19:00:12 #startmeeting 19:00:12 Meeting started Wed Jul 9 19:00:12 2025 UTC. The chair is carnil. Information about MeetBot at http://wiki.debian.org/MeetBot. 19:00:12 Useful Commands: #action #agreed #help #info #idea #link #topic. 19:00:19 Hi 19:00:56 hi all, welcome to todays meeting, first an apology had bit stressfull afternoon only finished some minutes ago ammending the meeting preparation 19:01:17 #chair carnil bwh ukleinek waldi 19:01:17 Current chairs: bwh carnil ukleinek waldi 19:01:23 hi 19:01:40 before we dig into the bug list, I have a slight offtopic question 19:02:10 #topic Interest to look into ITS mitigation patches for 5.10.y 19:02:24 ITS? 19:02:27 a while back ITS mitigations were applied to various stable kernel series, but not back to 5.10.y 19:02:35 context: https://lore.kernel.org/stable/aEQHkmGXOel3BcOF@eldamar.lan/ 19:02:40 https://lore.kernel.org/stable/20250610-its-5-10-v1-0-64f0ae98c98d@linux.intel.com/ 19:02:51 I meant to look at this but I think I forgot 19:03:05 questions goes maybe mainly to bwh: is there interest to look into these or should we state we do not have the capacities? 19:03:38 Pawan wrote to look into that 19:03:40 I can spend some time to review and do a little testing 19:04:10 #action bwh will review and test ITS mtiigation backport to 5.10 19:04:10 ukleinek: he did and posted a RFC series, but would explicitly want some feedback before submitting for real and Greg accepting them 19:04:17 ack thanks bwh 19:04:23 so ready to go to the buglist? 19:04:28 I think you mentioned another mitigation that was missing for 5.10? 19:04:31 03initramfs-tools 05debian/latest 06Benjamin Drung * [update] merge request !172: Draft: Use 3cpio for unmkinitramfs/lsinitramfs if available * 14https://salsa.debian.org/kernel-team/initramfs-tools/-/merge_requests/172 19:05:17 bwh: yes 19:05:34 the Native BHI vulneraility mitigations were as well not backported 19:05:47 so the question was if it is worth at all since it's in similar class 19:06:01 right 19:06:15 well I can look at that too and see if it's feasible 19:06:24 many thanks Ben 19:06:43 I have done some of these backports myself before, but just haven't been "in the loop" regarding recent CPU mitigations 19:06:56 #action bwh will look at feasibility of mitigated BHI in 5.10 19:06:59 yes same for me, I was not included neither this time 19:07:23 #topic #1088522: nouveau: Unable to boot with 3 monitors on Nvidia GPU 19:07:35 @ukleinek pinged the reporter about further debugging, waiting for reply. -> i.e. wait 19:07:47 * ukleinek nods 19:08:01 OK 19:08:04 #topic #1100153: linux: hangs on resume from suspend or hibernation with 6.12-6.13 on ThinkPad X1 Carbon Gen 10 19:08:18 Some incosistent results from bisect. ukleinek sked back to test three commits explicitly to narrow down if the bisect is correct respetively to gather a better understanding. 19:08:36 I think here as well there is at the moment nothing more we can do until hearing back that the bisect was done correctly 19:08:36 no reply :-\ 19:08:48 * ukleinek nods again 19:09:01 the next one at least is actionable :) 19:09:14 #topic #1107142: linux-image-6.12.27-amd64: mpt3sas LSI SAS2008 [1000:0072] probe failure - BAR 1 reservation error 19:09:28 I would have hoped that upstream chimes in 19:09:31 we got boot logs with the requested variants booting with pci=realloc=off and not 19:09:57 (upstream is as well in the loop on that) 19:10:41 I'll ping Bjorn again in irc, where we talked before and he requested the two bootlogs 19:10:43 ukleinek: would it help sorting out the differnces and re-iterate them into a reply to gather the attention of the upstream people on the linux-pici list? 19:11:10 #action ukleinek will ping Bjorn again in irc regarding #1107142 and the provided boot logs 19:11:25 * ukleinek was about to type something similar 19:11:47 #topic #1107521: linux-image-6.12.27+bpo-amd64: ath12k_pci errors and eventual kernel panic starting with linux-image-6.12.27+bpo-amd64 19:12:16 we asked here the rporter to bisect IIRC, but reporter when the other way and started to cherry-pick patches 19:12:25 Reporter said he continues to test his patched branch to make sure it is not a false positive with the applied patches 19:12:57 I would suggest we wait maybe one more week and then ask back on the status (and tag it moreinfo then) 19:13:02 other ideas welcome :) 19:13:40 ack for waiting 19:14:14 #topic #1108082: linux-image-6.12.32-amd64: Crash 19:14:23 Love the subject line 19:14:40 another waiting task. Reporter sucessfully set up