=== BenC__ is now known as BenC [02:31] Hi there. Natty 2.6.38-8 panics on me about once a day. Where should I report this? [02:32] Launchpad. [02:32] Yes, but where abouts? [02:32] And, I guess you've reported it here, too. Does apport not pop up suggesting you file a bug? === jk-- is now known as jk- [02:33] No, the machine dies and there's no apport dialog when it comes back up. [02:33] “ubuntu-bug linux” is the requisite incantation for bug reporting if apport's not catching it. [02:33] Ah, ta. Running now... [02:34] I have a photo of the panic backtrace. How can I attach it? [02:35] Ah, I see there's a browser window opening up... [02:35] Once the bug report is on launchpad, attach the file :) [02:36] now we just need to "close the loop" and get LP to fix the bug [02:38] I'm all for creating a sentient bug-fixing launchpad. [02:38] Chop chop! [02:39] but does the strong anthropic principle apply? [02:40] That a sentient bug-fixing launchpad would only exist in a world filled with bugs? [02:42] sentient? I propose RAOF gets the job of holding his hand over the big red killswitch in case LP goes all skynet on us [02:43] Any sufficiently advanced AI malevolent enough to require use of the killswitch is malevolent enough to remove the ability to press the killswitch :) [02:43] your job with the killswitch includes monitoring the killswitch [02:44] :D [06:56] * apw yawns [06:57] Want some more bees to gargle? :) [07:00] i like my women like i like my coffee covered in bees [07:02] RAOF, hows things looking for release from your point of view [07:02] By and large, pretty good. [07:03] there are some bugs out there, even some nasty ones, but not a huge heap of them [07:03] There are some nouveau crazies, but for some reason laptop manufacturers like to put totally insane bioses on nvidia systems. [07:03] There's one big bug involving compiz freezing that I thought we'd fixed on Friday, but sadly not. [07:03] that is so helpful of them, you'd think that the nvidia would specify the bios really [07:03] It might be in the kernel (missing vblank events) or it might be in libx11. [07:04] hrm [07:04] which grphics? or is it all of them? [07:04] intel! [07:04] No one uses it. It's pretty obscure :) [07:04] yeah minor corner case [07:05] * apw shakes head [07:05] Oh, and btrfs remains molasses-slow, but that's hardly news :/ [07:06] yeah, that one has yet to live up to its billing, though that is the common case of all new filesystems [07:06] they are really quick until the data loss bugs are gone [07:06] Heh. [07:06] ext4 and ext3 went trhough being as fast as anything, but your data was on a knife edge [07:07] Which is a pity, because btrfs has some all sorts of awesome in it. Once it's fast enough for regular use, I guess. [07:08] yeah, it has a nice model if it can ever be made performant [07:08] great for rollback and the like [07:09] Hot data migration for ssds would be nice, too. [07:09] And nice schroots and such. [07:09] so there doesn't seem to be anything for i915 at all in either 2.6.38.3 (which is pending for the first SRU) or in the stable queue [07:10] Yeah, I don't think there's anything anywhere for it. I'm not yet sure it *is* an i915 bug. [07:38] today is our last window for an upload really, and right now there is nothing pending which warrents the risk of actually doing one as far as i can see [07:38] RAOF, ^^ [07:45] apw: I concur. === michaelh1 is now known as michaelh1|away === michaelh1|away is now known as michaelh1 === lag is now known as Guest60197 === smb` is now known as smb [08:48] * smb waves to apw [08:48] smb, moin, hows you [08:48] smb, do we still use the cve bzr thingy? [08:48] apw, Good, just slow to get to work. :) [08:49] * amitk read that as CVS bzr thingy and shivered... [08:49] apw, I guess we use the bzr branch but more on a manual approach [08:49] amitk, hehe [08:49] smb, ok