[07:21]  * apw waves
[07:25]  * abogani waves back
[07:36]  * apw waves to smb
[07:36]  * smb waves back into general directions
[08:55] <apw> ppisati, try sudo apt-get install ubuntu-desktop^
[09:20]  * smb wonders whether he is the only one seeing badly corrupted graphics inside a kvm with Oneiric
[10:33]  * apw has no kvm installs of oneiric
[10:43]  * ppisati -> rush out to get some food
[11:20] <ehw> hi, guys, i'm looking to rebuild the natty backport kernel for lucid with a couple of patches; is it radically different than https://wiki.ubuntu.com/Kernel/BuildYourOwnKernel ?
[11:24] <apw> ehw not radically different, the backport kenel is in the lucid repo on the appropriate lts-backport branch
[11:30] <ehw> apw: ah, ok; so if i git clone the lucid tree, it's one of the branches?
[11:30]  * ehw <- gitn00b
[11:44] <apw> ewh yep
[12:29] <herton> apw: can you verify bug 805494? (well, it's a build issue and we could mark as ok, I think just check what code is currently in proposed, don't know...)
[12:29] <ubot2> Launchpad bug 805494 in linux "ubuntu/rtl8192se driver breaks build when running 3.0 and above kernels" [Low,Fix committed] https://launchpad.net/bugs/805494
[12:30] <tgardner> herton, I think that one is obvious IIRC
[12:31] <herton> tgardner: indeed, I think I can mark he verification-done then straight away
[12:31] <tgardner> herton, tag changed
[12:31] <herton> ok thanks
[12:32] <herton> smb, tgardner: last one needing verification, trivial also, is bug 809838
[12:32] <ubot2> Launchpad bug 809838 in linux "Re-enable RODATA for virtual i386" [Medium,Fix committed] https://launchpad.net/bugs/809838
[12:33] <smb> herton, That one actually should be ovious when the qa-regression suite does not fail anymore for virtual...
[12:33] <herton> smb: yeah, can you ack that? (change the tag)
[12:33] <smb> herton, yep
[12:54] <lamont> any chance we could get 690370 fixed for oneiric?
[12:58] <apw> bug #690370
[12:58] <ubot2> Launchpad bug 690370 in linux-ti-omap4 "Strange out of memory on pandaboard" [Undecided,Confirmed] https://launchpad.net/bugs/690370
[12:58] <apw> ppisati, ^^
[12:58] <tgardner> lamont, you should hassle ppisati
[12:59]  * lamont would just like a time horizon for when he can drop the ugly hack he has for $arch=armel
[13:00] <lamont> since ISTR it turned out to be "usb on arm", not "panda" :(
[13:00] <apw> lamont, whats the ugly hack
[13:03] <lamont> vm.min_free_kbytes = 16384
[13:03] <lamont> which also doesn't quite always provide enough to survive
[13:04] <lamont> most of the arm builders are running with just 8192, but the image builders turned out to need, well ,more
[13:07] <apw> that potential patch is utterly vile, a nice cacheline bouncer
[13:38] <ppisati> lamont: last i heard, our builders were ok
[13:38] <ppisati> lamont: but i could be wrong
[13:39] <lamont> ppisati: because we hack around the problem
[13:39] <ppisati> lamont: ah, i wasn't aware
[13:39] <lamont> given that rebooting most of them involves someone physically in the data center, I'm not really inclined to test until someone tells me they've fixed the bug
[13:40] <apw> lamont, i thought we now had new ones
[13:40] <lamont> apw: new is relative
[13:40] <apw> lamont, well ones we can remotly cycle
[13:40] <lamont> *aceae are remotely powerstabbable, the others are not
[13:40] <lamont> panda boards are a current ticket being worked on this week, might even get done by then, though I'
[13:40] <ppisati> lamont: does it mean that we are running the builders with that hack, otherwise we experience the problem?
[13:41] <lamont> m getting less confident in friday as the ready date
[13:41] <ppisati> lamont: did you try ming lei patch?
