20:00:02 <carnil> #startmeeting
20:00:02 <MeetBot> Meeting started Wed Nov 11 20:00:02 2026 UTC.  The chair is carnil. Information about MeetBot at https://wiki.debian.org/MeetBot.
20:00:02 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
20:00:20 <waldi> hi
20:00:20 <yunseong[m]> Hello
20:00:23 <carnil> #chair carnil ukleinek waldi
20:00:23 <MeetBot> Current chairs: carnil ukleinek waldi
20:00:40 <carnil> hi all, bwh said he could not partecipate tonight unfortunately
20:00:54 <carnil> (I'm not very well prepared, sorry)
20:01:01 <carnil> is there anything to discuss before we dig into the bugs?
20:01:15 <ukleinek> idk
20:01:45 <waldi> not from my side
20:02:00 <carnil> okay then let's start with the bugs
20:02:08 <carnil> #topic #1093775: (i, ) linux-image-6.12.9-amd64: Linux kernel option CONFIG_SND_USB_AUDIO_MIDI_V2 is not set
20:02:19 <ukleinek> looks easy to address
20:02:20 <carnil> this one re-appeared again after a while because someone followed up
20:02:43 <ukleinek> #action ukleinek will create a MR for #1093775
20:03:21 <carnil> thanks
20:03:35 <carnil> #topic #1149109: (i, M) linux-image-7.1.8+deb13-amd64: amdgpu flip_done timed out and freeze on RX 6600 (regression from 7.1.3)
20:04:02 <carnil> still no reply from the reporter; reminder we wasked to test a nwer kernel and then to report upstream and loop back the upstream issue
20:04:13 <ukleinek> wait
20:04:16 <carnil> so wait, if nothing happens for a couple of cycles we might as well close the bug
20:04:26 * ukleinek nods
20:04:28 <carnil> #topic #1151377: (i, u+U) linux: lpfc boot hang since 6.1.178 - use-after-free of mboxq in lpfc_sli4_driver_resource_setup() backport (merged with #1151558, #1153112)
20:04:55 <carnil> known regression in 6.1.y kernel, the change is pending in my MR https://salsa.debian.org/kernel-team/linux/-/merge_requests/2142 (with a bug closer)
20:05:06 <carnil> #topic #1151916: (i, Mu) linux: xhci AMD Renoir [1022:1639] USB hotplug detection broken after boot on 6.12.113
20:05:29 <carnil> ukleinek: from the last meeting we had an action to you to interpret the bisect results and maybe to come up with a better candidate
20:05:36 <carnil> I assume it is okay to leave the action to you?
20:05:49 <ukleinek> should be marked as forwarded
20:05:56 <carnil> I would not have better ideas at the moment for other things to do
20:06:13 <ukleinek> yeah, #1151916 is still open in my mailbox
20:06:24 <carnil> ukleinek: why forwarded?
20:06:36 <ukleinek> carnil: the previous bug
20:06:47 <carnil> ukleinek: ah well that one, yes once the reporter has filled the issue
20:07:01 <carnil> ah nevermind, yes
20:07:06 <ukleinek> the one with "the change is pending in my MR
20:07:08 <ukleinek> "
20:07:36 <ukleinek> i.e. #1151377
20:08:02 <carnil> marked it forwarded to https://lore.kernel.org/r/202610
20:08:03 <carnil> 08184328.2857037-1-andrey.misyrov@gmail.com
20:09:29 <carnil> okay back to #1151916, then ukleinek keeps the action
20:09:46 * ukleinek nods
20:09:47 <carnil> #topic #1152007: (i, u+M) linux-image-7.2.9+deb13-amd64: kernel panic on resume (s2idle) in ath11k after order-8 allocation failure in ath11k_dp_alloc
20:10:12 <carnil> yunseong[m]: asked questions, no reply from reporter yet
20:10:43 <carnil> so we can wait here as well
20:10:44 <ukleinek> what does it mean for a bug to be marked both patch and moreinfo?
