=== zopa is now known as azop === kamalmostafa is now known as kamal [01:48] Hi I have a bug to report, but needed some info on how to. I have encountered a Kernel regression\ [01:50] I am only able to run 10.04 and previous versions without the acpi=off option. once I upgraded to 10.10, 11.04, 11.10 I had to use the acpi=off option. I will provide whatever info is needed, but I currently only run Ubuntu 10.04 on this machine (Compaq Presario R3000). Thanks [02:21] can anyone see my messages? === smb` is now known as smb [08:33] * smb drag himself and a cup of coffee in [08:39] heh a vision of you dragging in a coffee is not easy [08:41] ohsix, i thought our perf had a fix in it for that to find the symbols, or to take an option so you can specify them [08:41] morning * [08:41] ppisati, morning [08:42] apw, In a cup :) [08:42] ppisati, apw morning [08:43] smb, thats a big cup if it needs dragging [08:44] apw, Oh you know how difficult even little things are in the morning [08:53] apw: could be, id idn't look at patches, and it wasn't finding vmlinux without me copying it (also, it was using kallsyms at record time but saying it was invalid at report time) [09:09] smb, http://kernel.ubuntu.com/~kernel-ppa/info/kernel-version-map.html [10:03] do any of you have an lvm disk/partition? [10:03] ppisati, Sometimes [10:03] i just found out that [10:04] after installing tha latest kernel [10:04] the grub logic that tries to determine the partition type [10:04] confuses my lvm partition as a fuse partition [10:04] tries to mount it [10:05] ppisati, latest kernel? 3.2.0-4 ? [10:05] * smb thinks he saw some bug about it... if he could remember where [10:05] yes [10:05] hmmm [10:05] i launched this morning the update [10:05] got the new kernel [10:05] but the installation was stuck [10:05] but i noticed that a process was sucking 100% cpu [10:05] and eating all my mem [10:06] down to the point where i had to manually kill -9 some of them [10:06] fuse partition to my mind doesn't mean anything [10:06] wait [10:06] i fill an lp bug with some details [10:07] ppisati, what kernel were you running before [10:07] apw: Linux newluxor 3.0.0-13-generic #22-Ubuntu [10:07] still running this one [10:07] didn't reboot yet [10:07] oh so this is an oneiric -> precise update ? [10:07] nope [10:08] i'm still in oneiric [10:08] but was installing... [10:08] but was upgrading to precice ? [10:08] 3.0.0-14 [10:08] nope [10:08] apt-get update [10:08] oh its an oneiric bug [10:08] to me [10:08] it's a grub bug [10:09] if i execute [10:09] sudo /usr/lib/os-probes/50mounted-tests /dev/sdd1 [10:09] where sdd1 is an lvm partition [10:09] i end up having: [10:11] nothing? [10:11] http://paste.ubuntu.com/767722/ [10:11] i had to kill immediately the script [10:11] else [10:11] it would suck all my mem [10:11] and my box would start swapping [10:11] trashing like hell [10:12] and the last line of /var/log/syslog say: [10:12] Dec 12 11:09:40 newluxor 50mounted-tests: debug: mounted using GRUB [10:12] Dec 12 11:09:40 newluxor 50mounted-tests: debug: running subtest /usr/lib/os-probes/mounted/10freedos [10:12] Dec 12 11:09:40 newluxor 10freedos: debug: /dev/sdd1 is a FUSE partition [10:12] what the HECK is a FUSE partition? [10:13] apw, probably not partition [10:13] fuse is used to access ntfs afaik [10:14] btw, that partition is mounted as /dev/mapper/storage-storage (and it has ext3 on it) [10:15] it's not a partition, but they do get mounted :> [10:15] one of the endpoints is a program and it shows up in /proc/mounts [10:17] Wonder whether it could be some interaction with the userspace fs helpers .... I am still trying to find something relevant there [10:20] i'm filling the lp bug with some more details... [10:27] ppisati, btw, are we talking about a 32 or 64 bit system? [10:29] ok, here it is: [10:29] lp 903125 [10:29] Launchpad bug 903125 in grub2 "50mounted-tests incorrectly handles lvm partition" [Undecided,New] https://launchpad.net/bugs/903125 [10:29] smb: 64bit [10:42] ppisati, Hm, seems I cannot immediately reproduce this in