Details
- Reviewers
mcgrathr phosek vitalybuka
Diff Detail
- Repository
- rG LLVM Github Monorepo
Unit Tests
Event Timeline
compiler-rt/lib/asan/asan_allocator.h | ||
---|---|---|
129–142 | i suspect these smaller values has something to do with ability allocate shadow, which is not an issue for lsan |
compiler-rt/lib/asan/asan_allocator.h | ||
---|---|---|
129–142 | Do you recommend I should remove them? |
No objections, but some small concerns that it will be annoying to split them again if we need different setups for these sanitizers.
compiler-rt/lib/asan/asan_allocator.h | ||
---|---|---|
129–142 | we can't just remove them for asan, because there probably was a reason | |
compiler-rt/lib/lsan/lsan_common.h | ||
80 | SANITIZER_ARM64 and the rest? | |
compiler-rt/lib/sanitizer_common/sanitizer_platform.h | ||
252–257 | this part looks a little bit unrelated, would it be better to have a separate patch? Can you swticth to stuff like SANITIZER_ARM? Also maybe it's better to have condition SANITIZER_WORDSIZE != 64 || the rest? |
compiler-rt/lib/asan/asan_allocator.h | ||
---|---|---|
129–142 | Oh I see, yeah perhaps we can instead just "unify" the values that are already matching in the asan and lsan configs rather than moving everything. | |
compiler-rt/lib/sanitizer_common/sanitizer_platform.h | ||
252–257 | Uploaded https://reviews.llvm.org/D126825 which instead just tells lsan to use SANITIZER_CAN_USE_ALLOCATOR64 (which should just be a no-op). Then I'll have a followup patch that sets this macro to 1 for aarch64. |
i suspect these smaller values has something to do with ability allocate shadow, which is not an issue for lsan