christoph/wamr
1
Fork 0
Code Releases Activity

Refine Fast JIT call indirect and call native process (#1620)

Browse Source 
Translate call_indirect opcode by calling wasm functions with Fast JIT IRs instead of
calling jit_call_indirect runtime API, so as to improve the performance.

Translate call native function process with Fast JIT IRs to validate each pointer argument
and convert it into native address, and then call the native function directly instead
of calling jit_invoke_native runtime API, so as to improve the performance.
This commit is contained in:
Wenyong Huang
2022-10-19 17:11:38 +08:00
committed by GitHub
parent 73809efb5d
commit 1d4cbfceac
4 changed files with 525 additions and 161 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
core/iwasm/compilation/aot_emit_variable.c
View File

@ -120,16 +120,9 @@ compile_global(AOTCompContext *comp_ctx, AOTFuncContext *func_ctx,
LLVMValueRef offset, global_ptr, global, res;
LLVMTypeRef ptr_type = NULL;
if (comp_ctx->is_jit_mode) {
global_base_offset =
offsetof(WASMModuleInstance, global_table_data.bytes)
+ sizeof(WASMMemoryInstance) * comp_ctx->comp_data->memory_count;
}
else {
global_base_offset =
offsetof(AOTModuleInstance, global_table_data.bytes)
+ sizeof(AOTMemoryInstance) * comp_ctx->comp_data->memory_count;
}
global_base_offset =
offsetof(AOTModuleInstance, global_table_data.bytes)
+ sizeof(AOTMemoryInstance) * comp_ctx->comp_data->memory_count;
bh_assert(global_idx < import_global_count + comp_data->global_count);