477 Commits

Author	SHA1	Message	Date
Marcin Kolny	c883fafead	Merge pull request #4203 from jammar1/jammar1/fix-br-v128 ... Add missing V128 handling in WASM_OP_BR, reported in #4173	2025-04-22 14:47:08 +01:00
James Marsh	0702f788fd	Add missing V128 handling in WASM_OP_BR, reported in #4173	2025-04-17 10:43:28 +01:00
liang.he	996758cd4a	Remove the dlen to optimize it. (#4193) ... There are two reasons for this optimization: - The value of dlen can equal 0x1_0000_0000, even in wasm32 mode, because it is derived from (4G-0). This results in a truncation when it is passed to b_memmove_s(). Consequently, s1max becomes 0 and n is greater than s1max. To correct this, a longer type is required. - The dlen is only used to check if there is enough space in b_memmove_s(). However, from a different angle, after confirming that both src+len and dst+len are within the memory range, we can be assured and there is no need for this explicit check.	2025-04-17 15:21:02 +08:00
YAMAMOTO Takashi	46ec863da3	fix false native stack overflow detections with HW_BOUND_CHECK (#4196) ... In call_wasm_with_hw_bound_check/call_native_with_hw_bound_check, ensure to set up the stack boundary (wasm_exec_env_set_thread_info) before checking the overflow. It seems that the problem was introduced by: https://github.com/bytecodealliance/wasm-micro-runtime/pull/2940	2025-04-15 11:48:48 +08:00
Huang Qi	cc1903603d	fix: Remove unused variables in SIMD_v128_const case (#4197) ... Fix compiler warnings about unused variables `high` and `low` in the `SIMD_v128_const` case. These variables are only needed inside the `WASM_ENABLE_FAST_INTERP != 0` conditional block, but were incorrectly declared outside of it, causing unused variable warnings. Signed-off-by: Huang Qi <huangqi3@xiaomi.com>	2025-04-15 09:47:18 +08:00
liang.he	751cdcf073	improve variable naming and code clarity in SIMD operations (#4157) ... Fix compilation warning about shadow, like ```sh declaration of ‘val’ shadows a previous local [-Wshadow] ```	2025-04-10 14:36:20 +08:00
liang.he	9aaf3599ec	prevent frame_offset underflow in wasm_loader (#4165)	2025-04-10 12:06:06 +08:00
liang.he	793135b41c	Fix few integer overflowing (#4161) ... - fix(interpreter): correct offset calculations in wasm_loader_get_const_offset function - fix(mem-alloc): update offset calculation in gc_migrate for memory migration - add pointer-overflow sanitizer	2025-04-10 12:04:56 +08:00
TianlongLiang	8fe98f64c1	Add import memory/table flag assert check for miniloader (#4179)	2025-04-10 12:00:23 +08:00
Marcin Kolny	efa8019bdb	Merge dev/simd for fast-interp (#4131) ... * Implement the first few SIMD opcodes for fast interpreter (v128.const, v128.any_true) (#3818) Tested on the following code: ``` (module (import "wasi_snapshot_preview1" "proc_exit" (func $proc_exit (param i32))) (memory (export "memory") 1) ;; WASI entry point (func $main (export "_start") v128.const i8x16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 v128.any_true if unreachable end v128.const i8x16 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 v128.any_true i32.const 0 i32.eq if unreachable end i32.const 0 call $proc_exit ) ) ``` * implement POP_V128() This is to simplify the simd implementation for fast interpreter * Add all SIMD operations into wasm_interp_fast switch * Add V128 comparison operations Tested using ``` (module (import "wasi_snapshot_preview1" "proc_exit" (func $proc_exit (param i32))) (memory (export "memory") 1) (func $assert_true (param v128) local.get 0 v128.any_true i32.eqz if unreachable end ) (func $main (export "_start") ;; Test v128.not v128.const i8x16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 v128.not v128.const i8x16 