Bug 234356 - REGRESSION (r287068): [macOS arm64] TestWebKitAPI.GraphicsContextGLCocoaTest.UnrecycledDisplayBuffersNoLeaks is failing
Summary: REGRESSION (r287068): [macOS arm64] TestWebKitAPI.GraphicsContextGLCocoaTest....
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: ANGLE (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Kimmo Kinnunen
URL:
Keywords: InRadar
Depends on: 233328 234317
Blocks:
  Show dependency tree / graph
 
Reported: 2021-12-15 11:21 PST by Ryan Haddad
Modified: 2021-12-17 16:36 PST (History)
10 users (show)

See Also:

Attachments
Patch (4.14 KB, patch)
2021-12-16 05:37 PST, Kimmo Kinnunen 		no flags Details | Formatted Diff | Diff
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Ryan Haddad 2021-12-15 11:21:37 PST 
TestWebKitAPI.GraphicsContextGLCocoaTest.UnrecycledDisplayBuffersNoLeaks is failing on arm64 macOS bots after https://trac.webkit.org/changeset/287068/webkit, x86 bots are not affected.

    TestWebKitAPI.GraphicsContextGLCocoaTest.UnrecycledDisplayBuffersNoLeaks
        
        /Volumes/Data/worker/monterey-release/build/Tools/TestWebKitAPI/Tests/WebCore/cocoa/TestGraphicsContextGLCocoa.mm:242
        Value of: memoryFootprintChangedBy(lastFootprint, footprintChange, footprintError)
          Actual: false (Footprint changed by 827590144. Expected at most 0+-157286400)
        Expected: true

https://results.webkit.org/?suite=api-tests&test=TestWebKitAPI.GraphicsContextGLCocoaTest.UnrecycledDisplayBuffersNoLeaks
Comment 1 Radar WebKit Bug Importer 2021-12-15 11:21:49 PST 
<rdar://problem/86533892>
Comment 2 Kenneth Russell 2021-12-15 12:00:22 PST 
To be clear: that commit was the ANGLE roll to 2021-12-14 in Bug 234317.

We'd strongly prefer to not revert that roll. Is there any indication what might have caused this?
Comment 3 Ryan Haddad 2021-12-15 12:41:28 PST 
(In reply to Kenneth Russell from comment #2)
> To be clear: that commit was the ANGLE roll to 2021-12-14 in Bug 234317.
> 
> We'd strongly prefer to not revert that roll. Is there any indication what
> might have caused this?
Kyle is investigating.
Comment 4 Kimmo Kinnunen 2021-12-15 12:43:02 PST 
I cannot attend to it until tomorrow. If it's a problem, naming it DISABLED_Unrecycled.. would skip it temporarily
Comment 5 Kimmo Kinnunen 2021-12-16 05:37:51 PST 
Created attachment 447348 [details]
Patch
Comment 6 EWS Watchlist 2021-12-16 05:39:14 PST 
Note that there are important steps to take when updating ANGLE. See https://trac.webkit.org/wiki/UpdatingANGLE
Comment 7 Kyle Piddington 2021-12-16 14:50:54 PST 
Comment on attachment 447348 [details]
Patch

LGTM, great catch.
Comment 8 Kenneth Russell 2021-12-16 18:38:50 PST 
Amazing work Kimmo tracking this down! r+ from me as well.
Comment 9 Kenneth Russell 2021-12-16 18:39:13 PST 
Filed https://bugs.chromium.org/p/angleproject/issues/detail?id=6829 about upstreaming this to ANGLE. I'll help with that.
Comment 10 EWS 2021-12-17 03:18:13 PST 
kpiddington@apple.com does not have reviewer permissions according to https://raw.githubusercontent.com/WebKit/WebKit/main/metadata/contributors.json.

Rejecting attachment 447348 [details] from commit queue.
Comment 11 Antti Koivisto 2021-12-17 03:21:06 PST 
Comment on attachment 447348 [details]
Patch

rs=me, actual review by Kyle Piddington
Comment 12 EWS 2021-12-17 04:00:36 PST 
Committed r287178 (245347@main): <https://commits.webkit.org/245347@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 447348 [details].
Comment 13 Kenneth Russell 2021-12-17 16:36:47 PST 
For bookkeeping purposes, the new test was added under Bug 233328.