctz(val);
+#   else
+        return ZSTD_countTrailingZeros32_fallback(val);
+#   endif
+}
+
+MEM_STATIC unsigned ZSTD_countLeadingZeros32_fallback(U32 val) {
+    assert(val != 0);
+    {
+        static const U32 DeBruijnClz[32] = {0, 9, 1, 10, 13, 21, 2, 29,
+                                            11, 14, 16, 18, 22, 25, 3, 30,
+                                            8, 12, 20, 28, 15, 17, 24, 7,
+                                            19, 27, 23, 6, 26, 5, 4, 31};
+        val |= val >> 1;
+        val |= val >> 2;
+        val |= val >> 4;
+        val |= val >> 8;
+        val |= val >> 16;
+        return 31 - DeBruijnClz[(val * 0x07C4ACDDU) >> 27];
+    }
+}
+
+MEM_STATIC unsigned ZSTD_countLeadingZeros32(U32 val)
+{
+    assert(val != 0);
+#   if defined(_MSC_VER)
+#       if STATIC_BMI2 == 1
+            return _lzcnt_u32(val);
+#       else
+            if (val != 0) {
+                unsigned long r;
+                _BitScanReverse(&r, val);
+                return (unsigned)(31 - r);
+            } else {
+                /* Should not reach this code path */
+                __assume(0);
+            }
+#       endif
+#   elif defined(__GNUC__) && (__GNUC__ >= 4)
+        return (unsigned)__builtin_clz(val);
+#   else
+        return ZSTD_countLeadingZeros32_fallback(val);
+#   endif
+}
+
+MEM_STATIC unsigned ZSTD_countTrailingZeros64(U64 val)
+{
+    assert(val != 0);
+#   if defined(_MSC_VER) && defined(_WIN64)
+#       if STATIC_BMI2 == 1
+            return _tzcnt_u64(val);
+#       else
+            if (val != 0) {
+                unsigned long r;
+                _BitScanForward64(&r, val);
+                return (unsigned)r;
+            } else {
+                /* Should not reach this code path */
+                __assume(0);
+            }
+#       endif
+#   elif defined(__GNUC__) && (__GNUC__ >= 4) && defined(__LP64__)
+        return (unsigned)__builtin_ctzll(val);
+#   else
+        {
+            U32 mostSignificantWord = (U32)(val >> 32);
+            U32 leastSignificantWord = (U32)val;
+            if (leastSignificantWord == 0) {
+                return 32 + ZSTD_countTrailingZeros32(mostSignificantWord);
+            } else {
+                return ZSTD_countTrailingZeros32(leastSignificantWord);
+            }
+        }
+#   endif
+}
+
+MEM_STATIC unsigned ZSTD_countLeadingZeros64(U64 val)
+{
+    assert(val != 0);
+#   if defined(_MSC_VER) && defined(_WIN64)
+#       if STATIC_BMI2 == 1
+            return _lzcnt_u64(val);
+#       else
+            if (val != 0) {
+                unsigned long r;
+                _BitScanReverse64(&r, val);
+                return (unsigned)(63 - r);
+            } else {
+                /* Should not reach this code path */
+                __assume(0);
+            }
+#       endif
+#   elif defined(__GNUC__) && (__GNUC__ >= 4)
+        return (unsigned)(__builtin_clzll(val));
+#   else
+        {
+            U32 mostSignificantWord = (U32)(val >> 32);
+            U32 leastSignificantWord = (U32)val;
+            if (mostSignificantWord == 0) {
+                return 32 + ZSTD_countLeadingZeros32(leastSignificantWord);
+            } else {
+                return ZSTD_countLeadingZeros32(mostSignificantWord);
+            }
+        }
+#   endif
+}
+
+MEM_STATIC unsigned ZSTD_NbCommonBytes(size_t val)
+{
+    if (MEM_isLittleEndian()) {
+        if (MEM_64bits()) {
+            return ZSTD_countTrailingZeros64((U64)val) >> 3;
+        } else {
+            return ZSTD_countTrailingZeros32((U32)val) >> 3;
+        }
+    } else {  /* Big Endian CPU */
+        if (MEM_64bits()) {
+            return ZSTD_countLeadingZeros64((U64)val) >> 3;
+        } else {
+            return ZSTD_countLeadingZeros32((U32)val) >> 3;
+        }
+    }
+}
+
+MEM_STATIC unsigned ZSTD_highbit32(U32 val)   /* compress, dictBuilder, decodeCorpus */
+{
+    assert(val != 0);
+    return 31 - ZSTD_countLeadingZeros32(val);
+}
+
+#endif /* ZSTD_BITS_H */

external/zstd/common/bitstream.h

@@ -0,0 +1,437 @@
+/* ******************************************************************
+ * bitstream
+ * Part of FSE library
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ *
+ * You can contact the author at :
+ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+#ifndef BITSTREAM_H_MODULE
+#define BITSTREAM_H_MODULE
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+/*
+*  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 "mem.h"            /* unaligned access routines */
+#include "compiler.h"       /* UNLIKELY() */
+#include "debug.h"          /* assert(), DEBUGLOG(), RAWLOG() */
+#include "error_private.h"  /* error codes and messages */
+#include "bits.h"           /* ZSTD_highbit32 */
+
+
+/*=========================================
+*  Target specific
+=========================================*/
+#ifndef ZSTD_NO_INTRINSICS
+#  if (defined(__BMI__) || defined(__BMI2__)) && defined(__GNUC__)
+#    include <immintrin.h>   /* support for bextr (experimental)/bzhi */
+#  elif defined(__ICCARM__)
+#    include <intrinsics.h>
+#  endif
+#endif
+
+#define STREAM_ACCUMULATOR_MIN_32  25
+#define STREAM_ACCUMULATOR_MIN_64  57
+#define STREAM_ACCUMULATOR_MIN    ((U32)(MEM_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;
+    unsigned bitPos;
+    char*  startPtr;
+    char*  ptr;
+    char*  endPtr;
+} BIT_CStream_t;
+
+MEM_STATIC size_t BIT_initCStream(BIT_CStream_t* bitC, void* dstBuffer, size_t dstCapacity);
+MEM_STATIC void   BIT_addBits(BIT_CStream_t* bitC, size_t value, unsigned nbBits);
+MEM_STATIC void   BIT_flushBits(BIT_CStream_t* bitC);
+MEM_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;
+    const char* limitPtr;
+} 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 ... :( */
+
+MEM_STATIC size_t   BIT_initDStream(BIT_DStream_t* bitD, const void* srcBuffer, size_t srcSize);
+MEM_STATIC size_t   BIT_readBits(BIT_DStream_t* bitD, unsigned nbBits);
+MEM_STATIC BIT_DStream_status BIT_reloadDStream(BIT_DStream_t* bitD);
+MEM_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
+******************************************/
+MEM_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 */
+
+MEM_STATIC void BIT_flushBitsFast(BIT_CStream_t* bitC);
+/* unsafe version; does not check buffer overflow */
+
+MEM_STATIC size_t BIT_readBitsFast(BIT_DStream_t* bitD, unsigned nbBits);
+/* faster, but works only if nbBits >= 1 */
+
+/*=====    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, 0x7FFFFFF, 0xFFFFFFF, 0x1FFFFFFF,
+    0x3FFFFFFF, 0x7FFFFFFF}; /* up to 31 bits */
+#define BIT_MASK_SIZE (sizeof(BIT_mask) / sizeof(BIT_mask[0]))
+
+/*-**************************************************************
+*  bitStream encoding
+****************************************************************/
+/*! BIT_initCStream() :
+ *  `dstCapacity` must be > sizeof(size_t)
+ *  @return : 0 if success,
+ *            otherwise an error code (can be tested using ERR_isError()) */
+MEM_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->bitContainer);
+    if (dstCapacity <= sizeof(bitC->bitContainer)) return ERROR(dstSize_tooSmall);
+    return 0;
+}
+
+MEM_STATIC FORCE_INLINE_ATTR size_t BIT_getLowerBits(size_t bitContainer, U32 const nbBits)
+{
+#if defined(STATIC_BMI2) && STATIC_BMI2 == 1 && !defined(ZSTD_NO_INTRINSICS)
+    return  _bzhi_u64(bitContainer, nbBits);
+#else
+    assert(nbBits < BIT_MASK_SIZE);
+    return bitContainer & BIT_mask[nbBits];
+#endif
+}
+
+/*! BIT_addBits() :
+ *  can add up to 31 bits into `bitC`.
+ *  Note : does not check for register overflow ! */
+MEM_STATIC void BIT_addBits(BIT_CStream_t* bitC,
+                            size_t value, unsigned nbBits)
+{
+    DEBUG_STATIC_ASSERT(BIT_MASK_SIZE == 32);
+    assert(nbBits < BIT_MASK_SIZE);
+    assert(nbBits + bitC->bitPos < sizeof(bitC->bitContainer) * 8);
+    bitC->bitContainer |= BIT_getLowerBits(value, nbBits) << bitC->bitPos;
+    bitC->bitPos += nbBits;
+}
+
+/*! BIT_addBitsFast() :
+ *  works only if `value` is _clean_,
+ *  meaning all high bits above nbBits are 0 */
+MEM_STATIC void BIT_addBitsFast(BIT_CStream_t* bitC,
+                                size_t value, unsigned nbBits)
+{
+    assert((value>>nbBits) == 0);
+    assert(nbBits + bitC->bitPos < sizeof(bitC->bitContainer) * 8);
+    bitC->bitContainer |= value << bitC->bitPos;
+    bitC->bitPos += nbBits;
+}
+
+/*! BIT_flushBitsFast() :
+ *  assumption : bitContainer has not overflowed
+ *  unsafe version; does not check buffer overflow */
+MEM_STATIC void BIT_flushBitsFast(BIT_CStream_t* bitC)
+{
+    size_t const nbBytes = bitC->bitPos >> 3;
+    assert(bitC->bitPos < sizeof(bitC->bitContainer) * 8);
+    assert(bitC->ptr <= bitC->endPtr);
+    MEM_writeLEST(bitC->ptr, bitC->bitContainer);
+    bitC->ptr += nbBytes;
+    bitC->bitPos &= 7;
+    bitC->bitContainer >>= nbBytes*8;
+}
+
+/*! BIT_flushBits() :
+ *  assumption : bitContainer has not overflowed
+ *  safe version; check for buffer overflow, and prevents it.
+ *  note : does not signal buffer overflow.
+ *  overflow will be revealed later on using BIT_closeCStream() */
+MEM_STATIC void BIT_flushBits(BIT_CStream_t* bitC)
+{
+    size_t const nbBytes = bitC->bitPos >> 3;
+    assert(bitC->bitPos < sizeof(bitC->bitContainer) * 8);
+    assert(bitC->ptr <= bitC->endPtr);
+    MEM_writeLEST(bitC->ptr, bitC->bitContainer);
+    bitC->ptr += nbBytes;
+    if (bitC->ptr > bitC->endPtr) bitC->ptr = bitC->endPtr;
+    bitC->bitPos &= 7;
+    bitC->bitContainer >>= nbBytes*8;
+}
+
+/*! BIT_closeCStream() :
+ *  @return : size of CStream, in bytes,
+ *            or 0 if it could not fit into dstBuffer */
+MEM_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; /* overflow detected */
+    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
+ */
+MEM_STATIC size_t BIT_initDStream(BIT_DStream_t* bitD, const void* srcBuffer, size_t srcSize)
+{
+    if (srcSize < 1) { ZSTD_memset(bitD, 0, sizeof(*bitD)); return ERROR(srcSize_wrong); }
+
+    bitD->start = (const char*)srcBuffer;
+    bitD->limitPtr = bitD->start + sizeof(bitD->bitContainer);
+
+    if (srcSize >=  sizeof(bitD->bitContainer)) {  /* normal case */
+        bitD->ptr   = (const char*)srcBuffer + srcSize - sizeof(bitD->bitContainer);
+        bitD->bitContainer = MEM_readLEST(bitD->ptr);
+        { BYTE const lastByte = ((const BYTE*)srcBuffer)[srcSize-1];
+          bitD->bitsConsumed = lastByte ? 8 - ZSTD_highbit32(lastByte) : 0;  /* ensures bitsConsumed is always set */
+          if (lastByte == 0) return ERROR(GENERIC); /* endMark not present */ }
+    } else {
+        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);
+                ZSTD_FALLTHROUGH;
+
+        case 6: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[5]) << (sizeof(bitD->bitContainer)*8 - 24);
+                ZSTD_FALLTHROUGH;
+
+        case 5: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[4]) << (sizeof(bitD->bitContainer)*8 - 32);
+                ZSTD_FALLTHROUGH;
+
+        case 4: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[3]) << 24;
+                ZSTD_FALLTHROUGH;
+
+        case 3: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[2]) << 16;
+                ZSTD_FALLTHROUGH;
+
+        case 2: bitD->bitContainer += (size_t)(((const BYTE*)(srcBuffer))[1]) <<  8;
+                ZSTD_FALLTHROUGH;
+
+        default: break;
+        }
+        {   BYTE const lastByte = ((const BYTE*)srcBuffer)[srcSize-1];
+            bitD->bitsConsumed = lastByte ? 8 - ZSTD_highbit32(lastByte) : 0;
+            if (lastByte == 0) return ERROR(corruption_detected);  /* endMark not present */
+        }
+        bitD->bitsConsumed += (U32)(sizeof(bitD->bitContainer) - srcSize)*8;
+    }
+
+    return srcSize;
+}
+
+MEM_STATIC FORCE_INLINE_ATTR size_t BIT_getUpperBits(size_t bitContainer, U32 const start)
+{
+    return bitContainer >> start;
+}
+
+MEM_STATIC FORCE_INLINE_ATTR size_t BIT_getMiddleBits(size_t bitContainer, U32 const start, U32 const nbBits)
+{
+    U32 const regMask = sizeof(bitContainer)*8 - 1;
+    /* if start > regMask, bitstream is corrupted, and result is undefined */
+    assert(nbBits < BIT_MASK_SIZE);
+    /* x86 transform & ((1 << nbBits) - 1) to bzhi instruction, it is better
+     * than accessing memory. When bmi2 instruction is not present, we consider
+     * such cpus old (pre-Haswell, 2013) and their performance is not of that
+     * importance.
+     */
+#if defined(__x86_64__) || defined(_M_X86)
+    return (bitContainer >> (start & regMask)) & ((((U64)1) << nbBits) - 1);
+#else
+    return (bitContainer >> (start & regMask)) & BIT_mask[nbBits];
+#endif
+}
+
+/*! 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 */
+MEM_STATIC  FORCE_INLINE_ATTR size_t BIT_lookBits(const BIT_DStream_t*  bitD, U32 nbBits)
+{
+    /* arbitrate between double-shift and shift+mask */
+#if 1
+    /* if bitD->bitsConsumed + nbBits > sizeof(bitD->bitContainer)*8,
+     * bitstream is likely corrupted, and result is undefined */
+    return BIT_getMiddleBits(bitD->bitContainer, (sizeof(bitD->bitContainer)*8) - bitD->bitsConsumed - nbBits, nbBits);
+#else
+    /* this code path is slower on my os-x laptop */
+    U32 const regMask = sizeof(bitD->bitContainer)*8 - 1;
+    return ((bitD->bitContainer << (bitD->bitsConsumed & regMask)) >> 1) >> ((regMask-nbBits) & regMask);
+#endif
+}
+
+/*! BIT_lookBitsFast() :
+ *  unsafe version; only works if nbBits >= 1 */
+MEM_STATIC size_t BIT_lookBitsFast(const BIT_DStream_t* bitD, U32 nbBits)
+{
+    U32 const regMask = sizeof(bitD->bitContainer)*8 - 1;
+    assert(nbBits >= 1);
+    return (bitD->bitContainer << (bitD->bitsConsumed & regMask)) >> (((regMask+1)-nbBits) & regMask);
+}
+
+MEM_STATIC FORCE_INLINE_ATTR 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. */
+MEM_STATIC FORCE_INLINE_ATTR size_t BIT_readBits(BIT_DStream_t* bitD, unsigned nbBits)
+{
+    size_t const value = BIT_lookBits(bitD, nbBits);
+    BIT_skipBits(bitD, nbBits);
+    return value;
+}
+
+/*! BIT_readBitsFast() :
+ *  unsafe version; only works if nbBits >= 1 */
+MEM_STATIC size_t BIT_readBitsFast(BIT_DStream_t* bitD, unsigned nbBits)
+{
+    size_t const value = BIT_lookBitsFast(bitD, nbBits);
+    assert(nbBits >= 1);
+    BIT_skipBits(bitD, nbBits);
+    return value;
+}
+
+/*! BIT_reloadDStreamFast() :
+ *  Similar to BIT_reloadDStream(), but with two differences:
+ *  1. bitsConsumed <= sizeof(bitD->bitContainer)*8 must hold!
+ *  2. Returns BIT_DStream_overflow when bitD->ptr < bitD->limitPtr, at this
+ *     point you must use BIT_reloadDStream() to reload.
+ */
+MEM_STATIC BIT_DStream_status BIT_reloadDStreamFast(BIT_DStream_t* bitD)
+{
+    if (UNLIKELY(bitD->ptr < bitD->limitPtr))
+        return BIT_DStream_overflow;
+    assert(bitD->bitsConsumed <= sizeof(bitD->bitContainer)*8);
+    bitD->ptr -= bitD->bitsConsumed >> 3;
+    bitD->bitsConsumed &= 7;
+    bitD->bitContainer = MEM_readLEST(bitD->ptr);
+    return BIT_DStream_unfinished;
+}
+
+/*! 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.
+ *           when status == BIT_DStream_unfinished, internal register is filled with at least 25 or 57 bits */
+MEM_STATIC BIT_DStream_status BIT_reloadDStream(BIT_DStream_t* bitD)
+{
+    if (bitD->bitsConsumed > (sizeof(bitD->bitContainer)*8))  /* overflow detected, like end of stream */
+        return BIT_DStream_overflow;
+
+    if (bitD->ptr >= bitD->limitPtr) {
+        return BIT_reloadDStreamFast(bitD);
+    }
+    if (bitD->ptr == bitD->start) {
+        if (bitD->bitsConsumed < sizeof(bitD->bitContainer)*8) return BIT_DStream_endOfBuffer;
+        return BIT_DStream_completed;
+    }
+    /* start < ptr < limitPtr */
+    {   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 = MEM_readLEST(bitD->ptr);   /* reminder : srcSize > sizeof(bitD->bitContainer), otherwise bitD->ptr == bitD->start */
+        return result;
+    }
+}
+
+/*! BIT_endOfDStream() :
+ * @return : 1 if DStream has _exactly_ reached its end (all bits consumed).
+ */
+MEM_STATIC unsigned BIT_endOfDStream(const BIT_DStream_t* DStream)
+{
+    return ((DStream->ptr == DStream->start) && (DStream->bitsConsumed == sizeof(DStream->bitContainer)*8));
+}
+
+#if defined (__cplusplus)
+}
+#endif
+
+#endif /* BITSTREAM_H_MODULE */

external/zstd/common/compiler.h

@@ -0,0 +1,343 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_COMPILER_H
+#define ZSTD_COMPILER_H
+
+#include "portability_macros.h"
+
+/*-*******************************************************
+*  Compiler specifics
+*********************************************************/
+/* force inlining */
+
+#if !defined(ZSTD_NO_INLINE)
+#if (defined(__GNUC__) && !defined(__STRICT_ANSI__)) || defined(__cplusplus) || defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L   /* C99 */
+#  define INLINE_KEYWORD inline
+#else
+#  define INLINE_KEYWORD
+#endif
+
+#if defined(__GNUC__) || defined(__ICCARM__)
+#  define FORCE_INLINE_ATTR __attribute__((always_inline))
+#elif defined(_MSC_VER)
+#  define FORCE_INLINE_ATTR __forceinline
+#else
+#  define FORCE_INLINE_ATTR
+#endif
+
+#else
+
+#define INLINE_KEYWORD
+#define FORCE_INLINE_ATTR
+
+#endif
+
+/**
+  On MSVC qsort requires that functions passed into it use the __cdecl calling conversion(CC).
+  This explicitly marks such functions as __cdecl so that the code will still compile
+  if a CC other than __cdecl has been made the default.
+*/
+#if  defined(_MSC_VER)
+#  define WIN_CDECL __cdecl
+#else
+#  define WIN_CDECL
+#endif
+
+/**
+ * FORCE_INLINE_TEMPLATE is used to define C "templates", which take constant
+ * parameters. They must be inlined for the compiler to eliminate the constant
+ * branches.
+ */
+#define FORCE_INLINE_TEMPLATE static INLINE_KEYWORD FORCE_INLINE_ATTR
+/**
+ * HINT_INLINE is used to help the compiler generate better code. It is *not*
+ * used for "templates", so it can be tweaked based on the compilers
+ * performance.
+ *
+ * gcc-4.8 and gcc-4.9 have been shown to benefit from leaving off the
+ * always_inline attribute.
+ *
+ * clang up to 5.0.0 (trunk) benefit tremendously from the always_inline
+ * attribute.
+ */
+#if !defined(__clang__) && defined(__GNUC__) && __GNUC__ >= 4 && __GNUC_MINOR__ >= 8 && __GNUC__ < 5
+#  define HINT_INLINE static INLINE_KEYWORD
+#else
+#  define HINT_INLINE static INLINE_KEYWORD FORCE_INLINE_ATTR
+#endif
+
+/* UNUSED_ATTR tells the compiler it is okay if the function is unused. */
+#if defined(__GNUC__)
+#  define UNUSED_ATTR __attribute__((unused))
+#else
+#  define UNUSED_ATTR
+#endif
+
+/* force no inlining */
+#ifdef _MSC_VER
+#  define FORCE_NOINLINE static __declspec(noinline)
+#else
+#  if defined(__GNUC__) || defined(__ICCARM__)
+#    define FORCE_NOINLINE static __attribute__((__noinline__))
+#  else
+#    define FORCE_NOINLINE static
+#  endif
+#endif
+
+
+/* target attribute */
+#if defined(__GNUC__) || defined(__ICCARM__)
+#  define TARGET_ATTRIBUTE(target) __attribute__((__target__(target)))
+#else
+#  define TARGET_ATTRIBUTE(target)
+#endif
+
+/* Target attribute for BMI2 dynamic dispatch.
+ * Enable lzcnt, bmi, and bmi2.
+ * We test for bmi1 & bmi2. lzcnt is included in bmi1.
+ */
+#define BMI2_TARGET_ATTRIBUTE TARGET_ATTRIBUTE("lzcnt,bmi,bmi2")
+
+/* prefetch
+ * can be disabled, by declaring NO_PREFETCH build macro */
+#if defined(NO_PREFETCH)
+#  define PREFETCH_L1(ptr)  (void)(ptr)  /* disabled */
+#  define PREFETCH_L2(ptr)  (void)(ptr)  /* disabled */
+#else
+#  if defined(_MSC_VER) && (defined(_M_X64) || defined(_M_I86))  /* _mm_prefetch() is not defined outside of x86/x64 */
+#    include <mmintrin.h>   /* https://msdn.microsoft.com/fr-fr/library/84szxsww(v=vs.90).aspx */
+#    define PREFETCH_L1(ptr)  _mm_prefetch((const char*)(ptr), _MM_HINT_T0)
+#    define PREFETCH_L2(ptr)  _mm_prefetch((const char*)(ptr), _MM_HINT_T1)
+#  elif defined(__GNUC__) && ( (__GNUC__ >= 4) || ( (__GNUC__ == 3) && (__GNUC_MINOR__ >= 1) ) )
+#    define PREFETCH_L1(ptr)  __builtin_prefetch((ptr), 0 /* rw==read */, 3 /* locality */)
+#    define PREFETCH_L2(ptr)  __builtin_prefetch((ptr), 0 /* rw==read */, 2 /* locality */)
+#  elif defined(__aarch64__)
+#    define PREFETCH_L1(ptr)  __asm__ __volatile__("prfm pldl1keep, %0" ::"Q"(*(ptr)))
+#    define PREFETCH_L2(ptr)  __asm__ __volatile__("prfm pldl2keep, %0" ::"Q"(*(ptr)))
+#  else
+#    define PREFETCH_L1(ptr) (void)(ptr)  /* disabled */
+#    define PREFETCH_L2(ptr) (void)(ptr)  /* disabled */
+#  endif
+#endif  /* NO_PREFETCH */
+
+#define CACHELINE_SIZE 64
+
+#define PREFETCH_AREA(p, s)  {            \
+    const char* const _ptr = (const char*)(p);  \
+    size_t const _size = (size_t)(s);     \
+    size_t _pos;                          \
+    for (_pos=0; _pos<_size; _pos+=CACHELINE_SIZE) {  \
+        PREFETCH_L2(_ptr + _pos);         \
+    }                                     \
+}
+
+/* vectorization
+ * older GCC (pre gcc-4.3 picked as the cutoff) uses a different syntax,
+ * and some compilers, like Intel ICC and MCST LCC, do not support it at all. */
+#if !defined(__INTEL_COMPILER) && !defined(__clang__) && defined(__GNUC__) && !defined(__LCC__)
+#  if (__GNUC__ == 4 && __GNUC_MINOR__ > 3) || (__GNUC__ >= 5)
+#    define DONT_VECTORIZE __attribute__((optimize("no-tree-vectorize")))
+#  else
+#    define DONT_VECTORIZE _Pragma("GCC optimize(\"no-tree-vectorize\")")
+#  endif
+#else
+#  define DONT_VECTORIZE
+#endif
+
+/* Tell the compiler that a branch is likely or unlikely.
+ * Only use these macros if it causes the compiler to generate better code.
+ * If you can remove a LIKELY/UNLIKELY annotation without speed changes in gcc
+ * and clang, please do.
+ */
+#if defined(__GNUC__)
+#define LIKELY(x) (__builtin_expect((x), 1))
+#define UNLIKELY(x) (__builtin_expect((x), 0))
+#else
+#define LIKELY(x) (x)
+#define UNLIKELY(x) (x)
+#endif
+
+#if __has_builtin(__builtin_unreachable) || (defined(__GNUC__) && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 5)))
+#  define ZSTD_UNREACHABLE { assert(0), __builtin_unreachable(); }
+#else
+#  define ZSTD_UNREACHABLE { assert(0); }
+#endif
+
+/* disable warnings */
+#ifdef _MSC_VER    /* Visual Studio */
+#  include <intrin.h>                    /* For Visual 2005 */
+#  pragma warning(disable : 4100)        /* disable: C4100: unreferenced formal parameter */
+#  pragma warning(disable : 4127)        /* disable: C4127: conditional expression is constant */
+#  pragma warning(disable : 4204)        /* disable: C4204: non-constant aggregate initializer */
+#  pragma warning(disable : 4214)        /* disable: C4214: non-int bitfields */
+#  pragma warning(disable : 4324)        /* disable: C4324: padded structure */
+#endif
+
+/*Like DYNAMIC_BMI2 but for compile time determination of BMI2 support*/
+#ifndef STATIC_BMI2
+#  if defined(_MSC_VER) && (defined(_M_X64) || defined(_M_I86))
+#    ifdef __AVX2__  //MSVC does not have a BMI2 specific flag, but every CPU that supports AVX2 also supports BMI2
+#       define STATIC_BMI2 1
+#    endif
+#  elif defined(__BMI2__) && defined(__x86_64__) && defined(__GNUC__)
+#    define STATIC_BMI2 1
+#  endif
+#endif
+
+#ifndef STATIC_BMI2
+    #define STATIC_BMI2 0
+#endif
+
+/* compile time determination of SIMD support */
+#if !defined(ZSTD_NO_INTRINSICS)
+#  if defined(__SSE2__) || defined(_M_AMD64) || (defined (_M_IX86) && defined(_M_IX86_FP) && (_M_IX86_FP >= 2))
+#    define ZSTD_ARCH_X86_SSE2
+#  endif
+#  if defined(__ARM_NEON) || defined(_M_ARM64)
+#    define ZSTD_ARCH_ARM_NEON
+#  endif
+#
+#  if defined(ZSTD_ARCH_X86_SSE2)
+#    include <emmintrin.h>
+#  elif defined(ZSTD_ARCH_ARM_NEON)
+#    include <arm_neon.h>
+#  endif
+#endif
+
+/* C-language Attributes are added in C23. */
+#if defined(__STDC_VERSION__) && (__STDC_VERSION__ > 201710L) && defined(__has_c_attribute)
+# define ZSTD_HAS_C_ATTRIBUTE(x) __has_c_attribute(x)
+#else
+# define ZSTD_HAS_C_ATTRIBUTE(x) 0
+#endif
+
+/* Only use C++ attributes in C++. Some compilers report support for C++
+ * attributes when compiling with C.
+ */
+#if defined(__cplusplus) && defined(__has_cpp_attribute)
+# define ZSTD_HAS_CPP_ATTRIBUTE(x) __has_cpp_attribute(x)
+#else
+# define ZSTD_HAS_CPP_ATTRIBUTE(x) 0
+#endif
+
+/* Define ZSTD_FALLTHROUGH macro for annotating switch case with the 'fallthrough' attribute.
+ * - C23: https://en.cppreference.com/w/c/language/attributes/fallthrough
+ * - CPP17: https://en.cppreference.com/w/cpp/language/attributes/fallthrough
+ * - Else: __attribute__((__fallthrough__))
+ */
+#ifndef ZSTD_FALLTHROUGH
+# if ZSTD_HAS_C_ATTRIBUTE(fallthrough)
+#  define ZSTD_FALLTHROUGH [[fallthrough]]
+# elif ZSTD_HAS_CPP_ATTRIBUTE(fallthrough)
+#  define ZSTD_FALLTHROUGH [[fallthrough]]
+# elif __has_attribute(__fallthrough__)
+/* Leading semicolon is to satisfy gcc-11 with -pedantic. Without the semicolon
+ * gcc complains about: a label can only be part of a statement and a declaration is not a statement.
+ */
+#  define ZSTD_FALLTHROUGH ; __attribute__((__fallthrough__))
+# else
+#  define ZSTD_FALLTHROUGH
+# endif
+#endif
+
+/*-**************************************************************
+*  Alignment check
+*****************************************************************/
+
+/* this test was initially positioned in mem.h,
+ * but this file is removed (or replaced) for linux kernel
+ * so it's now hosted in compiler.h,
+ * which remains valid for both user & kernel spaces.
+ */
+
+#ifndef ZSTD_ALIGNOF
+# if defined(__GNUC__) || defined(_MSC_VER)
+/* covers gcc, clang & MSVC */
+/* note : this section must come first, before C11,
+ * due to a limitation in the kernel source generator */
+#  define ZSTD_ALIGNOF(T) __alignof(T)
+
+# elif defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L)
+/* C11 support */
+#  include <stdalign.h>
+#  define ZSTD_ALIGNOF(T) alignof(T)
+
+# else
+/* No known support for alignof() - imperfect backup */
+#  define ZSTD_ALIGNOF(T) (sizeof(void*) < sizeof(T) ? sizeof(void*) : sizeof(T))
+
+# endif
+#endif /* ZSTD_ALIGNOF */
+
+/*-**************************************************************
+*  Sanitizer
+*****************************************************************/
+
+#if ZSTD_MEMORY_SANITIZER
+/* Not all platforms that support msan provide sanitizers/msan_interface.h.
+ * We therefore declare the functions we need ourselves, rather than trying to
+ * include the header file... */
+#include <stddef.h>  /* size_t */
+#define ZSTD_DEPS_NEED_STDINT
+#include "zstd_deps.h"  /* intptr_t */
+
+/* Make memory region fully initialized (without changing its contents). */
+void __msan_unpoison(const volatile void *a, size_t size);
+
+/* Make memory region fully uninitialized (without changing its contents).
+   This is a legacy interface that does not update origin information. Use
+   __msan_allocated_memory() instead. */
+void __msan_poison(const volatile void *a, size_t size);
+
+/* Returns the offset of the first (at least partially) poisoned byte in the
+   memory range, or -1 if the whole range is good. */
+intptr_t __msan_test_shadow(const volatile void *x, size_t size);
+#endif
+
+#if ZSTD_ADDRESS_SANITIZER
+/* Not all platforms that support asan provide sanitizers/asan_interface.h.
+ * We therefore declare the functions we need ourselves, rather than trying to
+ * include the header file... */
+#include <stddef.h>  /* size_t */
+
+/**
+ * Marks a memory region (<c>[addr, addr+size)</c>) as unaddressable.
+ *
+ * This memory must be previously allocated by your program. Instrumented
+ * code is forbidden from accessing addresses in this region until it is
+ * unpoisoned. This function is not guaranteed to poison the entire region -
+ * it could poison only a subregion of <c>[addr, addr+size)</c> due to ASan
+ * alignment restrictions.
+ *
+ * \note This function is not thread-safe because no two threads can poison or
+ * unpoison memory in the same memory region simultaneously.
+ *
+ * \param addr Start of memory region.
+ * \param size Size of memory region. */
+void __asan_poison_memory_region(void const volatile *addr, size_t size);
+
+/**
+ * Marks a memory region (<c>[addr, addr+size)</c>) as addressable.
+ *
+ * This memory must be previously allocated by your program. Accessing
+ * addresses in this region is allowed until this region is poisoned again.
+ * This function could unpoison a super-region of <c>[addr, addr+size)</c> due
+ * to ASan alignment restrictions.
+ *
+ * \note This function is not thread-safe because no two threads can
+ * poison or unpoison memory in the same memory region simultaneously.
+ *
+ * \param addr Start of memory region.
+ * \param size Size of memory region. */
+void __asan_unpoison_memory_region(void const volatile *addr, size_t size);
+#endif
+
+#endif /* ZSTD_COMPILER_H */

external/zstd/common/cpu.h

@@ -0,0 +1,213 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_COMMON_CPU_H
+#define ZSTD_COMMON_CPU_H
+
+/**
+ * Implementation taken from folly/CpuId.h
+ * https://github.com/facebook/folly/blob/master/folly/CpuId.h
+ */
+
+#include "mem.h"
+
+#ifdef _MSC_VER
+#include <intrin.h>
+#endif
+
+typedef struct {
+    U32 f1c;
+    U32 f1d;
+    U32 f7b;
+    U32 f7c;
+} ZSTD_cpuid_t;
+
+MEM_STATIC ZSTD_cpuid_t ZSTD_cpuid(void) {
+    U32 f1c = 0;
+    U32 f1d = 0;
+    U32 f7b = 0;
+    U32 f7c = 0;
+#if defined(_MSC_VER) && (defined(_M_X64) || defined(_M_IX86))
+    int reg[4];
+    __cpuid((int*)reg, 0);
+    {
+        int const n = reg[0];
+        if (n >= 1) {
+            __cpuid((int*)reg, 1);
+            f1c = (U32)reg[2];
+            f1d = (U32)reg[3];
+        }
+        if (n >= 7) {
+            __cpuidex((int*)reg, 7, 0);
+            f7b = (U32)reg[1];
+            f7c = (U32)reg[2];
+        }
+    }
+#elif defined(__i386__) && defined(__PIC__) && !defined(__clang__) && defined(__GNUC__)
+    /* The following block like the normal cpuid branch below, but gcc
+     * reserves ebx for use of its pic register so we must specially
+     * handle the save and restore to avoid clobbering the register
+     */
+    U32 n;
+    __asm__(
+        "pushl %%ebx\n\t"
+        "cpuid\n\t"
+        "popl %%ebx\n\t"
+        : "=a"(n)
+        : "a"(0)
+        : "ecx", "edx");
+    if (n >= 1) {
+      U32 f1a;
+      __asm__(
+          "pushl %%ebx\n\t"
+          "cpuid\n\t"
+          "popl %%ebx\n\t"
+          : "=a"(f1a), "=c"(f1c), "=d"(f1d)
+          : "a"(1));
+    }
+    if (n >= 7) {
+      __asm__(
+          "pushl %%ebx\n\t"
+          "cpuid\n\t"
+          "movl %%ebx, %%eax\n\t"
+          "popl %%ebx"
+          : "=a"(f7b), "=c"(f7c)
+          : "a"(7), "c"(0)
+          : "edx");
+    }
+#elif defined(__x86_64__) || defined(_M_X64) || defined(__i386__)
+    U32 n;
+    __asm__("cpuid" : "=a"(n) : "a"(0) : "ebx", "ecx", "edx");
+    if (n >= 1) {
+      U32 f1a;
+      __asm__("cpuid" : "=a"(f1a), "=c"(f1c), "=d"(f1d) : "a"(1) : "ebx");
+    }
+    if (n >= 7) {
+      U32 f7a;
+      __asm__("cpuid"
+              : "=a"(f7a), "=b"(f7b), "=c"(f7c)
+              : "a"(7), "c"(0)
+              : "edx");
+    }
+#endif
+    {
+        ZSTD_cpuid_t cpuid;
+        cpuid.f1c = f1c;
+        cpuid.f1d = f1d;
+        cpuid.f7b = f7b;
+        cpuid.f7c = f7c;
+        return cpuid;
+    }
+}
+
+#define X(name, r, bit)                                                        \
+  MEM_STATIC int ZSTD_cpuid_##name(ZSTD_cpuid_t const cpuid) {                 \
+    return ((cpuid.r) & (1U << bit)) != 0;                                     \
+  }
+
+/* cpuid(1): Processor Info and Feature Bits. */
+#define C(name, bit) X(name, f1c, bit)
+  C(sse3, 0)
+  C(pclmuldq, 1)
+  C(dtes64, 2)
+  C(monitor, 3)
+  C(dscpl, 4)
+  C(vmx, 5)
+  C(smx, 6)
+  C(eist, 7)
+  C(tm2, 8)
+  C(ssse3, 9)
+  C(cnxtid, 10)
+  C(fma, 12)
+  C(cx16, 13)
+  C(xtpr, 14)
+  C(pdcm, 15)
+  C(pcid, 17)
+  C(dca, 18)
+  C(sse41, 19)
+  C(sse42, 20)
+  C(x2apic, 21)
+  C(movbe, 22)
+  C(popcnt, 23)
+  C(tscdeadline, 24)
+  C(aes, 25)
+  C(xsave, 26)
+  C(osxsave, 27)
+  C(avx, 28)
+  C(f16c, 29)
+  C(rdrand, 30)
+#undef C
+#define D(name, bit) X(name, f1d, bit)
+  D(fpu, 0)
+  D(vme, 1)
+  D(de, 2)
+  D(pse, 3)
+  D(tsc, 4)
+  D(msr, 5)
+  D(pae, 6)
+  D(mce, 7)
+  D(cx8, 8)
+  D(apic, 9)
+  D(sep, 11)
+  D(mtrr, 12)
+  D(pge, 13)
+  D(mca, 14)
+  D(cmov, 15)
+  D(pat, 16)
+  D(pse36, 17)
+  D(psn, 18)
+  D(clfsh, 19)
+  D(ds, 21)
+  D(acpi, 22)
+  D(mmx, 23)
+  D(fxsr, 24)
+  D(sse, 25)
+  D(sse2, 26)
+  D(ss, 27)
+  D(htt, 28)
+  D(tm, 29)
+  D(pbe, 31)
+#undef D
+
+/* cpuid(7): Extended Features. */
+#define B(name, bit) X(name, f7b, bit)
+  B(bmi1, 3)
+  B(hle, 4)
+  B(avx2, 5)
+  B(smep, 7)
+  B(bmi2, 8)
+  B(erms, 9)
+  B(invpcid, 10)
+  B(rtm, 11)
+  B(mpx, 14)
+  B(avx512f, 16)
+  B(avx512dq, 17)
+  B(rdseed, 18)
+  B(adx, 19)
+  B(smap, 20)
+  B(avx512ifma, 21)
+  B(pcommit, 22)
+  B(clflushopt, 23)
+  B(clwb, 24)
+  B(avx512pf, 26)
+  B(avx512er, 27)
+  B(avx512cd, 28)
+  B(sha, 29)
+  B(avx512bw, 30)
+  B(avx512vl, 31)
+#undef B
+#define C(name, bit) X(name, f7c, bit)
+  C(prefetchwt1, 0)
+  C(avx512vbmi, 1)
+#undef C
+
+#undef X
+
+#endif /* ZSTD_COMMON_CPU_H */

external/zstd/common/debug.c

@@ -0,0 +1,24 @@
+/* ******************************************************************
+ * debug
+ * Part of FSE library
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ *
+ * You can contact the author at :
+ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+
+
+/*
+ * This module only hosts one global variable
+ * which can be used to dynamically influence the verbosity of traces,
+ * such as DEBUGLOG and RAWLOG
+ */
+
+#include "debug.h"
+
+int g_debuglevel = DEBUGLEVEL;

external/zstd/common/debug.h

@@ -0,0 +1,107 @@
+/* ******************************************************************
+ * debug
+ * Part of FSE library
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ *
+ * You can contact the author at :
+ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+
+
+/*
+ * The purpose of this header is to enable debug functions.
+ * They regroup assert(), DEBUGLOG() and RAWLOG() for run-time,
+ * and DEBUG_STATIC_ASSERT() for compile-time.
+ *
+ * By default, DEBUGLEVEL==0, which means run-time debug is disabled.
+ *
+ * Level 1 enables assert() only.
+ * Starting level 2, traces can be generated and pushed to stderr.
+ * The higher the level, the more verbose the traces.
+ *
+ * It's possible to dynamically adjust level using variable g_debug_level,
+ * which is only declared if DEBUGLEVEL>=2,
+ * and is a global variable, not multi-thread protected (use with care)
+ */
+
+#ifndef DEBUG_H_12987983217
+#define DEBUG_H_12987983217
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+
+/* static assert is triggered at compile time, leaving no runtime artefact.
+ * static assert only works with compile-time constants.
+ * Also, this variant can only be used inside a function. */
+#define DEBUG_STATIC_ASSERT(c) (void)sizeof(char[(c) ? 1 : -1])
+
+
+/* DEBUGLEVEL is expected to be defined externally,
+ * typically through compiler command line.
+ * Value must be a number. */
+#ifndef DEBUGLEVEL
+#  define DEBUGLEVEL 0
+#endif
+
+
+/* recommended values for DEBUGLEVEL :
+ * 0 : release mode, no debug, all run-time checks disabled
+ * 1 : enables assert() only, no display
+ * 2 : reserved, for currently active debug path
+ * 3 : events once per object lifetime (CCtx, CDict, etc.)
+ * 4 : events once per frame
+ * 5 : events once per block
+ * 6 : events once per sequence (verbose)
+ * 7+: events at every position (*very* verbose)
+ *
+ * It's generally inconvenient to output traces > 5.
+ * In which case, it's possible to selectively trigger high verbosity levels
+ * by modifying g_debug_level.
+ */
+
+#if (DEBUGLEVEL>=1)
+#  define ZSTD_DEPS_NEED_ASSERT
+#  include "zstd_deps.h"
+#else
+#  ifndef assert   /* assert may be already defined, due to prior #include <assert.h> */
+#    define assert(condition) ((void)0)   /* disable assert (default) */
+#  endif
+#endif
+
+#if (DEBUGLEVEL>=2)
+#  define ZSTD_DEPS_NEED_IO
+#  include "zstd_deps.h"
+extern int g_debuglevel; /* the variable is only declared,
+                            it actually lives in debug.c,
+                            and is shared by the whole process.
+                            It's not thread-safe.
+                            It's useful when enabling very verbose levels
+                            on selective conditions (such as position in src) */
+
+#  define RAWLOG(l, ...) {                                       \
+                if (l<=g_debuglevel) {                           \
+                    ZSTD_DEBUG_PRINT(__VA_ARGS__);               \
+            }   }
+#  define DEBUGLOG(l, ...) {                                     \
+                if (l<=g_debuglevel) {                           \
+                    ZSTD_DEBUG_PRINT(__FILE__ ": " __VA_ARGS__); \
+                    ZSTD_DEBUG_PRINT(" \n");                     \
+            }   }
+#else
+#  define RAWLOG(l, ...)      {}    /* disabled */
+#  define DEBUGLOG(l, ...)    {}    /* disabled */
+#endif
+
+
+#if defined (__cplusplus)
+}
+#endif
+
+#endif /* DEBUG_H_12987983217 */

