Disable sw-wr on non-x86 Linux in late beta and release.
Categories
(Core :: Graphics: WebRender, defect)
Tracking
()
People
(Reporter: jrmuizel, Assigned: jrmuizel)
References
Details
Attachments
(2 files)
48 bytes,
text/x-phabricator-request
|
pascalc
:
approval-mozilla-beta+
pascalc
:
approval-mozilla-release+
|
Details | Review |
48 bytes,
text/x-phabricator-request
|
Details | Review |
This avoids bug 1714064 (SPARC) and bug 1714511 (ARM)
Assignee | ||
Comment 1•3 years ago
|
||
This avoids bug 1714064 (SPARC) and bug 1714511 (ARM)
Assignee | ||
Comment 2•3 years ago
|
||
Comment on attachment 9226414 [details]
Bug 1715895 - Disable sw-wr on non-x86 Linux in late beta and release.
Beta/Release Uplift Approval Request
- User impact if declined: Crashes on SPARC and video corruption on ARM Linux
- Is this code covered by automated tests?: No
- Has the fix been verified in Nightly?: No
- Needs manual test from QE?: No
- If yes, steps to reproduce:
- List of other uplifts needed: None
- Risk to taking this patch: Low
- Why is the change risky/not risky? (and alternatives if risky): This just disables sw-wr on 89 (switches back to the same as 88) on non-x86, it has no impact on builds that we ship.
- String changes made/needed:
Comment 4•3 years ago
|
||
bugherder |
Comment 5•3 years ago
|
||
Jeff, I am assuming that this would fix bug 1714511, correct?
I saw that the uplift request in bug 1714069 was removed today, would that also be fixed with this one?
Assignee | ||
Comment 6•3 years ago
|
||
Nope, this won't help with bug 1714069. We're hoping to use bug 1715902 for that instead.
Comment 7•3 years ago
|
||
Comment on attachment 9226414 [details]
Bug 1715895 - Disable sw-wr on non-x86 Linux in late beta and release.
Approved for beta and release, thanks.
Comment 8•3 years ago
|
||
bugherder uplift |
Comment 9•3 years ago
|
||
Jeff, this patch does not graft cleanly to release, could you provide a rebased patch please? Thanks
Assignee | ||
Comment 10•3 years ago
|
||
This avoids bug 1714064 (SPARC) and bug 1714511 (ARM)
Assignee | ||
Updated•3 years ago
|
Comment 11•3 years ago
|
||
bugherder uplift |
Comment 12•3 years ago
|
||
I'm confused. There is EARLY_BETA_OR_EARLIER in:
https://hg.mozilla.org/mozilla-central/rev/6d9f9d90a767f4dd696b5648e5ebb0956108e3d4
which can be valid also on SPARC and thus it causes problem again...
But original fix pointed in https://bugzilla.mozilla.org/show_bug.cgi?id=1715895#c11 was missing EARLY_BETA_OR_EARLIER.
Assignee | ||
Comment 13•3 years ago
|
||
Indeed. sw-wr is disabled in release but it is just temporary to give non-x86 platforms a chance to get their bugs sorted out. The #ifdef will likely be going away in Firefox 91 or 92.
Comment 14•3 years ago
|
||
I would prefer to remove the #ifdef in Firefox 92 (after ESR version)...
Comment 15•3 years ago
|
||
(In reply to Jeff Muizelaar [:jrmuizel] from comment #13)
Indeed. sw-wr is disabled in release but it is just temporary to give non-x86 platforms a chance to get their bugs sorted out. The #ifdef will likely be going away in Firefox 91 or 92.
While Firefox 91 still contains the ifdef:
it no longer seems to workaround SPARC Rust issue (wr_dp_push_stacking_context() gets wrong parameters).
Is there still possibility to run Firefox 91 without WebRender API? Even for: firefox -P.
Assignee | ||
Comment 16•3 years ago
|
||
You should be able gfx.webrender.force-legacy-layers=true to disable WebRender in 91
Assignee | ||
Comment 17•3 years ago
|
||
Actually, that's wrong. gfx.webrender.force-disabled=true should still work in 91. gfx.webrender.force-legacy-layers=true is just for 92
Comment 18•3 years ago
|
||
(In reply to Jeff Muizelaar [:jrmuizel] from comment #17)
Actually, that's wrong. gfx.webrender.force-disabled=true should still work in 91. gfx.webrender.force-legacy-layers=true is just for 92
Sorry. What that just mean? Is it expected that function wr_dp_push_stacking_context() is called in Firefox 91 then?
I'm trying to workaround SPARC Rust issue (https://github.com/rust-lang/rust/issues/86163) and avoid to call this function (and similar which passes Box2D structure as argument).
Assignee | ||
Comment 19•3 years ago
|
||
When gfx.webrender.force-disabled=true is set in Firefox 91, wr_dp_push_stacking_context should not be called.
Comment 20•3 years ago
|
||
(In reply to Jeff Muizelaar [:jrmuizel] from comment #19)
When gfx.webrender.force-disabled=true is set in Firefox 91, wr_dp_push_stacking_context should not be called.
Thank you! Starting Firefox 91 with MOZ_WEBRENDER=0 seems to be the best option now.
Description
•