Le Wed, Mar 12, 2025 at 01:47:09PM +0100, Marcus Glocker a écrit : > On Wed, Mar 12, 2025 at 11:30:03AM GMT, Marcus Glocker wrote: > > > On Wed, Mar 12, 2025 at 10:24:04AM GMT, Landry Breuil wrote: > > > > > Le Wed, Mar 12, 2025 at 10:05:30AM +0100, Landry Breuil a ?crit : > > > > hi, > > > > > > > > updated my t470s from '#564: Tue Feb 25 21:20:18 MST 2025' to '#591: Tue > > > > Mar 11 20:48:18 MDT 2025' and my webcam now shows weird 'stripes' on top > > > > of the image (that works). > > > > > > the 'weird stripes' only show up from some webrtc clients (ive noticed > > > it with jitsi in firefox 137.0b4), using video(1) i havent been able to > > > replicate (yet).. > > > > > > no issue in https://mozilla.github.io/webrtc-landing/gum_test.html so > > > might be the way jitsi queries the video device. > > > > > > Landry > > > > I have the same cam in my X1 Carbon. I can't reproduce the issue on > > video(1) either, because it's using YUYV422 as the default input format. > > When using ffplay with MJPEG as input format, I can reproduce the issue. > > > > I'll try to figure out which of the recent commits has introduced the > > regression. > > The issue was introduced by: > > *** > > Revision 1.247 / (download) - annotate - [select for diffs], Sat Mar 1 > 12:30:19 2025 UTC (11 days ago) by mglocker > Branch: MAIN > Changes since 1.246: +55 -32 lines > Diff to previous 1.246 (colored) > > Copy frames directly in to the mmap'ed buffer. This saves us one bcopy() > down the frame queuing path. > > Original idea and diff from kirill@, with some modifications. > > ok kirill@ > > *** > > Unfortunately, I couldn't figure out yet what the issue is with this > commit. Hence, following the back out diff for uvideo.c 1.247. It > fixes the issue for me. What about you?
that diff fixes the issue for me, yes.
