Abhinav Anil Sharma
Abhinav Anil Sharma
Oh actually the failure in previous comment is due to some other regex mismatch (the AVX message was in the expected template, not output by the test run): an extra...
`code_api,thread_private|common.decode` failed on x86-32 `Internal Error: DynamoRIO debug check failure: /home/runner/work/dynamorio/dynamorio/core/vmareas.c:10664 TEST(FRAG_WAS_DELETED, f->flags) || info.overlap` https://github.com/DynamoRIO/dynamorio/runs/2823222611 The test failed once before as documented above, but with a different message. `prof_pcs,thread_private|common.nativeexec_retakeover_opt`...
`code_api,thread_private,tracedump_binary|common.fib` failed in `vs2017-32` https://github.com/DynamoRIO/dynamorio/runs/2897150100 ```Internal Error: DynamoRIO debug check failure: ..\core\vmareas.c:8312 !ok || !TEST(MEMPROT_WRITE, prot2) || TEST(FRAG_SELFMOD_SANDBOXED, *flags) || !INTERNAL_OPTION(hw_cache_consistency)```
`code_api,thread_private,tracedump_binary|common.fib` failed in `vs2017-32` again: https://github.com/DynamoRIO/dynamorio/runs/2955116184
`code_api,thread_private,tracedump_binary|common.floatpc_xl8all` failed on vs2017-64: https://github.com/DynamoRIO/dynamorio/runs/3023943403 `FXSAVE64 inter is **incorrectly** handled`
`|common.nativeexec_exe` failed on x86-64: https://github.com/DynamoRIO/dynamorio/runs/2964238871 `Internal Error: DynamoRIO debug check failure: /home/runner/work/dynamorio/dynamorio/core/native_exec.c:272 i < MAX_NATIVE_RETSTACK`
`code_api,thread_private|common.fib` failed in `vs2017-32`: https://github.com/DynamoRIO/dynamorio/runs/3242195135 `code_api,tracedump_text,tracedump_origins|common.fib` failed in `vs2017-64`: https://github.com/DynamoRIO/dynamorio/runs/3242195000
We had an offline chat where we discussed that option (A) is doable. Note that ensuring a single instr entry gets trickier if the repstr execution is interrupted by a...
> that will always re-enter the original block from the start but now with a different resumption state Oh, so after the fault, DR will not resume at the faulting...
I see that x86 prefixes are somewhat similar to OP_sys on AArch64. But there are two key differences that I think justify a different design (instead of splitting the opcodes):...