New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 887543 link

Starred by 1 user

Issue metadata

Status: Duplicate
Merged: issue 877205
Owner: ----
Closed: Sep 20
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 3
Type: Bug

Blocking:
issue 887545



Sign in to add a comment

Unreadable callstack in FindIt's task reports

Project Member Reported by lukasza@chromium.org, Sep 20

Issue description

Callstacks in FindIt's task reports are unreadable:

- EXPECTED BEHAVIOR: logs from test runs include human-readable callstacks

- ACTUAL BEHAVIOR: callstacks in logs from test runs include only stack addresses


For example see:

- Overall FindIt report is here: https://findit-for-me.appspot.com/waterfall/flake/flake-culprit?key=ag9zfmZpbmRpdC1mb3ItbWVyQwsSDEZsYWtlQ3VscHJpdCIxY2hyb21pdW0vODBlYzhlMmJmYWZlNDllNzdiZTI5OTZlN2FhOWM3MTNjNjg1M2FiMww

- Test-specific FindIt report is here: https://findit-for-me.appspot.com/waterfall/flake?key=ag9zfmZpbmRpdC1mb3ItbWVy1QELEhdNYXN0ZXJGbGFrZUFuYWx5c2lzUm9vdCKeAWNocm9taXVtLm1lbW9yeS9MaW51eCBDaHJvbWl1bSBPUyBBU2FuIExTYW4gVGVzdHMgKDEpLzI4OTI2L3NpbmdsZV9wcm9jZXNzX21hc2hfYnJvd3Nlcl90ZXN0cy9VMlZ6YzJsdmJsSmxjM1J2Y21WVVpYTjBMazFsYlc5eWVWQnlaWE56ZFhKbFRHOWhaSE5PYjNSQmJHeFVZV0p6DAsSE01hc3RlckZsYWtlQW5hbHlzaXMYAQw

- Results of FindIt task report are here: https://chromium-swarm.appspot.com/task?id=3fea225e0ca53c10&refresh=10&show_raw=1

- The FindIt task report seems to include a crash, but the callstack is unreadable:

