[00:48] linux-tools is uninstallable on current natty :'( [00:51] \o/ === ericm-Zzz is now known as ericm === ogra is now known as Guest2736 === Guest2736 is now known as ogra_ [01:34] * jjohansen reboot === MTecknology is now known as MT-TakinBreak === _LibertyZero is now known as LibertyZero [06:05] hello, I am having the same issues building the kernel modules virtualbox via dkms that this user is: http://lists.debian.org/debian-user/2011/01/msg01508.html ... however I am using the stock pae ubuntu kernel instead of a custom built one. I installed the pae headers, but I am still having the same issue. Is there some other change I need to make so that I can build this module? === _LibertyZero is now known as LibertyZero === smb` is now known as smb === ericm__ is now known as ericm-dinner === diwic is now known as diwic_afk === ogra_ is now known as ogra === diwic_afk is now known as diwic [12:39] hi, how different is -6 kernel to rc8? [12:48] any idea when the fix for bug #605042 will be uploaded to lucid? [12:48] Launchpad bug 605042 in linux-fsl-imx51 "[armel] java fails to start with eglibc-2.12-0ubuntu4" [Critical,In progress] https://launchpad.net/bugs/605042 [13:09] what is the equivalent library in kernel space for unistd.h? === ericm-dinner is now known as ericm|ubuntu [13:12] doko_, I saw that patch just today submitted for sru. I don't think it was committed yet and then it would need another upload which I am also not sure when that could be. [13:12] sconklin, bjf Would you happen to know about fsl plans for Lucid? [13:14] smb: we need the kernel on the buildd to fix it in eglibc ... [13:14] doko_, working on it... [13:15] thanks === diwic is now known as diwic_afk [14:43] Hi, can I call smp_processor_id() in user space? [14:47] fairuz: No [14:48] fairuz: Also, unless you have affinity, it's not a meaningful thing to do [14:48] fairuz: You could be rescheduled to a different CPU between making the call and doing something with it [14:48] mjg59: I'm running a dual core cortex-A9 [14:48] and I can set the affinity of a process [14:49] fairuz: Then sched_getaffinity will give you your affinity [14:49] mjg59: ah ok [14:49] mjg59: so a setaffinity will affect only one process right? [14:49] fairuz: Yes [14:49] See man sched_setaffinity [14:50] mjg59: ok ty [14:50] The first argument is the pid [14:52] mjg59: so basically we cant set an affinity for a program that not yet running? [14:53] fairuz: No, but you can launch it with taskset [14:53] mjg59: hmm whats that [14:53] fairuz: What are you actually trying to achieve? [14:53] fairuz: A command that launches a process on a specified set of CPUs [14:53] mjg59: actually i have a program..lets call it X [14:53] that i dont want to modify its source code [14:54] and I have another program that sets some hardware registers [14:54] so basically the flow is [14:54] 1) call my program, and set something [14:54] 2) call program X [14:54] 3) call my program again to reset something and colelct data [14:54] *collect [14:55] so if possible I want all of them to run on a processor of my choice [14:55] wrap all that in a shell script that you start with taskset [14:55] I already made a shell script to launch all that [14:55] unless you mean the two need to run on different CPUs, taskset is still the solution but you use it in the shell script [14:56] Krunch: ok, I think I have all the elements, So right now I just need to launch the script from taskset [14:57] and the affinity will be inherited by its children [14:57] what taskset will do exactly? [14:57] fairuz: taskset will set the affinity and then exec your process [14:57] fairuz: Which will then inherit that affinity [14:58] So taskset -c 1 myscript.sh will run myscript on cpu 1 [14:58] And taskset -c 0 myscript.sh will run it on cpu 0 [14:58] mjg59: Ah ok, so I set the affinity on a higher level [14:58] because right now I jsut set the affinity in my program, and not in program X [14:58] s/myscript/myscript and the processes it spawns/ [14:58] and I know its wrong [14:59] Krunch: mjg59: ok thanks.. I think I get it now [14:59] well you can have the 1st program set affinity of program X as well if you know its pid but that's just more work while you can make it simpler by taking advantage of affinity inheritance [15:00] Krunch: but I dont think thats possible, because at the moment the 1st program runs, the program X is not yet launched thus it does not have any PID yet, right? [15:02] i am sure you can devise some ipc mechanism to give it the pid [15:02] but doing it that way is really useless work and complexity [15:02] Krunch: ok. I'll try taskset [15:03] So all programs launch by a script are considered the children