1867566 - Suddenly Nightly page loads don't finish on Linux.

Status: RESOLVED WORKSFORME

(Core :: DOM: Navigation, defect)

Description

[David Koppelman]

Steps to reproduce:

Tried to load a page. (The problem has occurred when a page is loaded various ways, such as a refresh, opening a new URL, or submitting a form. After a minute or two I clicked on the WM X button to close Nightly.

Actual results:

The window does not change (maybe showing old contents) and the activity indicator in the tab shows activity, but the page does not load. I can switch to other tabs, but pages don't load or reload in those. After selecting exit there is a long delay, and the crash reporter pops up. (I'll provide crash IDs further below.)

Expected results:

The page should have loaded. After selecting quit Nightly should have completed execution within a second.

Comment 1

[David Koppelman]

Attachment: Stack at time of page load hang, and list of threads.

I've attached a stack at the time of the hang of thread 1, FWIW. After collecting that stack Nightly crashed with
Crash ID: f0326fde-4d16-4008-b2d6-d076e0231130

Comment 2

[BugBot [:suhaib / :marco/ :calixte]]

The Bugbug bot thinks this bug should belong to the 'Core::DOM: Navigation' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.

Comment 3

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)(away 4/27 - 5/6)]

bp-f0326fde-4d16-4008-b2d6-d076e0231130

MOZ_CRASH Reason (Sanitized) | Shutdown hanging at step AppShutdownNetTeardown. Something is blocking the main-thread.

jstuttle: I'm not sure who can check the crash report dipper, could you take a look or redirect to somebody else?

Comment 4

[Jens Stutte [:jstutte]]

The crash report points to AppShutdownNetTeardown but this might just be the secondary effect of already hanging and then asking for shutdown. So it is not clear if the stacks in the crash report can point us to the root cause.

Maybe the reporter could try to capture a startup profile to see where we are hanging. Probably best to switch on all threads in the profiler settings panel (if your Firefox is responsive enough to do so).

Comment 5

[Jens Stutte [:jstutte]]

(and you could try with mozregression to find the exact version that caused this)

Comment 6

[tgn-ff]

This sounds like Bug 1868070.

Comment 7

[Jens Stutte [:jstutte]]

(In reply to tgn-ff from comment #6)

This sounds like Bug 1868070.

The linked crash is from build 20231130093957 while that patch landed in 20231208094241. That would mean, the reporter could update nightly and re-check, I assume. Thanks for your support!

Comment 8

[David Koppelman]

I've been updating Nightly daily. My last crash was on 7 December, so a 2023-12-08 build could very well have fixed the problem for me. I'm not 100% sure because I never had a sure-fire STR. I'll report back if the problem re occurs.

Comment 9

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)(away 4/27 - 5/6)]

Thank you everybody!

I think we can close this bug as WORKSFORME for now. David Koppelman, feel free to reopen this bug if you see new crash in a newer Nightly build.