20:00:01 <carnil> #startmeeting 20:00:01 <MeetBot> Meeting started Wed Nov 26 20:00:01 2025 UTC. The chair is carnil. Information about MeetBot at https://wiki.debian.org/MeetBot. 20:00:01 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic. 20:00:05 <carnil> hi 20:00:23 <bwh> Hi 20:00:28 <ukleinek> o/ 20:00:30 <bwh> Sorry for flaking last week 20:00:35 <carnil> #chair bwh ukleinek waldi 20:00:35 <MeetBot> Current chairs: bwh carnil ukleinek waldi 20:00:51 <carnil> bwh: welcome back :) 20:01:10 <ukleinek> bwh: we were flaky, too. We talked in German and didn't go through all the bugs :-D 20:01:32 <carnil> before we start, is there anything which you would like to bring on the table for discussion? 20:01:57 <bwh> Nothing urgent from me 20:02:43 <carnil> okay then let's just start with what is on the agenda 20:02:52 <carnil> #topic Removal of iptables-legacy 20:03:00 <ukleinek> looking at the calendar, do we skip the meeting on 2025-12-24 and 2025-12-29, or do we pick another day in these weeks? 20:03:46 <bwh> Last time I think we just skipped 12-25 and 01-01 20:03:58 <ukleinek> s/29/31/ 20:04:09 <waldi> hi 20:04:24 <ukleinek> I think skipping is fine 20:04:24 <carnil> ukleinek: ah yes I cannot immagine that any of us prefers a jitsi/irc meeting on 24th instead of staying away from computer :) 20:04:49 <bwh> Whoever prepares the agenda for the following meeting should take care to tell find-recent-urgent to look back a larger number of days 20:05:05 <ukleinek> ack 20:05:32 <carnil> ack, let's not forget to plan for that in the upcoming meetings and fill as well in the dates for the next year 20:06:02 <carnil> so we already drifted away from the topic :) 20:06:12 <ukleinek> I wonder if looking back one week for normal meetings is sensible, I assume things fall through the cracks. But I don't have a better proposal ... 20:06:40 <bwh> OK, so from the feedback it seems like there are a number of false positives for code paths that would use iptables-legacy but are not execued 20:07:23 <bwh> waldi: Do you think you can exclude those, manually or automatically? 20:07:23 <ukleinek> less packages = better, so \o/ 20:07:44 <carnil> #link https://lists.debian.org/debian-devel/2025/11/msg00303.html 20:08:14 <waldi> bwh: i have to recheck all anyway, as i did not note what they actually do 20:08:55 <carnil> so the plan would be to do a refinement and check all first, then re-propose the MBF with the dd-list 20:08:58 <carnil> ? 20:09:04 <waldi> no, then write the bugs 20:09:18 <ukleinek> sounds sensible 20:09:27 <carnil> or so, though another feedback loop would make sense, but fine 20:09:56 <carnil> #action waldi rechecks all candidates and then fills bugs for the iptables-legacy removal 20:09:58 <ukleinek> if the set of packages is strictly less than announced I don't see a need to reannounce 20:10:07 <bwh> I agree 20:10:30 <waldi> all the detection is loosy anyway, as we don't have dependencies to check, but just code that uses it or in most cases runs update-alternatives automatically 20:11:22 * ukleinek nods and holds his breath for the next topic 20:11:32 <carnil> the bug list 20:11:46 <carnil> #topic #1105211: mt7921e: stuck at low transmit power 20:12:27 <carnil> while re-reading the bug I realized there might have just been some confusion on how to properly pass the parameter to the right module. I might be wrong. But asked some questions back to the reporter and gave hints on possible workaround configuration. 