a netconsole, now waiting that the crash occures to gather the logs 19:14:46 bwh: ack, yes maybe we should retitle it 19:15:09 but so far we do not even know where it crashes 19:15:37 we hoped to gather more actionable information by getting the logs through the netconsole 19:15:46 right 19:15:56 so I have no idea what to write into subject 19:16:23 It's ok to keep the (technically correct) subject for now IMHO 19:16:36 in any case, action is "wait" 19:16:44 * ukleinek nods once more 19:16:56 #topic #1108168: linux-image-6.12.32-amd64: black screen, then crash in nouveau when rebooting 19:17:25 thanks bwh for digging into it. bwh asked back to Vincent if he can reliably reporduce which would help for further debugging afater the analysis. 19:17:27 +1 for `bts tag 1108168 + moreinfo` 19:17:55 ah, already happend 19:17:55 It is... 19:18:16 ..ooOO(Even in the agenda! :-) 19:18:33 yes my fault 19:18:49 In die Ecke mit Dir! 19:18:59 :-) 19:19:30 so wait ... 19:19:51 while we wait we can do some more useful things ;-) 19:19:59 #topic #1108494: USB 3.2 connector Gen 1 port not working, kernel panics with mouse on right USB port 19:20:12 Waiting for reporter to do the additional requested testing. 19:20:15 right 19:20:35 * ukleinek nods 19:20:37 #topic #1108598: amdgpu: incorrectly redrawn screen after resuming from hibernation with Ryzen 5 2400G 19:20:41 03initramfs-tools 05pipeline 06Ben Hutchings 890166 * [27 minutes and 16 seconds] 03success (shellcheck: success; blhc: success; missing-breaks: failed; piuparts: success; extract-source: success; build i386: success; build source: success; build: success; test-build-all: success; reprotest: success; test-build-any: success; lintian: success; autopkgtest: success) 19:20:46 this one might benefit from some input. 19:21:03 the repoter (AFAIU) said he would be fine to just run the backports kernel 19:21:18 not sure if we should be satisfied with that or try to invest more time 19:21:39 OTOH we have so many amdgpu related bugs 19:22:01 we can mark the issue fixed with the bpo kernel version and it moves away from our radar 19:22:32 sounds good. If only all amdgpu bugs were fixed on 6.12 .... 19:22:55 * ukleinek agrees that our time is better spent elsewhere 19:23:13 I don't have any suggestion 19:23:15 #action carnil marks #1108598 as fixed with 6.12.32-1 19:23:22 ^^ so let's do that 19:23:35 maybe also tell him that we spend our energy elsewhere 19:24:15 #action carnil tells reporter of #1108598 that if running bpo is fine for reporter we move on on the issue and close it with 6.12.32-1 19:24:25 ukleinek: ^^ better? 19:24:46 #topic #1108965: linux: Enable RTL8851BE module 19:24:58 sure, I would have trusted you about that even without explict statement 19:25:30 bwh: prepared merge request. Intended to take it for merging, as I intent to upload 6.12.37 once released with the new TSA mitigations. 19:25:42 (along with the other one bwh assigned already to me) 19:25:59 03linux 05debian/latest 06Uwe Kleine-König * [approved] merge request !1570: rtw89: Enable RTW89_8851BE, RTW89_8852BTE as modules * 14https://salsa.debian.org/kernel-team/linux/-/merge_requests/1570 19:26:04 if that's fine with everybody obviously. 19:26:15 OK 19:26:28 +1 19:26:46 #topic #1104269: linux-image-6.12.22-amd64: amdgpu flip_done and commit wait timeouts during normal use - causes session lockup 19:26:56 03linux 05debian/latest 06Uwe Kleine-König * [approved] merge request !1571: rtw88: Enable RTW88_8821AU, RTW88_8812AU, RTW88_8814AE, RTW88_8814AU as modules * 14https://salsa.debian.org/kernel-team/linux/-/merge_requests/1571 19:27:06 there is a legtimate question from the reporter: 19:27:33 he finds there are so many flip related issues upstream in amdgpu, is it still worth reporting a separate issue? 19:27:57 from the reporter's POV I'd claim yes. 19:27:59 I awas not able to find an exact duplicate, but have not spend a lot of time last week on it 19:28:44 ukleinek: then we suggest the reporter to still fill an issue, ok? 19:28:50 I think so 19:28:53 yupp 19:29:08 should I since I was already in contact. 19:29:35 The issue linked is reported as present up to and including 6.15.3, while the reporter says 6.15.4 is stable. So *maybe* it is the same and is fixed in 6.15.4, but it could well be different. 