gentle ping

In D158978#4622400, @vitalybuka wrote:

What is tsan behavior?
And ubsan?

This sounds strange to me.

Intended usecase is to run but log errors.
But In multi-process environment failing process at the end will likely fails all parents.

So I'd rather reconsider behavior of msan and hwasan,

Thanks! I have run the test on 32-bit platforms successfully.

Thanks for testing.

Thanks for working on this, this is a real useful feature!

LGTM. I am not familiar with CMake, wait for @phosek's opinion ?

update from @fmayer comment

Thanks for your reply!
Internal symbolizer is a very useful feature, which lets us symbolize sanitizer error report without the dependency of external symbolizer tool.
I have tested the build_symbolizer.sh approach in my local environment, we will try to deploy the build_symbolizer.sh approach to our sanitizer buildbot.
The cmake approach may be more compilcated according to https://bugs.llvm.org/show_bug.cgi?id=30750, I will give it a try.

Hi @vitalybuka @pcc
is there any progress about this problem? Is symbolizer.o can be buildable with a regular cmake rule possible now? Thanks!

update

update

update from comments

update

gentle ping

Update from MaskRay's comments.

address review comments

Though, I guess you wouldn't ship coverage enabled in production.

Sure.

add comment about the intent of the test

address review comments

We are going to use D150708 to incrementally add interceptors

Thanks for you review!

Why do we want to skip sanitizing these loads and stores?

When use asan and gcov together, asan will instrument the loads and stores inserted by gcov which increases the code size a lot.
As the loads and stores inserted by gcov are unlikey to have bugs, so I add nosanitize metadata to these loads and stores like we do in SanitizerCoverage.

Also, gocv is misspelled in the commit description, please fix that up.

Thanks for pointing this.

address review comments

do not use match-all memset/memcpy/memmove for kernel mode HWASAN

do not use match-all callback for kernel mode HWASAN

gentle ping :)

rebase

update

implement __hwasan_load_match_all/__hwasan_store_match_all and handle hwasan-match-all-tag flag in them.

update

update

UNSUPPORTED: HWAddressSanitizer-x86_64 :: TestCases/short-granule-and-match-all-tag.cpp (19669 of 70430)

Pre-merge checks buildbot does not support HWASAN test, so we can not see this test failure in premerge checks .
Do you mean I create another differential only containing this testcase, commit it to llvm trunk and check if sanitizer-aarch64-linux-bootstrap-hwasan(https://lab.llvm.org/buildbot/#/builders/236) will fail?

remove additional test

just curious when stack will be allocated using regular allocator :)

• only test

update

I'm fine with waiting for HWASAN includes sanitizer_common_interceptors.inc :)

We encountered a crash due to this data race during the process of enabling hwasan for our server-side applications on AArch64/Linux. With this change, the crash will be fixed.

gentle ping, @eugenis PTAL, thanks!

format

address review comments

clang-format

update

rebase