external/zstd/common/entropy_common.c

@@ -0,0 +1,341 @@
+/* ******************************************************************
+ * Common functions of New Generation Entropy library
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ *
+ *  You can contact the author at :
+ *  - FSE+HUF source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *  - Public forum : https://groups.google.com/forum/#!forum/lz4c
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+
+/* *************************************
+*  Dependencies
+***************************************/
+#include "mem.h"
+#include "error_private.h"       /* ERR_*, ERROR */
+#define FSE_STATIC_LINKING_ONLY  /* FSE_MIN_TABLELOG */
+#include "fse.h"
+#define HUF_STATIC_LINKING_ONLY  /* HUF_TABLELOG_ABSOLUTEMAX */
+#include "huf.h"
+#include "bits.h"                /* ZSDT_highbit32, ZSTD_countTrailingZeros32 */
+
+
+/*===   Version   ===*/
+unsigned FSE_versionNumber(void) { return FSE_VERSION_NUMBER; }
+
+
+/*===   Error Management   ===*/
+unsigned FSE_isError(size_t code) { return ERR_isError(code); }
+const char* FSE_getErrorName(size_t code) { return ERR_getErrorName(code); }
+
+unsigned HUF_isError(size_t code) { return ERR_isError(code); }
+const char* HUF_getErrorName(size_t code) { return ERR_getErrorName(code); }
+
+
+/*-**************************************************************
+*  FSE NCount encoding-decoding
+****************************************************************/
+FORCE_INLINE_TEMPLATE
+size_t FSE_readNCount_body(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;
+    unsigned const maxSV1 = *maxSVPtr + 1;
+    int previous0 = 0;
+
+    if (hbSize < 8) {
+        /* This function only works when hbSize >= 8 */
+        char buffer[8] = {0};
+        ZSTD_memcpy(buffer, headerBuffer, hbSize);
+        {   size_t const countSize = FSE_readNCount(normalizedCounter, maxSVPtr, tableLogPtr,
+                                                    buffer, sizeof(buffer));
+            if (FSE_isError(countSize)) return countSize;
+            if (countSize > hbSize) return ERROR(corruption_detected);
+            return countSize;
+    }   }
+    assert(hbSize >= 8);
+
+    /* init */
+    ZSTD_memset(normalizedCounter, 0, (*maxSVPtr+1) * sizeof(normalizedCounter[0]));   /* all symbols not present in NCount have a frequency of 0 */
+    bitStream = MEM_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++;
+
+    for (;;) {
+        if (previous0) {
+            /* Count the number of repeats. Each time the
+             * 2-bit repeat code is 0b11 there is another
+             * repeat.
+             * Avoid UB by setting the high bit to 1.
+             */
+            int repeats = ZSTD_countTrailingZeros32(~bitStream | 0x80000000) >> 1;
+            while (repeats >= 12) {
+                charnum += 3 * 12;
+                if (LIKELY(ip <= iend-7)) {
+                    ip += 3;
+                } else {
+                    bitCount -= (int)(8 * (iend - 7 - ip));
+                    bitCount &= 31;
+                    ip = iend - 4;
+                }
+                bitStream = MEM_readLE32(ip) >> bitCount;
+                repeats = ZSTD_countTrailingZeros32(~bitStream | 0x80000000) >> 1;
+            }
+            charnum += 3 * repeats;
+            bitStream >>= 2 * repeats;
+            bitCount += 2 * repeats;
+
+            /* Add the final repeat which isn't 0b11. */
+            assert((bitStream & 3) < 3);
+            charnum += bitStream & 3;
+            bitCount += 2;
+
+            /* This is an error, but break and return an error
+             * at the end, because returning out of a loop makes
+             * it harder for the compiler to optimize.
+             */
+            if (charnum >= maxSV1) break;
+
+            /* We don't need to set the normalized count to 0
+             * because we already memset the whole buffer to 0.
+             */
+
+            if (LIKELY(ip <= iend-7) || (ip + (bitCount>>3) <= iend-4)) {
+                assert((bitCount >> 3) <= 3); /* For first condition to work */
+                ip += bitCount>>3;
+                bitCount &= 7;
+            } else {
+                bitCount -= (int)(8 * (iend - 4 - ip));
+                bitCount &= 31;
+                ip = iend - 4;
+            }
+            bitStream = MEM_readLE32(ip) >> bitCount;
+        }
+        {
+            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 */
+            /* When it matters (small blocks), this is a
+             * predictable branch, because we don't use -1.
+             */
+            if (count >= 0) {
+                remaining -= count;
+            } else {
+                assert(count == -1);
+                remaining += count;
+            }
+            normalizedCounter[charnum++] = (short)count;
+            previous0 = !count;
+
+            assert(threshold > 1);
+            if (remaining < threshold) {
+                /* This branch can be folded into the
+                 * threshold update condition because we
+                 * know that threshold > 1.
+                 */
+                if (remaining <= 1) break;
+                nbBits = ZSTD_highbit32(remaining) + 1;
+                threshold = 1 << (nbBits - 1);
+            }
+            if (charnum >= maxSV1) break;
+
+            if (LIKELY(ip <= iend-7) || (ip + (bitCount>>3) <= iend-4)) {
+                ip += bitCount>>3;
+                bitCount &= 7;
+            } else {
+                bitCount -= (int)(8 * (iend - 4 - ip));
+                bitCount &= 31;
+                ip = iend - 4;
+            }
+            bitStream = MEM_readLE32(ip) >> bitCount;
+    }   }
+    if (remaining != 1) return ERROR(corruption_detected);
+    /* Only possible when there are too many zeros. */
+    if (charnum > maxSV1) return ERROR(maxSymbolValue_tooSmall);
+    if (bitCount > 32) return ERROR(corruption_detected);
+    *maxSVPtr = charnum-1;
+
+    ip += (bitCount+7)>>3;
+    return ip-istart;
+}
+
+/* Avoids the FORCE_INLINE of the _body() function. */
+static size_t FSE_readNCount_body_default(
+        short* normalizedCounter, unsigned* maxSVPtr, unsigned* tableLogPtr,
+        const void* headerBuffer, size_t hbSize)
+{
+    return FSE_readNCount_body(normalizedCounter, maxSVPtr, tableLogPtr, headerBuffer, hbSize);
+}
+
+#if DYNAMIC_BMI2
+BMI2_TARGET_ATTRIBUTE static size_t FSE_readNCount_body_bmi2(
+        short* normalizedCounter, unsigned* maxSVPtr, unsigned* tableLogPtr,
+        const void* headerBuffer, size_t hbSize)
+{
+    return FSE_readNCount_body(normalizedCounter, maxSVPtr, tableLogPtr, headerBuffer, hbSize);
+}
+#endif
+
+size_t FSE_readNCount_bmi2(
+        short* normalizedCounter, unsigned* maxSVPtr, unsigned* tableLogPtr,
+        const void* headerBuffer, size_t hbSize, int bmi2)
+{
+#if DYNAMIC_BMI2
+    if (bmi2) {
+        return FSE_readNCount_body_bmi2(normalizedCounter, maxSVPtr, tableLogPtr, headerBuffer, hbSize);
+    }
+#endif
+    (void)bmi2;
+    return FSE_readNCount_body_default(normalizedCounter, maxSVPtr, tableLogPtr, headerBuffer, hbSize);
+}
+
+size_t FSE_readNCount(
+        short* normalizedCounter, unsigned* maxSVPtr, unsigned* tableLogPtr,
+        const void* headerBuffer, size_t hbSize)
+{
+    return FSE_readNCount_bmi2(normalizedCounter, maxSVPtr, tableLogPtr, headerBuffer, hbSize, /* bmi2 */ 0);
+}
+
+
+/*! 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(BYTE* huffWeight, size_t hwSize, U32* rankStats,
+                     U32* nbSymbolsPtr, U32* tableLogPtr,
+                     const void* src, size_t srcSize)
+{
+    U32 wksp[HUF_READ_STATS_WORKSPACE_SIZE_U32];
+    return HUF_readStats_wksp(huffWeight, hwSize, rankStats, nbSymbolsPtr, tableLogPtr, src, srcSize, wksp, sizeof(wksp), /* bmi2 */ 0);
+}
+
+FORCE_INLINE_TEMPLATE size_t
+HUF_readStats_body(BYTE* huffWeight, size_t hwSize, U32* rankStats,
+                   U32* nbSymbolsPtr, U32* tableLogPtr,
+                   const void* src, size_t srcSize,
+                   void* workSpace, size_t wkspSize,
+                   int bmi2)
+{
+    U32 weightTotal;
+    const BYTE* ip = (const BYTE*) src;
+    size_t iSize;
+    size_t oSize;
+
+    if (!srcSize) return ERROR(srcSize_wrong);
+    iSize = ip[0];
+    /* ZSTD_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);
+        /* max (hwSize-1) values decoded, as last one is implied */
+        oSize = FSE_decompress_wksp_bmi2(huffWeight, hwSize-1, ip+1, iSize, 6, workSpace, wkspSize, bmi2);
+        if (FSE_isError(oSize)) return oSize;
+    }
+
+    /* collect weight stats */
+    ZSTD_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 = ZSTD_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 << ZSTD_highbit32(rest);
+            U32 const lastWeight = ZSTD_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;
+}
+
+/* Avoids the FORCE_INLINE of the _body() function. */
+static size_t HUF_readStats_body_default(BYTE* huffWeight, size_t hwSize, U32* rankStats,
+                     U32* nbSymbolsPtr, U32* tableLogPtr,
+                     const void* src, size_t srcSize,
+                     void* workSpace, size_t wkspSize)
+{
+    return HUF_readStats_body(huffWeight, hwSize, rankStats, nbSymbolsPtr, tableLogPtr, src, srcSize, workSpace, wkspSize, 0);
+}
+
+#if DYNAMIC_BMI2
+static BMI2_TARGET_ATTRIBUTE size_t HUF_readStats_body_bmi2(BYTE* huffWeight, size_t hwSize, U32* rankStats,
+                     U32* nbSymbolsPtr, U32* tableLogPtr,
+                     const void* src, size_t srcSize,
+                     void* workSpace, size_t wkspSize)
+{
+    return HUF_readStats_body(huffWeight, hwSize, rankStats, nbSymbolsPtr, tableLogPtr, src, srcSize, workSpace, wkspSize, 1);
+}
+#endif
+
+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 wkspSize,
+                     int bmi2)
+{
+#if DYNAMIC_BMI2
+    if (bmi2) {
+        return HUF_readStats_body_bmi2(huffWeight, hwSize, rankStats, nbSymbolsPtr, tableLogPtr, src, srcSize, workSpace, wkspSize);
+    }
+#endif
+    (void)bmi2;
+    return HUF_readStats_body_default(huffWeight, hwSize, rankStats, nbSymbolsPtr, tableLogPtr, src, srcSize, workSpace, wkspSize);
+}

external/zstd/common/error_private.c

@@ -0,0 +1,60 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+/* The purpose of this file is to have a single list of error strings embedded in binary */
+
+#include "error_private.h"
+
+const char* ERR_getErrorString(ERR_enum code)
+{
+#ifdef ZSTD_STRIP_ERROR_STRINGS
+    (void)code;
+    return "Error strings stripped";
+#else
+    static const char* const notErrorCode = "Unspecified error code";
+    switch( code )
+    {
+    case PREFIX(no_error): return "No error detected";
+    case PREFIX(GENERIC):  return "Error (generic)";
+    case PREFIX(prefix_unknown): return "Unknown frame descriptor";
+    case PREFIX(version_unsupported): return "Version not supported";
+    case PREFIX(frameParameter_unsupported): return "Unsupported frame parameter";
+    case PREFIX(frameParameter_windowTooLarge): return "Frame requires too much memory for decoding";
+    case PREFIX(corruption_detected): return "Data corruption detected";
+    case PREFIX(checksum_wrong): return "Restored data doesn't match checksum";
+    case PREFIX(literals_headerWrong): return "Header of Literals' block doesn't respect format specification";
+    case PREFIX(parameter_unsupported): return "Unsupported parameter";
+    case PREFIX(parameter_combination_unsupported): return "Unsupported combination of parameters";
+    case PREFIX(parameter_outOfBound): return "Parameter is out of bound";
+    case PREFIX(init_missing): return "Context should be init first";
+    case PREFIX(memory_allocation): return "Allocation error : not enough memory";
+    case PREFIX(workSpace_tooSmall): return "workSpace buffer is not large enough";
+    case PREFIX(stage_wrong): return "Operation not authorized at current processing stage";
+    case PREFIX(tableLog_tooLarge): return "tableLog requires too much memory : unsupported";
+    case PREFIX(maxSymbolValue_tooLarge): return "Unsupported max Symbol Value : too large";
+    case PREFIX(maxSymbolValue_tooSmall): return "Specified maxSymbolValue is too small";
+    case PREFIX(stabilityCondition_notRespected): return "pledged buffer stability condition is not respected";
+    case PREFIX(dictionary_corrupted): return "Dictionary is corrupted";
+    case PREFIX(dictionary_wrong): return "Dictionary mismatch";
+    case PREFIX(dictionaryCreation_failed): return "Cannot create Dictionary from provided samples";
+    case PREFIX(dstSize_tooSmall): return "Destination buffer is too small";
+    case PREFIX(srcSize_wrong): return "Src size is incorrect";
+    case PREFIX(dstBuffer_null): return "Operation on NULL destination buffer";
+        /* following error codes are not stable and may be removed or changed in a future version */
+    case PREFIX(frameIndex_tooLarge): return "Frame index is too large";
+    case PREFIX(seekableIO): return "An I/O error occurred when reading/seeking";
+    case PREFIX(dstBuffer_wrong): return "Destination buffer is wrong";
+    case PREFIX(srcBuffer_wrong): return "Source buffer is wrong";
+    case PREFIX(externalMatchFinder_failed): return "External matchfinder returned an error code";
+    case PREFIX(maxCode):
+    default: return notErrorCode;
+    }
+#endif
+}

external/zstd/common/error_private.h

@@ -0,0 +1,159 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+/* Note : this module is expected to remain private, do not expose it */
+
+#ifndef ERROR_H_MODULE
+#define ERROR_H_MODULE
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+
+/* ****************************************
+*  Dependencies
+******************************************/
+#include "../zstd_errors.h"  /* enum list */
+#include "compiler.h"
+#include "debug.h"
+#include "zstd_deps.h"       /* size_t */
+
+
+/* ****************************************
+*  Compiler-specific
+******************************************/
+#if defined(__GNUC__)
+#  define ERR_STATIC static __attribute__((unused))
+#elif defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */)
+#  define ERR_STATIC static inline
+#elif defined(_MSC_VER)
+#  define ERR_STATIC static __inline
+#else
+#  define ERR_STATIC static  /* this version may generate warnings for unused static functions; disable the relevant warning */
+#endif
+
+
+/*-****************************************
+*  Customization (error_public.h)
+******************************************/
+typedef ZSTD_ErrorCode ERR_enum;
+#define PREFIX(name) ZSTD_error_##name
+
+
+/*-****************************************
+*  Error codes handling
+******************************************/
+#undef ERROR   /* already defined on Visual Studio */
+#define ERROR(name) ZSTD_ERROR(name)
+#define ZSTD_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); }
+
+/* check and forward error code */
+#define CHECK_V_F(e, f) size_t const e = f; if (ERR_isError(e)) return e
+#define CHECK_F(f)   { CHECK_V_F(_var_err__, f); }
+
+
+/*-****************************************
+*  Error Strings
+******************************************/
+
+const char* ERR_getErrorString(ERR_enum code);   /* error_private.c */
+
+ERR_STATIC const char* ERR_getErrorName(size_t code)
+{
+    return ERR_getErrorString(ERR_getErrorCode(code));
+}
+
+/**
+ * Ignore: this is an internal helper.
+ *
+ * This is a helper function to help force C99-correctness during compilation.
+ * Under strict compilation modes, variadic macro arguments can't be empty.
+ * However, variadic function arguments can be. Using a function therefore lets
+ * us statically check that at least one (string) argument was passed,
+ * independent of the compilation flags.
+ */
+static INLINE_KEYWORD UNUSED_ATTR
+void _force_has_format_string(const char *format, ...) {
+  (void)format;
+}
+
+/**
+ * Ignore: this is an internal helper.
+ *
+ * We want to force this function invocation to be syntactically correct, but
+ * we don't want to force runtime evaluation of its arguments.
+ */
+#define _FORCE_HAS_FORMAT_STRING(...) \
+  if (0) { \
+    _force_has_format_string(__VA_ARGS__); \
+  }
+
+#define ERR_QUOTE(str) #str
+
+/**
+ * Return the specified error if the condition evaluates to true.
+ *
+ * In debug modes, prints additional information.
+ * In order to do that (particularly, printing the conditional that failed),
+ * this can't just wrap RETURN_ERROR().
+ */
+#define RETURN_ERROR_IF(cond, err, ...) \
+  if (cond) { \
+    RAWLOG(3, "%s:%d: ERROR!: check %s failed, returning %s", \
+           __FILE__, __LINE__, ERR_QUOTE(cond), ERR_QUOTE(ERROR(err))); \
+    _FORCE_HAS_FORMAT_STRING(__VA_ARGS__); \
+    RAWLOG(3, ": " __VA_ARGS__); \
+    RAWLOG(3, "\n"); \
+    return ERROR(err); \
+  }
+
+/**
+ * Unconditionally return the specified error.
+ *
+ * In debug modes, prints additional information.
+ */
+#define RETURN_ERROR(err, ...) \
+  do { \
+    RAWLOG(3, "%s:%d: ERROR!: unconditional check failed, returning %s", \
+           __FILE__, __LINE__, ERR_QUOTE(ERROR(err))); \
+    _FORCE_HAS_FORMAT_STRING(__VA_ARGS__); \
+    RAWLOG(3, ": " __VA_ARGS__); \
+    RAWLOG(3, "\n"); \
+    return ERROR(err); \
+  } while(0);
+
+/**
+ * If the provided expression evaluates to an error code, returns that error code.
+ *
+ * In debug modes, prints additional information.
+ */
+#define FORWARD_IF_ERROR(err, ...) \
+  do { \
+    size_t const err_code = (err); \
+    if (ERR_isError(err_code)) { \
+      RAWLOG(3, "%s:%d: ERROR!: forwarding error in %s: %s", \
+             __FILE__, __LINE__, ERR_QUOTE(err), ERR_getErrorName(err_code)); \
+      _FORCE_HAS_FORMAT_STRING(__VA_ARGS__); \
+      RAWLOG(3, ": " __VA_ARGS__); \
+      RAWLOG(3, "\n"); \
+      return err_code; \
+    } \
+  } while(0);
+
+#if defined (__cplusplus)
+}
+#endif
+
+#endif /* ERROR_H_MODULE */

external/zstd/common/fse.h

@@ -0,0 +1,717 @@
+/* ******************************************************************
+ * FSE : Finite State Entropy codec
+ * Public Prototypes declaration
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ *
+ * You can contact the author at :
+ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+#ifndef FSE_H
+#define FSE_H
+
+
+/*-*****************************************
+*  Dependencies
+******************************************/
+#include "zstd_deps.h"    /* size_t, ptrdiff_t */
+
+
+/*-*****************************************
+*  FSE_PUBLIC_API : control library symbols visibility
+******************************************/
+#if defined(FSE_DLL_EXPORT) && (FSE_DLL_EXPORT==1) && defined(__GNUC__) && (__GNUC__ >= 4)
+#  define FSE_PUBLIC_API __attribute__ ((visibility ("default")))
+#elif defined(FSE_DLL_EXPORT) && (FSE_DLL_EXPORT==1)   /* Visual expected */
+#  define FSE_PUBLIC_API __declspec(dllexport)
+#elif defined(FSE_DLL_IMPORT) && (FSE_DLL_IMPORT==1)
+#  define FSE_PUBLIC_API __declspec(dllimport) /* It isn't required but allows to generate better code, saving a function pointer load from the IAT and an indirect jump.*/
+#else
+#  define FSE_PUBLIC_API
+#endif
+
+/*------   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 */
+
+
+/*-****************************************
+*  FSE simple functions
+******************************************/
+/*! FSE_compress() :
+    Compress content of buffer 'src', of size 'srcSize', into destination buffer 'dst'.
+    'dst' buffer must be already allocated. Compression runs faster is dstCapacity >= FSE_compressBound(srcSize).
+    @return : size of compressed data (<= dstCapacity).
+    Special values : if return == 0, srcData is not compressible => Nothing is stored within dst !!!
+                     if return == 1, srcData is a single byte symbol * srcSize times. Use RLE compression instead.
+                     if FSE_isError(return), compression failed (more details using FSE_getErrorName())
+*/
+FSE_PUBLIC_API size_t FSE_compress(void* dst, size_t dstCapacity,
+                             const void* src, size_t srcSize);
+
+/*! FSE_decompress():
+    Decompress FSE data from buffer 'cSrc', of size 'cSrcSize',
+    into already allocated destination buffer 'dst', of size 'dstCapacity'.
+    @return : size of regenerated data (<= maxDstSize),
+              or an error code, which can be tested using FSE_isError() .
+
+    ** Important ** : FSE_decompress() does not decompress non-compressible nor RLE data !!!
+    Why ? : making this distinction requires a header.
+    Header management is intentionally delegated to the user layer, which can better manage special cases.
+*/
+FSE_PUBLIC_API size_t FSE_decompress(void* dst,  size_t dstCapacity,
+                               const void* cSrc, size_t cSrcSize);
+
+
+/*-*****************************************
+*  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_PUBLIC_API const char* FSE_getErrorName(size_t code);   /* provides error code string (useful for debugging) */
+
+
+/*-*****************************************
+*  FSE advanced functions
+******************************************/
+/*! FSE_compress2() :
+    Same as FSE_compress(), but allows the selection of 'maxSymbolValue' and 'tableLog'
+    Both parameters can be defined as '0' to mean : use default value
+    @return : size of compressed data
+    Special values : if return == 0, srcData is not compressible => Nothing is stored within cSrc !!!
+                     if return == 1, srcData is a single byte symbol * srcSize times. Use RLE compression.
+                     if FSE_isError(return), it's an error code.
+*/
+FSE_PUBLIC_API size_t FSE_compress2 (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog);
+
+
+/*-*****************************************
+*  FSE detailed API
+******************************************/
+/*!
+FSE_compress() does the following:
+1. count symbol occurrence from source[] into table count[] (see hist.h)
+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).
+    useLowProbCount is a boolean parameter which trades off compressed size for
+    faster header decoding. When it is set to 1, the compressed data will be slightly
+    smaller. And when it is set to 0, FSE_readNCount() and FSE_buildDTable() will be
+    faster. If you are compressing a small amount of data (< 2 KB) then useLowProbCount=0
+    is a good default, since header deserialization makes a big speed difference.
+    Otherwise, useLowProbCount=1 is a good default, since the speed difference is small.
+    @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, unsigned useLowProbCount);
+
+/*! 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_PUBLIC_API FSE_CTable* FSE_createCTable (unsigned maxSymbolValue, unsigned tableLog);
+FSE_PUBLIC_API void        FSE_freeCTable (FSE_CTable* ct);
+
+/*! FSE_buildCTable():
+    Builds `ct`, which must be already allocated, using FSE_createCTable().
+    @return : 0, or an errorCode, which can be tested using FSE_isError() */
+FSE_PUBLIC_API size_t FSE_buildCTable(FSE_CTable* ct, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog);
+
+/*! 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);
+
+/*! FSE_readNCount_bmi2():
+ * Same as FSE_readNCount() but pass bmi2=1 when your CPU supports BMI2 and 0 otherwise.
+ */
+FSE_PUBLIC_API size_t FSE_readNCount_bmi2(short* normalizedCounter,
+                           unsigned* maxSymbolValuePtr, unsigned* tableLogPtr,
+                           const void* rBuffer, size_t rBuffSize, int bmi2);
+
+/*! 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_PUBLIC_API FSE_DTable* FSE_createDTable(unsigned tableLog);
+FSE_PUBLIC_API void        FSE_freeDTable(FSE_DTable* dt);
+
+/*! 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 (FSE_DTable* dt, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog);
+
+/*! 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)
+*/
+
+#endif  /* FSE_H */
+
+#if defined(FSE_STATIC_LINKING_ONLY) && !defined(FSE_H_FSE_STATIC_LINKING_ONLY)
+#define FSE_H_FSE_STATIC_LINKING_ONLY
+
+/* *** Dependency *** */
+#include "bitstream.h"
+
+
+/* *****************************************
+*  Static allocation
+*******************************************/
+/* FSE buffer bounds */
+#define FSE_NCOUNTBOUND 512
+#define FSE_BLOCKBOUND(size) ((size) + ((size)>>7) + 4 /* fse states */ + sizeof(size_t) /* bitContainer */)
+#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)))
+
+/* or use the size to malloc() space directly. Pay attention to alignment restrictions though */
+#define FSE_CTABLE_SIZE(maxTableLog, maxSymbolValue)   (FSE_CTABLE_SIZE_U32(maxTableLog, maxSymbolValue) * sizeof(FSE_CTable))
+#define FSE_DTABLE_SIZE(maxTableLog)                   (FSE_DTABLE_SIZE_U32(maxTableLog) * sizeof(FSE_DTable))
+
+
+/* *****************************************
+ *  FSE advanced API
+ ***************************************** */
+
+unsigned FSE_optimalTableLog_internal(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue, unsigned minus);
+/**< same as FSE_optimalTableLog(), which used `minus==2` */
+
+/* FSE_compress_wksp() :
+ * Same as FSE_compress2(), but using an externally allocated scratch buffer (`workSpace`).
+ * FSE_COMPRESS_WKSP_SIZE_U32() provides the minimum size required for `workSpace` as a table of FSE_CTable.
+ */
+#define FSE_COMPRESS_WKSP_SIZE_U32(maxTableLog, maxSymbolValue)   ( FSE_CTABLE_SIZE_U32(maxTableLog, maxSymbolValue) + ((maxTableLog > 12) ? (1 << (maxTableLog - 2)) : 1024) )
+size_t FSE_compress_wksp (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog, void* workSpace, size_t wkspSize);
+
+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 >= `FSE_BUILD_CTABLE_WORKSPACE_SIZE_U32(maxSymbolValue, tableLog)` of `unsigned`.
+ * See FSE_buildCTable_wksp() for breakdown of workspace usage.
+ */
+#define FSE_BUILD_CTABLE_WORKSPACE_SIZE_U32(maxSymbolValue, tableLog) (((maxSymbolValue + 2) + (1ull << (tableLog)))/2 + sizeof(U64)/sizeof(U32) /* additional 8 bytes for potential table overwrite */)
+#define FSE_BUILD_CTABLE_WORKSPACE_SIZE(maxSymbolValue, tableLog) (sizeof(unsigned) * FSE_BUILD_CTABLE_WORKSPACE_SIZE_U32(maxSymbolValue, tableLog))
+size_t FSE_buildCTable_wksp(FSE_CTable* ct, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog, void* workSpace, size_t wkspSize);
+
+#define FSE_BUILD_DTABLE_WKSP_SIZE(maxTableLog, maxSymbolValue) (sizeof(short) * (maxSymbolValue + 1) + (1ULL << maxTableLog) + 8)
+#define FSE_BUILD_DTABLE_WKSP_SIZE_U32(maxTableLog, maxSymbolValue) ((FSE_BUILD_DTABLE_WKSP_SIZE(maxTableLog, maxSymbolValue) + sizeof(unsigned) - 1) / sizeof(unsigned))
+FSE_PUBLIC_API size_t FSE_buildDTable_wksp(FSE_DTable* dt, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog, void* workSpace, size_t wkspSize);
+/**< Same as FSE_buildDTable(), using an externally allocated `workspace` produced with `FSE_BUILD_DTABLE_WKSP_SIZE_U32(maxSymbolValue)` */
+
+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 */
+
+#define FSE_DECOMPRESS_WKSP_SIZE_U32(maxTableLog, maxSymbolValue) (FSE_DTABLE_SIZE_U32(maxTableLog) + 1 + FSE_BUILD_DTABLE_WKSP_SIZE_U32(maxTableLog, maxSymbolValue) + (FSE_MAX_SYMBOL_VALUE + 1) / 2 + 1)
+#define FSE_DECOMPRESS_WKSP_SIZE(maxTableLog, maxSymbolValue) (FSE_DECOMPRESS_WKSP_SIZE_U32(maxTableLog, maxSymbolValue) * sizeof(unsigned))
+size_t FSE_decompress_wksp(void* dst, size_t dstCapacity, const void* cSrc, size_t cSrcSize, unsigned maxLog, void* workSpace, size_t wkspSize);
+/**< same as FSE_decompress(), using an externally allocated `workSpace` produced with `FSE_DECOMPRESS_WKSP_SIZE_U32(maxLog, maxSymbolValue)` */
+
+size_t FSE_decompress_wksp_bmi2(void* dst, size_t dstCapacity, const void* cSrc, size_t cSrcSize, unsigned maxLog, void* workSpace, size_t wkspSize, int bmi2);
+/**< Same as FSE_decompress_wksp() but with dynamic BMI2 support. Pass 1 if your CPU supports BMI2 or 0 if it doesn't. */
+
+typedef enum {
+   FSE_repeat_none,  /**< Cannot use the previous table */
+   FSE_repeat_check, /**< Can use the previous table but it must be checked */
+   FSE_repeat_valid  /**< Can use the previous table and it is assumed to be valid */
+ } FSE_repeat;
+
+/* *****************************************
+*  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 */
+
+MEM_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 = MEM_read16(ptr);
+    statePtr->value = (ptrdiff_t)1<<tableLog;
+    statePtr->stateTable = u16ptr+2;
+    statePtr->symbolTT = 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 */
+MEM_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];
+    }
+}
+
+MEM_STATIC void FSE_encodeSymbol(BIT_CStream_t* bitC, FSE_CState_t* statePtr, unsigned symbol)
+{
+    FSE_symbolCompressionTransform const symbolTT = ((const FSE_symbolCompressionTransform*)(statePtr->symbolTT))[symbol];
+    const U16* const stateTable = (const U16*)(statePtr->stateTable);
+    U32 const nbBitsOut  = (U32)((statePtr->value + symbolTT.deltaNbBits) >> 16);
+    BIT_addBits(bitC, statePtr->value, nbBitsOut);
+    statePtr->value = stateTable[ (statePtr->value >> nbBitsOut) + symbolTT.deltaFindState];
+}
+
+MEM_STATIC void FSE_flushCState(BIT_CStream_t* bitC, const FSE_CState_t* statePtr)
+{
+    BIT_addBits(bitC, statePtr->value, statePtr->stateLog);
+    BIT_flushBits(bitC);
+}
+
+
+/* FSE_getMaxNbBits() :
+ * Approximate maximum cost of a symbol, in bits.
+ * Fractional get rounded up (i.e. a symbol with a normalized frequency of 3 gives the same result as a frequency of 2)
+ * note 1 : assume symbolValue is valid (<= maxSymbolValue)
+ * note 2 : if freq[symbolValue]==0, @return a fake cost of tableLog+1 bits */
+MEM_STATIC U32 FSE_getMaxNbBits(const void* symbolTTPtr, U32 symbolValue)
+{
+    const FSE_symbolCompressionTransform* symbolTT = (const FSE_symbolCompressionTransform*) symbolTTPtr;
+    return (symbolTT[symbolValue].deltaNbBits + ((1<<16)-1)) >> 16;
+}
+
+/* FSE_bitCost() :
+ * Approximate symbol cost, as fractional value, using fixed-point format (accuracyLog fractional bits)
+ * note 1 : assume symbolValue is valid (<= maxSymbolValue)
+ * note 2 : if freq[symbolValue]==0, @return a fake cost of tableLog+1 bits */
+MEM_STATIC U32 FSE_bitCost(const void* symbolTTPtr, U32 tableLog, U32 symbolValue, U32 accuracyLog)
+{
+    const FSE_symbolCompressionTransform* symbolTT = (const FSE_symbolCompressionTransform*) symbolTTPtr;
+    U32 const minNbBits = symbolTT[symbolValue].deltaNbBits >> 16;
+    U32 const threshold = (minNbBits+1) << 16;
+    assert(tableLog < 16);
+    assert(accuracyLog < 31-tableLog);  /* ensure enough room for renormalization double shift */
+    {   U32 const tableSize = 1 << tableLog;
+        U32 const deltaFromThreshold = threshold - (symbolTT[symbolValue].deltaNbBits + tableSize);
+        U32 const normalizedDeltaFromThreshold = (deltaFromThreshold << accuracyLog) >> tableLog;   /* linear interpolation (very approximate) */
+        U32 const bitMultiplier = 1 << accuracyLog;
+        assert(symbolTT[symbolValue].deltaNbBits + tableSize <= threshold);
+        assert(normalizedDeltaFromThreshold <= bitMultiplier);
+        return (minNbBits+1)*bitMultiplier - normalizedDeltaFromThreshold;
+    }
+}
+
+
+/* ======    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 */
+
+MEM_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;
+}
+
+MEM_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;
+}
+
+MEM_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;
+}
+
+MEM_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% */
+MEM_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;
+}
+
+MEM_STATIC unsigned FSE_endOfDState(const FSE_DState_t* DStatePtr)
+{
+    return DStatePtr->state == 0;
+}
+
+
+
+#ifndef FSE_COMMONDEFS_ONLY
+
+/* **************************************************************
+*  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
+#if (FSE_DEFAULT_MEMORY_USAGE > FSE_MAX_MEMORY_USAGE)
+#  error "FSE_DEFAULT_MEMORY_USAGE must be <= FSE_MAX_MEMORY_USAGE"
+#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
+
+
+#endif   /* !FSE_COMMONDEFS_ONLY */
+
+
+/* ***************************************************************
+*  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_STATIC_LINKING_ONLY */
+
+
+#if defined (__cplusplus)
+}
+#endif

external/zstd/common/fse_decompress.c

@@ -0,0 +1,405 @@
+/* ******************************************************************
+ * FSE : Finite State Entropy decoder
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ *
+ *  You can contact the author at :
+ *  - FSE source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *  - Public forum : https://groups.google.com/forum/#!forum/lz4c
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+
+
+/* **************************************************************
+*  Includes
+****************************************************************/
+#include "debug.h"      /* assert */
+#include "bitstream.h"
+#include "compiler.h"
+#define FSE_STATIC_LINKING_ONLY
+#include "fse.h"
+#include "error_private.h"
+#define ZSTD_DEPS_NEED_MALLOC
+#include "zstd_deps.h"
+#include "bits.h"       /* ZSTD_highbit32 */
+
+
+/* **************************************************************
+*  Error Management
+****************************************************************/
+#define FSE_isError ERR_isError
+#define FSE_STATIC_ASSERT(c) DEBUG_STATIC_ASSERT(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_DTable* FSE_createDTable (unsigned tableLog)
+{
+    if (tableLog > FSE_TABLELOG_ABSOLUTE_MAX) tableLog = FSE_TABLELOG_ABSOLUTE_MAX;
+    return (FSE_DTable*)ZSTD_malloc( FSE_DTABLE_SIZE_U32(tableLog) * sizeof (U32) );
+}
+
+void FSE_freeDTable (FSE_DTable* dt)
+{
+    ZSTD_free(dt);
+}
+
+static size_t FSE_buildDTable_internal(FSE_DTable* dt, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog, void* workSpace, size_t wkspSize)
+{
+    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;
+    BYTE* spread = (BYTE*)(symbolNext + maxSymbolValue + 1);
+
+    U32 const maxSV1 = maxSymbolValue + 1;
+    U32 const tableSize = 1 << tableLog;
+    U32 highThreshold = tableSize-1;
+
+    /* Sanity Checks */
+    if (FSE_BUILD_DTABLE_WKSP_SIZE(tableLog, maxSymbolValue) > wkspSize) return ERROR(maxSymbolValue_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];
+        }   }   }
+        ZSTD_memcpy(dt, &DTableH, sizeof(DTableH));
+    }
+
+    /* Spread symbols */
+    if (highThreshold == tableSize - 1) {
+        size_t const tableMask = tableSize-1;
+        size_t const step = FSE_TABLESTEP(tableSize);
+        /* First lay down the symbols in order.
+         * We use a uint64_t to lay down 8 bytes at a time. This reduces branch
+         * misses since small blocks generally have small table logs, so nearly
+         * all symbols have counts <= 8. We ensure we have 8 bytes at the end of
+         * our buffer to handle the over-write.
+         */
+        {
+            U64 const add = 0x0101010101010101ull;
+            size_t pos = 0;
+            U64 sv = 0;
+            U32 s;
+            for (s=0; s<maxSV1; ++s, sv += add) {
+                int i;
+                int const n = normalizedCounter[s];
+                MEM_write64(spread + pos, sv);
+                for (i = 8; i < n; i += 8) {
+                    MEM_write64(spread + pos + i, sv);
+                }
+                pos += n;
+            }
+        }
+        /* Now we spread those positions across the table.
+         * The benefit of doing it in two stages is that we avoid the
+         * variable size inner loop, which caused lots of branch misses.
+         * Now we can run through all the positions without any branch misses.
+         * We unroll the loop twice, since that is what empirically worked best.
+         */
+        {
+            size_t position = 0;
+            size_t s;
+            size_t const unroll = 2;
+            assert(tableSize % unroll == 0); /* FSE_MIN_TABLELOG is 5 */
+            for (s = 0; s < (size_t)tableSize; s += unroll) {
+                size_t u;
+                for (u = 0; u < unroll; ++u) {
+                    size_t const uPosition = (position + (u * step)) & tableMask;
+                    tableDecode[uPosition].symbol = spread[s + u];
+                }
+                position = (position + (unroll * step)) & tableMask;
+            }
+            assert(position == 0);
+        }
+    } else {
+        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);
+            U32 const nextState = symbolNext[symbol]++;
+            tableDecode[u].nbBits = (BYTE) (tableLog - ZSTD_highbit32(nextState) );
+            tableDecode[u].newState = (U16) ( (nextState << tableDecode[u].nbBits) - tableSize);
+    }   }
+
+    return 0;
+}
+
+size_t FSE_buildDTable_wksp(FSE_DTable* dt, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog, void* workSpace, size_t wkspSize)
+{
+    return FSE_buildDTable_internal(dt, normalizedCounter, maxSymbolValue, tableLog, workSpace, wkspSize);
+}
+
+
+#ifndef FSE_COMMONDEFS_ONLY
+
+/*-*******************************************************
+*  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_TEMPLATE 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 wkspSize)
+{
+    return FSE_decompress_wksp_bmi2(dst, dstCapacity, cSrc, cSrcSize, maxLog, workSpace, wkspSize, /* bmi2 */ 0);
+}
+
+typedef struct {
+    short ncount[FSE_MAX_SYMBOL_VALUE + 1];
+    FSE_DTable dtable[1]; /* Dynamically sized */
+} FSE_DecompressWksp;
+
+
+FORCE_INLINE_TEMPLATE size_t FSE_decompress_wksp_body(
+        void* dst, size_t dstCapacity,
+        const void* cSrc, size_t cSrcSize,
+        unsigned maxLog, void* workSpace, size_t wkspSize,
+        int bmi2)
+{
+    const BYTE* const istart = (const BYTE*)cSrc;
+    const BYTE* ip = istart;
+    unsigned tableLog;
+    unsigned maxSymbolValue = FSE_MAX_SYMBOL_VALUE;
+    FSE_DecompressWksp* const wksp = (FSE_DecompressWksp*)workSpace;
+
+    DEBUG_STATIC_ASSERT((FSE_MAX_SYMBOL_VALUE + 1) % 2 == 0);
+    if (wkspSize < sizeof(*wksp)) return ERROR(GENERIC);
+
+    /* normal FSE decoding mode */
+    {
+        size_t const NCountLength = FSE_readNCount_bmi2(wksp->ncount, &maxSymbolValue, &tableLog, istart, cSrcSize, bmi2);
+        if (FSE_isError(NCountLength)) return NCountLength;
+        if (tableLog > maxLog) return ERROR(tableLog_tooLarge);
+        assert(NCountLength <= cSrcSize);
+        ip += NCountLength;
+        cSrcSize -= NCountLength;
+    }
+
+    if (FSE_DECOMPRESS_WKSP_SIZE(tableLog, maxSymbolValue) > wkspSize) return ERROR(tableLog_tooLarge);
+    assert(sizeof(*wksp) + FSE_DTABLE_SIZE(tableLog) <= wkspSize);
+    workSpace = (BYTE*)workSpace + sizeof(*wksp) + FSE_DTABLE_SIZE(tableLog);
+    wkspSize -= sizeof(*wksp) + FSE_DTABLE_SIZE(tableLog);
+
+    CHECK_F( FSE_buildDTable_internal(wksp->dtable, wksp->ncount, maxSymbolValue, tableLog, workSpace, wkspSize) );
+
+    {
+        const void* ptr = wksp->dtable;
+        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, dstCapacity, ip, cSrcSize, wksp->dtable, 1);
+        return FSE_decompress_usingDTable_generic(dst, dstCapacity, ip, cSrcSize, wksp->dtable, 0);
+    }
+}
+
+/* Avoids the FORCE_INLINE of the _body() function. */
+static size_t FSE_decompress_wksp_body_default(void* dst, size_t dstCapacity, const void* cSrc, size_t cSrcSize, unsigned maxLog, void* workSpace, size_t wkspSize)
+{
+    return FSE_decompress_wksp_body(dst, dstCapacity, cSrc, cSrcSize, maxLog, workSpace, wkspSize, 0);
+}
+
+#if DYNAMIC_BMI2
+BMI2_TARGET_ATTRIBUTE static size_t FSE_decompress_wksp_body_bmi2(void* dst, size_t dstCapacity, const void* cSrc, size_t cSrcSize, unsigned maxLog, void* workSpace, size_t wkspSize)
+{
+    return FSE_decompress_wksp_body(dst, dstCapacity, cSrc, cSrcSize, maxLog, workSpace, wkspSize, 1);
+}
+#endif
+
+size_t FSE_decompress_wksp_bmi2(void* dst, size_t dstCapacity, const void* cSrc, size_t cSrcSize, unsigned maxLog, void* workSpace, size_t wkspSize, int bmi2)
+{
+#if DYNAMIC_BMI2
+    if (bmi2) {
+        return FSE_decompress_wksp_body_bmi2(dst, dstCapacity, cSrc, cSrcSize, maxLog, workSpace, wkspSize);
+    }
+#endif
+    (void)bmi2;
+    return FSE_decompress_wksp_body_default(dst, dstCapacity, cSrc, cSrcSize, maxLog, workSpace, wkspSize);
+}
+
+
+typedef FSE_DTable DTable_max_t[FSE_DTABLE_SIZE_U32(FSE_MAX_TABLELOG)];
+
+#ifndef ZSTD_NO_UNUSED_FUNCTIONS
+size_t FSE_buildDTable(FSE_DTable* dt, const short* normalizedCounter, unsigned maxSymbolValue, unsigned tableLog) {
+    U32 wksp[FSE_BUILD_DTABLE_WKSP_SIZE_U32(FSE_TABLELOG_ABSOLUTE_MAX, FSE_MAX_SYMBOL_VALUE)];
+    return FSE_buildDTable_wksp(dt, normalizedCounter, maxSymbolValue, tableLog, wksp, sizeof(wksp));
+}
+
+size_t FSE_decompress(void* dst, size_t dstCapacity, const void* cSrc, size_t cSrcSize)
+{
+    /* Static analyzer seems unable to understand this table will be properly initialized later */
+    U32 wksp[FSE_DECOMPRESS_WKSP_SIZE_U32(FSE_MAX_TABLELOG, FSE_MAX_SYMBOL_VALUE)];
+    return FSE_decompress_wksp(dst, dstCapacity, cSrc, cSrcSize, FSE_MAX_TABLELOG, wksp, sizeof(wksp));
+}
+#endif
+
+
+#endif   /* FSE_COMMONDEFS_ONLY */