20:12:09 <ukleinek> but ack for wait
20:12:24 <ukleinek> (that is, on the reporter, not before progressing to the next bug :-)
20:12:37 <carnil> #topic #1152896: (i, ) ahci: regression on Intel 9-series [8086:8c82] causing host bus fatal errors and ICRC ABRT in 7.x
20:12:54 <waldi> waiting for OP
20:13:06 <carnil> waldi: asked the question to test with iommu=off and report back, but no followup when I last checked
20:13:09 <carnil> wait
20:13:21 * ukleinek nods
20:13:23 <carnil> #topic #1152945: (i, ) linux-image-7.2.8+deb14-amd64: nouveau CTXSW_TIMEOUT on GM107 (Quadro K620) kills Xorg channel and freezes desktop
20:13:46 <carnil> ukleinek: asked back if that is a regression and to test with 7.3.y kernel
20:13:50 <carnil> wait as well
20:14:11 <carnil> #topic #1153232: (i, ) linux-image-7.2.15+deb14-amd64: kernel NULL pointer dereference in nouveau during systemd upgrade
20:14:17 <carnil> a new report
20:15:23 <carnil> not exactly sure if he can reproduce as it showed on a system update, involving systemd I assue
20:15:39 <carnil> as it is even stated in the first sentence.
20:16:02 <ukleinek> also involving fwupd
20:16:16 <ukleinek> looking on lore.k.o there is no similar report
20:17:15 <waldi> it killed fwupd in it's path. so something fwupd did to the drm device, nouveau did not like
20:18:04 <waldi> but a simple open on a device should not kill anything
20:20:22 <carnil> hmm
20:20:31 <ukleinek> probably needs some optimism to assume this is fixable.
20:21:00 <ukleinek> maybe reach out to fwupd people asking what that process does to the nouveau device?
20:21:59 <carnil> okay that would be maybe a question to Mario Limonciello to see if he has an idea.
20:22:27 <ukleinek> that's a nouveau guy, right?
20:22:43 <carnil> he is maintainer for fwupd
20:23:12 <ukleinek> then ack
20:23:44 <carnil> I will drop him a mail if he can briefly skimm over the bug (resp. CC him on the bug)
20:24:33 <carnil> #action carnil ask Mario Limonciello (fwupd maintainer) through looping in ##1153232 if he has ideas on what the fwupd process might do to the nouveau device
20:24:47 <carnil> #topic #1153524: (i, u) linux: nfsd 6.12 regression: blocking NFSv4 LOCK set without FL_SLEEP, CB_NOTIFY_LOCK never sent, cross-client waiters stall 30 s
20:25:09 <carnil> I wanted to try to replicate this, but then think if it would be best handled upstream at the linux-nfs list
20:25:54 <ukleinek> in my experience: Don't use different kernel versions on nfs server and nfs client
20:26:07 <ukleinek> (still a bug of course)
20:26:13 <carnil> you will have that all the time in some environments :)
20:26:59 <ukleinek> ah, but that problem occured when the server was upgraded to the version already running on the clients ....
20:28:14 <ukleinek> ack for involving upstream
20:28:26 <carnil> if nobody objects I will try to replicate, but then quickly involve upstream seems best
20:29:02 <carnil> #action carnil tries to replicate #1153524 and involve upstream to clarify
20:29:12 <carnil> #topic 1153534: (i, ) linux-image-6.12.116+deb13-amd64: regression: snd_hda_intel soft lockup in azx_probe_work at boot (Jasper Lake, HP 15s-fq3xxx), system hangs
20:29:22 <carnil> New report. If the problem is reproducible then might be worth asking for a bisection?
20:29:53 <ukleinek> yeah
20:29:55 <carnil> is reported as regressing from 6.12.107 to 6.12.116
20:30:21 <carnil> and is then not the known audio regression which was fixed in 6.12.116-1
20:30:48 <ukleinek> 5401 patches to bisect
20:30:56 <carnil> I will ask the two questions
20:31:17 <ukleinek> great, thanks
20:31:29 <carnil> #action carnil ask if #1153534 is reproduible under 6.12.116-1 and if yes ask for bisection between 6.12.107..6.12.116
20:31:55 <carnil> #topic #1153579: (i, ) list_del corruption with RTL8922AE in MLO mode after some time
20:31:58 <carnil> new report
20:33:08 <ukleinek> I wonder what makes the reporter think it's related to MLO
20:34:23 <ukleinek> #action ukleinek asks the reporter a few questions about #1153579
20:34:28 <carnil> https://lore.kernel.org/all/05d59bc0ceeb4dab92979504626ca868@realtek.com/ migth be related?
