build: propagate ASAN/TSAN flags down to cmake_external deps.

[htuch]

While we await a solution for bazel-contrib/rules_foreign_cc#154 (comment), this PR provides a
temporary workaround to get full ASAN/TSAN propagation as needed.

This continues #6061, which had accumulated too much
complicated merge history to work with after the recent gperftools/LuaJIT migration to
rules_foreign_cc.

Risk level: Low
Testing: --config=clang-{tsan,asan}. This is broken in the latest CI images due to the mismatched
linker/compiler flags we previously had been sending to cmake_external builds.

Signed-off-by: Harvey Tuch htuch@google.com

[htuch]

OK, this is great, we don't need the gperftools patch after moving over to rules_foreign_cc 👍

[PiotrSikora]

It looks that the Lua tests are failing with ASan/TSan, could you investigate? Though, the issue is most likely with Lua itself, and not with your changes, which only uncovered it.

[htuch]

@PiotrSikora I know, I'm digging into this now. I'm hoping I can figure out a way to workaround these Envoy side, if not upstream, in a timely way, this is blocker for the coverage work atm.

[PiotrSikora]

@htuch another way to approach this, assuming Lua issues are going to take more than a day or two to resolve, is to temporarily exclude Lua tests from ASan and TSan runs, since this PR is an improvement, all things considered, and it's not the reason for the breakage.

[lizan]

or blacklist Lua itself from being instrumented by ASAN? https://github.com/google/sanitizers/wiki/AddressSanitizer#turning-off-instrumentation

[htuch]

@lizan I think I'll blacklist, I've been looking at these errors and there are a few different varieties. We need someone to spend some serious time looking at this, it's not a high priority right now for us though.

[htuch]

Sigh, now we have some unrelated LSAN issue with host protoc, I guess we can disable LSAN but not sure why it started happening now..

[lizan]

@htuch I see that on some other PRs as well, so it might not directly related to this PR. Could be some docker capabilities issue with Circle.

/retest

[repokitteh-read-only]

🔨 rebuilding ci/circleci: asan (failed build)

🐱

Caused by: a #6196 (comment) was created by @lizan.

see: more, trace.

[htuch]

@lizan I think I have a fix based on what you did for LuaJIT. This should hopefully help other PRs.

[lizan]

@htuch seems it didn't help, LeakSanitizer has encountered a fatal error is probably something that LSAN bug or LSAN itself couldn't recover... I'm not sure why protoc is build with ASAN, as the flags are not set to host_cflags, no?

[htuch]

@lizan I don't think there is a clear split of host/target in https://github.com/protocolbuffers/protobuf/blob/master/BUILD. I will continue to investigate in this PR.

[stale]

This pull request has been automatically marked as stale because it has not had activity in the last 7 days. It will be closed in 7 days if no further activity occurs. Please feel free to give a status update now, ping for review, or re-open when it's ready. Thank you for your contributions!

[stale]

This pull request has been automatically closed because it has not had activity in the last 14 days. Please feel free to give a status update now, ping for review, or re-open when it's ready. Thank you for your contributions!