external/zstd/common/huf.h

@@ -0,0 +1,377 @@
+/* ******************************************************************
+ * huff0 huffman codec,
+ * part of Finite State Entropy library
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ *
+ * You can contact the author at :
+ * - Source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+#ifndef HUF_H_298734234
+#define HUF_H_298734234
+
+/* *** Dependencies *** */
+#include "zstd_deps.h"    /* size_t */
+
+
+/* *** library symbols visibility *** */
+/* Note : when linking with -fvisibility=hidden on gcc, or by default on Visual,
+ *        HUF symbols remain "private" (internal symbols for library only).
+ *        Set macro FSE_DLL_EXPORT to 1 if you want HUF symbols visible on DLL interface */
+#if defined(FSE_DLL_EXPORT) && (FSE_DLL_EXPORT==1) && defined(__GNUC__) && (__GNUC__ >= 4)
+#  define HUF_PUBLIC_API __attribute__ ((visibility ("default")))
+#elif defined(FSE_DLL_EXPORT) && (FSE_DLL_EXPORT==1)   /* Visual expected */
+#  define HUF_PUBLIC_API __declspec(dllexport)
+#elif defined(FSE_DLL_IMPORT) && (FSE_DLL_IMPORT==1)
+#  define HUF_PUBLIC_API __declspec(dllimport)  /* not required, just to generate faster code (saves a function pointer load from IAT and an indirect jump) */
+#else
+#  define HUF_PUBLIC_API
+#endif
+
+
+/* ========================== */
+/* ***  simple functions  *** */
+/* ========================== */
+
+/** HUF_compress() :
+ *  Compress content from buffer 'src', of size 'srcSize', into buffer 'dst'.
+ * 'dst' buffer must be already allocated.
+ *  Compression runs faster if `dstCapacity` >= HUF_compressBound(srcSize).
+ * `srcSize` must be <= `HUF_BLOCKSIZE_MAX` == 128 KB.
+ * @return : size of compressed data (<= `dstCapacity`).
+ *  Special values : if return == 0, srcData is not compressible => Nothing is stored within dst !!!
+ *                   if HUF_isError(return), compression failed (more details using HUF_getErrorName())
+ */
+HUF_PUBLIC_API size_t HUF_compress(void* dst, size_t dstCapacity,
+                             const void* src, size_t srcSize);
+
+/** HUF_decompress() :
+ *  Decompress HUF data from buffer 'cSrc', of size 'cSrcSize',
+ *  into already allocated buffer 'dst', of minimum size 'dstSize'.
+ * `originalSize` : **must** be the ***exact*** size of original (uncompressed) data.
+ *  Note : in contrast with FSE, HUF_decompress can regenerate
+ *         RLE (cSrcSize==1) and uncompressed (cSrcSize==dstSize) data,
+ *         because it knows size to regenerate (originalSize).
+ * @return : size of regenerated data (== originalSize),
+ *           or an error code, which can be tested using HUF_isError()
+ */
+HUF_PUBLIC_API size_t HUF_decompress(void* dst,  size_t originalSize,
+                               const void* cSrc, size_t cSrcSize);
+
+
+/* ***   Tool functions *** */
+#define HUF_BLOCKSIZE_MAX (128 * 1024)                  /**< maximum input size for a single block compressed with HUF_compress */
+HUF_PUBLIC_API size_t HUF_compressBound(size_t size);   /**< maximum compressed size (worst case) */
+
+/* Error Management */
+HUF_PUBLIC_API unsigned    HUF_isError(size_t code);       /**< tells if a return value is an error code */
+HUF_PUBLIC_API const char* HUF_getErrorName(size_t code);  /**< provides error code string (useful for debugging) */
+
+
+/* ***   Advanced function   *** */
+
+/** HUF_compress2() :
+ *  Same as HUF_compress(), but offers control over `maxSymbolValue` and `tableLog`.
+ * `maxSymbolValue` must be <= HUF_SYMBOLVALUE_MAX .
+ * `tableLog` must be `<= HUF_TABLELOG_MAX` . */
+HUF_PUBLIC_API size_t HUF_compress2 (void* dst, size_t dstCapacity,
+                               const void* src, size_t srcSize,
+                               unsigned maxSymbolValue, unsigned tableLog);
+
+/** HUF_compress4X_wksp() :
+ *  Same as HUF_compress2(), but uses externally allocated @workSpace.
+ * @workSpace's size, aka @wkspSize, must be >= HUF_WORKSPACE_SIZE
+ * @srcSize must be >= 6
+ */
+#define HUF_WORKSPACE_SIZE ((8 << 10) + 512 /* sorting scratch space */)
+#define HUF_WORKSPACE_SIZE_U64 (HUF_WORKSPACE_SIZE / sizeof(U64))
+HUF_PUBLIC_API size_t HUF_compress4X_wksp (void* dst, size_t dstCapacity,
+                                     const void* src, size_t srcSize,
+                                     unsigned maxSymbolValue, unsigned tableLog,
+                                     void* workSpace, size_t wkspSize);
+
+#endif   /* HUF_H_298734234 */
+
+/* ******************************************************************
+ *  WARNING !!
+ *  The following section contains advanced and experimental definitions
+ *  which shall never be used in the context of a dynamic library,
+ *  because they are not guaranteed to remain stable in the future.
+ *  Only consider them in association with static linking.
+ * *****************************************************************/
+#if defined(HUF_STATIC_LINKING_ONLY) && !defined(HUF_H_HUF_STATIC_LINKING_ONLY)
+#define HUF_H_HUF_STATIC_LINKING_ONLY
+
+/* *** Dependencies *** */
+#include "mem.h"   /* U32 */
+#define FSE_STATIC_LINKING_ONLY
+#include "fse.h"
+
+
+/* *** Constants *** */
+#define HUF_TABLELOG_MAX      12      /* max runtime value of tableLog (due to static allocation); can be modified up to HUF_TABLELOG_ABSOLUTEMAX */
+#define HUF_TABLELOG_DEFAULT  11      /* default tableLog value when none specified */
+#define HUF_SYMBOLVALUE_MAX  255
+
+#define HUF_TABLELOG_ABSOLUTEMAX  12  /* 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 when incompressible is 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 */
+/* this is a private definition, just exposed for allocation and strict aliasing purpose. never EVER access its members directly */
+typedef size_t HUF_CElt;   /* consider it an incomplete type */
+#define HUF_CTABLE_SIZE_ST(maxSymbolValue)   ((maxSymbolValue)+2)   /* Use tables of size_t, for proper alignment */
+#define HUF_CTABLE_SIZE(maxSymbolValue)       (HUF_CTABLE_SIZE_ST(maxSymbolValue) * sizeof(size_t))
+#define HUF_CREATE_STATIC_CTABLE(name, maxSymbolValue) \
+    HUF_CElt name[HUF_CTABLE_SIZE_ST(maxSymbolValue)] /* no final ; */
+
+/* static allocation of HUF's DTable */
+typedef U32 HUF_DTable;
+#define HUF_DTABLE_SIZE(maxTableLog)   (1 + (1<<(maxTableLog)))
+#define HUF_CREATE_STATIC_DTABLEX1(DTable, maxTableLog) \
+        HUF_DTable DTable[HUF_DTABLE_SIZE((maxTableLog)-1)] = { ((U32)((maxTableLog)-1) * 0x01000001) }
+#define HUF_CREATE_STATIC_DTABLEX2(DTable, maxTableLog) \
+        HUF_DTable DTable[HUF_DTABLE_SIZE(maxTableLog)] = { ((U32)(maxTableLog) * 0x01000001) }
+
+
+/* ****************************************
+*  Advanced decompression functions
+******************************************/
+size_t HUF_decompress4X1 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);   /**< single-symbol decoder */
+#ifndef HUF_FORCE_DECOMPRESS_X1
+size_t HUF_decompress4X2 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);   /**< double-symbols decoder */
+#endif
+
+size_t HUF_decompress4X_DCtx (HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);   /**< decodes RLE and uncompressed */
+size_t HUF_decompress4X_hufOnly(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize); /**< considers RLE and uncompressed as errors */
+size_t HUF_decompress4X_hufOnly_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize); /**< considers RLE and uncompressed as errors */
+size_t HUF_decompress4X1_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);   /**< single-symbol decoder */
+size_t HUF_decompress4X1_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize);   /**< single-symbol decoder */
+#ifndef HUF_FORCE_DECOMPRESS_X1
+size_t HUF_decompress4X2_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);   /**< double-symbols decoder */
+size_t HUF_decompress4X2_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize);   /**< double-symbols decoder */
+#endif
+
+
+/* ****************************************
+ *  HUF detailed API
+ * ****************************************/
+#define HUF_OPTIMAL_DEPTH_THRESHOLD ZSTD_btultra
+typedef enum {
+   HUF_depth_fast, /** Use heuristic to find the table depth**/
+   HUF_depth_optimal /** Test possible table depths to find the one that produces the smallest header + encoded size**/
+ } HUF_depth_mode;
+
+/*! HUF_compress() does the following:
+ *  1. count symbol occurrence from source[] into table count[] using FSE_count() (exposed within "fse.h")
+ *  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()
+ *  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.
+ */
+unsigned HUF_minTableLog(unsigned symbolCardinality);
+unsigned HUF_cardinality(const unsigned* count, unsigned maxSymbolValue);
+unsigned HUF_optimalTableLog(unsigned maxTableLog, size_t srcSize, unsigned maxSymbolValue, void* workSpace,
+ size_t wkspSize, HUF_CElt* table, const unsigned* count, HUF_depth_mode depthMode); /* table is used as scratch space for building and testing tables, not a return value */
+size_t HUF_buildCTable (HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue, unsigned maxNbBits);   /* @return : maxNbBits; CTable and count can overlap. In which case, CTable will overwrite count content */
+size_t HUF_writeCTable (void* dst, size_t maxDstSize, const HUF_CElt* CTable, unsigned maxSymbolValue, unsigned huffLog);
+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);
+size_t HUF_compress4X_usingCTable_bmi2(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable, int bmi2);
+size_t HUF_estimateCompressedSize(const HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue);
+int HUF_validateCTable(const HUF_CElt* CTable, const unsigned* count, unsigned maxSymbolValue);
+
+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 assumed 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.
+ *  If suspectUncompressible then some sampling checks will be run to potentially skip huffman coding */
+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,    /**< `workSpace` must be aligned on 4-bytes boundaries, `wkspSize` must be >= HUF_WORKSPACE_SIZE */
+                       HUF_CElt* hufTable, HUF_repeat* repeat, int preferRepeat, int bmi2,
+                       int suspectUncompressible, HUF_depth_mode depthMode);
+
+/** HUF_buildCTable_wksp() :
+ *  Same as HUF_buildCTable(), but using externally allocated scratch buffer.
+ * `workSpace` must be aligned on 4-bytes boundaries, and its size must be >= HUF_CTABLE_WORKSPACE_SIZE.
+ */
+#define HUF_CTABLE_WORKSPACE_SIZE_U32 (2*HUF_SYMBOLVALUE_MAX +1 +1)
+#define HUF_CTABLE_WORKSPACE_SIZE (HUF_CTABLE_WORKSPACE_SIZE_U32 * sizeof(unsigned))
+size_t HUF_buildCTable_wksp (HUF_CElt* tree,
+                       const unsigned* 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(BYTE* huffWeight, size_t hwSize,
+                     U32* rankStats, U32* nbSymbolsPtr, U32* tableLogPtr,
+                     const void* src, size_t srcSize);
+
+/*! HUF_readStats_wksp() :
+ * Same as HUF_readStats() but takes an external workspace which must be
+ * 4-byte aligned and its size must be >= HUF_READ_STATS_WORKSPACE_SIZE.
+ * If the CPU has BMI2 support, pass bmi2=1, otherwise pass bmi2=0.
+ */
+#define HUF_READ_STATS_WORKSPACE_SIZE_U32 FSE_DECOMPRESS_WKSP_SIZE_U32(6, HUF_TABLELOG_MAX-1)
+#define HUF_READ_STATS_WORKSPACE_SIZE (HUF_READ_STATS_WORKSPACE_SIZE_U32 * sizeof(unsigned))
+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 wkspSize,
+                          int bmi2);
+
+/** HUF_readCTable() :
+ *  Loading a CTable saved with HUF_writeCTable() */
+size_t HUF_readCTable (HUF_CElt* CTable, unsigned* maxSymbolValuePtr, const void* src, size_t srcSize, unsigned *hasZeroWeights);
+
+/** HUF_getNbBitsFromCTable() :
+ *  Read nbBits from CTable symbolTable, for symbol `symbolValue` presumed <= HUF_SYMBOLVALUE_MAX
+ *  Note 1 : is not inlined, as HUF_CElt definition is private */
+U32 HUF_getNbBitsFromCTable(const HUF_CElt* symbolTable, U32 symbolValue);
+
+/*
+ * HUF_decompress() does the following:
+ * 1. select the decompression algorithm (X1, X2) based on pre-computed heuristics
+ * 2. build Huffman table from save, using HUF_readDTableX?()
+ * 3. decode 1 or 4 segments in parallel using HUF_decompress?X?_usingDTable()
+ */
+
+/** HUF_selectDecoder() :
+ *  Tells which decoder is likely to decode faster,
+ *  based on a set of pre-computed metrics.
+ * @return : 0==HUF_decompress4X1, 1==HUF_decompress4X2 .
+ *  Assumption : 0 < dstSize <= 128 KB */
+U32 HUF_selectDecoder (size_t dstSize, size_t cSrcSize);
+
+/**
+ *  The minimum workspace size for the `workSpace` used in
+ *  HUF_readDTableX1_wksp() and HUF_readDTableX2_wksp().
+ *
+ *  The space used depends on HUF_TABLELOG_MAX, ranging from ~1500 bytes when
+ *  HUF_TABLE_LOG_MAX=12 to ~1850 bytes when HUF_TABLE_LOG_MAX=15.
+ *  Buffer overflow errors may potentially occur if code modifications result in
+ *  a required workspace size greater than that specified in the following
+ *  macro.
+ */
+#define HUF_DECOMPRESS_WORKSPACE_SIZE ((2 << 10) + (1 << 9))
+#define HUF_DECOMPRESS_WORKSPACE_SIZE_U32 (HUF_DECOMPRESS_WORKSPACE_SIZE / sizeof(U32))
+
+#ifndef HUF_FORCE_DECOMPRESS_X2
+size_t HUF_readDTableX1 (HUF_DTable* DTable, const void* src, size_t srcSize);
+size_t HUF_readDTableX1_wksp (HUF_DTable* DTable, const void* src, size_t srcSize, void* workSpace, size_t wkspSize);
+#endif
+#ifndef HUF_FORCE_DECOMPRESS_X1
+size_t HUF_readDTableX2 (HUF_DTable* DTable, const void* src, size_t srcSize);
+size_t HUF_readDTableX2_wksp (HUF_DTable* DTable, const void* src, size_t srcSize, void* workSpace, size_t wkspSize);
+#endif
+
+size_t HUF_decompress4X_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);
+#ifndef HUF_FORCE_DECOMPRESS_X2
+size_t HUF_decompress4X1_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);
+#endif
+#ifndef HUF_FORCE_DECOMPRESS_X1
+size_t HUF_decompress4X2_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);
+#endif
+
+
+/* ====================== */
+/* single stream variants */
+/* ====================== */
+
+size_t HUF_compress1X (void* dst, size_t dstSize, const void* src, size_t srcSize, unsigned maxSymbolValue, unsigned tableLog);
+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_WORKSPACE_SIZE_U64 U64 */
+size_t HUF_compress1X_usingCTable(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable);
+size_t HUF_compress1X_usingCTable_bmi2(void* dst, size_t dstSize, const void* src, size_t srcSize, const HUF_CElt* CTable, int bmi2);
+/** 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.
+ *  If suspectUncompressible then some sampling checks will be run to potentially skip huffman coding */
+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,   /**< `workSpace` must be aligned on 4-bytes boundaries, `wkspSize` must be >= HUF_WORKSPACE_SIZE */
+                       HUF_CElt* hufTable, HUF_repeat* repeat, int preferRepeat, int bmi2,
+                       int suspectUncompressible, HUF_depth_mode depthMode);
+
+size_t HUF_decompress1X1 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);   /* single-symbol decoder */
+#ifndef HUF_FORCE_DECOMPRESS_X1
+size_t HUF_decompress1X2 (void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);   /* double-symbol decoder */
+#endif
+
+size_t HUF_decompress1X_DCtx (HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);
+size_t HUF_decompress1X_DCtx_wksp (HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize);
+#ifndef HUF_FORCE_DECOMPRESS_X2
+size_t HUF_decompress1X1_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);   /**< single-symbol decoder */
+size_t HUF_decompress1X1_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize);   /**< single-symbol decoder */
+#endif
+#ifndef HUF_FORCE_DECOMPRESS_X1
+size_t HUF_decompress1X2_DCtx(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize);   /**< double-symbols decoder */
+size_t HUF_decompress1X2_DCtx_wksp(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize);   /**< double-symbols decoder */
+#endif
+
+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 */
+#ifndef HUF_FORCE_DECOMPRESS_X2
+size_t HUF_decompress1X1_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);
+#endif
+#ifndef HUF_FORCE_DECOMPRESS_X1
+size_t HUF_decompress1X2_usingDTable(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable);
+#endif
+
+/* BMI2 variants.
+ * If the CPU has BMI2 support, pass bmi2=1, otherwise pass bmi2=0.
+ */
+size_t HUF_decompress1X_usingDTable_bmi2(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable, int bmi2);
+#ifndef HUF_FORCE_DECOMPRESS_X2
+size_t HUF_decompress1X1_DCtx_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int bmi2);
+#endif
+size_t HUF_decompress4X_usingDTable_bmi2(void* dst, size_t maxDstSize, const void* cSrc, size_t cSrcSize, const HUF_DTable* DTable, int bmi2);
+size_t HUF_decompress4X_hufOnly_wksp_bmi2(HUF_DTable* dctx, void* dst, size_t dstSize, const void* cSrc, size_t cSrcSize, void* workSpace, size_t wkspSize, int bmi2);
+#ifndef HUF_FORCE_DECOMPRESS_X2
+size_t HUF_readDTableX1_wksp_bmi2(HUF_DTable* DTable, const void* src, size_t srcSize, void* workSpace, size_t wkspSize, int bmi2);
+#endif
+#ifndef HUF_FORCE_DECOMPRESS_X1
+size_t HUF_readDTableX2_wksp_bmi2(HUF_DTable* DTable, const void* src, size_t srcSize, void* workSpace, size_t wkspSize, int bmi2);
+#endif
+
+#endif /* HUF_STATIC_LINKING_ONLY */
+
+#if defined (__cplusplus)
+}
+#endif

external/zstd/common/mem.h

@@ -0,0 +1,435 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef MEM_H_MODULE
+#define MEM_H_MODULE
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+/*-****************************************
+*  Dependencies
+******************************************/
+#include <stddef.h>  /* size_t, ptrdiff_t */
+#include "compiler.h"  /* __has_builtin */
+#include "debug.h"  /* DEBUG_STATIC_ASSERT */
+#include "zstd_deps.h"  /* ZSTD_memcpy */
+
+
+/*-****************************************
+*  Compiler specifics
+******************************************/
+#if defined(_MSC_VER)   /* Visual Studio */
+#   include <stdlib.h>  /* _byteswap_ulong */
+#   include <intrin.h>  /* _byteswap_* */
+#endif
+#if defined(__GNUC__)
+#  define MEM_STATIC static __inline __attribute__((unused))
+#elif defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */)
+#  define MEM_STATIC static inline
+#elif defined(_MSC_VER)
+#  define MEM_STATIC static __inline
+#else
+#  define MEM_STATIC static  /* this version may generate warnings for unused static functions; disable the relevant warning */
+#endif
+
+/*-**************************************************************
+*  Basic Types
+*****************************************************************/
+#if  !defined (__VMS) && (defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) )
+#  if defined(_AIX)
+#    include <inttypes.h>
+#  else
+#    include <stdint.h> /* intptr_t */
+#  endif
+  typedef   uint8_t BYTE;
+  typedef   uint8_t U8;
+  typedef    int8_t S8;
+  typedef  uint16_t U16;
+  typedef   int16_t S16;
+  typedef  uint32_t U32;
+  typedef   int32_t S32;
+  typedef  uint64_t U64;
+  typedef   int64_t S64;
+#else
+# include <limits.h>
+#if CHAR_BIT != 8
+#  error "this implementation requires char to be exactly 8-bit type"
+#endif
+  typedef unsigned char      BYTE;
+  typedef unsigned char      U8;
+  typedef   signed char      S8;
+#if USHRT_MAX != 65535
+#  error "this implementation requires short to be exactly 16-bit type"
+#endif
+  typedef unsigned short      U16;
+  typedef   signed short      S16;
+#if UINT_MAX != 4294967295
+#  error "this implementation requires int to be exactly 32-bit type"
+#endif
+  typedef unsigned int        U32;
+  typedef   signed int        S32;
+/* note : there are no limits defined for long long type in C90.
+ * limits exist in C99, however, in such case, <stdint.h> is preferred */
+  typedef unsigned long long  U64;
+  typedef   signed long long  S64;
+#endif
+
+
+/*-**************************************************************
+*  Memory I/O API
+*****************************************************************/
+/*=== Static platform detection ===*/
+MEM_STATIC unsigned MEM_32bits(void);
+MEM_STATIC unsigned MEM_64bits(void);
+MEM_STATIC unsigned MEM_isLittleEndian(void);
+
+/*=== Native unaligned read/write ===*/
+MEM_STATIC U16 MEM_read16(const void* memPtr);
+MEM_STATIC U32 MEM_read32(const void* memPtr);
+MEM_STATIC U64 MEM_read64(const void* memPtr);
+MEM_STATIC size_t MEM_readST(const void* memPtr);
+
+MEM_STATIC void MEM_write16(void* memPtr, U16 value);
+MEM_STATIC void MEM_write32(void* memPtr, U32 value);
+MEM_STATIC void MEM_write64(void* memPtr, U64 value);
+
+/*=== Little endian unaligned read/write ===*/
+MEM_STATIC U16 MEM_readLE16(const void* memPtr);
+MEM_STATIC U32 MEM_readLE24(const void* memPtr);
+MEM_STATIC U32 MEM_readLE32(const void* memPtr);
+MEM_STATIC U64 MEM_readLE64(const void* memPtr);
+MEM_STATIC size_t MEM_readLEST(const void* memPtr);
+
+MEM_STATIC void MEM_writeLE16(void* memPtr, U16 val);
+MEM_STATIC void MEM_writeLE24(void* memPtr, U32 val);
+MEM_STATIC void MEM_writeLE32(void* memPtr, U32 val32);
+MEM_STATIC void MEM_writeLE64(void* memPtr, U64 val64);
+MEM_STATIC void MEM_writeLEST(void* memPtr, size_t val);
+
+/*=== Big endian unaligned read/write ===*/
+MEM_STATIC U32 MEM_readBE32(const void* memPtr);
+MEM_STATIC U64 MEM_readBE64(const void* memPtr);
+MEM_STATIC size_t MEM_readBEST(const void* memPtr);
+
+MEM_STATIC void MEM_writeBE32(void* memPtr, U32 val32);
+MEM_STATIC void MEM_writeBE64(void* memPtr, U64 val64);
+MEM_STATIC void MEM_writeBEST(void* memPtr, size_t val);
+
+/*=== Byteswap ===*/
+MEM_STATIC U32 MEM_swap32(U32 in);
+MEM_STATIC U64 MEM_swap64(U64 in);
+MEM_STATIC size_t MEM_swapST(size_t in);
+
+
+/*-**************************************************************
+*  Memory I/O Implementation
+*****************************************************************/
+/* MEM_FORCE_MEMORY_ACCESS : For accessing unaligned memory:
+ * Method 0 : always use `memcpy()`. Safe and portable.
+ * Method 1 : Use compiler extension to set unaligned access.
+ * Method 2 : direct access. This method is portable but violate C standard.
+ *            It can generate buggy code on targets depending on alignment.
+ * Default  : method 1 if supported, else method 0
+ */
+#ifndef MEM_FORCE_MEMORY_ACCESS   /* can be defined externally, on command line for example */
+#  ifdef __GNUC__
+#    define MEM_FORCE_MEMORY_ACCESS 1
+#  endif
+#endif
+
+MEM_STATIC unsigned MEM_32bits(void) { return sizeof(size_t)==4; }
+MEM_STATIC unsigned MEM_64bits(void) { return sizeof(size_t)==8; }
+
+MEM_STATIC unsigned MEM_isLittleEndian(void)
+{
+#if defined(__BYTE_ORDER__) && defined(__ORDER_LITTLE_ENDIAN__) && (__BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__)
+    return 1;
+#elif defined(__BYTE_ORDER__) && defined(__ORDER_BIG_ENDIAN__) && (__BYTE_ORDER__ == __ORDER_BIG_ENDIAN__)
+    return 0;
+#elif defined(__clang__) && __LITTLE_ENDIAN__
+    return 1;
+#elif defined(__clang__) && __BIG_ENDIAN__
+    return 0;
+#elif defined(_MSC_VER) && (_M_AMD64 || _M_IX86)
+    return 1;
+#elif defined(__DMC__) && defined(_M_IX86)
+    return 1;
+#else
+    const union { U32 u; BYTE c[4]; } one = { 1 };   /* don't use static : performance detrimental  */
+    return one.c[0];
+#endif
+}
+
+#if defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==2)
+
+/* violates C standard, by lying on structure alignment.
+Only use if no other choice to achieve best performance on target platform */
+MEM_STATIC U16 MEM_read16(const void* memPtr) { return *(const U16*) memPtr; }
+MEM_STATIC U32 MEM_read32(const void* memPtr) { return *(const U32*) memPtr; }
+MEM_STATIC U64 MEM_read64(const void* memPtr) { return *(const U64*) memPtr; }
+MEM_STATIC size_t MEM_readST(const void* memPtr) { return *(const size_t*) memPtr; }
+
+MEM_STATIC void MEM_write16(void* memPtr, U16 value) { *(U16*)memPtr = value; }
+MEM_STATIC void MEM_write32(void* memPtr, U32 value) { *(U32*)memPtr = value; }
+MEM_STATIC void MEM_write64(void* memPtr, U64 value) { *(U64*)memPtr = value; }
+
+#elif defined(MEM_FORCE_MEMORY_ACCESS) && (MEM_FORCE_MEMORY_ACCESS==1)
+
+typedef __attribute__((aligned(1))) U16 unalign16;
+typedef __attribute__((aligned(1))) U32 unalign32;
+typedef __attribute__((aligned(1))) U64 unalign64;
+typedef __attribute__((aligned(1))) size_t unalignArch;
+
+MEM_STATIC U16 MEM_read16(const void* ptr) { return *(const unalign16*)ptr; }
+MEM_STATIC U32 MEM_read32(const void* ptr) { return *(const unalign32*)ptr; }
+MEM_STATIC U64 MEM_read64(const void* ptr) { return *(const unalign64*)ptr; }
+MEM_STATIC size_t MEM_readST(const void* ptr) { return *(const unalignArch*)ptr; }
+
+MEM_STATIC void MEM_write16(void* memPtr, U16 value) { *(unalign16*)memPtr = value; }
+MEM_STATIC void MEM_write32(void* memPtr, U32 value) { *(unalign32*)memPtr = value; }
+MEM_STATIC void MEM_write64(void* memPtr, U64 value) { *(unalign64*)memPtr = value; }
+
+#else
+
+/* default method, safe and standard.
+   can sometimes prove slower */
+
+MEM_STATIC U16 MEM_read16(const void* memPtr)
+{
+    U16 val; ZSTD_memcpy(&val, memPtr, sizeof(val)); return val;
+}
+
+MEM_STATIC U32 MEM_read32(const void* memPtr)
+{
+    U32 val; ZSTD_memcpy(&val, memPtr, sizeof(val)); return val;
+}
+
+MEM_STATIC U64 MEM_read64(const void* memPtr)
+{
+    U64 val; ZSTD_memcpy(&val, memPtr, sizeof(val)); return val;
+}
+
+MEM_STATIC size_t MEM_readST(const void* memPtr)
+{
+    size_t val; ZSTD_memcpy(&val, memPtr, sizeof(val)); return val;
+}
+
+MEM_STATIC void MEM_write16(void* memPtr, U16 value)
+{
+    ZSTD_memcpy(memPtr, &value, sizeof(value));
+}
+
+MEM_STATIC void MEM_write32(void* memPtr, U32 value)
+{
+    ZSTD_memcpy(memPtr, &value, sizeof(value));
+}
+
+MEM_STATIC void MEM_write64(void* memPtr, U64 value)
+{
+    ZSTD_memcpy(memPtr, &value, sizeof(value));
+}
+
+#endif /* MEM_FORCE_MEMORY_ACCESS */
+
+MEM_STATIC U32 MEM_swap32_fallback(U32 in)
+{
+    return  ((in << 24) & 0xff000000 ) |
+            ((in <<  8) & 0x00ff0000 ) |
+            ((in >>  8) & 0x0000ff00 ) |
+            ((in >> 24) & 0x000000ff );
+}
+
+MEM_STATIC U32 MEM_swap32(U32 in)
+{
+#if defined(_MSC_VER)     /* Visual Studio */
+    return _byteswap_ulong(in);
+#elif (defined (__GNUC__) && (__GNUC__ * 100 + __GNUC_MINOR__ >= 403)) \
+  || (defined(__clang__) && __has_builtin(__builtin_bswap32))
+    return __builtin_bswap32(in);
+#else
+    return MEM_swap32_fallback(in);
+#endif
+}
+
+MEM_STATIC U64 MEM_swap64_fallback(U64 in)
+{
+     return  ((in << 56) & 0xff00000000000000ULL) |
+            ((in << 40) & 0x00ff000000000000ULL) |
+            ((in << 24) & 0x0000ff0000000000ULL) |
+            ((in << 8)  & 0x000000ff00000000ULL) |
+            ((in >> 8)  & 0x00000000ff000000ULL) |
+            ((in >> 24) & 0x0000000000ff0000ULL) |
+            ((in >> 40) & 0x000000000000ff00ULL) |
+            ((in >> 56) & 0x00000000000000ffULL);
+}
+
+MEM_STATIC U64 MEM_swap64(U64 in)
+{
+#if defined(_MSC_VER)     /* Visual Studio */
+    return _byteswap_uint64(in);
+#elif (defined (__GNUC__) && (__GNUC__ * 100 + __GNUC_MINOR__ >= 403)) \
+  || (defined(__clang__) && __has_builtin(__builtin_bswap64))
+    return __builtin_bswap64(in);
+#else
+    return MEM_swap64_fallback(in);
+#endif
+}
+
+MEM_STATIC size_t MEM_swapST(size_t in)
+{
+    if (MEM_32bits())
+        return (size_t)MEM_swap32((U32)in);
+    else
+        return (size_t)MEM_swap64((U64)in);
+}
+
+/*=== Little endian r/w ===*/
+
+MEM_STATIC U16 MEM_readLE16(const void* memPtr)
+{
+    if (MEM_isLittleEndian())
+        return MEM_read16(memPtr);
+    else {
+        const BYTE* p = (const BYTE*)memPtr;
+        return (U16)(p[0] + (p[1]<<8));
+    }
+}
+
+MEM_STATIC void MEM_writeLE16(void* memPtr, U16 val)
+{
+    if (MEM_isLittleEndian()) {
+        MEM_write16(memPtr, val);
+    } else {
+        BYTE* p = (BYTE*)memPtr;
+        p[0] = (BYTE)val;
+        p[1] = (BYTE)(val>>8);
+    }
+}
+
+MEM_STATIC U32 MEM_readLE24(const void* memPtr)
+{
+    return (U32)MEM_readLE16(memPtr) + ((U32)(((const BYTE*)memPtr)[2]) << 16);
+}
+
+MEM_STATIC void MEM_writeLE24(void* memPtr, U32 val)
+{
+    MEM_writeLE16(memPtr, (U16)val);
+    ((BYTE*)memPtr)[2] = (BYTE)(val>>16);
+}
+
+MEM_STATIC U32 MEM_readLE32(const void* memPtr)
+{
+    if (MEM_isLittleEndian())
+        return MEM_read32(memPtr);
+    else
+        return MEM_swap32(MEM_read32(memPtr));
+}
+
+MEM_STATIC void MEM_writeLE32(void* memPtr, U32 val32)
+{
+    if (MEM_isLittleEndian())
+        MEM_write32(memPtr, val32);
+    else
+        MEM_write32(memPtr, MEM_swap32(val32));
+}
+
+MEM_STATIC U64 MEM_readLE64(const void* memPtr)
+{
+    if (MEM_isLittleEndian())
+        return MEM_read64(memPtr);
+    else
+        return MEM_swap64(MEM_read64(memPtr));
+}
+
+MEM_STATIC void MEM_writeLE64(void* memPtr, U64 val64)
+{
+    if (MEM_isLittleEndian())
+        MEM_write64(memPtr, val64);
+    else
+        MEM_write64(memPtr, MEM_swap64(val64));
+}
+
+MEM_STATIC size_t MEM_readLEST(const void* memPtr)
+{
+    if (MEM_32bits())
+        return (size_t)MEM_readLE32(memPtr);
+    else
+        return (size_t)MEM_readLE64(memPtr);
+}
+
+MEM_STATIC void MEM_writeLEST(void* memPtr, size_t val)
+{
+    if (MEM_32bits())
+        MEM_writeLE32(memPtr, (U32)val);
+    else
+        MEM_writeLE64(memPtr, (U64)val);
+}
+
+/*=== Big endian r/w ===*/
+
+MEM_STATIC U32 MEM_readBE32(const void* memPtr)
+{
+    if (MEM_isLittleEndian())
+        return MEM_swap32(MEM_read32(memPtr));
+    else
+        return MEM_read32(memPtr);
+}
+
+MEM_STATIC void MEM_writeBE32(void* memPtr, U32 val32)
+{
+    if (MEM_isLittleEndian())
+        MEM_write32(memPtr, MEM_swap32(val32));
+    else
+        MEM_write32(memPtr, val32);
+}
+
+MEM_STATIC U64 MEM_readBE64(const void* memPtr)
+{
+    if (MEM_isLittleEndian())
+        return MEM_swap64(MEM_read64(memPtr));
+    else
+        return MEM_read64(memPtr);
+}
+
+MEM_STATIC void MEM_writeBE64(void* memPtr, U64 val64)
+{
+    if (MEM_isLittleEndian())
+        MEM_write64(memPtr, MEM_swap64(val64));
+    else
+        MEM_write64(memPtr, val64);
+}
+
+MEM_STATIC size_t MEM_readBEST(const void* memPtr)
+{
+    if (MEM_32bits())
+        return (size_t)MEM_readBE32(memPtr);
+    else
+        return (size_t)MEM_readBE64(memPtr);
+}
+
+MEM_STATIC void MEM_writeBEST(void* memPtr, size_t val)
+{
+    if (MEM_32bits())
+        MEM_writeBE32(memPtr, (U32)val);
+    else
+        MEM_writeBE64(memPtr, (U64)val);
+}
+
+/* code only tested on 32 and 64 bits systems */
+MEM_STATIC void MEM_check(void) { DEBUG_STATIC_ASSERT((sizeof(size_t)==4) || (sizeof(size_t)==8)); }
+
+
+#if defined (__cplusplus)
+}
+#endif
+
+#endif /* MEM_H_MODULE */

