1811775 - Exiting PiP causes video to freeze with hardware video decoding

Status: UNCONFIRMED

(Core :: Audio/Video: Playback, defect)

Description

[Almtst]

Attachment: about:support

User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:109.0) Gecko/20100101 Firefox/110.0

Steps to reproduce:

Sometimes exiting PiP causes the video to freeze while the audio continues to play normally, this can be forced by quickly entering and exiting PiP several times.

I ran mozregression and got "Bug 1762125. Allow DXVA on newer Intel drivers", pushlog "https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=6c5ac0c99219dfa6396841f69b13c47e3e559c37&tochange=ed124d06d67623a6227a41ecffd8533733755976" as the cause.

I also captured a profile "https://share.firefox.dev/3iX72rP"

Actual results:

Video freezes

Expected results:

Video continues to play normally

Comment 1

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

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

Comment 2

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

:jrmuizel, since you are the author of the regressor, bug 1762125, could you take a look? Also, could you set the severity field?

For more information, please visit auto_nag documentation.

Comment 3

[kpatenio]

Thanks for attaching a profile and a mozregression log. Given the regression, I'll move this over to Core::Audio/Video Playback to see if the folks there can investigate this further.

Comment 4

[Jeff Muizelaar [:jrmuizel]]

Almtst, can you try downgrading you Intel driver to something earlier (with a x.x.100.x instead of x.x.101.x) and try getting a new regression window?

Comment 5

[Almtst]

I couldn't find a driver version x.x.100.x on Intel's website, but there was one from 8/11/2020, version 21.20.16.5174, I ran mozregression with that driver and it seems that the problem is present since the introduction of PiP, that made me think that it's possible that I got Bug 1762125 as the regression because before that bug fix, HW video decoding is disabled for newer drivers, so I disabled HW acceleration with the driver updated (31.0.101.2115) and latest FF beta and the problem doesn't occur this way, only with HW acceleration enabled.