=== _LibertyZero is now known as LibertyZero === tjaalton_ is now known as tjaalton [07:49] Morning [08:37] hi! what's the magic git command to figure if a commit is in the master-next branch of ubuntu-natty (and thus will be in the Natty final kernel), or if I have to ask for it explicitly on the kernel-team ML? [08:38] diwic, You know you just can look at that branch? git log origin/master-next [08:39] (from a natty tree that is) [08:39] smb, right, then I can pipe that to grep. Didn't think of that, thanks. [08:40] diwic, there would even be a --grep="bla" for git log [08:42] I believe there's even a “git contains” somewhere... let me se. [08:42] RAOF, You probably mean git describe --contains [08:43] smb: I do indeed. [08:44] The downside there is that you need a sha1 on natty maybe still be able to be rebased makes it hard to know that [08:45] Actually master-next is never linear === doko_ is now known as doko [11:05] hey. i just opened bug 752352 [11:05] is this expected behavior ? [11:06] it was not listed in the changelog. [11:41] hi. is it possible to build the linux-tools package with make-kpkg? [11:51] smb, you know anything about bug 752352 ? [12:19] smoser Could be a security feature... I remember kees was asking about some file permissions recently [12:19] its not listed in changelog. [12:19] and the files are publically available on launchpad or apt repository [12:20] so just plain hiding their content isn't all that useful. [12:20] ie, if i can't read it in /boot i can download the deb myself [12:22] Well, I am the wrong person to argue about that. ;) But it makes it harder for automated attacks I guess. [12:23] See "UBUNTU: [Config] packaging: adjust perms on vmlinuz as well" [12:26] smoser, Is that a real problem. At least for starting instances one needs to be root anyway... [12:26] not a problem for starting anything [12:26] but caused my builds to break [12:26] they assume that they can read contents of /boot [12:26] to copy the kernels out [12:27] they can very easily be root to do that, the scripts just be root as little as they have to [12:27] smoser, Doh! [12:29] smoser, The reasoning from the patch is that exploits could get the location of kernel addresses from there. I agree that they can have locations prepared per kernel versions... [12:30] so that *is* intended behavior ? [12:31] smoser, It was a change that was requested by the security team, yes [12:31] and, by design not listed in the changelog [12:31] :-( [12:31] It should be in the changelog [12:31] It is at least referenced there in the file [12:32] linux (2.6.38-8.40) natty; ... [12:32] ... [12:32] * [Config] packaging: adjust perms on vmlinuz as well [12:33] I am ready to admit that this is not that obvious [12:33] ah. sorry. i looked in the specific binary's config [12:34] oh, and it is there also... i just looked at the top entry [12:34] sorry for the spam [12:34] no worries. There is a lot of text in the changelog at the end of a day [12:40] thank you smb [12:59] It seems that the latest -8 kernel makes update-grub not create the subvol=@ entry (for btrfs), are there any changes which could be related to that? [13:10] i have made small "hello" type driver.. now i would like to start with desigining drivers for ubuntu. please suggest me how should i start? [13:12] sconklin, i have read the book you suggested me and tried the basic examples. now please suggest me to start development with community. :) [13:13] aakshay: which book? LDD? [13:13] fairuz, yes.. [13:14] fairuz, it gave me the basic idea only.. how to go further ?... :) [13:14] aakshay: for me the book is complete :D So what did you managed to do until now [13:15] external module? [13:15] fairuz, i designed the basic initialization module. [13:16] and what did you initialized? [13:16] Is there any http acces to the git repos of ubuntu by the way? [13:16] fairuz, a simple function... :p [13:16] which displayed some info like " entering the module" [13:17] fairuz, means very basic.. :D [13:17] aakshay: So have you tried to do some communication between user space and kernel space? [13:17] fairuz, nah.. not yet :( [13:17] aakshay: so i think you better