Closed
Bug 1780084
Opened 2 years ago
Closed 7 months ago
`mozilla::webgpu::Buffer::MapAsync` doesn't check `mValid`
Categories
(Core :: Graphics: WebGPU, defect, P3)
Core
Graphics: WebGPU
Tracking
()
RESOLVED
FIXED
120 Branch
Tracking | Status | |
---|---|---|
firefox120 | --- | fixed |
People
(Reporter: jimb, Assigned: nical, NeedInfo)
References
Details
Attachments
(1 file)
mozilla::webgpu::Buffer::MapAsync
will call SetMapped
even if mValid
is false, which seems wrong.
The WebGPU spec currently says that buffer validation should happen on the content timeline, but apparently that's supposed to change. But even if error reporting should happen on the Device timeline in the usual way, we should not be doing anything to the internal state of our Buffer
object.
Assignee | ||
Comment 1•2 years ago
|
||
Moving to webgpu-v1 milestone since this is more about spec decisions than a bug I think. The pending buffer patches should handle things properly whether or not mValid is false.
Comment 2•2 years ago
|
||
The severity field is not set for this bug.
:jimb, could you have a look please?
For more information, please visit auto_nag documentation.
Flags: needinfo?(jimb)
Updated•11 months ago
|
Severity: -- → S3
Priority: -- → P3
Updated•8 months ago
|
Assignee: nobody → nical.bugzilla
Status: NEW → ASSIGNED
Assignee | ||
Comment 3•7 months ago
|
||
Pushed by nsilva@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/c06cdccafe8f Sanity-check for mValid before mapping the buffer. r=webgpu-reviewers,ErichDonGubler
Comment 5•7 months ago
|
||
bugherder |
Status: ASSIGNED → RESOLVED
Closed: 7 months ago
status-firefox120:
--- → fixed
Resolution: --- → FIXED
Target Milestone: --- → 120 Branch
You need to log in
before you can comment on or make changes to this bug.
Description
•