external/zstd/common/pool.c

@@ -0,0 +1,369 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+
+/* ======   Dependencies   ======= */
+#include "zstd_deps.h" /* size_t */
+#include "debug.h"     /* assert */
+#include "zstd_internal.h"  /* ZSTD_customCalloc, ZSTD_customFree */
+#include "pool.h"
+
+/* ======   Compiler specifics   ====== */
+#if defined(_MSC_VER)
+#  pragma warning(disable : 4204)        /* disable: C4204: non-constant aggregate initializer */
+#endif
+
+
+#ifdef ZSTD_MULTITHREAD
+
+#include "threading.h"   /* pthread adaptation */
+
+/* A job is a function and an opaque argument */
+typedef struct POOL_job_s {
+    POOL_function function;
+    void *opaque;
+} POOL_job;
+
+struct POOL_ctx_s {
+    ZSTD_customMem customMem;
+    /* Keep track of the threads */
+    ZSTD_pthread_t* threads;
+    size_t threadCapacity;
+    size_t threadLimit;
+
+    /* The queue is a circular buffer */
+    POOL_job *queue;
+    size_t queueHead;
+    size_t queueTail;
+    size_t queueSize;
+
+    /* The number of threads working on jobs */
+    size_t numThreadsBusy;
+    /* Indicates if the queue is empty */
+    int queueEmpty;
+
+    /* The mutex protects the queue */
+    ZSTD_pthread_mutex_t queueMutex;
+    /* Condition variable for pushers to wait on when the queue is full */
+    ZSTD_pthread_cond_t queuePushCond;
+    /* Condition variables for poppers to wait on when the queue is empty */
+    ZSTD_pthread_cond_t queuePopCond;
+    /* Indicates if the queue is shutting down */
+    int shutdown;
+};
+
+/* POOL_thread() :
+ * Work thread for the thread pool.
+ * Waits for jobs and executes them.
+ * @returns : NULL on failure else non-null.
+ */
+static void* POOL_thread(void* opaque) {
+    POOL_ctx* const ctx = (POOL_ctx*)opaque;
+    if (!ctx) { return NULL; }
+    for (;;) {
+        /* Lock the mutex and wait for a non-empty queue or until shutdown */
+        ZSTD_pthread_mutex_lock(&ctx->queueMutex);
+
+        while ( ctx->queueEmpty
+            || (ctx->numThreadsBusy >= ctx->threadLimit) ) {
+            if (ctx->shutdown) {
+                /* even if !queueEmpty, (possible if numThreadsBusy >= threadLimit),
+                 * a few threads will be shutdown while !queueEmpty,
+                 * but enough threads will remain active to finish the queue */
+                ZSTD_pthread_mutex_unlock(&ctx->queueMutex);
+                return opaque;
+            }
+            ZSTD_pthread_cond_wait(&ctx->queuePopCond, &ctx->queueMutex);
+        }
+        /* Pop a job off the queue */
+        {   POOL_job const job = ctx->queue[ctx->queueHead];
+            ctx->queueHead = (ctx->queueHead + 1) % ctx->queueSize;
+            ctx->numThreadsBusy++;
+            ctx->queueEmpty = (ctx->queueHead == ctx->queueTail);
+            /* Unlock the mutex, signal a pusher, and run the job */
+            ZSTD_pthread_cond_signal(&ctx->queuePushCond);
+            ZSTD_pthread_mutex_unlock(&ctx->queueMutex);
+
+            job.function(job.opaque);
+
+            /* If the intended queue size was 0, signal after finishing job */
+            ZSTD_pthread_mutex_lock(&ctx->queueMutex);
+            ctx->numThreadsBusy--;
+            ZSTD_pthread_cond_signal(&ctx->queuePushCond);
+            ZSTD_pthread_mutex_unlock(&ctx->queueMutex);
+        }
+    }  /* for (;;) */
+    assert(0);  /* Unreachable */
+}
+
+/* ZSTD_createThreadPool() : public access point */
+POOL_ctx* ZSTD_createThreadPool(size_t numThreads) {
+    return POOL_create (numThreads, 0);
+}
+
+POOL_ctx* POOL_create(size_t numThreads, size_t queueSize) {
+    return POOL_create_advanced(numThreads, queueSize, ZSTD_defaultCMem);
+}
+
+POOL_ctx* POOL_create_advanced(size_t numThreads, size_t queueSize,
+                               ZSTD_customMem customMem)
+{
+    POOL_ctx* ctx;
+    /* Check parameters */
+    if (!numThreads) { return NULL; }
+    /* Allocate the context and zero initialize */
+    ctx = (POOL_ctx*)ZSTD_customCalloc(sizeof(POOL_ctx), customMem);
+    if (!ctx) { return NULL; }
+    /* Initialize the job queue.
+     * It needs one extra space since one space is wasted to differentiate
+     * empty and full queues.
+     */
+    ctx->queueSize = queueSize + 1;
+    ctx->queue = (POOL_job*)ZSTD_customCalloc(ctx->queueSize * sizeof(POOL_job), customMem);
+    ctx->queueHead = 0;
+    ctx->queueTail = 0;
+    ctx->numThreadsBusy = 0;
+    ctx->queueEmpty = 1;
+    {
+        int error = 0;
+        error |= ZSTD_pthread_mutex_init(&ctx->queueMutex, NULL);
+        error |= ZSTD_pthread_cond_init(&ctx->queuePushCond, NULL);
+        error |= ZSTD_pthread_cond_init(&ctx->queuePopCond, NULL);
+        if (error) { POOL_free(ctx); return NULL; }
+    }
+    ctx->shutdown = 0;
+    /* Allocate space for the thread handles */
+    ctx->threads = (ZSTD_pthread_t*)ZSTD_customCalloc(numThreads * sizeof(ZSTD_pthread_t), customMem);
+    ctx->threadCapacity = 0;
+    ctx->customMem = customMem;
+    /* Check for errors */
+    if (!ctx->threads || !ctx->queue) { POOL_free(ctx); return NULL; }
+    /* Initialize the threads */
+    {   size_t i;
+        for (i = 0; i < numThreads; ++i) {
+            if (ZSTD_pthread_create(&ctx->threads[i], NULL, &POOL_thread, ctx)) {
+                ctx->threadCapacity = i;
+                POOL_free(ctx);
+                return NULL;
+        }   }
+        ctx->threadCapacity = numThreads;
+        ctx->threadLimit = numThreads;
+    }
+    return ctx;
+}
+
+/*! POOL_join() :
+    Shutdown the queue, wake any sleeping threads, and join all of the threads.
+*/
+static void POOL_join(POOL_ctx* ctx) {
+    /* Shut down the queue */
+    ZSTD_pthread_mutex_lock(&ctx->queueMutex);
+    ctx->shutdown = 1;
+    ZSTD_pthread_mutex_unlock(&ctx->queueMutex);
+    /* Wake up sleeping threads */
+    ZSTD_pthread_cond_broadcast(&ctx->queuePushCond);
+    ZSTD_pthread_cond_broadcast(&ctx->queuePopCond);
+    /* Join all of the threads */
+    {   size_t i;
+        for (i = 0; i < ctx->threadCapacity; ++i) {
+            ZSTD_pthread_join(ctx->threads[i]);  /* note : could fail */
+    }   }
+}
+
+void POOL_free(POOL_ctx *ctx) {
+    if (!ctx) { return; }
+    POOL_join(ctx);
+    ZSTD_pthread_mutex_destroy(&ctx->queueMutex);
+    ZSTD_pthread_cond_destroy(&ctx->queuePushCond);
+    ZSTD_pthread_cond_destroy(&ctx->queuePopCond);
+    ZSTD_customFree(ctx->queue, ctx->customMem);
+    ZSTD_customFree(ctx->threads, ctx->customMem);
+    ZSTD_customFree(ctx, ctx->customMem);
+}
+
+/*! POOL_joinJobs() :
+ *  Waits for all queued jobs to finish executing.
+ */
+void POOL_joinJobs(POOL_ctx* ctx) {
+    ZSTD_pthread_mutex_lock(&ctx->queueMutex);
+    while(!ctx->queueEmpty || ctx->numThreadsBusy > 0) {
+        ZSTD_pthread_cond_wait(&ctx->queuePushCond, &ctx->queueMutex);
+    }
+    ZSTD_pthread_mutex_unlock(&ctx->queueMutex);
+}
+
+void ZSTD_freeThreadPool (ZSTD_threadPool* pool) {
+  POOL_free (pool);
+}
+
+size_t POOL_sizeof(const POOL_ctx* ctx) {
+    if (ctx==NULL) return 0;  /* supports sizeof NULL */
+    return sizeof(*ctx)
+        + ctx->queueSize * sizeof(POOL_job)
+        + ctx->threadCapacity * sizeof(ZSTD_pthread_t);
+}
+
+
+/* @return : 0 on success, 1 on error */
+static int POOL_resize_internal(POOL_ctx* ctx, size_t numThreads)
+{
+    if (numThreads <= ctx->threadCapacity) {
+        if (!numThreads) return 1;
+        ctx->threadLimit = numThreads;
+        return 0;
+    }
+    /* numThreads > threadCapacity */
+    {   ZSTD_pthread_t* const threadPool = (ZSTD_pthread_t*)ZSTD_customCalloc(numThreads * sizeof(ZSTD_pthread_t), ctx->customMem);
+        if (!threadPool) return 1;
+        /* replace existing thread pool */
+        ZSTD_memcpy(threadPool, ctx->threads, ctx->threadCapacity * sizeof(*threadPool));
+        ZSTD_customFree(ctx->threads, ctx->customMem);
+        ctx->threads = threadPool;
+        /* Initialize additional threads */
+        {   size_t threadId;
+            for (threadId = ctx->threadCapacity; threadId < numThreads; ++threadId) {
+                if (ZSTD_pthread_create(&threadPool[threadId], NULL, &POOL_thread, ctx)) {
+                    ctx->threadCapacity = threadId;
+                    return 1;
+            }   }
+    }   }
+    /* successfully expanded */
+    ctx->threadCapacity = numThreads;
+    ctx->threadLimit = numThreads;
+    return 0;
+}
+
+/* @return : 0 on success, 1 on error */
+int POOL_resize(POOL_ctx* ctx, size_t numThreads)
+{
+    int result;
+    if (ctx==NULL) return 1;
+    ZSTD_pthread_mutex_lock(&ctx->queueMutex);
+    result = POOL_resize_internal(ctx, numThreads);
+    ZSTD_pthread_cond_broadcast(&ctx->queuePopCond);
+    ZSTD_pthread_mutex_unlock(&ctx->queueMutex);
+    return result;
+}
+
+/**
+ * Returns 1 if the queue is full and 0 otherwise.
+ *
+ * When queueSize is 1 (pool was created with an intended queueSize of 0),
+ * then a queue is empty if there is a thread free _and_ no job is waiting.
+ */
+static int isQueueFull(POOL_ctx const* ctx) {
+    if (ctx->queueSize > 1) {
+        return ctx->queueHead == ((ctx->queueTail + 1) % ctx->queueSize);
+    } else {
+        return (ctx->numThreadsBusy == ctx->threadLimit) ||
+               !ctx->queueEmpty;
+    }
+}
+
+
+static void
+POOL_add_internal(POOL_ctx* ctx, POOL_function function, void *opaque)
+{
+    POOL_job const job = {function, opaque};
+    assert(ctx != NULL);
+    if (ctx->shutdown) return;
+
+    ctx->queueEmpty = 0;
+    ctx->queue[ctx->queueTail] = job;
+    ctx->queueTail = (ctx->queueTail + 1) % ctx->queueSize;
+    ZSTD_pthread_cond_signal(&ctx->queuePopCond);
+}
+
+void POOL_add(POOL_ctx* ctx, POOL_function function, void* opaque)
+{
+    assert(ctx != NULL);
+    ZSTD_pthread_mutex_lock(&ctx->queueMutex);
+    /* Wait until there is space in the queue for the new job */
+    while (isQueueFull(ctx) && (!ctx->shutdown)) {
+        ZSTD_pthread_cond_wait(&ctx->queuePushCond, &ctx->queueMutex);
+    }
+    POOL_add_internal(ctx, function, opaque);
+    ZSTD_pthread_mutex_unlock(&ctx->queueMutex);
+}
+
+
+int POOL_tryAdd(POOL_ctx* ctx, POOL_function function, void* opaque)
+{
+    assert(ctx != NULL);
+    ZSTD_pthread_mutex_lock(&ctx->queueMutex);
+    if (isQueueFull(ctx)) {
+        ZSTD_pthread_mutex_unlock(&ctx->queueMutex);
+        return 0;
+    }
+    POOL_add_internal(ctx, function, opaque);
+    ZSTD_pthread_mutex_unlock(&ctx->queueMutex);
+    return 1;
+}
+
+
+#else  /* ZSTD_MULTITHREAD  not defined */
+
+/* ========================== */
+/* No multi-threading support */
+/* ========================== */
+
+
+/* We don't need any data, but if it is empty, malloc() might return NULL. */
+struct POOL_ctx_s {
+    int dummy;
+};
+static POOL_ctx g_poolCtx;
+
+POOL_ctx* POOL_create(size_t numThreads, size_t queueSize) {
+    return POOL_create_advanced(numThreads, queueSize, ZSTD_defaultCMem);
+}
+
+POOL_ctx*
+POOL_create_advanced(size_t numThreads, size_t queueSize, ZSTD_customMem customMem)
+{
+    (void)numThreads;
+    (void)queueSize;
+    (void)customMem;
+    return &g_poolCtx;
+}
+
+void POOL_free(POOL_ctx* ctx) {
+    assert(!ctx || ctx == &g_poolCtx);
+    (void)ctx;
+}
+
+void POOL_joinJobs(POOL_ctx* ctx){
+    assert(!ctx || ctx == &g_poolCtx);
+    (void)ctx;
+}
+
+int POOL_resize(POOL_ctx* ctx, size_t numThreads) {
+    (void)ctx; (void)numThreads;
+    return 0;
+}
+
+void POOL_add(POOL_ctx* ctx, POOL_function function, void* opaque) {
+    (void)ctx;
+    function(opaque);
+}
+
+int POOL_tryAdd(POOL_ctx* ctx, POOL_function function, void* opaque) {
+    (void)ctx;
+    function(opaque);
+    return 1;
+}
+
+size_t POOL_sizeof(const POOL_ctx* ctx) {
+    if (ctx==NULL) return 0;  /* supports sizeof NULL */
+    assert(ctx == &g_poolCtx);
+    return sizeof(*ctx);
+}
+
+#endif  /* ZSTD_MULTITHREAD */

external/zstd/common/pool.h

@@ -0,0 +1,90 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef POOL_H
+#define POOL_H
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+
+#include "zstd_deps.h"
+#define ZSTD_STATIC_LINKING_ONLY   /* ZSTD_customMem */
+#include "../zstd.h"
+
+typedef struct POOL_ctx_s POOL_ctx;
+
+/*! POOL_create() :
+ *  Create a thread pool with at most `numThreads` threads.
+ * `numThreads` must be at least 1.
+ *  The maximum number of queued jobs before blocking is `queueSize`.
+ * @return : POOL_ctx pointer on success, else NULL.
+*/
+POOL_ctx* POOL_create(size_t numThreads, size_t queueSize);
+
+POOL_ctx* POOL_create_advanced(size_t numThreads, size_t queueSize,
+                               ZSTD_customMem customMem);
+
+/*! POOL_free() :
+ *  Free a thread pool returned by POOL_create().
+ */
+void POOL_free(POOL_ctx* ctx);
+
+
+/*! POOL_joinJobs() :
+ *  Waits for all queued jobs to finish executing.
+ */
+void POOL_joinJobs(POOL_ctx* ctx);
+
+/*! POOL_resize() :
+ *  Expands or shrinks pool's number of threads.
+ *  This is more efficient than releasing + creating a new context,
+ *  since it tries to preserve and re-use existing threads.
+ * `numThreads` must be at least 1.
+ * @return : 0 when resize was successful,
+ *           !0 (typically 1) if there is an error.
+ *    note : only numThreads can be resized, queueSize remains unchanged.
+ */
+int POOL_resize(POOL_ctx* ctx, size_t numThreads);
+
+/*! POOL_sizeof() :
+ * @return threadpool memory usage
+ *  note : compatible with NULL (returns 0 in this case)
+ */
+size_t POOL_sizeof(const POOL_ctx* ctx);
+
+/*! POOL_function :
+ *  The function type that can be added to a thread pool.
+ */
+typedef void (*POOL_function)(void*);
+
+/*! POOL_add() :
+ *  Add the job `function(opaque)` to the thread pool. `ctx` must be valid.
+ *  Possibly blocks until there is room in the queue.
+ *  Note : The function may be executed asynchronously,
+ *         therefore, `opaque` must live until function has been completed.
+ */
+void POOL_add(POOL_ctx* ctx, POOL_function function, void* opaque);
+
+
+/*! POOL_tryAdd() :
+ *  Add the job `function(opaque)` to thread pool _if_ a queue slot is available.
+ *  Returns immediately even if not (does not block).
+ * @return : 1 if successful, 0 if not.
+ */
+int POOL_tryAdd(POOL_ctx* ctx, POOL_function function, void* opaque);
+
+
+#if defined (__cplusplus)
+}
+#endif
+
+#endif

external/zstd/common/portability_macros.h

@@ -0,0 +1,148 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_PORTABILITY_MACROS_H
+#define ZSTD_PORTABILITY_MACROS_H
+
+/**
+ * This header file contains macro definitions to support portability.
+ * This header is shared between C and ASM code, so it MUST only
+ * contain macro definitions. It MUST not contain any C code.
+ *
+ * This header ONLY defines macros to detect platforms/feature support.
+ *
+ */
+
+
+/* compat. with non-clang compilers */
+#ifndef __has_attribute
+  #define __has_attribute(x) 0
+#endif
+
+/* compat. with non-clang compilers */
+#ifndef __has_builtin
+#  define __has_builtin(x) 0
+#endif
+
+/* compat. with non-clang compilers */
+#ifndef __has_feature
+#  define __has_feature(x) 0
+#endif
+
+/* detects whether we are being compiled under msan */
+#ifndef ZSTD_MEMORY_SANITIZER
+#  if __has_feature(memory_sanitizer)
+#    define ZSTD_MEMORY_SANITIZER 1
+#  else
+#    define ZSTD_MEMORY_SANITIZER 0
+#  endif
+#endif
+
+/* detects whether we are being compiled under asan */
+#ifndef ZSTD_ADDRESS_SANITIZER
+#  if __has_feature(address_sanitizer)
+#    define ZSTD_ADDRESS_SANITIZER 1
+#  elif defined(__SANITIZE_ADDRESS__)
+#    define ZSTD_ADDRESS_SANITIZER 1
+#  else
+#    define ZSTD_ADDRESS_SANITIZER 0
+#  endif
+#endif
+
+/* detects whether we are being compiled under dfsan */
+#ifndef ZSTD_DATAFLOW_SANITIZER
+# if __has_feature(dataflow_sanitizer)
+#  define ZSTD_DATAFLOW_SANITIZER 1
+# else
+#  define ZSTD_DATAFLOW_SANITIZER 0
+# endif
+#endif
+
+/* Mark the internal assembly functions as hidden  */
+#ifdef __ELF__
+# define ZSTD_HIDE_ASM_FUNCTION(func) .hidden func
+#else
+# define ZSTD_HIDE_ASM_FUNCTION(func)
+#endif
+
+/* Enable runtime BMI2 dispatch based on the CPU.
+ * Enabled for clang & gcc >=4.8 on x86 when BMI2 isn't enabled by default.
+ */
+#ifndef DYNAMIC_BMI2
+  #if ((defined(__clang__) && __has_attribute(__target__)) \
+      || (defined(__GNUC__) \
+          && (__GNUC__ >= 5 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 8)))) \
+      && (defined(__x86_64__) || defined(_M_X64)) \
+      && !defined(__BMI2__)
+  #  define DYNAMIC_BMI2 1
+  #else
+  #  define DYNAMIC_BMI2 0
+  #endif
+#endif
+
+/**
+ * Only enable assembly for GNUC compatible compilers,
+ * because other platforms may not support GAS assembly syntax.
+ *
+ * Only enable assembly for Linux / MacOS, other platforms may
+ * work, but they haven't been tested. This could likely be
+ * extended to BSD systems.
+ *
+ * Disable assembly when MSAN is enabled, because MSAN requires
+ * 100% of code to be instrumented to work.
+ */
+#if defined(__GNUC__)
+#  if defined(__linux__) || defined(__linux) || defined(__APPLE__)
+#    if ZSTD_MEMORY_SANITIZER
+#      define ZSTD_ASM_SUPPORTED 0
+#    elif ZSTD_DATAFLOW_SANITIZER
+#      define ZSTD_ASM_SUPPORTED 0
+#    else
+#      define ZSTD_ASM_SUPPORTED 1
+#    endif
+#  else
+#    define ZSTD_ASM_SUPPORTED 0
+#  endif
+#else
+#  define ZSTD_ASM_SUPPORTED 0
+#endif
+
+/**
+ * Determines whether we should enable assembly for x86-64
+ * with BMI2.
+ *
+ * Enable if all of the following conditions hold:
+ * - ASM hasn't been explicitly disabled by defining ZSTD_DISABLE_ASM
+ * - Assembly is supported
+ * - We are compiling for x86-64 and either:
+ *   - DYNAMIC_BMI2 is enabled
+ *   - BMI2 is supported at compile time
+ */
+#if !defined(ZSTD_DISABLE_ASM) &&                                 \
+    ZSTD_ASM_SUPPORTED &&                                         \
+    defined(__x86_64__) &&                                        \
+    (DYNAMIC_BMI2 || defined(__BMI2__))
+# define ZSTD_ENABLE_ASM_X86_64_BMI2 1
+#else
+# define ZSTD_ENABLE_ASM_X86_64_BMI2 0
+#endif
+
+/*
+ * For x86 ELF targets, add .note.gnu.property section for Intel CET in
+ * assembly sources when CET is enabled.
+ */
+#if defined(__ELF__) && (defined(__x86_64__) || defined(__i386__)) \
+    && defined(__has_include)
+# if __has_include(<cet.h>)
+#  include <cet.h>
+# endif
+#endif
+
+#endif /* ZSTD_PORTABILITY_MACROS_H */

external/zstd/common/threading.c

@@ -0,0 +1,176 @@
+/**
+ * Copyright (c) 2016 Tino Reichardt
+ * All rights reserved.
+ *
+ * You can contact the author at:
+ * - zstdmt source repository: https://github.com/mcmilk/zstdmt
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+/**
+ * This file will hold wrapper for systems, which do not support pthreads
+ */
+
+#include "threading.h"
+
+/* create fake symbol to avoid empty translation unit warning */
+int g_ZSTD_threading_useless_symbol;
+
+#if defined(ZSTD_MULTITHREAD) && defined(_WIN32)
+
+/**
+ * Windows minimalist Pthread Wrapper
+ */
+
+
+/* ===  Dependencies  === */
+#include <process.h>
+#include <errno.h>
+
+
+/* ===  Implementation  === */
+
+typedef struct {
+    void* (*start_routine)(void*);
+    void* arg;
+    int initialized;
+    ZSTD_pthread_cond_t initialized_cond;
+    ZSTD_pthread_mutex_t initialized_mutex;
+} ZSTD_thread_params_t;
+
+static unsigned __stdcall worker(void *arg)
+{
+    void* (*start_routine)(void*);
+    void* thread_arg;
+
+    /* Inialized thread_arg and start_routine and signal main thread that we don't need it
+     * to wait any longer.
+     */
+    {
+        ZSTD_thread_params_t*  thread_param = (ZSTD_thread_params_t*)arg;
+        thread_arg = thread_param->arg;
+        start_routine = thread_param->start_routine;
+
+        /* Signal main thread that we are running and do not depend on its memory anymore */
+        ZSTD_pthread_mutex_lock(&thread_param->initialized_mutex);
+        thread_param->initialized = 1;
+        ZSTD_pthread_cond_signal(&thread_param->initialized_cond);
+        ZSTD_pthread_mutex_unlock(&thread_param->initialized_mutex);
+    }
+
+    start_routine(thread_arg);
+
+    return 0;
+}
+
+int ZSTD_pthread_create(ZSTD_pthread_t* thread, const void* unused,
+            void* (*start_routine) (void*), void* arg)
+{
+    ZSTD_thread_params_t thread_param;
+    (void)unused;
+
+    thread_param.start_routine = start_routine;
+    thread_param.arg = arg;
+    thread_param.initialized = 0;
+    *thread = NULL;
+
+    /* Setup thread initialization synchronization */
+    if(ZSTD_pthread_cond_init(&thread_param.initialized_cond, NULL)) {
+        /* Should never happen on Windows */
+        return -1;
+    }
+    if(ZSTD_pthread_mutex_init(&thread_param.initialized_mutex, NULL)) {
+        /* Should never happen on Windows */
+        ZSTD_pthread_cond_destroy(&thread_param.initialized_cond);
+        return -1;
+    }
+
+    /* Spawn thread */
+    *thread = (HANDLE)_beginthreadex(NULL, 0, worker, &thread_param, 0, NULL);
+    if (!thread) {
+        ZSTD_pthread_mutex_destroy(&thread_param.initialized_mutex);
+        ZSTD_pthread_cond_destroy(&thread_param.initialized_cond);
+        return errno;
+    }
+
+    /* Wait for thread to be initialized */
+    ZSTD_pthread_mutex_lock(&thread_param.initialized_mutex);
+    while(!thread_param.initialized) {
+        ZSTD_pthread_cond_wait(&thread_param.initialized_cond, &thread_param.initialized_mutex);
+    }
+    ZSTD_pthread_mutex_unlock(&thread_param.initialized_mutex);
+    ZSTD_pthread_mutex_destroy(&thread_param.initialized_mutex);
+    ZSTD_pthread_cond_destroy(&thread_param.initialized_cond);
+
+    return 0;
+}
+
+int ZSTD_pthread_join(ZSTD_pthread_t thread)
+{
+    DWORD result;
+
+    if (!thread) return 0;
+
+    result = WaitForSingleObject(thread, INFINITE);
+    CloseHandle(thread);
+
+    switch (result) {
+    case WAIT_OBJECT_0:
+        return 0;
+    case WAIT_ABANDONED:
+        return EINVAL;
+    default:
+        return GetLastError();
+    }
+}
+
+#endif   /* ZSTD_MULTITHREAD */
+
+#if defined(ZSTD_MULTITHREAD) && DEBUGLEVEL >= 1 && !defined(_WIN32)
+
+#define ZSTD_DEPS_NEED_MALLOC
+#include "zstd_deps.h"
+
+int ZSTD_pthread_mutex_init(ZSTD_pthread_mutex_t* mutex, pthread_mutexattr_t const* attr)
+{
+    *mutex = (pthread_mutex_t*)ZSTD_malloc(sizeof(pthread_mutex_t));
+    if (!*mutex)
+        return 1;
+    return pthread_mutex_init(*mutex, attr);
+}
+
+int ZSTD_pthread_mutex_destroy(ZSTD_pthread_mutex_t* mutex)
+{
+    if (!*mutex)
+        return 0;
+    {
+        int const ret = pthread_mutex_destroy(*mutex);
+        ZSTD_free(*mutex);
+        return ret;
+    }
+}
+
+int ZSTD_pthread_cond_init(ZSTD_pthread_cond_t* cond, pthread_condattr_t const* attr)
+{
+    *cond = (pthread_cond_t*)ZSTD_malloc(sizeof(pthread_cond_t));
+    if (!*cond)
+        return 1;
+    return pthread_cond_init(*cond, attr);
+}
+
+int ZSTD_pthread_cond_destroy(ZSTD_pthread_cond_t* cond)
+{
+    if (!*cond)
+        return 0;
+    {
+        int const ret = pthread_cond_destroy(*cond);
+        ZSTD_free(*cond);
+        return ret;
+    }
+}
+
+#endif

external/zstd/common/threading.h

@@ -0,0 +1,150 @@
+/**
+ * Copyright (c) 2016 Tino Reichardt
+ * All rights reserved.
+ *
+ * You can contact the author at:
+ * - zstdmt source repository: https://github.com/mcmilk/zstdmt
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef THREADING_H_938743
+#define THREADING_H_938743
+
+#include "debug.h"
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+#if defined(ZSTD_MULTITHREAD) && defined(_WIN32)
+
+/**
+ * Windows minimalist Pthread Wrapper
+ */
+#ifdef WINVER
+#  undef WINVER
+#endif
+#define WINVER       0x0600
+
+#ifdef _WIN32_WINNT
+#  undef _WIN32_WINNT
+#endif
+#define _WIN32_WINNT 0x0600
+
+#ifndef WIN32_LEAN_AND_MEAN
+#  define WIN32_LEAN_AND_MEAN
+#endif
+
+#undef ERROR   /* reported already defined on VS 2015 (Rich Geldreich) */
+#include <windows.h>
+#undef ERROR
+#define ERROR(name) ZSTD_ERROR(name)
+
+
+/* mutex */
+#define ZSTD_pthread_mutex_t           CRITICAL_SECTION
+#define ZSTD_pthread_mutex_init(a, b)  ((void)(b), InitializeCriticalSection((a)), 0)
+#define ZSTD_pthread_mutex_destroy(a)  DeleteCriticalSection((a))
+#define ZSTD_pthread_mutex_lock(a)     EnterCriticalSection((a))
+#define ZSTD_pthread_mutex_unlock(a)   LeaveCriticalSection((a))
+
+/* condition variable */
+#define ZSTD_pthread_cond_t             CONDITION_VARIABLE
+#define ZSTD_pthread_cond_init(a, b)    ((void)(b), InitializeConditionVariable((a)), 0)
+#define ZSTD_pthread_cond_destroy(a)    ((void)(a))
+#define ZSTD_pthread_cond_wait(a, b)    SleepConditionVariableCS((a), (b), INFINITE)
+#define ZSTD_pthread_cond_signal(a)     WakeConditionVariable((a))
+#define ZSTD_pthread_cond_broadcast(a)  WakeAllConditionVariable((a))
+
+/* ZSTD_pthread_create() and ZSTD_pthread_join() */
+typedef HANDLE ZSTD_pthread_t;
+
+int ZSTD_pthread_create(ZSTD_pthread_t* thread, const void* unused,
+                   void* (*start_routine) (void*), void* arg);
+
+int ZSTD_pthread_join(ZSTD_pthread_t thread);
+
+/**
+ * add here more wrappers as required
+ */
+
+
+#elif defined(ZSTD_MULTITHREAD)    /* posix assumed ; need a better detection method */
+/* ===   POSIX Systems   === */
+#  include <pthread.h>
+
+#if DEBUGLEVEL < 1
+
+#define ZSTD_pthread_mutex_t            pthread_mutex_t
+#define ZSTD_pthread_mutex_init(a, b)   pthread_mutex_init((a), (b))
+#define ZSTD_pthread_mutex_destroy(a)   pthread_mutex_destroy((a))
+#define ZSTD_pthread_mutex_lock(a)      pthread_mutex_lock((a))
+#define ZSTD_pthread_mutex_unlock(a)    pthread_mutex_unlock((a))
+
+#define ZSTD_pthread_cond_t             pthread_cond_t
+#define ZSTD_pthread_cond_init(a, b)    pthread_cond_init((a), (b))
+#define ZSTD_pthread_cond_destroy(a)    pthread_cond_destroy((a))
+#define ZSTD_pthread_cond_wait(a, b)    pthread_cond_wait((a), (b))
+#define ZSTD_pthread_cond_signal(a)     pthread_cond_signal((a))
+#define ZSTD_pthread_cond_broadcast(a)  pthread_cond_broadcast((a))
+
+#define ZSTD_pthread_t                  pthread_t
+#define ZSTD_pthread_create(a, b, c, d) pthread_create((a), (b), (c), (d))
+#define ZSTD_pthread_join(a)         pthread_join((a),NULL)
+
+#else /* DEBUGLEVEL >= 1 */
+
+/* Debug implementation of threading.
+ * In this implementation we use pointers for mutexes and condition variables.
+ * This way, if we forget to init/destroy them the program will crash or ASAN
+ * will report leaks.
+ */
+
+#define ZSTD_pthread_mutex_t            pthread_mutex_t*
+int ZSTD_pthread_mutex_init(ZSTD_pthread_mutex_t* mutex, pthread_mutexattr_t const* attr);
+int ZSTD_pthread_mutex_destroy(ZSTD_pthread_mutex_t* mutex);
+#define ZSTD_pthread_mutex_lock(a)      pthread_mutex_lock(*(a))
+#define ZSTD_pthread_mutex_unlock(a)    pthread_mutex_unlock(*(a))
+
+#define ZSTD_pthread_cond_t             pthread_cond_t*
+int ZSTD_pthread_cond_init(ZSTD_pthread_cond_t* cond, pthread_condattr_t const* attr);
+int ZSTD_pthread_cond_destroy(ZSTD_pthread_cond_t* cond);
+#define ZSTD_pthread_cond_wait(a, b)    pthread_cond_wait(*(a), *(b))
+#define ZSTD_pthread_cond_signal(a)     pthread_cond_signal(*(a))
+#define ZSTD_pthread_cond_broadcast(a)  pthread_cond_broadcast(*(a))
+
+#define ZSTD_pthread_t                  pthread_t
+#define ZSTD_pthread_create(a, b, c, d) pthread_create((a), (b), (c), (d))
+#define ZSTD_pthread_join(a)         pthread_join((a),NULL)
+
+#endif
+
+#else  /* ZSTD_MULTITHREAD not defined */
+/* No multithreading support */
+
+typedef int ZSTD_pthread_mutex_t;
+#define ZSTD_pthread_mutex_init(a, b)   ((void)(a), (void)(b), 0)
+#define ZSTD_pthread_mutex_destroy(a)   ((void)(a))
+#define ZSTD_pthread_mutex_lock(a)      ((void)(a))
+#define ZSTD_pthread_mutex_unlock(a)    ((void)(a))
+
+typedef int ZSTD_pthread_cond_t;
+#define ZSTD_pthread_cond_init(a, b)    ((void)(a), (void)(b), 0)
+#define ZSTD_pthread_cond_destroy(a)    ((void)(a))
+#define ZSTD_pthread_cond_wait(a, b)    ((void)(a), (void)(b))
+#define ZSTD_pthread_cond_signal(a)     ((void)(a))
+#define ZSTD_pthread_cond_broadcast(a)  ((void)(a))
+
+/* do not use ZSTD_pthread_t */
+
+#endif /* ZSTD_MULTITHREAD */
+
+#if defined (__cplusplus)
+}
+#endif
+
+#endif /* THREADING_H_938743 */

external/zstd/common/xxhash.c

@@ -0,0 +1,24 @@
+/*
+ *  xxHash - Fast Hash algorithm
+ *  Copyright (c) Meta Platforms, Inc. and affiliates.
+ *
+ *  You can contact the author at :
+ *  - xxHash homepage: https://cyan4973.github.io/xxHash/
+ *  - xxHash source repository : https://github.com/Cyan4973/xxHash
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+*/
+
+
+
+/*
+ * xxhash.c instantiates functions defined in xxhash.h
+ */
+
+#define XXH_STATIC_LINKING_ONLY   /* access advanced declarations */
+#define XXH_IMPLEMENTATION   /* access definitions */
+
+#include "xxhash.h"

external/zstd/common/zstd_common.c

@@ -0,0 +1,83 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+
+
+/*-*************************************
+*  Dependencies
+***************************************/
+#define ZSTD_DEPS_NEED_MALLOC
+#include "zstd_deps.h"   /* ZSTD_malloc, ZSTD_calloc, ZSTD_free, ZSTD_memset */
+#include "error_private.h"
+#include "zstd_internal.h"
+
+
+/*-****************************************
+*  Version
+******************************************/
+unsigned ZSTD_versionNumber(void) { return ZSTD_VERSION_NUMBER; }
+
+const char* ZSTD_versionString(void) { return ZSTD_VERSION_STRING; }
+
+
+/*-****************************************
+*  ZSTD Error Management
+******************************************/
+#undef ZSTD_isError   /* defined within zstd_internal.h */
+/*! ZSTD_isError() :
+ *  tells if a return value is an error code
+ *  symbol is required for external callers */
+unsigned ZSTD_isError(size_t code) { return ERR_isError(code); }
+
+/*! ZSTD_getErrorName() :
+ *  provides error code string from function result (useful for debugging) */
+const char* ZSTD_getErrorName(size_t code) { return ERR_getErrorName(code); }
+
+/*! ZSTD_getError() :
+ *  convert a `size_t` function result into a proper ZSTD_errorCode enum */
+ZSTD_ErrorCode ZSTD_getErrorCode(size_t code) { return ERR_getErrorCode(code); }
+
+/*! ZSTD_getErrorString() :
+ *  provides error code string from enum */
+const char* ZSTD_getErrorString(ZSTD_ErrorCode code) { return ERR_getErrorString(code); }
+
+
+
+/*=**************************************************************
+*  Custom allocator
+****************************************************************/
+void* ZSTD_customMalloc(size_t size, ZSTD_customMem customMem)
+{
+    if (customMem.customAlloc)
+        return customMem.customAlloc(customMem.opaque, size);
+    return ZSTD_malloc(size);
+}
+
+void* ZSTD_customCalloc(size_t size, ZSTD_customMem customMem)
+{
+    if (customMem.customAlloc) {
+        /* calloc implemented as malloc+memset;
+         * not as efficient as calloc, but next best guess for custom malloc */
+        void* const ptr = customMem.customAlloc(customMem.opaque, size);
+        ZSTD_memset(ptr, 0, size);
+        return ptr;
+    }
+    return ZSTD_calloc(1, size);
+}
+
+void ZSTD_customFree(void* ptr, ZSTD_customMem customMem)
+{
+    if (ptr!=NULL) {
+        if (customMem.customFree)
+            customMem.customFree(customMem.opaque, ptr);
+        else
+            ZSTD_free(ptr);
+    }
+}

external/zstd/common/zstd_deps.h

@@ -0,0 +1,111 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+/* This file provides common libc dependencies that zstd requires.
+ * The purpose is to allow replacing this file with a custom implementation
+ * to compile zstd without libc support.
+ */
+
+/* Need:
+ * NULL
+ * INT_MAX
+ * UINT_MAX
+ * ZSTD_memcpy()
+ * ZSTD_memset()
+ * ZSTD_memmove()
+ */
+#ifndef ZSTD_DEPS_COMMON
+#define ZSTD_DEPS_COMMON
+
+#include <limits.h>
+#include <stddef.h>
+#include <string.h>
+
+#if defined(__GNUC__) && __GNUC__ >= 4
+# define ZSTD_memcpy(d,s,l) __builtin_memcpy((d),(s),(l))
+# define ZSTD_memmove(d,s,l) __builtin_memmove((d),(s),(l))
+# define ZSTD_memset(p,v,l) __builtin_memset((p),(v),(l))
+#else
+# define ZSTD_memcpy(d,s,l) memcpy((d),(s),(l))
+# define ZSTD_memmove(d,s,l) memmove((d),(s),(l))
+# define ZSTD_memset(p,v,l) memset((p),(v),(l))
+#endif
+
+#endif /* ZSTD_DEPS_COMMON */
+
+/* Need:
+ * ZSTD_malloc()
+ * ZSTD_free()
+ * ZSTD_calloc()
+ */
+#ifdef ZSTD_DEPS_NEED_MALLOC
+#ifndef ZSTD_DEPS_MALLOC
+#define ZSTD_DEPS_MALLOC
+
+#include <stdlib.h>
+
+#define ZSTD_malloc(s) malloc(s)
+#define ZSTD_calloc(n,s) calloc((n), (s))
+#define ZSTD_free(p) free((p))
+
+#endif /* ZSTD_DEPS_MALLOC */
+#endif /* ZSTD_DEPS_NEED_MALLOC */
+
+/*
+ * Provides 64-bit math support.
+ * Need:
+ * U64 ZSTD_div64(U64 dividend, U32 divisor)
+ */
+#ifdef ZSTD_DEPS_NEED_MATH64
+#ifndef ZSTD_DEPS_MATH64
+#define ZSTD_DEPS_MATH64
+
+#define ZSTD_div64(dividend, divisor) ((dividend) / (divisor))
+
+#endif /* ZSTD_DEPS_MATH64 */
+#endif /* ZSTD_DEPS_NEED_MATH64 */
+
+/* Need:
+ * assert()
+ */
+#ifdef ZSTD_DEPS_NEED_ASSERT
+#ifndef ZSTD_DEPS_ASSERT
+#define ZSTD_DEPS_ASSERT
+
+#include <assert.h>
+
+#endif /* ZSTD_DEPS_ASSERT */
+#endif /* ZSTD_DEPS_NEED_ASSERT */
+
+/* Need:
+ * ZSTD_DEBUG_PRINT()
+ */
+#ifdef ZSTD_DEPS_NEED_IO
+#ifndef ZSTD_DEPS_IO
+#define ZSTD_DEPS_IO
+
+#include <stdio.h>
+#define ZSTD_DEBUG_PRINT(...) fprintf(stderr, __VA_ARGS__)
+
+#endif /* ZSTD_DEPS_IO */
+#endif /* ZSTD_DEPS_NEED_IO */
+
+/* Only requested when <stdint.h> is known to be present.
+ * Need:
+ * intptr_t
+ */
+#ifdef ZSTD_DEPS_NEED_STDINT
+#ifndef ZSTD_DEPS_STDINT
+#define ZSTD_DEPS_STDINT
+
+#include <stdint.h>
+
+#endif /* ZSTD_DEPS_STDINT */
+#endif /* ZSTD_DEPS_NEED_STDINT */