of that script? [15:03] ## [15:03] ## Ubuntu Kernel Team Meeting - Today @ 17:00 UTC - #ubuntu-meeting [15:03] ## agenda: https://wiki.ubuntu.com/KernelTeam/Meeting [15:03] ## [15:03] fairuz: Yes [15:04] mjg59: I hope taskset is not arch dependant since I'm using ARM [15:04] some tools are not working on ARM [15:04] =( [15:04] I hope that too! [15:05] mjg59: Krunch: Thanks [15:11] So I imagine if I set taskset -c 3 myprog [15:12] then it up to the kernel to choose which processor it want myprog to run? === ericm|ubuntu is now known as ericm-Zzz [15:32] "-c3" means to run only on the 4th cpu [15:38] Krunch: I thougth it means cpu 0 and cpu 1 are chosen> [15:40] fairuz: No, -c is a list of CPUs, not a mask [15:44] mjg59: ok [15:48] pgraner, can you accept the natty nomination on bug 508516 [15:48] Launchpad bug 508516 in linux "wakeup from sleep fails on toshiba nb305" [Undecided,In progress] https://launchpad.net/bugs/508516 [16:30] ## [16:30] ## Ubuntu Kernel Team Meeting - in 30 minutes - #ubuntu-meeting [16:30] ## === jepler- is now known as jepler [17:13] hi [17:13] please i need your help [17:13] i just tried to update samba on my server and i f*** it up badly [17:14] my server is ubuntu 10.10 and i tried to update to samba 3.5.8 from here [17:14] ftp://ftp.sernet.de/pub/samba/3.5/debian/dists/lenny/main/binary-amd64/ [17:14] now i am left without samba [17:14] and this: [17:14] <-lunch [17:14] bbiab [17:15] Ntemis, you tried to update from a lenny package? i suspect that is not the best plan [17:15] http://pastebin.com/NadhUFQg [17:16] please help me unistall the conflicting packages [17:16] i want my samba back :) [17:16] could not find any deb for 3.5.8 version [17:17] my kernel is 2.6.37.3 :) [17:17] so i wanted to use the latest samba also, my fault [17:18] fortunately i backed up my samba.conf [17:18] Ntemis, i would try and install the official version of samba [17:18] as the first step [17:19] no go [17:19] have a look in my pastebin [17:20] packages have unmet dependencies [17:21] jjohansen, just a heads up, you've been dropped from the meeting agenda :-) [17:22] bjf: okay [17:22] bjf: shooting for 6 min next week :) [17:22] jjohansen, heh, something like that === doko_ is now known as doko === bjf changed the topic of #ubuntu-kernel to: Home: https://wiki.ubuntu.com/Kernel/ || Natty Kernel Version: 2.6.38 || Ubuntu Kernel Team Meeting - March-15 - 17:00 UTC || If you have a question just ask, and do wait around for an answer! [17:47] JFo: you here? === MT-TakinBreak is now known as MTecknology [18:22] sconklin, yep [18:22] am now [19:34] jjohansen: fyi, the natty toolchain problem (".size expression ...") seems to be fixed in the archive now. natty kernel build is no longer failing on Launchpad. [19:39] kamal: oh, /me was just starting to look as I did an upgrade this morning and was seeing the failure locally [19:39] * jjohansen goes to update again [19:39] kamal: btw thanks for the heads up [19:40] jjohansen: sure :-). apw uploaded a new natty kernel, and it sees to be building fine on LP now. I've successfully built amd64 on tangerine, and am testing i386 there now. [19:41] s/sees/seems/ [19:46] jjohansen: natty kernel {i386, amd64} now building fine on tangerine again [19:47] nice [19:48] I've just sucked in the new tool chain locally and have fired off a test [19:48] jjohansen: pretty massive breakage... I'm surprised it got into the archive at all in that state. anyway ... back to the old grind. ;-) [19:49] yeah, well accidents happen, /me has had his fair share too === ogra is now known as Guest59853 === Guest59853 is now known as ogra [22:53] apw, have you guys thought about backporting the support for ATI Cayman cards (http://www.phoronix.com/scan.php?page=news_item&px=OTE2NA) [22:58] bryceh_: sigh, it will probably happen [22:59] I am looking for the source for the linux-image-generic-lts-backport-maverick kernel. After reading https://help.ubuntu.com/community/Kernel/Compile and https://wiki.ubuntu.com/KernelTeam/KernelGitGuide I am unable to locate which Git tree to clone. [23:00] I am trying diagnose a problem with the Mantis DVB driver and want to apply three patches from Linus' tree to test the issue. [23:00] Any help regarding the location of the sources of linux-image-generic-lts-backport-maverick will be very helpful. Thanks. [23:14] rishi: its in the lts-backport-maverick branch in the ubuntu-lucid repo [23:19] jjohansen: Ok. Ok. [23:20] I am just curious, what is the difference between http://kernel.ubuntu.com/~kernel-ppa/mainline/ and https://launchpad.net/~kernel-ppa/+archive/ppa ? Both seem to go by the name kernel-ppa, but have different content. === yofel_ is now known as yofel