=== achiang` is now known as achiang === robbiew1 is now known as robbiew === panda is now known as Guest23706 [07:41] smb, morning ... this amd quirk, can you remind me of the bug for that one === GrueMaster_ is now known as GrueMaster [07:43] apw, hello, could a patch like this be considered for inclusion? https://patchwork.kernel.org/patch/244201/ [07:45] ricotz, is the patch upstream yet? and in part the answer depends on where you are looking to put it [07:47] apw, this patch isnt upstream, and the current natty kernel (38rc6) breaks the usage of my dvb card, including this patch solved it for me [07:49] the patch doesn't claim for fix anything, just speed things up ... so thats a little unexpected [07:49] * apw drop for a few mins [07:51] apw, yes, the current driver gives me timeouts when switching channels/transponders mostly everytime, so this "speed-up" patch solved it for me [09:28] * apw sighs ... [09:29] wassup? [09:29] more plumbing probs? [09:29] * cking has to reboot [09:29] heh ... no this is the 'fix' for the problems [09:30] ah [09:30] but the plumber was late, so i was waiting around like a lost puppy [09:30] apw, Though we thought they *had* been fixed already [09:30] that are a law to themselves [09:30] no _i_ bodged it till it could be fixed [09:31] Hm, old plumbers law: first it bodges then it comes off === ogra_ is now known as ogra [09:42] smb, something like that :) [10:36] Hello all! [10:37] does the regression found in the -proposed kernel for ec2 invalidates this kernel in -proposed? [10:37] https://bugs.launchpad.net/bugs/723335 [10:37] Launchpad bug 723335 in linux "linux: 2.6.35-27.48 -proposed tracker" [Medium,Triaged] [10:37] apw, ^ ? [10:38] ara, Hm, is that another one than bug 725089? [10:38] Launchpad bug 725089 in linux "QRT test-kernel-security causes kernel panic on EC2 AMD64 image" [High,New] https://launchpad.net/bugs/725089 [10:38] smb, no, that one [10:38] smb, the one above is the tracker, the bug is the one you mention [10:38] ara, I will try to get to the bottom of that today [10:39] smb, cool, if the regression is confirmed, then the linux (not linux-ec2) package will also be dropped? [10:39] I looked at it on Friday and was not sure what could have caused it. Actually it seemed that the kernel before that may already have been causing issues [10:41] ara, Unfortunately for Maverick there is no ec2 package [10:41] smb, so would that be -virtual which has failed then ? [10:41] We use the virtual image generated from the normal linux package [10:41] In some way yes. [10:42] But I am not sure one can fail virtual without failing the main package. [11:07] smb, indeed if -virtual is toast they all are === diwic is now known as diwic_afk [12:31] I'm trying to examine some regression with alsa or kernel, what is fastest way of finding it, shoud I install different kernels from packages or compiling from source? I read it can be done with git bisect but could I use git with ubuntu kernels compiled like this https://help.ubuntu.com/community/Kernel/Compile? Or should I use this method https://wiki.ubuntu.com/KernelTeam/GitKernelBuild [12:41] nusch, i would start by picking the pre-build kerenls from launchpad [12:42] to narrow it as much as possible, then probabally try the mainline builds for finer bisection in between [12:42] nusch, which release as this appeared and when? [12:47] dma_addr_t dma_map_single (struct device * dev, void * cpu_addr, size_t size, enum dma_data_direction dir); .. is cpu_addr a physical addr? [12:49] fairuz, looks like a kernel virtual addy to me [12:51] apw: I'm talking about this bug https://bugs.launchpad.net/ubuntu/+source/alsa-driver/+bug/662009 and somebody already found this was broken between 2.6.33 and 2.6.35 [12:52] Launchpad bug 662009 in alsa-driver "[Lenovo Y530 - Realtek ALC888] Regression from 10.04: No sound with snd_hda_intel model=lenovo-sky" [Undecided,Triaged] [12:52] so I wanted to examine it deeper [12:52] nusch, so i would seee if .33 and .35 mainline builds show the same pattern [12:52] and then use the -rc builds in the mainline archive to find a finer location [12:52] quicker than building ones own [12:53] Could you provide m with some details how to do that? [12:53] https://wiki.ubuntu.com/Kernel/MainlineBuilds [12:53] that has the docs on where to find them and how