Issue metadata
Sign in to add a comment
|
browser_tests on Ubuntu-12.04 failing on chromium.linux/Linux Tests (dbg)(1) |
||||||||||||||||||||||
Issue descriptionbrowser_tests on Ubuntu-12.04 failing on chromium.linux/Linux Tests (dbg)(1) Type: build-failure Builders failed on: - Linux Tests (dbg)(1): https://build.chromium.org/p/chromium.linux/builders/Linux%20Tests%20%28dbg%29%281%29
,
Nov 16 2016
Bisecting... OK 432328 4f55b637f4b2f0dc5887b1e35b6ca7f7788c1701 NG 432358 862c7245d9b3f295df0db82f6e17c730d3f54825
,
Nov 16 2016
humm.. I can't reproduce the error on my local machine.
,
Nov 16 2016
,
Nov 16 2016
I found strange stack trace. https://luci-logdog.appspot.com/v/?s=chromium%2Fbb%2Fchromium.linux%2FLinux_Tests__dbg__1_%2F58836%2F%2B%2Frecipes%2Fsteps%2Fbrowser_side_navigation_browser_tests_on_Ubuntu-12.04%2F0%2Flogs%2FPasswordAutofillAgentTest.PreviewSuggestionIfUsernameReadOnly%2F0 PasswordAutofillAgentTest.PreviewSuggestionIfUsernameReadOnly (run #1): [ RUN ] PasswordAutofillAgentTest.PreviewSuggestionIfUsernameReadOnly Received signal 11 SEGV_MAPERR 0000000004e2 #0 0x7fa1db4e896e base::debug::StackTrace::StackTrace() #1 0x7fa1db4e84af base::debug::(anonymous namespace)::StackDumpSignalHandler() #2 0x7fa1db933cb0 <unknown> #3 0x000000e2d8c8 logging::CheckEQImpl<>() #4 0x00000104cca2 base::Pickle::headerT<>() #5 0x00000104cc45 IPC::Message::header() #6 0x00000384bb35 IPC::Message::is_reply() #7 0x000003849f01 content::MockRenderThread::Send() #8 0x7fa1d42cf8af content::RenderFrameImpl::Send() #9 0x7fa1d25c6ea2 content::RenderFrameObserver::Send() #10 0x00000140952d autofill::PasswordAutofillAgentTest::SimulateOnFillPasswordForm() #11 0x0000013fc19b autofill::PasswordAutofillAgentTest_PreviewSuggestionIfUsernameReadOnly_Test::TestBody() #12 0x000001051e8e _ZN7testing8internal12InvokeHelperIvSt5tupleIJEEE12InvokeMethodI14ServiceProcessMS6_FvvEEEvPT_T0_RKS3_ #13 0x000003999532 testing::internal::HandleExceptionsInMethodIfSupported<>() #14 0x00000398e3c6 testing::Test::Run() #15 0x00000398eb7d testing::TestInfo::Run() #16 0x00000398f11f testing::TestCase::Run() #17 0x00000399446c testing::internal::UnitTestImpl::RunAllTests() #18 0x0000010514be _ZN7testing8internal12InvokeHelperIbSt5tupleIJEEE12InvokeMethodI20MockServiceIPCServerMS6_FbvEEEbPT_T0_RKS3_ #19 0x00000399abe2 testing::internal::HandleExceptionsInMethodIfSupported<>() #20 0x00000399410f testing::UnitTest::Run() #21 0x0000026f2f51 RUN_ALL_TESTS() #22 0x0000026f1e62 base::TestSuite::Run() #23 0x000002699260 ChromeTestSuiteRunner::RunTestSuite() #24 0x000002699359 ChromeTestLauncherDelegate::RunTestSuite() #25 0x00000387b882 content::LaunchTests() #26 0x0000026996e9 LaunchChromeTests() #27 0x000002699185 main #28 0x7fa1c33677ed __libc_start_main #29 0x0000007906a5 <unknown> r8: 0000000000000038 r9: 0101010101010101 r10: bd62c998560c8b00 r11: 00007fa1c33d0c52 r12: 000000000079067c r13: 00007ffcfb5cd5d0 r14: 0000000000000000 r15: 0000000000000000 di: 00000000000004e2 si: 00007ffcfb5cbfc0 bp: 00007ffcfb5cbf90 bx: 0000000000000000 dx: 00000000000004e2 ax: 00000000000004d2 cx: bd62c998560c8b00 sp: 00007ffcfb5cbf70 ip: 0000000000e2d8c8 efl: 0000000000010202 cgf: bd62000000000033 erf: 0000000000000004 trp: 000000000000000e msk: 0000000000000000 cr2: 00000000000004e2 [end of stack trace] I think PasswordAutofillAgentTest::SimulateOnFillPasswordForm() doesn't call RenderFrameObserver::Send().
