CrWinAsan(dll) has been failing to compile for a while now |
|||
Issue descriptionStarted here: https://build.chromium.org/p/chromium.fyi/builders/CrWinAsan%28dll%29/builds/5251 (Nov 17) [4721/45555] ACTION //components/resources:compressed_about_credits(//build/toolchain/win:clang_x86) FAILED: gen/components/resources/about_credits.bro C:/b/depot_tools/python276_bin/python.exe ../../build/gn_run_binary.py bro.exe --force --input gen/components/resources/about_credits.html --output gen/components/resources/about_credits.bro ==3776==AddressSanitizer CHECK failed: C:\b\c\b\CrWinAsan_dll_\src\third_party\llvm\projects\compiler-rt\lib\asan\asan_globals.cc:195 "((g->beg)) != (0)" (0x0, 0x0) #0 0x6b2a9e59 (C:\b\c\b\CrWinAsan_dll_\src\out\Release\clang_rt.asan_dynamic-i386.dll+0x10029e59) #1 0x6b28e2e3 (C:\b\c\b\CrWinAsan_dll_\src\out\Release\clang_rt.asan_dynamic-i386.dll+0x1000e2e3) #2 0x6b2a0013 (C:\b\c\b\CrWinAsan_dll_\src\out\Release\clang_rt.asan_dynamic-i386.dll+0x10020013) #3 0x6b2a0975 (C:\b\c\b\CrWinAsan_dll_\src\out\Release\clang_rt.asan_dynamic-i386.dll+0x10020975) #4 0x98a1b7 (C:\b\c\b\CrWinAsan_dll_\src\out\Release\bro.exe+0x4ba1b7) #5 0x98a3bb (C:\b\c\b\CrWinAsan_dll_\src\out\Release\bro.exe+0x4ba3bb) #6 0x75673389 (C:\Windows\syswow64\kernel32.dll+0x7dd73389) #7 0x77a09a01 (C:\Windows\SysWOW64\ntdll.dll+0x7dea9a01) #8 0x77a099d4 (C:\Windows\SysWOW64\ntdll.dll+0x7dea99d4) rnk, likely your r287246? Do we need to revert our roll over this? Does clusterfuzz do component builds?
,
Nov 23 2016
r287780 should fix the problem, but I'm not sure how much we can rely on the 256 byte padding heuristic.
,
Nov 24 2016
It now crashes elsewhere:
FAILED: gen/components/resources/about_credits.bro
C:/b/depot_tools/python276_bin/python.exe ../../build/gn_run_binary.py bro.exe --force --input gen/components/resources/about_credits.html --output gen/components/resources/about_credits.bro
==492==AddressSanitizer CHECK failed: C:\b\c\b\CrWinAsan_dll_\src\third_party\llvm\projects\compiler-rt\lib\asan\asan_globals.cc:362 "((globals[i].size == 0 && globals[i].size_with_redzone == 0 && globals[i].name == nullptr && globals[i].module_name == nullptr && globals[i].odr_indicator == 0)) != (0)" (0x0, 0x0)
#0 0x6a799e79 (C:\b\c\b\CrWinAsan_dll_\src\out\Release\clang_rt.asan_dynamic-i386.dll+0x10029e79)
#1 0x6a77e2e3 (C:\b\c\b\CrWinAsan_dll_\src\out\Release\clang_rt.asan_dynamic-i386.dll+0x1000e2e3)
#2 0x6a7909ab (C:\b\c\b\CrWinAsan_dll_\src\out\Release\clang_rt.asan_dynamic-i386.dll+0x100209ab)
#3 0x14893a7 (C:\b\c\b\CrWinAsan_dll_\src\out\Release\bro.exe+0x4b93a7)
#4 0x14895ab (C:\b\c\b\CrWinAsan_dll_\src\out\Release\bro.exe+0x4b95ab)
#5 0x75933389 (C:\Windows\syswow64\kernel32.dll+0x7dd73389)
#6 0x77ca9a01 (C:\Windows\SysWOW64\ntdll.dll+0x7dea9a01)
#7 0x77ca99d4 (C:\Windows\SysWOW64\ntdll.dll+0x7dea99d4)
,
Nov 24 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/7dd55ef76fd36f1840677c3277ee0e598a50806a commit 7dd55ef76fd36f1840677c3277ee0e598a50806a Author: thakis <thakis@chromium.org> Date: Thu Nov 24 20:08:24 2016 Roll clang 287685:287780. Ran `tools/clang/scripts/upload_revision.py 287780`. Quick follow-up to https://codereview.chromium.org/2523073002, picks up fixes for a PCH crash and for asan x86 dll builds. BUG= 667891 , 668212 TBR=rnk Review-Url: https://codereview.chromium.org/2528733002 Cr-Commit-Position: refs/heads/master@{#434386} [modify] https://crrev.com/7dd55ef76fd36f1840677c3277ee0e598a50806a/tools/clang/scripts/update.py
,
Nov 29 2016
Let's try r288096.
,
Nov 30 2016
The builder went green when it got that revision in https://build.chromium.org/p/chromium.fyi/builders/CrWinAsan%28dll%29/builds/5301. However, the tester is purple because browser_tests isn't running. We probably need to copy the ASan runtime DLL around or something.
,
Nov 30 2016
Looks like HID.DLL is in the way of the shadow memory mapping. We have an issue that we closed as stale for that ( http://crbug.com/515604 ). Let's close this and reopen that.
,
Dec 1 2016
|
|||
►
Sign in to add a comment |
|||
Comment 1 by r...@chromium.org
, Nov 23 2016