to instal them etc [12:56] ok and should I install also new alsa package like this from ppa linux-headers-alsa-driver-2.6.35-25-generic with kernel version included or get rid of it and use standard ubuntu ones ? [12:57] i'd not expect the headers to make any difference to a sound issue [12:57] sory I meant linux-alsa-driver-modules-2.6.35-25-generic [12:58] those are from ppa:ubuntu-audio-dev/ppa [12:58] well that only applies to 35-25, so as long as you arn't booting that you won't be using it [13:00] I'm saying about whole set of packages from ppa:ubuntu-audio-dev/ppa which has version number included in name, so It will need installing new alsa with every kernel, or should I uninstall it and allow system use standard packages [13:00] nusch, when you're git bisecting, make sure this package from ubuntu-audio-dev is NOT installed. [13:01] if you are looking for a general bug in the kernel, you don't want any unecesary packages like that [13:01] they just add variables you don't need [13:01] But I'm not sure the problem exists in kernel or alsa modules [13:02] well from a kernel point of view, it includes alsa, and its that one we care if it works or not [13:03] ok thx, I'll inform you after testing [13:05] apw: any thoughts on the baseCommitRegex stuff? [13:07] ? [13:07] jk-, where? [13:07] kernel-team [13:08] apw, In those rfc mails you tend to ignore ... ;-P [13:08] nthing with that in the title [13:08] [RFC, PATCH 0/2] UBUNTU: lucid: allow printchanges to accept varying release commit messages [13:08] smb :) [13:09] jk-, Not that I did much more response. I think the only thing I had was that if we start to change it should be Natty and then potentially go back, but that was already said [13:09] well i don't see any inherant issue with them, though what form are you using thats not being picked up [13:10] smb: yeah, that's fine; I was starting with lucid because that's the kernel I was most recently working with === diwic_afk is now known as diwic === shadeslayer is now known as kshadeslayer [13:11] apw: "UBUNTU: (Release) 2.6.35-6.11 for " [13:12] so if you used "UBUNTU: release 2.6.35-6.11" it would work as it is [13:12] so if you used "UBUNTU: release Ubuntu-2.6.35-6.11" it would work as it is [13:13] I'll check with Ike about why he's used the different format [13:16] jk-, but overall nothing bad in there. i've replied as such and mentioned the flexibility we do have with the current pattern [13:16] apw: ok, cheers. I'll sort out our commit conventions and get back to you [13:20] jk-, it doesn't seem like a bad change overall ... the branch specific file could be useful without your use [13:20] might simplify some of our existing stuff === zul_ is now known as zul === Quintasan_ is now known as Quintasan === _LibertyZero is now known as LibertyZero === yofel_ is now known as yofel === sconklin-gone is now known as sconklin [15:48] <_ruben> any clues on how i could persue this issue: http://pastebin.ubuntu.com/573509/ .. apparently using "local my.ip.ad.dr" when creating a sit tunnel "breaks" the link local (/128 instead of /64) === herton is now known as herton_lunch [16:29] sconklin, Hm, I found one line on irc from Thursday where jj mentions regressions for Lucid and Maverick. But not which those are. I cannot remember anything about Lucid (beside some load accounting problems maybe which seem to come and go). But we should ask him when he comes online. [16:29] ok, thanks! [16:30] by the way all - high winds and bad wx here, so if I drop offline that's why [16:50] bjf, you have likely noticed by now, but I got your nominations done on bug 723945 [16:50] Launchpad bug 723945 in linux-ti-omap4 "CVE-2010-4258" [Undecided,New] https://launchpad.net/bugs/723945 [16:50] JFo, heh, thanks! [16:50] bjf, my pleasure === herton_lunch is now known as herton [17:03] jjohansen, Hey sconklin refreshed my memory today about you mentioning regressions on ec2 in Lucid and Maverick. But I cannot remember what the regression in Lucid was. Do you still remember? [17:05] smb: yeah, just a sec I am looking up the bug # [17:11] smb: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/725310 [17:11] Launchpad bug 725310 in linux "Lucid 2.6.32-29.58 /dev/mem is readable (dup-of: 725308)" [Undecided,New] [17:11] Launchpad bug 725308 in linux "Lucid 2.6.32-29.58 /dev/mem is