DCHECK in ProcessMemoryDump::AddAllocatorDumpInternal while attempting to upload accumulated crash dumps on Android |
||||
Issue descriptionI saw this several times while investigating Nexus 5X GPU.FYI bot flakiness, but lost track of the other occurrences. Here is one I've got: https://build.chromium.org/p/tryserver.chromium.android/builders/android_optional_gpu_tests_rel/builds/759 I usually see in the log: I cr_MinidmpUploadService: Attempting to upload accumulated crash dumps. and then: F chromium: [FATAL:process_memory_dump.cc(183)] Check failed: insertion_result.second. Duplicate name: gpu/gl/renderbuffers/client_3/renderbuffer_1 Attached is the full log.
,
Nov 2 2016
primiano@, from "git blame" looks like you should be familiar with this code. Could you please take a look, maybe you will know why this is happening?
,
Nov 3 2016
Hmm that is because the GPU memory probe is creating duplicate entries. Ericrk IIRC you or somebody in your team did add the gpu/ memory dump provider? Can you help triaging this? Looks like we get duplicate gpu/gl/renderbuffers/client_X/renderbuffer_Y items.
,
Nov 4 2016
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/b03ea3889fdf345f5042a3112e2585266a951ebe commit b03ea3889fdf345f5042a3112e2585266a951ebe Author: ericrk <ericrk@chromium.org> Date: Fri Nov 04 23:08:42 2016 Use share group tracing ID for renderbuffer memory dump We previously used client id. This is not guaranteed to be unique. Use the share group tracing ID instead. BUG= 661436 CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel Review-Url: https://codereview.chromium.org/2473983003 Cr-Commit-Position: refs/heads/master@{#430058} [modify] https://crrev.com/b03ea3889fdf345f5042a3112e2585266a951ebe/gpu/command_buffer/service/renderbuffer_manager.cc
,
Nov 4 2016
Thanks a lot Eric! Marking this as fixed. Please reopen if this is not the case. |
||||
►
Sign in to add a comment |
||||
Comment 1 by ynovikov@chromium.org
, Nov 2 2016