MacIOSurfaceTextureClientOGL not properly implementing locking

RESOLVED WONTFIX

Status

()

RESOLVED WONTFIX
3 years ago
3 years ago

People

(Reporter: jya, Unassigned)

Tracking

Trunk
Points:
---

Firefox Tracking Flags

(firefox43 affected)

Details

(Whiteboard: [gfx-noted])

Attachments

(1 attachment)

(Reporter)

Description

3 years ago
Upon reviewing bug 1072313 ; this caught my attention.

There's no actual locking put in place with MacIOSurface ; and the logic to report if a surface is locked appears broken.

No idea if it's worth doing, but this will apply the same logic as what is done for other GPU surface (D3D9, D3D11 and others).
(Reporter)

Comment 1

3 years ago
Created attachment 8662194 [details] [diff] [review]
Implement proper MacIOSSurfaceTextureClientOGL locking.
Attachment #8662194 - Flags: review?(matt.woodrow)
Comment on attachment 8662194 [details] [diff] [review]
Implement proper MacIOSSurfaceTextureClientOGL locking.

Review of attachment 8662194 [details] [diff] [review]:
-----------------------------------------------------------------

We never call Lock() on a MacIOSurfaceTextureClientOGL, and MacIOSurface::Lock probably doesn't really do what we'd want anyway.

How about we just assert that this isn't called?
Attachment #8662194 - Flags: review?(matt.woodrow)
It'd be nice for Lock()/Unlock() to return the status based on the io surface lock, rather than assume the lock succeeded in that case.  Non-withstanding Matt's comment to assert, or even better, MOZ_CRASH.
Whiteboard: [gfx-noted]
(Reporter)

Updated

3 years ago
Status: NEW → RESOLVED
Last Resolved: 3 years ago
Resolution: --- → WONTFIX
You need to log in before you can comment on or make changes to this bug.