Implement 2-level Multi-tier JIT engine: tier-up from Fast JIT to LLVM JIT to
get quick cold startup by Fast JIT and better performance by gradually
switching to LLVM JIT when the LLVM JIT functions are compiled by the
backend threads.
Refer to:
https://github.com/bytecodealliance/wasm-micro-runtime/issues/1302
Error was reported when building docker image:
/usr/bin/ld: libaotclib.a(aot_llvm_extra.cpp.o): relocation R_X86_64_32S against symbol
`_ZN4llvm30TargetTransformInfoWrapperPass2IDE' can not be used when making a PIE object;
recompile with -fPIC
Add `-fPIC` to `CMAKE_CXX_FLAGS` and `CMAKE_SHARED_LIBRARY_LINK_C_FLAGS` to fix it.
Enable bulk memory by default since it is a finished wasm spec proposal
and is enabled by default in latest wasi-sdk. Developer often encounters
"invalid section id" or "unsupported opcode" error if it isn't enabled.
Use sha256 to hash binary file content. If the incoming wasm binary is
cached before, wasm_module_new() simply returns the existed one.
Use -DWAMR_BUILD_WASM_CACHE=0/1 to control the feature.
OpenSSL 1.1.1 is required if the feature is enabled.
Add an option to pass user data to the allocator functions. It is common to
do this so that the host embedder can pass a struct as user data and access
that struct from the allocator, which gives the host embedder the ability to
do things such as track allocation statistics within the allocator.
Compile with `cmake -DWASM_MEM_ALLOC_WITH_USER_DATA=1` to enable
the option, and the allocator functions provided by the host embedder should
be like below (an extra argument `data` is added):
void *malloc(void *data, uint32 size) { .. }
void *realloc(void *data, uint32 size) { .. }
void free(void *data, void *ptr) { .. }
Signed-off-by: Andrew Chambers <ncham@amazon.com>
Clear warnings like below when building LLVM JIT and wamrc:
llvm-project/llvm/include/llvm/ExecutionEngine/Orc/IndirectionUtils.h:73:25:
warning: redundant move in return statement [-Wredundant-move]
73 | return std::move(Err);
Add a new options to control the native stack hw bound check feature:
- Besides the original option `cmake -DWAMR_DISABLE_HW_BOUND_CHECK=1/0`,
add a new option `cmake -DWAMR_DISABLE_STACK_HW_BOUND_CHECK=1/0`
- When the linear memory hw bound check is disabled, the stack hw bound check
will be disabled automatically, no matter what the input option is
- When the linear memory hw bound check is enabled, the stack hw bound check
is enabled/disabled according to the value of input option
- Besides the original option `--bounds-checks=1/0`, add a new option
`--stack-bounds-checks=1/0` for wamrc
Refer to: https://github.com/bytecodealliance/wasm-micro-runtime/issues/1677
Add CIs to enable the release process of a new version of WAMR,
and build and publish the binary files when a version is released,
including iwasm, wamrc, lldb, vscode-extension and wamr-ide for
Ubuntu-20.04, Ubuntu-22.04 and MacOS.
And refine the CIs to test spec cases.
Use the cmake variable `WAMR_BUILD_GLOBAL_HEAP_POOL` and
`WAMR_BUILD_GLOBAL_HEAP_SIZE` to enable/disable the global heap pool
and set its size. And set the default global heap size in core/config.h and
the cmake files.
As a result, the developers who build iwasm can easily enable/disable the
global heap pool and change its size regardless of the iwasm implementation,
without manually finding and patching the right location for that value.
