Intermittent headless-spi SUMMARY: ThreadSanitizer: data race /builds/worker/checkouts/gecko/security/nss/lib/pki/pki3hack.c:750:15 in fill_CERTCertificateFields
Categories
(Core :: Security: PSM, defect)
Tracking
()
| Tracking | Status | |
|---|---|---|
| firefox111 | --- | fixed |
People
(Reporter: intermittent-bug-filer, Assigned: keeler)
References
(Blocks 2 open bugs)
Details
(Keywords: csectype-race, intermittent-failure)
Attachments
(3 files)
Filed by: rvandermeulen [at] mozilla.com
Parsed log: https://treeherder.mozilla.org/logviewer?job_id=401517600&repo=try
Full log: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/DRzPh0-DSB2taJPxRenN2g/runs/0/artifacts/public/logs/live_backing.log
Comment 1•3 years ago
•
|
||
This hits about 50% of the time when running mochitest-plain in headless-spi (socketprocess) mode (--headless --setpref=network.process.enabled=true --setpref=network.http.network_access_on_socket_process.enabled=true) under TSAN. Seems to hit consistently when running dom/base/test/test_bug466080.html. This isn't currently enabled by default in CI, but I've been running Try pushes in the hopes of getting it green enough to do so.
Log link:
https://treeherder.mozilla.org/logviewer?job_id=401496165&repo=try&lineNumber=3120
Pernosco trace:
https://pernos.co/debug/m_IHUARELHepEATm7BITwQ/index.html#f{m[A98G,AV0_,t[CQ,CFU_,f{e[A8zk,zt4_,s{afv/yALAA,bAdc,u7As,o8ao___/
Comment 2•3 years ago
|
||
Updated•3 years ago
|
Comment 3•3 years ago
|
||
This test does change the value of the pref network.cors_preflight.allow_client_cert. I don't know if that might be causing the race or not.
Comment 4•3 years ago
|
||
Looks like FindClientCertificatesWithPrivateKeys can create CERTCertificates in some cases, and it is being called off the socket thread in SelectClientAuthCertificate::DoSelectClientAuthCertificate. Arguably an NSS bug, but I think our current approach is to fix this sort of issue by dispatching to the socket thread, so I'm moving this to PSM. I don't think it's a security issue---there's no attacker-controllable data and the two threads should write identical data to the CERTCertificate's fields.
Updated•3 years ago
|
| Comment hidden (Intermittent Failures Robot) |
| Assignee | ||
Comment 6•3 years ago
|
||
Updated•3 years ago
|
Comment 8•3 years ago
|
||
| bugherder | ||
Comment 9•3 years ago
|
||
This appears to still be happening on m-c tip:
https://treeherder.mozilla.org/logviewer?job_id=402780486&repo=try&lineNumber=3131
| Assignee | ||
Comment 10•3 years ago
|
||
| Comment hidden (Intermittent Failures Robot) |
Comment 12•3 years ago
|
||
Comment 13•3 years ago
|
||
| bugherder | ||
| Assignee | ||
Updated•3 years ago
|
Description
•