external/zstd/common/zstd_internal.h

@@ -0,0 +1,396 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_CCOMMON_H_MODULE
+#define ZSTD_CCOMMON_H_MODULE
+
+/* this module contains definitions which must be identical
+ * across compression, decompression and dictBuilder.
+ * It also contains a few functions useful to at least 2 of them
+ * and which benefit from being inlined */
+
+/*-*************************************
+*  Dependencies
+***************************************/
+#include "compiler.h"
+#include "cpu.h"
+#include "mem.h"
+#include "debug.h"                 /* assert, DEBUGLOG, RAWLOG, g_debuglevel */
+#include "error_private.h"
+#define ZSTD_STATIC_LINKING_ONLY
+#include "../zstd.h"
+#define FSE_STATIC_LINKING_ONLY
+#include "fse.h"
+#define HUF_STATIC_LINKING_ONLY
+#include "huf.h"
+#ifndef XXH_STATIC_LINKING_ONLY
+#  define XXH_STATIC_LINKING_ONLY  /* XXH64_state_t */
+#endif
+#include "xxhash.h"                /* XXH_reset, update, digest */
+#ifndef ZSTD_NO_TRACE
+#  include "zstd_trace.h"
+#else
+#  define ZSTD_TRACE 0
+#endif
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+/* ---- static assert (debug) --- */
+#define ZSTD_STATIC_ASSERT(c) DEBUG_STATIC_ASSERT(c)
+#define ZSTD_isError ERR_isError   /* for inlining */
+#define FSE_isError  ERR_isError
+#define HUF_isError  ERR_isError
+
+
+/*-*************************************
+*  shared macros
+***************************************/
+#undef MIN
+#undef MAX
+#define MIN(a,b) ((a)<(b) ? (a) : (b))
+#define MAX(a,b) ((a)>(b) ? (a) : (b))
+#define BOUNDED(min,val,max) (MAX(min,MIN(val,max)))
+
+
+/*-*************************************
+*  Common constants
+***************************************/
+#define ZSTD_OPT_NUM    (1<<12)
+
+#define ZSTD_REP_NUM      3                 /* number of repcodes */
+static UNUSED_ATTR 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 UNUSED_ATTR const size_t ZSTD_fcs_fieldSize[4] = { 0, 2, 4, 8 };
+static UNUSED_ATTR const size_t ZSTD_did_fieldSize[4] = { 0, 1, 2, 4 };
+
+#define ZSTD_FRAMEIDSIZE 4   /* magic number size */
+
+#define ZSTD_BLOCKHEADERSIZE 3   /* C standard doesn't allow `static const` variable to be init using another `static const` variable */
+static UNUSED_ATTR const size_t ZSTD_blockHeaderSize = ZSTD_BLOCKHEADERSIZE;
+typedef enum { bt_raw, bt_rle, bt_compressed, bt_reserved } blockType_e;
+
+#define ZSTD_FRAMECHECKSUMSIZE 4
+
+#define MIN_SEQUENCES_SIZE 1 /* nbSeq==0 */
+#define MIN_CBLOCK_SIZE (1 /*litCSize*/ + 1 /* RLE or RAW */)   /* for a non-null block */
+#define MIN_LITERALS_FOR_4_STREAMS 6
+
+typedef enum { set_basic, set_rle, set_compressed, set_repeat } symbolEncodingType_e;
+
+#define LONGNBSEQ 0x7F00
+
+#define MINMATCH 3
+
+#define Litbits  8
+#define LitHufLog 11
+#define MaxLit ((1<<Litbits) - 1)
+#define MaxML   52
+#define MaxLL   35
+#define DefaultMaxOff 28
+#define MaxOff  31
+#define MaxSeq MAX(MaxLL, MaxML)   /* Assumption : MaxOff < MaxLL,MaxML */
+#define MLFSELog    9
+#define LLFSELog    9
+#define OffFSELog   8
+#define MaxFSELog  MAX(MAX(MLFSELog, LLFSELog), OffFSELog)
+
+#define ZSTD_MAX_HUF_HEADER_SIZE 128 /* header + <= 127 byte tree description */
+/* Each table cannot take more than #symbols * FSELog bits */
+#define ZSTD_MAX_FSE_HEADERS_SIZE (((MaxML + 1) * MLFSELog + (MaxLL + 1) * LLFSELog + (MaxOff + 1) * OffFSELog + 7) / 8)
+
+static UNUSED_ATTR const U8 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 UNUSED_ATTR 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 UNUSED_ATTR const U32 LL_defaultNormLog = LL_DEFAULTNORMLOG;
+
+static UNUSED_ATTR const U8 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 UNUSED_ATTR 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 UNUSED_ATTR const U32 ML_defaultNormLog = ML_DEFAULTNORMLOG;
+
+static UNUSED_ATTR const S16 OF_defaultNorm[DefaultMaxOff+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 UNUSED_ATTR const U32 OF_defaultNormLog = OF_DEFAULTNORMLOG;
+
+
+/*-*******************************************
+*  Shared functions to include for inlining
+*********************************************/
+static void ZSTD_copy8(void* dst, const void* src) {
+#if defined(ZSTD_ARCH_ARM_NEON)
+    vst1_u8((uint8_t*)dst, vld1_u8((const uint8_t*)src));
+#else
+    ZSTD_memcpy(dst, src, 8);
+#endif
+}
+#define COPY8(d,s) { ZSTD_copy8(d,s); d+=8; s+=8; }
+
+/* Need to use memmove here since the literal buffer can now be located within
+   the dst buffer. In circumstances where the op "catches up" to where the
+   literal buffer is, there can be partial overlaps in this call on the final
+   copy if the literal is being shifted by less than 16 bytes. */
+static void ZSTD_copy16(void* dst, const void* src) {
+#if defined(ZSTD_ARCH_ARM_NEON)
+    vst1q_u8((uint8_t*)dst, vld1q_u8((const uint8_t*)src));
+#elif defined(ZSTD_ARCH_X86_SSE2)
+    _mm_storeu_si128((__m128i*)dst, _mm_loadu_si128((const __m128i*)src));
+#elif defined(__clang__)
+    ZSTD_memmove(dst, src, 16);
+#else
+    /* ZSTD_memmove is not inlined properly by gcc */
+    BYTE copy16_buf[16];
+    ZSTD_memcpy(copy16_buf, src, 16);
+    ZSTD_memcpy(dst, copy16_buf, 16);
+#endif
+}
+#define COPY16(d,s) { ZSTD_copy16(d,s); d+=16; s+=16; }
+
+#define WILDCOPY_OVERLENGTH 32
+#define WILDCOPY_VECLEN 16
+
+typedef enum {
+    ZSTD_no_overlap,
+    ZSTD_overlap_src_before_dst
+    /*  ZSTD_overlap_dst_before_src, */
+} ZSTD_overlap_e;
+
+/*! ZSTD_wildcopy() :
+ *  Custom version of ZSTD_memcpy(), can over read/write up to WILDCOPY_OVERLENGTH bytes (if length==0)
+ *  @param ovtype controls the overlap detection
+ *         - ZSTD_no_overlap: The source and destination are guaranteed to be at least WILDCOPY_VECLEN bytes apart.
+ *         - ZSTD_overlap_src_before_dst: The src and dst may overlap, but they MUST be at least 8 bytes apart.
+ *           The src buffer must be before the dst buffer.
+ */
+MEM_STATIC FORCE_INLINE_ATTR
+void ZSTD_wildcopy(void* dst, const void* src, ptrdiff_t length, ZSTD_overlap_e const ovtype)
+{
+    ptrdiff_t diff = (BYTE*)dst - (const BYTE*)src;
+    const BYTE* ip = (const BYTE*)src;
+    BYTE* op = (BYTE*)dst;
+    BYTE* const oend = op + length;
+
+    if (ovtype == ZSTD_overlap_src_before_dst && diff < WILDCOPY_VECLEN) {
+        /* Handle short offset copies. */
+        do {
+            COPY8(op, ip)
+        } while (op < oend);
+    } else {
+        assert(diff >= WILDCOPY_VECLEN || diff <= -WILDCOPY_VECLEN);
+        /* Separate out the first COPY16() call because the copy length is
+         * almost certain to be short, so the branches have different
+         * probabilities. Since it is almost certain to be short, only do
+         * one COPY16() in the first call. Then, do two calls per loop since
+         * at that point it is more likely to have a high trip count.
+         */
+        ZSTD_copy16(op, ip);
+        if (16 >= length) return;
+        op += 16;
+        ip += 16;
+        do {
+            COPY16(op, ip);
+            COPY16(op, ip);
+        }
+        while (op < oend);
+    }
+}
+
+MEM_STATIC size_t ZSTD_limitCopy(void* dst, size_t dstCapacity, const void* src, size_t srcSize)
+{
+    size_t const length = MIN(dstCapacity, srcSize);
+    if (length > 0) {
+        ZSTD_memcpy(dst, src, length);
+    }
+    return length;
+}
+
+/* define "workspace is too large" as this number of times larger than needed */
+#define ZSTD_WORKSPACETOOLARGE_FACTOR 3
+
+/* when workspace is continuously too large
+ * during at least this number of times,
+ * context's memory usage is considered wasteful,
+ * because it's sized to handle a worst case scenario which rarely happens.
+ * In which case, resize it down to free some memory */
+#define ZSTD_WORKSPACETOOLARGE_MAXDURATION 128
+
+/* Controls whether the input/output buffer is buffered or stable. */
+typedef enum {
+    ZSTD_bm_buffered = 0,  /* Buffer the input/output */
+    ZSTD_bm_stable = 1     /* ZSTD_inBuffer/ZSTD_outBuffer is stable */
+} ZSTD_bufferMode_e;
+
+
+/*-*******************************************
+*  Private declarations
+*********************************************/
+typedef struct seqDef_s {
+    U32 offBase;   /* offBase == Offset + ZSTD_REP_NUM, or repcode 1,2,3 */
+    U16 litLength;
+    U16 mlBase;    /* mlBase == matchLength - MINMATCH */
+} seqDef;
+
+/* Controls whether seqStore has a single "long" litLength or matchLength. See seqStore_t. */
+typedef enum {
+    ZSTD_llt_none = 0,             /* no longLengthType */
+    ZSTD_llt_literalLength = 1,    /* represents a long literal */
+    ZSTD_llt_matchLength = 2       /* represents a long match */
+} ZSTD_longLengthType_e;
+
+typedef struct {
+    seqDef* sequencesStart;
+    seqDef* sequences;      /* ptr to end of sequences */
+    BYTE*  litStart;
+    BYTE*  lit;             /* ptr to end of literals */
+    BYTE*  llCode;
+    BYTE*  mlCode;
+    BYTE*  ofCode;
+    size_t maxNbSeq;
+    size_t maxNbLit;
+
+    /* longLengthPos and longLengthType to allow us to represent either a single litLength or matchLength
+     * in the seqStore that has a value larger than U16 (if it exists). To do so, we increment
+     * the existing value of the litLength or matchLength by 0x10000.
+     */
+    ZSTD_longLengthType_e longLengthType;
+    U32                   longLengthPos;  /* Index of the sequence to apply long length modification to */
+} seqStore_t;
+
+typedef struct {
+    U32 litLength;
+    U32 matchLength;
+} ZSTD_sequenceLength;
+
+/**
+ * Returns the ZSTD_sequenceLength for the given sequences. It handles the decoding of long sequences
+ * indicated by longLengthPos and longLengthType, and adds MINMATCH back to matchLength.
+ */
+MEM_STATIC ZSTD_sequenceLength ZSTD_getSequenceLength(seqStore_t const* seqStore, seqDef const* seq)
+{
+    ZSTD_sequenceLength seqLen;
+    seqLen.litLength = seq->litLength;
+    seqLen.matchLength = seq->mlBase + MINMATCH;
+    if (seqStore->longLengthPos == (U32)(seq - seqStore->sequencesStart)) {
+        if (seqStore->longLengthType == ZSTD_llt_literalLength) {
+            seqLen.litLength += 0x10000;
+        }
+        if (seqStore->longLengthType == ZSTD_llt_matchLength) {
+            seqLen.matchLength += 0x10000;
+        }
+    }
+    return seqLen;
+}
+
+/**
+ * Contains the compressed frame size and an upper-bound for the decompressed frame size.
+ * Note: before using `compressedSize`, check for errors using ZSTD_isError().
+ *       similarly, before using `decompressedBound`, check for errors using:
+ *          `decompressedBound != ZSTD_CONTENTSIZE_ERROR`
+ */
+typedef struct {
+    size_t nbBlocks;
+    size_t compressedSize;
+    unsigned long long decompressedBound;
+} ZSTD_frameSizeInfo;   /* decompress & legacy */
+
+const seqStore_t* ZSTD_getSeqStore(const ZSTD_CCtx* ctx);   /* compress & dictBuilder */
+void ZSTD_seqToCodes(const seqStore_t* seqStorePtr);   /* compress, dictBuilder, decodeCorpus (shouldn't get its definition from here) */
+
+/* custom memory allocation functions */
+void* ZSTD_customMalloc(size_t size, ZSTD_customMem customMem);
+void* ZSTD_customCalloc(size_t size, ZSTD_customMem customMem);
+void ZSTD_customFree(void* ptr, ZSTD_customMem customMem);
+
+
+/* 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);   /* zstdmt, adaptive_compression (shouldn't get this definition from here) */
+
+
+typedef struct {
+    blockType_e blockType;
+    U32 lastBlock;
+    U32 origSize;
+} blockProperties_t;   /* declared here for decompress and fullbench */
+
+/*! ZSTD_getcBlockSize() :
+ *  Provides the size of compressed block from block header `src` */
+/* Used by: decompress, fullbench (does not get its definition from here) */
+size_t ZSTD_getcBlockSize(const void* src, size_t srcSize,
+                          blockProperties_t* bpPtr);
+
+/*! ZSTD_decodeSeqHeaders() :
+ *  decode sequence header from src */
+/* Used by: decompress, fullbench (does not get its definition from here) */
+size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx* dctx, int* nbSeqPtr,
+                       const void* src, size_t srcSize);
+
+/**
+ * @returns true iff the CPU supports dynamic BMI2 dispatch.
+ */
+MEM_STATIC int ZSTD_cpuSupportsBmi2(void)
+{
+    ZSTD_cpuid_t cpuid = ZSTD_cpuid();
+    return ZSTD_cpuid_bmi1(cpuid) && ZSTD_cpuid_bmi2(cpuid);
+}
+
+#if defined (__cplusplus)
+}
+#endif
+
+#endif   /* ZSTD_CCOMMON_H_MODULE */

external/zstd/common/zstd_trace.h

@@ -0,0 +1,163 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_TRACE_H
+#define ZSTD_TRACE_H
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+#include <stddef.h>
+
+/* weak symbol support
+ * For now, enable conservatively:
+ * - Only GNUC
+ * - Only ELF
+ * - Only x86-64, i386 and aarch64
+ * Also, explicitly disable on platforms known not to work so they aren't
+ * forgotten in the future.
+ */
+#if !defined(ZSTD_HAVE_WEAK_SYMBOLS) && \
+    defined(__GNUC__) && defined(__ELF__) && \
+    (defined(__x86_64__) || defined(_M_X64) || defined(__i386__) || defined(_M_IX86) || defined(__aarch64__)) && \
+    !defined(__APPLE__) && !defined(_WIN32) && !defined(__MINGW32__) && \
+    !defined(__CYGWIN__) && !defined(_AIX)
+#  define ZSTD_HAVE_WEAK_SYMBOLS 1
+#else
+#  define ZSTD_HAVE_WEAK_SYMBOLS 0
+#endif
+#if ZSTD_HAVE_WEAK_SYMBOLS
+#  define ZSTD_WEAK_ATTR __attribute__((__weak__))
+#else
+#  define ZSTD_WEAK_ATTR
+#endif
+
+/* Only enable tracing when weak symbols are available. */
+#ifndef ZSTD_TRACE
+#  define ZSTD_TRACE ZSTD_HAVE_WEAK_SYMBOLS
+#endif
+
+#if ZSTD_TRACE
+
+struct ZSTD_CCtx_s;
+struct ZSTD_DCtx_s;
+struct ZSTD_CCtx_params_s;
+
+typedef struct {
+    /**
+     * ZSTD_VERSION_NUMBER
+     *
+     * This is guaranteed to be the first member of ZSTD_trace.
+     * Otherwise, this struct is not stable between versions. If
+     * the version number does not match your expectation, you
+     * should not interpret the rest of the struct.
+     */
+    unsigned version;
+    /**
+     * Non-zero if streaming (de)compression is used.
+     */
+    unsigned streaming;
+    /**
+     * The dictionary ID.
+     */
+    unsigned dictionaryID;
+    /**
+     * Is the dictionary cold?
+     * Only set on decompression.
+     */
+    unsigned dictionaryIsCold;
+    /**
+     * The dictionary size or zero if no dictionary.
+     */
+    size_t dictionarySize;
+    /**
+     * The uncompressed size of the data.
+     */
+    size_t uncompressedSize;
+    /**
+     * The compressed size of the data.
+     */
+    size_t compressedSize;
+    /**
+     * The fully resolved CCtx parameters (NULL on decompression).
+     */
+    struct ZSTD_CCtx_params_s const* params;
+    /**
+     * The ZSTD_CCtx pointer (NULL on decompression).
+     */
+    struct ZSTD_CCtx_s const* cctx;
+    /**
+     * The ZSTD_DCtx pointer (NULL on compression).
+     */
+    struct ZSTD_DCtx_s const* dctx;
+} ZSTD_Trace;
+
+/**
+ * A tracing context. It must be 0 when tracing is disabled.
+ * Otherwise, any non-zero value returned by a tracing begin()
+ * function is presented to any subsequent calls to end().
+ *
+ * Any non-zero value is treated as tracing is enabled and not
+ * interpreted by the library.
+ *
+ * Two possible uses are:
+ * * A timestamp for when the begin() function was called.
+ * * A unique key identifying the (de)compression, like the
+ *   address of the [dc]ctx pointer if you need to track
+ *   more information than just a timestamp.
+ */
+typedef unsigned long long ZSTD_TraceCtx;
+
+/**
+ * Trace the beginning of a compression call.
+ * @param cctx The dctx pointer for the compression.
+ *             It can be used as a key to map begin() to end().
+ * @returns Non-zero if tracing is enabled. The return value is
+ *          passed to ZSTD_trace_compress_end().
+ */
+ZSTD_WEAK_ATTR ZSTD_TraceCtx ZSTD_trace_compress_begin(
+    struct ZSTD_CCtx_s const* cctx);
+
+/**
+ * Trace the end of a compression call.
+ * @param ctx The return value of ZSTD_trace_compress_begin().
+ * @param trace The zstd tracing info.
+ */
+ZSTD_WEAK_ATTR void ZSTD_trace_compress_end(
+    ZSTD_TraceCtx ctx,
+    ZSTD_Trace const* trace);
+
+/**
+ * Trace the beginning of a decompression call.
+ * @param dctx The dctx pointer for the decompression.
+ *             It can be used as a key to map begin() to end().
+ * @returns Non-zero if tracing is enabled. The return value is
+ *          passed to ZSTD_trace_compress_end().
+ */
+ZSTD_WEAK_ATTR ZSTD_TraceCtx ZSTD_trace_decompress_begin(
+    struct ZSTD_DCtx_s const* dctx);
+
+/**
+ * Trace the end of a decompression call.
+ * @param ctx The return value of ZSTD_trace_decompress_begin().
+ * @param trace The zstd tracing info.
+ */
+ZSTD_WEAK_ATTR void ZSTD_trace_decompress_end(
+    ZSTD_TraceCtx ctx,
+    ZSTD_Trace const* trace);
+
+#endif /* ZSTD_TRACE */
+
+#if defined (__cplusplus)
+}
+#endif
+
+#endif /* ZSTD_TRACE_H */

external/zstd/decompress/huf_decompress_amd64.S

@@ -0,0 +1,574 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#include "../common/portability_macros.h"
+
+/* Stack marking
+ * ref: https://wiki.gentoo.org/wiki/Hardened/GNU_stack_quickstart
+ */
+#if defined(__ELF__) && defined(__GNUC__)
+.section .note.GNU-stack,"",%progbits
+#endif
+
+#if ZSTD_ENABLE_ASM_X86_64_BMI2
+
+/* Calling convention:
+ *
+ * %rdi contains the first argument: HUF_DecompressAsmArgs*.
+ * %rbp isn't maintained (no frame pointer).
+ * %rsp contains the stack pointer that grows down.
+ *      No red-zone is assumed, only addresses >= %rsp are used.
+ * All register contents are preserved.
+ *
+ * TODO: Support Windows calling convention.
+ */
+
+ZSTD_HIDE_ASM_FUNCTION(HUF_decompress4X1_usingDTable_internal_bmi2_asm_loop)
+ZSTD_HIDE_ASM_FUNCTION(HUF_decompress4X2_usingDTable_internal_bmi2_asm_loop)
+ZSTD_HIDE_ASM_FUNCTION(_HUF_decompress4X2_usingDTable_internal_bmi2_asm_loop)
+ZSTD_HIDE_ASM_FUNCTION(_HUF_decompress4X1_usingDTable_internal_bmi2_asm_loop)
+.global HUF_decompress4X1_usingDTable_internal_bmi2_asm_loop
+.global HUF_decompress4X2_usingDTable_internal_bmi2_asm_loop
+.global _HUF_decompress4X1_usingDTable_internal_bmi2_asm_loop
+.global _HUF_decompress4X2_usingDTable_internal_bmi2_asm_loop
+.text
+
+/* Sets up register mappings for clarity.
+ * op[], bits[], dtable & ip[0] each get their own register.
+ * ip[1,2,3] & olimit alias var[].
+ * %rax is a scratch register.
+ */
+
+#define op0    rsi
+#define op1    rbx
+#define op2    rcx
+#define op3    rdi
+
+#define ip0    r8
+#define ip1    r9
+#define ip2    r10
+#define ip3    r11
+
+#define bits0  rbp
+#define bits1  rdx
+#define bits2  r12
+#define bits3  r13
+#define dtable r14
+#define olimit r15
+
+/* var[] aliases ip[1,2,3] & olimit
+ * ip[1,2,3] are saved every iteration.
+ * olimit is only used in compute_olimit.
+ */
+#define var0   r15
+#define var1   r9
+#define var2   r10
+#define var3   r11
+
+/* 32-bit var registers */
+#define vard0  r15d
+#define vard1  r9d
+#define vard2  r10d
+#define vard3  r11d
+
+/* Calls X(N) for each stream 0, 1, 2, 3. */
+#define FOR_EACH_STREAM(X) \
+    X(0);                  \
+    X(1);                  \
+    X(2);                  \
+    X(3)
+
+/* Calls X(N, idx) for each stream 0, 1, 2, 3. */
+#define FOR_EACH_STREAM_WITH_INDEX(X, idx) \
+    X(0, idx);                             \
+    X(1, idx);                             \
+    X(2, idx);                             \
+    X(3, idx)
+
+/* Define both _HUF_* & HUF_* symbols because MacOS
+ * C symbols are prefixed with '_' & Linux symbols aren't.
+ */
+_HUF_decompress4X1_usingDTable_internal_bmi2_asm_loop:
+HUF_decompress4X1_usingDTable_internal_bmi2_asm_loop:
+    /* Save all registers - even if they are callee saved for simplicity. */
+    push %rax
+    push %rbx
+    push %rcx
+    push %rdx
+    push %rbp
+    push %rsi
+    push %rdi
+    push %r8
+    push %r9
+    push %r10
+    push %r11
+    push %r12
+    push %r13
+    push %r14
+    push %r15
+
+    /* Read HUF_DecompressAsmArgs* args from %rax */
+    movq %rdi, %rax
+    movq  0(%rax), %ip0
+    movq  8(%rax), %ip1
+    movq 16(%rax), %ip2
+    movq 24(%rax), %ip3
+    movq 32(%rax), %op0
+    movq 40(%rax), %op1
+    movq 48(%rax), %op2
+    movq 56(%rax), %op3
+    movq 64(%rax), %bits0
+    movq 72(%rax), %bits1
+    movq 80(%rax), %bits2
+    movq 88(%rax), %bits3
+    movq 96(%rax), %dtable
+    push %rax      /* argument */
+    push 104(%rax) /* ilimit */
+    push 112(%rax) /* oend */
+    push %olimit   /* olimit space */
+
+    subq $24, %rsp
+
+.L_4X1_compute_olimit:
+    /* Computes how many iterations we can do safely
+     * %r15, %rax may be clobbered
+     * rbx, rdx must be saved
+     * op3 & ip0 mustn't be clobbered
+     */
+    movq %rbx, 0(%rsp)
+    movq %rdx, 8(%rsp)
+
+    movq 32(%rsp), %rax /* rax = oend */
+    subq %op3,    %rax  /* rax = oend - op3 */
+
+    /* r15 = (oend - op3) / 5 */
+    movabsq $-3689348814741910323, %rdx
+    mulq %rdx
+    movq %rdx, %r15
+    shrq $2, %r15
+
+    movq %ip0,     %rax /* rax = ip0 */
+    movq 40(%rsp), %rdx /* rdx = ilimit */
+    subq %rdx,     %rax /* rax = ip0 - ilimit */
+    movq %rax,     %rbx /* rbx = ip0 - ilimit */
+
+    /* rdx = (ip0 - ilimit) / 7 */
+    movabsq $2635249153387078803, %rdx
+    mulq %rdx
+    subq %rdx, %rbx
+    shrq %rbx
+    addq %rbx, %rdx
+    shrq $2, %rdx
+
+    /* r15 = min(%rdx, %r15) */
+    cmpq %rdx, %r15
+    cmova %rdx, %r15
+
+    /* r15 = r15 * 5 */
+    leaq (%r15, %r15, 4), %r15
+
+    /* olimit = op3 + r15 */
+    addq %op3, %olimit
+
+    movq 8(%rsp), %rdx
+    movq 0(%rsp), %rbx
+
+    /* If (op3 + 20 > olimit) */
+    movq %op3, %rax    /* rax = op3 */
+    addq $20,  %rax    /* rax = op3 + 20 */
+    cmpq %rax, %olimit /* op3 + 20 > olimit */
+    jb .L_4X1_exit
+
+    /* If (ip1 < ip0) go to exit */
+    cmpq %ip0, %ip1
+    jb .L_4X1_exit
+
+    /* If (ip2 < ip1) go to exit */
+    cmpq %ip1, %ip2
+    jb .L_4X1_exit
+
+    /* If (ip3 < ip2) go to exit */
+    cmpq %ip2, %ip3
+    jb .L_4X1_exit
+
+/* Reads top 11 bits from bits[n]
+ * Loads dt[bits[n]] into var[n]
+ */
+#define GET_NEXT_DELT(n)                \
+    movq $53, %var##n;                  \
+    shrxq %var##n, %bits##n, %var##n;   \
+    movzwl (%dtable,%var##n,2),%vard##n
+
+/* var[n] must contain the DTable entry computed with GET_NEXT_DELT
+ * Moves var[n] to %rax
+ * bits[n] <<= var[n] & 63
+ * op[n][idx] = %rax >> 8
+ * %ah is a way to access bits [8, 16) of %rax
+ */
+#define DECODE_FROM_DELT(n, idx)       \
+    movq %var##n, %rax;                \
+    shlxq %var##n, %bits##n, %bits##n; \
+    movb %ah, idx(%op##n)
+
+/* Assumes GET_NEXT_DELT has been called.
+ * Calls DECODE_FROM_DELT then GET_NEXT_DELT
+ */
+#define DECODE_AND_GET_NEXT(n, idx) \
+    DECODE_FROM_DELT(n, idx);       \
+    GET_NEXT_DELT(n)                \
+
+/* // ctz & nbBytes is stored in bits[n]
+ * // nbBits is stored in %rax
+ * ctz  = CTZ[bits[n]]
+ * nbBits  = ctz & 7
+ * nbBytes = ctz >> 3
+ * op[n]  += 5
+ * ip[n]  -= nbBytes
+ * // Note: x86-64 is little-endian ==> no bswap
+ * bits[n] = MEM_readST(ip[n]) | 1
+ * bits[n] <<= nbBits
+ */
+#define RELOAD_BITS(n)             \
+    bsfq %bits##n, %bits##n;       \
+    movq %bits##n, %rax;           \
+    andq $7, %rax;                 \
+    shrq $3, %bits##n;             \
+    leaq 5(%op##n), %op##n;        \
+    subq %bits##n, %ip##n;         \
+    movq (%ip##n), %bits##n;       \
+    orq $1, %bits##n;              \
+    shlx %rax, %bits##n, %bits##n
+
+    /* Store clobbered variables on the stack */
+    movq %olimit, 24(%rsp)
+    movq %ip1, 0(%rsp)
+    movq %ip2, 8(%rsp)
+    movq %ip3, 16(%rsp)
+
+    /* Call GET_NEXT_DELT for each stream */
+    FOR_EACH_STREAM(GET_NEXT_DELT)
+
+    .p2align 6
+
+.L_4X1_loop_body:
+    /* Decode 5 symbols in each of the 4 streams (20 total)
+     * Must have called GET_NEXT_DELT for each stream
+     */
+    FOR_EACH_STREAM_WITH_INDEX(DECODE_AND_GET_NEXT, 0)
+    FOR_EACH_STREAM_WITH_INDEX(DECODE_AND_GET_NEXT, 1)
+    FOR_EACH_STREAM_WITH_INDEX(DECODE_AND_GET_NEXT, 2)
+    FOR_EACH_STREAM_WITH_INDEX(DECODE_AND_GET_NEXT, 3)
+    FOR_EACH_STREAM_WITH_INDEX(DECODE_FROM_DELT, 4)
+
+    /* Load ip[1,2,3] from stack (var[] aliases them)
+     * ip[] is needed for RELOAD_BITS
+     * Each will be stored back to the stack after RELOAD
+     */
+    movq 0(%rsp), %ip1
+    movq 8(%rsp), %ip2
+    movq 16(%rsp), %ip3
+
+    /* Reload each stream & fetch the next table entry
+     * to prepare for the next iteration
+     */
+    RELOAD_BITS(0)
+    GET_NEXT_DELT(0)
+
+    RELOAD_BITS(1)
+    movq %ip1, 0(%rsp)
+    GET_NEXT_DELT(1)
+
+    RELOAD_BITS(2)
+    movq %ip2, 8(%rsp)
+    GET_NEXT_DELT(2)
+
+    RELOAD_BITS(3)
+    movq %ip3, 16(%rsp)
+    GET_NEXT_DELT(3)
+
+    /* If op3 < olimit: continue the loop */
+    cmp %op3, 24(%rsp)
+    ja .L_4X1_loop_body
+
+    /* Reload ip[1,2,3] from stack */
+    movq 0(%rsp), %ip1
+    movq 8(%rsp), %ip2
+    movq 16(%rsp), %ip3
+
+    /* Re-compute olimit */
+    jmp .L_4X1_compute_olimit
+
+#undef GET_NEXT_DELT
+#undef DECODE_FROM_DELT
+#undef DECODE
+#undef RELOAD_BITS
+.L_4X1_exit:
+    addq $24, %rsp
+
+    /* Restore stack (oend & olimit) */
+    pop %rax /* olimit */
+    pop %rax /* oend */
+    pop %rax /* ilimit */
+    pop %rax /* arg */
+
+    /* Save ip / op / bits */
+    movq %ip0,  0(%rax)
+    movq %ip1,  8(%rax)
+    movq %ip2, 16(%rax)
+    movq %ip3, 24(%rax)
+    movq %op0, 32(%rax)
+    movq %op1, 40(%rax)
+    movq %op2, 48(%rax)
+    movq %op3, 56(%rax)
+    movq %bits0, 64(%rax)
+    movq %bits1, 72(%rax)
+    movq %bits2, 80(%rax)
+    movq %bits3, 88(%rax)
+
+    /* Restore registers */
+    pop %r15
+    pop %r14
+    pop %r13
+    pop %r12
+    pop %r11
+    pop %r10
+    pop %r9
+    pop %r8
+    pop %rdi
+    pop %rsi
+    pop %rbp
+    pop %rdx
+    pop %rcx
+    pop %rbx
+    pop %rax
+    ret
+
+_HUF_decompress4X2_usingDTable_internal_bmi2_asm_loop:
+HUF_decompress4X2_usingDTable_internal_bmi2_asm_loop:
+    /* Save all registers - even if they are callee saved for simplicity. */
+    push %rax
+    push %rbx
+    push %rcx
+    push %rdx
+    push %rbp
+    push %rsi
+    push %rdi
+    push %r8
+    push %r9
+    push %r10
+    push %r11
+    push %r12
+    push %r13
+    push %r14
+    push %r15
+
+    movq %rdi, %rax
+    movq  0(%rax), %ip0
+    movq  8(%rax), %ip1
+    movq 16(%rax), %ip2
+    movq 24(%rax), %ip3
+    movq 32(%rax), %op0
+    movq 40(%rax), %op1
+    movq 48(%rax), %op2
+    movq 56(%rax), %op3
+    movq 64(%rax), %bits0
+    movq 72(%rax), %bits1
+    movq 80(%rax), %bits2
+    movq 88(%rax), %bits3
+    movq 96(%rax), %dtable
+    push %rax      /* argument */
+    push %rax      /* olimit */
+    push 104(%rax) /* ilimit */
+
+    movq 112(%rax), %rax
+    push %rax /* oend3 */
+
+    movq %op3, %rax
+    push %rax /* oend2 */
+
+    movq %op2, %rax
+    push %rax /* oend1 */
+
+    movq %op1, %rax
+    push %rax /* oend0 */
+
+    /* Scratch space */
+    subq $8, %rsp
+
+.L_4X2_compute_olimit:
+    /* Computes how many iterations we can do safely
+     * %r15, %rax may be clobbered
+     * rdx must be saved
+     * op[1,2,3,4] & ip0 mustn't be clobbered
+     */
+    movq %rdx, 0(%rsp)
+
+    /* We can consume up to 7 input bytes each iteration. */
+    movq %ip0,     %rax  /* rax = ip0 */
+    movq 40(%rsp), %rdx  /* rdx = ilimit */
+    subq %rdx,     %rax  /* rax = ip0 - ilimit */
+    movq %rax,    %r15   /* r15 = ip0 - ilimit */
+
+    /* rdx = rax / 7 */
+    movabsq $2635249153387078803, %rdx
+    mulq %rdx
+    subq %rdx, %r15
+    shrq %r15
+    addq %r15, %rdx
+    shrq $2, %rdx
+
+    /* r15 = (ip0 - ilimit) / 7 */
+    movq %rdx, %r15
+
+    /* r15 = min(r15, min(oend0 - op0, oend1 - op1, oend2 - op2, oend3 - op3) / 10) */
+    movq 8(%rsp),  %rax /* rax = oend0 */
+    subq %op0,     %rax /* rax = oend0 - op0 */
+    movq 16(%rsp), %rdx /* rdx = oend1 */
+    subq %op1,     %rdx /* rdx = oend1 - op1 */
+
+    cmpq  %rax,    %rdx
+    cmova %rax,    %rdx /* rdx = min(%rdx, %rax) */
+
+    movq 24(%rsp), %rax /* rax = oend2 */
+    subq %op2,     %rax /* rax = oend2 - op2 */
+
+    cmpq  %rax,    %rdx
+    cmova %rax,    %rdx /* rdx = min(%rdx, %rax) */
+
+    movq 32(%rsp), %rax /* rax = oend3 */
+    subq %op3,     %rax /* rax = oend3 - op3 */
+
+    cmpq  %rax,    %rdx
+    cmova %rax,    %rdx /* rdx = min(%rdx, %rax) */
+
+    movabsq $-3689348814741910323, %rax
+    mulq %rdx
+    shrq $3,       %rdx /* rdx = rdx / 10 */
+
+    /* r15 = min(%rdx, %r15) */
+    cmpq  %rdx, %r15
+    cmova %rdx, %r15
+
+    /* olimit = op3 + 5 * r15 */
+    movq %r15, %rax
+    leaq (%op3, %rax, 4), %olimit
+    addq %rax, %olimit
+
+    movq 0(%rsp), %rdx
+
+    /* If (op3 + 10 > olimit) */
+    movq %op3, %rax    /* rax = op3 */
+    addq $10,  %rax    /* rax = op3 + 10 */
+    cmpq %rax, %olimit /* op3 + 10 > olimit */
+    jb .L_4X2_exit
+
+    /* If (ip1 < ip0) go to exit */
+    cmpq %ip0, %ip1
+    jb .L_4X2_exit
+
+    /* If (ip2 < ip1) go to exit */
+    cmpq %ip1, %ip2
+    jb .L_4X2_exit
+
+    /* If (ip3 < ip2) go to exit */
+    cmpq %ip2, %ip3
+    jb .L_4X2_exit
+
+#define DECODE(n, idx)              \
+    movq %bits##n, %rax;            \
+    shrq $53, %rax;                 \
+    movzwl 0(%dtable,%rax,4),%r8d;  \
+    movzbl 2(%dtable,%rax,4),%r15d; \
+    movzbl 3(%dtable,%rax,4),%eax;  \
+    movw %r8w, (%op##n);            \
+    shlxq %r15, %bits##n, %bits##n; \
+    addq %rax, %op##n
+
+#define RELOAD_BITS(n)              \
+    bsfq %bits##n, %bits##n;        \
+    movq %bits##n, %rax;            \
+    shrq $3, %bits##n;              \
+    andq $7, %rax;                  \
+    subq %bits##n, %ip##n;          \
+    movq (%ip##n), %bits##n;        \
+    orq $1, %bits##n;               \
+    shlxq %rax, %bits##n, %bits##n
+
+
+    movq %olimit, 48(%rsp)
+
+    .p2align 6
+
+.L_4X2_loop_body:
+    /* We clobber r8, so store it on the stack */
+    movq %r8, 0(%rsp)
+
+    /* Decode 5 symbols from each of the 4 streams (20 symbols total). */
+    FOR_EACH_STREAM_WITH_INDEX(DECODE, 0)
+    FOR_EACH_STREAM_WITH_INDEX(DECODE, 1)
+    FOR_EACH_STREAM_WITH_INDEX(DECODE, 2)
+    FOR_EACH_STREAM_WITH_INDEX(DECODE, 3)
+    FOR_EACH_STREAM_WITH_INDEX(DECODE, 4)
+
+    /* Reload r8 */
+    movq 0(%rsp), %r8
+
+    FOR_EACH_STREAM(RELOAD_BITS)
+
+    cmp %op3, 48(%rsp)
+    ja .L_4X2_loop_body
+    jmp .L_4X2_compute_olimit
+
+#undef DECODE
+#undef RELOAD_BITS
+.L_4X2_exit:
+    addq $8, %rsp
+    /* Restore stack (oend & olimit) */
+    pop %rax /* oend0 */
+    pop %rax /* oend1 */
+    pop %rax /* oend2 */
+    pop %rax /* oend3 */
+    pop %rax /* ilimit */
+    pop %rax /* olimit */
+    pop %rax /* arg */
+
+    /* Save ip / op / bits */
+    movq %ip0,  0(%rax)
+    movq %ip1,  8(%rax)
+    movq %ip2, 16(%rax)
+    movq %ip3, 24(%rax)
+    movq %op0, 32(%rax)
+    movq %op1, 40(%rax)
+    movq %op2, 48(%rax)
+    movq %op3, 56(%rax)
+    movq %bits0, 64(%rax)
+    movq %bits1, 72(%rax)
+    movq %bits2, 80(%rax)
+    movq %bits3, 88(%rax)
+
+    /* Restore registers */
+    pop %r15
+    pop %r14
+    pop %r13
+    pop %r12
+    pop %r11
+    pop %r10
+    pop %r9
+    pop %r8
+    pop %rdi
+    pop %rsi
+    pop %rbp
+    pop %rdx
+    pop %rcx
+    pop %rbx
+    pop %rax
+    ret
+
+#endif

external/zstd/decompress/zstd_ddict.c

@@ -0,0 +1,244 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+/* zstd_ddict.c :
+ * concentrates all logic that needs to know the internals of ZSTD_DDict object */
+
+/*-*******************************************************
+*  Dependencies
+*********************************************************/
+#include "../common/zstd_deps.h"   /* ZSTD_memcpy, ZSTD_memmove, ZSTD_memset */
+#include "../common/cpu.h"         /* bmi2 */
+#include "../common/mem.h"         /* low level memory routines */
+#define FSE_STATIC_LINKING_ONLY
+#include "../common/fse.h"
+#define HUF_STATIC_LINKING_ONLY
+#include "../common/huf.h"
+#include "zstd_decompress_internal.h"
+#include "zstd_ddict.h"
+
+#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>=1)
+#  include "../legacy/zstd_legacy.h"
+#endif
+
+
+
+/*-*******************************************************
+*  Types
+*********************************************************/
+struct ZSTD_DDict_s {
+    void* dictBuffer;
+    const void* dictContent;
+    size_t dictSize;
+    ZSTD_entropyDTables_t entropy;
+    U32 dictID;
+    U32 entropyPresent;
+    ZSTD_customMem cMem;
+};  /* typedef'd to ZSTD_DDict within "zstd.h" */
+
+const void* ZSTD_DDict_dictContent(const ZSTD_DDict* ddict)
+{
+    assert(ddict != NULL);
+    return ddict->dictContent;
+}
+
+size_t ZSTD_DDict_dictSize(const ZSTD_DDict* ddict)
+{
+    assert(ddict != NULL);
+    return ddict->dictSize;
+}
+
+void ZSTD_copyDDictParameters(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict)
+{
+    DEBUGLOG(4, "ZSTD_copyDDictParameters");
+    assert(dctx != NULL);
+    assert(ddict != NULL);
+    dctx->dictID = ddict->dictID;
+    dctx->prefixStart = ddict->dictContent;
+    dctx->virtualStart = ddict->dictContent;
+    dctx->dictEnd = (const BYTE*)ddict->dictContent + ddict->dictSize;
+    dctx->previousDstEnd = dctx->dictEnd;
+#ifdef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION
+    dctx->dictContentBeginForFuzzing = dctx->prefixStart;
+    dctx->dictContentEndForFuzzing = dctx->previousDstEnd;
+#endif
+    if (ddict->entropyPresent) {
+        dctx->litEntropy = 1;
+        dctx->fseEntropy = 1;
+        dctx->LLTptr = ddict->entropy.LLTable;
+        dctx->MLTptr = ddict->entropy.MLTable;
+        dctx->OFTptr = ddict->entropy.OFTable;
+        dctx->HUFptr = ddict->entropy.hufTable;
+        dctx->entropy.rep[0] = ddict->entropy.rep[0];
+        dctx->entropy.rep[1] = ddict->entropy.rep[1];
+        dctx->entropy.rep[2] = ddict->entropy.rep[2];
+    } else {
+        dctx->litEntropy = 0;
+        dctx->fseEntropy = 0;
+    }
+}
+
+
+static size_t
+ZSTD_loadEntropy_intoDDict(ZSTD_DDict* ddict,
+                           ZSTD_dictContentType_e dictContentType)
+{
+    ddict->dictID = 0;
+    ddict->entropyPresent = 0;
+    if (dictContentType == ZSTD_dct_rawContent) return 0;
+
+    if (ddict->dictSize < 8) {
+        if (dictContentType == ZSTD_dct_fullDict)
+            return ERROR(dictionary_corrupted);   /* only accept specified dictionaries */
+        return 0;   /* pure content mode */
+    }
+    {   U32 const magic = MEM_readLE32(ddict->dictContent);
+        if (magic != ZSTD_MAGIC_DICTIONARY) {
+            if (dictContentType == ZSTD_dct_fullDict)
+                return ERROR(dictionary_corrupted);   /* only accept specified dictionaries */
+            return 0;   /* pure content mode */
+        }
+    }
+    ddict->dictID = MEM_readLE32((const char*)ddict->dictContent + ZSTD_FRAMEIDSIZE);
+
+    /* load entropy tables */
+    RETURN_ERROR_IF(ZSTD_isError(ZSTD_loadDEntropy(
+            &ddict->entropy, ddict->dictContent, ddict->dictSize)),
+        dictionary_corrupted, "");
+    ddict->entropyPresent = 1;
+    return 0;
+}
+
+
+static size_t ZSTD_initDDict_internal(ZSTD_DDict* ddict,
+                                      const void* dict, size_t dictSize,
+                                      ZSTD_dictLoadMethod_e dictLoadMethod,
+                                      ZSTD_dictContentType_e dictContentType)
+{
+    if ((dictLoadMethod == ZSTD_dlm_byRef) || (!dict) || (!dictSize)) {
+        ddict->dictBuffer = NULL;
+        ddict->dictContent = dict;
+        if (!dict) dictSize = 0;
+    } else {
+        void* const internalBuffer = ZSTD_customMalloc(dictSize, ddict->cMem);
+        ddict->dictBuffer = internalBuffer;
+        ddict->dictContent = internalBuffer;
+        if (!internalBuffer) return ERROR(memory_allocation);
+        ZSTD_memcpy(internalBuffer, dict, dictSize);
+    }
+    ddict->dictSize = dictSize;
+    ddict->entropy.hufTable[0] = (HUF_DTable)((ZSTD_HUFFDTABLE_CAPACITY_LOG)*0x1000001);  /* cover both little and big endian */
+
+    /* parse dictionary content */
+    FORWARD_IF_ERROR( ZSTD_loadEntropy_intoDDict(ddict, dictContentType) , "");
+
+    return 0;
+}
+
+ZSTD_DDict* ZSTD_createDDict_advanced(const void* dict, size_t dictSize,
+                                      ZSTD_dictLoadMethod_e dictLoadMethod,
+                                      ZSTD_dictContentType_e dictContentType,
+                                      ZSTD_customMem customMem)
+{
+    if ((!customMem.customAlloc) ^ (!customMem.customFree)) return NULL;
+
+    {   ZSTD_DDict* const ddict = (ZSTD_DDict*) ZSTD_customMalloc(sizeof(ZSTD_DDict), customMem);
+        if (ddict == NULL) return NULL;
+        ddict->cMem = customMem;
+        {   size_t const initResult = ZSTD_initDDict_internal(ddict,
+                                            dict, dictSize,
+                                            dictLoadMethod, dictContentType);
+            if (ZSTD_isError(initResult)) {
+                ZSTD_freeDDict(ddict);
+                return NULL;
+        }   }
+        return ddict;
+    }
+}
+
+/*! ZSTD_createDDict() :
+*   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_createDDict(const void* dict, size_t dictSize)
+{
+    ZSTD_customMem const allocator = { NULL, NULL, NULL };
+    return ZSTD_createDDict_advanced(dict, dictSize, ZSTD_dlm_byCopy, ZSTD_dct_auto, allocator);
+}
+
+/*! ZSTD_createDDict_byReference() :
+ *  Create a digested dictionary, to start decompression without startup delay.
+ *  Dictionary content is simply referenced, it will be accessed during decompression.
+ *  Warning : dictBuffer must outlive DDict (DDict must be freed before dictBuffer) */
+ZSTD_DDict* ZSTD_createDDict_byReference(const void* dictBuffer, size_t dictSize)
+{
+    ZSTD_customMem const allocator = { NULL, NULL, NULL };
+    return ZSTD_createDDict_advanced(dictBuffer, dictSize, ZSTD_dlm_byRef, ZSTD_dct_auto, allocator);
+}
+
+
+const ZSTD_DDict* ZSTD_initStaticDDict(
+                                void* sBuffer, size_t sBufferSize,
+                                const void* dict, size_t dictSize,
+                                ZSTD_dictLoadMethod_e dictLoadMethod,
+                                ZSTD_dictContentType_e dictContentType)
+{
+    size_t const neededSpace = sizeof(ZSTD_DDict)
+                             + (dictLoadMethod == ZSTD_dlm_byRef ? 0 : dictSize);
+    ZSTD_DDict* const ddict = (ZSTD_DDict*)sBuffer;
+    assert(sBuffer != NULL);
+    assert(dict != NULL);
+    if ((size_t)sBuffer & 7) return NULL;   /* 8-aligned */
+    if (sBufferSize < neededSpace) return NULL;
+    if (dictLoadMethod == ZSTD_dlm_byCopy) {
+        ZSTD_memcpy(ddict+1, dict, dictSize);  /* local copy */
+        dict = ddict+1;
+    }
+    if (ZSTD_isError( ZSTD_initDDict_internal(ddict,
+                                              dict, dictSize,
+                                              ZSTD_dlm_byRef, dictContentType) ))
+        return NULL;
+    return ddict;
+}
+
+
+size_t ZSTD_freeDDict(ZSTD_DDict* ddict)
+{
+    if (ddict==NULL) return 0;   /* support free on NULL */
+    {   ZSTD_customMem const cMem = ddict->cMem;
+        ZSTD_customFree(ddict->dictBuffer, cMem);
+        ZSTD_customFree(ddict, cMem);
+        return 0;
+    }
+}
+
+/*! ZSTD_estimateDDictSize() :
+ *  Estimate amount of memory that will be needed to create a dictionary for decompression.
+ *  Note : dictionary created by reference using ZSTD_dlm_byRef are smaller */
+size_t ZSTD_estimateDDictSize(size_t dictSize, ZSTD_dictLoadMethod_e dictLoadMethod)
+{
+    return sizeof(ZSTD_DDict) + (dictLoadMethod == ZSTD_dlm_byRef ? 0 : dictSize);
+}
+
+size_t ZSTD_sizeof_DDict(const ZSTD_DDict* ddict)
+{
+    if (ddict==NULL) return 0;   /* support sizeof on NULL */
+    return sizeof(*ddict) + (ddict->dictBuffer ? ddict->dictSize : 0) ;
+}
+
+/*! 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 ddict->dictID;
+}