so will update the branch then [08:50] apw, I must admit I have not done much cve lately. Too much time spent on other things [08:50] me also, just was updating the awsome tracker with some new cves was all [08:51] *sigh* there is always new [09:05] The cking is back. :) Was it next week when you are on vacation? [09:05] smb, no vacation as yet [09:06] Ah ok, saw it somewhere and was for some reason believing it would be this week [09:07] As this Fridays is a Good Friday. :-P [09:08] smb, yep, may take wednesday off - electrician is doing the final wiring [09:09] cking, Good idea. Not good to have computers running while someone pokes around with a screwdriver [09:09] So You can move around Easter? [09:10] still got to decorate the office - got to wait 5 weeks for the plaster to dry before I can apply paint [09:12] 5 weeks!? That is so close to torture... :-P [09:19] smb, well it's dry-ish now, but if put modern paints on it now it the pain will trap in moisture forever [09:21] cking, Understandable, though it feels very hard to have an office place that is connected and all but you have to wait for some plaster to dry. [10:36] apw, Can you confirm that bug 747090 fix, will land in the next kernel upload? [10:36] Launchpad bug 747090 in linux "wrong return address sometimes pushed for INT in kvm (not qemu)" [High,Fix committed] https://launchpad.net/bugs/747090 [10:36] yes that fix is committed and will be in the first SRU upload [10:37] Daviey, ^^ [10:39] apw, ah, super - thanks! === diwic is now known as diwic_afk [10:53] what about bug 762496 what will happen if compiling the vanilla kernel fixes the issue [10:53] Launchpad bug 762496 in linux "atheros wifi cards cause kernel panic when connecting to wpa2 secured wifi networks" [Undecided,New] https://launchpad.net/bugs/762496 === michaelh1 is now known as michaelh1|away [11:12] eagles0513875, if 2.6.38.3 fixes your issue then that is going to be released as the first sru for the kernel [11:12] which is normally 1-2 weeks after release [11:12] ok there were some patches released to fix that issue for 2.6.38.3 [11:12] eagles0513875, and have you tested that they fix your issue? [11:13] working on it as we speak [11:13] well thats good feedback if you can indicate which patches fixed the problem we can tag them and the bug will close out automatically when the SRU releases [11:14] apw: any chance to get it in before the first sru release [11:14] as there are going to be tons of disgruntled people [11:14] very unlikely now, if this was raised a week ago before the freeze we might have gotten it in, though the fix was only available friday i believe [11:15] ya i was talking to someone about that friday as it has been plaguing me big time [11:15] that is the problem with people waiting till the last minute to test, its all too late to do anything about [11:16] no chance to file a freeze exception [11:16] unless someone in the kernel team happens to have the h/w we simply don't find out [11:16] the problem is timing, a kernel rebuild takes close to 24 hours, and it has to be done before 9am tommorrow [11:17] which is basically impossible now [11:17] ahh [11:17] 24 hrs O_o [11:17] im suprised canonical doesnt have a dedicated high end machine to cut down the kernel packaging and compilation time [11:17] and then we have to respin the installer etc [11:18] eagles0513875, there are no such things as high end armel boxes [11:18] which is why our freezes are loudly advertised, and pretty much hard [11:18] and why testing early is sooo important [11:19] then in that case whats the earliest one can get in on upgrading to the next dev release [11:19] that isnt part of the canonical staff [11:19] eagles0513875, from the moment the archive opens, [11:19] which is usually how long after [11:19] release [11:19] there is no special treatment for canonical staff [11:19] usually