readable" [Undecided,New] [17:11] and I have really started looking at it yet [17:11] not? :) [17:12] smb: yes NOT [17:12] :) [17:12] smb: I did start looking into https://bugs.launchpad.net/ubuntu/+source/linux/+bug/725089 [17:12] Launchpad bug 725089 in linux "QRT test-kernel-security causes kernel panic on EC2 AMD64 image" [High,New] [17:13] which is the maverick bug [17:13] jjohansen, I did sort of do that too [17:13] smb: ah okay :) [17:13] jjohansen, Seems I was able to boot the latest proposed kernel there [17:13] Though I think I discovered another regression (but that is old) [17:14] smb: oh which regression is that [17:15] oh nm its in your bug comment [17:15] jjohansen, I noted in the bug report. But it seems running "reboot" in the instance does not work [17:15] yeah [17:15] Or I am not patient enough for it [17:16] But it seems to be consistent with what I see on my local test system [17:16] For a Lucid image I can do a "xm shutdown" and the domU goes away [17:17] For a Maverick image I do the same and it stays around until I do a "xm destroy" [17:17] * smb thinks the reboot / halt vectors may probably need better handling [17:18] hrmm, yeah [17:18] But as said, this is something that probably came with the change from Lucid to Maverick [17:18] apw: any reports of wireless just dying with the newest kernels [17:18] jjohansen, no that have reached me no [17:19] apw: hrmm, I have had my wireless just die a couple times since upgrading friday [17:19] what h/w you got ? [17:20] intel 4965 [17:21] hrm pretty common [17:21] i wonder if you got some upgraded firmware recently [17:22] hrmm, unlikely I haven't upgraded the firmware package [17:22] smb, jjohansen: so is it correct that we should not hold the maverick kernel for ec2, and that we should Lucid until further notice from jj? [17:23] apw: it could have an anomally, if it happens again I will file a bug [17:23] sconklin: uh no give a little bit more with maverick, I want to double check [17:23] apw, one more bug I noted to be marked as fixed by latest natty rebase: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/722925 [17:23] Launchpad bug 722925 in linux "bug on fs/inode.c:1422" [High,Triaged] [17:23] sconklin, jjohansen The Lucid thing could just be again something that never was working in ec2 (due to the duplicate files) [17:23] jjohansen, icirtianly worth watching, not someting i have heard from anyone else yet [17:23] apw, same reporter confirmed the fix at upstream bug https://bugzilla.kernel.org/show_bug.cgi?id=29792 [17:23] bugzilla.kernel.org bug 29792 in Block Layer "BUG: fs/inode.c:1422 when deleting device-mapper devices" [Normal,Closed: duplicate] [17:24] smb: all of qrt was definitely passing on lucid and maverick, the question is when kees added the test that is failing last year [17:25] smb: I couldn't remember, so the failure could have always been there [17:25] I was going to boot an older instance to test if it was there [17:25] to determine whether it was a regression or not [17:25] smb, jjohansen: could you please send me an email with as detailed a status as you have - I need to keep everyone informed on this, as we're slipping until this is resolved. I've forwarded Stefan's earlier status to you so you can build on that [17:25] jjohansen, Yeah, sounds good. Let me check something quick [17:26] herton, thanks for the pointer ... helpfiul [17:27] herton, marked in the changelog [17:27] cool [17:27] herton, bug -> Fix Committed [17:28] * apw notes herton is reading a lot of bugs :) [17:28] great, thanks [17:28] yeah, trying to follow them :) === diwic is now known as diwic_afk [17:40] jjohansen: let me know if I can help with the qrt thingy [17:40] kees: will do [17:42] jjohansen: fwiw, the SECCOMP test was added roughly in May 2009 [17:43] Question would be whether we ever saw it ok with xen... Or just went meh... [17:44] ... seems to check devmem_is_allowed and that is defined in init.c... for which xen has a copy... [17:47] which bug is thiat? devmem not working or seccomp killing it? [17:48] kees, jjohansen So there is a difference here. The x86 code says "if (!page_is_ram(pagenr)) -> ok" while xen has something like if the page is beyond the given pages, then it is ok" [17:49] if (mfn_to_local_pfn(pagenr) >= max_pfn) [17:49] smb: hrmmm, interesting [17:53] smb: the