Hi Salvatore,
Thank you for your reply and bissect explanation.
As a quick top-post Reply: 6.18.2-1~exp1 works for me \o/
I think, then, the bissection isn't worth it. But if you need it for the
record, just ask me.
Thanks,
Benoît
Le 19/12/2025 à 22:14, Salvatore Bonaccorso a écrit :
Control: tags -1 + moreinfo upstream
Hi Benoît,
On Thu, Dec 18, 2025 at 09:37:32PM +0100, Benoît Rouits wrote:
Package: src:linux
Version: 6.17.11-1
Severity: normal
File: /usr/share/bug/linux-image-6.17.11+deb14-amd64/presubj
X-Debbugs-Cc: [email protected], [email protected]
User: [email protected]
Usertags: amd64
Dear Maintainer,
I am experiencing a regression in linux image 6.17.11 (works with image 6.17.9)
* What led up to the situation?
Trying to use the webcam with e.g. firefox (or qv4l2 for testing purpose)
* What exactly did you do (or not do) that was effective (or
ineffective)?
Went to a video chat site (medical in my case).
* What was the outcome of this action?
firefox did not ask permissions for integrated webcam video capture, and no
capture was possible.
* What outcome did you expect instead?
firefox to display a popup for accepting webcam capture, and use the webcam.
When using qv4l2, the capture is a black screen.
various CLI checks:
$ lsusb | grep -i uvc
Bus 002 Device 003: ID 13d3:5205 IMC Networks USB2.0 UVC 2M WebCam
$ sudo dmesg | grep uvcvideo
[ 7.733171] usbcore: registered new interface driver uvcvideo
[ 1027.557731] uvcvideo 2-1.2:1.1: Failed to resubmit video URB (-1).
[ 1027.557798] uvcvideo 2-1.2:1.1: Failed to resubmit video URB (-1).
$ v4l2-compliance | grep -5 "fail:"
Control ioctls (Input 0):
test VIDIOC_QUERY_EXT_CTRL/QUERYMENU: OK
test VIDIOC_QUERYCTRL: OK
test VIDIOC_G/S_CTRL: OK
fail: v4l2-test-controls.cpp(980): ret != EINVAL (got 13)
test VIDIOC_G/S/TRY_EXT_CTRLS: FAIL
test VIDIOC_(UN)SUBSCRIBE_EVENT/DQEVENT: OK
test VIDIOC_G/S_JPEGCOMP: OK (Not Supported)
Standard Controls: 12 Private Controls: 0
If you need the same CLI output on the working linux image 6.17.9 for
comparison, or other CLI outputs, just ask me.
First of all as we will upload soon the last version from the 6.17.y
series, 6.17.13 and then move to 6.18.y there is no much point into
analyzing if the issue is gone with 6.18.2. Thus, can you please test
6.18.2-1~exp1 as uploaded to experimental and check if the problem
persist?
If yes, and given you are able to reproduce the behaviour
consistently, it would be great if you can bisect the problem between
the two good and bad versions. Then we can report the problem upstream
(and know it affects still 6.18.2). So that would involve compiling
and testing a few kernels:
git clone --single-branch -b linux-6.17.y
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git
cd linux-stable
git checkout v6.17.9
cp /boot/config-$(uname -r) .config
yes '' | make localmodconfig
make savedefconfig
mv defconfig arch/x86/configs/my_defconfig
# test 6.17.9 to ensure this is "good"
make my_defconfig
make -j $(nproc) bindeb-pkg
... install the resulting .deb package and confirm problem does not exist
# test 6.17.11 to ensure this is "bad"
git checkout v6.17.11
make my_defconfig
make -j $(nproc) bindeb-pkg
... install the resulting .deb package and confirm problem exists
With that confirmed, the bisection can start:
git bisect start
git bisect good v6.17.9
git bisect bad v6.17.11
In each bisection step git checks out a state between the oldest
known-bad and the newest known-good commit. In each step test using:
make my_defconfig
make -j $(nproc) bindeb-pkg
... install, verify if problem exists
and if the problem is hit run:
git bisect bad
and if the problem doesn't trigger run:
git bisect good
. Please pay attention to always select the just built kernel for
booting, it won't always be the default kernel picked up by grub.
Iterate until git announces to have identified the first bad commit.
Then provide the output of
git bisect log
In the course of the bisection you might have to uninstall previous
kernels again to not exhaust the disk space in /boot. Also in the end
uninstall all self-built kernels again.
Let us know on both first 6.18.2 (and then you can actually stop) or
if it still present, if you can identify a breaking commit with the
bisect procedure.
Regards,
Salvatore
