18:01:36 #startmeeting 18:01:36 Meeting started Fri Mar 22 18:01:36 2013 UTC. The chair is davej. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:01:36 Useful Commands: #action #agreed #halp #info #idea #link #topic. 18:01:37 #meetingname Fedora Kernel meeting 18:01:37 The meeting name has been set to 'fedora_kernel_meeting' 18:01:37 #meetingtopic Fedora Kernel meeting 18:01:41 #addchair jforbes jwb 18:02:25 Who's around ? 18:02:33 * jforbes 18:03:30 I suspect jwb will turn up soon. jforbes, want to jump the queue and do an f18 overview first ? 18:03:41 syre 18:03:43 sorry, here 18:03:54 fashionably late 18:04:08 * nirik is lurking around. 18:04:14 ok, let's do the usual order then 18:04:28 f17 then? 18:04:34 jwb: yep. how much has F17 sucked the life out of you this last week ? 18:04:41 heh 18:04:52 so F17 finally had 3.8.3 pushed to stable yesterday 18:05:01 this is after about a month of not getting any update to stable 18:05:14 the bug count is around 360 or so, which is down but not great 18:05:29 worth noting those are also f18 bugs 18:05:42 yeah. i expect some bugs to filter in from the rebase, but hopefully we caught most of them in f18 and fixed or at least known 18:06:31 we'll march along the 3.8.y train until it runs out. probably have a 3.9 rebase, but we'll have to see 18:06:45 could use a bit wider testing/karma reporting on f17 overall 18:06:48 i think that's about it 18:07:01 ok. jforbes, you're up 18:07:04 sure even... So we have 3.8.4 in updates-testing now. There was an annoying issue with display resolutions on 3.8.3-201 and that was fixed with the 3.8.3-203 update on Monday. If you have an issue with incorrect display resolution, update. 18:07:16 heh, typed that earlier and sat on it. 18:07:49 We still have a bit of a massive bugcount in F18, and there aren't any silver bullet, will close a lot of them fixes, they are pretty well spread out 18:08:09 Right now sitting at 423, though a few of those will close with 3.8.4 18:08:14 that display resolution thing only affected intel? or was it more wide? 18:08:43 nirik: that issue was intel only, there are a couple of other issues with other graphics 18:09:01 * nirik nods. ok 18:09:31 Trying to track down what we can. 18:10:03 I think that's it right now 18:10:09 so last month, we had 239 f18 bugs opened, and we closed out 99. this month, 184 opened, 110 closed. which feels like an improvement. 18:10:50 Yeah, it is when you look at it that way. Doesnt always seem so just looking at the bug lists though 18:11:08 also the number of f17 bugs incoming seems to have dropped dramatically. I don't know if that's us doing better, or just people migrating to f18 18:11:25 given the utter lack of karma... they're moving 18:11:33 we aren't doing anything different between f17 and f18 18:11:34 yeah, seems plausible 18:11:40 and the bugs are growing in f18 18:12:08 Good to see that F18 is compelling enough to upgrade 18:12:17 if bugzilla wasn't so shit, I'd like to just lump 17/18 bugs together 18:12:38 because I'm sure we have dupes amongst those 18:13:12 Yeah, might be good next week for me to sort by subject and compare f17/f18 directly 18:13:23 we have been closing some of them out (and leaving the newer one open, unless the older has more/better debugging info), but it's not fun work, so we haven't done much of that. 18:13:41 jforbes, i briefly looked at that today 18:13:57 it was.. not trivial and not something i was prepared to do on a friady 18:14:00 friday 18:14:21 we could still use abrt being smarter about filing things too 18:14:31 i've been trying to dupe across releases as i see stuff come in, but hard to do sometimes 18:14:38 like if it's an oops in drivers/wireless, *we* shouldn't have to do the reassigning to linville 18:14:41 ditto xorg-* etc 18:15:12 I'd really like to cut down on the triage bullshit like that so we can spend more time doing useful things 18:16:14 I would be very much in favor of that 18:16:49 ok, on to 19/rawhide ? 18:17:00 Yeah, nothing else on 18 here 18:17:49 now that we're branched, we starting to see f19 bugs. most interesting thing so far seems to be the new intel_pstate cpufreq driver is uh, in need of more fixing. 18:18:56 we're still on -rc3 in both 19/rawhide. I've not been doing daily snapshots, mostly because I've been busy actually chasing down a number of problems upstream so that we get them fixed for rc4 18:19:07 rc4 should land this weekend sometime, so I'll package that up on Monday 18:19:45 think that's about it 18:20:02 questions ? open floor ? 18:20:46 so, given that slub_debug=- makes most debug kernels ok... 18:20:57 is it still worthwhile to do nondebug ones? 18:21:02 I noticed that there were a few nodebug builds that didn't actually correspond to rawhide builds. They picked up new stuff that hadn't been in a build yet. 18:21:16 I think this is a good thing, but it was a bit surprising. 18:21:16 they're triggered off of commit 18:21:51 nirik, if it helps people testing, probably. adding a repo to f18 and updating is easier for people than editing a cmdline 18:21:53 brunowolff: it's a weird situation, usually triggered by when I think a build is happening, or often when I see a build finished message from koji 18:22:09 jwb: true. 18:22:18 brunowolff: if there were commits between when they buid was submitted and when I do a build, you end up with those commits in the nodebug build 18:22:50 So far 3.9 has been working well for me. I had some issues that I wasn't sure about, but ended up being dracut related. 18:23:12 same here. 18:23:14 brunowolff: yeah, there's been a few uglies there, that should be behind us now. 18:24:36 Do you guys care about slow path warnings? 18:25:05 not sure what those are ? example ? 18:25:50 I noticed during a reboot in the last week there was about that looked like a traceback, but mentioned something about slowpath. 18:26:07 I assumed it was a warning that may or may not be of interest to someone. 18:26:10 no idea without seeing it. 18:28:20 I have one still in /var/log/messages, its about 30 lines. Should I post it here? 18:28:27 fpaset 18:28:30 er, fpaste 18:28:58 Where is our fpaste server? (I haven't used it yet.) 18:29:08 paste.fedoraproject.org 18:29:11 paste.fedoraproject.org 18:29:13 or yum install fpaste, then cat foo | fpaste 18:30:43 I'll have it soon. 18:31:16 http://paste.fedoraproject.org/5813/97705713 18:31:53 ah. that got fixed 18:31:56 but in general, yes 18:32:11 OK. Next time I'll file a bug. 18:32:34 the warn_slowpath_common thing is a red herring there. the interesting part is the flush_to_ldisc 18:32:51 warn_slowpath_common will be in every trace that comes from a WARN() statement 18:34:05 OK. I thought the warning was about the slow path itself. 18:34:16 interesting how that bug started showing up in every day use for people after being there for a while. Fuzzing with trinity picked up that a while ago. 18:34:50 there's a bunch of tty stuff still pending too. all these years later, tty layer still sucks. shocking I know. 18:35:41 ok, anything else ? or shall we wrap up early ? 18:35:51 That was it from me. 18:35:58 * nirik has nothing more. 18:36:46 ok, let's call it done. 18:36:56 thanks for coming guys 18:37:01 #endmeeting