external/zstd/decompress/zstd_ddict.h

@@ -0,0 +1,44 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+
+#ifndef ZSTD_DDICT_H
+#define ZSTD_DDICT_H
+
+/*-*******************************************************
+ *  Dependencies
+ *********************************************************/
+#include "../common/zstd_deps.h"   /* size_t */
+#include "../zstd.h"     /* ZSTD_DDict, and several public functions */
+
+
+/*-*******************************************************
+ *  Interface
+ *********************************************************/
+
+/* note: several prototypes are already published in `zstd.h` :
+ * ZSTD_createDDict()
+ * ZSTD_createDDict_byReference()
+ * ZSTD_createDDict_advanced()
+ * ZSTD_freeDDict()
+ * ZSTD_initStaticDDict()
+ * ZSTD_sizeof_DDict()
+ * ZSTD_estimateDDictSize()
+ * ZSTD_getDictID_fromDict()
+ */
+
+const void* ZSTD_DDict_dictContent(const ZSTD_DDict* ddict);
+size_t ZSTD_DDict_dictSize(const ZSTD_DDict* ddict);
+
+void ZSTD_copyDDictParameters(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict);
+
+
+
+#endif /* ZSTD_DDICT_H */

external/zstd/decompress/zstd_decompress_block.h

@@ -0,0 +1,68 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+
+#ifndef ZSTD_DEC_BLOCK_H
+#define ZSTD_DEC_BLOCK_H
+
+/*-*******************************************************
+ *  Dependencies
+ *********************************************************/
+#include "../common/zstd_deps.h"   /* size_t */
+#include "../zstd.h"    /* DCtx, and some public functions */
+#include "../common/zstd_internal.h"  /* blockProperties_t, and some public functions */
+#include "zstd_decompress_internal.h"  /* ZSTD_seqSymbol */
+
+
+/* ===   Prototypes   === */
+
+/* note: prototypes already published within `zstd.h` :
+ * ZSTD_decompressBlock()
+ */
+
+/* note: prototypes already published within `zstd_internal.h` :
+ * ZSTD_getcBlockSize()
+ * ZSTD_decodeSeqHeaders()
+ */
+
+
+ /* Streaming state is used to inform allocation of the literal buffer */
+typedef enum {
+    not_streaming = 0,
+    is_streaming = 1
+} streaming_operation;
+
+/* ZSTD_decompressBlock_internal() :
+ * decompress block, starting at `src`,
+ * into destination buffer `dst`.
+ * @return : decompressed block size,
+ *           or an error code (which can be tested using ZSTD_isError())
+ */
+size_t ZSTD_decompressBlock_internal(ZSTD_DCtx* dctx,
+                               void* dst, size_t dstCapacity,
+                         const void* src, size_t srcSize, const int frame, const streaming_operation streaming);
+
+/* ZSTD_buildFSETable() :
+ * generate FSE decoding table for one symbol (ll, ml or off)
+ * this function must be called with valid parameters only
+ * (dt is large enough, normalizedCounter distribution total is a power of 2, max is within range, etc.)
+ * in which case it cannot fail.
+ * The workspace must be 4-byte aligned and at least ZSTD_BUILD_FSE_TABLE_WKSP_SIZE bytes, which is
+ * defined in zstd_decompress_internal.h.
+ * Internal use only.
+ */
+void ZSTD_buildFSETable(ZSTD_seqSymbol* dt,
+             const short* normalizedCounter, unsigned maxSymbolValue,
+             const U32* baseValue, const U8* nbAdditionalBits,
+                   unsigned tableLog, void* wksp, size_t wkspSize,
+                   int bmi2);
+
+
+#endif /* ZSTD_DEC_BLOCK_H */

external/zstd/decompress/zstd_decompress_internal.h

@@ -0,0 +1,237 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+
+/* zstd_decompress_internal:
+ * objects and definitions shared within lib/decompress modules */
+
+ #ifndef ZSTD_DECOMPRESS_INTERNAL_H
+ #define ZSTD_DECOMPRESS_INTERNAL_H
+
+
+/*-*******************************************************
+ *  Dependencies
+ *********************************************************/
+#include "../common/mem.h"             /* BYTE, U16, U32 */
+#include "../common/zstd_internal.h"   /* constants : MaxLL, MaxML, MaxOff, LLFSELog, etc. */
+
+
+
+/*-*******************************************************
+ *  Constants
+ *********************************************************/
+static UNUSED_ATTR 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 UNUSED_ATTR 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, 0x1FFFFFFD, 0x3FFFFFFD, 0x7FFFFFFD };
+
+static UNUSED_ATTR const U8 OF_bits[MaxOff+1] = {
+                     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 };
+
+static UNUSED_ATTR 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 };
+
+
+/*-*******************************************************
+ *  Decompression types
+ *********************************************************/
+ typedef struct {
+     U32 fastMode;
+     U32 tableLog;
+ } ZSTD_seqSymbol_header;
+
+ typedef struct {
+     U16  nextState;
+     BYTE nbAdditionalBits;
+     BYTE nbBits;
+     U32  baseValue;
+ } ZSTD_seqSymbol;
+
+ #define SEQSYMBOL_TABLE_SIZE(log)   (1 + (1 << (log)))
+
+#define ZSTD_BUILD_FSE_TABLE_WKSP_SIZE (sizeof(S16) * (MaxSeq + 1) + (1u << MaxFSELog) + sizeof(U64))
+#define ZSTD_BUILD_FSE_TABLE_WKSP_SIZE_U32 ((ZSTD_BUILD_FSE_TABLE_WKSP_SIZE + sizeof(U32) - 1) / sizeof(U32))
+#define ZSTD_HUFFDTABLE_CAPACITY_LOG 12
+
+typedef struct {
+    ZSTD_seqSymbol LLTable[SEQSYMBOL_TABLE_SIZE(LLFSELog)];    /* Note : Space reserved for FSE Tables */
+    ZSTD_seqSymbol OFTable[SEQSYMBOL_TABLE_SIZE(OffFSELog)];   /* is also used as temporary workspace while building hufTable during DDict creation */
+    ZSTD_seqSymbol MLTable[SEQSYMBOL_TABLE_SIZE(MLFSELog)];    /* and therefore must be at least HUF_DECOMPRESS_WORKSPACE_SIZE large */
+    HUF_DTable hufTable[HUF_DTABLE_SIZE(ZSTD_HUFFDTABLE_CAPACITY_LOG)];  /* can accommodate HUF_decompress4X */
+    U32 rep[ZSTD_REP_NUM];
+    U32 workspace[ZSTD_BUILD_FSE_TABLE_WKSP_SIZE_U32];
+} ZSTD_entropyDTables_t;
+
+typedef enum { ZSTDds_getFrameHeaderSize, ZSTDds_decodeFrameHeader,
+               ZSTDds_decodeBlockHeader, ZSTDds_decompressBlock,
+               ZSTDds_decompressLastBlock, ZSTDds_checkChecksum,
+               ZSTDds_decodeSkippableHeader, ZSTDds_skipFrame } ZSTD_dStage;
+
+typedef enum { zdss_init=0, zdss_loadHeader,
+               zdss_read, zdss_load, zdss_flush } ZSTD_dStreamStage;
+
+typedef enum {
+    ZSTD_use_indefinitely = -1,  /* Use the dictionary indefinitely */
+    ZSTD_dont_use = 0,           /* Do not use the dictionary (if one exists free it) */
+    ZSTD_use_once = 1            /* Use the dictionary once and set to ZSTD_dont_use */
+} ZSTD_dictUses_e;
+
+/* Hashset for storing references to multiple ZSTD_DDict within ZSTD_DCtx */
+typedef struct {
+    const ZSTD_DDict** ddictPtrTable;
+    size_t ddictPtrTableSize;
+    size_t ddictPtrCount;
+} ZSTD_DDictHashSet;
+
+#ifndef ZSTD_DECODER_INTERNAL_BUFFER
+#  define ZSTD_DECODER_INTERNAL_BUFFER  (1 << 16)
+#endif
+
+#define ZSTD_LBMIN 64
+#define ZSTD_LBMAX (128 << 10)
+
+/* extra buffer, compensates when dst is not large enough to store litBuffer */
+#define ZSTD_LITBUFFEREXTRASIZE  BOUNDED(ZSTD_LBMIN, ZSTD_DECODER_INTERNAL_BUFFER, ZSTD_LBMAX)
+
+typedef enum {
+    ZSTD_not_in_dst = 0,  /* Stored entirely within litExtraBuffer */
+    ZSTD_in_dst = 1,           /* Stored entirely within dst (in memory after current output write) */
+    ZSTD_split = 2            /* Split between litExtraBuffer and dst */
+} ZSTD_litLocation_e;
+
+struct ZSTD_DCtx_s
+{
+    const ZSTD_seqSymbol* LLTptr;
+    const ZSTD_seqSymbol* MLTptr;
+    const ZSTD_seqSymbol* OFTptr;
+    const HUF_DTable* HUFptr;
+    ZSTD_entropyDTables_t entropy;
+    U32 workspace[HUF_DECOMPRESS_WORKSPACE_SIZE_U32];   /* space needed when building huffman tables */
+    const void* previousDstEnd;   /* detect continuity */
+    const void* prefixStart;      /* start of current segment */
+    const void* virtualStart;     /* virtual start of previous segment if it was just before current one */
+    const void* dictEnd;          /* end of previous segment */
+    size_t expected;
+    ZSTD_frameHeader fParams;
+    U64 processedCSize;
+    U64 decodedSize;
+    blockType_e bType;            /* used in ZSTD_decompressContinue(), store blockType between block header decoding and block decompression stages */
+    ZSTD_dStage stage;
+    U32 litEntropy;
+    U32 fseEntropy;
+    XXH64_state_t xxhState;
+    size_t headerSize;
+    ZSTD_format_e format;
+    ZSTD_forceIgnoreChecksum_e forceIgnoreChecksum;   /* User specified: if == 1, will ignore checksums in compressed frame. Default == 0 */
+    U32 validateChecksum;         /* if == 1, will validate checksum. Is == 1 if (fParams.checksumFlag == 1) and (forceIgnoreChecksum == 0). */
+    const BYTE* litPtr;
+    ZSTD_customMem customMem;
+    size_t litSize;
+    size_t rleSize;
+    size_t staticSize;
+#if DYNAMIC_BMI2 != 0
+    int bmi2;                     /* == 1 if the CPU supports BMI2 and 0 otherwise. CPU support is determined dynamically once per context lifetime. */
+#endif
+
+    /* dictionary */
+    ZSTD_DDict* ddictLocal;
+    const ZSTD_DDict* ddict;     /* set by ZSTD_initDStream_usingDDict(), or ZSTD_DCtx_refDDict() */
+    U32 dictID;
+    int ddictIsCold;             /* if == 1 : dictionary is "new" for working context, and presumed "cold" (not in cpu cache) */
+    ZSTD_dictUses_e dictUses;
+    ZSTD_DDictHashSet* ddictSet;                    /* Hash set for multiple ddicts */
+    ZSTD_refMultipleDDicts_e refMultipleDDicts;     /* User specified: if == 1, will allow references to multiple DDicts. Default == 0 (disabled) */
+
+    /* streaming */
+    ZSTD_dStreamStage streamStage;
+    char*  inBuff;
+    size_t inBuffSize;
+    size_t inPos;
+    size_t maxWindowSize;
+    char*  outBuff;
+    size_t outBuffSize;
+    size_t outStart;
+    size_t outEnd;
+    size_t lhSize;
+#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>=1)
+    void* legacyContext;
+    U32 previousLegacyVersion;
+    U32 legacyVersion;
+#endif
+    U32 hostageByte;
+    int noForwardProgress;
+    ZSTD_bufferMode_e outBufferMode;
+    ZSTD_outBuffer expectedOutBuffer;
+
+    /* workspace */
+    BYTE* litBuffer;
+    const BYTE* litBufferEnd;
+    ZSTD_litLocation_e litBufferLocation;
+    BYTE litExtraBuffer[ZSTD_LITBUFFEREXTRASIZE + WILDCOPY_OVERLENGTH]; /* literal buffer can be split between storage within dst and within this scratch buffer */
+    BYTE headerBuffer[ZSTD_FRAMEHEADERSIZE_MAX];
+
+    size_t oversizedDuration;
+
+#ifdef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION
+    void const* dictContentBeginForFuzzing;
+    void const* dictContentEndForFuzzing;
+#endif
+
+    /* Tracing */
+#if ZSTD_TRACE
+    ZSTD_TraceCtx traceCtx;
+#endif
+};  /* typedef'd to ZSTD_DCtx within "zstd.h" */
+
+MEM_STATIC int ZSTD_DCtx_get_bmi2(const struct ZSTD_DCtx_s *dctx) {
+#if DYNAMIC_BMI2 != 0
+	return dctx->bmi2;
+#else
+    (void)dctx;
+	return 0;
+#endif
+}
+
+/*-*******************************************************
+ *  Shared internal functions
+ *********************************************************/
+
+/*! ZSTD_loadDEntropy() :
+ *  dict : must point at beginning of a valid zstd dictionary.
+ * @return : size of dictionary header (size of magic number + dict ID + entropy tables) */
+size_t ZSTD_loadDEntropy(ZSTD_entropyDTables_t* entropy,
+                   const void* const dict, size_t const dictSize);
+
+/*! ZSTD_checkContinuity() :
+ *  check if next `dst` follows previous position, where decompression ended.
+ *  If yes, do nothing (continue on current segment).
+ *  If not, classify previous segment as "external dictionary", and start a new segment.
+ *  This function cannot fail. */
+void ZSTD_checkContinuity(ZSTD_DCtx* dctx, const void* dst, size_t dstSize);
+
+
+#endif /* ZSTD_DECOMPRESS_INTERNAL_H */

external/zstd/libzstd.mk

@@ -0,0 +1,214 @@
+# ################################################################
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under both the BSD-style license (found in the
+# LICENSE file in the root directory of this source tree) and the GPLv2 (found
+# in the COPYING file in the root directory of this source tree).
+# You may select, at your option, one of the above-listed licenses.
+# ################################################################
+
+##################################################################
+# Input Variables
+##################################################################
+
+# Zstd lib directory
+LIBZSTD ?= ./
+
+# ZSTD_LIB_MINIFY is a helper variable that
+# configures a bunch of other variables to space-optimized defaults.
+ZSTD_LIB_MINIFY ?= 0
+
+# Legacy support
+ifneq ($(ZSTD_LIB_MINIFY), 0)
+  ZSTD_LEGACY_SUPPORT ?= 0
+else
+  ZSTD_LEGACY_SUPPORT ?= 5
+endif
+ZSTD_LEGACY_MULTITHREADED_API ?= 0
+
+# Build size optimizations
+ifneq ($(ZSTD_LIB_MINIFY), 0)
+  HUF_FORCE_DECOMPRESS_X1 ?= 1
+  HUF_FORCE_DECOMPRESS_X2 ?= 0
+  ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT ?= 1
+  ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG ?= 0
+  ZSTD_NO_INLINE ?= 1
+  ZSTD_STRIP_ERROR_STRINGS ?= 1
+else
+  HUF_FORCE_DECOMPRESS_X1 ?= 0
+  HUF_FORCE_DECOMPRESS_X2 ?= 0
+  ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT ?= 0
+  ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG ?= 0
+  ZSTD_NO_INLINE ?= 0
+  ZSTD_STRIP_ERROR_STRINGS ?= 0
+endif
+
+# Assembly support
+ZSTD_NO_ASM ?= 0
+
+##################################################################
+# libzstd helpers
+##################################################################
+
+VOID ?= /dev/null
+
+# Make 4.3 doesn't support '\#' anymore (https://lwn.net/Articles/810071/)
+NUM_SYMBOL := \#
+
+# define silent mode as default (verbose mode with V=1 or VERBOSE=1)
+$(V)$(VERBOSE).SILENT:
+
+# When cross-compiling from linux to windows,
+# one might need to specify TARGET_SYSTEM as "Windows."
+# Building from Fedora fails without it.
+# (but Ubuntu and Debian don't need to set anything)
+TARGET_SYSTEM ?= $(OS)
+
+# Version numbers
+LIBVER_SRC := $(LIBZSTD)/zstd.h
+LIBVER_MAJOR_SCRIPT:=`sed -n '/define ZSTD_VERSION_MAJOR/s/.*[[:blank:]]\([0-9][0-9]*\).*/\1/p' < $(LIBVER_SRC)`
+LIBVER_MINOR_SCRIPT:=`sed -n '/define ZSTD_VERSION_MINOR/s/.*[[:blank:]]\([0-9][0-9]*\).*/\1/p' < $(LIBVER_SRC)`
+LIBVER_PATCH_SCRIPT:=`sed -n '/define ZSTD_VERSION_RELEASE/s/.*[[:blank:]]\([0-9][0-9]*\).*/\1/p' < $(LIBVER_SRC)`
+LIBVER_SCRIPT:= $(LIBVER_MAJOR_SCRIPT).$(LIBVER_MINOR_SCRIPT).$(LIBVER_PATCH_SCRIPT)
+LIBVER_MAJOR := $(shell echo $(LIBVER_MAJOR_SCRIPT))
+LIBVER_MINOR := $(shell echo $(LIBVER_MINOR_SCRIPT))
+LIBVER_PATCH := $(shell echo $(LIBVER_PATCH_SCRIPT))
+LIBVER := $(shell echo $(LIBVER_SCRIPT))
+CCVER := $(shell $(CC) --version)
+ZSTD_VERSION?= $(LIBVER)
+
+ifneq ($(ZSTD_LIB_MINIFY), 0)
+  HAVE_CC_OZ ?= $(shell echo "" | $(CC) -Oz -x c -c - -o /dev/null 2> /dev/null && echo 1 || echo 0)
+ifneq ($(HAVE_CC_OZ), 0)
+    # Some compilers (clang) support an even more space-optimized setting.
+    CFLAGS += -Oz
+else
+    CFLAGS += -Os
+endif
+  CFLAGS += -fno-stack-protector -fomit-frame-pointer -fno-ident \
+            -DDYNAMIC_BMI2=0 -DNDEBUG
+else
+  CFLAGS ?= -O3
+endif
+
+DEBUGLEVEL ?= 0
+CPPFLAGS += -DXXH_NAMESPACE=ZSTD_ -DDEBUGLEVEL=$(DEBUGLEVEL)
+ifeq ($(TARGET_SYSTEM),Windows_NT)   # MinGW assumed
+  CPPFLAGS += -D__USE_MINGW_ANSI_STDIO   # compatibility with %zu formatting
+endif
+DEBUGFLAGS= -Wall -Wextra -Wcast-qual -Wcast-align -Wshadow \
+            -Wstrict-aliasing=1 -Wswitch-enum -Wdeclaration-after-statement \
+            -Wstrict-prototypes -Wundef -Wpointer-arith \
+            -Wvla -Wformat=2 -Winit-self -Wfloat-equal -Wwrite-strings \
+            -Wredundant-decls -Wmissing-prototypes -Wc++-compat
+CFLAGS   += $(DEBUGFLAGS) $(MOREFLAGS)
+ASFLAGS  += $(DEBUGFLAGS) $(MOREFLAGS) $(CFLAGS)
+LDFLAGS  += $(MOREFLAGS)
+FLAGS     = $(CPPFLAGS) $(CFLAGS) $(ASFLAGS) $(LDFLAGS)
+
+ifndef ALREADY_APPENDED_NOEXECSTACK
+export ALREADY_APPENDED_NOEXECSTACK := 1
+ifeq ($(shell echo "int main(int argc, char* argv[]) { (void)argc; (void)argv; return 0; }" | $(CC) $(FLAGS) -z noexecstack -x c -Werror - -o $(VOID) 2>$(VOID) && echo 1 || echo 0),1)
+LDFLAGS += -z noexecstack
+endif
+ifeq ($(shell echo | $(CC) $(FLAGS) -Wa,--noexecstack -x assembler -Werror -c - -o $(VOID) 2>$(VOID) && echo 1 || echo 0),1)
+CFLAGS  += -Wa,--noexecstack
+# CFLAGS are also added to ASFLAGS
+else ifeq ($(shell echo | $(CC) $(FLAGS) -Qunused-arguments -Wa,--noexecstack -x assembler -Werror -c - -o $(VOID) 2>$(VOID) && echo 1 || echo 0),1)
+# See e.g.: https://github.com/android/ndk/issues/171
+CFLAGS  += -Qunused-arguments -Wa,--noexecstack
+# CFLAGS are also added to ASFLAGS
+endif
+endif
+
+ifeq ($(shell echo "int main(int argc, char* argv[]) { (void)argc; (void)argv; return 0; }" | $(CC) $(FLAGS) -z cet-report=error -x c -Werror - -o $(VOID) 2>$(VOID) && echo 1 || echo 0),1)
+LDFLAGS += -z cet-report=error
+endif
+
+HAVE_COLORNEVER = $(shell echo a | grep --color=never a > /dev/null 2> /dev/null && echo 1 || echo 0)
+GREP_OPTIONS ?=
+ifeq ($(HAVE_COLORNEVER), 1)
+  GREP_OPTIONS += --color=never
+endif
+GREP = grep $(GREP_OPTIONS)
+
+ZSTD_COMMON_FILES := $(sort $(wildcard $(LIBZSTD)/common/*.c))
+ZSTD_COMPRESS_FILES := $(sort $(wildcard $(LIBZSTD)/compress/*.c))
+ZSTD_DECOMPRESS_FILES := $(sort $(wildcard $(LIBZSTD)/decompress/*.c))
+ZSTD_DICTBUILDER_FILES := $(sort $(wildcard $(LIBZSTD)/dictBuilder/*.c))
+ZSTD_DEPRECATED_FILES := $(sort $(wildcard $(LIBZSTD)/deprecated/*.c))
+ZSTD_LEGACY_FILES :=
+
+ZSTD_DECOMPRESS_AMD64_ASM_FILES := $(sort $(wildcard $(LIBZSTD)/decompress/*_amd64.S))
+
+ifneq ($(ZSTD_NO_ASM), 0)
+  CPPFLAGS += -DZSTD_DISABLE_ASM
+else
+  # Unconditionally add the ASM files they are disabled by
+  # macros in the .S file.
+  ZSTD_DECOMPRESS_FILES += $(ZSTD_DECOMPRESS_AMD64_ASM_FILES)
+endif
+
+ifneq ($(HUF_FORCE_DECOMPRESS_X1), 0)
+  CFLAGS += -DHUF_FORCE_DECOMPRESS_X1
+endif
+
+ifneq ($(HUF_FORCE_DECOMPRESS_X2), 0)
+  CFLAGS += -DHUF_FORCE_DECOMPRESS_X2
+endif
+
+ifneq ($(ZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT), 0)
+  CFLAGS += -DZSTD_FORCE_DECOMPRESS_SEQUENCES_SHORT
+endif
+
+ifneq ($(ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG), 0)
+  CFLAGS += -DZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG
+endif
+
+ifneq ($(ZSTD_NO_INLINE), 0)
+  CFLAGS += -DZSTD_NO_INLINE
+endif
+
+ifneq ($(ZSTD_STRIP_ERROR_STRINGS), 0)
+  CFLAGS += -DZSTD_STRIP_ERROR_STRINGS
+endif
+
+ifneq ($(ZSTD_LEGACY_MULTITHREADED_API), 0)
+  CFLAGS += -DZSTD_LEGACY_MULTITHREADED_API
+endif
+
+ifneq ($(ZSTD_LEGACY_SUPPORT), 0)
+ifeq ($(shell test $(ZSTD_LEGACY_SUPPORT) -lt 8; echo $$?), 0)
+  ZSTD_LEGACY_FILES += $(shell ls $(LIBZSTD)/legacy/*.c | $(GREP) 'v0[$(ZSTD_LEGACY_SUPPORT)-7]')
+endif
+endif
+CPPFLAGS  += -DZSTD_LEGACY_SUPPORT=$(ZSTD_LEGACY_SUPPORT)
+
+UNAME := $(shell uname)
+
+ifndef BUILD_DIR
+ifeq ($(UNAME), Darwin)
+  ifeq ($(shell md5 < /dev/null > /dev/null; echo $$?), 0)
+    HASH ?= md5
+  endif
+else ifeq ($(UNAME), FreeBSD)
+  HASH ?= gmd5sum
+else ifeq ($(UNAME), NetBSD)
+  HASH ?= md5 -n
+else ifeq ($(UNAME), OpenBSD)
+  HASH ?= md5
+endif
+HASH ?= md5sum
+
+HASH_DIR = conf_$(shell echo $(CC) $(CPPFLAGS) $(CFLAGS) $(LDFLAGS) $(ZSTD_FILES) | $(HASH) | cut -f 1 -d " " )
+HAVE_HASH :=$(shell echo 1 | $(HASH) > /dev/null && echo 1 || echo 0)
+ifeq ($(HAVE_HASH),0)
+  $(info warning : could not find HASH ($(HASH)), needed to differentiate builds using different flags)
+  BUILD_DIR := obj/generic_noconf
+endif
+endif # BUILD_DIR
+
+ZSTD_SUBDIR := $(LIBZSTD)/common $(LIBZSTD)/compress $(LIBZSTD)/decompress $(LIBZSTD)/dictBuilder $(LIBZSTD)/legacy $(LIBZSTD)/deprecated
+vpath %.c $(ZSTD_SUBDIR)
+vpath %.S $(ZSTD_SUBDIR)

external/zstd/libzstd.pc.in

@@ -0,0 +1,16 @@
+#   ZSTD - standard compression algorithm
+#   Copyright (c) Meta Platforms, Inc. and affiliates.
+#   BSD 2-Clause License (https://opensource.org/licenses/bsd-license.php)
+
+prefix=@PREFIX@
+exec_prefix=@EXEC_PREFIX@
+includedir=@INCLUDEDIR@
+libdir=@LIBDIR@
+
+Name: zstd
+Description: fast lossless compression algorithm library
+URL: https://facebook.github.io/zstd/
+Version: @VERSION@
+Libs: -L${libdir} -lzstd
+Libs.private: @LIBS_PRIVATE@
+Cflags: -I${includedir}

external/zstd/module.modulemap

@@ -0,0 +1,35 @@
+module libzstd [extern_c] {
+    header "zstd.h"
+    export *
+    config_macros [exhaustive] \
+        /* zstd.h */ \
+        ZSTD_STATIC_LINKING_ONLY, \
+        ZSTDLIB_VISIBILITY, \
+        ZSTDLIB_VISIBLE, \
+        ZSTDLIB_HIDDEN, \
+        ZSTD_DLL_EXPORT, \
+        ZSTDLIB_STATIC_API, \
+        ZSTD_DISABLE_DEPRECATE_WARNINGS, \
+        ZSTD_CLEVEL_DEFAULT, \
+        /* zdict.h */ \
+        ZDICT_STATIC_LINKING_ONLY, \
+        ZDICTLIB_VISIBLE, \
+        ZDICTLIB_HIDDEN, \
+        ZDICTLIB_VISIBILITY, \
+        ZDICTLIB_STATIC_API, \
+        ZDICT_DISABLE_DEPRECATE_WARNINGS, \
+        /* zstd_errors.h */ \
+        ZSTDERRORLIB_VISIBLE, \
+        ZSTDERRORLIB_HIDDEN, \
+        ZSTDERRORLIB_VISIBILITY
+
+    module dictbuilder [extern_c] {
+        header "zdict.h"
+        export *
+    }
+
+    module errors [extern_c] {
+        header "zstd_errors.h"
+        export *
+    }
+}

external/zstd/zdict.h

@@ -0,0 +1,474 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+#ifndef ZSTD_ZDICT_H
+#define ZSTD_ZDICT_H
+
+/*======  Dependencies  ======*/
+#include <stddef.h>  /* size_t */
+
+
+/* =====   ZDICTLIB_API : control library symbols visibility   ===== */
+#ifndef ZDICTLIB_VISIBLE
+   /* Backwards compatibility with old macro name */
+#  ifdef ZDICTLIB_VISIBILITY
+#    define ZDICTLIB_VISIBLE ZDICTLIB_VISIBILITY
+#  elif defined(__GNUC__) && (__GNUC__ >= 4) && !defined(__MINGW32__)
+#    define ZDICTLIB_VISIBLE __attribute__ ((visibility ("default")))
+#  else
+#    define ZDICTLIB_VISIBLE
+#  endif
+#endif
+
+#ifndef ZDICTLIB_HIDDEN
+#  if defined(__GNUC__) && (__GNUC__ >= 4) && !defined(__MINGW32__)
+#    define ZDICTLIB_HIDDEN __attribute__ ((visibility ("hidden")))
+#  else
+#    define ZDICTLIB_HIDDEN
+#  endif
+#endif
+
+#if defined(ZSTD_DLL_EXPORT) && (ZSTD_DLL_EXPORT==1)
+#  define ZDICTLIB_API __declspec(dllexport) ZDICTLIB_VISIBLE
+#elif defined(ZSTD_DLL_IMPORT) && (ZSTD_DLL_IMPORT==1)
+#  define ZDICTLIB_API __declspec(dllimport) ZDICTLIB_VISIBLE /* It isn't required but allows to generate better code, saving a function pointer load from the IAT and an indirect jump.*/
+#else
+#  define ZDICTLIB_API ZDICTLIB_VISIBLE
+#endif
+
+/*******************************************************************************
+ * Zstd dictionary builder
+ *
+ * FAQ
+ * ===
+ * Why should I use a dictionary?
+ * ------------------------------
+ *
+ * Zstd can use dictionaries to improve compression ratio of small data.
+ * Traditionally small files don't compress well because there is very little
+ * repetition in a single sample, since it is small. But, if you are compressing
+ * many similar files, like a bunch of JSON records that share the same
+ * structure, you can train a dictionary on ahead of time on some samples of
+ * these files. Then, zstd can use the dictionary to find repetitions that are
+ * present across samples. This can vastly improve compression ratio.
+ *
+ * When is a dictionary useful?
+ * ----------------------------
+ *
+ * Dictionaries are useful when compressing many small files that are similar.
+ * The larger a file is, the less benefit a dictionary will have. Generally,
+ * we don't expect dictionary compression to be effective past 100KB. And the
+ * smaller a file is, the more we would expect the dictionary to help.
+ *
+ * How do I use a dictionary?
+ * --------------------------
+ *
+ * Simply pass the dictionary to the zstd compressor with
+ * `ZSTD_CCtx_loadDictionary()`. The same dictionary must then be passed to
+ * the decompressor, using `ZSTD_DCtx_loadDictionary()`. There are other
+ * more advanced functions that allow selecting some options, see zstd.h for
+ * complete documentation.
+ *
+ * What is a zstd dictionary?
+ * --------------------------
+ *
+ * A zstd dictionary has two pieces: Its header, and its content. The header
+ * contains a magic number, the dictionary ID, and entropy tables. These
+ * entropy tables allow zstd to save on header costs in the compressed file,
+ * which really matters for small data. The content is just bytes, which are
+ * repeated content that is common across many samples.
+ *
+ * What is a raw content dictionary?
+ * ---------------------------------
+ *
+ * A raw content dictionary is just bytes. It doesn't have a zstd dictionary
+ * header, a dictionary ID, or entropy tables. Any buffer is a valid raw
+ * content dictionary.
+ *
+ * How do I train a dictionary?
+ * ----------------------------
+ *
+ * Gather samples from your use case. These samples should be similar to each
+ * other. If you have several use cases, you could try to train one dictionary
+ * per use case.
+ *
+ * Pass those samples to `ZDICT_trainFromBuffer()` and that will train your
+ * dictionary. There are a few advanced versions of this function, but this
+ * is a great starting point. If you want to further tune your dictionary
+ * you could try `ZDICT_optimizeTrainFromBuffer_cover()`. If that is too slow
+ * you can try `ZDICT_optimizeTrainFromBuffer_fastCover()`.
+ *
+ * If the dictionary training function fails, that is likely because you
+ * either passed too few samples, or a dictionary would not be effective
+ * for your data. Look at the messages that the dictionary trainer printed,
+ * if it doesn't say too few samples, then a dictionary would not be effective.
+ *
+ * How large should my dictionary be?
+ * ----------------------------------
+ *
+ * A reasonable dictionary size, the `dictBufferCapacity`, is about 100KB.
+ * The zstd CLI defaults to a 110KB dictionary. You likely don't need a
+ * dictionary larger than that. But, most use cases can get away with a
+ * smaller dictionary. The advanced dictionary builders can automatically
+ * shrink the dictionary for you, and select the smallest size that doesn't
+ * hurt compression ratio too much. See the `shrinkDict` parameter.
+ * A smaller dictionary can save memory, and potentially speed up
+ * compression.
+ *
+ * How many samples should I provide to the dictionary builder?
+ * ------------------------------------------------------------
+ *
+ * We generally recommend passing ~100x the size of the dictionary
+ * in samples. A few thousand should suffice. Having too few samples
+ * can hurt the dictionaries effectiveness. Having more samples will
+ * only improve the dictionaries effectiveness. But having too many
+ * samples can slow down the dictionary builder.
+ *
+ * How do I determine if a dictionary will be effective?
+ * -----------------------------------------------------
+ *
+ * Simply train a dictionary and try it out. You can use zstd's built in
+ * benchmarking tool to test the dictionary effectiveness.
+ *
+ *   # Benchmark levels 1-3 without a dictionary
+ *   zstd -b1e3 -r /path/to/my/files
+ *   # Benchmark levels 1-3 with a dictionary
+ *   zstd -b1e3 -r /path/to/my/files -D /path/to/my/dictionary
+ *
+ * When should I retrain a dictionary?
+ * -----------------------------------
+ *
+ * You should retrain a dictionary when its effectiveness drops. Dictionary
+ * effectiveness drops as the data you are compressing changes. Generally, we do
+ * expect dictionaries to "decay" over time, as your data changes, but the rate
+ * at which they decay depends on your use case. Internally, we regularly
+ * retrain dictionaries, and if the new dictionary performs significantly
+ * better than the old dictionary, we will ship the new dictionary.
+ *
+ * I have a raw content dictionary, how do I turn it into a zstd dictionary?
+ * -------------------------------------------------------------------------
+ *
+ * If you have a raw content dictionary, e.g. by manually constructing it, or
+ * using a third-party dictionary builder, you can turn it into a zstd
+ * dictionary by using `ZDICT_finalizeDictionary()`. You'll also have to
+ * provide some samples of the data. It will add the zstd header to the
+ * raw content, which contains a dictionary ID and entropy tables, which
+ * will improve compression ratio, and allow zstd to write the dictionary ID
+ * into the frame, if you so choose.
+ *
+ * Do I have to use zstd's dictionary builder?
+ * -------------------------------------------
+ *
+ * No! You can construct dictionary content however you please, it is just
+ * bytes. It will always be valid as a raw content dictionary. If you want
+ * a zstd dictionary, which can improve compression ratio, use
+ * `ZDICT_finalizeDictionary()`.
+ *
+ * What is the attack surface of a zstd dictionary?
+ * ------------------------------------------------
+ *
+ * Zstd is heavily fuzz tested, including loading fuzzed dictionaries, so
+ * zstd should never crash, or access out-of-bounds memory no matter what
+ * the dictionary is. However, if an attacker can control the dictionary
+ * during decompression, they can cause zstd to generate arbitrary bytes,
+ * just like if they controlled the compressed data.
+ *
+ ******************************************************************************/
+
+
+/*! ZDICT_trainFromBuffer():
+ *  Train a dictionary from an array of samples.
+ *  Redirect towards ZDICT_optimizeTrainFromBuffer_fastCover() single-threaded, with d=8, steps=4,
+ *  f=20, and accel=1.
+ *  Samples must be stored concatenated in a single flat buffer `samplesBuffer`,
+ *  supplied with an array of sizes `samplesSizes`, providing the size of each sample, in order.
+ *  The resulting dictionary will be saved into `dictBuffer`.
+ * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
+ *          or an error code, which can be tested with ZDICT_isError().
+ *  Note:  Dictionary training will fail if there are not enough samples to construct a
+ *         dictionary, or if most of the samples are too small (< 8 bytes being the lower limit).
+ *         If dictionary training fails, you should use zstd without a dictionary, as the dictionary
+ *         would've been ineffective anyways. If you believe your samples would benefit from a dictionary
+ *         please open an issue with details, and we can look into it.
+ *  Note: ZDICT_trainFromBuffer()'s memory usage is about 6 MB.
+ *  Tips: In general, a reasonable dictionary has a size of ~ 100 KB.
+ *        It's possible to select smaller or larger size, just by specifying `dictBufferCapacity`.
+ *        In general, it's recommended to provide a few thousands samples, though this can vary a lot.
+ *        It's recommended that total size of all samples be about ~x100 times the target size of dictionary.
+ */
+ZDICTLIB_API size_t ZDICT_trainFromBuffer(void* dictBuffer, size_t dictBufferCapacity,
+                                    const void* samplesBuffer,
+                                    const size_t* samplesSizes, unsigned nbSamples);
+
+typedef struct {
+    int      compressionLevel;   /**< optimize for a specific zstd compression level; 0 means default */
+    unsigned notificationLevel;  /**< Write log to stderr; 0 = none (default); 1 = errors; 2 = progression; 3 = details; 4 = debug; */
+    unsigned dictID;             /**< force dictID value; 0 means auto mode (32-bits random value)
+                                  *   NOTE: The zstd format reserves some dictionary IDs for future use.
+                                  *         You may use them in private settings, but be warned that they
+                                  *         may be used by zstd in a public dictionary registry in the future.
+                                  *         These dictionary IDs are:
+                                  *           - low range  : <= 32767
+                                  *           - high range : >= (2^31)
+                                  */
+} ZDICT_params_t;
+
+/*! ZDICT_finalizeDictionary():
+ * Given a custom content as a basis for dictionary, and a set of samples,
+ * finalize dictionary by adding headers and statistics according to the zstd
+ * dictionary format.
+ *
+ * Samples must be stored concatenated in a flat buffer `samplesBuffer`,
+ * supplied with an array of sizes `samplesSizes`, providing the size of each
+ * sample in order. The samples are used to construct the statistics, so they
+ * should be representative of what you will compress with this dictionary.
+ *
+ * The compression level can be set in `parameters`. You should pass the
+ * compression level you expect to use in production. The statistics for each
+ * compression level differ, so tuning the dictionary for the compression level
+ * can help quite a bit.
+ *
+ * You can set an explicit dictionary ID in `parameters`, or allow us to pick
+ * a random dictionary ID for you, but we can't guarantee no collisions.
+ *
+ * The dstDictBuffer and the dictContent may overlap, and the content will be
+ * appended to the end of the header. If the header + the content doesn't fit in
+ * maxDictSize the beginning of the content is truncated to make room, since it
+ * is presumed that the most profitable content is at the end of the dictionary,
+ * since that is the cheapest to reference.
+ *
+ * `maxDictSize` must be >= max(dictContentSize, ZSTD_DICTSIZE_MIN).
+ *
+ * @return: size of dictionary stored into `dstDictBuffer` (<= `maxDictSize`),
+ *          or an error code, which can be tested by ZDICT_isError().
+ * Note: ZDICT_finalizeDictionary() will push notifications into stderr if
+ *       instructed to, using notificationLevel>0.
+ * NOTE: This function currently may fail in several edge cases including:
+ *         * Not enough samples
+ *         * Samples are uncompressible
+ *         * Samples are all exactly the same
+ */
+ZDICTLIB_API size_t ZDICT_finalizeDictionary(void* dstDictBuffer, size_t maxDictSize,
+                                const void* dictContent, size_t dictContentSize,
+                                const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples,
+                                ZDICT_params_t parameters);
+
+
+/*======   Helper functions   ======*/
+ZDICTLIB_API unsigned ZDICT_getDictID(const void* dictBuffer, size_t dictSize);  /**< extracts dictID; @return zero if error (not a valid dictionary) */
+ZDICTLIB_API size_t ZDICT_getDictHeaderSize(const void* dictBuffer, size_t dictSize);  /* returns dict header size; returns a ZSTD error code on failure */
+ZDICTLIB_API unsigned ZDICT_isError(size_t errorCode);
+ZDICTLIB_API const char* ZDICT_getErrorName(size_t errorCode);
+
+#endif   /* ZSTD_ZDICT_H */
+
+#if defined(ZDICT_STATIC_LINKING_ONLY) && !defined(ZSTD_ZDICT_H_STATIC)
+#define ZSTD_ZDICT_H_STATIC
+
+/* This can be overridden externally to hide static symbols. */
+#ifndef ZDICTLIB_STATIC_API
+#  if defined(ZSTD_DLL_EXPORT) && (ZSTD_DLL_EXPORT==1)
+#    define ZDICTLIB_STATIC_API __declspec(dllexport) ZDICTLIB_VISIBLE
+#  elif defined(ZSTD_DLL_IMPORT) && (ZSTD_DLL_IMPORT==1)
+#    define ZDICTLIB_STATIC_API __declspec(dllimport) ZDICTLIB_VISIBLE
+#  else
+#    define ZDICTLIB_STATIC_API ZDICTLIB_VISIBLE
+#  endif
+#endif
+
+/* ====================================================================================
+ * The definitions in this section are considered experimental.
+ * They should never be used with a dynamic library, as they may change in the future.
+ * They are provided for advanced usages.
+ * Use them only in association with static linking.
+ * ==================================================================================== */
+
+#define ZDICT_DICTSIZE_MIN    256
+/* Deprecated: Remove in v1.6.0 */
+#define ZDICT_CONTENTSIZE_MIN 128
+
+/*! ZDICT_cover_params_t:
+ *  k and d are the only required parameters.
+ *  For others, value 0 means default.
+ */
+typedef struct {
+    unsigned k;                  /* Segment size : constraint: 0 < k : Reasonable range [16, 2048+] */
+    unsigned d;                  /* dmer size : constraint: 0 < d <= k : Reasonable range [6, 16] */
+    unsigned steps;              /* Number of steps : Only used for optimization : 0 means default (40) : Higher means more parameters checked */
+    unsigned nbThreads;          /* Number of threads : constraint: 0 < nbThreads : 1 means single-threaded : Only used for optimization : Ignored if ZSTD_MULTITHREAD is not defined */
+    double splitPoint;           /* Percentage of samples used for training: Only used for optimization : the first nbSamples * splitPoint samples will be used to training, the last nbSamples * (1 - splitPoint) samples will be used for testing, 0 means default (1.0), 1.0 when all samples are used for both training and testing */
+    unsigned shrinkDict;         /* Train dictionaries to shrink in size starting from the minimum size and selects the smallest dictionary that is shrinkDictMaxRegression% worse than the largest dictionary. 0 means no shrinking and 1 means shrinking  */
+    unsigned shrinkDictMaxRegression; /* Sets shrinkDictMaxRegression so that a smaller dictionary can be at worse shrinkDictMaxRegression% worse than the max dict size dictionary. */
+    ZDICT_params_t zParams;
+} ZDICT_cover_params_t;
+
+typedef struct {
+    unsigned k;                  /* Segment size : constraint: 0 < k : Reasonable range [16, 2048+] */
+    unsigned d;                  /* dmer size : constraint: 0 < d <= k : Reasonable range [6, 16] */
+    unsigned f;                  /* log of size of frequency array : constraint: 0 < f <= 31 : 1 means default(20)*/
+    unsigned steps;              /* Number of steps : Only used for optimization : 0 means default (40) : Higher means more parameters checked */
+    unsigned nbThreads;          /* Number of threads : constraint: 0 < nbThreads : 1 means single-threaded : Only used for optimization : Ignored if ZSTD_MULTITHREAD is not defined */
+    double splitPoint;           /* Percentage of samples used for training: Only used for optimization : the first nbSamples * splitPoint samples will be used to training, the last nbSamples * (1 - splitPoint) samples will be used for testing, 0 means default (0.75), 1.0 when all samples are used for both training and testing */
+    unsigned accel;              /* Acceleration level: constraint: 0 < accel <= 10, higher means faster and less accurate, 0 means default(1) */
+    unsigned shrinkDict;         /* Train dictionaries to shrink in size starting from the minimum size and selects the smallest dictionary that is shrinkDictMaxRegression% worse than the largest dictionary. 0 means no shrinking and 1 means shrinking  */
+    unsigned shrinkDictMaxRegression; /* Sets shrinkDictMaxRegression so that a smaller dictionary can be at worse shrinkDictMaxRegression% worse than the max dict size dictionary. */
+
+    ZDICT_params_t zParams;
+} ZDICT_fastCover_params_t;
+
+/*! ZDICT_trainFromBuffer_cover():
+ *  Train a dictionary from an array of samples using the COVER algorithm.
+ *  Samples must be stored concatenated in a single flat buffer `samplesBuffer`,
+ *  supplied with an array of sizes `samplesSizes`, providing the size of each sample, in order.
+ *  The resulting dictionary will be saved into `dictBuffer`.
+ * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
+ *          or an error code, which can be tested with ZDICT_isError().
+ *          See ZDICT_trainFromBuffer() for details on failure modes.
+ *  Note: ZDICT_trainFromBuffer_cover() requires about 9 bytes of memory for each input byte.
+ *  Tips: In general, a reasonable dictionary has a size of ~ 100 KB.
+ *        It's possible to select smaller or larger size, just by specifying `dictBufferCapacity`.
+ *        In general, it's recommended to provide a few thousands samples, though this can vary a lot.
+ *        It's recommended that total size of all samples be about ~x100 times the target size of dictionary.
+ */
+ZDICTLIB_STATIC_API size_t ZDICT_trainFromBuffer_cover(
+          void *dictBuffer, size_t dictBufferCapacity,
+    const void *samplesBuffer, const size_t *samplesSizes, unsigned nbSamples,
+          ZDICT_cover_params_t parameters);
+
+/*! ZDICT_optimizeTrainFromBuffer_cover():
+ * The same requirements as above hold for all the parameters except `parameters`.
+ * This function tries many parameter combinations and picks the best parameters.
+ * `*parameters` is filled with the best parameters found,
+ * dictionary constructed with those parameters is stored in `dictBuffer`.
+ *
+ * All of the parameters d, k, steps are optional.
+ * If d is non-zero then we don't check multiple values of d, otherwise we check d = {6, 8}.
+ * if steps is zero it defaults to its default value.
+ * If k is non-zero then we don't check multiple values of k, otherwise we check steps values in [50, 2000].
+ *
+ * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
+ *          or an error code, which can be tested with ZDICT_isError().
+ *          On success `*parameters` contains the parameters selected.
+ *          See ZDICT_trainFromBuffer() for details on failure modes.
+ * Note: ZDICT_optimizeTrainFromBuffer_cover() requires about 8 bytes of memory for each input byte and additionally another 5 bytes of memory for each byte of memory for each thread.
+ */
+ZDICTLIB_STATIC_API size_t ZDICT_optimizeTrainFromBuffer_cover(
+          void* dictBuffer, size_t dictBufferCapacity,
+    const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples,
+          ZDICT_cover_params_t* parameters);
+
+/*! ZDICT_trainFromBuffer_fastCover():
+ *  Train a dictionary from an array of samples using a modified version of COVER algorithm.
+ *  Samples must be stored concatenated in a single flat buffer `samplesBuffer`,
+ *  supplied with an array of sizes `samplesSizes`, providing the size of each sample, in order.
+ *  d and k are required.
+ *  All other parameters are optional, will use default values if not provided
+ *  The resulting dictionary will be saved into `dictBuffer`.
+ * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
+ *          or an error code, which can be tested with ZDICT_isError().
+ *          See ZDICT_trainFromBuffer() for details on failure modes.
+ *  Note: ZDICT_trainFromBuffer_fastCover() requires 6 * 2^f bytes of memory.
+ *  Tips: In general, a reasonable dictionary has a size of ~ 100 KB.
+ *        It's possible to select smaller or larger size, just by specifying `dictBufferCapacity`.
+ *        In general, it's recommended to provide a few thousands samples, though this can vary a lot.
+ *        It's recommended that total size of all samples be about ~x100 times the target size of dictionary.
+ */
+ZDICTLIB_STATIC_API size_t ZDICT_trainFromBuffer_fastCover(void *dictBuffer,
+                    size_t dictBufferCapacity, const void *samplesBuffer,
+                    const size_t *samplesSizes, unsigned nbSamples,
+                    ZDICT_fastCover_params_t parameters);
+
+/*! ZDICT_optimizeTrainFromBuffer_fastCover():
+ * The same requirements as above hold for all the parameters except `parameters`.
+ * This function tries many parameter combinations (specifically, k and d combinations)
+ * and picks the best parameters. `*parameters` is filled with the best parameters found,
+ * dictionary constructed with those parameters is stored in `dictBuffer`.
+ * All of the parameters d, k, steps, f, and accel are optional.
+ * If d is non-zero then we don't check multiple values of d, otherwise we check d = {6, 8}.
+ * if steps is zero it defaults to its default value.
+ * If k is non-zero then we don't check multiple values of k, otherwise we check steps values in [50, 2000].
+ * If f is zero, default value of 20 is used.
+ * If accel is zero, default value of 1 is used.
+ *
+ * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
+ *          or an error code, which can be tested with ZDICT_isError().
+ *          On success `*parameters` contains the parameters selected.
+ *          See ZDICT_trainFromBuffer() for details on failure modes.
+ * Note: ZDICT_optimizeTrainFromBuffer_fastCover() requires about 6 * 2^f bytes of memory for each thread.
+ */
+ZDICTLIB_STATIC_API size_t ZDICT_optimizeTrainFromBuffer_fastCover(void* dictBuffer,
+                    size_t dictBufferCapacity, const void* samplesBuffer,
+                    const size_t* samplesSizes, unsigned nbSamples,
+                    ZDICT_fastCover_params_t* parameters);
+
+typedef struct {
+    unsigned selectivityLevel;   /* 0 means default; larger => select more => larger dictionary */
+    ZDICT_params_t zParams;
+} ZDICT_legacy_params_t;
+
+/*! ZDICT_trainFromBuffer_legacy():
+ *  Train a dictionary from an array of samples.
+ *  Samples must be stored concatenated in a single flat buffer `samplesBuffer`,
+ *  supplied with an array of sizes `samplesSizes`, providing the size of each sample, in order.
+ *  The resulting dictionary will be saved into `dictBuffer`.
+ * `parameters` is optional and can be provided with values set to 0 to mean "default".
+ * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
+ *          or an error code, which can be tested with ZDICT_isError().
+ *          See ZDICT_trainFromBuffer() for details on failure modes.
+ *  Tips: In general, a reasonable dictionary has a size of ~ 100 KB.
+ *        It's possible to select smaller or larger size, just by specifying `dictBufferCapacity`.
+ *        In general, it's recommended to provide a few thousands samples, though this can vary a lot.
+ *        It's recommended that total size of all samples be about ~x100 times the target size of dictionary.
+ *  Note: ZDICT_trainFromBuffer_legacy() will send notifications into stderr if instructed to, using notificationLevel>0.
+ */
+ZDICTLIB_STATIC_API size_t ZDICT_trainFromBuffer_legacy(
+    void* dictBuffer, size_t dictBufferCapacity,
+    const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples,
+    ZDICT_legacy_params_t parameters);
+
+
+/* Deprecation warnings */
+/* It is generally possible to disable deprecation warnings from compiler,
+   for example with -Wno-deprecated-declarations for gcc
+   or _CRT_SECURE_NO_WARNINGS in Visual.
+   Otherwise, it's also possible to manually define ZDICT_DISABLE_DEPRECATE_WARNINGS */
+#ifdef ZDICT_DISABLE_DEPRECATE_WARNINGS
+#  define ZDICT_DEPRECATED(message) /* disable deprecation warnings */
+#else
+#  define ZDICT_GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__)
+#  if defined (__cplusplus) && (__cplusplus >= 201402) /* C++14 or greater */
+#    define ZDICT_DEPRECATED(message) [[deprecated(message)]]
+#  elif defined(__clang__) || (ZDICT_GCC_VERSION >= 405)
+#    define ZDICT_DEPRECATED(message) __attribute__((deprecated(message)))
+#  elif (ZDICT_GCC_VERSION >= 301)
+#    define ZDICT_DEPRECATED(message) __attribute__((deprecated))
+#  elif defined(_MSC_VER)
+#    define ZDICT_DEPRECATED(message) __declspec(deprecated(message))
+#  else
+#    pragma message("WARNING: You need to implement ZDICT_DEPRECATED for this compiler")
+#    define ZDICT_DEPRECATED(message)
+#  endif
+#endif /* ZDICT_DISABLE_DEPRECATE_WARNINGS */
+
+ZDICT_DEPRECATED("use ZDICT_finalizeDictionary() instead")
+ZDICTLIB_STATIC_API
+size_t ZDICT_addEntropyTablesFromBuffer(void* dictBuffer, size_t dictContentSize, size_t dictBufferCapacity,
+                                  const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples);
+
+
+#endif   /* ZSTD_ZDICT_H_STATIC */
+
+#if defined (__cplusplus)
+}
+#endif