my vm... [10:44] smb: no worries, i'll try to debug it a bit more [10:58] ppisati, You should add the output of sudo fdisk -l [10:59] ppisati, I mean you got it for the one disk, but just for all of the partitions [11:03] hm... 50mounted-tests... where does that come from? [11:04] apw, http://paste.ubuntu.com/767761/ is the dmidecode output from my machine, to see if it should be suspending or not. I *think* that the machine is failing to suspend, rather than suspending and failing to resume, because the screen never turns off (and the RTC trick puts a Magic number in dmesg but no "hash matched" lines at all) [11:06] smb: so you have an lvm partition, right? [11:06] smb: sudo grub-mount /dev/sdd1 /media/ [11:06] smb: and then ls -1 /media [11:06] smb: watch memory and cpu [11:07] ppisati, yes, for / and I even added a second disk an put a single pv on that [11:08] smb: nothing happens until i do some activity on it [11:09] smb: e.g. now i mounted it [11:09] smb: and the box is ok [11:09] smb: grub-mount on /media type fuse.grub-mount (rw,nosuid,nodev) [11:09] ppisati, It tells me unknown fs and nothing happens. No process running [11:09] uhm [11:11] cking: hi! i've done the quick ALPM test with my thinkpad t510, and just want to be sure: the longer test with i/o load should be done on battery too, right? [11:12] ppisati, Just to make sure, are the partition/contents reported for sda-sdc correct (as in the report)? [11:12] what you mean? [11:12] ah [11:12] ppisati, All that WinXP and Win2K found there [11:13] yep [11:15] lvdisplay uses only labels [11:15] is there a way to pass a /dev? [11:15] or any other tools to inspect the content of /dev/sdd? [11:15] just realised that I probably shouldn't have directed the above specifically to ap-w :) [11:15] ppisati, Ok. So just for reference. In the vm the grub version are the same as yours. lvdisplay /dev// [11:16] ppisati, but lvdisplay is for lvs not partitions directly [11:16] yep [11:16] I've got a new Lenovo U300s, which doesn't seem to suspend correctly. http://paste.ubuntu.com/767761/ is the dmidecode output, to see if it should be working or not. I *think* that the machine is failing to suspend, rather than suspending and failing to resume, because the screen never turns off (and the RTC trick from https://wiki.ubuntu.com/DebuggingKernelSuspend puts a Magic number in dmesg but no "hash matc [11:16] hed" lines at all) [11:17] crap [11:17] grub-mount is a binary [11:17] ppisati, I sourced the common scripts from 50mounted-tests and ran fs_type... or use blkid [11:18] ppisati, Actually sudo blkid should give us quite a bit of info [11:19] uuid and type [11:19] [flag@newluxor ~]$ sudo blkid /dev/sdd1 [11:19] /dev/sdd1: UUID="ew7mIr-AC0i-WDIU-CmjU-psOF-PU1z-WuQuDf" TYPE="LVM2_member" [11:20] and on sdd ? [11:20] nothing [11:20] apw, that is nothing [11:20] [flag@newluxor ~]$ sudo blkid /dev/sdd [11:20] [flag@newluxor ~]$ [11:20] to me it looks like, grub-mount is fooled [11:20] As soon as there is partitions [11:20] mount it as $something [11:20] why is the volume on a partition [11:20] and when some activity is done in the mount point, fuse starts sucking cpu/mem [11:21] ohsix, choice [11:21] weird [11:21] i think i did that when i was playing with it, along with some looped files; but to do that when you have storage ... :] [11:21] ohsix, It would not matter that much on x86, but its safe and historically you used to give it a lvm partition type, too [11:22] ohsix: this is an old disk i got while retiring/disassembling another box [11:22] right but the partition table is superfluous [11:22] ohsix: i put in my workstation since i wanted to reuse the disk [11:23] ohsix, If you use the whole disk as pv. But you could only use parts of it for example [11:23] ohsix: for what is worth, it came froma debian server [11:23] i think it was a lenny server IIRC [11:23] ppisati, so do i remember you saying you haven't booted onto the new kernel you got? and do we know what else you did get [11:23] so, long ago [11:23] right, don't let me digress i just thought