20:12:28 <ukleinek> sounds like a wait 20:12:31 <carnil> I propose to wait 20:12:41 <bwh> agreed 20:12:55 <carnil> #topic #1106558: linux-image-6.12.22+bpo-amd64: Wifi module fails to load 20:13:06 <carnil> wait; asked back if still reproducible with 6.12.57-1 as in trixie 20:13:24 <carnil> #topic #1116554: linux-image-6.12.43+deb12-amd64: hard freeze during normal operation with stacktraces in the syslog 20:13:43 <carnil> wait :-/ Asked back if still reproducible with the newest 6.12.57-1 in trixie. Asked as well for tests with kernel in unstable. -> wait now for feedback from reporter 20:13:57 <bwh> OK 20:14:21 <carnil> #topic #1116643: UBSAN: shift-out-of-bounds in .../drivers/gpu/drm/display/drm_dp_mst_topology.c (shift exponent -1) 20:14:55 <carnil> Found potential right upstream report with a proposed patch. Asked Vincent if problem easy reproducible and if he can test the patch. But maybe if someone has spare cycles it would be good to double check that seems a potential good candidate. In any case if Vincent is able to reproduce he will try to test the patch. 20:15:03 <ukleinek> sounds easy 20:15:25 <carnil> ukleinek: so you are going to double check :) 20:15:26 <ukleinek> (that is, to look at the upstream patch and judge if it fixes the problem) 20:15:43 <carnil> ukleinek: yes please do 20:15:46 <ukleinek> #action ukleinek looks at #1116643 20:16:06 <carnil> #topic #1118437: linux-image-6.17.2-amd64: kernel crash in interrupt after receiving an ip packet on veth from xsk from user space 20:16:25 <carnil> we are at v7 of proposed patchset -> Wait for fixes to land IMHO 20:16:43 * ukleinek nods 20:17:09 <carnil> #topic #1120554: linux-image-6.12.48+deb13-amd64: SOF driver fails on Intel Comet Lake - CSE_IPC_RESET_PHASE_1 timeout 20:17:36 <carnil> ukleinek: asked questions back (in particular to confirm firmware-sof-signed is installed and if not install it and test with it) 20:17:59 <carnil> Additionally to test with newer kernels, but so far no reply from reporter 20:18:06 * ukleinek nods again 20:18:18 <carnil> So with as many of our bugs: wait for feedback 20:18:24 <bwh> We should add that to the bug control file 20:18:46 <ukleinek> having many bugs that don't need an action is good1 20:19:06 <carnil> bwh: you mean the information on firmware-sof-signed? 20:19:17 <bwh> #action bwh will add firmware-sof-signed and other firmware packages to Package-Status field of the reportbug control file 20:19:28 <carnil> +1 sounds good 20:19:44 <ukleinek> ..ooOO(Can I have an irssi pluging that simplifies nodding?) 20:20:07 <bwh> carnil: I think we have a (possibly outdated) list of firmware binaries built from firmware-{non,}free but not those from elsewhere 20:20:37 <waldi> should we retrieve all binary packages from non-free-firmware and just add them? 20:21:08 <bwh> That might make sense but I'm not sure whether everything from there is relevant 20:21:11 <ukleinek> maybe review that list. 20:21:24 <bwh> I will certainly start with that list 20:21:37 <ukleinek> /nod 20:21:49 <waldi> does it matter if some not completely relevant packages show up in this list? 20:22:17 <waldi> we talk about 49 right now 20:22:23 * carnil has no idea how long the list is 20:22:27 <carnil> ok 20:23:04 <bwh> Next bug? 20:23:11 <carnil> I guess it would be a good stat and then refine it as needed 20:23:12 <carnil> yes 20:23:14 <ukleinek> it doesn't really hurt, but obviously useless packages (if any) should still be skipped. I think it's easy to spot those 20:23:29 <carnil> #topic #1120598: nfs: ls input/output error ("NFS: readdir(/) returns -5") on krb5 NFSv4 client using SHA2 20:24:35 <carnil> patch in progress upstream (and the problem is only relevant under certain conditions where as Chuck Lever describes it in the proposed patch. So we can simply wait here that patch lands in mainline and then goes to the stable series as needed. 