20:35:54 <ukleinek> great, will consider that when replying
20:36:29 <carnil> there is a debug patch attached in the thread which might help to further debug
20:36:40 <carnil> and indeed the trace looks same
20:36:45 * ukleinek nods
20:36:45 <carnil> okay thanks ukleinek
20:37:03 <carnil> #topic #1146033: (n, Mu) linux-image-7.1.8+deb14.1-amd64: cannot shutdown laptop
20:37:21 <carnil> From last meeting: @ukleinek followed up asking for a bisection of the non---first-parent history.
20:37:40 <carnil> which was done
20:37:51 <ukleinek> and the reporter found a plausible commit
20:38:10 <ukleinek> # first bad commit: [310326bb7df4bba094a3fc60364c641c547fd923] nouveau: pci: quiesce GPU on shutdown
20:38:48 <ukleinek> https://lore.kernel.org/all/20261025213027.842978-1-marex@nabladev.com/
20:39:49 <carnil> ah nice. Should we mark it forwarded there to keep track of the progress?
20:40:07 <ukleinek> #action ukleinek follows up on #1146033 mentioning https://lore.kernel.org/all/20261025213027.842978-1-marex@nabladev.com/
20:40:55 <carnil> #topic #1149377: (n, u) linux-image-6.12.111+deb13-amd64: Bluetooth no longer works after kernel upgrade
20:41:07 <carnil> this is pending now in 6.12.117-1 (I added a bug closer)
20:41:16 <carnil> #topic #1149988: (n, M) linux-image-7.2.9+deb14-amd64: sound does not work in any 7.2 kernels
20:41:24 * ukleinek nods
20:41:33 <carnil> ukleinek asked if iommu=on/off is the relevant difference between works and works not
20:41:44 <carnil> -> wait
20:41:51 <ukleinek> that was only today, so yes, wait
20:42:00 <carnil> #topic #1151553: (n, u+) linux: fastrpc: Reduce console noise from expected DSP info, reserved memory and session messages (forwarded: git.kernel.org, linux!2135)
20:42:26 <ukleinek> next
20:42:36 <carnil> I think we agreed last week we will not backport to older stable series as the commit was explicitly not meant to be a bugfix nor marked accoredingly, we keep it open for now
20:42:44 <carnil> #topic #1152418: (n, M) linux-base-7.2.11+deb14-amd64: Nouveau doesn't get recognized for hardware acceleration
20:42:59 <carnil> Waiting for reply from reporter.
20:43:17 <carnil> #topic #1152943: (n, ) 6.12.113-1 to 6.12.115-2: MediaSupervisor segfault
20:44:00 <carnil> reporter updated to 6.12.116-1 and will report back if problem reoccurs. I tend to already close it and ask reporter to reopen *if* it reappers
20:44:14 * ukleinek shrugs
20:44:37 <carnil> (I was initally faily confident it was the regression with usb audio devices)
20:44:56 <carnil> ukleinek: I understand you do not agree? We can it keep as well open
20:45:20 <ukleinek> no, I don't think it matters, no strong opinion
20:46:35 <carnil> the problem I see is it might be kept open if reporter will not confirm to be closed. As only said "Currently on the 6.12.116-1 kernel, will update if this recurs"
20:46:46 <carnil> but let's keep it open for now, we can close later
20:46:55 <ukleinek> well, closing makes sure the bug keeps open indefinetly while being obsolete
20:47:21 <ukleinek> agreed on reasoning and procedure
20:47:37 <carnil> #topic #1152950: (n, ) linux-image-6.12.115+deb13.1-amd64: bnxt_en driver stops working unless iommu=pt is used
20:47:58 <carnil> I had an action here last week, but it was not topmost on the list, I have not done it yet
20:48:45 <carnil> I can keep an action to recheck the related commits between 6.12.113..6.12.115 and might ask for a bisection
20:48:52 * ukleinek nods
20:49:19 <carnil> #action carnil checks 6.12.113..6.12.115 for changes to the relevant driver (maybe ask for bisection)
20:49:30 <carnil> #action carnil checks 6.12.113..6.12.115 for changes to the relevant driver (maybe ask for bisection) for #1152950
20:49:41 <carnil> #topic #1153147: (n, ) NFS client: mkdir on existing, uncached dir returns EROFS instead of EEXIST under ro bind mount
20:49:57 <waldi> need to refer to linux-nfs or?