[13:42] <lamont> ppisati: I'm saying that when they started panicking, I was pointed at that bug as the likely cause, and setting min_free_bytes to 8192 fixed things, so we went with that as the cause and have been happy since, other than a brief period when we lit up some beaglexm image builders and discoverd that 8192 doesn't seem to be enough for them
[13:43] <lamont> ppisati: only noticed the patch today.  if someone wanted to roll me a kernel to try on one of the builders, I could drop the hack there and give it a new kernel and we could see how long it takes to fall over, though at this point, I'm not aware of what build would be a good testcase
[13:46] <lamont> afk for a few
[14:11] <ppisati> lamont: do the builders use NFS for building? or do they have local usb disks?
[14:23]  * ogasawara back in 20
[14:25] <lamont> ppisati: the image builders use nfs, the non-image builders use local disk, the pandas-to-be will be devmapper with local disk over a read-only nfs
[14:25] <lamont> actually, the pandas-to-be may be localdisk only, haven't decided completely yet
[14:26] <tgardner> herton, do you remember what bug it was that you worked with Lamont on the 'af_unix: Only allow recv on connected seqpacket sockets.' regression ? Its in the proposed 2.6.35 stable update, so I thought I should point out to the list the possibility for regression.
[14:26] <ppisati> lamont: uhm, and i guess image builders show this problem while the non-image builders don't have it (or don't show it so often)
[14:27] <lamont> ppisati: it was much more repeatable (at 8192) on the image builders, than it was on the non-image builders with no hack
[14:28] <ppisati> lamont: i see
[14:28] <herton> tgardner: let me check here, looking
[14:28] <lamont> ppisati: from that, you can infer the timeline of the hack
[14:29] <ppisati> lamont: i see, ok, if i roll a new kernel with tha patch, can you test it (without the hack)?
[14:30] <herton> tgardner: bug 791512
[14:30] <ubot2> Launchpad bug 791512 in linux-meta "tcp connections hang in forwarding machine" [Undecided,Invalid] https://launchpad.net/bugs/791512
[14:30] <ppisati> lamont: on one of the image builders i mean
[14:30] <tgardner> herton, thanks
[14:30] <lamont> ppisati: I believe so
[14:31] <ppisati> lamont: actually the patch will just limit the numer of simultaneous skb
[14:31] <ppisati> lamont: so it won't fix it
[14:31] <lamont> heh
[14:31] <ppisati> lamont: when you it the problem, it will start to drop pkts
[14:31] <lamont> dropping packets is better than dropping the  machine
[14:31] <ppisati> lamont: and i guess with NFS that would be trigger another problem
[14:31] <ppisati> lamont: but does it panic?
[14:32] <lamont> oh yeah.
[14:32] <lamont> "trying to allocate memory on interrupt stack with no free memory" or such
[14:32] <lamont> I'll take kernel panic for $200, bob.
[14:33] <ppisati> i see, ok so, let's try with that patch
[14:33]  * ppisati goes to compile a new natty kernel
[14:33] <ppisati> lamont: natty right?
[14:33] <lamont> hahaha
[14:33] <lamont> maverick on those machines
[14:33] <ppisati> ah ok :)
[14:33] <lamont> the pandafarm will be natty
[14:34] <lamont> and if #distro is willing to sign off on building lucid bits on a natty kernel (y'all factor into that answer...), then the greater-we will be happier and maybe we can get all the arm builders up to natty
[14:35] <lamont> but we'll probably want to hold off on oneiric until the next LTS, and then flash forward to the LTS
[14:35]  * lamont heads to town, afk for a bit
[14:37]  * ppisati wonders if there's a way (zone tunable, sysctl, /sys/*, etcetc) to tweak/limit/throttle skb allocations by size/requester
[14:40] <apw> ppisati, it doesn't look like it, but worth trying to find out from more expert people
[14:41] <ppisati> apw: i'm pretty sure there's some kind of mechanism
[14:41] <ppisati> uhm...