try that first [13:17] should i go for that now? [13:17] ok :) [13:17] aakshay: http://www.freesoftwaremagazine.com/articles/drivers_linux?page=0%2C3 [13:17] a good tutorial [13:18] aakshay: http://www.freesoftwaremagazine.com/articles/drivers_linux?page=0%2C0 this is the first page [13:18] okiez.. i will do according to this tutorial [13:18] :) [13:18] okiez.. [13:18] but to run these modules, do i need my own build kernel?.. ;p [13:19] this tutorial is for our of tree modules [13:19] it means you just need the kernel source to build it [13:19] no need to recompile your kernel [13:19] *out [13:19] *out of tree [13:20] fairuz, ok but please tell me something. i am designing all these codes in the kernel directory of my normaly ubuntu i am using [13:20] i have not built any separate kernel [13:20] no problem with that [13:21] ok.. [13:21] you compile a module for a specific kernel [13:21] fairuz, thankyou so much... [13:21] if you are compiling using your currrent kernel, then it will run only on the current kernel [13:21] okiez... but what specific kernel means here? :-o [13:21] the kernel that you are using :D [13:21] type uname -r in your terminal to see which kernel you are using [13:22] fairuz, okiez... :p [13:22] fairuz, thanks a lot.. can you please tell me how long will it take to complete with this tutorial? :p [13:22] aakshay: 30 minutes ? [13:23] or less [13:23] ohhhh!!!! .... :D [13:23] fairuz, so let me try this to complete this in 30 minutes [13:23] :P [13:23] aakshay: ok, but you need more time to understand... dont just copy paste the code [13:24] fairuz, yes.. thanks a lot... :) [13:24] do page 1 to page 7 of the tutorial [13:24] aakshay: np [13:24] fairuz, okiez.. let me start noe itself [13:24] :P [13:24] *now [13:49] It seems that the latest -8 kernel makes update-grub not create the subvol=@ entry (for btrfs), are there any changes which could be related to that? Also, is there any git http access so one could clone the repos and have a look? (I'm badly firewalled). [13:49] arand, http://kernel.ubuntu.com/git [13:54] Hmm, I don't see any http links there, and using git clone http://kernel.ubuntu.com/ubuntu/ubuntu-natty.git fails... === maxb_ is now known as maxb [14:01] arand: git clone git://kernel.ubuntu.com/ubuntu/ubuntu-natty.git [14:02] fairuz: I am firewalled and can't use the git:// protocol (unless I go through tor, at 60K/s ...). [14:13] diwic: bug 732445, looks like you mention the patch has landed upstream in the takashi's sound-2.6 tree, did you want it applied as pre-stable to natty? [14:14] ogasawara, let me see [14:15] diwic: if so, care to just send a quick email to the list [14:15] ogasawara, ok, I want it in [14:18] ogasawara, I think it's already in [14:19] ogasawara, it's just that since Takashi chose a different solution he took away my buglink in the process [14:19] diwic: ok cool [14:20] ogasawara, if git log origin/master-next --grep 584c0c4c359 says it's in, it's in, right? [14:20] diwic: yep, it's in [14:21] git describe --contains 24449f05b663f1984ee6ecdba5a0d2ff2a1e0462 [14:21] Ubuntu-2.6.38-8.40~661 [14:27] * ogasawara back in 20 [15:00] fairuz, am back. :) [15:00] understood the coe [15:00] *code [15:00] aakshay__: make it worked then? [15:00] but while running , i am getting the error :( [15:00] "error: expected declaration specifiers or ‘...’ before ‘sizt_t’ [15:00] " [15:00] fairuz, :( [15:01] aakshay__: try to debug it :D it should give you the line number [15:01] fairuz, its taking ssize_t as 'sizt_t' [15:01] yes its line 17 and 18 [15:01] aakshay__: check your code. You do copy paste from the site or write it on your own? [15:02] i took the same code but wrote myself.. [15:02] fairuz, so that i will learn :p [15:03] whats on line 16 to 18? [15:03] please wait.. let me copy the code on pastebin.com [15:03] :p [15:04] fairuz, its "http://paste.ubuntu.com/590256/" [15:05] aakshay__: As i said before, take a good look at the code and you will find it [15:05] you miss spell something on line 17 and 18 [15:05] oops!! .. let me check.. ;p [15:06] hehe.. :D.. such a silly mistake.. n took lot of time to get corrected.. :p.. thanks.. [15:06] let me