Page MenuHomePhabricator

earthdok (Sergey Matveev)
User

Projects

User does not belong to any projects.

User Details

User Since
Feb 4 2013, 5:58 AM (341 w, 5 d)

Recent Activity

May 6 2015

earthdok added a comment to D9228: [sanitizer] Extend sancov.py to show which PCs are missing from coverage..
In D9228#167162, @kcc wrote:

LGTM, thanks!

Once committed, please also update the docs at http://clang.llvm.org/docs/SanitizerCoverage.html#how-good-is-the-coverage

May 6 2015, 2:12 PM
earthdok added a comment to D9228: [sanitizer] Extend sancov.py to show which PCs are missing from coverage..

Committed r236637

May 6 2015, 1:52 PM
earthdok updated the diff for D9228: [sanitizer] Extend sancov.py to show which PCs are missing from coverage..
  • fix meaningless sorts
May 6 2015, 12:04 PM
earthdok added a comment to D9228: [sanitizer] Extend sancov.py to show which PCs are missing from coverage..
In D9228#166797, @kcc wrote:

We already have tests that run sancov.py

May 6 2015, 11:31 AM
earthdok updated the diff for D9228: [sanitizer] Extend sancov.py to show which PCs are missing from coverage..
  • update regexp
  • fix regexp: accept "call", accept "@plt"
  • testcase
May 6 2015, 11:29 AM

May 5 2015

earthdok added a comment to D8318: [sanitizer][MIPS] Implement clone for MIPS.

Stoptheworld isn't actually used on 32-bit platforms, and that's the only thing you need clone for. But since you already have the implementation, let's land it. rslgtm

May 5 2015, 1:24 PM
earthdok added a comment to D9228: [sanitizer] Extend sancov.py to show which PCs are missing from coverage..
In D9228#165634, @kcc wrote:

looks good, but I wonder if a test can be added?

May 5 2015, 8:11 AM

Apr 24 2015

earthdok added a comment to D9159: [lsan] Add an interface function for on-demand leak checking..

Committed r235728

Apr 24 2015, 9:57 AM
earthdok updated the diff for D9159: [lsan] Add an interface function for on-demand leak checking..
  • change "extra" to "recoverable"
Apr 24 2015, 9:50 AM

Apr 23 2015

earthdok updated the diff for D9228: [sanitizer] Extend sancov.py to show which PCs are missing from coverage..
  • update regexp
Apr 23 2015, 11:34 AM
earthdok retitled D9228: [sanitizer] Extend sancov.py to show which PCs are missing from coverage. from to [sanitizer] Extend sancov.py to show which PCs are missing from coverage..
Apr 23 2015, 10:58 AM

Apr 21 2015

earthdok updated the diff for D9159: [lsan] Add an interface function for on-demand leak checking..
  • address comments
Apr 21 2015, 1:13 PM
earthdok added a comment to D9159: [lsan] Add an interface function for on-demand leak checking..

Leaking symbolization is an issue... Let me know if you want to investigate it, or delegate it to me.

Fixed in r235411

Apr 21 2015, 1:12 PM
earthdok added a comment to D9159: [lsan] Add an interface function for on-demand leak checking..

Our symbolization code still leaks. :( When I call this function in an infinite loop I observe a small but steady increase in memory usage (as shown by smem) - but only when symbolization is enabled.

Apr 21 2015, 11:48 AM
earthdok added a comment to D9159: [lsan] Add an interface function for on-demand leak checking..

Incidentally this will call the summary callback every time leaks are detected. Do you foresee any problems with that?

Apr 21 2015, 10:19 AM
earthdok added a comment to D9159: [lsan] Add an interface function for on-demand leak checking..

As always, I hate the name, so any suggestions are welcome.

Apr 21 2015, 10:15 AM
earthdok retitled D9159: [lsan] Add an interface function for on-demand leak checking. from to [lsan] Add an interface function for on-demand leak checking..
Apr 21 2015, 10:14 AM
earthdok accepted D8318: [sanitizer][MIPS] Implement clone for MIPS.

lgtm

Apr 21 2015, 7:08 AM

Apr 17 2015

earthdok added a comment to D8318: [sanitizer][MIPS] Implement clone for MIPS.
In D8318#154546, @sagar wrote:

Hi @earthdok,

Note also that the glibc implementation appears to have a few differences between 32 and 64 bit mips, so please double check those to make sure they don't affect us.

The only difference between 32-bit and 64-bit in glibc implementation I can see is argument passing to clone syscall. This patch was written according to n64 abi of mips. In case of n64 abi we have argument registers a0 to a5 for syscalls. Therefore this patch will work fine for LSan 64-bit on Mips.