less than a week [11:19] but even testing the alphas would be good [11:20] those appear very early in the cycle [11:20] ya the problem for me is time atm :( [11:24] apw: mind if i pm ya a sec [11:24] eagles0513875, yep, a problem for a lot of people, but it means you'll likely have to wait a couple of weeks longer than the rest to upgrade [11:24] eagles0513875, ok [11:24] im already on natty i just cant use the wifi at home === diwic_afk is now known as diwic [11:37] eagles0513875, you might want to test the kernel which is in the pre-proposed archive [11:37] https://launchpad.net/~kernel-ppa/+archive/pre-proposed [11:37] i was gonna recompile as there are some default things which on a netbook i dont need [11:37] that should be similar to the SRU kernel and contain the .3 fixes [11:38] :) will give that a shot. is it possible if that has the fixes to get that pushed today before release [11:38] that is simply built from the tip of our tree, that is what is committed for the next upload [11:39] i will discuss it, but i suspect we are unlikely to be uploading before release now [11:39] the biggest prevention you have here is disgruntled users [11:39] upon release [11:40] eagles0513875, i have to balance risk, known breakage for you and the (relativly) small number of atheros users against the risk of making any change which can affect all other users [11:40] true [11:40] as esentially any change now invalidates all of the testing we have done to now [11:41] now in theory its low risk, but ... non-zero and we have been bitten by just recompiling breaking things [11:41] i will try and get a specific kernel with those atheros patches today for you to test at least [11:42] im trying the preproposed as you mentioned above if that containes the 38.3 fixes [11:42] i dont think kde 4.6 is ready for prime time though im getting crashes left right and center [11:44] you are 64 bit yes ? [11:44] all im noticing right now somethign is seriously up with networking stack all of a sudden here :-/ [11:44] yep [11:44] maverick 64bit is rock solid [11:44] is the pre proposed up or down? [11:44] i ran apt-get update and im getting an error no address associated with hostename [11:48] eagles0513875, was there when i looked just now [11:49] ok let me reboot im having some connectivity issues [11:58] now for the moment of testing truth [11:58] apw: what i did find funny in testing [11:58] if i leave my wired network plugged in and then connect to wpa2 connection it doesnt kernel panic [12:00] eagles0513875, odd indeed [12:00] now knetworkmanager is crashing non stop for me [12:00] woohoo [12:05] apw: ill keep working on this later im off for lunch for now === diwic is now known as diwic_afk === yofel_ is now known as yofel [12:44] eagles0513875, can you test http://people.canonical.com/~apw/lp762496-natty/ and report back [12:44] * apw drops for a bit [12:58] hi.. i have written code to copy data from kernel to user space.. But its not showin the required output.. :( code is on http://paste.ubuntu.com/595486/ ..... please suggest.. [13:08] M_gerorge: hmm.. to copy from or to user space you should use write and read function [13:09] module_init will only be called once when you insert the module using insmod [13:10] fairuz: ok.. but i have written these copy_[to/from]_user functions in read and write but they are still not called.. [13:10] you have to have a write and read function in kernel space too! [13:10] it's not magic :D [13:11] fairuz: :D.... but copy_to/from_user are inbuilt or not? [13:11] :p [13:12] M_gerorge: Yes, but how you will call it from user space? [13:12] fairuz: yes.. this is the basic which i am not getting. can you please tell me how can i explore about this more? [13:13] M_gerorge: For me, a simple module needs to have at least 4 functions [13:13] init, exit, write and read [13:14] hmm and also, open and release