it odd; it should work even if it's a partition [11:23] apw: yep, not rebooted yet [11:24] what else you did get with the kernel in the update [11:24] ohsix, Not that it matters, but if you do that on the mainframe on dasd devices you will be doomed [11:24] apw: ah [11:24] apw: i don't remember exactly, some other packages [11:24] grub2 is 10 weeks old, so unlikely that [11:24] nope [11:24] there is a log in /var/log which has the list [11:25] udev and util-linux are also simiarly old [11:25] /var/log/apt/history.log? [11:25] sounds believable :) [11:26] apw: http://paste.ubuntu.com/767779/ [11:26] ppisati, seems os-proper is its own package... [11:26] ah, so it's not grub [11:27] os-prober [11:27] i had to hack at it a bit on my netbook with the way it added osx and xp :] [11:27] ok so you only gotten the kernel and ghostscript, but as you arn't running the new kernel yet ... and ghostscript seems unlikely as a cause [11:28] i am somewhat perplexed as to how this could be new [11:28] you can disable it in /etc/default and it'll not generate the new style boot menu [11:28] no no [11:28] apw: it all started now because i installed this disk [11:28] apw: in my workstation, before this disk was working on another box [11:28] apw: a debian box, that was bot updated since... [11:29] 10:03:48 ppisati | i just found out that │ apw [11:29] months? years pehaps [11:29] 10:04:03 ppisati | after installing tha latest kernel │ aquarius [11:29] because the new kernel triggers grub-update [11:29] and i never tried grub-update after i installed this disk [11:29] thats not how you worded it at the beginning, there was no mention at all of the lvm being new [11:29] So summary, this disk may contain something os-prober does not expect and confuses it. And its nothing update related. [11:29] nope [11:30] * apw resets problem space [11:30] the update just triggered the problem [11:30] smb, sounds about right [11:32] ppisati, Maybe you could run the initial steps of 50mounted-tests and see that fstype is what you expect for start [11:33] or read it, heh; strace will tell you what it's doing, the stuff in os-prober is pretty comprehensible [12:15] * ppisati - [12:15] < reboot [12:36] * ppisati -> lunch [13:23] Has anybody reported nivida binary drivers not functioning with the latest kernel in oneiric? [13:24] Daviey, I think I've seen a few bugs regarding nvidia drivers in Oneiric [13:27] jsalisbury: Is this not something currently tested? [13:31] Not sure for sure, but even if, there is a lot of hw and a lot of subtle problems to not find... [13:34] Daviey: is this breaking _all_ nvidia users currently? [13:36] mdeslaur: NFI, i can only speak for myself. [13:39] Daviey: do you have a bug#? [13:39] cking: hi! should the ALPM long term test be done on battery? [13:40] mdeslaur: I haven't raised one as yet, and not come across ones jsalisbury might have seen, [13:40] Trying to poll my loco to see if others have noticed. [13:40] htorque, yes, since the power saving mode only works when on battery, or unless one forces it [13:40] Daviey: ok, I'll throw oneiric on a nvidia laptop, see if I see it too [13:40] htorque, you can force it on using: sudo pm-powersave true [13:40] Daviey, I recall two bugs recently. Let me check for the bug numbers. [13:41] thanks! [13:41] mdeslaur: note, nvidia _binary_ drivers [13:41] cking: thanks, will force it then. :-) [13:41] Daviey: yes, thanks [13:41] htorque, thanks for testing [13:42] yw :-) [13:42] Daviey, hmm, actually they may have been for precise. Here's one: bug 897102 [13:42] Launchpad bug 897102 in linux "precise can't boot with linux 3.2.0.1.1" [High,Confirmed] https://launchpad.net/bugs/897102 [13:42] Oh [13:43] Daviey, yeah, sorry, the other one was for precise too: bug 901386 [13:43] Launchpad bug 901386 in nvidia-graphics-drivers "Ubuntu 3.2 kernels vs. nvidia-current vs GTX 560 (Mainline works)" [High,Confirmed] https://launchpad.net/bugs/901386 [13:43] dammit, just had confirmation from someone with the same hardware as me that it is working. [13:45] Daviey, Not wanting to worry you, but the last time I had strange gfx trouble it was the beginning of my mainboard dissolving itself... :-P [13:46] smb: "thanks", thankfully - booting into the 3.