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 i8x16.eq call $assert_true ;; Test v128.and v128.const i8x16 255 255 255 255 0 0 0 0 255 255 255 255 0 0 0 0 v128.const i8x16 255 255 0 0 255 255 0 0 255 255 0 0 255 255 0 0 v128.and v128.const i8x16 255 255 0 0 0 0 0 0 255 255 0 0 0 0 0 0 i8x16.eq call $assert_true ;; Test v128.andnot v128.const i8x16 255 255 255 255 0 0 0 0 255 255 255 255 0 0 0 0 v128.const i8x16 255 255 0 0 255 255 0 0 255 255 0 0 255 255 0 0 v128.andnot v128.const i8x16 0 0 255 255 0 0 0 0 0 0 255 255 0 0 0 0 i8x16.eq call $assert_true ;; Test v128.or v128.const i8x16 255 255 0 0 0 0 255 255 255 255 0 0 0 0 255 0 v128.const i8x16 0 0 255 255 255 255 0 0 0 0 255 255 255 255 0 0 v128.or v128.const i8x16 255 255 255 255 255 255 255 255 255 255 255 255 255 255 255 0 i8x16.eq call $assert_true ;; Test v128.xor v128.const i8x16 255 255 0 0 255 255 0 0 255 255 0 0 255 255 0 0 v128.const i8x16 255 255 255 255 0 0 0 0 255 255 255 255 0 0 0 0 v128.xor v128.const i8x16 0 0 255 255 255 255 0 0 0 0 255 255 255 255 0 0 i8x16.eq call $assert_true i32.const 0 call $proc_exit ) ) ``` * Add first NEON SIMD opcode implementations to fast interpreter (#3859) Add some implementations of SIMD opcodes using NEON instructions. Tested using: ```wast (module (import "wasi_snapshot_preview1" "proc_exit" (func $proc_exit (param i32))) (memory (export "memory") 1) (func $assert_true (param v128) local.get 0 v128.any_true i32.eqz if unreachable end ) (func $main (export "_start") i32.const 0 i32.const 32 memory.grow drop i32.const 0 v128.const i8x16 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 v128.store i32.const 0 v128.load v128.const i8x16 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 i8x16.eq call $assert_true i32.const 16 v128.const i8x16 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 v128.store i32.const 16 v128.load v128.const i8x16 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 i8x16.eq call $assert_true i32.const 0 v128.load v128.const i8x16 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 i8x16.eq call $assert_true drop i32.const 0 i32.const 1 memory.grow drop i32.const 0 i64.const 0x7F80FF017E02FE80 i64.store i32.const 0 v128.load8x8_s v128.const i16x8 127 -128 -1 1 126 2 -2 -128 i16x8.eq call $assert_true i32.const 0 i64.const 0x80FE027E01FF807F i64.store i32.const 0 v128.load8x8_u v128.const i16x8 128 254 2 126 1 255 128 127 i16x8.eq call $assert_true i32.const 0 i64.const 0x8000FFFE7FFF0001 i64.store i32.const 0 v128.load16x4_s v128.const i32x4 -32768 -2 32767 1 i32x4.eq call $assert_true i32.const 0 i64.const 0x8000FFFE7FFF0001 i64.store i32.const 0 v128.load16x4_u v128.const i32x4 32768 65534 32767 1 i32x4.eq call $assert_true i32.const 0 i64.const 0x8000000000000001 i64.store i32.const 0 v128.load32x2_s v128.const i64x2 -2147483648 1 i64x2.eq call $assert_true i32.const 0 i64.const 0x8000000000000001 i64.store i32.const 0 v128.load32x2_u v128.const i64x2 2147483648 1 i64x2.eq call $assert_true call $proc_exit ) ) ``` * Emit imm for lane extract and replace (#3906) * Fix replacement value not being correct (#3919) * Implement load lanes opcodes for wasm (#3942) * Implement final SIMD opcodes: store lane (#4001) * Fix load/store (#4054) * Correctly use unsigned functions (#4055) * implement local and function calls for v128 in the fast interpreter * Fix splat opcodes, add V128 handling in preserve_referenced_local and reserve_block_ret * Fix incorrect memory overflow values + SIMD ifdefs * Fix load/load_splat macros * correct endif wasm loader * Update core/iwasm/interpreter/wasm_opcode.h * Fix spec tests when WASM_CPU_SUPPORTS_UNALIGNED_ADDR_ACCESS is 0 * Resolve merge conflicts arising from main -> dev/simd_for_interp and implement fast interpreter const offset loader support for V128 * Enable SIMDe tests on CI * Document WAMR_BUILD_LIB_SIMDE --------- Co-authored-by: James Marsh <mrshnja@amazon.co.uk> Co-authored-by: jammar1 <108334558+jammar1@users.noreply.github.com> Co-authored-by: Maks Litskevich <makslit@amazon.com> Co-authored-by: Marcin Kolny <marcin.kolny@gmail.com> Co-authored-by: Wenyong Huang <wenyong.huang@intel.com>	2025-03-20 14:23:20 +08:00
