As reported in #3500, when debug interpreter is enabled, the classic interpreter
performs a lock operation to read `exec_env->current_status->signal_flag` and
do further handling before fetching next opcode, which makes the interpreter
run slower.
This PR atomic loads the `exec_env->current_status->signal_flag` without mutex
lock when 32-bit atomic load is supported, and only adding lock for further
handling when the signal_flag is WAMR_SIG_SINGSTEP, which improves the
performance.
Some host environment may also create an signal alternate stack and access
it after the wasm runtime exits, the runtime should backup the stack info and
restore it before thread exits.
The issue was found in golang:
```
signal 23 received on thread with on signal stack
fatal error: non-Go code disabled signaltstack
```
Error was reported when building jetstream benchmark with emsdk:
```
clang: error: unsupported option '-msse2' for target 'wasm32-wasi'
```
Error was reported when building bwa workload:
```
fatal: reference is not a tree: 139f68fc4c3747813783a488aef2adc86626b01b
```
Fix several issues of GC and exception handling in wasm loader:
- Should restore param_reftype_maps/param_reftype_map_count/param_count
in the handling of opcode throw
- Should set wasm_ref_type when pushing param types of tag type and block type
if the type is a multi-byte type
- Should set init_values.data as NULL for opcode struct.new_default in load_init_expr
This PR fixes the issues reported in #3411.
The compilation errors were introduced by #3515 and occur in debug building
when wasm mini loader is compiled or GC is enabled.
And remove two wasm files in standalone test-running-modes case,
which will be generated by run.sh.
In wasm_loader_pop_frame_offset, when the stack is in polymorphic state
and the stack top operand is VALUE_TYPE_ANY, if we popping I64/F64
operand, we should pop one offset but not two offsets.
The issue was reported in #3513 and #3514.
When datacount section exists, loader will check whether the data count
read from data segment section is same with the data count read from
datacount section, but the value of latter can be 0, loader should not
skip the check when the latter is 0. This fixes#3491.
And fix handle_name_section return value not checked issue and early
return true issue after handle_name_section.
And also add the failed case in #3491 to ba-issues.
Fix several issues in wasm loader:
- Parse a block's type index with leb int32 instead leb uint32
- Correct dst dynamic offset of loop block arguments for opcode br
when copying the stack operands to the arguments of loop block
- Free each frame_csp's param_frame_offsets when destroy loader ctx
- Fix compilation error in wasm_mini_loader.c
- Add test cases of failed issues
This PR fixes issue #3467 and #3468.
When copying two cells from src offsets to dst offsets in RECOVER_BR_INFO,
the offsets may be overlapped and the src data may be overwritten, use
GET_I64_FROM_ADDR and then SET_I64_FROM_ADDR instead to resolve it.
And handling VALUE_TYPE_FUNCREF/VALUE_TYPE_EXTERNREF for opcode
local.set and local.tee when reference types feature is enabled.
This PR fixes issue #3401 and #3402.
If there is no else branch, make a virtual else opcode for easier integrity
check and to copy the correct results to the block return address for
fast-interp mode: change if block from `if ... end` to `if ... else end`.
Reported in issue #3386, #3387, #3388.
Fix aot debugger compilation error on windows as reported in #3184.
And update the stack size configuration for product-mini zephyr sample
since the native stack overflow check was enhanced and the zephyr-sdk
was also upgraded.
Fix the integer overflow issue when checking target branch depth in opcode
br_table, and fix is_32bit_type not check VALUE_TYPE_ANY issue, which may
cause wasm_loader_push_frame_offset push extra unneeded offset.
Fix wasm loader integrity checks for opcode ref.func and opcode else:
for opcode ref.func, the function must be an import, exported, or present in a
table elem segment or global initializer to be used as the operand to ref.func,
for opcode else, there must not be an else opcode previously.
Reported in #3336 and #3337.
And fix mini loader PUSH_MEM_OFFSET/POP_MEM_OFFSET macro
definitions due to the introducing of memory64 feature.
Enhance the GC subtyping checks:
- Fix issues in the type equivalence check
- Enable the recursive type subtyping check
- Add a equivalence type flag in defined types of aot file, if there is an
equivalence type before, just set it true and re-use the previous type
- Normalize the defined types for interpreter and AOT
- Enable spec test case type-equivalence.wast and type-subtyping.wast,
and enable some commented cases
- Enable set WAMR_BUILD_SANITIZER from cmake variable
The current frame was freed before tail calling to an import or native function
and the prev_frame was set as exec_env's cur_frame, so after the tail calling,
we should recover context from prev_frame but not current frame.