*-generic-(n-1) it works :) [13:48] Daviey, Hehe. Ok, make sure that the nvidia module loads. As it is dkms, there are sometimes issues when the compile is done at the wrong time and you end up with it complaining about wrong symbols [13:50] The manual steps to recover are a bit ... inconvenient ... Meaning you have to run the dkms commands to uninstall, build and install the module (all with a lot of arguments to specify module name, version and kernel version) [13:52] smb: right, i'm on it :) [13:52] thanks [13:57] smb: STAND DOWN, http://pb.daviey.com/JXf4/ \o/ .. life is probably good. [14:11] hi there, could someone else try to help me with this bug? https://bugs.launchpad.net/ubuntu/+source/linux/+bug/898615 [14:11] Launchpad bug 898615 in linux "[Presario C500] Hibernate command fails and returns to userspace" [Medium,Confirmed] [14:30] tgardner: you gonna attend the tech board meeting today? [14:32] ogasawara, I have not decided. I've got a conflict. I don't really have much more to add to the discussion other then what I've already written. All I've received in that thread (as far as I'm concerned) is opinion and hyperbole. Nobody has any real numbers. [14:33] I've just tried to report a suspend bug with "ubuntu-bug linux", but apport claims I'm not running an Ubuntu package; I'm running precise, installed from yesterday's nightly build and upgraded today. What might I be doing wrong? [14:33] ogasawara, are you interested in going to bat for it? [14:35] tgardner: I'm willing to attend, but I'm not as compassionate about the topic as you. I do see value in us just building the flavor and having it shoved to universe and not officially supporting it. [14:36] ogasawara, I think its interesting the nobody has expressed an interest in maintaining a universe kernel. that tells me something... [14:37] plus, I think you meant compassionate --> passionate [14:37] tgardner: indeed :) [14:37] kernel dudes don't have compassion :) === yofel_ is now known as yofel [15:07] jsalisbury - hi [15:08] brendand_, hello === brendand_ is now known as brendand [15:09] jsalisbury - i added an item to the kernel team agenda. hope that's alright [15:10] brendand, sure. Is that the Hardware Certification - Testing Pools entry? [15:10] jsalisbury - yep [15:11] brendand, ok, noted [15:11] ogasawara, *sigh* 802.11n seems to be broken in -rc4 on my Thinkpad x120e. 802.11g still works ok (rtl8111/8168B) [15:11] tgardner: -rc4 or do you mean -rc5? [15:12] ogasawara, right [15:12] I really had other plans for my day. [15:16] brendand, is the topic you added a one time event? [15:16] jsalisbury - should be [15:16] brendand, cool, thanks [15:27] drat - fresh i386 precise server install panics sometime after trying to set up networking. [15:28] feh, guess i need to try under kvm. it scrolled fast for 1-2 seconds before ending up here, so i don't think screendumped panic is helpful [15:33] * ogasawara back in 20 [15:41] tgardner, apw re: bug 894768 According to Raphael in comment #33, this has been happening in prior releases. Are there plans to backport the fix to prior releases? [15:41] Launchpad bug 894768 in linux "Installation randomly fails with: File "/usr/lib/ubiquity/ubiquity/install_misc.py", line 621, in copy_file targetfh.write(buf) IOError: [Errno 22] Invalid argument " [High,Fix released] https://launchpad.net/bugs/894768 [16:00] jsalisbury, the ultimate fix was from upstream, so perhaps they ought to be driven through stable [16:01] tgardner, ok, that makes sense. [16:01] hm, kvm boots fine. hardware does not. maybe i goofed, i suppose... [16:02] jsalisbury, lemme make sure those fixes really did land in -rc5 [16:02] tgardner, great, thanks [16:05] jsalisbury, ok, 3.2-4 is still carrying apw's fix, but upstream is developing a different approach that has yet to be finalized. [16:05] tgardner, ah, ok. [16:18] jsalisbury, yep am waiting for