Apr 17 2015, 7:24 AM

Apr 7 2015

earthdok added a comment to D8318: [sanitizer][MIPS] Implement clone for MIPS.

lgtm

Apr 7 2015, 10:24 AM
earthdok added a comment to D8802: [libc++] Fix PR22606 - Leak pthread_key with static storage duration to ensure all of thread-local destructors are called..

I'd still leave a short comment though.

Apr 7 2015, 6:02 AM

Apr 2 2015

earthdok added a comment to D8318: [sanitizer][MIPS] Implement clone for MIPS.

Hi, sorry for the delay.

Apr 2 2015, 9:25 AM

Mar 10 2015

earthdok added a comment to D7013: [LSan] [MIPS] adding support of LSan for mips64/mips64el arch.

Just a reminder:

Mar 10 2015, 7:57 AM

Mar 5 2015

earthdok added a comment to D8032: tsan: fix signal handling during stop-the-world.

Still kinda wary about this approach, but let's give it a try. LGTM

Mar 5 2015, 7:10 AM

Mar 4 2015

earthdok added a comment to D7827: [compiler-rt] Spring cleaning of compiler-rt symbolizers.

Just saw this CL.

Mar 4 2015, 12:18 PM
earthdok added a comment to D8032: tsan: fix signal handling during stop-the-world.

We need a comment in sanitizer_stoptheworld.h documenting the fact that the callback must exit by returning and not by calling exit() or _exit().

Mar 4 2015, 11:16 AM

Mar 3 2015

earthdok added inline comments to D8032: tsan: fix signal handling during stop-the-world.
Mar 3 2015, 8:30 AM
earthdok added inline comments to D8032: tsan: fix signal handling during stop-the-world.
Mar 3 2015, 7:09 AM

Mar 2 2015

earthdok added a comment to D8005: asan: fix signal handling during stoptheworld.

lgtm w/nit

Mar 2 2015, 9:09 AM

Feb 19 2015

earthdok added inline comments to D7723: tsan: fix PTRACE_ATTACH handling during stop-the-world.
Feb 19 2015, 5:20 AM

Feb 18 2015

earthdok added a comment to D7577: [LSan] Make parent tool responsible for initializing LSan flags..

lgtm

Feb 18 2015, 12:32 PM
earthdok added a comment to D7013: [LSan] [MIPS] adding support of LSan for mips64/mips64el arch.

Sorry, this fell off my radar. Do you need me to land this?

Feb 18 2015, 12:08 PM
earthdok added inline comments to D7723: tsan: fix PTRACE_ATTACH handling during stop-the-world.
Feb 18 2015, 10:33 AM
earthdok added inline comments to D7723: tsan: fix PTRACE_ATTACH handling during stop-the-world.
Feb 18 2015, 10:00 AM
earthdok added a comment to D7722: tsan: fix signal handling during stop-the-world.

The general approach LGTM

Feb 18 2015, 6:34 AM
earthdok added inline comments to D7722: tsan: fix signal handling during stop-the-world.
Feb 18 2015, 6:33 AM

Jan 28 2015

earthdok added reviewers for D7227: Make asan_symbolize.py not crash on Windows.: earthdok, glider.
Jan 28 2015, 8:55 AM
earthdok added a comment to D7227: Make asan_symbolize.py not crash on Windows..

Wouldn't it be better to do this in the Chromium-specific wrapper (src/tools/valgrind/asan_symbolize.py)?

Jan 28 2015, 8:55 AM

Jan 23 2015

earthdok accepted D7124: [LSan][MIPS64] Enable LSan testing for mips64/mips64el.

lgtm

Jan 23 2015, 12:00 PM
earthdok accepted D7013: [LSan] [MIPS] adding support of LSan for mips64/mips64el arch.

lgtm with nits

Jan 23 2015, 11:58 AM

Jan 16 2015

earthdok added inline comments to D7013: [LSan] [MIPS] adding support of LSan for mips64/mips64el arch.
Jan 16 2015, 1:12 PM

Jan 15 2015

earthdok added a comment to D5616: [lsan] [mips] adding support of lsan for mips64/mips64el arch.

The point of confusion here is whether the TCB is considered a part of TLS or not. From LSan's point of view, we have to include the TCB in the rootset. One reason is because thread-specific storage (i.e. pointers stored using pthread_setspecific()) is only reachable through a static array which is a part of the TCB structure. So LSan includes the TCB into the TLS.

Jan 15 2015, 10:08 AM

Jan 14 2015

earthdok added a comment to D5616: [lsan] [mips] adding support of lsan for mips64/mips64el arch.

Sorry about the long wait. I had a long vacation as well.

Jan 14 2015, 12:19 PM

Dec 18 2014

earthdok closed D6710: Refactor sanitizer build scripts..