liang.he	766f378590	Merge pull request #4033 from g0djan/godjan/iterate_callstack ... Copy callstack API	2025-03-11 10:31:56 +08:00
Huang Qi	412631ac13	fix: correct typos and improve comments across multiple files by codespell (#4116) ... Signed-off-by: Huang Qi <huangqi3@xiaomi.com>	2025-03-07 08:21:54 +08:00
Georgii Rylov	e488345607	fix bug for return value when skip_n is passed	2025-03-05 11:47:56 +00:00
Georgii Rylov	56bb7e715b	last fixes	2025-03-05 09:11:21 +00:00
Georgii Rylov	32d0f5503e	spare line	2025-03-03 14:06:45 +00:00
Georgii Rylov	0b5084cb63	remove spare diff line	2025-02-27 15:06:31 +00:00
Georgii Rylov	fc3077b74d	address comments	2025-02-27 14:32:17 +00:00
Georgii Rylov	857e6b73c8	formatting	2025-02-26 10:57:50 +00:00
liang.he	f2e3348305	wasm_loader allocates more spaces for elements (#4099) ... - allocate memory for array initialization based on length - update reference type mapping for struct initialization	2025-02-26 12:58:45 +08:00
Georgii Rylov	188eb1c941	remove unnecessary includes	2025-02-25 17:42:06 +00:00
Georgii Rylov :slightly_smiling_face	32338bb7d6	Copy read only API behind a flag instead of using user defined callback	2025-02-24 17:22:05 +00:00
Jérôme Vouillon	159f5890a6	[gc] Subtyping fix (#4075)	2025-02-17 11:55:58 +08:00
Georgii Rylov	b9039f986e	Merge branch 'main' into godjan/iterate_callstack	2025-01-28 13:35:47 +00:00
Georgii Rylov	478b373cda	update signature	2025-01-28 13:27:01 +00:00
Georgii Rylov	b6daacb836	ignore frames with no function	2025-01-28 11:25:45 +00:00
Georgii Rylov	6bfc08849a	Calculate func_index instead of adding an extra field to wasm frame	2025-01-28 11:09:32 +00:00
Georgii Rylov	1f4d3dd4d4	clang-format	2025-01-27 11:31:02 +00:00
TianlongLiang	b6dea221a6	Fix wasm loader check data segment count (#4039) ... correctly report error when datacount section has non-zero data segment count while the data section is not present	2025-01-21 13:08:09 +08:00
liang.he	e3ddbd58f7	Synchronize the GC spec tests to the commit from December 9. 2024. (#4022) ... - Synchronize the GC spec tests to the commit from December 9. 2024. - Revise the error messages to be consistent with the spec test cases. - bypass gc spec test on the nuttx platform as a workaround	2025-01-20 09:41:47 +08:00
Wenyong Huang	831e4bbfd5	Refine getting const offsets in wasm loader of fast-interp (#4012) ... - Refine const offsets in loader for fast-interp - handle const cell num overflow - Use const array, remove list	2025-01-20 09:39:32 +08:00
Georgii Rylov	68e4534822	Iterate callstack API	2025-01-17 16:16:45 +00:00
Wenyong Huang	9c3807e124	Refine read leb int wasm loader of fast interpreter (#4017)	2025-01-14 17:43:29 +08:00
liang.he	53da420c41	Enable shrunk memory by default and add related configurations (#4008) ... - Enable shrunk memory by default and add related configurations - Improve error messages for memory access alignment checks - Add documentation for WAMR shrunk memory build option - Update NuttX workflow to disable shrunk memory build option	2025-01-13 07:09:04 +08:00
liang.he	02683d2eed	Improve stack consistency by ensuring sufficient space for dummy offsets (#4011) ... One more corner case: if the `frame_offset` increases and becomes equal to the `frame_offset_boundary` after the last assignment within the for loop.	2025-01-09 13:11:25 +08:00