19:29:36 #action carnil follows up on #1104269 asking to still fill a new upstream issue and report back. 19:30:01 For upstream having active reporters/testers is also beneficial 19:30:59 I just realize again, it's not the same reporter, but we discussed that last time I think already 19:32:10 so let's do as with action above and let him report upstream a new issue 19:32:15 OK 19:32:25 #topic #1108324: linux-image-6.15-rt-amd64: i915 module nonexistent 19:32:33 Progress: https://lists.debian.org/debian-kernel/2025/07/msg00098.html 19:33:24 Right. There's enough information there that I could consider writing a patch myself. But I wouldn't be able to test until I get home. 19:34:07 the reporter of that bug should be a willing tester 19:34:44 (Without implying anything about the timeframe for a patch) 19:35:08 bwh: want to formulate this as an action? 19:35:55 #action bwh will try to make progress on i915 vs PREEMPT_RT, possibly by proposing his own patch 19:36:12 bwh: thanks! 19:36:31 #topic #1108860: linux-image-6.1.0-34-amd64: Wireguard fragmentation fails with VXLAN since kernel 6.1.0-34, causing network timeouts 19:36:34 ukleinek: As this is going to be quite a big change I think I would feel bad about asking someone else to test it when I can do so 19:36:59 bwh: fair 19:37:16 ..ooOO(beta tester then) 19:37:48 reporter confirmed they are going to try to bisect, but will need time as servers are in production. So we have to wait here, but I expect we will get back with a result of finding offending changes. 19:37:51 Hmm, I saw a recent kernel fix related to PMTU discovery but I don't know if it could be related t othis 19:38:15 bwh: in version after 6.12.35 and 6.15.4? 19:38:24 actually not that recent but I saw it recently 19:38:57 funny, I think some freifunk people also do vxlan over wireguard, they might be affected, too. 19:39:57 bwh: the one change I see related to PMTU would be b88786ea2c8f3d5ca1c7c9a69c0b0b7befdfbae8 tunnels: Accept PACKET_HOST in skb_tunnel_check_pmtu(). 19:40:06 and that falls exactly in the range the reporter claims it breaks 19:40:29 above is in 6.1.134, and reporter said 6.1.133-1 was fine and 6.1.135-1 broke 19:40:37 That is much more likely to be it 19:40:56 s/DevCamp/DebCamp/ 19:41:06 So let's ask them to test 6.1.$new with b88786ea2c8f3d5ca1c7c9a69c0b0b7befdfbae8 reverted? 19:41:54 that would make sense indeed and would shortcut a bit the testing 19:42:09 I think I would just ask that 19:42:17 👍 19:42:44 #action carnil ask reporter of #1108860 to test additonally 6.1.y + revert of b88786ea2c8f3d5ca1c7c9a69c0b0b7befdfbae8 19:43:04 oh we are progressing fast timewise 19:43:12 #topic #1100105: linux-image-amd64: Issues with Framework audio module 19:43:29 Feedback from the reporter arrived. Removing moreinfo tag to put it back on our radar. 19:43:55 ukleinek: do you have ideas on that? 19:44:51 hmm, no, not even an idea if this is a software or hardware/firmware problem. 19:46:10 it's strange that 1 out of (at least) 3 usb devices has problems that cannot be cured by neither replugging nor bus driver reload. 19:46:25 yes 19:47:19 is kexec easy to do for the reporter? That could give a hint if it's software or hardware related 19:47:52 kexec brings its own set of driver bugs 19:48:19 or might it be related to userspace somehow? Then a systemd restart? (There was something like that, wasn't it?) 19:49:09 Maybe restart the wireplumber service? 19:49:22 (assuming that's being used) 19:50:05 ukleinek: do you mean systemctl deamon-reexec? 19:50:36 #action ukleinek to follow up on #1100105 with systemctl daemon-reexec, stopping wireplumber before suspend ... 19:50:45 It's not totally clear to me now whether the kernel is detecting the sound device or not... 19:51:44 carnil: no, that's not it. IIRC there was something that does the userspace-part of system shutdown and then make systemd start all services again. 19:52:14 systemctl soft-reboot, I think 19:52:25 yes, just found that, too. 19:52:50 That does seem worth trying 19:52:55 ack 19:53:35 * ukleinek will take care 19:53:52 have you still enough energy? we are approaching 22:00 so either skip the remining bugs or conclude the list and move to MRs and AoB 19:54:35 #topic #1108037: linux-perf: perf report --demangle only works on some(?) symbols 19:54:39 I'm OK to continue through the list 19:54:50 * ukleinek is still ready to tackle more, but won't argue if we complete the meeting soon 19:55:02 reporter commented today with further information. bwh I saw you re on it, correct? 