Command: ./browser_tests --brave-new-test-launcher --test-launcher-bot-mode --cfi-diag=0 --enable-features=SingleProcessMash --override-use-software-gl-for-tests --test-launcher-print-test-stdio=always --test-launcher-batch-limit=1 --test-launcher-summary-output=/b/s/w/ioulZK7K/output.json --gtest_filter=SessionRestoreTest.MemoryPressureLoadsNotAllTabs --gtest_repeat=20 --test-launcher-retry-limit=0 --gtest_also_run_disabled_tests --no-sandbox
...
BrowserTestBase received signal: Segmentation fault. Backtrace:
#0 0x000000f75c21 (/b/s/w/ir/out/Release/browser_tests+0xf75c20)
#1 0x0000112ea76c (/b/s/w/ir/out/Release/browser_tests+0x112ea76b)
#2 0x000012a5b2d6 (/b/s/w/ir/out/Release/browser_tests+0x12a5b2d5)
#3 0x7f87d4cdccb0 (/lib/x86_64-linux-gnu/libc-2.19.so+0x36caf)
#4 0x0000066e13c8 (/b/s/w/ir/out/Release/browser_tests+0x66e13c7)
#5 0x000011e830f1 (/b/s/w/ir/out/Release/browser_tests+0x11e830f0)
#6 0x0000123fb875 (/b/s/w/ir/out/Release/browser_tests+0x123fb874)
#7 0x000012621811 (/b/s/w/ir/out/Release/browser_tests+0x12621810)
#8 0x00001261ea45 (/b/s/w/ir/out/Release/browser_tests+0x1261ea44)
#9 0x00001262334a (/b/s/w/ir/out/Release/browser_tests+0x12623349)
#10 0x0000059cef59 (/b/s/w/ir/out/Release/browser_tests+0x59cef58)
#11 0x0000059cf2d5 (/b/s/w/ir/out/Release/browser_tests+0x59cf2d4)
#12 0x00001123a928 (/b/s/w/ir/out/Release/browser_tests+0x1123a927)
#13 0x00001123ab55 (/b/s/w/ir/out/Release/browser_tests+0x1123ab54)
#14 0x000011350d65 (/b/s/w/ir/out/Release/browser_tests+0x11350d64)
#15 0x0000110bea86 (/b/s/w/ir/out/Release/browser_tests+0x110bea85)
#16 0x0000110bf779 (/b/s/w/ir/out/Release/browser_tests+0x110bf778)
#17 0x00001133c561 (/b/s/w/ir/out/Release/browser_tests+0x1133c560)
#18 0x0000110bd8e5 (/b/s/w/ir/out/Release/browser_tests+0x110bd8e4)
#19 0x00001115b89e (/b/s/w/ir/out/Release/browser_tests+0x1115b89d)
#20 0x0000114cc7ef (/b/s/w/ir/out/Release/browser_tests+0x114cc7ee)
#21 0x0000114cc3a8 (/b/s/w/ir/out/Release/browser_tests+0x114cc3a7)
#22 0x000012a58e24 (/b/s/w/ir/out/Release/browser_tests+0x12a58e23)
#23 0x00001163d536 (/b/s/w/ir/out/Release/browser_tests+0x1163d535)
#24 0x00001163a67f (/b/s/w/ir/out/Release/browser_tests+0x1163a67e)
#25 0x000005d0551c (/b/s/w/ir/out/Release/browser_tests+0x5d0551b)
#26 0x00000a6af748 (/b/s/w/ir/out/Release/browser_tests+0xa6af747)
#27 0x00000b998b8c (/b/s/w/ir/out/Release/browser_tests+0xb998b8b)
#28 0x00000a6aaf0f (/b/s/w/ir/out/Release/browser_tests+0xa6aaf0e)
#29 0x00000a6b8067 (/b/s/w/ir/out/Release/browser_tests+0xa6b8066)
#30 0x00000a6a3a80 (/b/s/w/ir/out/Release/browser_tests+0xa6a3a7f)
#31 0x000010ac8175 (/b/s/w/ir/out/Release/browser_tests+0x10ac8174)
#32 0x000018e9e223 (/b/s/w/ir/out/Release/browser_tests+0x18e9e222)
#33 0x000010ac3013 (/b/s/w/ir/out/Release/browser_tests+0x10ac3012)
#34 0x000012a57518 (/b/s/w/ir/out/Release/browser_tests+0x12a57517)
#35 0x0000114c6be6 (/b/s/w/ir/out/Release/browser_tests+0x114c6be5)
#36 0x00000704ab03 (/b/s/w/ir/out/Release/browser_tests+0x704ab02)
#37 0x00000704ce55 (/b/s/w/ir/out/Release/browser_tests+0x704ce54)
#38 0x00000704e227 (/b/s/w/ir/out/Release/browser_tests+0x704e226)
#39 0x000007075c97 (/b/s/w/ir/out/Release/browser_tests+0x7075c96)
#40 0x000007074ee6 (/b/s/w/ir/out/Release/browser_tests+0x7074ee5)
#41 0x000011519fdb (/b/s/w/ir/out/Release/browser_tests+0x11519fda)
#42 0x0000110282d2 (/b/s/w/ir/out/Release/browser_tests+0x110282d1)
#43 0x000012aebb06 (/b/s/w/ir/out/Release/browser_tests+0x12aebb05)
#44 0x0000110290d7 (/b/s/w/ir/out/Release/browser_tests+0x110290d6)
#45 0x000011027def (/b/s/w/ir/out/Release/browser_tests+0x11027dee)
#46 0x7f87d4cc7f45 (/lib/x86_64-linux-gnu/libc-2.19.so+0x21f44)
#47 0x000000f286ea (/b/s/w/ir/out/Release/browser_tests+0xf286e9)




 
Cc: dpranke@google.com thakis@chromium.org
Components: Tools>Test>FindIt Internals>Logging Infra
+thakis@ and +dpranke@ who I hope can route this bug to the right team (not sure if this is a problem with the compiler? test code? infra?)
Blocking: 887545
Components: -Infra Infra>Client>Chrome Build
The original failure on the waterfall was not symbolized either https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Linux%20Chromium%20OS%20ASan%20LSan%20Tests%20%281%29/28926

A build setup problem?
No, this appears to be intentional: https://bugs.chromium.org/p/chromium/issues/detail?id=877205
Mergedinto: 877205
Status: Duplicate (was: Untriaged)
re #c1: unfortunately, client-side support for crash stacks is kinda unowned in Chromium these days. In theory we have owners for crashpad itself, but not really owners for making sure sure that every test binary produces usable stacks.

I think this is something yihongg's team is hoping to address in 2019 if we get additional staffing.

Sign in to add a comment