18:03:03 #startmeeting 18:03:03 Meeting started Fri Jun 22 18:03:03 2012 UTC. The chair is jwb. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:03:03 Useful Commands: #action #agreed #halp #info #idea #link #topic. 18:03:14 #meetingname Fedora-Kernel 18:03:14 The meeting name has been set to 'fedora-kernel' 18:03:23 #meetingtopic Fedora Kernel 18:03:31 #topic init 18:03:43 anybody other than me around? 18:03:51 jforbes said he would possibly be late 18:03:57 davej ? 18:03:57 yeah, I'm here 18:03:58 I am mostly planning on lurking. 18:04:20 release runthrough ? 18:04:22 sure 18:04:25 #topic F15 18:04:35 [dead parrot joke] 18:04:56 F15 finally had it's final update pushed today. it wound up at the equivalent of 3.3.8, which was the last stable kernel in the 3.3 series 18:05:16 aside from that, a lot of bug cleanup earlier in the month. it goes EOL next week 18:05:35 99 bugs at eol is the lowest we've ever had iirc. 18:05:42 sweet. progress :) 18:05:50 anything else for F15? 18:06:06 will be interesting to see if any of those continue to migrate to be 16/17 bugs 18:06:07 #topic F16/F17 18:06:19 davej, some did during the cleanup 18:06:29 yeah, sometimes people just need a kick 18:06:40 f16/f17 are currently at 3.4.3 18:06:48 there's been a couple of notable issues with 3.4 18:07:14 namely iwlwifi, UDL/multiseat interaction, and apparently a bunch of suspend/won't boot regressions 18:07:37 581 f16 bugs open, which is pretty harsh 18:07:38 airlied suggested we disable the DRM udl module this morning to fix the second thing, so we did that 18:07:52 davej, right. those need going through again i think 18:08:02 yeah 18:08:37 we also had some miscommunication and systemtap lost some functionality 18:09:10 Anton Arapov submitted some backport patches for uprobes that are in 3.5 and upstream -tip tree 18:09:39 it's good to at least see that moving in the right direction, instead of utrace coming back 18:09:39 those are in Fedora git now, so hopefully systemtap is back to whatever state it was or was planning to be with uprobes 18:10:08 yeah, definitely. the patches should be pretty short-lived 18:10:24 3.4.4 should be out soon. we'll get that included as well 18:10:57 there are a few packaging type bugs in F17 at the moment related to custom kernel builds 18:11:11 if someone in the community wanted to take a look at those and clean them up, that might be nice 18:11:49 i know one has a patch attached, so that probably just needs committing. the other is failure to build if you turn off EFI support. the vgaarb patches in F17 need fixing for that 18:12:59 something we've started doing as we're sweeping through the 16/17 bugs (though just 17 so far) is adding tags to the whiteboard to indicate the subsystem the bug belongs to. This is something else we'd appreciate help on. 18:13:38 yeah. we don't have set standards yet, but they'll probably work themselves out as we continue to do this 18:14:42 anything else we want to mention on f16/f17? 18:15:02 don't think so 18:15:06 #rawhide 18:15:11 oops, heh 18:15:15 #topic Rawhide 18:15:29 chugging along. at 3.5-rc3 18:15:35 3.5 seems pretty unexciting now that we're into -rc. It was pretty ugly early on. 18:15:50 yeah, though i think you had the worst experiences because of trinity 18:15:59 indeed. 18:16:15 a number of silent memory corruption bugs now got fixed due to that 18:17:37 as always, we need more people testing the rawhide kernel 18:17:53 It's working well for me. 18:17:54 there's nothing from a technical perspective that would prevent anyone from running it on f16/f17 18:18:18 I've not seen teh 18 schedule yet, how long until we branch ? 18:18:49 7th August. 18:18:51 ok 18:19:51 summer is a bit tardy with everyone at conferences etc, but I think we can still squeeze another rebase into 18 before final freeze. 18:20:05 I see the branch on July 31st on Robyn's calendar. 18:20:10 yeah, a bit of time yet. which is good because if the Secure Boot feature is approved we have a number of things to work on 18:20:19 brunowolff: I was looking at https://fedoraproject.org/wiki/Releases/18/Schedule 18:21:23 I use http://rbergero.fedorapeople.org/schedules/f-18/f-18-all-milestones.ics 18:21:43 I don't know which is correct, but they should be the same in theory. 18:22:00 I point this out to Robyn. 18:25:09 I think that's about it for rawhide ? 18:25:23 yeah, think so 18:25:33 i have one topic to cover 18:25:42 anyone have something else they want to get to first? 18:25:57 worth mentioning that the autoqa stuff is starting to take shape 18:26:09 #topic Kernel QA 18:26:21 right. so we have the kernel-tests git tree with actual tests in it now 18:26:27 you can grab that at git.fedorahosted.org/git/kernel-tests.git 18:26:58 it's basic right now, but we have some lofty ambitions to flesh it out over time into something even more useful 18:26:58 the default set of tests runs fairly quickly for now 18:28:28 #info kernel test repo can be found at git.fedorahosted.org/git/kernel-tests.git . contributions welcome 18:28:40 anything else on that? 18:28:57 props to jforbes for doing most of the lifting to get that off the ground. 18:29:00 that's all I got 18:29:14 ok 18:29:24 #topic Kernel updates and bodhi 18:29:36 so just wanted to mention a few quick things on kernel updates 18:29:58 we (re)discovered last week that bodhi karma requirements are per package, not per update/release 18:30:36 which means it's not flexible enough to set a karma requirement of, say, 5 for a rebase in f16 and 3 for a security fix in f17 18:31:01 it makes it harder to automatically hold things back when it's a "big" update 18:31:39 combine that with the fact that it takes 14 days before an update can be manually pushed to stable, and things can start lagging 18:31:45 so 18:32:09 what we're going to try for a bit is setting the karma requirement to 1, but disabling the automatic promotion 18:32:49 that should allow us to push it manually before 14 days if it gets 1 or more positive karma, while keeping it fairly flexible between releases 18:33:09 means more diligence on our part is needed, but hopefully it makes things better 18:33:33 Are there particular things you like to see in the karma notes? 18:33:49 I know there are things we don't like to see ;) 18:34:12 yeah. positive karma isn't necessarily something that requires a lot of info 18:34:14 we for eg get people giving negative karma because their pet bug didn't get fixed, even though we made no mention of it. 18:34:41 right 18:34:49 unfortunatly we're not going to fix all 1000 bugs in one update. 18:35:01 if someone wants to list all of the testing they did in a positive karma note, that's OK too. but it's not required 18:35:26 if it's negative karma, we'd ideally want it pointing to a filed bug 18:36:05 Usually I just note arches and that there wasn't anything obviously wrong after an hour or more of normal use. 18:36:22 yeah, that's perfectly fine 18:37:12 any more questions/comments on updates and what we're planning to try? 18:37:55 #topic Open Floor 18:39:42 i'll give it a couple of more minutes in case we have shy people 18:40:13 Nothing more from me, other than just a "good job, and keep up the good work" 18:40:20 thank you 18:41:01 ok, let's wrap it up 18:41:06 thanks everyone 18:41:09 #endmeeting