external/zstd/zstd_errors.h

@@ -0,0 +1,111 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_ERRORS_H_398273423
+#define ZSTD_ERRORS_H_398273423
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+/*===== dependency =====*/
+#include <stddef.h>   /* size_t */
+
+
+/* =====   ZSTDERRORLIB_API : control library symbols visibility   ===== */
+#ifndef ZSTDERRORLIB_VISIBLE
+   /* Backwards compatibility with old macro name */
+#  ifdef ZSTDERRORLIB_VISIBILITY
+#    define ZSTDERRORLIB_VISIBLE ZSTDERRORLIB_VISIBILITY
+#  elif defined(__GNUC__) && (__GNUC__ >= 4) && !defined(__MINGW32__)
+#    define ZSTDERRORLIB_VISIBLE __attribute__ ((visibility ("default")))
+#  else
+#    define ZSTDERRORLIB_VISIBLE
+#  endif
+#endif
+
+#ifndef ZSTDERRORLIB_HIDDEN
+#  if defined(__GNUC__) && (__GNUC__ >= 4) && !defined(__MINGW32__)
+#    define ZSTDERRORLIB_HIDDEN __attribute__ ((visibility ("hidden")))
+#  else
+#    define ZSTDERRORLIB_HIDDEN
+#  endif
+#endif
+
+#if defined(ZSTD_DLL_EXPORT) && (ZSTD_DLL_EXPORT==1)
+#  define ZSTDERRORLIB_API __declspec(dllexport) ZSTDERRORLIB_VISIBLE
+#elif defined(ZSTD_DLL_IMPORT) && (ZSTD_DLL_IMPORT==1)
+#  define ZSTDERRORLIB_API __declspec(dllimport) ZSTDERRORLIB_VISIBLE /* It isn't required but allows to generate better code, saving a function pointer load from the IAT and an indirect jump.*/
+#else
+#  define ZSTDERRORLIB_API ZSTDERRORLIB_VISIBLE
+#endif
+
+/*-*********************************************
+ *  Error codes list
+ *-*********************************************
+ *  Error codes _values_ are pinned down since v1.3.1 only.
+ *  Therefore, don't rely on values if you may link to any version < v1.3.1.
+ *
+ *  Only values < 100 are considered stable.
+ *
+ *  note 1 : this API shall be used with static linking only.
+ *           dynamic linking is not yet officially supported.
+ *  note 2 : Prefer relying on the enum than on its value whenever possible
+ *           This is the only supported way to use the error list < v1.3.1
+ *  note 3 : ZSTD_isError() is always correct, whatever the library version.
+ **********************************************/
+typedef enum {
+  ZSTD_error_no_error = 0,
+  ZSTD_error_GENERIC  = 1,
+  ZSTD_error_prefix_unknown                = 10,
+  ZSTD_error_version_unsupported           = 12,
+  ZSTD_error_frameParameter_unsupported    = 14,
+  ZSTD_error_frameParameter_windowTooLarge = 16,
+  ZSTD_error_corruption_detected = 20,
+  ZSTD_error_checksum_wrong      = 22,
+  ZSTD_error_literals_headerWrong = 24,
+  ZSTD_error_dictionary_corrupted      = 30,
+  ZSTD_error_dictionary_wrong          = 32,
+  ZSTD_error_dictionaryCreation_failed = 34,
+  ZSTD_error_parameter_unsupported   = 40,
+  ZSTD_error_parameter_combination_unsupported = 41,
+  ZSTD_error_parameter_outOfBound    = 42,
+  ZSTD_error_tableLog_tooLarge       = 44,
+  ZSTD_error_maxSymbolValue_tooLarge = 46,
+  ZSTD_error_maxSymbolValue_tooSmall = 48,
+  ZSTD_error_stabilityCondition_notRespected = 50,
+  ZSTD_error_stage_wrong       = 60,
+  ZSTD_error_init_missing      = 62,
+  ZSTD_error_memory_allocation = 64,
+  ZSTD_error_workSpace_tooSmall= 66,
+  ZSTD_error_dstSize_tooSmall = 70,
+  ZSTD_error_srcSize_wrong    = 72,
+  ZSTD_error_dstBuffer_null   = 74,
+  /* following error codes are __NOT STABLE__, they can be removed or changed in future versions */
+  ZSTD_error_frameIndex_tooLarge = 100,
+  ZSTD_error_seekableIO          = 102,
+  ZSTD_error_dstBuffer_wrong     = 104,
+  ZSTD_error_srcBuffer_wrong     = 105,
+  ZSTD_error_externalMatchFinder_failed = 106,
+  ZSTD_error_maxCode = 120  /* never EVER use this value directly, it can change in future versions! Use ZSTD_isError() instead */
+} ZSTD_ErrorCode;
+
+/*! ZSTD_getErrorCode() :
+    convert a `size_t` function result into a `ZSTD_ErrorCode` enum type,
+    which can be used to compare with enum list published above */
+ZSTDERRORLIB_API ZSTD_ErrorCode ZSTD_getErrorCode(size_t functionResult);
+ZSTDERRORLIB_API const char* ZSTD_getErrorString(ZSTD_ErrorCode code);   /**< Same as ZSTD_getErrorName, but using a `ZSTD_ErrorCode` enum argument */
+
+
+#if defined (__cplusplus)
+}
+#endif
+
+#endif /* ZSTD_ERRORS_H_398273423 */

modules.yaml

@@ -1,334 +0,0 @@
-name: jsix
-templates: scripts/templates
-modules:
-    kernel:
-        kind: exe
-        output: jsix.elf
-        target: host
-        deps:
-            - cpu
-            - kutil
-        includes:
-            - src/kernel
-        source:
-            - src/kernel/apic.cpp
-            - src/kernel/ap_startup.s
-            - src/kernel/assert.cpp
-            - src/kernel/boot.s
-            - src/kernel/clock.cpp
-            - src/kernel/console.cpp
-            - src/kernel/cpprt.cpp
-            - src/kernel/cpu.cpp
-            - src/kernel/debug.cpp
-            - src/kernel/debug.s
-            - src/kernel/device_manager.cpp
-            - src/kernel/frame_allocator.cpp
-            - src/kernel/fs/gpt.cpp
-            - src/kernel/gdt.cpp
-            - src/kernel/gdtidt.s
-            - src/kernel/hpet.cpp
-            - src/kernel/idt.cpp
-            - src/kernel/interrupts.cpp
-            - src/kernel/interrupts.s
-            - src/kernel/io.cpp
-            - src/kernel/log.cpp
-            - src/kernel/main.cpp
-            - src/kernel/memory_bootstrap.cpp
-            - src/kernel/msr.cpp
-            - src/kernel/objects/channel.cpp
-            - src/kernel/objects/endpoint.cpp
-            - src/kernel/objects/kobject.cpp
-            - src/kernel/objects/thread.cpp
-            - src/kernel/objects/process.cpp
-            - src/kernel/objects/system.cpp
-            - src/kernel/objects/vm_area.cpp
-            - src/kernel/page_table.cpp
-            - src/kernel/page_tree.cpp
-            - src/kernel/pci.cpp
-            - src/kernel/scheduler.cpp
-            - src/kernel/serial.cpp
-            - src/kernel/symbol_table.cpp
-            - src/kernel/syscall.cpp
-            - src/kernel/syscall.s
-            - src/kernel/syscalls/channel.cpp
-            - src/kernel/syscalls/endpoint.cpp
-            - src/kernel/syscalls/object.cpp
-            - src/kernel/syscalls/process.cpp
-            - src/kernel/syscalls/system.cpp
-            - src/kernel/syscalls/thread.cpp
-            - src/kernel/syscalls/vm_area.cpp
-            - src/kernel/task.s
-            - src/kernel/tss.cpp
-            - src/kernel/vm_space.cpp
-
-    boot:
-        kind: exe
-        target: boot
-        output: boot.efi
-        deps:
-            - cpu
-        source:
-            - src/boot/main.cpp
-            - src/boot/console.cpp
-            - src/boot/error.cpp
-            - src/boot/fs.cpp
-            - src/boot/hardware.cpp
-            - src/boot/loader.cpp
-            - src/boot/memory.cpp
-            - src/boot/paging.cpp
-            - src/boot/status.cpp
-            - src/boot/support.cpp
-
-    nulldrv:
-        kind: exe
-        target: user
-        output: nulldrv.elf
-        deps:
-          - libc
-        source:
-            - src/drivers/nulldrv/io.cpp
-            - src/drivers/nulldrv/main.cpp
-            - src/drivers/nulldrv/serial.cpp
-
-    fb:
-        kind: exe
-        target: user
-        output: fb.elf
-        deps:
-          - libc
-        source:
-            - src/drivers/fb/font.cpp
-            - src/drivers/fb/main.cpp
-            - src/drivers/fb/screen.cpp
-            - src/drivers/fb/scrollback.cpp
-
-    kutil:
-        kind: lib
-        output: libkutil.a
-        includes:
-            - src/libraries/kutil/include
-        source:
-            - src/libraries/kutil/assert.cpp
-            - src/libraries/kutil/bip_buffer.cpp
-            - src/libraries/kutil/heap_allocator.cpp
-            - src/libraries/kutil/logger.cpp
-            - src/libraries/kutil/memory.cpp
-            - src/libraries/kutil/printf.c
-            - src/libraries/kutil/spinlock.cpp
-
-    cpu:
-        kind: lib
-        output: libcpu.a
-        includes:
-            - src/libraries/cpu/include
-        source:
-            - src/libraries/cpu/cpu_id.cpp
-
-    j6:
-      kind: lib
-      output: libj6.a
-      includes:
-        - src/libraries/j6/include
-      target: user
-      source:
-        - src/libraries/j6/syscalls.s
-
-    libc:
-        kind: lib
-        output: libc.a
-        includes:
-          - src/libraries/libc/include
-        deps:
-            - j6
-        target: user
-        defines:
-            - DISABLE_SSE
-            - LACKS_UNISTD_H
-            - LACKS_FCNTL_H
-            - LACKS_SYS_PARAM_H
-            - LACKS_SYS_MMAN_H
-            - LACKS_SCHED_H
-            - LACKS_STRINGS_H
-            - HAVE_MMAP=0
-              #- LACKS_STRING_H
-              #- LACKS_ERRNO_H
-              #- LACKS_STDLIB_H
-              #- LACKS_TIME_H
-        source:
-            - src/libraries/libc/arch/x86_64/_Exit.s
-            - src/libraries/libc/arch/x86_64/crt0.s
-            - src/libraries/libc/arch/x86_64/init_libc.c
-            - src/libraries/libc/ctype/isalnum.c
-            - src/libraries/libc/ctype/isalpha.c
-            - src/libraries/libc/ctype/isblank.c
-            - src/libraries/libc/ctype/iscntrl.c
-            - src/libraries/libc/ctype/isdigit.c
-            - src/libraries/libc/ctype/isgraph.c
-            - src/libraries/libc/ctype/islower.c
-            - src/libraries/libc/ctype/isprint.c
-            - src/libraries/libc/ctype/ispunct.c
-            - src/libraries/libc/ctype/isspace.c
-            - src/libraries/libc/ctype/isupper.c
-            - src/libraries/libc/ctype/isxdigit.c
-            - src/libraries/libc/ctype/tolower.c
-            - src/libraries/libc/ctype/toupper.c
-            - src/libraries/libc/inttypes/imaxabs.c
-            - src/libraries/libc/inttypes/imaxdiv.c
-            - src/libraries/libc/inttypes/strtoimax.c
-            - src/libraries/libc/inttypes/strtoumax.c
-            - src/libraries/libc/locale/localeconv.c
-            - src/libraries/libc/locale/setlocale.c
-            - src/libraries/libc/j6libc/assert.c
-            - src/libraries/libc/j6libc/errno.c
-            - src/libraries/libc/j6libc/allocpages.c
-            - src/libraries/libc/j6libc/atomax.c
-            - src/libraries/libc/j6libc/closeall.c
-            - src/libraries/libc/j6libc/close.c
-            - src/libraries/libc/j6libc/digits.c
-            - src/libraries/libc/j6libc/filemode.c
-            - src/libraries/libc/j6libc/fillbuffer.c
-            - src/libraries/libc/j6libc/flushbuffer.c
-            - src/libraries/libc/j6libc/is_leap.c
-            - src/libraries/libc/j6libc/load_lc_collate.c
-            - src/libraries/libc/j6libc/load_lc_ctype.c
-            - src/libraries/libc/j6libc/load_lc_messages.c
-            - src/libraries/libc/j6libc/load_lc_monetary.c
-            - src/libraries/libc/j6libc/load_lc_numeric.c
-            - src/libraries/libc/j6libc/load_lc_time.c
-            - src/libraries/libc/j6libc/load_lines.c
-            - src/libraries/libc/j6libc/open.c
-            - src/libraries/libc/j6libc/prepread.c
-            - src/libraries/libc/j6libc/prepwrite.c
-            - src/libraries/libc/j6libc/print.c
-            - src/libraries/libc/j6libc/rename.c
-            - src/libraries/libc/j6libc/scan.c
-            - src/libraries/libc/j6libc/seed.c
-            - src/libraries/libc/j6libc/seek.c
-            - src/libraries/libc/j6libc/stdinit.c
-            - src/libraries/libc/j6libc/strtox_main.c
-            - src/libraries/libc/j6libc/strtox_prelim.c
-            - src/libraries/libc/j6libc/sbrk.c
-            - src/libraries/libc/signal/raise.c
-            - src/libraries/libc/signal/signal.c
-            - src/libraries/libc/stdio/clearerr.c
-            - src/libraries/libc/stdio/fclose.c
-            - src/libraries/libc/stdio/feof.c
-            - src/libraries/libc/stdio/ferror.c
-            - src/libraries/libc/stdio/fflush.c
-            - src/libraries/libc/stdio/fgetc.c
-            - src/libraries/libc/stdio/fgetpos.c
-            - src/libraries/libc/stdio/fgets.c
-            - src/libraries/libc/stdio/fopen.c
-            - src/libraries/libc/stdio/fprintf.c
-            - src/libraries/libc/stdio/fputc.c
-            - src/libraries/libc/stdio/fputs.c
-            - src/libraries/libc/stdio/fread.c
-            - src/libraries/libc/stdio/freopen.c
-            - src/libraries/libc/stdio/fscanf.c
-            - src/libraries/libc/stdio/fseek.c
-            - src/libraries/libc/stdio/fsetpos.c
-            - src/libraries/libc/stdio/ftell.c
-            - src/libraries/libc/stdio/fwrite.c
-            - src/libraries/libc/stdio/getc.c
-            - src/libraries/libc/stdio/getchar.c
-            - src/libraries/libc/stdio/perror.c
-            - src/libraries/libc/stdio/printf.c
-            - src/libraries/libc/stdio/putc.c
-            - src/libraries/libc/stdio/putchar.c
-            - src/libraries/libc/stdio/puts.c
-            - src/libraries/libc/stdio/remove.c
-            - src/libraries/libc/stdio/rename.c
-            - src/libraries/libc/stdio/rewind.c
-            - src/libraries/libc/stdio/scanf.c
-            - src/libraries/libc/stdio/setbuf.c
-            - src/libraries/libc/stdio/setvbuf.c
-            - src/libraries/libc/stdio/snprintf.c
-            - src/libraries/libc/stdio/sprintf.c
-            - src/libraries/libc/stdio/sscanf.c
-            - src/libraries/libc/stdio/tmpfile.c
-            - src/libraries/libc/stdio/tmpnam.c
-            - src/libraries/libc/stdio/ungetc.c
-            - src/libraries/libc/stdio/vfprintf.c
-            - src/libraries/libc/stdio/vfscanf.c
-            - src/libraries/libc/stdio/vprintf.c
-            - src/libraries/libc/stdio/vscanf.c
-            - src/libraries/libc/stdio/vsnprintf.c
-            - src/libraries/libc/stdio/vsprintf.c
-            - src/libraries/libc/stdio/vsscanf.c
-            - src/libraries/libc/stdlib/abort.c
-            - src/libraries/libc/stdlib/abs.c
-            - src/libraries/libc/stdlib/atexit.c
-            - src/libraries/libc/stdlib/atoi.c
-            - src/libraries/libc/stdlib/atol.c
-            - src/libraries/libc/stdlib/atoll.c
-            - src/libraries/libc/stdlib/bsearch.c
-            - src/libraries/libc/stdlib/div.c
-            - src/libraries/libc/stdlib/exit.c
-            - src/libraries/libc/stdlib/_Exit.c
-            - src/libraries/libc/stdlib/getenv.c
-            - src/libraries/libc/stdlib/labs.c
-            - src/libraries/libc/stdlib/ldiv.c
-            - src/libraries/libc/stdlib/llabs.c
-            - src/libraries/libc/stdlib/lldiv.c
-            - src/libraries/libc/stdlib/malloc.c
-            - src/libraries/libc/stdlib/qsort.c
-            - src/libraries/libc/stdlib/rand.c
-            - src/libraries/libc/stdlib/srand.c
-            - src/libraries/libc/stdlib/strtol.c
-            - src/libraries/libc/stdlib/strtoll.c
-            - src/libraries/libc/stdlib/strtoul.c
-            - src/libraries/libc/stdlib/strtoull.c
-            - src/libraries/libc/stdlib/system.c
-            - src/libraries/libc/string/memchr.c
-            - src/libraries/libc/string/memcmp.c
-            - src/libraries/libc/string/memcpy.c
-            - src/libraries/libc/string/memmove.c
-            - src/libraries/libc/string/memset.c
-            - src/libraries/libc/string/strcat.c
-            - src/libraries/libc/string/strchr.c
-            - src/libraries/libc/string/strcmp.c
-            - src/libraries/libc/string/strcoll.c
-            - src/libraries/libc/string/strcpy.c
-            - src/libraries/libc/string/strcspn.c
-            - src/libraries/libc/string/strerror.c
-            - src/libraries/libc/string/strlen.c
-            - src/libraries/libc/string/strncat.c
-            - src/libraries/libc/string/strncmp.c
-            - src/libraries/libc/string/strncpy.c
-            - src/libraries/libc/string/strpbrk.c
-            - src/libraries/libc/string/strrchr.c
-            - src/libraries/libc/string/strspn.c
-            - src/libraries/libc/string/strstr.c
-            - src/libraries/libc/string/strtok.c
-            - src/libraries/libc/string/strxfrm.c
-            - src/libraries/libc/time/asctime.c
-            - src/libraries/libc/time/clock.c
-            - src/libraries/libc/time/ctime.c
-            - src/libraries/libc/time/difftime.c
-            - src/libraries/libc/time/gmtime.c
-            - src/libraries/libc/time/localtime.c
-            - src/libraries/libc/time/mktime.c
-            - src/libraries/libc/time/strftime.c
-            - src/libraries/libc/time/time.c
-            - src/libraries/libc/time/timespec_get.c
-
-    tests:
-        kind: exe
-        target: native
-        output: tests
-        deps:
-            - kutil
-        source:
-            - src/tests/constexpr_hash.cpp
-            - src/tests/linked_list.cpp
-            - src/tests/logger.cpp
-            - src/tests/heap_allocator.cpp
-            - src/tests/main.cpp
-            - src/tests/map.cpp
-            - src/tests/vector.cpp
-
-overlays:
-    - url: https://f000.backblazeb2.com/file/jsix-os/sysroot-llvm8-20190706.tar.bz2
-      path: sysroot
-    - url: https://f000.backblazeb2.com/file/jsix-os/sysroot-j6libc-8cb8ce7.tar.bz2
-      path: sysroot

other_software.md

@@ -33,42 +33,18 @@ Boost Software license:
 > ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
 > DEALINGS IN THE SOFTWARE.
 
-## cpptoml
-
-jsix uses the [cpptoml][] library for parsing TOML configuration files. cpptoml
-is released under the terms of the MIT license:
-
-[cpptoml]: https://github.com/skystrife/cpptoml
-
-> Copyright (c) 2014 Chase Geigle
->
-> Permission is hereby granted, free of charge, to any person obtaining a copy of
-> this software and associated documentation files (the "Software"), to deal in
-> the Software without restriction, including without limitation the rights to
-> use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of
-> the Software, and to permit persons to whom the Software is furnished to do so,
-> subject to the following conditions:
-> 
-> The above copyright notice and this permission notice shall be included in all
-> copies or substantial portions of the Software.
-> 
-> THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
-> IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS
-> FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
-> COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
-> IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
-> CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
-
 ## printf
 
-jsix uses Marco Paland's tiny [printf][] library for its `*printf()` functions,
+The implementation of jsix's `*printf()` family of functions is derived from
-which is also released under the terms of the MIT license:
+Marco Paland and Eyal Rozenberg's tiny [printf][] library, which is also
+released under the terms of the MIT license:
 
-[printf]: https://github.com/mpaland/printf
+[printf]: https://github.com/eyalroz/printf
 
 > The MIT License (MIT)
 >
 > Copyright (c) 2014 Marco Paland
+> Copyright (c) 2021 Eyal Rozenberg
 >
 > Permission is hereby granted, free of charge, to any person obtaining a copy
 > of this software and associated documentation files (the "Software"), to deal
@@ -87,3 +63,43 @@ which is also released under the terms of the MIT license:
 > LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 > OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 > SOFTWARE.
+
+## Zstandard
+
+The implementation of jsix's initrd contains zstd-compressed data. Zstd
+decompression is handled by the reference [zstd][] library, which is released
+under either a BSD or GPL license. I have chosen to use it under the BSD
+license terms, shown below:
+
+[zstd]: https://github.com/facebook/zstd
+
+> BSD License
+> 
+> For Zstandard software
+> 
+> Copyright (c) Meta Platforms, Inc. and affiliates. All rights reserved.
+> 
+> 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.
+>
+> * Neither the name Facebook, nor Meta, nor the names of its contributors may
+>   be used to endorse or promote products derived from this software without
+>   specific prior written permission.
+>
+> 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 HOLDER 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.

qemu.sh

@@ -1,80 +1,151 @@
 #!/usr/bin/env bash
 
-build="$(dirname $0)/build"
+root=$(dirname $0)
-assets="$(dirname $0)/assets"
+build="${root}/build"
+assets="${root}/assets"
+
+no_build=""
 debug=""
+isaexit='-device isa-debug-exit,iobase=0xf4,iosize=0x04'
 debugtarget="${build}/jsix.elf"
-flash_name="ovmf_vars"
 gfx="-nographic"
 vga="-vga none"
+log=""
 kvm=""
-cpu="Broadwell,+pdpe1gb"
+debugcon_cmd=""
+cpu_features=",+pdpe1gb,+invtsc,+hypervisor,+x2apic,+xsavec,+xsaves,+xsaveopt"
+cpu="Broadwell"
+smp=4
 
-for arg in $@; do
+while true; do
-	case "${arg}" in
+    case "$1" in
-		--debugboot)
+        -b | --debugboot)
-			debug="-s -S"
+            debug="-s -S"
-			debugtarget="${build}/boot/boot.efi"
+            debugtarget="${build}/boot/boot.efi"
-			flash_name="ovmf_vars_d"
+            shift
-			;;
+            ;;
-		--debug)
+        -d | --debug)
-			debug="-s -S"
+            debug="-s -S"
-			flash_name="ovmf_vars_d"
+            isaexit=""
-			;;
+            shift
-		--gfx)
+            ;;
-			gfx=""
+        -g | --gfx)
-			vga=""
+            gfx=""
-			;;
+            vga=""
-		--vga)
+            shift
-			vga=""
+            ;;
-			;;
+        -v | --vga)
-		--kvm)
+            vga=""
-			kvm="-enable-kvm"
+            shift
-			cpu="host"
+            ;;
-			;;
+        -r | --remote)
-		*)
+            shift
-			build="${arg}"
+            vnchost="${1:-${VNCHOST:-"localhost:5500"}}"
-			;;
+            gfx="-vnc ${vnchost},reverse"
-	esac
+            vga=""
+            shift
+            ;;
+        -k | --kvm)
+            kvm="-enable-kvm"
+            cpu="max"
+            shift
+            ;;
+        -c | --cpus)
+            smp=$2
+            shift 2
+            ;;
+        -l | --log)
+            log="-d mmu,int,guest_errors -D ${root}/jsix.log"
+            shift
+            ;;
+        -x | --debugcon)
+            debugcon_cmd="less --follow-name -R +F debugcon.log"
+            shift
+            ;;
+        --no-build)
+            no_build="yes"
+            shift
+            ;;
+        *)
+            if [ -d "$1" ]; then
+                build="$1"
+                shift
+            fi
+            break
+            ;;
+    esac
 done
 
 if [[ ! -c /dev/kvm ]]; then
-	kvm=""
+    kvm=""
 fi
 
-if ! ninja -C "${build}"; then
+[[ -z "${no_build}" ]] && if ! ninja -C "${build}"; then
-	exit 1
+    exit 1
 fi
 
 if [[ -n $TMUX ]]; then
-	if [[ -n $debug ]]; then
+    cols=$(tput cols)
-		tmux split-window -h "gdb ${debugtarget}" &
+    log_width=100
-	else
+
-		tmux split-window -h -l 80 "sleep 1; telnet localhost 45455" &
+    if [[ -n $debug ]]; then
-		tmux last-pane
+        log_cols=1
-		tmux split-window -l 10 "sleep 1; telnet localhost 45454" &
+        if [[ $debugcon_cmd ]]; then
-	fi
+            log_cols=2
+        fi
+
+        gdb_width=$(($cols - $log_cols * $log_width))
+
+        if (($gdb_width < 150)); then
+            stack=1
+            gdb_width=$(($cols - $log_width))
+            tmux split-window -h -l $gdb_width "gdb ${debugtarget}"
+            if [[ $debugcon_cmd ]]; then
+                tmux select-pane -t .left
+                tmux split-window -v "$debugcon_cmd"
+            fi
+        else
+            if [[ $debugcon_cmd ]]; then
+                tmux split-window -h -l $(($log_width + $gdb_width)) "$debugcon_cmd"
+            fi
+            tmux split-window -h -l $gdb_width "gdb ${debugtarget}"
+            tmux select-pane -t .left
+            tmux select-pane -t .right
+        fi
+    else
+        if [[ $debugcon_cmd ]]; then
+            tmux split-window -h -l $log_width "$debugcon_cmd"
+            tmux last-pane
+        fi
+        tmux split-window -l 10 "sleep 1; telnet localhost 45454"
+    fi
 elif [[ $DESKTOP_SESSION = "i3" ]]; then
-	if [[ -n $debug ]]; then
+    if [[ -n $debug ]]; then
-		i3-msg exec i3-sensible-terminal -- -e "gdb ${PWD}/${build}/jsix.elf" &
+        i3-msg exec i3-sensible-terminal -- -e "gdb ${debugtarget}" &
-	else
+    else
-		i3-msg exec i3-sensible-terminal -- -e 'telnet localhost 45454' &
+        i3-msg exec i3-sensible-terminal -- -e 'telnet localhost 45454' &
-	fi
+    fi
 fi
 