[14:41] <apw> ppisati, i was saying at the level the code is being modded, above it there may be, i'd take my input as "don't know, you should find out"
[14:41] <ppisati> yep
[15:20]  * bjf -> dentist
[15:28] <davmor2> ogasawara: Hey give me five to grab the netbook out the other room and I'll fire it up for you
[15:28] <ogasawara> davmor2: awesome, thanks
[15:36] <ppisati> lamont: http://people.canonical.com/~ppisati/lp690370/linux-image-2.6.35-903-omap4_2.6.35-903.23~throttleusbnetskb_armel.deb
[15:37] <davmor2> ogasawara: sorry for the delay needed to install some missing packages to get the the desktop
[15:37] <ppisati> lamont: please test it on a builder (that shows that bug) and don't forget to remove the min_free hack
[15:37] <davmor2> ogasawara: kernel .5 generic has working wifi
[15:38] <ogasawara> davmor2: ack, k gimme a few minutes to fire up a kernel build to start bisecting the set of patches
[15:41] <davmor2> ogasawara: just trying .6 does too
[15:41] <ogasawara> davmor2: eh? I thought .6 was broken for you
[15:42] <davmor2> .7 doesn't
[15:42] <davmor2> ogasawara: just gone throught them all
[15:43] <ogasawara> davmor2: ah, thanks for clarifying
[15:43]  * ogasawara restarts to bisect between the right versions
[15:43] <davmor2> ogasawara: sorry for the confusion :)
[15:46] <ogasawara> davmor2: so could I get you to now test http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.0-rc7-oneiric/
[15:46] <ogasawara> davmor2: it's the upstream v3.0-rc6 kernel which the Ubuntu 3.0.0-6.7 kernel was based on
[15:46] <ogasawara> davmor2: just want to confirm that works
[15:47] <ogasawara> bah, s/v3.0-rc6/v3.0-rc7/
[15:48] <ogasawara> davmor2: ^^
[15:48] <davmor2> give me a second
[16:03]  * herton ->lunch
[16:09] <davmor2> ogasawara: last package installing now man this is now a fast machine :(
[16:09] <ogasawara> davmor2: heh, no worries
[16:14] <davmor2> ogasawara: right 3.0.0-300rc7-generic has no wifi running
[16:14] <davmor2> 0300 even
[16:15] <ogasawara> davmor2: hrm, really?  well that's not what I was expecting
[16:16] <ogasawara> davmor2: and you installed the header packages too?
[16:16] <davmor2> ogasawara: all 3 .deb packages
[16:18] <jmburgess> Hey kernel team.  I was wondering of bug 814323 should he marked as triaged. Its a known problem with the virtualbox module and I've forwarded it on upstream
[16:18] <ubot2> Launchpad bug 814323 in linux "Virtualbox 4.1 (vboxdrv) Blocks Machine Suspend" [Undecided,Confirmed] https://launchpad.net/bugs/814323
[16:19] <ogasawara> davmor2: hrm, k lets see if I can find a different way to bisect then.  can you try http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.0-oneiric/
[16:36] <davmor2> ogasawara: I'm just double checking should I be setting these up from a kernel with the wifi working? I'm assuming is shouldn't matter but as I've found out before assumption is the best policy
[16:36] <davmor2> isn't
[16:36] <ogasawara> davmor2: it shouldn't matter
[16:37] <davmor2> ogasawara: that's what I thought, right this is the last one up and running now,   no wifi
[16:37] <apw> jmburgess, possibly so yes
[16:38] <ogasawara> davmor2: and you said 3.0.0-6.7 was working right
[16:39] <davmor2> ogasawara: let me grab you the uname of the one that was working
[16:39] <ogasawara> davmor2: cool
[16:48] <davmor2> ogasawara: http://paste.ubuntu.com/653236/
[16:49] <davmor2> ogasawara: if you want something other than uname -a let me know :)
[16:49] <ogasawara> davmor2: nope that confirms 3.0.0-6.7
[16:50] <ogasawara> davmor2: "3.0.0-6-generic #7"
[16:50] <davmor2> ogasawara: that's fine then :)
[16:52] <ogasawara> davmor2: K, give me a about 30min.  I'm going to rebase the patches we were carrying for -6.7 onto v3.0 and build you a kernel to retest
[16:53] <davmor2> ogasawara: I'm off at 18:30 and it's 17:52, I'll happy try it tomorrow for you if you drop the link/package on the bug
[16:54] <ogasawara> davmor2: sounds good, I'll post to the bug.  thanks for the testing.
[16:54] <davmor2> ogasawara: no probs
[16:55] <davmor2> ogasawara: while we are at it is there any other relevant info you want for 815064 or does that cover everything you guys need?