Committed r224520

Dec 18 2014, 6:40 AM
earthdok added inline comments to D6710: Refactor sanitizer build scripts..
Dec 18 2014, 6:14 AM
earthdok updated the diff for D6710: Refactor sanitizer build scripts..
  • address nits
Dec 18 2014, 6:14 AM

Dec 17 2014

earthdok retitled D6710: Refactor sanitizer build scripts. from to Refactor sanitizer build scripts..
Dec 17 2014, 12:15 PM
earthdok closed D6689: [zorg] Add sanitizer-bootstrap-lite bot..
Dec 17 2014, 10:23 AM
earthdok accepted D6689: [zorg] Add sanitizer-bootstrap-lite bot..

Committed r224445

Dec 17 2014, 10:23 AM
earthdok updated the diff for D6689: [zorg] Add sanitizer-bootstrap-lite bot..
  • fix the "build stage1 clang" step
Dec 17 2014, 10:06 AM
earthdok added a comment to D6689: [zorg] Add sanitizer-bootstrap-lite bot..

No, in the first case the line is:
(cd clang_build && cmake -DCMAKE_BUILD_TYPE=RelWithDebInfo \

${CMAKE_COMMON_OPTIONS} $LLVM_CHECKOUT)
Dec 17 2014, 10:00 AM
earthdok added a comment to D6689: [zorg] Add sanitizer-bootstrap-lite bot..

Another thing I don't understand is how fatal and non-fatal steps are configured. I.e. here:

Dec 17 2014, 6:20 AM
earthdok updated the diff for D6689: [zorg] Add sanitizer-bootstrap-lite bot..
  • rename
  • add a missing bit
Dec 17 2014, 6:13 AM

Dec 16 2014

earthdok added inline comments to D6689: [zorg] Add sanitizer-bootstrap-lite bot..
Dec 16 2014, 4:21 PM
earthdok added inline comments to D6689: [zorg] Add sanitizer-bootstrap-lite bot..
Dec 16 2014, 12:17 PM
earthdok retitled D6689: [zorg] Add sanitizer-bootstrap-lite bot. from to [zorg] Add sanitizer-bootstrap-lite bot..
Dec 16 2014, 12:14 PM

Dec 8 2014

earthdok added a comment to D5616: [lsan] [mips] adding support of lsan for mips64/mips64el arch.
In D5616#76, @kumarsukhani wrote:

@earthdok: I am not sure if mips stores the value of pthread struct size anywhere the way x86 does to write a UNIT test for pthread struct size.
Btw why x86_64 is using the hard codes values, even though they have direct access to pthread structure size?

Dec 8 2014, 7:40 AM

Dec 5 2014

earthdok closed D704: Clear LD_PRELOAD when forking external symbolizer..
Dec 5 2014, 9:43 AM
earthdok closed D1025: Revert to C-style callbacks for iteration over allocator chunks. Also clean up LSan code, fix some comments and replace void* with uptr to bring down the number of reinterpret_casts..
Dec 5 2014, 9:43 AM
earthdok closed D1036: [lsan] Define interceptors more correctly. Also, always clear allocated memory..
Dec 5 2014, 9:43 AM
earthdok closed D1037: [tsan] Move some suppressions-related code to common. Factor out code to be reused in LSan. Also switch from linked list to vector..
Dec 5 2014, 9:43 AM
earthdok closed D1057: [sanitizer] Support padding with spaces in Printf..
Dec 5 2014, 9:43 AM
earthdok closed D1060: [lsan] Add suppression support..
Dec 5 2014, 9:41 AM
earthdok closed D1105: [lsan] Handle fork() correctly..
Dec 5 2014, 9:41 AM
earthdok closed D1193: [lsan] Make __lsan_do_leak_check() honor the detect_leaks flag..
Dec 5 2014, 9:41 AM
earthdok closed D1258: [lsan] Add leak_check_at_exit flag..
Dec 5 2014, 9:41 AM
earthdok closed D1473: [lsan] Add a stress test..
Dec 5 2014, 9:41 AM
earthdok closed D1496: [sanitizer] Add a fast version of StackDepotGet() for use in LSan. Add a class that holds a snapshot of the StackDepot optimized for querying by ID. This allows us to speed up LSan dramatically..
Dec 5 2014, 9:40 AM
earthdok closed D1515: [sanitizer] Handle Die() in StopTheWorld..
Dec 5 2014, 9:40 AM
earthdok closed D1558: [sanitizer_common] Add internal_clone()..
Dec 5 2014, 9:40 AM
earthdok closed D1578: [lsan] Colorize LSan reports..
Dec 5 2014, 9:39 AM
earthdok closed D1859: [sanitizer] Fix deadlock in StopTheWorld. The tracer thread in StopTheWorld could wait on a mutex forever if the parent process died before unlocking it. Use PR_SET_PDEATHSIG so that the parent would kill the child in this scenario..
Dec 5 2014, 9:39 AM
earthdok closed D1895: [sanitizer] Define kPthreadDestructorIterations on POSIX systems..
Dec 5 2014, 9:39 AM
earthdok closed D1896: [asan] Improve thread lifetime tracking on POSIX systems..
Dec 5 2014, 9:38 AM
earthdok closed D1901: [lsan] Support ASan's stack-use-after-return mode in LSan..
Dec 5 2014, 9:38 AM
earthdok closed D2221: [sanitizer] Expose __sanitizer_print_stack_trace()..
Dec 5 2014, 9:38 AM
earthdok closed D2417: [lsan] Introduce __lsan_(un)register_root_region()..
Dec 5 2014, 9:38 AM
earthdok closed D2629: [msan] Rewrite strto* interceptors and add a few more..
Dec 5 2014, 9:38 AM
earthdok closed D2642: [msan] Intercept *getxattr and *listxattr..
Dec 5 2014, 9:37 AM
earthdok closed D2659: [sanitizer] Add most ioctls from /usr/include/sound/..
Dec 5 2014, 9:37 AM
earthdok closed D6441: [ASan, LSan] Improve tracking of thread creation..
Dec 5 2014, 9:37 AM
earthdok closed D6544: [ASan] Refactor thread starting code..
Dec 5 2014, 9:32 AM
earthdok added a comment to D6544: [ASan] Refactor thread starting code..

