1698524 - Clear memory counter each time we re-enable the profiler

Status: RESOLVED FIXED

(Core :: Gecko Profiler, defect, P2)

Description

[Randell Jesup [:jesup] (needinfo me)]

We noticed this issue a long time ago, but I never put the patch up for review. Resetting the memory counter on profiler restart makes far more sense.

Comment 1

[Randell Jesup [:jesup] (needinfo me)]

Attachment: Bug 1698524: Clear profiler memory counter on re-enabling of the profiler r=gerald

Comment 2

[BugBot [:suhaib / :marco/ :calixte]]

There's a r+ patch which didn't land and no activity in this bug for 2 weeks.
:jesup, could you have a look please?
For more information, please visit auto_nag documentation.

Comment 3

[Gerald Squelart (he/him) (not at Mozilla since 2022-09-15)]

Still r+ from me.

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/95e43f573edb
Clear profiler memory counter on re-enabling of the profiler r=gerald

Comment 5

[Natalia Csoregi [:nataliaCs]]

Backed out for causing assertion failure on ProfilerCounts.h

backout: https://hg.mozilla.org/integration/autoland/rev/dfdb488633e22d1dd19b869da177428263a7efbe

push: https://treeherder.mozilla.org/jobs?repo=autoland&revision=95e43f573edb8ddf1211ce8f4704dabf90876d0c&group_state=expanded&selectedTaskRun=DBiwnKmFQIGyRdS14IF7sg.0

failure log: https://treeherder.mozilla.org/logviewer?job_id=335471707&repo=autoland&lineNumber=5757

[task 2021-04-05T18:47:35.956Z] 18:47:35 INFO - GECKO(1536) | Assertion failure: aNumber >= mPrevNumber, at /builds/worker/workspace/obj-build/dist/include/mozilla/ProfilerCounts.h:108
[task 2021-04-05T18:47:35.956Z] 18:47:35 INFO - GECKO(1536) | #01: SamplerThread::Run() [tools/profiler/core/platform.cpp:3468]
[task 2021-04-05T18:47:35.956Z] 18:47:35 INFO - GECKO(1536) | #02: ThreadEntry(void*) [tools/profiler/core/platform-linux-android.cpp:460]
[task 2021-04-05T18:47:35.959Z] 18:47:35 INFO - GECKO(1536) | #03: ??? [/lib/x86_64-linux-gnu/libpthread.so.0 + 0x76db]
[task 2021-04-05T18:47:35.959Z] 18:47:35 INFO - GECKO(1536) | #04: clone [/lib/x86_64-linux-gnu/libc.so.6 + 0x121a3f]
[task 2021-04-05T18:47:35.960Z] 18:47:35 INFO - GECKO(1536) | #05: ??? (???:???)
[task 2021-04-05T18:47:36.016Z] 18:47:36 INFO - GECKO(1536) | [Parent 1536, Breakpad Server] WARNING: Resource acquired is being released in non-LIFO order; why?
[task 2021-04-05T18:47:36.016Z] 18:47:36 INFO - GECKO(1536) | : file /builds/worker/checkouts/gecko/xpcom/threads/BlockingResourceBase.cpp:293
[task 2021-04-05T18:47:36.017Z] 18:47:36 INFO - GECKO(1536) | --- Mutex : dumpSafetyLock (currently acquired)

Firstly noticed on : https://treeherder.mozilla.org/jobs?repo=autoland&searchStr=linux%2C18.04%2Cx64%2Cdebug%2Cmochitests%2Ctest-linux1804-64%2Fdebug-mochitest-devtools-chrome-e10s%2Cdt11&revision=af733fba4ae4a59de6b617109dbd0135cb28c7b3&selectedTaskRun=C0X92zi1RfCXjquBfEjO8w.0

Comment 6

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/78b33c83dcd6
Clear profiler memory counter on re-enabling of the profiler r=gerald

Comment 7

[Noemi Erli[:noemi_erli]]

Backed out changeset 78b33c83dcd6 (Bug 1698524) for causing assertion CLOSED TREE

Log: https://treeherder.mozilla.org/logviewer?job_id=335492986&repo=autoland&lineNumber=5438

Backout: https://hg.mozilla.org/integration/autoland/rev/91727265484ddabe292cf3b4d74bf3c24ecee207

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/d81407c4ae4b
Clear profiler memory counter on re-enabling of the profiler r=gerald

Comment 9

[Greg Tatum [:gregtatum]]

Is this a duplicate of Bug 1520587? It would be good to close the other one too. Also, there is a front-end mitigation that could be removed after this rides the trains.

https://github.com/firefox-devtools/profiler/blob/15f5bf5a9bae8965d0036084d1469bb9b0ba4ed8/src/profile-logic/profile-data.js#L1380-L1391

Comment 10

[Atila Butkovits]

https://hg.mozilla.org/mozilla-central/rev/d81407c4ae4b