After installing this update it is required that you logout of
your current user session and log back in to ensure the changes
supplied by this update are applied properly.
@mstransky I recommend disabling karma autopush on this update, otherwise it gets pushed stable before you manage to gather extensive feedback on this. It might even skip updates-testing completely.
This update's test gating status has been changed to 'passed'.
Okay, this build restored the video playback functionality but the performance is horrible running at only few FPS, now including other web-sites as well, like YouTube with vp9 and ac1 codecs. This is on RX 6700XT.
BZ#2094319 No video playback on Firefox 101 on some sites
Okay, this build restored the video playback functionality but the performance is horrible running at only few FPS, now including other web-sites as well, like YouTube with vp9 and ac1 codecs. This is on RX 6700XT.
Please file that at bugzilla.redhat.com. I don't see that on RX 6600XT. Also we're supposed to switch to SW decode when VA-API fails.
Just quickly tested youtube playback on two different machines:
Kabylake (9th gen): VA-API is working and 1440p60 is playing smoothly with hardware acceleration.
Tigerlake (12th gen): VA-API seems to initialize just fine with the iHD_drv_video driver but no hardware acceleration. This laptop does also have an RTX3050 with the nvidia driver loaded but that isn't used for Firefox.
BZ#2094953 VA-API does not work even though it is supposed to according to update description
Just quickly tested youtube playback on two different machines:
Tigerlake (12th gen): VA-API seems to initialize just fine with the iHD_drv_video driver but no hardware acceleration This laptop does also have an RTX3050 with the nvidia driver loaded but that isn't used for Firefox.
I was doing a zoom with this today, and oomd keeps killing the scope the zoom is running in. Also sounds like a memory leak as xvitaly says. Should I open a bug for this? There's only one thing to go on in the logs.
Jun 10 09:45:50 fovo.local systemd-oomd[662]: Killed /user.slice/user-1000.slice/user@1000.service/app.slice/app-gnome-firefox-4369.scope/4369 due to memory used (16076365824) / total (16403767296) and swap used (7871721472) / total (8589930496) being more than 90.00%
That's a metricftonne of execessive resources. Nothing is logged until the kill. If I disable oomd, it consistently siphons away enough resources to freeze the computer completely - can't get to a tty, mouse arrow frozen, doesn't recover at all for 1 minute so I forced power off.
BZ#2094319 No video playback on Firefox 101 on some sites
BZ#2094953 VA-API does not work even though it is supposed to according to update description
BZ#2095324 regression: firefox 101 breaks video on some sites
This update has been submitted for testing by stransky.
This update's test gating status has been changed to 'waiting'.
stransky edited this update.
This update's test gating status has been changed to 'failed'.
This update's test gating status has been changed to 'waiting'.
VAAPI accelaration is working well for me on youtube with Thinkpad P1 gen3:
@mstransky I recommend disabling karma autopush on this update, otherwise it gets pushed stable before you manage to gather extensive feedback on this. It might even skip updates-testing completely.
This update's test gating status has been changed to 'passed'.
My RAM gets flooded with this build as well. The quickest reproducer I was able to find is loading an HTML game. No such issue with firefox-101.0-1.
This update has been submitted for stable by bodhi.
Okay, this build restored the video playback functionality but the performance is horrible running at only few FPS, now including other web-sites as well, like YouTube with vp9 and ac1 codecs. This is on RX 6700XT.
stransky edited this update.
Please file that at bugzilla.redhat.com. I don't see that on RX 6600XT. Also we're supposed to switch to SW decode when VA-API fails.
Opened a BZ here: https://bugzilla.redhat.com/show_bug.cgi?id=2095743
Just quickly tested youtube playback on two different machines:
It's because you hit https://bugzilla.mozilla.org/show_bug.cgi?id=1773377 On systems with more GPU the VA-API is broken now but should fall back to SW decode.
Leaks all available RAM after 2-3 hours, then earlyoom kills it.
I was doing a zoom with this today, and oomd keeps killing the scope the zoom is running in. Also sounds like a memory leak as xvitaly says. Should I open a bug for this? There's only one thing to go on in the logs.
Jun 10 09:45:50 fovo.local systemd-oomd[662]: Killed /user.slice/user-1000.slice/user@1000.service/app.slice/app-gnome-firefox-4369.scope/4369 due to memory used (16076365824) / total (16403767296) and swap used (7871721472) / total (8589930496) being more than 90.00%
That's a metricftonne of execessive resources. Nothing is logged until the kill. If I disable oomd, it consistently siphons away enough resources to freeze the computer completely - can't get to a tty, mouse arrow frozen, doesn't recover at all for 1 minute so I forced power off.
extreme memory leak, consumes all available memory and swap https://bugzilla.redhat.com/show_bug.cgi?id=2095791
I can see we have a build no.2, could we push that through to test memleak fix please?
Yes, but more fixes are coming - https://bugzilla.mozilla.org/show_bug.cgi?id=1773377 for multiple GPU scenario and https://bugzilla.mozilla.org/show_bug.cgi?id=1774075 for radeon WebGL fixes.
Feel free to just test builds from koji, I'd like to submit all fixes together. And thanks for testing, it's very helpful!
Oh, cool. Thanks!
stransky edited this update.
New build(s):
Removed build(s):
Karma has been reset.
This update has been submitted for testing by stransky.
Shouldn't BZ#2095718 be added as well? If not, I can close it manually.
stransky edited this update.
This update has been pushed to testing.
This update can be pushed to stable now if the maintainer wishes
This update has been submitted for stable by stransky.
VAAPI accelaration is working well for me on youtube with Thinkpad P1 gen3:
Working fine.
Appears to work fine with Youtube and such
Does anyone running this version notice any pixelation and color tearing in the incoming video in WebRTC calls after a short while?
This update has been pushed to stable.