Intermittent SUMMARY: ThreadSanitizer: data race /builds/worker/checkouts/gecko/netwerk/protocol/http/nsHttpHandler.cpp:1975:27 in mozilla::net::nsHttpHandler::PrefsChanged(char const*)
Categories
(Core :: DOM: Device Interfaces, defect, P5)
Tracking
()
Tracking | Status | |
---|---|---|
firefox86 | --- | fixed |
People
(Reporter: intermittent-bug-filer, Assigned: KrisWright)
References
(Blocks 1 open bug)
Details
(Keywords: intermittent-failure)
Attachments
(4 files)
Filed by: rmaries [at] mozilla.com
Parsed log: https://treeherder.mozilla.org/logviewer.html#?job_id=306295631&repo=autoland
Full log: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/XGXFzS9HR2KxhcKoH3RhfQ/runs/0/artifacts/public/logs/live_backing.log
[task 2020-06-14T15:26:15.183Z] 15:26:15 INFO - TEST-START | dom/credentialmanagement/tests/mochitest/test_credman_iframes.html
[task 2020-06-14T15:26:18.364Z] 15:26:18 INFO - GECKO(1612) | #39 do_main /builds/worker/checkouts/gecko/browser/app/nsBrowserApp.cpp:217:22 (firefox+0xca433)
[task 2020-06-14T15:26:18.364Z] 15:26:18 INFO - GECKO(1612) | #40 main /builds/worker/checkouts/gecko/browser/app/nsBrowserApp.cpp:331:16 (firefox+0xca433)
[task 2020-06-14T15:26:18.365Z] 15:26:18 INFO - GECKO(1612) | SUMMARY: ThreadSanitizer: data race /builds/worker/checkouts/gecko/netwerk/protocol/http/nsHttpHandler.cpp:1975:27 in mozilla::net::nsHttpHandler::PrefsChanged(char const*)
[task 2020-06-14T15:26:18.365Z] 15:26:18 INFO - GECKO(1612) | ==================
[task 2020-06-14T15:26:18.366Z] 15:26:18 INFO - GECKO(1612) | Crash Annotation GraphicsCriticalError: |[C0][GFX1-]: Receive IPC close with reason=AbnormalShutdown (t=11.3857) [GFX1-]: Receive IPC close with reason=AbnormalShutdown
[task 2020-06-14T15:26:18.367Z] 15:26:18 INFO - GECKO(1612) | Exiting due to channel error.
[task 2020-06-14T15:26:18.367Z] 15:26:18 INFO - GECKO(1612) | Crash Annotation GraphicsCriticalError: |[C0][GFX1-]: Receive IPC close with reason=AbnormalShutdown (t=14.1084) [GFX1-]: Receive IPC close with reason=AbnormalShutdown
[task 2020-06-14T15:26:18.367Z] 15:26:18 INFO - GECKO(1612) | Exiting due to channel error.
[task 2020-06-14T15:26:18.368Z] 15:26:18 INFO - GECKO(1612) | Crash Annotation GraphicsCriticalError: |[C0][GFX1-]: Receive IPC close with reason=AbnormalShutdown (t=25.7987) [GFX1-]: Receive IPC close with reason=AbnormalShutdown
[task 2020-06-14T15:26:18.368Z] 15:26:18 INFO - GECKO(1612) | Exiting due to channel error.
[task 2020-06-14T15:26:18.369Z] 15:26:18 INFO - GECKO(1612) | Exiting due to channel error.
[task 2020-06-14T15:26:19.378Z] 15:26:19 INFO - GECKO(1612) | Crash Annotation GraphicsCriticalError: |[C0][GFX1-]: Receive IPC close with reason=AbnormalShutdown (t=24.8695)
[task 2020-06-14T15:26:19.413Z] 15:26:19 INFO - TEST-INFO | Main app process: killed by SIGIOT
Comment hidden (Intermittent Failures Robot) |
Comment 2•4 years ago
|
||
https://wiki.mozilla.org/Bug_Triage#Intermittent_Test_Failure_Cleanup
For more information, please visit auto_nag documentation.
Comment 3•4 years ago
|
||
recent failure: https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=311449432&repo=autoland&lineNumber=2947
Comment hidden (Intermittent Failures Robot) |
Comment 5•4 years ago
|
||
Comment 6•4 years ago
|
||
Assignee | ||
Comment 7•4 years ago
|
||
This is specifically caused by the bool mResponseTimeoutEnabled, which we are reading on the socket thread during nsHTTPConnection::Activate. We're always going to call PrefsChanged
on the main thread, so to avoid racing with the socket thread we probably ought to make mResponseTimeoutEnabled
atomic or something. I haven't checked to see if anything else being read off that prefs list would have the same issue.
Comment 8•4 years ago
|
||
Gonna suppress this as I think Bug 1683730 makes it worse. That said, I think we should also send the naive fix to the netwerk team's desk and see what they think.
Comment 9•4 years ago
|
||
Depends on D100268
Updated•4 years ago
|
Updated•4 years ago
|
Comment 10•4 years ago
|
||
Comment 11•4 years ago
|
||
bugherder |
Updated•4 years ago
|
Assignee | ||
Comment 12•4 years ago
|
||
It appears two values are accessed offthread in nsHttpConnection
. While it is unlikely that UpdateMirror
and nsHttpConnection::Activate
would ever overlap, the possibility is nonzero and could potentially lead to a race.
Assignee | ||
Comment 13•4 years ago
|
||
Assignee | ||
Comment 14•4 years ago
|
||
I'm going to go ahead and land this, and continue to leave the bug open until I remove the suppression.
Comment 15•4 years ago
|
||
Comment 16•4 years ago
|
||
bugherder |
Assignee | ||
Updated•4 years ago
|
Comment 17•4 years ago
|
||
Updated•4 years ago
|
Comment 18•4 years ago
|
||
bugherder |
Description
•