Found in https://github.com/bytecodealliance/wasm-micro-runtime/issues/3279.
When thread manager is enabled, the aux stack of exec_env may be allocated
by wasm_cluster_allocate_aux_stack or disabled by setting aux_stack_bottom
as UINTPTR_MAX directly. For the latter, no need to free it.
And fix an issue when paring `--gc-heap-size=n` argument for iwasm, and
fix a variable shadowed warning in fast-jit.
Fix the warnings and issues reported:
- in Windows platform
- by CodeQL static code analyzing
- by Coverity static code analyzing
And update CodeQL script to build exception handling and memory features.
Adding a new cmake flag (cache variable) `WAMR_BUILD_MEMORY64` to enable
the memory64 feature, it can only be enabled on the 64-bit platform/target and
can only use software boundary check. And when it is enabled, it can support both
i32 and i64 linear memory types. The main modifications are:
- wasm loader & mini-loader: loading and bytecode validating process
- wasm runtime: memory instantiating process
- classic-interpreter: wasm code executing process
- Support memory64 memory in related runtime APIs
- Modify main function type check when it's memory64 wasm file
- Modify `wasm_runtime_invoke_native` and `wasm_runtime_invoke_native_raw` to
handle registered native function pointer argument when memory64 is enabled
- memory64 classic-interpreter spec test in `test_wamr.sh` and in CI
Currently, it supports memory64 memory wasm file that uses core spec
(including bulk memory proposal) opcodes and threads opcodes.
ps.
https://github.com/bytecodealliance/wasm-micro-runtime/issues/3091https://github.com/bytecodealliance/wasm-micro-runtime/pull/3240https://github.com/bytecodealliance/wasm-micro-runtime/pull/3260
The PR #3259 reverted PR #3192, it fixes#3210 but makes #3170 failed again.
The workaround is that we should update `ctx->dynamic_offset` only for opcode br
and should not update it for opcode br_if. This PR fixes both issue #3170 and #3210.
Some environment may call wasm_runtime_full_init/wasm_runtime_init multiple
times without knowing that runtime is initialized or not, it is better to add lock
and increase reference count during initialization.
ps. https://github.com/bytecodealliance/wasm-micro-runtime/discussions/3253.
Add cmake variable `-DWAMR_BUILD_AOT_INTRINSICS=1/0` to enable/disable
the aot intrinsic functions, which are normally used by AOT XIP feature, and
can be disabled to reduce the aot runtime binary size.
And refactor the code in aot_intrinsics.h/.c.
Should not update `ctx->dynamic_offset` in emit_br_info, since the `Part e` only
sets the dst offsets, the operand stack should not be changed, e.g., the stack
operands are to be used by the opcodes followed by `br_if` opcode.
Reported in https://github.com/bytecodealliance/wasm-micro-runtime/issues/3210.
- Fix nightly run CI failure which was introduced by PR #3239 and now it must set
WAMR_BUILD_TARGET when building iwasm for Android platform
- Remove building llvm, wamrc and jit in CodeQL CI, since it will
do static code analyzing for llvm project and cause CodeQL run
failed: `Oops! A fatal internal error occurred.
This particular kind of error most often happens as a side effect of running out of
disk space.`
Enhance CodeQL Code Security Analysis:
- Add more compilation combinations to build iwasm with different kinds of features
- Disable run on PR created and keep nightly run, since the whole time is very long,
and will check how to restore run on PR created in the future
The symbols in windows 32-bit may start with '_' and can not be found
when resolving the relocations to them. This PR ignores the underscore
when handling the relocation name of AOT_FUNC_INTERNAL_PREFIX, and
redirect the relocation with name "_aot_stack_sizes" to the relocation with
name ".aot_stack_sizes" (the name of the data section created).
ps.
https://github.com/bytecodealliance/wasm-micro-runtime/issues/3216
Merge branch dev/rust_sdk to main.
The wamr-rust-sdk has been migrated to the standalone repo:
https://github.com/bytecodealliance/wamr-rust-sdk
So here we just update the document to add the introduction.
Fix the errors reported in the sanitizer test of nightly run CI.
When the stack is in polymorphic state, the stack operands may be changed
after pop and push operations (e.g. stack is empty but pop op can succeed
in polymorphic, and the push op can push a new operand to stack), this may
impact the following checks to other target blocks of the br_table opcode.
As an original design rule, the code in `core/shared/platform` should not
rely on the code in `core/share/utils`. In the current implementation,
platform layer calls function `bh_memory_remap_slow` in utils layer.