19:55:07 This looks like Clang introduced an extension that the GNU toolchain doesn't yet support 19:55:21 I will continue to take care of this 19:55:56 #action bwh takes care of following up on #1108037 19:56:08 #topic #1108294: kernel 6.12 breaks cgroup awareness of openjdk 21 19:56:35 MR from ben, ready to go, will merge after the meeting. Side-question was if we should report regression upstream, with some doupts though 19:56:44 (which I see ben as well already did :)) 19:57:14 #topic #1108496: linux-image-6.12.30+bpo-amd64: USB audio glitches: retire_capture_urb 19:57:24 waiting for user feedback after questions from Ben 19:57:39 #topic #1108609: Very often fails do detects external monitors connected via Dock and USB-C 19:57:53 issue is reported upstream, https://gitlab.freedesktop.org/drm/amd/-/issues/4384 but so far no action there 19:57:58 -> wait? 19:58:15 * ukleinek nods (and considers creating a macro for nodding) 19:58:18 OK 19:58:54 last src:linux issue 19:59:03 #topic 851404: linux-image-4.8.0-2-amd64: Internal speakers don't work after standby on a Toshiba z20t-c-121 19:59:38 questions back from ukleinek, no reply from reporter so far, tagged moreinfo until getting a reply 19:59:39 * ukleinek did `bts tag 851404 + moreinfo` 19:59:56 dang, already was tagged 20:00:20 #topic 20:00:20 #1109015: (G, ) Wifi not working anymore with new atheros firmware from experimental 20:00:27 this is new from today 20:00:35 but only affecting experimental 20:00:42 #action bwh will forward #1109015 upstream 20:01:06 thanks :) 20:01:08 #Topic 20:01:08 #1108250: (n, M) firmware-intel-graphics: i915 timeout errors after firmware upgrade 20:01:21 dang, sorry for typo 20:01:43 waiting for more information from reporter (questions from bwh) 20:01:45 #topic #1108250: (n, M) firmware-intel-graphics: i915 timeout errors after firmware upgrad 20:01:46 I think this is actually driver flakiness because the GPU doesn't use firmware, but I may be mistaken. I asked for more info including the PCI ID 20:02:01 i.e. wait now 20:02:04 yes 20:02:25 bwh: we skip the two initramfs-tools issues, as you are on them, correct? 20:02:31 OK 20:02:42 #topic migration excuses + new upstream versions 20:03:04 All of the new upstream versions are either in experimental or in an MR (linux) 20:03:21 as mentioned above would like to to a 6.12.37 upload once it is out. 6.12.35-1 is now 6 days in unstable, wonder if I should ask release team first to migrate to testing then only go for 6.12.37 20:03:27 or irrelevant (firmware-free) 20:03:28 or just superseed 6.12.37 20:03:38 (superseed with) 20:04:03 bwh: thanks for !1569 (the MR for 6.16-rc5) 20:04:29 bwh: yes correct 20:04:31 carnil: I agree with getting the latest version into testing before superseding it 20:05:15 bwh: then will ask release team to make migration time shorter (+ asking kibi for d-i plans) 20:05:26 sounds good 20:05:43 I have btw, no idea how we are approaching the full freeze or not, but maybe you have more insights from Debcamp :) 20:06:08 I haven't talked to the release team so far 20:06:16 ok 20:06:23 so semi-last topic 20:06:28 #topic merge requests 20:06:37 any particular outstanding there to look at? 20:06:48 https://bugs.debian.org/release-critical/ says 157 rc bugs left 20:07:14 Review of the 6.16 update would be appreciated 20:08:07 I think I missed to look into the arm64 MRs that I promised to handle last week. Luckily they don't appear in the list, so nobody notices 20:09:03 ukleinek: you get a new chance this week :) 20:09:10 bwh: I'll try 6.16 on my mediatek board, wanted to test that anyhow 20:09:25 * carnil looks at the ones related tagged 6.12 backport needed along with the next upload 20:10:10 ..ooOO(But the PWM is broken on that board, have to care for that first for some reasons) 20:11:15 I think we are almost through then 20:11:29 #topic AoB? 20:11:37 how is Debcamp, bwh + waldi? 20:11:41 ;-) 20:12:16 not yet, i'll be there in friday 20:12:18 We should maybe discuss the RFC about debug packages, but not in this meeting - either on the list or at next week's meeting 20:12:38 DebCamp is pretty nice and relaxed from my point of view 20:12:54 bwh: ok (re the RFC on debug packages) 20:13:45 fwiw, since the next meeting is on jitsi, I might not be able to attend, not sure 100% yet 20:14:02 can we please stop full quotes in bug reports? just today i found several instances of hundreds of lines of log, just to say that i have nothing to say about the quoted log 20:14:23 I need to go catch a bus now, but I am willing to be chair next week if it's my turn 20:14:26 who is taking care of chairing next week? 20:14:31 i might also be in transit, not sure when i'll be home 20:14:43 #action bwh will chair next weeks meeting 20:14:50 bwh: ok catch the bus! 20:14:58 i will end the meeting then 20:15:02 #endmeeting