20:50:08 <carnil> waldi followed up on the bug already, question from the reporter to where report the problem
20:50:42 * ukleinek marked it forwarded
20:51:54 <carnil> so we wait to see what upstream say
20:52:01 <ukleinek> no relevant reply upstream yet, so yes, wait
20:52:33 <carnil> ukleinek, waldi, yunseong[m] still have energy or should we skip the remaining normal and wishlist bugs?
20:53:01 <ukleinek> #1153514 seems important and not too complicated
20:53:22 <carnil> #topic #1153514: (n, ) radeon driver fails to find installed firmware due to Debian-specific patch
20:53:30 <carnil> let's look at this one and skip the rest
20:53:43 <carnil> so this only needs to have MR with waldi's changes right?
20:53:50 <waldi> we have two options: fix the patch or drop it
20:53:53 <ukleinek> that's my understanding, yes
20:54:03 <waldi> it is a workaround for machines without firmware
20:54:40 <ukleinek> that patch used to cover more hardwares. I wonder what actually happend to consider it not needed for these any more.
20:55:26 <ukleinek> anyhow, the low bar todo is to keep it and apply waldi's fix
20:56:23 <ukleinek> waldi: Will you push that forward?
20:56:23 <carnil> has bwh expressed an opinion already?
20:57:14 <waldi> i would say we drop it. 13 years of workaround
20:58:23 <ukleinek> #action ukleinek will create a MR dropping debian/patches/bugfix/all/radeon-amdgpu-firmware-is-required-for-drm-and-kms-on-r600-onward.patch
20:59:04 <carnil> ukleinek: please add as well /cc to @benh or reviewer, I think we cannot have multiple reviwers in our salsa gitlab
20:59:17 <ukleinek> ack
20:59:29 <carnil> okay let's skip the rest of the bugs
20:59:46 <carnil> #topic Migration excuses
20:59:59 <carnil> linux, we have to wait, too new yet
21:00:22 <carnil> #topic New upstream versions
21:00:31 <carnil> yunseong[m]: prepared an update for ktls-utils
21:00:55 <carnil> kup has wip and for linux waldi has a MR for 7.4-rc1
21:01:04 <yunseong[m]> Yes, Last week at the MiniDebConf in Cambridge, I met Ben. With Ben's guidance, I created an MR to update the kTLS version.
21:01:24 <yunseong[m]> * the kTLS util version.
21:01:24 <ukleinek> RL events are great
21:01:29 <carnil> very nice :)
21:01:45 <carnil> #topic Merge requests
21:01:52 <carnil> any specific MR do discuss?
21:01:59 <ukleinek> #action ukleinek will look into !2155
21:02:49 <carnil> ok!
21:03:03 <ukleinek> no further urge on my side
21:03:09 <carnil> #topic AoB
21:03:23 <carnil> most important question here, who will chair next weeks meeting
21:03:41 * ukleinek won't attend next week's meeting, will give a talk at my local LUG
21:04:05 <carnil> will it be recorded? ;-)
21:04:15 <ukleinek> carnil: no
21:04:37 <waldi> i should be able to handle it. but will be on a conference the days before
21:05:10 <ukleinek> but will be live-streamed: https://lug-freiburg.de/vortrag-hybrid-veranstaltung-das-entwicklungsmodell-des-linux-kernel.html (in German)
21:05:10 <carnil> waldi: thanks. if you find yourself to not have time, ping me and I will try to take over
21:06:04 <carnil> #action waldi should be able to chair next weeks meeting (fallback: carnil)
21:06:15 <carnil> so I think we are done for tonight
21:06:21 * ukleinek is
21:06:34 <carnil> thanks all for attending and contributing
21:06:37 <carnil> #endmeeting