devmem checks we never validated on Xen, that's for sure. and those were added in Oct 2010 [17:55] kees, smb: yeah I think we haven't run qrt since devmem checks were added [17:55] err that is on lucid [17:56] * jjohansen isn't sure because he does remember running qrt last fall for some other bugs around that time frame [17:56] jjohansen: yeah, I don't think the devmem test was run on Xen. again, it's likely my test sucks. :) [17:57] jjohansen: I'm trying to work up a more definitive test. right now it's a bit "just try a whole range of addresses" [17:57] Well at least it may not take into account xen strangeness or xen needs some tweaking [17:57] kees: is brute forcing the address space again :) [17:57] But either way. I also thought to have seen something fail on Lucid before [17:58] jjohansen: hehe [17:58] smb: really, my recollections of xen qrt are that it was clean [17:58] that is one of the things we made sure of before the release [17:59] but there have been a few tests added since then, and those recollection are a lot more fuzzy [18:00] jjohansen, The only thing one can be sure of is that one cannot be sure... [18:01] smb: yep [18:01] Probably we quickly can check that with a current ec2 instance for Lucid. The latest kernel should not be in there. If the test fails too, then its at least not a new regression [18:01] And the SECCOMP thing seems to be something that not always happens too. But feel free to get it a second try [18:07] smoser, hey ... you have a bug open which is getting dogpiled, and as reporter i'd like your feedback: bug #625364 [18:07] Launchpad bug 625364 in pm-utils "lenovo/thinkpad R500/T6x/T400[s]/T500/W500/W700/X60/X200 suspend fails" [Undecided,Confirmed] https://launchpad.net/bugs/625364 [18:07] apw, i think there are 2 bugs there in the dogpile [18:07] one is the tpm bug [18:07] one is, i think intel video related [18:08] smoser, tend to agree. if you are working though i want to close out the tpm bit you were suffereing [18:08] so i can get them to make new ones [18:08] smb: that is what I am doing, I fired up some fresh instances and am giving it another test [18:08] my T400 was suspending and resuming fine in natty. [18:08] smb: oh, also note that I disabled the /dev/mem check on ec2 for the time being [18:08] i was sufffering i think from the intel video bit. === sforshee is now known as sforshee-lunch [18:08] i haven't suspended and resumed a lot in several upgrades on natty [18:09] but it *was* working when i originally jumped to natty [18:09] yeah we've had a number of bugs come and go [18:09] cking, did you find that patch ? [18:10] apw, the S4 one? [18:12] cking, yeah === evilshadeslayer is now known as shadeslayer [18:56] I asked afternoon about searching for regresions in kernel, I'm working with bug https://bugs.launchpad.net/ubuntu/+source/alsa-driver/+bug/662009 [18:56] I tested it on my Y530, and last working kernel-ppa version is 2.6.34.7-maverick and in 2.6.35-rc2-maverick which is next it's already broken - I hear music very silent when I connect external speakers and set them to full volume. What next, should I download git version of those or test vanilia kernel? [18:56] Launchpad bug 662009 in alsa-driver "[Lenovo Y530 - Realtek ALC888] Regression from 10.04: No sound with snd_hda_intel model=lenovo-sky" [Undecided,Triaged] === sforshee-lunch is now known as sforshee === ogra is now known as Guest25882 === Guest25882 is now known as ogra_ === ogra_ is now known as ogra [20:46] bjf, sconklin: neither are regressions release the kernels [20:46] jjohansen: ok, thanks. Are they bugs which are not regressions? [20:47] sconklin: correct I sent an email with a little more detail [20:47] very little more [20:47] jjohansen: thanks much! We'll crank the machine back up ;) [20:48] sorry it took so long the one really did look like a regression, passed on old first time I tried failed, on new started bisecting and then it just got weird [21:24] JFo, if you could bless bug #726796, i'd be so grateful [21:24] Launchpad bug 726796 in linux "linux: 2.6.35-28.49 -proposed tracker" [Medium,In progress] https://launchpad.net/bugs/726796 === bo is now known as Guest26312 [21:26] bjf, approved [21:26] apw, thanks === calc is now known as Guest54789 === soren_ is now known as soren === sconklin is now known as sconklin-gone