This PR adds inline function `os_mremap_slow` in platform_api_vmcore.h,
and lets os_remap call it if mremap fails. And remove bh_memutils.h/c as
as they are unused.
And resolve the compilation warning in wamrc:
```bash
core/shared/platform/common/posix/posix_memmap.c:255:16:
warning: implicit declaration of function ‘bh_memory_remap_slow’
255 | return bh_memory_remap_slow(old_addr, old_size, new_size);
```
Implement the GC (Garbage Collection) feature for interpreter mode,
AOT mode and LLVM-JIT mode, and support most features of the latest
spec proposal, and also enable the stringref feature.
Use `cmake -DWAMR_BUILD_GC=1/0` to enable/disable the feature,
and `wamrc --enable-gc` to generate the AOT file with GC supported.
And update the AOT file version from 2 to 3 since there are many AOT
ABI breaks, including the changes of AOT file format, the changes of
AOT module/memory instance layouts, the AOT runtime APIs for the
AOT code to invoke and so on.
In linux x86-64, developer can use cmake variable to configure whether
to enable writing linear memory base address to x86 GS register or not:
- `cmake -DWAMR_DISABLE_WRITE_GS_BASE=1`: disabled it
- `cmake -DWAMR_DISABLE_WRITE_GS_BASE=0`: enabled it
- `cmake` without `-DWAMR_DISABLE_WRITE_GS_BASE=1/0`:
auto-detected by the compiler
Check whether the arguments are NULL before calling bh_hash_map_find,
or lots of "HashMap find elem failed: map or key is NULL" warnings may
be dumped. Reported in #3053.
Though SIMD isn't supported by interpreter, when JIT is enabled,
developer may run `iwasm --interp <wasm_file>` to trigger the SIMD
opcode in interpreter, which isn't handled before this PR.
- Enable quick aot entry when hw bound check is disabled
- Remove unnecessary ret_type argument in the quick aot entries
- Declare detailed prototype of aot function to call in each quick aot entry
After PR #2922, the caller of wasm-c-api wasm_func_call should ensure
the params and results are correctly created, runtime doesn't check
their format again.
Not know how to make `__attribute__((no_sanitize_address))` take effect
on gcc in Ubuntu-20.04, just disable quick AOT entry for wasm-c-api
sample in nightly-run CI.
Compilation error was reported when `cmake -DWAMR_BUILD_LIBC_WASI=0`
on linux-sgx platform:
```
core/shared/platform/linux-sgx/sgx_socket.c:8:10:
fatal error: libc_errno.h: No such file or directory
8 | #include "libc_errno.h"
| ^~~~~~~~~~~~~~
```
After fixing, both `cmake -DWAMR_BUILD_LIBC_WASI=1` and
`WAMR_BUILD_LIBC_WASI=0` work good.
Allow to invoke the quick call entry wasm_runtime_quick_invoke_c_api_import to
call the wasm-c-api import functions to speedup the calling process, which reduces
the data copying.
Use `wamrc --invoke-c-api-import` to generate the optimized AOT code, and set
`jit_options->quick_invoke_c_api_import` true in wasm_engine_new when LLVM JIT
is enabled.
In some scenarios there may be lots of callings to AOT/JIT functions from the
host embedder, which expects good performance for the calling process, while
in the current implementation, runtime calls the wasm_runtime_invoke_native
to prepare the array of registers and stacks for the invokeNative assemble code,
and the latter then puts the elements in the array to physical registers and
native stacks and calls the AOT/JIT function, there may be many data copying
and handlings which impact the performance.
This PR registers some quick AOT/JIT entries for some simple wasm signatures,
and let runtime call the entry to directly invoke the AOT/JIT function instead of
calling wasm_runtime_invoke_native, which speedups the calling process.
We may extend the mechanism next to allow the developer to register his quick
AOT/JIT entries to speedup the calling process of invoking the AOT/JIT functions
for some specific signatures.
Add an API to set segue flags for wasm-c-api LLVM JIT mode:
```C
wasm_config_t *
wasm_config_set_segue_flags(wasm_config_t *config, uint32 segue_flags);
```
- Don't allocate the implicit/unused frame when calling the LLVM JIT function
- Don't set exec_env's thread handle and stack boundary in the recursive
calling from host, since they have been set in the first time calling
- Fix frame not freed in llvm_jit_call_func_bytecode
The host embedder may also want to terminate the wasm instance
for single-threading mode, and it should work by setting exception
to the wasm instance.
