Bug 224854 - ANGLE should be compiled with thread-safe statics
Summary: ANGLE should be compiled with thread-safe statics
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: ANGLE (show other bugs)
Version: Other
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Kimmo Kinnunen
URL:
Keywords: InRadar
Depends on:
Blocks:
 
Reported: 2021-04-20 23:11 PDT by Kimmo Kinnunen
Modified: 2021-04-28 00:00 PDT (History)
8 users (show)

See Also:

Attachments
Patch (1.56 KB, patch)
2021-04-20 23:33 PDT, 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 Kimmo Kinnunen 2021-04-20 23:11:28 PDT 
ANGLE should be compiled with thread-safe statics
The code assumes that, as that's the c++ semantics and ANGLE is a c++ project.
Comment 1 Kimmo Kinnunen 2021-04-20 23:33:32 PDT 
Created attachment 426651 [details]
Patch
Comment 2 EWS Watchlist 2021-04-20 23:34:35 PDT 
Note that there are important steps to take when updating ANGLE. See https://trac.webkit.org/wiki/UpdatingANGLE
Comment 3 Kenneth Russell 2021-04-23 14:15:06 PDT 
Comment on attachment 426651 [details]
Patch

Seems fine. r+
Comment 4 Darin Adler 2021-04-25 13:18:07 PDT 
Comment on attachment 426651 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=426651&action=review

> Source/ThirdParty/ANGLE/ChangeLog:11
> +        Compile ANGLE with normal thread-safe c++ local statics.
> +        No known threading failures. Based on ad-hoc risk vs benefit
> +        evaluation of today and after future merges, it appears
> +        better to err in the side of caution.

Excellent that we can/will do this.

This comment addresses the safety side of the tradeoff. What about the performance side?
Comment 5 Radar WebKit Bug Importer 2021-04-27 23:12:13 PDT 
<rdar://problem/77250303>
Comment 6 Kimmo Kinnunen 2021-04-27 23:49:35 PDT 
(In reply to Darin Adler from comment #4)
> Excellent that we can/will do this.
> 
> This comment addresses the safety side of the tradeoff. What about the
> performance side?

Based on ad hoc testing of WebGL, I could not spot any difference.
Comment 7 EWS 2021-04-28 00:00:56 PDT 
Committed r276693 (237107@main): <https://commits.webkit.org/237107@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 426651 [details].