compile again.. :) [15:07] fairuz, yipii.. done!! [15:07] :) [15:07] aakshay__: When it says error at line 17, don't look at line 40.. =) [15:09] fairuz, :p... i looked at line 17 only.. :D... i was only lookin at ssize_t , not on required variable.. :D [15:15] fairuz, its done.. :).. and i understood it well.. please tell me what i can do further? [15:15] aakshay__: you tried to insmod and rmmod your module? it worked? [15:15] yes [15:15] i removed the modules [15:15] an checked them using the lsmod command [15:16] :) [15:16] *and [15:16] aakshay__: you tried to communicate with your module from user space? [15:17] fairuz, dont know [15:17] :p [15:18] aakshay__: so you dont completed the tutorial -.-" [15:18] i entered the data. it was from the user space then its copied to kernel space to be accessed from /dev/memory [15:19] :p [15:19] fairuz, is this was not the concept?.. :) [15:21] fairuz, oh yes.. you are right.. the data in the tutorial is sent directly using the "echo" command [15:22] there is no use of functions "copy_from_user" and so [15:22] :P [15:25] fairuz, i have not done the "parallel port" part... [15:25] fairuz, how can we make use of the functions "copy_from_user" and others ? [15:35] bryceh_, no joy on bug #686388 [15:43] fairuz, thanks for help.. :).. i have to leave. good bye.. . :) [15:45] smb, do you remember how to skip failing mounts at boot up? mountall is hanging 'cause /dev/md0p1 hasn't been enabled when /etc/fstab is read. [15:46] tgardner, typing s [15:46] though it might work only with splash [15:46] smb, k, I'll go try that. === tgardner is now known as tgardner-server- === tgardner-server- is now known as tgardner [15:54] smb, huh, it even tells you what to do, but you have to be able to see the console. [15:55] tgardner, Sort of makes sense as you are usually in front of it to type the letter [15:57] tgardner, This is where I start to like boards with integrated ipmi module. The only thing I have not yet got running is to get the java based graphical console over a tunnel when I am not in my home network. :) [15:58] smb, I suppose this emerald has that, but I'm just too lazy to get it working. I'll only have it for a limited time. [16:01] tgardner, Too true. It was relative simple here as it got its own lan port. The only annoying part is that it only runs with sun-java installed (he graphical console) [16:19] <-lunch and a few errands [16:19] bbiab [16:46] ogasawara, these are probably already on your radar: http://people.canonical.com/~kernel/reports/regressions-proposed-report.html [16:48] bjf: thanks, I'll browse through em to just make sure there's no kittens dieing [16:48] ogasawara, no pink kittens this time around? how about green? [16:49] or is that dyeing ? [16:49] I can never remember [16:49] heh, I can never remember either [16:59] ogasawara, ooh! I fixed bug #747364. do I get brownie points ? I used the "Won't Fix" club. [17:00] tgardner: heh, I'll buy ya a beer in budapest [17:00] ogasawara, big spender. it'll only cost you $1.80 [17:01] tgardner: diapers are expensive these days, it's all I can afford :) [17:01] ogasawara, go cloth === herton is now known as herton_lunch [17:02] * smb tries not to think too much about the term brownie points now [17:02] * tgardner thinks smb is a sick puppie [17:02] ogasawara, i don't know if "fixing a bug" by marking it "won't fix" constitutes a beer, but maybe a 1.80 beer === cmagina is now known as cmagina-lunch === sconklin is now known as sconklin-lunch [17:46] ogasawara, how are your oneiric armel build issues coming? [17:46] tgardner: should be fixed, final build is just finishing up [17:46] tgardner: I pushed the bits to the repo [17:46] ogasawara, cool [17:46] tgardner: am about to rebase to 2.6.39-rc2 [17:47] ogasawara, which will introduce a whole new raft of issues :) [17:48] ogasawara, don't forget to pick up master-next commits [17:48] tgardner: ack === herton_lunch is now known as herton === cmagina-lunch is now known as cmagina === sforshee is now known as sforshee-lunch [18:22] jjohansen: ping [18:25] kristian-aalborg: pong [18:26] I've been thinking.... the install I tested my custum kernels on