it to hit mainline, some is slated to go to stable anyhow and depending what does we should teat that and confirm it fixes older releases [16:18] jsalisbury, the instant fix we have is not applicable to older releases as the code at fault did not exist back there [16:18] apw, tgardner, thanks for explaining [16:19] apw, note that there is someone in the bug report that says 3.2-4 doesn't fix the problem [16:20] tgardner, assuming they are not simply wrong, then there must be more than one bug, i cannot repro it at all [16:21] tgardner, but we should get them to retest with the 4 patch set as well before trying to debug, as that does fix two issues one of which is mine [16:21] apw, this onl;y happens with the installer, right ? [16:22] no anything could trigger it, its not doing anything special, just a lot of it [16:22] tgardner, will get them a test kernel with the expected fixes [16:22] apw, just wondered if we should upload a new kernel with the upstream patches in order to get some testing in the dailies. [16:23] tgardner, let me find out what upstream has done, its supposed to be in for the -rcs ... it may be in the next one already [16:23] apw, ack [16:48] * apw pops out on an errand [16:52] ogasawara, so 802.11n isn't quite as bad as I thought with 3.2-4. Its just a little racy. Given the lock patch on the wireless list I can see why. [16:53] tgardner: ack [17:05] bjf - do you remember the model of server you asked me to test last week? i figured out why it wasn't booting (i think) [17:06] brendand: it was a hp PowerEdge 2940 (i think) [17:06] bjf - probably Dell PowerEdge [17:07] bjf - i remember now. PowerEdge 2950 it was [17:08] bjf - i'll give it a go again and let you know. it was with natty right? [17:08] brendand: i'll buy that [17:08] brendand: i've got to find the bug [17:09] bjf - as far as i can remember it was supposed to not boot with the latest Natty kernel in updates, right? [17:10] brendand: better let me find it so you don't waste your time booting the wrong one [17:12] bjf - sure [17:14] Does anyone know what gen6 / ironlake mean in the i915 driver? Is gen6 just sandybridge, and ironlake a subset? [17:15] arges: ironlake iirc gen5, arrandale === vanhoof_ is now known as vanhoof [17:15] ivy bridge is GEN7 IIRC [17:15] tgardner: ack [17:16] tgardner: sandybridge is gen6 [17:16] vanhoof, ah thanks [17:16] ironlake is part of the westmere platform, which would be clarkdale/arrandale cpus [17:17] and IIRC that is gen5 [17:17] vanhoof, so would an x220 with a sandybridge process call some ironlake functions? [17:17] i guess thats dependent on how the code is organized [17:18] arges: i'd imagine so, unless it's conditionalized like a lot of things I've seen lately have been if gen6 || gen7 ... [17:18] arges: but i'd bounce that off of someone who knows more :) [17:18] brendand: it was Oneiric server (bug 897243) [17:18] Launchpad bug 897243 in linux "Dell PowerEdge 2900/2950 crashing with Ubuntu Server 11.10" [High,Triaged] https://launchpad.net/bugs/897243 [17:20] vanhoof, yea, def. anyway thanks! [17:21] bjf - ah. that's a totally different story. [19:07] * tgardner -> lunch [19:25] arges, I know that in at least some areas of the code i915 calls ironlake functions for anything with a PCH, so that would include sandybridge [19:25] so yes, an x220 with sandybridge would be calling at least some ironlake functions [20:37] hggdh, i've added a comment to bug 238208. also, at least one person reported that it fixed the issue for them. i believe the commit should remain. [20:37] Launchpad bug 238208 in linux "Need MemoryStick driver Ricoh R5C592 (part of R5C832/822chipset)" [Wishlist,Fix released] https://launchpad.net/bugs/238208 [20:42] bjf, sometimes its the media that doesn't work. I've found quite a bit of variation in those SD cards [20:43] tgardner: thanks, i'll add that as a comment [21:13] * cking battening down the hatches, incoming storm [21:55] anyone know what %w0 means in inline assembler? [21:57] bjf: thank you [22:04] * tgardner -> EOD === maxb_ is now known as maxb