functions [13:14] fairuz:yes.. as i am running the scull driver example it is simply loaded and exited.. but how can i use the functions scull_read and scull_write? [13:14] You need to have a device nod [13:15] fairuz: is this the dev_t *dev? [13:15] M_gerorge: no.. something like /dev/mydevice [13:15] it's a file that you create with mknod [13:16] fairuz: ok .. i can do this.. then what to do next?.. [13:16] :P [13:16] hmm.. wait, I give a link [13:16] http://www.freesoftwaremagazine.com/articles/drivers_linux?page=0%2C0 [13:17] follow the tutorial here and come back if you encounter some problems :D [13:18] fairuz: thanks a lot and i used this tutorial also. but not get my answer.. :( [13:18] fairuz: in this the dev/memory is created [13:18] but how to use the memory_read and memory_write function?.. :( [13:19] Ok, I'll explain. (but remember that I'm not an expert myself and maybe it's a bit off course) :D [13:20] fairuz: it wil be my pleasure to be taught by you..:) [13:20] M_gerorge: Are you sure you followed the tutorial? I dont see any major number in your code === lag is now known as Guest9916 [13:21] fairuz: yes.. actually thats is the normal code i wrote myself to use copy_functions [13:21] fairuz: can i paste the other codes? :p [13:22] to pastebin, sure :D [13:22] thanks.. let me do this [13:24] fairuz: please see http://paste.ubuntu.com/595492/ [13:25] M_gerorge: if you are new to device drivers, I afraid that's maybe a little bit complicated? [13:25] fairuz: i will try my best to understand [13:26] M_gerorge: Can you follow the memory device driver tutorial I gave you? [13:27] fairuz: yes.. [13:27] Come back to me if there are something that you don't understand [13:30] fairuz: i read it and understood it completely.. :p.. but it doesnt gave how to access the memory_read and memory_write functions.. :(.. while doing "echo -n abcdef >/dev/memory" and "cat...." these commands doesn't use these memory_read and write too... :( [13:30] of course it does :D [13:31] when you create a module, you associated a major number to it [13:31] let say 70 [13:31] yes [13:31] And apart from that [13:31] yes [13:31] you need to create a device nod using the same major number [13:31] yes [13:31] let say /dev/mydev [13:32] so in your module, normally you specify the file operations struct [13:32] yes [13:32] e.g struct file_operations pmnc_fops = { [13:32] write: pmnc_write, [13:32] open: pmnc_open, [13:32] release:pmnc_release, [13:32] unlocked_ioctl: device_ioctl [13:32] }; [13:32] so from user space [13:33] when you write to the device nod, it's like you call the write function in the module [13:33] e.g when you do echo "123" > /dev/mydev [13:33] fairuz: ok.... :p [13:34] in the backend, it will call the write function of your module with "123"'s pointer as it's buffer [13:34] ok.. [13:34] yes [13:34] and same with the " [13:35] "cat ..." it calls the read function [13:35] :P [13:35] for read, it's the same principals [13:35] yes [13:35] that is in command line [13:35] but if you want to use it user space program [13:35] yes [13:35] you can open /dev/mydev like normal files [13:36] that use fread and fwrite functions [13:36] ok... [13:36] fairuz: now one issue i found that i added printk lines for debugging.. and they dont get printed.. [13:37] so i concluded that these functions are not called.. :P [13:37] if you do dmesg | tail [13:37] yes [13:37] you will found them :D [13:37] means in "/var/log/message" [13:37] ? [13:37] or tail /var/log/messages [13:38] yes.. i dint find them in that.. :p [13:38] but i may be wrong.. i will check once more [13:38] fairuz: let me check once more this [13:38] printk commands [13:38] M_gerorge: ok [13:38] use pr_info instead [13:39] :D [13:39] ok. thanks.. :).. let me try [13:39] :P [13:48] fairuz: it's called and shown too.. [13:49] nice [13:49] fairuz: thanks a lot for giving