Refactor LLVM JIT for some purposes:
- To simplify the source code of JIT compilation
- To simplify the JIT modes
- To align with LLVM latest changes
- To prepare for the Multi-tier JIT compilation, refer to #1302
The changes mainly include:
- Remove the MCJIT mode, replace it with ORC JIT eager mode
- Remove the LLVM legacy pass manager (only keep the LLVM new pass manager)
- Change the lazy mode's LLVM module/function binding:
change each function in an individual LLVM module into all functions in a single LLVM module
- Upgraded ORC JIT to ORCv2 JIT to enable lazy compilation
Refer to #1468
Refactor the layout of interpreter and AOT module instance:
- Unify the interp/AOT module instance, use the same WASMModuleInstance/
WASMMemoryInstance/WASMTableInstance data structures for both interpreter
and AOT
- Make the offset of most fields the same in module instance for both interpreter
and AOT, append memory instance structure, global data and table instances to
the end of module instance for interpreter mode (like AOT mode)
- For extra fields in WASM module instance, use WASMModuleInstanceExtra to
create a field `e` for interpreter
- Change the LLVM JIT module instance creating process, LLVM JIT uses the WASM
module and module instance same as interpreter/Fast-JIT mode. So that Fast JIT
and LLVM JIT can access the same data structures, and make it possible to
implement the Multi-tier JIT (tier-up from Fast JIT to LLVM JIT) in the future
- Unify some APIs: merge some APIs for module instance and memory instance's
related operations (only implement one copy)
Note that the AOT ABI is same, the AOT file format, AOT relocation types, how AOT
code accesses the AOT module instance and so on are kept unchanged.
Refer to:
https://github.com/bytecodealliance/wasm-micro-runtime/issues/1384
Initial integration of WASI-NN based on #1225:
- Implement the library core/iwasm/libraries/wasi-nn
- Support TensorFlow, CPU, F32 at the first stage
- Add cmake variable `-DWAMR_BUILD_WASI_NN`
- Add test case based on Docker image and update document
Refer to #1573
This PR integrates an Intel SGX feature called Intel Protection File System Library (IPFS)
into the runtime to create, operate and delete files inside the enclave, while guaranteeing
the confidentiality and integrity of the data persisted. IPFS can be referred to here:
https://www.intel.com/content/www/us/en/developer/articles/technical/overview-of-intel-protected-file-system-library-using-software-guard-extensions.html
Introduce a cmake variable `WAMR_BUILD_SGX_IPFS`, when enabled, the files interaction
API of WASI will leverage IPFS, instead of the regular POSIX OCALLs. The implementation
has been written with light changes to sgx platform layer, so all the security aspects
WAMR relies on are conserved.
In addition to this integration, the following changes have been made:
- The CI workflow has been adapted to test the compilation of the runtime and sample
with the flag `WAMR_BUILD_SGX_IPFS` set to true
- Introduction of a new sample that demonstrates the interaction of the files (called `file`),
- Documentation of this new feature
And enable classic interpreter instead fast interpreter when llvm jit is enabled,
so as to fix the issue that llvm jit cannot handle opcode drop_64/select_64.
Use the semantic versioning (https://semver.org) to replace the current date
versioning system, which is more general and is requested by some developers,
e.g. issue #1357.
There are three parts in the new version string:
- major. Any incompatible modification on ABIs and APIs will lead to an increment
in the value of major, which mainly includes: AOT calling conventions, AOT file
format, wasm_export.h, wasm_c_api.h, and so on.
- minor. It represents new features, including MVP/POST-MVP features, libraries,
WAMR private ones, and so one.
- patch. It represents patches.
The new version will start from 1.0.0. Update the help info and version showing for
iwasm and wamrc.
Add a new option WAMR_BUILD_STACK_GUARD_SIZE to set the custom
stack guard size. For most RTOS systems, we use the native stack base
address as the check boundary which may be not safe as POSIX based
systems (like Linux).
Import WAMR Fast JIT which is a lightweight JIT with quick startup, small footprint,
relatively good performance (~40% to ~50% of LLVM JIT) and good portability.
Platforms supported: Linux, MacOS and Linux SGX.
Arch supported: x86-64.
Fix build script to enable hw bound check for interpreter when
AOT is disabled, so as to enable spec cases test for interp with
hw bound check. And fix the issues found.