20:24:57 <carnil> #topic #1120602: hyper-v: BUG: kernel NULL pointer dereference, address: 00000000000000a0 20:25:10 <carnil> patch submitted to stable (already in 6.12.58) and pending for 6.1.y 20:25:22 <carnil> so will be in the next uploads 20:25:34 <carnil> -> wait and make sure to add bug closer on stable series import 20:25:42 <bwh> So this can be tagged patch fixed-upstream, right? 20:26:25 <carnil> ah hes potentially. it has not landed yet in older stable series than 6.12.y and it did not affect mainline or upper stable series 20:26:41 <carnil> s/hes/yes/ 20:26:44 <bwh> Well I will tag it patch anyway 20:26:51 <carnil> thanks 20:27:04 <ukleinek> [x] done 20:27:26 <ukleinek> (patch + fixed-upstream) 20:27:27 <carnil> #topic #1120718: linux-image-6.12.43+deb13-amd64: hung tasks on mutex acquisition during rtnetlink ops 20:27:48 <carnil> reporter should try to identify last kernel without problems (has hints on snapshot.d.o service) 20:28:01 <carnil> -> wait for user feedback 20:28:08 <bwh> Did you find out what the OOT module is there? 20:28:29 <bwh> The hang messages should OE taint but no module with those flags is in the module list 20:29:34 <ukleinek> firmware-nvidia-graphics is installed 20:29:34 <carnil> oh I might have missed that because the module list does not contain the information. 20:30:00 <ukleinek> hmm, but that might be autoinstalled?! 20:30:08 <bwh> Oh the later log has not tainted, so never mind 20:31:07 <carnil> ok so we remain with that single action to first wait for the reporter 20:31:30 <bwh> agreed 20:31:38 <carnil> #topic #1121006: linux: reported optimal_io_size from mpt3sas devices results in 4GB raid10 optimal_io_size 20:32:51 <carnil> this one is intersting, there is a upstream thread relating to the problem but which stalled. I was planning to ping mpt3sas folks about it and see if they can chime in and get I/O people's attention. But I was wondering if someone else in the team has other ideas to tackle apart form that path. 20:33:31 <carnil> I believe it would be best to get Christoph's Hellwig's attention on that thread again 20:33:50 <carnil> and best thing I can immagine is to try to bring the people again on that thread 20:33:57 <carnil> it stalled in august afaics 20:34:12 <bwh> yes I think that's a good plan 20:34:46 <carnil> Then I take care of it (as I was in contact for it with Filippo and Moritz Muehlenhoff) 20:34:52 <ukleinek> How das Christoph Hellwig come into play, he's in the stalled upstream thread I assume? 20:34:56 <bwh> yes 20:35:39 <carnil> #topic #1121227: linux-image-6.12.48+deb13-amd64: Black screen with Nouveau TRAP_PROP [RT-FAULT] on NVIDIA GT 230M (GT216M) 20:35:53 <carnil> I think we need a bit to "handhold" the reporter here to get to the information we need 20:36:25 <carnil> Asked to followup with the reportbug collected information but we got then another bugreport against general withouth information 20:36:43 <carnil> se we have neither full kernel logs nor information on other components of the system 20:37:14 <carnil> but can we deduce already something here from the information posted in the first message? 20:37:28 <bwh> I think they might be confused by your not top-posting 20:38:13 <bwh> So, maybe try asking again but without quoting...? 