me your precious time... :p [13:50] fairuz: will come back to you if i will have more problems.. :D.. [13:52] fairuz: good bye... === diwic_afk is now known as diwic === lag is now known as Guest96979 [15:22] rebooting for updates... brb === fairuz is now known as fairuz-away [17:03] ogra_, linux-meta-ti-omap4 2.6.38.1208.7 accepted === sforshee is now known as sforshee-afk [17:03] tgardner, awesome ! === herton is now known as herton_lunch === diwic is now known as diwic_afk === sforshee-afk is now known as sforshee [17:41] there may be an issue with either apport or ubuntu-but (I assume this is apport too?) as shown in bug 764341 https://bugs.launchpad.net/ubuntu/+source/linux/+bug/764341 [17:41] Launchpad bug 764341 in linux "WARNING: at /build/buildd/linux-2.6.38/net/sched/sch_generic.c:256 dev_watchdog+0x213/0x220()" [Medium,Incomplete] [17:41] Launchpad bug 764341 in linux "WARNING: at /build/buildd/linux-2.6.38/net/sched/sch_generic.c:256 dev_watchdog+0x213/0x220()" [Medium,Incomplete] https://launchpad.net/bugs/764341 [17:41] oh good, the bot is back :) [17:45] who feeds and tends the bot? [17:49] nobody, apparently. Thus it goes off on its own? ;-) === sforshee is now known as sforshee-lunch [18:08] tgardner, ever seen this? bug 764273 [18:08] Launchpad bug 764273 in linux "WARNING: at /build/buildd/linux-2.6.38/ubuntu/aufs/plink.c:450 au_plink_put 0x4c/0xa0 aufs " [Undecided,New] https://launchpad.net/bugs/764273 [18:09] JFo, I'm wondering why aufs is even loaded? [18:10] *shrug [18:10] unless he needs it for whatever he is building (mozilla maintainer) [18:11] I confess I know next to nothing about aufs [18:11] tgardner,: micahg, that plink error you reported against the kernel... it happens every time the chroot is destroyed? JFo: yes, but I think it might be related to being in /dev/shm [18:12] JFo, well, that makes two of us. apw is a bit more familiar with it. [18:12] heh [18:13] JFo, he's not even running our bits: aufs 2.1-standalone.tree-38-rcN-20110207 [18:14] hmmm [18:14] JFo, the file that the WARNING is coming from is only 396 lines in our version [18:16] tgardner, JFo: well, jdstrand added a way to use sbuild in /dev/shm, so I thought I'd try it [18:16] he says he is using the latest kernel: JFo: 2.6.38-8.42 [18:16] uhh, I certainly am not using any other code. all I do is put the sbuild overlay in /dev/shm somewhere [18:16] but where would he have gotten that aufs? [18:17] jdstrand, I think the aufs may be the issue [18:17] but no more than I know... that is like saying the stars are made of really hot cheese === herton_lunch is now known as herton === bjf changed the topic of #ubuntu-kernel to: Home: https://wiki.ubuntu.com/Kernel/ || Natty Kernel Version: 2.6.38 (Kernel is Frozen) || Ubuntu Kernel Team Meeting - April-19 - 17:00 UTC || If you have a question just ask, and do wait around for an answer! [18:31] eagles0513875: did that kernel work? [18:33] bah got bounced [18:39] dang, where has the day gone? [18:39] <-lunch [18:40] it's hiding in the shadowes [18:43] apw: I was just checking our aufs bits. The SHA1 in the BOM doesn't appear to be valid in the upstream repo. [18:44] will cjeck it out shortly, there is an algo, and two repos to che k [18:45] * smb imagines apw_ trying to hit the softkeys while humping along... [18:45] the mind boggles [18:45] yep not as easy as it could be, br in front of a stationary one shortly [18:46] apw: are you someplace weird? [18:46] he was in France earlier [18:46] cking, oh, a train? [18:46] I guess a british trains counts as weird... [18:46] on a train for the next half or so [18:47] apw: ah, well it can wait. [18:47] yep woll look shortly [19:03] * tgardner --> lunch [19:07] apw: hey quick question where can i download a newer version of knetwork manager === sforshee-lunch is now known as sforshee [19:20] * bjf -> lunch [19:49] JFo, that bug