When WAMR_BUILD_TARGET isn't set, choosing right target is decided
by checking `CMAKE_SIZEOF_VOID_P` variable. However, choosing `X86_32`
target is not doing specifically checking size of void pointer. It is kind
a fallback target for others.
This patch explicitly checks the size of void pointer before setting the target
to `X86_32` to fix the issue.
Increase default/min native stack size when UVWASI is enabled as
UVWASI requires larger native stack size.
Increase the reserved bytes to the native thread stack boundary to
better detect the native stack overflow.
Set WASM_DISABLE_HW_BOUND_CHECK to 0 when interpreter is
enabled and AOT is disabled, as memory access boundary check
with hardware trap is only enabled in AOT/JIT mode.
Use LLVM new pass manager for wamrc to replace the legacy pass manger,
so as to gain better performance and reduce the compilation time.
Reference links:
- https://llvm.org/docs/NewPassManager.html
- https://blog.llvm.org/posts/2021-03-26-the-new-pass-manager
And add an option to use the legacy pm mode when building wamrc:
cmake .. -DWAMR_BUILD_LLVM_LEGACY_PM=1
For JIT mode, keep it unchanged as it only runs several function passes and
using new pass manager will increase the compilation time.
And refactor the codes of applying LLVM passes.
Refactor LLVM Orc JIT to actually enable the lazy compilation and speedup
the launching process:
https://llvm.org/docs/ORCv2.html#laziness
Main modifications:
- Create LLVM module for each wasm function, wrap it with thread safe module
so that the modules can be compiled parallelly
- Lookup function from aot module instance's func_ptrs but not directly call the
function to decouple the module relationship
- Compile the function when it is first called and hasn't been compiled
- Create threads to pre-compile the WASM functions parallelly when loading
- Set Lazy JIT as default, update document and build/test scripts
Various fixes and beautifications coordinated with @1c3t3a,
fixes 2 of the 3 all remaining issues from #892:
- enable to os_mmap executable memory
- fix os_malloc/os_realloc/os_free issues
- implement os_thread_get_stack_boundary
- add build scripts to include with esp-idf to use wamr as
an ESP-IDF component
- update sample and document
When compile 32-bit targets on some 64-bit systems, the "-m32" flag might be
unrecognized by some gcc compilers, e.g. compiling arm32 in aarch64 system,
compiling riscv32 in riscv64 system.
Add check before adding "-m32" flag to gcc, and only add it if it is supported.
Enhance workflows by caching llvm directories directly instead of
caching the llvm-xxx.tar.gz packages, so as to reduce the time
consumption of unpacking the cached LLVM packages.
- fix data race issue between debug control thread and main thread
- fix possible memory leaks in breakpoints list
- fix memory uninitialized issues
- remove unused data structures
- add more checks when handling packet and args
- fix mini-loader issues
- fix config_common.cmake fast interp prompt issue
Implement source debugging feature for classic interpreter and AOT:
- use `cmake -DWAMR_BUILD_DEBUG_INTERP=1` to enable interpreter debugging
- use `cmake -DWAMR_BUILD_DEBUG_AOT=1` to enable AOT debugging
See doc/source_debugging.md for more details.
Implement the latest SIMD opcodes and update LLVM 13.0,
update the llvm build scripts, update the sample workloads‘ build scripts,
and build customized wasi-sdk to build some workloads.
Also refine the CI rules.
Signed-off-by: Wenyong Huang <wenyong.huang@intel.com>
Implement XIP (Execution In Place) feature for AOT mode to enable running the AOT code inside AOT file directly, without memory mapping the executable memory for AOT code and applying relocations for text section. Developer can use wamrc with "--enable-indirect-mode --disable-llvm-intrinsics" flags to generate the AOT file and run iwasm with "--xip" flag. Known issues: there might still be some relocations in the text section which access the ".rodata" like sections.
And also enable ARC target support for both interpreter mode and AOT mode.
Signed-off-by: Wenyong Huang <wenyong.huang@intel.com>