,
Nov 16 2016
It looks like caused by memory corruption. I'm trying to reproduce using ASAN.
,
Nov 16 2016
Thanks, horo@, indeed, this looks like corrupted vtable of password_autofill_agent_, perhaps it has been destroyed? vasilii@, would you find time to have a look at this? I'm afraid I'm unable to until this week's training is over. The blame range is https://chromium.googlesource.com/chromium/src/+log/4f55b637f4b2f0dc5887b1e35b6ca7f7788c1701..862c7245d9b3f295df0db82f6e17c730d3f54825
,
Nov 16 2016
Also note that this has something to do with site isolation -- the build step is not "browser_test" here, but site_per_process_browser_tests (not sure which build settings this corresponds to, though). Also, attaching the full log in case it disappears from the bot. It's a shame that Findit throws HTTP 500 for this: https://findit-for-me.appspot.com/waterfall/build-failure?url=https%3A%2F%2Fbuild.chromium.org%2Fp%2Fchromium.linux%2Fbuilders%2FLinux%2520Tests%2520%2528dbg%2529%25281%2529%2Fbuilds%2F58836
,
Nov 16 2016
I hit the following violation which was already reported in Issue 660479 . I think it is not related to this bug. ================================================================= ==17643==ERROR: AddressSanitizer: odr-violation (0x7f84619230e0): [1] size=37 'blink::mojom::blink::OffscreenCanvasSurface::Name_' gen/third_party/WebKit/public/platform/modules/offscreencanvas/offscreen_canvas_surface.mojom-blink.cc:58:36 [2] size=37 'blink::mojom::blink::OffscreenCanvasSurface::Name_' gen/third_party/WebKit/public/platform/modules/offscreencanvas/offscreen_canvas_surface.mojom-blink.cc:58:36 These globals were registered at these points: [1]: #0 0x920537 (/usr/local/google/home/horo/chromium/src/out/asan/browser_tests+0x920537) #1 0x7f84593c91dd (/usr/local/google/home/horo/chromium/src/out/asan/./libblink_core.so+0x701c1dd) [2]: #0 0x920537 (/usr/local/google/home/horo/chromium/src/out/asan/browser_tests+0x920537) #1 0x7f8460bfa83d (/usr/local/google/home/horo/chromium/src/out/asan/./libblink_platform.so+0x148783d) ==17643==HINT: if you don't care about these errors you may set ASAN_OPTIONS=detect_odr_violation=0 SUMMARY: AddressSanitizer: odr-violation: global 'blink::mojom::blink::OffscreenCanvasSurface::Name_' at gen/third_party/WebKit/public/platform/modules/offscreencanvas/offscreen_canvas_surface.mojom-blink.cc:58:36 ==17643==ABORTING
,
Nov 16 2016
The failure has gone after build #58842. https://build.chromium.org/p/chromium.linux/builders/Linux%20Tests%20%28dbg%29%281%29
,
Nov 17 2016
,
Nov 17 2016
Thanks for the additional information, horo@. Looking at the fixing build's blame range, I don't see which of the many CLs could fix this. I'm marking this as WontFix as we are not sure what the fix was and whether this is fixed for good. If we ever need to restart investigation, here are some initial pointers: Blamelist which introduced the failure: https://chromium.googlesource.com/chromium/src/+log/4f55b637f4b2f0dc5887b1e35b6ca7f7788c1701..862c7245d9b3f295df0db82f6e17c730d3f54825 Blamelist which made the failure disappear: https://chromium.googlesource.com/chromium/src/+log/ff0c1a487b941b11478790642d92e94a1e765248%5E..73a9f66a9a89f66628ea6554f33750f3a17fd823 My observation in #8 about site isolation is a red herring, it failed in browser_tests step also (see attached log from the last failing build).
,
Nov 29
|
|||||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||||
Comment 1 by horo@chromium.org
, Nov 16 2016