• Updates on bug list
  • Open discussion (time permitting)



04:01   BenC    Welcome everyone
04:01   BenC    only person we're missing is Tim, and he should be here any second
04:01   kylem   moo.
04:01   BenC    kylem: Please leave cows outside the channel
04:02   BenC    I leave my in the pastures, you can do the same :)
04:02   BenC    *mine
04:02   kylem   haha.
04:02   BenC    This is going to be short, no set agenda, so we'll be doing an open format
04:02   BenC    core team business first
04:03   BenC    kylem, pkl, rtg: Any issues, concerns, comments, suggestions?
04:03   kylem   BenC, have you seen the paravirt patches on linux-kernel from ingo today to clean up the issues about vdso?
04:03   pkl_    no, making slow progress on the bugs.
04:03   zul     morning
04:03   BenC    I know I still need to do the bug triage wiki :)
04:04   BenC    probably get to that this evening when we get back
04:04   rtg     Suspend to RAM has lots of problems.
04:04   pkl_    not going out in the town then? :-)
04:04   pkl_    in -> on
04:04   BenC    kylem: No, definitely interested in them though
04:05   zul     I might have a copy of i9t
04:05   BenC    pkl_: Only if invited, and only if they pay the tab :)
04:05   BenC    rtg: Yeah, we need to discuss your proposal on kernel-team@l.u.c
04:06   zul     I say disable suspend to ram ;)
04:06   BenC    rtg: I've skimmed your comments, and I think it's probably the right thing, but I need to see about what to do in userspace to enforce it
04:06   BenC    enforce it, and allow it to be enabled easily for those who want to use it
04:06   rtg     zul: I suggested that S2R be marked as experimental.
04:07   BenC    kylem: Want to look at pulling those patches in?
04:07   zul     rtg: ah cool
04:08   rtg     The patches Ben is talking about will help me track down misbehaving drivers with regard to S2R.
04:08   BenC    kylem: Oh, we already forced vdso back on for paravirt
04:08   pkl_    rtg:  The idea is to apply them to our Feisty kernel, and hopefully get them accepted upstream?
04:08   cjwatson        whoa, what's that about disabling suspend to RAM?
04:08   zul     BenC: are we thinking about xen paravirt-ops for feisty?
04:09   pkl_    rtg: Still no response from upstream?
04:09   rtg     pkl: Eventually.
04:09   BenC    zul: Has someone cleaned it up so it doesn't look like someone took a huge dump on the core kernel?
04:09   rtg     pkl: I've had some response, but have yet to propose the patches.
04:09   zul     BenC: I believe so, I havent checked recently. I can look at it this week
04:10   BenC    cjwatson: It's a huge matter of it not working for a lot of people, but this is only a suggestion at this point
04:10   cjwatson        right, but for those people for whom it does work, disabling it isn't so good ...
04:10   cjwatson        (and there are a lot of those too)
04:10   rtg     cjwatson: I'm not proposing that we disable S2R, just mark it as experimental. Its a huge regression  from Edgy.
04:11   BenC    cjwatson: Right, which is why 99% probability it wont even be considered :)
04:12   BenC    cjwatson: We're trying to get things fixed, but we're also looking at alternatives
04:12   cjwatson        given our historical user base, I think it's definitely worthwhile putting resources into fixing it where possible
04:12   rtg     cjwatson: Its all I've been working on for 3 weeks.
04:12   pkl_    Any idea when it broke?
04:12   BenC    cjwatson: mainly there are drivers breaking things, and finding them is hard
04:13   BenC    it's not S2R that is broken, it's drivers and kernel bits that are keeping things from suspending well
04:13   rtg     pkl: The breakages are due to many reasons, not just a single driver.
04:13   BenC    it locks up before the suspend is complete
04:13   pkl_    rtg: but it apparently used to work much better?  Do you know what kernel changes caused the regression?
04:14   rtg     Actually, most platforms suspend OK, it is the resume that has problems.
04:14   cjwatson        can we maybe maintain a blacklist of drivers that are known to break resume and avoid offering suspend if they're there?
04:15   cjwatson        (we only offer it conditionally as it is)
04:15   BenC    cjwatson: That's the problem...finding out which driver is causing it
04:15   cjwatson        ok
04:15   BenC    rtg's laptop produces at least one form of the problem, so he and I will work on bisecting it this week
04:16   rtg     My Dell XPS exhibits problems in core console code. thats a hard one to blacklist.
04:16   BenC    likely we'll start to uncover some similar issues elsewhere after finding one
04:16   rtg     lib_pata has been causing problems as well.
04:17   zul     do we have a list of regressions from edgy?
04:17   BenC    I think we may need to look into the libata-acpi patches
04:17   BenC    zul: launchpad.net/ubuntu/+source/linux-source-2.6.20/+bugs :)
04:17   zul     BenC: heh besides that
04:18   BenC    we don't have a concise list, no
04:18   zul     I can start on it.
04:19   BenC    Ok, bugs, bugs, bugs
04:19   BenC    let's work on getting the list of Unconfirmed bugs way down
04:20   BenC    as soon as I get the wiki page done, I'll post to kernel-team@ and hopefully we can get the stagnant bugs taken care of
04:20   BenC    Moving on...
04:20   BenC    Any community input, questions?
04:20   zul     I dont think so
04:21   BenC    I know we have a lot of community members interested in helping, but not sure how to help
04:21   BenC    I hope to add some more info on that to the wiki
04:21   BenC    and the bug triaging wiki page will go a long way to steering you folks into the abyss we call kernel-development :)
04:22   BenC    Looks like we'll be finished quite early today
04:23   BenC    I'll remain on until the official end of the meeting (other core are invited to as well) in case anyone pops in for a question
04:23   BenC    Thanks everyone, and have a good week
04:24   pkl_    OK.  Hope you have a fun time where you are :

MeetingLogs/Kernel/20070305 (last edited 2008-08-06 16:31:58 by localhost)