is kind of a mess since I have experimented a lot [18:27] yeah that happens [18:28] fun times indeed... but it may be a factor in my kernels messing up [18:29] or anyway, that thought crossed my mind [18:29] yes that is possible [18:29] oh hey jjohansen, you pinged me the other night but I just missed you when I responded. Do you still need me? [18:30] generally we try to build inside a chroot to keep the environment consistent [18:30] JFo: hrmm, I can't remember why I pinged you, so you are in the clear :) [18:30] oh good :-) [18:31] JFo: hrmm maybe it was for a bug nomination [18:31] could be [18:31] if you remember it I am happy to fix it [18:31] or rather approve it [18:32] jjohansen, was it to add a linux(Ubuntu) task to this one? https://bugs.launchpad.net/linux/+bug/744419 [18:33] JFo: actually it was Bug #748656 [18:34] * JFo looks [18:34] jjohansen, done === kentb is now known as kentb[dell] [18:35] JFo: so Bug #744419 is not kernel related, we traced it down to eucalyptus [18:35] or at least that is the current working theory [18:36] ah hah [18:36] kvm runs fine, and this is failing when we put in the maverick kernel [18:36] well, I added a kernel task just in case [18:36] if it isn't needed it is an easy thing to close [18:36] just don't want it to disappear [18:36] so for the moment lets mark the linux task invalid, and we can change it back if needed [18:36] ok [18:37] JFo: arg no wait [18:37] * jjohansen was looking at the wrong bug [18:37] heh [18:37] JFo: that one is linux, its the rtl8192ce driver, traced that one down on the weekend [18:38] ah, I just read your comment [18:38] it bugs 746751, that isn't the kernel [18:38] we were chasing that on friday and the start of this week [18:39] * jjohansen is getting confused by having too many bug tabs open in his browser [18:39] I know the feeling :-( [18:40] ogasawara: just a heads up on this one too, bugs 746751 appears not to be kernel related at this time serge is chasing the eucalyptus path [18:43] jjohansen: cool, thanks [18:45] ogasawara: did smb update you on bug 634487 [18:46] jjohansen: he mentioned it the other day and I've been following his comments to the bug [18:46] ogasawara: okay, great === sconklin-lunch is now known as sconklin [18:53] tgardner, yeah, although upstream had alluded to known issues with macbooks. I wish Intel were a large enough company to employ more than one engineer in their QA department. [18:57] bryceh_, thats the only feedback so far. is the macbook a regression? [18:59] tgardner, upstreams exact words were "Those two patches have been reported by one user to have fixed the issue for [18:59] him, but I need a few more testers since they seem to foul up a MacBook (but [18:59] then there are more than one issue at play with MacBooks...)" [19:00] so, I would interpret it to mean that yes it is a regression on macbooks [19:00] bryceh_, gah, what a mess. [19:02] maybe if Intel can make some profits they can buy that guy a macbook to test [19:03] tgardner, anyway I'm thinking maybe we wait until there's something actually posted to drm-next. These one off patches seem sketchy too often. [19:03] bryceh_, ack [19:04] smoser: yeah, while the /boot perms obviously don't stop someone from just going and getting the deb themself, it does stop some "simple" attacks that will be dumbfounded by the lack of kernel symbols. [19:05] smoser: it's crude and effective for a certain set of attacks, and does nothing to stop others. I felt it was worth it; sorry about the breakage it caused you! === sforshee-lunch is now known as sforshee [19:08] i really don't think it woudl stop an attack [19:09] * tgardner --> lunch [19:09] i really can't imagine someone saying "oh bother, i can't read the kernel symbols for this kernel, i guess i wont' insert a root exploit today" [19:09] "i'd have to do a 'wget' in order to do it, and thats really just too much work" [19:18] I'm going to install a backup kernel, I need linux-headers-xxx, linux-image-xxx... and anything more? [20:05] * jjohansen -> lunch [21:56] * ogasawara back on later === sconklin is now known as sconklin-gone [23:27] JFo: got a kernel regression for you... 731878