TianlongLiang	a653746b7b	Check whether related table has funcref elem in opcode call_indirect (#3999) ... * check whether table has funcref elem in call_indirect * check whether table has funcref elem in call_indirect when gc is enabled	2025-01-06 13:55:43 +08:00
liang.he	9989b1cc1b	[fuzzing] Use software bound-check during fuzzing (#4003) ... * Update CMakeLists.txt of fuzzing - enable software bound-check - enable wasi - disable libc builtin and multiple modules * Fix off-by-one error in result offset calculation for function calls	2025-01-06 11:36:11 +08:00
liang.he	1958808a24	Fix table index calculations in wasm_loader and wasm_mini_loader (#4004)	2025-01-05 15:27:40 +08:00
liang.he	099056b076	Ensure __heap_base and __data_end global indices are validated against import count (#3996)	2025-01-03 14:37:09 +08:00
Marcin Kolny	31ff576edf	Error message improvement (#4000) ... Improve error message in the scenario where the runtime was built with ref types disabled but the module uses reference types feature.	2025-01-03 10:44:25 +08:00
TianlongLiang	1d111a38d6	Fix loader small bug (#3928)	2024-11-26 10:08:51 +08:00
WenLY1	b0c6d5c23a	add testcases for shared heap and fix POP_MEM_OFFSET of memory64 (#3916) ... - add testcases for shared_heap - fix POP_MEM_OFFSET and POP_TBL_ELEM_IDX of memory64 Signed-off-by: wenlingyun1 <wenlingyun1@xiaomi.com>	2024-11-24 11:34:38 +08:00
James Ring	dbdf3df60b	Use plain assignment rather than bh_memcpy_s (#3924)	2024-11-24 11:32:34 +08:00
James Ring	00c2aa10a8	Drop declarative elements on module instantiation (#3922)	2024-11-24 11:30:00 +08:00
TianlongLiang	0e4dffc479	Fix a leak in wasm_loader_emit_br_info (#3900) ... Reference Info: 377955855 wamr:wasm_mutator_fuzz_loader: Direct-leak in wasm_loader_emit_br_info https://issues.oss-fuzz.com/issues/377955855	2024-11-13 15:16:13 +08:00
James Ring	fdda259d36	Fix linked global initialization in multimodule (#3905) ... While resolving linked globals in multi-module mode, WAMR tries to copy the linked global's initial value into the destination global in the current module. However, a bug in the implementation causes the copy to be done from the InitializerExpression struct, not from its WASMValue field. This did not come up in WAMR's spec test runner because those are built with WASM_ENABLE_SPEC_TEST, which means these globals are resolved as builtins, not linked globals, which goes through a different (presumably not faulty) path.	2024-11-13 14:52:27 +08:00
TianlongLiang	bf78863c56	Wasm loader enhancement: check code size in code entry (#3892) ... add wasm loader check: in code entry, the code size should match the size of vec(locals) + expr, and expr should end with opcode end	2024-11-07 13:38:42 +08:00
Maks Litskevich	6426fc4993	Fix out of bounds issues after memory.grow on non-aot non-threads builds (#3872) ... Co-authored-by: Deniz Sokmen <dsokmen@amazon.com>	2024-10-23 14:48:40 +08:00
James Ring	3ad95303d6	Fix quadratic runtime for duplicate export name detection (#3861) ... Previously, the loader would check the name of a new export against all existing exports, leading to a quadratic running time. This change makes the loader parse the entire export section. The exports are then sorted by name, then adjacent exports are checked for uniqueness.	2024-10-21 11:02:30 +08:00
Wenyong Huang	87588caa7f	Fix lookup function issue reported in nightly run (#3868)	2024-10-21 09:57:42 +08:00
Wenyong Huang	48eaa2286a	Refine wasm/aot function instance lookup (#3865) ... Sort the module instance's export functions with the function name, and use binary search to lookup the wasm/aot function.	2024-10-18 15:15:33 +08:00