-exec qemu-system-x86_64 \
+if [[ -n "${debug}" ]]; then
-	-drive "if=pflash,format=raw,readonly,file=${assets}/ovmf/x64/ovmf_code.fd" \
+    (sleep 1; echo "Debugging ready.") &
-	-drive "if=pflash,format=raw,file=${build}/${flash_name}.fd" \
+fi
-	-drive "format=raw,file=${build}/jsix.img" \
+
-	-device "isa-debug-exit,iobase=0xf4,iosize=0x04" \
+qemu-system-x86_64 \
-	-monitor telnet:localhost:45454,server,nowait \
+    -drive "if=pflash,format=raw,readonly=on,file=${assets}/ovmf/x64/ovmf_code.fd" \
-	-serial stdio \
+    -drive "if=pflash,format=raw,file=${build}/ovmf_vars.fd" \
-	-serial telnet:localhost:45455,server,nowait \
+    -drive "format=raw,file=${build}/jsix.img" \
-	-smp 4 \
+    -chardev "file,path=${root}/debugcon.log,id=jsix_debug" \
-	-m 512 \
+    -device "isa-debugcon,iobase=0x6600,chardev=jsix_debug" \
-	-d mmu,int,guest_errors \
+    -monitor telnet:localhost:45454,server,nowait \
-	-D jsix.log \
+    -serial stdio \
-	-cpu "${cpu}" \
+    -smp "${smp}" \
-	-M q35 \
+    -m 4096 \
-	-no-reboot \
+    -cpu "${cpu}${cpu_features}" \
-	$gfx $vga $kvm $debug
+    -M q35 \
+    -no-reboot \
+    -name jsix \
+    $isaexit $log $gfx $vga $kvm $debug
+
+((result = ($? >> 1) - 1))
+exit $result

requirements.txt

@@ -0,0 +1,9 @@
+cogapp >= 3
+ninja >= 1.10.2
+peru >= 1.2.1
+pyyaml >= 5.4
+lark == 0.12.0
+pure-cdb == 4
+pyzstd == 0.15
+pyelftools
+iced-x86

scripts/bonnibel/__init__.py

@@ -0,0 +1,25 @@
+from os.path import join
+
+class BonnibelError(Exception):
+    def __init__(self, message):
+        self.message = message
+
+def load_config(filename):
+    from yaml import safe_load
+    with open(filename, 'r') as infile:
+        return safe_load(infile.read())
+
+def mod_rel(path):
+    return join("${module_dir}", path)
+
+def target_rel(path, module=None):
+    if module:
+        return join("${target_dir}", module + ".dir", path)
+    else:
+        return join("${target_dir}", path)
+
+def include_rel(path, module=None):
+    if module:
+        return join("${build_root}", "include", module, path)
+    else:
+        return join("${build_root}", "include", path)

scripts/bonnibel/config.py

@@ -0,0 +1,68 @@
+_config_cache = {}
+
+def _load(filename, depfiles):
+    from . import load_config
+    if not filename in _config_cache:
+        if filename.exists():
+            depfiles.add(filename)
+            data = load_config(filename)
+            _config_cache[filename] = data
+    return _config_cache.get(filename, dict())
+
+
+def _build_config(chain, depfiles):
+    config = {}
+    for d in [_load(c, depfiles) for c in chain]:
+        for k, v in d.items():
+            if isinstance(v, (list, tuple)):
+                config[k] = config.get(k, list()) + list(v)
+            elif isinstance(v, dict):
+                config[k] = config.get(k, dict())
+                config[k].update(v)
+            else:
+                config[k] = v
+
+    return config
+
+
+def _make_ninja_config(outfile, config, files):
+    from os import makedirs
+    from ninja.ninja_syntax import Writer
+
+    makedirs(outfile.parent, exist_ok=True)
+
+    with open(outfile, "w") as buildfile:
+        build = Writer(buildfile)
+        build.comment("This file is automatically generated by bonnibel")
+        build.comment("Source config files:")
+        for f in files:
+            build.comment(f"  - {f}")
+
+        build.newline()
+        for k, v in config.items():
+            build.variable(k, v)
+
+ 
+def generate_configs(root, output, config, targets, kinds):
+
+    assets = root / "assets" / "build"
+    base = ["global.yaml", f"config.{config}.yaml"]
+
+    depfiles = set()
+
+    for target in targets:
+        chain = base + [f"target.{target}.yaml"]
+
+        files = [assets / f for f in chain]
+        config = _build_config(files, depfiles)
+        outfile = output / target / f"config.ninja"
+        _make_ninja_config(outfile, config, files)
+
+        for kind in kinds:
+            custom = [f"kind.{kind}.yaml", f"target.{target}.{kind}.yaml"]
+            files = [assets / f for f in chain + custom]
+            config = _build_config(files, depfiles)
+            outfile = output / target / f"config.{kind}.ninja"
+            _make_ninja_config(outfile, config, files)
+
+    return depfiles

scripts/bonnibel/manifest.py

@@ -0,0 +1,132 @@
+from . import BonnibelError
+
+class Manifest:
+    from collections import namedtuple
+    Entry = namedtuple("Entry", ("module", "target", "output", "flags"))
+
+    flags = {
+        "graphical": 0x01,
+        "symbols":   0x80,
+    }
+
+    boot_flags = {
+        "debug": 0x01,
+        "test":  0x02,
+    }
+
+    def __init__(self, path, modules):
+        from . import load_config
+
+        config = load_config(path)
+
+        self.location = config.get("location", "jsix")
+
+        self.kernel = self.__build_entry(modules,
+                config.get("kernel", dict()),
+                name="kernel", target="kernel")
+
+        self.init = self.__build_entry(modules,
+                config.get("init", None), target="init")
+
+        self.panics = [self.__build_entry(modules, i, target="kernel")
+                for i in config.get("panic", tuple())]
+
+        self.services = [self.__build_entry(modules, i)
+                for i in config.get("services", tuple())]
+
+        self.drivers = [self.__build_entry(modules, i)
+                for i in config.get("drivers", tuple())]
+
+        libs = set(config.get("libs", tuple()))
+        libs.update(self.__libdeps([modules[e.module] for e in self.services]))
+        libs.update(self.__libdeps([modules[e.module] for e in self.drivers]))
+
+        self.libs = [self.__build_entry(modules, i)
+                for i in libs]
+
+        self.flags = config.get("flags", tuple())
+
+        self.symbols = ""
+
+        initrd = config.get("initrd", dict())
+        self.initrd = {
+            "name": initrd.get("name", "initrd.dat"),
+            "format": initrd.get("format", "zstd"),
+        }
+
+        self.data = []
+        for d in config.get("data", tuple()):
+            self.add_data(**d)
+
+    def __build_entry(self, modules, config, target="user", name=None):
+        flags = tuple()
+
+        if isinstance(config, str):
+            name = config
+        elif isinstance(config, dict):
+            name = config.get("name", name)
+            target = config.get("target", target)
+            flags = config.get("flags", tuple())
+            if isinstance(flags, str):
+                flags = flags.split()
+
+        mod = modules.get(name)
+        if not mod:
+            raise BonnibelError(f"Manifest specifies unknown module '{name}'")
+
+        for f in flags:
+            if not f in Manifest.flags:
+                raise BonnibelError(f"Manifest specifies unknown flag '{f}'")
+
+        return Manifest.Entry(name, target, mod.get_output(), flags)
+
+    def __libdeps(self, modules):
+        deps = set([m.name for m in modules if m.kind == "lib"])
+        for m in modules:
+            if m.static: continue
+            deps.update(self.__libdeps(m.depmods))
+        return deps
+
+    def add_data(self, output, desc, flags=tuple()):
+        e = Manifest.Entry(None, None, output, flags)
+        self.data.append(e)
+        return e
+
+    def write_boot_config(self, path):
+        from os.path import join
+        import struct
+
+        with open(path, 'wb') as outfile:
+            magic = "jsixboot".encode("utf-8") # magic string
+            version = 1
+
+            bootflags = sum([Manifest.boot_flags.get(s, 0) for s in self.flags])
+
+            outfile.write(struct.pack("<8s BBH",
+                magic,
+                version, len(self.panics), bootflags))
+
+            def write_str(s):
+                data = s.encode("utf-16le")
+                outfile.write(struct.pack("<H", len(data)+2))
+                outfile.write(data)
+                outfile.write(b"\0\0")
+
+            def write_path(name):
+                write_str(join(self.location, name).replace('/','\\'))
+
+            def write_ent(ent):
+                flags = 0
+                for f in ent.flags:
+                    flags |= Manifest.flags[f]
+
+                outfile.write(struct.pack("<H", flags))
+                write_path(ent.output)
+
+            write_ent(self.kernel)
+            write_ent(self.init)
+            write_path(self.initrd["name"])
+            write_path(self.symbols)
+
+            for p in self.panics:
+                write_ent(p)

scripts/bonnibel/module.py

@@ -0,0 +1,352 @@
+from . import include_rel, mod_rel, target_rel
+from . import BonnibelError
+
+def resolve(path):
+    if path.startswith('/') or path.startswith('$'):
+        return path
+    from pathlib import Path
+    return str(Path(path).resolve())
+
+class BuildOptions:
+    def __init__(self, includes=tuple(), local=tuple(), late=tuple(), libs=tuple(), order_only=tuple(), ld_script=None):
+        self.includes = list(includes)
+        self.local = list(local)
+        self.late = list(late)
+        self.libs = list(libs)
+        self.order_only = list(order_only)
+        self.ld_script = ld_script and str(ld_script)
+
+    @property
+    def implicit(self):
+        libfiles = list(map(lambda x: x[0], self.libs))
+        if self.ld_script is not None:
+            return libfiles + [self.ld_script]
+        else:
+            return libfiles
+
+    @property
+    def linker_args(self):
+        from pathlib import Path
+        def arg(path, static):
+            if static: return path
+            return "-l:" + Path(path).name
+        return [arg(*l) for l in self.libs]
+
+
+class Module:
+    __fields = {
+        # name: (type, default)
+        "kind": (str, "exe"),
+        "outfile": (str, None),
+        "basename": (str, None),
+        "targets": (set, ()),
+        "deps":  (set, ()),
+        "public_headers":  (set, ()),
+        "copy_headers": (bool, False),
+        "includes": (tuple, ()),
+        "include_phase": (str, "normal"),
+        "sources": (tuple, ()),
+        "drivers": (tuple, ()),
+        "variables": (dict, ()),
+        "default": (bool, False),
+        "description": (str, None),
+        "no_libc": (bool, False),
+        "ld_script": (str, None),
+        "static": (bool, False),
+    }
+
+    def __init__(self, name, modfile, root, **kwargs):
+        from pathlib import Path
+        from .source import make_source, make_copy_source
+
+        # Required fields
+        self.root = Path(root)
+        self.name = name
+        self.modfile = modfile
+
+        for name, data in self.__fields.items():
+            ctor, default = data
+            value = kwargs.get(name, default)
+            if value is not None:
+                value = ctor(value)
+
+            setattr(self, name, value)
+
+        for name in kwargs:
+            if not name in self.__fields:
+                raise AttributeError(f"No attribute named {name} on Module ({modfile})")
+
+        if not self.no_libc:
+            self.deps.add("libc_free")
+
+        # Turn strings into real Source objects
+        self.sources = [make_source(root, f) for f in self.sources]
+
+        header_source = lambda f: make_source(root, Path("include") / f)
+        if self.copy_headers:
+            header_source = lambda f: make_copy_source(root, f, "include")
+        self.public_headers = [header_source(f) for f in self.public_headers]
+
+        # Filled by Module.update
+        self.depmods = set()
+
+    def __repr__(self):
+        return f"<Module {self.kind} {self.name}>"
+
+    @property
+    def basename(self):
+        if self.__basename is not None:
+            return self.__basename
+        if self.kind == "lib":
+            return f"lib{self.name}"
+        else:
+            return self.name
+
+    @basename.setter
+    def basename(self, value):
+        self.__basename = value
+
+    def get_output(self, static=False):
+        if self.outfile is not None:
+            return self.outfile
+        elif self.kind == "headers":
+            return None
+        else:
+            ext = dict(exe=".elf", driver=".drv", lib=(static and ".a" or ".so"))
+            return self.basename + ext.get(self.kind, "")
+
+    @classmethod
+    def update(cls, mods):
+        from . import BonnibelError
+
+        def resolve(source, modlist):
+            resolved = set()
+            for dep in modlist:
+                if not dep in mods:
+                    raise BonnibelError(f"module '{source.name}' references unknown module '{dep}'")
+                mod = mods[dep]
+                resolved.add(mod)
+            return resolved
+
+        for mod in mods.values():
+            mod.depmods = resolve(mod, mod.deps)
+
+        target_mods = [mod for mod in mods.values() if mod.targets]
+        for mod in target_mods:
+            closed = set()
+            children = set(mod.depmods)
+            while children:
+                child = children.pop()
+                closed.add(child)
+                child.targets |= mod.targets
+                children |= {m for m in child.depmods if not m in closed}
+
+    def generate(self, output):
+        from pathlib import Path
+        from collections import defaultdict
+        from ninja.ninja_syntax import Writer
+
+        def walk_deps(deps, static, results):
+            for mod in deps:
+                if static or mod.name not in results:
+                    results[mod.name] = (mod, static)
+                walk_deps(mod.depmods, static or mod.static, results)
+
+        all_deps = {}
+        walk_deps(self.depmods, self.static, all_deps)
+        all_deps = all_deps.values()
+
+        def gather_phony(build, deps, child_rel):
+            phony = ".headers.phony"
+            child_phony = [child_rel(phony, module=c.name)
+                    for c, _ in all_deps]
+
+            build.build(
+                rule = "touch",
+                outputs = [mod_rel(phony)],
+                implicit = child_phony,
+                order_only = list(map(mod_rel, deps)),
+            )
+
+        filename = str(output / f"module.{self.name}.ninja")
+        with open(filename, "w") as buildfile:
+            build = Writer(buildfile)
+
+            build.comment("This file is automatically generated by bonnibel")
+            build.newline()
+
+            build.variable("module_dir", target_rel(self.name + ".dir"))
+            build.variable("module_kind", self.kind)
+            build.newline()
+
+            build.include(f"${{target_dir}}/config.{self.kind}.ninja")
+            build.newline()
+
+            modopts = BuildOptions(
+                local = [self.root, "${module_dir}"],
+                ld_script = self.ld_script and self.root / self.ld_script,
+                )
+            if self.public_headers:
+                modopts.includes += [
+                    self.root / "include",
+                    f"${{target_dir}}/{self.name}.dir/include",
+                ]
+
+            for key, value in self.variables.items():
+                build.variable(key, value)
+            build.newline()
+
+            for include in self.includes:
+                p = Path(include)
+                if p.is_absolute():
+                    if not p in modopts.includes:
+                        modopts.includes.append(str(p.resolve()))
+                elif include != ".":
+                    incpath = self.root / p
+                    destpath = mod_rel(p)
+                    for header in incpath.rglob("*.h"):
+                        dest_header = f"{destpath}/" + str(header.relative_to(incpath))
+                    modopts.includes.append(str(incpath))
+                    modopts.includes.append(destpath)
+
+            for dep, static in all_deps:
+                if dep.public_headers:
+                    if dep.include_phase == "normal":
+                        modopts.includes += [dep.root / "include", f"${{target_dir}}/{dep.name}.dir/include"]
+                    elif dep.include_phase == "late":
+                        modopts.late += [dep.root / "include", f"${{target_dir}}/{dep.name}.dir/include"]
+                    else:
+                        from . import BonnibelError
+                        raise BonnibelError(f"Module {dep.name} has invalid include_phase={dep.include_phase}")
+
+                if dep.kind == "headers":
+                    continue
+                elif dep.kind == "lib":
+                    modopts.libs.append((target_rel(dep.get_output(static)), static))
+                else:
+                    modopts.order_only.append(target_rel(dep.get_output(static)))
+
+            cc_includes = []
+            if modopts.local:
+                cc_includes += [f"-iquote{i}" for i in modopts.local]
+
+            if modopts.includes:
+                cc_includes += [f"-I{i}" for i in modopts.includes]
+
+            if modopts.late:
+                cc_includes += [f"-idirafter{i}" for i in modopts.late]
+
+            if cc_includes:
+                build.variable("ccflags", ["${ccflags}"] + cc_includes)
+
+            as_includes = [f"-I{d}" for d in modopts.local + modopts.includes + modopts.late]
+            if as_includes:
+                build.variable("asflags", ["${asflags}"] + as_includes)
+
+            if modopts.libs:
+                build.variable("libs", ["-L${target_dir}", "${libs}"] + modopts.linker_args)
+
+            if modopts.ld_script:
+                build.variable("ldflags", ["${ldflags}"] + ["-T", modopts.ld_script])
+
+            header_deps = []
+            inputs = []
+            headers = set(self.public_headers)
+            while headers:
+                source = headers.pop()
+                headers.update(source.next)
+
+                if source.action:
+                    build.newline()
+                    build.build(rule=source.action, **source.args)
+
+                if source.gather:
+                    header_deps += list(source.outputs)
+
+                if source.input:
+                    inputs.extend(map(mod_rel, source.outputs))
+
+            build.newline()
+
+            inputs = []
+            sources = set(self.sources)
+            while sources:
+                source = sources.pop()
+                sources.update(source.next)
+
+                if source.action:
+                    build.newline()
+                    build.build(rule=source.action, **source.args)
+
+                if source.gather:
+                    header_deps += list(source.outputs)
+
+                if source.input:
+                    inputs.extend(map(mod_rel, source.outputs))
+
+            gather_phony(build, header_deps, target_rel)
+
+            if self.kind == "headers":
+                # Header-only, don't output a build rule
+                return
+
+            output = target_rel(self.get_output())
+            build.newline()
+            build.build(
+                rule = self.kind,
+                outputs = output,
+                inputs = inputs,
+                implicit = modopts.implicit,
+                order_only = modopts.order_only,
+                variables = {"name": self.name,
+                             "soname": self.get_output()},
+            )
+
+            dump = output + ".dump"
+            build.newline()
+            build.build(
+                rule = "dump",
+                outputs = dump,
+                inputs = output,
+                variables = {"name": self.name},
+            )
+
+            s_output = target_rel(self.get_output(static=True))
+            if s_output != output:
+                build.newline()
+                build.build(
+                    rule = self.kind + "_static",
+                    outputs = s_output,
+                    inputs = inputs,
+                    order_only = modopts.order_only,
+                    variables = {"name": self.name},
+                )
+
+                dump = s_output + ".dump"
+                build.newline()
+                build.build(
+                    rule = "dump",
+                    outputs = dump,
+                    inputs = s_output,
+                    variables = {"name": self.name},
+                )
+
+            if self.default:
+                build.newline()
+                build.default(output)
+                build.default(dump)
+
+    def add_input(self, path, **kwargs):
+        from .source import make_source
+        s = make_source(self.root, path, **kwargs)
+        self.sources.append(s)
+        return s.outputs
+
+    def add_depends(self, paths, deps):
+        for source in self.sources:
+            if source.path in paths:
+                source.add_deps(deps)
+
+        for source in self.public_headers:
+            if source.path in paths:
+                source.add_deps(deps)

scripts/bonnibel/project.py

@@ -0,0 +1,293 @@
+from . import BonnibelError
+
+class Project:
+    def __init__(self, root):
+        from .version import git_version
+
+        self.root = root
+        self.version = git_version(root)
+
+    def __str__(self):
+        return f"{self.name} {self.version.major}.{self.version.minor}.{self.version.patch}-{self.version.sha}"
+
+    def generate(self, root, output, modules, config, manifest_file):
+        import sys
+        from os.path import join
+        from ninja.ninja_syntax import Writer
+
+        targets = set()
+        kinds = set()
+        for mod in modules.values():
+            targets.update(mod.targets)
+            kinds.add(mod.kind)
+
+        from .config import generate_configs
+        config_deps = generate_configs(root, output, config, targets, kinds)
+
+        with open(output / "build.ninja", "w") as buildfile:
+            build = Writer(buildfile)
+
+            build.comment("This file is automatically generated by bonnibel")
+            build.variable("ninja_required_version", "1.3")
+            build.variable("build_root", output)
+            build.variable("source_root", root)
+            build.variable("build_config", config)
+            build.newline()
+
+            build.include(root / "assets/build/rules.ninja")
+            build.newline()
+
+            build.variable("version_major", self.version.major)
+            build.variable("version_minor", self.version.minor)
+            build.variable("version_patch", self.version.patch)
+            build.variable("version_sha", self.version.sha)
+            build.newline()
+
+            build.variable("cogflags", [
+                "-I", "${source_root}/scripts",
+                "-D", "definitions_path=${source_root}/definitions",
+                ])
+            build.newline()
+
+            for target in targets:
+                build.subninja(output / target / "target.ninja")
+            build.newline()
+
+            build.build(
+                rule = "touch",
+                outputs = "${build_root}/.all_headers",
+                implicit = [f"${{build_root}}/include/{m.name}/.headers.phony"
+                    for m in modules.values() if m.public_headers],
+                )
+            build.build(
+                rule = "phony",
+                outputs = ["all-headers"],
+                inputs = ["${build_root}/.all_headers"])
+
+            from .manifest import Manifest
+            manifest = Manifest(manifest_file, modules)
+
+            debugroot = output / ".debug"
+            debugroot.mkdir(exist_ok=True)
+
+            fatroot = output / "fatroot"
+            fatroot.mkdir(exist_ok=True)
+
+            (fatroot / manifest.location).mkdir(exist_ok=True)
+
+            initrdroot = output / "initrd_root"
+            initrdroot.mkdir(exist_ok=True)
+
+            fatroot_content = []
+            initrd_content = []
+
+            def add_fatroot(source, name):
+                output = join(manifest.location, name)
+                fatroot_output = f"${{build_root}}/fatroot/{output}"
+
+                build.build(
+                    rule = "cp",
+                    outputs = [fatroot_output],
+                    inputs = [source],
+                    variables = {
+                        "description": f"Installing {output}",
+                    })
+
+                fatroot_content.append(fatroot_output)
+                build.newline()
+
+            def add_fatroot_exe(entry):
+                input_path = f"${{build_root}}/{entry.target}/{entry.output}"
+                intermediary = f"${{build_root}}/{entry.output}"
+
+                build.build(
+                    rule = "strip",
+                    outputs = [intermediary],
+                    inputs = [input_path],
+                    implicit = [f"{input_path}.dump"],
+                    variables = {
+                        "name": f"Stripping {entry.module}",
+                        "debug": f"${{build_root}}/.debug/{entry.output}.debug",
+                    })
+
+                add_fatroot(intermediary, entry.output)
+
+            def add_initrd_content(root, name):
+                output = join(root, name)
+                initrd_output = f"${{build_root}}/initrd_root/{output}"
+
+                build.build(
+                    rule = "cp",
+                    outputs = [initrd_output],
+                    inputs = [f"${{build_root}}/{name}"],
+                    variables = {
+                        "description": f"Installing {name}",
+                    })
+
+                initrd_content.append(initrd_output)
+                build.newline()
+
+            def add_initrd_stripped(root, entry):
+                input_path = f"${{build_root}}/{entry.target}/{entry.output}"
+                intermediary = f"${{build_root}}/{entry.output}"
+
+                build.build(
+                    rule = "strip",
+                    outputs = [intermediary],
+                    inputs = [input_path],
+                    implicit = [f"{input_path}.dump"],
+                    variables = {
+                        "name": f"Stripping {entry.module}",
+                        "debug": f"${{build_root}}/.debug/{entry.output}.debug",
+                    })
+
+                add_initrd_content(root, entry.output)
+
+            add_fatroot_exe(manifest.kernel)
+            add_fatroot_exe(manifest.init)
+            for program in manifest.panics:
+                add_fatroot_exe(program)
+
+            for program in manifest.services:
+                add_initrd_stripped("jsix/services", program)
+
+            for program in manifest.drivers:
+                add_initrd_stripped("jsix/drivers", program)
+
+            for program in manifest.libs:
+                add_initrd_stripped("jsix/lib", program)
+
+            syms = manifest.add_data("symbol_table.dat",
+                    "Symbol table", ("symbols",))
+
+            syms_file = "jsix.symbols"
+            syms_path = join(manifest.location, syms_file );
+            syms_out = join(fatroot, syms_path)
+            build.build(
+                rule = "makest",
+                outputs = [syms_out],
+                inputs = [f"${{build_root}}/kernel/{modules['kernel'].get_output(static=True)}"],
+                )
+            fatroot_content.append(syms_out)
+            manifest.symbols = syms_file
+
+            bootloader = "${build_root}/fatroot/efi/boot/bootx64.efi"
+            build.build(
+                rule = "cp",
+                outputs = [bootloader],
+                inputs = ["${build_root}/boot/boot.efi"],
+                variables = {
+                    "description": "Installing bootloader",
+                })
+            build.newline()
+
+            boot_config = join(fatroot, "jsix", "boot.conf")
+            manifest.write_boot_config(boot_config)
+
+            initrd = str(fatroot / manifest.location / manifest.initrd["name"])
+            build.build(
+                rule = "makeinitrd",
+                outputs = [initrd],
+                inputs = [str(initrdroot)],
+                implicit = initrd_content + ["${source_root}/scripts/mkj6romfs.py"],
+                variables = {"format": manifest.initrd["format"]},
+            )
+            build.newline()
+
+            fatroot_content.append(initrd)
+
+            build.build(
+                rule = "makefat",
+                outputs = ["${build_root}/jsix.img"],
+                inputs = ["${source_root}/assets/diskbase.img"],
+                implicit = fatroot_content + [bootloader],
+                variables = {"name": "jsix.img"},
+                )
+            build.newline()
+
+            default_assets = {
+                "UEFI Variables": ("ovmf/x64/ovmf_vars.fd", "ovmf_vars.fd"),
+                "GDB Debug Helpers": ("debugging/jsix.elf-gdb.py", "jsix.elf-gdb.py"),
+            }
+
+            for name, assets in default_assets.items():
+                p = root / "assets" / assets[0]
+                out = f"${{build_root}}/{assets[1]}"
+                build.build(
+                    rule = "cp",
+                    outputs = [out],
+                    inputs = [str(p)],
+                    variables = {"name": name},
+                    )
+                build.default([out])
+                build.newline()
+
+            compdb = "${source_root}/compile_commands.json"
+
+            build.rule("regen",
+                command = " ".join([str(root / 'configure')] + sys.argv[1:]),
+                description = "Regenerate build files",
+                generator = True,
+                )
+            build.newline()
+
+            regen_implicits = \
+                [f"{self.root}/configure", str(manifest_file)] + \
+                [str(mod.modfile) for mod in modules.values()]
+
+            regen_implicits += list(map(str, config_deps))
+
+            build.build(
+                rule = "compdb",
+                outputs = [compdb],
+                implicit = regen_implicits,
+                )
+            build.default([compdb])
+            build.newline()
+
+            build.build(
+                rule = "regen",
+                outputs = ['build.ninja'],
+                implicit = regen_implicits,
+                implicit_outputs = 
+                    [f"module.{mod.name}.ninja" for mod in modules.values()] +
+                    [f"{target}/target.ninja" for target in targets] +
+                    [boot_config],
+                )
+
+            build.newline()
+            build.default(["${build_root}/jsix.img"])
+
+        for target in targets:
+            mods = [m.name for m in modules.values() if target in m.targets]
+
+            targetdir = output / target
+            targetdir.mkdir(exist_ok=True)
+
+            buildfilename = str(targetdir / "target.ninja")
+            with open(buildfilename, "w") as buildfile:
+                build = Writer(buildfile)
+                build.comment("This file is automatically generated by bonnibel")
+                build.newline()
+
+                build.variable("target", target)
+                build.variable("target_dir", output / target)
+                build.newline()
+
+                build.include(f"{target}/config.ninja")
+
+                build.newline()
+                for kind in ('defs', 'run'):
+                    for lang in ('c', 'cpp'):
+                        deffile = str(output / target / f"{lang}.{kind}")
+
+                        build.build(
+                            rule = f"dump_{lang}_{kind}",
+                            outputs = [deffile],
+                            implicit = [buildfilename],
+                        )
+                        build.default(deffile)
+                        build.newline()
+
+                for mod in mods:
+                    build.subninja(f"module.{mod}.ninja")

scripts/bonnibel/source.py

@@ -0,0 +1,121 @@
+from os.path import join, splitext
+from . import mod_rel
+
+def _resolve(path):
+    if path.startswith('/') or path.startswith('$'):
+        return path
+    from pathlib import Path
+    return str(Path(path).resolve())
+
+def _dynamic_action(name):
+    def prop(self):
+        root, suffix = splitext(self.path)
+        return f"{name}{suffix}"
+    return prop
+
+
+class Source:
+    next = tuple()
+    action = None
+    args = dict()
+    gather = False
+    outputs = tuple()
+    input = False
+
+    def __init__(self, path, root = "${module_dir}", deps=tuple()): 
+        self.path = path
+        self.root = root
+        self.deps = deps
+
+    def add_deps(self, deps):
+        self.deps += tuple(deps)
+
+    @property
+    def fullpath(self):
+        return join(self.root, self.path)
+
+class ParseSource(Source):
+    action = property(_dynamic_action("parse"))
+
+    @property
+    def output(self):
+        root, _ = splitext(self.path)
+        return root
+
+    @property
+    def outputs(self):
+        return (self.output,)
+
+    @property
+    def gather(self):
+        _, suffix = splitext(self.output)
+        return suffix in (".h", ".hh", ".inc")
+
+    @property
+    def next(self):
+        _, suffix = splitext(self.output)
+        nextType = {
+            ".s": CompileSource,
+            ".cpp": CompileSource,
+            }.get(suffix)
+
+        if nextType:
+            return (nextType(self.output),)
+        return tuple()
+
+    @property
+    def args(self):
+        return dict(
+            outputs = list(map(mod_rel, self.outputs)),
+            inputs = [self.fullpath],
+            implicit = list(map(_resolve, self.deps)),
+            variables = dict(name=self.path),
+            )
+
+class CopySource(ParseSource):
+    action = "cp"
+
+    def __init__(self, path, root = "${module_dir}", deps=tuple(), prefix = ""): 
+        self.path = path
+        self.root = root
+        self.deps = deps
+        self.prefix = prefix
+
+    @property
+    def output(self):
+        from pathlib import Path
+        return Path(self.prefix) / self.path
+
+class HeaderSource(Source): pass
+
+class CompileSource(Source):
+    action = property(_dynamic_action("compile"))
+    input = True
+
+    @property
+    def outputs(self):
+        return (self.path + ".o",)
+
+    @property
+    def args(self):
+        return dict(
+            outputs = list(map(mod_rel, self.outputs)),
+            inputs = [join(self.root, self.path)],
+            implicit = list(map(_resolve, self.deps)) + [mod_rel(".headers.phony")],
+            variables = dict(name=self.path),
+            )
+
+def make_source(root, path):
+    _, suffix = splitext(path)
+
+    if suffix in (".s", ".c", ".cpp"):
+        return CompileSource(path, root)
+    elif suffix in (".cog",):
+        return ParseSource(path, root)
+    elif suffix in (".h", ".hh", ".inc"):
+        return HeaderSource(path, root)
+    else:
+        raise RuntimeError(f"{path} has no Source type")
+
+def make_copy_source(root, path, prefix = ""):
+    return CopySource(path, root, prefix=prefix)

scripts/bonnibel/version.py

@@ -0,0 +1,41 @@
+from collections import namedtuple as _namedtuple
+
+version = _namedtuple('version', [
+    'major',
+    'minor',
+    'patch',
+    'sha',
+    'dirty',
+    ])
+
+
+def _run_git(root, *args):
+    from subprocess import run
+
+    git = run(['git', '-C', root] + list(args),
+            check=True, capture_output=True)
+    return git.stdout.decode('utf-8').strip()
+
+
+def git_version(root):
+    full_version = _run_git(root, 'describe', '--dirty')
+    full_sha = _run_git(root, 'rev-parse', 'HEAD')
+
+    dirty = False
+    parts1 = full_version.split('-')
+    if parts1[-1] == "dirty":
+        dirty = True
+        parts1 = parts1[:-1]
+
+    if parts1[0][0] == 'v':
+        parts1[0] = parts1[0][1:]
+
+    parts2 = parts1[0].split('.')
+
+    return version(
+        parts2[0],
+        parts2[1],
+        parts2[2],
+        full_sha[:7],
+        dirty)
+

scripts/build_symbol_table.py

@@ -4,37 +4,43 @@
 # is as follows:
 #
 # <num_entires> : 8 bytes
-# <index>       : 16 * N bytes
+# <index>       : 24 * N bytes
 # <name data>   : variable
 #
 # Each index entry has the format
 # <symbol address> : 8 bytes
+# <symbol size>    : 8 bytes
 # <offset of name> : 8 bytes
 #
 # Name offsets are from the start of the symbol table as a whole. (ie,
 # where <num_entries> is located.)
 
+import re
+
+sym_re = re.compile(r'([0-9a-fA-F]{16}) ([0-9a-fA-F]{16} )?[tTvVwW] (.*)')
+
 def parse_syms(infile):
     """Take the output of the `nm` command, and parse it into a tuple
     representing the symbols in the text segment of the binary. Returns
     a list of (address, symbol_name)."""
 
-    from cxxfilt import demangle, InvalidName
-
     syms = []
     for line in sys.stdin:
-        addr, t, mangled = line.split()
+        match = sym_re.match(line)
-        if t not in "tTvVwW": continue
+        if not match: continue
 
-        try:
+        addr = int(match.group(1), base=16)
-            name = demangle(mangled)
+        size = int(match.group(2) or "0", base=16)
-        except InvalidName:
+        name = match.group(3)
+
+        if size == 0:
+            if not "." in name:
+                print(f"SYMBOL WARNING: zero size for symbol {name}")
             continue
 
-        addr = int(addr, base=16)
+        syms.append([addr, size, name, 0])
-        syms.append((addr, name))
 
-    return sorted(syms)
+    return syms
 
 
 def write_table(syms, outfile):
@@ -46,30 +52,35 @@ def write_table(syms, outfile):
     outfile.write(struct.pack("@Q", len(syms)))
     index_pos = outfile.tell()
 
-    outfile.seek(struct.calcsize("@QQ") * len(syms), 1)
+    outfile.seek(struct.calcsize("@QQQ") * len(syms), 1)
     nul = b'\0'
 
-    positions = {}
     for s in syms:
-        addr, name = s
+        s[3] = outfile.tell()
-        positions[addr] = outfile.tell()
+        outfile.write(s[2].encode('utf-8'))
-
-        data = name.encode('utf-8')
-        outfile.write(name.encode('utf-8'))
         outfile.write(nul)
 
     outfile.seek(index_pos)
     for s in syms:
         addr = s[0]
-        pos = positions[addr]
+        size = s[1]
-        outfile.write(struct.pack("@QQ", addr, pos))
+        pos = s[3]
+        outfile.write(struct.pack("@QQQ", addr, size, pos))
+
+    return len(syms)
 
 
 if __name__ == "__main__":
     import sys
     if len(sys.argv) != 2:
-        print(f"Usage: {sys.argv[0]} <output>")
+        print(f"Usage: nm -n -S --demangle | {sys.argv[0]} <output>")
         sys.exit(1)
 
-    outfile = open(sys.argv[1], "wb")
+    syms = 0
-    write_table(parse_syms(sys.stdin), outfile)
+    size = 0
+    with open(sys.argv[1], "wb") as outfile:
+        syms = write_table(parse_syms(sys.stdin), outfile)
+        outfile.seek(0, 2)
+        size = outfile.tell()
+
+    print(f"Wrote {syms} symbols ({size/1024:.1f} KiB) to {sys.argv[1]}.")

scripts/build_sysroot.sh

@@ -1,131 +0,0 @@
-#!/usr/bin/env bash
-
-TARGET="x86_64-elf"
-LLVM_BRANCH="release_80"
-
-TOOLS="clang lld" # lld libunwind libcxxabi libcxx"
-PROJECTS="compiler-rt libcxxabi libcxx libunwind"
-#RUNTIMES="compiler-rt libcxxabi libcxx libunwind"
-
-set -e
-
-README=$(realpath "$(dirname $0)/readme_for_prebuilt_sysroots.md")
-SYSROOT=$(realpath "$(dirname $0)/../sysroot")
-WORK=$(realpath "$(dirname $0)/sysroot")
-mkdir -p "${SYSROOT}"
-mkdir -p "${WORK}"
-
-export CC=clang
-export CXX=clang++
-
-if [[ ! -d "${WORK}/llvm" ]]; then
-	echo "Downloading LLVM..."
-	git clone -q \
-		--branch "${LLVM_BRANCH}" \
-		--depth 1 \
-		"https://git.llvm.org/git/llvm.git" "${WORK}/llvm"
-fi
-
-for tool in ${TOOLS}; do
-	if [[ ! -d "${WORK}/llvm/tools/${tool}" ]]; then
-		echo "Downloading ${tool}..."
-		git clone -q \
-			--branch "${LLVM_BRANCH}" \
-			--depth 1 \
-			"https://git.llvm.org/git/${tool}.git" "${WORK}/llvm/tools/${tool}"
-	fi
-done
-
-if [[ ! -d "${WORK}/llvm/tools/clang/tools/extra" ]]; then
-	echo "Downloading clang-tools-extra..."
-	git clone -q \
-		--branch "${LLVM_BRANCH}" \
-		--depth 1 \
-		"https://git.llvm.org/git/clang-tools-extra.git" "${WORK}/llvm/tools/clang/tools/extra"
-fi
-
-for proj in ${PROJECTS}; do
-	if [[ ! -d "${WORK}/llvm/projects/${proj}" ]]; then
-		echo "Downloading ${proj}..."
-		git clone -q \
-			--branch "${LLVM_BRANCH}" \
-			--depth 1 \
-			"https://git.llvm.org/git/${proj}.git" "${WORK}/llvm/projects/${proj}"
-	fi
-done
-
-for proj in ${RUNTIMES}; do
-	if [[ ! -d "${WORK}/llvm/runtimes/${proj}" ]]; then
-		echo "Downloading ${proj}..."
-		git clone -q \
-			--branch "${LLVM_BRANCH}" \
-			--depth 1 \
-			"https://git.llvm.org/git/${proj}.git" "${WORK}/llvm/runtime/${proj}"
-	fi
-done
-
-mkdir -p "${WORK}/build/llvm"
-pushd "${WORK}/build/llvm"
-
-echo "Configuring LLVM..."
-
-cmake -G Ninja \
-	-DCLANG_DEFAULT_RTLIB=compiler-rt \
-	-DCLANG_DEFAULT_STD_C=c11 \
-	-DCLANG_DEFAULT_STD_CXX=cxx14 \
-	-DCMAKE_BUILD_TYPE=Release \
-	-DCMAKE_C_COMPILER="clang" \
-	-DCMAKE_CXX_COMPILER="clang++" \
-	-DCMAKE_CXX_FLAGS="-Wno-unused-parameter -D_LIBCPP_HAS_NO_ALIGNED_ALLOCATION -D_LIBUNWIND_IS_BAREMETAL=1 -U_LIBUNWIND_SUPPORT_DWARF_UNWIND" \
-	-DCMAKE_INSTALL_PREFIX="${SYSROOT}" \
-	-DCMAKE_MAKE_PROGRAM=`which ninja` \
-	-DDEFAULT_SYSROOT="${SYSROOT}" \
-	-DLIBCXX_CXX_ABI=libcxxabi \
-	-DLIBCXX_CXX_ABI_INCLUDE_PATHS="${WORK}/llvm/projects/libcxxabi/include" \
-	-DLIBCXX_CXX_ABI_LIBRARY_PATH=lib \
-	-DLIBCXX_ENABLE_EXPERIMENTAL_LIBRARY=OFF \
-	-DLIBCXX_ENABLE_NEW_DELETE_DEFINITIONS=ON \
-	-DLIBCXX_ENABLE_SHARED=OFF \
-	-DLIBCXX_ENABLE_STATIC_ABI_LIBRARY=ON \
-	-DLIBCXX_ENABLE_THREADS=OFF \
-	-DLIBCXX_INCLUDE_BENCHMARKS=OFF \
-	-DLIBCXX_USE_COMPILER_RT=ON \
-	-DLIBCXXABI_ENABLE_NEW_DELETE_DEFINITIONS=OFF \
-	-DLIBCXXABI_ENABLE_SHARED=OFF \
-	-DLIBCXXABI_ENABLE_STATIC_UNWINDER=ON \
-	-DLIBCXXABI_ENABLE_THREADS=OFF \
-	-DLIBCXXABI_LIBCXX_PATH="${WORK}/llvm/projects/libcxx" \
-	-DLIBCXXABI_USE_COMPILER_RT=ON \
-	-DLIBCXXABI_USE_LLVM_UNWINDER=ON \
-	-DLIBUNWIND_ENABLE_SHARED=OFF \
-	-DLIBUNWIND_ENABLE_THREADS=OFF \
-	-DLIBUNWIND_USE_COMPILER_RT=ON \
-	-DLLVM_CONFIG_PATH="${SYSROOT}/bin/llvm-config" \
-	-DLLVM_DEFAULT_TARGET_TRIPLE="x86_64-unknown-elf" \
-	-DLLVM_ENABLE_LIBCXX=ON \
-	-DLLVM_ENABLE_LLD=ON \
-	-DLLVM_ENABLE_PIC=OFF \
-	-DLLVM_ENABLE_THREADS=OFF \
-	-DLLVM_INSTALL_BINUTILS_SYMLINKS=ON \
-	-DLLVM_TARGETS_TO_BUILD="X86" \
-	${WORK}/llvm > cmake_configure.log
-
-#  -DCMAKE_ASM_COMPILER=nasm \
-#  -DCMAKE_LINKER="${SYSROOT}/bin/ld.lld" \
-#  -DCOMPILER_RT_ENABLE_LLD=ON \
-#  -DLIBCXX_ENABLE_LLD=ON \
-#  -DLIBCXX_ENABLE_STATIC_UNWINDER=ON \
-#  -DLIBCXXABI_ENABLE_LLD=ON \
-#  -DLIBUNWIND_ENABLE_LLD=ON \
-#  -DLLVM_ENABLE_PROJECTS="libcxx;libcxxabi;libunwind;compiler-rt" \
-#  -DCOMPILER_RT_BAREMETAL_BUILD=ON \
-#  -DLIBCXXABI_BAREMETAL=ON \
-
-echo "Building LLVM..."
-ninja && ninja install
-ninja cxx cxxabi compiler-rt
-ninja install-compiler-rt install-cxx install-cxxabi
-popd
-
-
-cp "${README}" "${SYSROOT}/README.md"

scripts/debug_buddy_alloc.gdb

@@ -0,0 +1,62 @@
+das -enabled off
+
+break heap_allocator.cpp:200
+commands
+silent
+printf "n %016lx %d\n", m_end, current
+continue
+end
+
+break heap_allocator.cpp:206
+commands
+silent
+printf "N %016lx %d\n", m_end, order
+continue
+end
+
+break heap_allocator::register_free_block
+commands
+silent
+printf "F %016lx %d\n", block, order
+continue
+end
+
+break heap_allocator.cpp:118
+commands
+silent
+printf "f %016lx %d\n", block, info->order
+continue
+end
+
+break heap_allocator.cpp:241
+commands
+silent
+printf "S %016lx %d\n", block, order
+continue
+end
+
+break heap_allocator.cpp:158
+commands
+silent
+printf "P %016lx %d\n", block, order
+continue
+end
+
+break heap_allocator.cpp:180
+commands
+silent
+printf "p %016lx %d\n", block, order
+continue
+end
+
+break heap_allocator.cpp:182
+commands
+silent
+printf "M %016lx %016lx %d\n", block, buddy, order
+continue
+end
+
+set logging file buddy_allocs.txt
+set logging overwrite on
+set logging enabled on
+continue

scripts/debug_frame_alloc.gdb

@@ -0,0 +1,18 @@
+das -enabled off
+break frame_allocator.cpp:62
+commands
+silent
+printf "+ %016lx %3d\n", *address, n
+continue
+end
+break frame_allocator.cpp:95
+commands
+silent
+printf "- %016lx %3d\n", address, count
+continue
+end
+
+set logging file frame_allocs.txt
+set logging overwrite on
+set logging enabled on
+continue

scripts/debug_heap_alloc.gdb

@@ -0,0 +1,29 @@
+das -enabled off
+break heap_allocator.cpp:81
+commands
+silent
+printf "+ %016lx %4d\n", block, length
+continue
+end
+break heap_allocator.cpp:86
+commands
+silent
+printf "+ %016lx %4d\n", block, length
+continue
+end
+break heap_allocator.cpp:120
+commands
+silent
+printf "- %016lx\n", p
+continue
+end
+break heap_allocator.cpp:140
+commands
+silent
+printf "> %016lx %4d %4d\n", p, old_length, new_length
+continue
+end
+set logging file heap_allocs.txt
+set logging overwrite on
+set logging enabled on
+continue

scripts/definitions/context.py

@@ -0,0 +1,83 @@
+class NotFound(Exception): pass
+
+class Context:
+    def __init__(self, path, verbose=False):
+        if isinstance(path, str):
+            self.__paths = [path]
+        else:
+            self.__paths = path
+
+        self.__closed = set()
+        self.__verbose = verbose
+
+        self.__deps = {}
+
+        self.objects = dict()
+        self.interfaces = dict()
+
+    verbose = property(lambda self: self.__verbose)
+
+    def find(self, filename):
+        from os.path import exists, isabs, join
+
+        if exists(filename) or isabs(filename):
+            return filename
+
+        for path in self.__paths:
+            full = join(path, filename)
+            if exists(full):
+                return full
+
+        raise NotFound(filename)
+
+    def parse(self, filename):
+        pending = set()
+        pending.add(filename)
+
+        from .parser import LarkError
+        from .parser import Lark_StandAlone as Parser
+        from .transformer import DefTransformer
+
+        objects = {}
+        interfaces = {}
+
+        while pending:
+            name = pending.pop()
+            self.__closed.add(name)
+
+            path = self.find(name)
+
+            parser = Parser(transformer=DefTransformer(name))
+
+            try:
+                imps, objs, ints = parser.parse(open(path, "r").read())
+            except LarkError as e:
+                import sys
+                import textwrap
+                print(f"\nError parsing {name}:", file=sys.stderr)
+                print(textwrap.indent(str(e), "    "), file=sys.stderr)
+                sys.exit(1)
+
+            objects.update(objs)
+            interfaces.update(ints)
+
+            self.__deps[name] = imps
+
+            pending.update(imps.difference(self.__closed))
+
+        from .types import ObjectRef
+        ObjectRef.connect(objects)
+
+        for obj in objects.values():
+            for method in obj.methods:
+                caps = method.options.get("cap", list())
+                for cap in caps:
+                    if not cap in obj.caps:
+                        from .errors import UnknownCapError
+                        raise UnknownCapError(f"Unknown capability {cap} on {obj.name}::{method.name}")
+
+        self.objects.update(objects)
+        self.interfaces.update(interfaces)
+
+    def deps(self):
+        return {self.find(k): tuple(map(self.find, v)) for k, v in self.__deps.items()}