is mean to be removed from our kernel [19:49] (as in the message) [19:49] apw, which? [19:50] JFo, the plink one [19:50] unless i messed up the kernel update last time [19:52] ok, anything in particular I need to do to it? [19:52] or wait and see? [19:57] JFo, oh its a different one, hrm [19:57] :) [19:58] I'm on a full pot of coffee it could have been anyth.... OMG pwnies! [20:00] JFo, heh i bet [20:00] :) [20:00] * jjohansen -> lunch [20:00] the sad part is, I didn't realize I drank the whole thing until I went for another refill :-/ [20:03] ok tgardner this version of aufs seems to be the one in the tree i have; it seems to be a few commits behind which i can look over for sru but it does correspond [20:04] apw: you're looking at bug #764273 ? [20:04] Launchpad bug 764273 in linux "WARNING: at /build/buildd/linux-2.6.38/ubuntu/aufs/plink.c:450 au_plink_put 0x4c/0xa0 aufs " [Medium,Incomplete] https://launchpad.net/bugs/764273 [20:05] well i was answering your previous comment that the BOM commit id didn't exist [20:05] (which i contend it does) and that WARN_ON seems to be on that line in the natty source [20:06] apw: hmm, then one of us is doing drugs. [20:07] COMMIT: 65835da20b77c98fb538c9114fc31f5de1328230 [20:07] is in the BOM [20:08] apw: agreed. [20:08] rtg@lochsa:~/proj/linux/aufs2-standalone$ git log -p 65835da20b77c98fb538c9114fc31f5de1328230 [20:08] fatal: bad object 65835da20b77c98fb538c9114fc31f5de1328230 [20:08] commit 65835da20b77c98fb538c9114fc31f5de1328230 [20:08] Author: J. R. Okajima [20:08] Date: Thu Feb 3 04:46:48 2011 +0900 [20:08] aufs: possible bugfix, exclude the freeing file [20:08] ahh yes, its not a direct commit id in that tree [20:08] its the commit id as listed in the Changelog [20:09] which is the id in the aufs main tree that the commit in the standalone tree was made from [20:10] apw: ok, I'll take your word for it. the real issue is that I don't think he's running _our_ version of aufs. 'WARNING: at /build/buildd/linux-2.6.38/ubuntu/aufs/plink.c:450 au_plink_put+0x4c/0xa0 [aufs]()' references a line number that doesn't exist. [20:10] tgardner, the problem is that the primary tree is increadibly slow [20:11] tgardner, in my tree here this is the plink.c:450 which look plausable [20:11] WARN(verbose && !list_empty(plink_list), "pseudo-link is not flushed"); [20:11] and the panic information looks like its 8.42 [20:11] apw: ya know, I might have been on the wrong branch. shit. [20:12] heh happens to the best of us [20:12] more coffee required methinks [20:13] maybe soo indeed [20:13] none for me thanks... [20:13] * JFo has the jitters [20:16] * apw slaps JFo ... [20:16] * JFo jumps 10 feet in the air [20:16] :) [20:17] now that i want on youtube [20:19] heh [20:21] dear bisected kernel, please don't misbehave, it makes bisecting very difficult... [20:21] cking, i hate those days [20:22] just when you think you've cornered it, it jumps out of reach [20:40] do we care about bug 763656 or should that go to skype? [20:40] Launchpad bug 763656 in linux "general protection fault: 0000 [#1] SMP " [Undecided,New] https://launchpad.net/bugs/763656 [20:40] since they were installing skype when it happened [20:40] well there is little we can actually do about it [20:41] I thought that might be the case === michaelh1|away is now known as michaelh1 === michaelh1 is now known as michaelh1|away === michaelh1|away is now known as michaelh1 [21:39] * JFo needs a bit of a break: bbiab [22:20] JFo, i've submitted a merge request for apport changes which remove all the user-input from the linux source package-hooks [22:20] pgraner, ^ [22:20] JFo, pgraner bug #765178 [22:20] Launchpad bug 765178 in apport "remove user-input questions from linux source package hooks" [Undecided,New] https://launchpad.net/bugs/765178 [22:28] bjf, ack [22:29] pgraner, by the way *I HATE BZR* [22:29] bjf, :-D [22:40] ok