[16:56] <ogasawara> bug 815064
[16:56] <ubot2> Launchpad bug 815064 in linux "No driver for the Network controller: Ralink corp. Device 5390" [Undecided,Confirmed] https://launchpad.net/bugs/815064
[16:57] <ogasawara> davmor2: at a quick glance, it looks like you posted everything we'd need.  I'll ping you otherwise.
[16:57] <davmor2> ogasawara: cool
[16:57] <davmor2> ogasawara: I tried to add everything I could think of :)
[17:04] <jmburgess> apw: does importance medium sound correct? If it does ill ask someone on the bug control team to mark it as triaged 
[17:04] <apw> jmburgess, i can mark it so
[17:04] <jmburgess> Great thanks
[17:05] <apw> jmburgess, done
[17:34] <jmburgess> I think bug 816110 is ready to be marked as triaged.  He's running oneric on kernel 3.0 and has all the apport info. is there anything else I should ask of him, or is it good to be marked triaged?
[17:34] <ubot2> Launchpad bug 816110 in linux "Don't work multitouch on acer 522" [Undecided,Confirmed] https://launchpad.net/bugs/816110
[17:37] <apw> jmburgess, i suspect that the answer to that one is, you have an ALPS, sorry
[17:39] <jmburgess> ALPS?
[17:42] <apw> jmburgess, make of touchpad, wherein the the manufacturer thinks their wire protocol is super secret
[17:44] <jmburgess> apw, sorry I'm new. How did you know its an alps, and does that mean this is a duplicate of bug 550625?
[17:44] <ubot2> Launchpad bug 550625 in linux "Alps touchpad is recognized but synaptics clients and scrolling do not work" [Medium,Confirmed] https://launchpad.net/bugs/550625
[17:45] <apw> jmburgess, i am not 100% sure as the device is matched as a ps/2 mouse, but it has an alps glide point so i am suspicious
[17:48] <jmburgess> ape: gotcha...i guess ill just mark it as a duplicate then
[17:49] <jmburgess> Sorry that was to apw, not ape
[17:50] <apw> jmburgess, hard to know if its is really a dup or not
[17:51] <apw> the other one implies it is recognised as a touchpad, which does not fit the symptoms here
[17:52] <jmburgess> Oh good point.
[17:53] <jmburgess> Let's see if I can find if this is a dup
[17:59] <jmburgess> Oooooo looks to be related/duplicate of bug 565543
[17:59] <ubot2> Launchpad bug 565543 in linux "Alps touchpad detected as ImPS/2 Generic Wheel Mouse(in VAIO E series) after the kernel upgrade" [Medium,Incomplete] https://launchpad.net/bugs/565543
[18:11] <jmburgess> Is the latest dev release running kernel 3.0.0.7-generic considered upstream mainline?
[18:14] <apw> jmburgess, thats an ubuntu kernel so not mainline, it has our delta on it
[19:21] <BenC> amitk, apw, cking, kees, pgraner, ogasawara, tgardner: Any of you going to linuxcon?
[19:21] <tgardner> BenC, none of the kernel team. kirkland will be there.
[19:22] <tgardner> plumbers for you ?
[19:22] <BenC> when is plumbers?
[19:22] <tgardner> sep7-9, santa rosa
[19:22] <BenC> I'll see, but no idea
[19:22] <kirkland> BenC: howdy, long time no see :-)
[19:23] <BenC> kirkland: hey, I will likely see you at linuxcon :)
[19:23] <BenC> First beer is on me...jaeger bombs are on you...
[19:23] <kirkland> BenC: nice, will be good to catchup
[19:23] <kirkland> X-o
[19:23] <BenC> hehe
[19:35] <apw> bug #817187
[19:35] <ubot2> Launchpad bug 817187 in linux "libvirt FTBFS on 2.6.39 and later kernels" [Undecided,Invalid] https://launchpad.net/bugs/817187
[20:40]  * tgardner -> EOD
[20:47] <amitk> BenC: going to plumbers, not linuxcon - jaegers are on you if you attend ;)
[21:27] <kees> BenC: I'll be in santa rosa for 1 day (for the linux security summit day)
[22:24] <jjohansen> rebooting