20:38:46 <carnil> and maybe give the concrete commands to run so we get the info on the next message 20:39:18 <carnil> if I should speed up things let me know 20:39:40 <ukleinek> #action ukleinek tries to handhold for #1121227 20:40:02 <carnil> ukleinek: ok perfect, I just cancelled the action was formulating :) 20:40:18 <carnil> #topic #1100105: linux-image-amd64: Issues with Framework audio module 20:40:44 <carnil> this one confuses me tbh. S "now" it work with 6.12.48-1 but fails with newer versions from testing/unstable if my interpretation of the bug log is correct 20:41:29 <bwh> I really think this is flakiness and not any software fix being added or removed 20:41:58 <ukleinek> matches my suspicion 20:41:59 <bwh> Every reboot brings a different random state of the hardware 20:42:52 <bwh> I suppose it's also possible that some quirk not directed at this specific hardware module is also affecting it 20:42:56 <ukleinek> Hmm, that means that finding the problem (or a fix) is really hard 20:43:02 <bwh> yes 20:43:45 <bwh> Has there ever been any upstream contact about this bug? 20:44:07 <ukleinek> the reporter talked to the laptop vendor 20:44:38 <ukleinek> I think upstream wasn't involved yet. 20:44:41 <bwh> right 20:44:53 <bwh> #action bwh will forward #1100105 upstream 20:44:59 <ukleinek> thanks 20:45:05 <carnil> thank you 20:45:21 <carnil> #topic #1116251: linux-image-6.12.48+deb13-rpi: Raspberry Pi Zero W does not boot 20:45:38 <carnil> there is some progress now, bdrung followed up and we have now output from the boot 20:46:02 <carnil> #link https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1116251#42 20:46:18 <ukleinek> right, I intended to look what changed between these kernels regarding CMA but didn't manage to actually do it 20:46:25 <carnil> I skimmed over the changes in the stable series, but found only 46efab01648a04082266115a8e917c3b26b97fa8 20:46:34 <carnil> would that make sense? 20:46:48 <carnil> cf. https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1116251#49 20:47:37 <ukleinek> letting bdrung test with that commit reverted definitively brings us forward 20:48:48 <carnil> okay so we wait for bdrung now to see if the commit reverted fixes the problem 20:49:17 <carnil> #topic #1117959: ipv6_route flags RTF_ADDRCONF and RTF_PREFIX_RT are not cleared when static on-link routes are added during IPv6 address configuration 20:49:36 <carnil> commit is in next, wait that it lands in mainline (and it is candidate for stable series and trickles down) 20:49:52 <bwh> OK 20:49:57 <carnil> #topic #1118349: Hanging the radeon video on AMD APU, reopening #879992 20:49:59 * ukleinek agrees to "Thank you so much Salvatore" 20:50:21 <carnil> closed now, as we really were not able to get on the same page with the reporter :( 20:50:43 <carnil> #topic #1119093: linux-image-6.16.3+deb13-amd64: UBSAN array-index-out-of-bounds in ath5k driver 20:50:53 <carnil> Fix submitted upstream. Wait for landing and monitor for inclusion in the stable series as needed. 20:50:56 <carnil> -> wait 20:51:25 <carnil> #topic #1120280: linux-image-6.12.57+deb13-amd64-unsigned: BUG: ENTRY_TRAMPOLINE stack guard page was hit at (maybe related with usage of make --load-average) 20:51:41 <carnil> questions from ukleinek, wait for feedback from reporter 20:51:47 <ukleinek> /nod 20:51:54 <carnil> #topic #1120854: linux-image-arm64: Enable dma engine for Renesas RZ platform 20:52:04 <carnil> this has a MR associated 20:52:11 <carnil> #link https://salsa.debian.org/kernel-team/linux/-/merge_requests/1720 20:52:17 <bwh> Yes and I should go and review that 20:52:28 <carnil> someone with better knowledge than me for arm64 feels to take care of reviewing it? 20:52:34 <carnil> ah bwh :) 20:53:00 <carnil> #action bwh reviews MR (!1720) for #1120854 20:53:23 <carnil> #topic #1121032: thunderbolt_net: missing ndo_set_mac_address prevents bonding mode 802.3ad 20:53:25 <ukleinek> Is bwh for Renesas now? The footer in the last mail suggests that :-D 20:53:35 <ukleinek> s/for/working for/ 20:53:50 <bwh> Huh? 20:54:10 <carnil> proposed patch waits for confirmation from reporter, but IMHO looks promising. So wait for feedback and wait it lands in mainline 20:54:23 <ukleinek> bwh: ah, no, the renesas guy uses strange quoting and so his footer looks like it was added by you 20:54:25 <carnil> ukleinek: really? 