WebGL/ANGLE interop with D3D layers requests incorrect extension

RESOLVED FIXED in Firefox 6

Status

()

Core
Canvas: WebGL
RESOLVED FIXED
6 years ago
6 years ago

People

(Reporter: jgilbert, Assigned: jgilbert)

Tracking

(Depends on: 1 bug, {regression})

unspecified
mozilla8
x86
Windows 7
regression
Points:
---
Dependency tree / graph

Firefox Tracking Flags

(firefox5 unaffected, firefox6 fixed, firefox7 fixed)

Details

(Whiteboard: [qa-])

Attachments

(1 attachment)

(Assignee)

Description

6 years ago
GLContextEGL requires extension "EGL_ANGLE_surface_d3d_share_handle", which no longer exists. According to the extension spec, the extension was renamed in version 2 to "EGL_ANGLE_surface_d3d_texture_2d_share_handle", but this change was evidently never factored back into GLContextEGL.

Refs:
http://www.khronos.org/registry/egl/extensions/ANGLE/EGL_ANGLE_surface_d3d_texture_2d_share_handle.txt
EGL_ANGLE_surface_d3d_share_handle only ever existed in a local patch that we used to have, not in upstream ANGLE.

When I upgraded ANGLE to r653 (bug 657748), I removed this patch as that functionality had been integrated upstream. I didn't notice that the extension name was different.
tracking-firefox6: --- → +
tracking-firefox7: --- → +
tracking-firefox8: --- → +
(Assignee)

Comment 2

6 years ago
Created attachment 549201 [details] [diff] [review]
Fix

Updates extension and relevant identifiers to match spec.
Fixed y-flip problem with shared textures.
Attachment #549201 - Flags: review?(bjacob)
Attachment #549201 - Flags: review?(bjacob) → review+
Summary: ANGLE layers acceleration requests incorrect extension → WebGL/ANGLE interop with D3D layers requests incorrect extension
Landed on central:
http://hg.mozilla.org/mozilla-central/rev/267ada59f86f

Please give us beta approval.
Attachment #549201 - Flags: approval-mozilla-beta?
Attachment #549201 - Flags: approval-mozilla-aurora?
Status: ASSIGNED → RESOLVED
Last Resolved: 6 years ago
Resolution: --- → FIXED

Updated

6 years ago
Attachment #549201 - Flags: approval-mozilla-beta?
Attachment #549201 - Flags: approval-mozilla-beta+
Attachment #549201 - Flags: approval-mozilla-aurora?
Attachment #549201 - Flags: approval-mozilla-aurora+
Blocks: 657748
Keywords: regression
Landed on aurora and beta.
http://hg.mozilla.org/releases/mozilla-aurora/rev/935e296aabf4
http://hg.mozilla.org/releases/mozilla-beta/rev/0bb77d0b73ad
tracking-firefox6: + → ---
tracking-firefox7: + → ---
tracking-firefox8: + → ---

Updated

6 years ago
status-firefox6: affected → fixed
status-firefox7: affected → fixed
status-firefox8: affected → ---
Target Milestone: --- → mozilla8

Comment 5

6 years ago
Mozilla/5.0 (Windows NT 6.1; rv:7.0) Gecko/20100101 Firefox/7.0

Can anyone please help me with a test case or with STR / guidelines that I can use to verify this fix?

Thank you
Actually, no, it's not very easy to verify this one. The only user-visible difference is speed: when we had this bug, running WebGL demos with a large window or full-screen was slower than normal.

In order to prevent this kind of regression in the future, we need WebGL performance tests: bug 676138.
Depends on: 676138
qa- as no QA fix verification needed
Whiteboard: [qa-]
You need to log in before you can comment on or make changes to this bug.