Committed r223483

Dec 5 2014, 9:31 AM

Dec 4 2014

earthdok added inline comments to D6280: [compiler-rt] Implement AddressSanitizer suppressions.
Dec 4 2014, 5:09 PM
earthdok retitled D6544: [ASan] Refactor thread starting code. from to [ASan] Refactor thread starting code..
Dec 4 2014, 4:59 PM
earthdok added inline comments to D6280: [compiler-rt] Implement AddressSanitizer suppressions.
Dec 4 2014, 4:25 PM
earthdok added a comment to D6441: [ASan, LSan] Improve tracking of thread creation..

Committed r223419

Dec 4 2014, 4:12 PM
earthdok added a comment to D6441: [ASan, LSan] Improve tracking of thread creation..

BTW, I think the ThreadRegistry::CreateThread() call should be moved into AsanThread::Create(). This code is duplicated across 4 uses of AsanThread::Create(), and anyway most other ThreadRegistry access is done via AsanThread (including ThreadRegistry::FinishThread()).

Dec 4 2014, 3:32 PM
earthdok added inline comments to D6441: [ASan, LSan] Improve tracking of thread creation..
Dec 4 2014, 3:27 PM
earthdok updated the diff for D6441: [ASan, LSan] Improve tracking of thread creation..
  • add comment and use nullptrs
Dec 4 2014, 3:27 PM
earthdok updated the diff for D6441: [ASan, LSan] Improve tracking of thread creation..
  • struct
Dec 4 2014, 9:05 AM

Dec 3 2014

earthdok added inline comments to D6441: [ASan, LSan] Improve tracking of thread creation..
Dec 3 2014, 8:12 AM

Dec 2 2014

earthdok added inline comments to D6441: [ASan, LSan] Improve tracking of thread creation..
Dec 2 2014, 5:37 AM

Nov 27 2014

earthdok updated the diff for D6441: [ASan, LSan] Improve tracking of thread creation..
  • test
Nov 27 2014, 10:40 AM
earthdok added a comment to D6441: [ASan, LSan] Improve tracking of thread creation..

TSan does this.

Nov 27 2014, 10:28 AM
earthdok added a comment to D6441: [ASan, LSan] Improve tracking of thread creation..

The problem with the solution I proposed in http://llvm.org/bugs/show_bug.cgi?id=21621#c6 is that we don't know the thread's OS PID until ThreadStart() has been called. So, instead of checking whether the thread that interests us is in "created" state, LSan would have to process all "created" threads separately. That's not super complicated, but I like this solution better. At least it brings us one step closer to a unified pthread_create interceptor in ASan/MSan/TSan.

Nov 27 2014, 9:38 AM
earthdok retitled D6441: [ASan, LSan] Improve tracking of thread creation. from to [ASan, LSan] Improve tracking of thread creation..
Nov 27 2014, 9:28 AM

Nov 26 2014

earthdok added a comment to D6394: Simplify Symbolizer::SymbolizePC() interface..

LSan and sanitizer_common lgtm

Nov 26 2014, 11:56 AM
earthdok added inline comments to D6394: Simplify Symbolizer::SymbolizePC() interface..
Nov 26 2014, 5:40 AM
earthdok added a comment to D6396: [ASan] Fix leak tests on 64-bit targets other than x86-64 (take two).

lgtm

Nov 26 2014, 4:56 AM