20:54:42 <carnil> #1121211: UBSAN: array-index-out-of-bounds in /build/reproducible-path/linux-6.17.8/drivers/clk/samsung/clk-exynos-clkout.c:178:18 20:54:49 <carnil> #topic #1121211: UBSAN: array-index-out-of-bounds in /build/reproducible-path/linux-6.17.8/drivers/clk/samsung/clk-exynos-clkout.c:178:18 20:54:52 <carnil> (sorry) 20:55:02 <carnil> reported upstream, patch tested, will be in 6.19 20:55:26 <carnil> (will later be backported to stable series, but not urgent, needs just monitoring for adding bug closer) 20:55:43 <bwh> I just added a patch tag to the last 2 20:55:43 <carnil> and just in time before the 1h exceeds we are done with the bug list :) 20:56:00 <carnil> bwh: thanks, I seem to have been bit negligent on proper tagging :( 20:56:09 <ukleinek> \o/, thanks carnil, great preparation 20:56:16 <carnil> #topic Migration excuses 20:56:33 <carnil> we are waiting for firmware-nonfree (too-young) and src:linux (as well to young, uploaded today) 20:56:42 <carnil> #topic New upstream versions 20:56:46 <carnil> nothing?! ;-) 20:56:58 <bwh> I think this may be broken due a uscan change (discussed on d-devel) 20:57:20 <ukleinek> 🙈 20:57:31 <bwh> but we might also really be up-to-date 20:57:46 <bwh> firmware-free should definitely be there even though none of the upstream changes are relevant to it 20:58:52 <carnil> yes right, that suprised me but I did not bother to look what might have been the cause. Se we might double check with the uscan change. 20:59:18 <carnil> #topic Merge requests 20:59:38 <carnil> the only one additionally which I see ready yet would be !1670 20:59:54 <bwh> waldi: Please answer my last comment on !1615 21:00:19 * ukleinek forgot what we said about a debug flavor. Did we consider that a good idea? 21:00:32 <KGB> 03linux 05debian/latest 06Bastian Blank * [update] merge request !1615: Support for drivers in Rust * 14https://salsa.debian.org/kernel-team/linux/-/merge_requests/1615 21:00:59 <bwh> waldi: Thank you! 21:01:15 <bwh> Then I think I can rebase and merge that one now? 21:01:18 <waldi> ukleinek: it makes sense, but IMHO we need to wait until we can get rid of the extra -cloud build 21:01:29 <carnil> ukleinek: maybe we should put that on the agenda topic as separate item to have a proper look? 21:02:06 <waldi> bwh: i would like to check again. i can then merge it 21:02:20 <bwh> waldi: OK 21:02:20 <ukleinek> if waldi has a strong opinion on the debug kernel, that's fine for me 21:02:38 <waldi> ukleinek: it's more the extra build time and most important space 21:02:53 <waldi> we break buildds 21:03:06 <ukleinek> IIRC I had doubts about its usefulness 21:04:05 <carnil> and some concerns about beeing clear to users that it's not used by default. But IIRC this was clarified how we can deal with it. 21:05:08 <carnil> ok but I guess most MRs are ongoing yet, and thanks bwh and waldi for the "support for drivers in rust" one 21:05:19 <carnil> anything else on the MRs? 21:05:24 <bwh> not from me 21:05:29 <waldi> not from me 21:05:32 <carnil> #topic AoB 21:05:46 * ukleinek still has 1337 on his list, but is stuck in contacting MediaTek 21:05:52 <bwh> My turn to chair? 21:06:26 <carnil> I hear no objections 21:06:32 <bwh> :-) 21:06:41 * ukleinek was about to ask the same, but it seems its my turn in two weeks then. 21:06:51 <carnil> thank you bwh! 21:07:04 <carnil> so I think we can conclude the meeting 21:07:10 <bwh> Thanks carnil 21:07:13 <carnil> #endmeeting