Closed Bug 1994319 Opened 1 year ago Closed 1 year ago

Categories

(Core :: Graphics: WebGPU, defect, P1)

defect

Tracking

()

VERIFIED FIXED
Tracking Status
firefox-esr115 --- disabled
firefox-esr140 --- disabled
firefox144 --- wontfix
firefox145 + verified
firefox146 --- verified

People

(Reporter: ErichDonGubler, Assigned: ErichDonGubler)

References

Details

(Keywords: sec-other, Whiteboard: [adv-main145-] patch for sec-high bugs)

Attachments

(1 external file, 1 obsolete file)

No description provided.
Attached file (secure) (obsolete) β€”
Attachment #9520076 - Flags: approval-mozilla-release?

Did you mean to request uplift to beta (145) rather than release (144)? Or do you really intend to ship this in a 144.0.x update? Wouldn't you still need a beta uplift request in that case? And where did these fixes land on nightly? Maybe bug 1992842?

Keywords: sec-other
Whiteboard: patch for sec-high bugs

Did you mean to request uplift to beta (145) rather than release (144)? Or do you really intend to ship this in a 144.0.x update? Wouldn't you still need a beta uplift request in that case?

I backported to Stable, but not Beta, because I'd erroneously thought that the fixes had already landed before soft freeze for Beta.

And where did these fixes land on nightly? Maybe bug 1992842?

Unfortunately, we're still waiting for Nightly fixes to land, which is why bug 1986431 and bug 1988488 are still open.


:dmeehan has asked me if there was a way to split the revendoring done by patches here, so we can associate them with their respective bugs. This is possible, just some significant effort to manage. I may close this bug and file patches directly against those bugs instead, depending on others' posture.

:ErichDonGubler, it's difficult to track what's need here and if it's still appropriate for the planned dot release.
Bug 1986431 and Bug 1988488 are still open, but there is a depends on bug (Bug 1995050) that fixes them?
Maybe Fx145 is a better target and there is time to get appropriate beta uplifts in?

Flags: needinfo?(egubler)

:dmeehan: As of half an hour before your message, we're just waiting on confirmation from the filer of those issues to confirm a fix in Nightly. I will file a patch stack for Beta with this bug now, so that things are ready to be landed once confirmed. You can make the call for which patches to take at that point. πŸ™‚

Flags: needinfo?(egubler)
Attached file (secure) β€”
Attachment #9521468 - Flags: approval-mozilla-beta?

I'll reject the release uplift request since Bug 1986431 is still open. This can ride the train with Fx145.

Attachment #9520076 - Attachment is obsolete: true
Attachment #9520076 - Flags: approval-mozilla-release?

(In reply to Donal Meehan [:dmeehan] from comment #8)

I'll reject the release uplift request since Bug 1986431 is still open. This can ride the train with Fx145.

The vendoring of the upstream in bug 1995050 landed on nightly 146 -- it's not riding any train to 145; this bug is trying to get cherry-picked patches for bug 1986431 and bug 1988488 uplifted to 145.

it would have been less confusing to have each cherry-picked patch attached to the security bug it's trying to fix. Is the build set up so that we can't just apply patches on top of the upstream code without having to vendor it first? That could really bite us if we ever have to fix something here in an emergency chemspill.

(In reply to Daniel Veditz [:dveditz] from comment #9)

This can ride the train with Fx145.

it's not riding any train to 145

This can "ride train" was in relation to going to release and just meant as a comment on why the release uplift was cancelled. It will still need an uplift to beta.

:ErichDonGubler, can you take a look at :dveditz questions in Comment 9?

Flags: needinfo?(egubler)

Erich, could you also file the uplift request form to document the riskiness of the patch? Thanks

firefox-beta Uplift Approval Request

  • User impact if declined: If declined, we leave users open to these vulnerabilities to two sec-high security vulnerabilities until the current Nightly train catches up with them.
  • Code covered by automated testing: no
  • Fix verified in Nightly: yes
  • Needs manual QE test: yes
  • Steps to reproduce for manual QE testing: See reproduction steps for bug 1986431 and bug 1988488.
  • Risk associated with taking this patch: low
  • Explanation of risk level: These have been verified to fix a problem, and only limit resource usage. There is little to no risk for existing control flow otherwise.
  • String changes made/needed: No.
  • Is Android affected?: no
Flags: qe-verify+

(In reply to Daniel Veditz [:dveditz] from comment #9)

it would have been less confusing to have each cherry-picked patch attached to the security bug it's trying to fix. Is the build set up so that we can't just apply patches on top of the upstream code without having to vendor it first? That could really bite us if we ever have to fix something here in an emergency chemspill.

I don't disagree. I did offer to do that to :dmeehan (which didn't get an explicit response, now that I think about it) via DM, but I wasn't sure if it was necessary or desirable, so I left the thought of splitting at that. Guess I know now it would be helpful to do so! πŸ˜… If we really need to, I can still split into multiple patches. It would have to be strictly sequenced, since we would have to consume a sequence of commits upstream to make this work.

:pascalc: Done!

Flags: needinfo?(egubler)
Attachment #9521468 - Flags: approval-mozilla-beta? → approval-mozilla-beta+
Status: ASSIGNED → RESOLVED
Closed: 1 year ago
Resolution: --- → FIXED
Group: gfx-core-security → core-security-release
QA Whiteboard: [sec] [uplift] [qa-triage-done-c146/b145]

Reproduced the crash on Firefox 145.0a1 (2025-09-15) on Ubuntu 24.04 with the attachment provided in Bug 1988488.

The crash is no longer reproducible on Firefox 145.0b9 (treeherder build) and Firefox 146.0a1 (2025-10-29) on the same system.

Status: RESOLVED → VERIFIED
QA Whiteboard: [sec] [uplift] [qa-triage-done-c146/b145] → [sec] [uplift] [qa-triage-done-c146/b145] [qa-ver-done-c146/b145]
Flags: qe-verify+

It probably makes more sense to track the advisories for this in bug 1986431 and bug 1988488.

Whiteboard: patch for sec-high bugs → [adv-main145-] patch for sec-high bugs

Note that :aleiserson has filed bug 2002393 to handle this question from :dveditz in the depth it deserves:

Is the build set up so that we can't just apply patches on top of the upstream code without having to vendor it first? That could really bite us if we ever have to fix something here in an emergency chemspill.

Sorry I dropped the ball on that, and thanks for picking it up, Andy!

Group: core-security-release
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: