/srv/irclogs.ubuntu.com/2011/03/24/#ubuntu-kernel.txt

kristian-aalborgafter building a (hopefully) working config, can I just "make menuconfig" and then Roosevelt stuff again?00:05
jjohansengenerally yes00:08
kristian-aalborgcool00:10
kristian-aalborgI guess that building kernels is a fairly good way to get the building process straight00:11
jjohansenhehe, yeah its a great way :)00:13
kristian-aalborgI might put a tutorial together... not on an expert level, obviously... but just how to get things running a bit better00:16
=== chuck_ is now known as zul
kristian-aalborgwill this building method work on any debian-derived distro?00:44
DrDetroitjjohansen: here is the result of my test01:02
DrDetroit2.6.32-28-generic #55-Ubuntu SMP Mon Jan 10 21:21:01 UTC 2011 i686 GNU/Linux01:02
DrDetroit1:40,  2 users,  load average: 0.42, 1.03, 1.2701:02
DrDetroitbut its getting pretty poor performance01:03
DrDetroitI thought 28 was good01:03
DrDetroitI sure don't know what to think or do anymore01:03
DrDetroiti guess i will go back to 27 i think i still have that on my menu01:06
DrDetroitDo you know how long folks have been reporting this issue?01:07
jjohansenDrDetroit: I had a user on the weekend reporting a similar issue but we didn't track it down01:20
jjohansenDrDetroit: I am not even sure it is the same problem, but the symptoms are similar01:21
DrDetroitjjohansen:  I am sort of unsure on how to proceed01:25
jjohansenDrDetroit: lets give -27 a good run and see if you see problems there01:26
DrDetroitjjohansen:  Ok I will do that01:26
jjohansenDrDetroit: from last night bisects I would say we weren't giving the problem enough time to surface01:27
DrDetroitI have been running that rotating torus screensaver01:27
DrDetroitI can tell when things slow down after awhile01:27
jjohansenDrDetroit: there is also the possibility the problem stems from something else that was updated at the same time as the kernel01:27
DrDetroitit ls almost like an old fashioned windows memory leak01:27
DrDetroitI have been running xchat, a terminal window, 3 instances olf mudlet (a mud client) and firefox01:28
jjohansenyes, except in the one I was chasing on the weekend there was plenty of free memory01:28
jjohansenwe should check that when you see the slow down01:28
DrDetroitok01:28
jjohansenDrDetroit: also we should try killing running processes one by one and see if that clears up the problem01:29
DrDetroitMem:    444440k total,   331720k used,   112720k free,    17080k buffers01:30
DrDetroitSwap:  1299448k total,        0k used,  1299448k free,   151360k cached01:30
DrDetroitthat is current with only xchat and a terminal window running01:30
jjohansenDrDetroit: so yeah good atm, and if we see the slow down check again01:30
DrDetroitjjohansen: its slow now01:31
jjohansenbefore we were just doing a bisect hunt for a kernel regression, this time we will be much more thorough01:32
DrDetroitjjohansen: with those stats01:32
jjohansenah01:32
jjohansenwhat does top report01:32
DrDetroitthat was top01:32
DrDetroit950 root      20   0  111m  39m 8692 S  5.8  9.1  25:24.79 Xorg               01:33
DrDetroit 1708 bobp      20   0  2548 1192  904 R  2.6  0.3   0:00.44 top                01:33
DrDetroit 1533 bobp      20   0 46296  12m 9724 S  1.3  3.0   0:32.65 gnome-terminal     01:33
DrDetroit  223 root      20   0     0    0    0 S  0.3  0.0   0:05.54 usb-storage        01:33
DrDetroit 1267 root      20   0  3612 1216 1036 S  0.3  0.3   0:05.28 hald-addon-stor    01:33
DrDetroit 1396 root      20   0  5180  968  688 S  0.3  0.2   0:08.67 udisks-daemon      01:33
DrDetroit 1581 bobp      20   0 49604  19m  12m S  0.3  4.5   1:35.80 xchat-gnome        01:33
DrDetroitthat is the maddening part01:33
jjohansenokay and the load figures01:33
DrDetroitnothing really shows up hogging cpu or memory01:33
jjohansenokay01:33
DrDetroitCpu(s):  7.2%us,  3.6%sy,  0.0%ni, 89.3%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st01:34
DrDetroitthats from top 01:34
DrDetroitits crappy poor at 90 percent idle01:34
DrDetroitjjohansen: let me proppse this01:34
DrDetroitI will reboot and go back to 27 but before i do that, i will double check my hardware to make sure it is all working properly including memory01:35
DrDetroiti think it is01:35
DrDetroitbut i can double check01:35
DrDetroitthen i will boot back to 27 and run that for a while01:35
DrDetroitI can report back later this evening or tomorrow early01:36
jjohansenwell yes that is worth doing, which kernel are you running /me thought you were on .2701:36
DrDetroitwhichever you prefer01:36
DrDetroitcurrently this is 28 which i thought was good01:36
jjohansenokay, well lets run a few commands first01:36
jjohansendo01:37
jjohansensar 1 10001:37
DrDetroitjjohansen:apparetnly its not installed01:39
DrDetroiti will install it hold on01:39
jjohansensudo apt-get install sysstat01:39
DrDetroitits running01:41
DrDetroitit looks like it is repeating the same stuff, is that what you wanted01:41
jjohansenyes, look at the iowait column01:42
jjohansenwhat kind of figures are you seeing01:42
DrDetroitAverage:        all     37.54      7.69     11.41      0.28      0.00     43.0801:43
DrDetroitthe are all 0.0001:43
DrDetroiti tanke that back i see one 7 a couple 6's and a couple 1's but the majority are 0.0001:44
jjohansenalright so we aren't seeing io backup causing the slow down01:44
jjohansennot a big deal, a few low numbers isn't going to be a problem01:44
DrDetroitthe average is .2801:44
DrDetroiti did run itop but i really didnt understand what it was telling me01:45
jjohansenif we where getting high numbers with high idle, it would indicate that processes are spending all their time blocked on io01:45
kristian-aalborgwill subsequent builds be significantly faster?01:45
jjohansenkristian-aalborg: they can be, it depends on what make determines needs to be rebuilt, based off Makefile dependencies and time stamps01:46
jjohansenso if you change a header file that gets included everywhere, everything needs to be rebuilt01:46
* kristian-aalborg screams!01:46
jjohansenchange a small section of a .c file and likely only 1 file gets rebuilt01:47
kristian-aalborgit *still* does not work :(01:47
jjohansenkristian-aalborg: are you doing make clean?01:47
kristian-aalborgstill goes straight to reboot01:47
kristian-aalborgno, fdr clean01:48
jjohansenah fdr clean will clean out the old build01:48
kristian-aalborgno no... I start out with fdr clean, then cp over the config file01:48
jjohansenif you want to do an incremental build with the build system do rm debian/stamps/stamps-build-XXXX01:49
jjohansenkristian-aalborg: right, 1 clean is needed, after which clean can be skipped if you want to do incremental builds01:49
kristian-aalborgI thought I was going to want that, but not before my kernel works :/01:50
jjohansenhowever depending on your config changes you may need to do a clean01:50
kristian-aalborgjjohansen: obviously, something is still very wrong with the .debs I produce01:51
jjohansenI am not sure how well the dependencies between the config file and the rest of the build system are tracked01:51
jjohansenkristian-aalborg: :(01:51
jjohansenDrDetroit: can you try running vmstat01:51
DrDetroitsure01:51
DrDetroitprocs -----------memory---------- ---swap-- -----io---- -system-- ----cpu----01:52
DrDetroit r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa01:52
DrDetroit 1  0      0  77948  27512 172400    0    0    31     6  318 1975 46 10 44  101:52
kristian-aalborgbuild system is Lubuntu 10.4, box to use kernel on is Ubuntu 10.4... should be the same01:52
jjohansenDrDetroit: what processes are running01:53
jjohansenkristian-aalborg: what problems are you seeing with your .debs?01:53
DrDetroitthree terminal windows, xchat01:53
DrDetroitand i just closed 1 terminal window01:53
kristian-aalborgthe debs look fine, I md5summed them too before install01:54
DrDetroitof course there is the stuff on my panel01:54
kristian-aalborgwell, one of them... 01:54
DrDetroittime date, sound etc01:54
jjohansenDrDetroit: try this log out of your X session and log back in, let see if reset a session, restarting the panels etc, has an effect01:55
kristian-aalborgjjohansen: it's what happens when I try to boot the kernel that I installed from those debs that things get weird :/01:55
DrDetroitok01:55
jjohansenkristian-aalborg: still rebooting?01:55
DrDetroitbrb01:56
kristian-aalborgjjohansen: yes01:56
jjohansenkristian-aalborg: and you are just changing the config, no other patches right?01:57
kristian-aalborgI did sudo dpkg -i linux-* this time, would be the same I reckon01:57
kristian-aalborgyes, I use the config that's in /boot on the old box01:57
jjohansenkristian-aalborg: can you send me the config, I will compare and try building a lucid kernel for you01:58
kristian-aalborghttp://pastebin.com/LUM4xbu9 this is what I get when I diff the two files01:59
jjohansenkristian-aalborg: diff your config and the lucid config?02:00
DrDetroitjjohansen: it is significently worse02:00
DrDetroitand my desktop has changed themes02:00
DrDetroithehe02:00
kristian-aalborgsorry for being unclear02:01
jjohansenDrDetroit: interesting, want to look at top and sar output again02:01
DrDetroitsure02:01
kristian-aalborgthis is the diff between the config from the working config and the non-working one - sent from the old machine02:01
DrDetroitjjohansen: here is top02:02
DrDetroitTasks: 140 total,   1 running, 139 sleeping,   0 stopped,   0 zombie02:03
DrDetroitCpu(s): 21.9%us,  9.4%sy, 24.1%ni, 43.8%id,  0.6%wa,  0.1%hi,  0.0%si,  0.0%st02:03
DrDetroitMem:    444440k total,   385744k used,    58696k free,    29016k buffers02:03
DrDetroitSwap:  1299448k total,        0k used,  1299448k free,   194280k cached02:03
jjohansenkristian-aalborg: thats it? one not having cgroups shouldn't make any difference02:03
jjohansenDrDetroit: and sar02:03
DrDetroitAverage:        all     29.36      0.00      7.80      0.44      0.00     62.4102:03
DrDetroitiowaits ave .4402:04
kristian-aalborgyes, it is weird02:04
jjohansenalright02:04
DrDetroittop was with sar running02:04
jjohansenkristian-aalborg: can you send me the working config02:05
kristian-aalborghttp://pastebin.com/XZjV4Fup02:05
kristian-aalborghere :)02:05
DrDetroitjjohansen: we can continue this tomorrow if you would like to work with others tonite02:06
jjohansenDrDetroit: its fine, we can keep going if you want02:06
kristian-aalborgI should get some sleep... 3AM here... 02:07
DrDetroithehe02:07
kristian-aalborgjjohansen: we could continue later, unless you're already building?02:07
jjohansenkristian-aalborg: yeah I kicked it off, but we can continue later it will take a bit to build02:08
kristian-aalborgdepends on what "a bit" is on your system - on mine, it's usually hours ;)02:09
jjohansenkristian-aalborg: say 20-30min02:09
kristian-aalborgI think the smart thing for me is going to bed then02:10
kristian-aalborg30 mins of building02:11
DrDetroitsleep....sleep....sleep02:11
* DrDetroit giggles02:11
jjohansenhehe02:11
kristian-aalborgthen spending either one hour trying to figure out what went wrong or the rest of the night tweaking ;)02:11
kristian-aalborgjjohansen: I'll catch up later, just remember not to delete the files ;)02:14
jjohansenkristian-aalborg: :)02:15
jjohansenDrDetroit: alright give the -27 kernel a spin02:15
kristian-aalborgsee ya, and thanks for helping out02:17
DrDetroitjjohansen: ok will do02:19
DrDetroitback later on02:19
DrDetroitwith a report02:19
DrDetroitjjohansen: ok on 27 will let it run for a few hours with hypertorus screensaver running02:28
DrDetroitI have my normal apps open and running also02:28
=== Theravadan|afk is now known as Theravadan
jjohansenDrDetroit: okay I'll check back in a few hours02:46
=== DrDetroit is now known as DrD_away
=== _LibertyZero is now known as LibertyZero
=== _LibertyZero is now known as LibertyZero
fairuzHi morning07:40
jjohansenmorning07:42
fairuzjjohansen: can i start with my questions? :D07:42
jjohansensure07:42
fairuzjjohansen: about /usr/src/linux-headers-xxxxxxx, I need one for each kernel i compile?07:43
fairuzjjohansen: because i see files in /lib/modules/xxxxxxxx/build point to the headers folder07:44
jjohansenfairuz: define need, they are good to have but the headers aren't "needed" to run the kernel07:44
fairuzjjohansen: yes, i mean for compiling external module07:45
fairuzjjohansen: for external module, i need all of those right?07:46
jjohansenyes, the headers are needed for compiling external modules07:46
fairuzjjohansen: and i suppose there are make install_headers to create this folder?07:47
fairuz:D07:47
jjohansenfairuz: the target is actually headers_install07:48
jjohansenfairuz: or maybe headers_install_all07:49
fairuzjjohansen: I think I already tried that, but no idea where the files go after it finishes07:49
jjohansenfor the work I've done I've never needed to do it, but yeah its there07:49
fairuzjjohansen: I compile the kernel with custom name and it boots ok07:53
fairuzjjohansen: just there is no display right now, which I think it's normal because I'm not installing the modules yet07:54
jjohansenfairuz: sounds about right07:54
fairuzjjohansen: i just get the plain console. I will try now to use sshfs to install the modules. I tried it yesterday but got some problems07:55
fairuzjjohansen: something about permissions07:56
jjohansenfairuz: hrmm, I think I might know what you hit, give me a minute to find it07:56
jjohansenfairuz: http://blog.sumostyle.net/robg/2006/08/07/some-useful-sshfs-options/08:05
fairuzjjohansen: ok i'll take a look at it08:05
jjohansenyou may need allow_root or allow_other, I can't remember which08:05
jjohansenI think I did the mount as my regular user and I had to use allow_root, to get the install to work08:05
fairuzjjohansen: i'll try that08:09
fairuzjjohansen: have to reboot with my working kernel first :D08:09
fairuzjjohansen: it's not a simple thing this whole kernel thing. To get it booted is one thing. To get it works with all the modules is other thing.08:11
jjohansenyeah it can be a pain, there are reasons to like monolithic kernels08:12
fairuzjjohansen: so you do sshfs without sudo ?08:16
jjohansenfairuz: that I can't remember, the remote install I was doing where a while ago08:17
jjohansenI know I did sudo make install ; sudo make modules_install08:17
fairuzjjohansen: ok. 08:18
jjohansenbut I think the fuse mount was probably done as a user08:18
jjohansenI know I did the compile as just a plain user08:18
fairuzjjohansen: it dont let me to use -o, I need to edit /etc/fuse.conf08:18
jjohansenfairuz: yeah, I think I had to do that too08:19
jjohansengetting it setup was a bit of a pain08:19
jjohansenbut it worked well once I had it08:19
fairuzjjohansen: one thing i find it weird, before it asking the remote password. It asks for my password 3 times. Each time.08:20
jjohansenhrmm, I don't remember it doing that, but its possible, /me needs to go back and try doing a remote install like this again some time08:21
fairuzjjohansen: it's installing the modules now :D08:52
fairuzjjohansen:  I use sudo sshfs -o allow_other,default_permissions <host> <mountpoint>08:53
jjohansenyeah that sounds like it should work08:54
=== DrD_away is now known as DrDetroit
DrDetroitand good morning to all08:54
fairuzjjohansen: i only have around 400Mb left on the SD, hope it's enough08:54
fairuzDrDetroit: morning08:54
jjohansenmorning DrDetroit08:55
DrDetroitjjohansen: I have run the 27 kernel since falling asleep last night08:55
DrDetroitit seems ok , have not seen any issues yet08:55
jjohansenhrmm okay08:55
DrDetroitlet me open up a few more apps and see what happens08:55
fairuzjjohansen: It finishes :D09:05
jjohansen:)09:05
fairuzjjohansen: but I think I should change the symlink for build and source09:06
fairuzjjohansen: right now it points to the mounpoint09:06
jjohansenhehe, yeah that might be an idea09:06
fairuzjjohansen: maybe I should change it directly to my source folder in the build machine?09:07
fairuzjjohansen: since it's accessible by network09:07
jjohansenfairuz: that should work09:08
fairuzjjohansen: i will be happy if this works :D09:08
fairuzjjohansen: hmm still no luck. I got FATAL: Could not load /lib/modules/2.6.35-980-omap4-fairuz/modules.dep: No such file or directory09:24
fairuzjjohansen: but the file is there!09:24
jjohansenfairuz: try running depmod -ae09:24
fairuzWARNING: -e needs -E or -F09:25
fairuzFATAL: Could not open /lib/modules/2.6.35-980-omap4-fairuz/modules.dep.temp for writing: Permission denied09:25
jjohansenfairuz: sudo depmod -a09:26
fairuzjjohansen: ok done. what it does?09:27
fairuzjjohansen: i reboot now?09:27
jjohansendepmod - program to generate modules.dep and map files.09:27
jjohansensure give it a try09:27
kristian-aalborgjjohansen: hi again09:32
jjohansenmorning kristian-aalborg09:32
DrDetroitmorning kristian-aalborg09:33
kristian-aalborgmorning09:33
kristian-aalborgwhat time is it where you people live?09:33
jjohansen02:3409:34
fairuz>:o09:34
DrDetroit4:3409:34
jjohansenkristian-aalborg: kernel.ubuntu.com/~jj/linux-image-2.6.32-31-generic_2.6.32-31.60~kristian_i386.deb 09:34
kristian-aalborgah, east and west coast I suppose09:35
kristian-aalborgjjohansen: thanks09:35
DrDetroitcentral here, I live n the Ozark mountains of Arkansas09:35
kristian-aalborgI have a load of wires all over the place atm, it will be a moment before I can download anything09:35
=== doko__ is now known as doko
fairuzjjohansen: still got FATAL: Could not load /lib/modules/2.6.35-980-omap4-fairuz/modules.dep: No such file or directory09:39
jjohansenfairuz: hrmm, is /lib mounted, at this point?09:40
fairuzjjohansen: I have no idea of this09:41
jjohansenfairuz: how is the filesystem layed out are you using an initramfs/initrd is /lib in the root filesystem09:42
fairuzjjohansen: I have no idea about initramfs, but now when it finishes botting, i can go in /lib yes09:47
jjohansenwhat does mount show09:47
fairuzjjohansen: oh there is not /lib09:47
fairuzproc on /proc type proc (rw)09:48
fairuznone on /proc/sys/fs/binfmt_misc type binfmt_misc (rw,noexec,nosuid,nodev)09:48
fairuznone on /sys type sysfs (rw,noexec,nosuid,nodev)09:48
fairuznone on /sys/kernel/debug type debugfs (rw)09:48
fairuznone on /sys/kernel/security type securityfs (rw)09:48
fairuznone on /dev type devtmpfs (rw,mode=0755)09:48
fairuznone on /dev/pts type devpts (rw,noexec,nosuid,gid=5,mode=0620)09:48
fairuznone on /dev/shm type tmpfs (rw,nosuid,nodev)09:48
fairuznone on /var/run type tmpfs (rw,nosuid,mode=0755)09:48
fairuznone on /var/lock type tmpfs (rw,noexec,nosuid,nodev)09:48
fairuzjjohansen: hmm ... but why with other kenel, it loads ok09:51
jjohansenfairuz: hrmm I not sure09:51
fairuzjjohansen: wait i go restart all machines and retry09:53
DrDetroitwell now, that was strange, my screensaver went off, but returned back to the regular screen09:59
fairuzjjohansen: is this something to do with initrd?10:01
fairuzjjohansen: at boot, i load uImage and also uInitrd10:01
jjohansenfairuz: yep, what files are in /boot/10:02
jjohansenDrDetroit: its not unheard of for the screen saver to have odd bugs10:02
DrDetroithehe, i know, its a computer, it's all smoke and magic to me10:02
jjohansenif you can replicate it is worth filing a bug over10:02
fairuzjjohansen: there are some vmlinuz files, system.map files, initrd.img-xxxxx10:03
DrDetroitnaw10:03
jjohansenfairuz: so you need to make an initrd10:03
fairuzjjohansen: there are files for my old kernel, no files for my new kernel10:03
fairuzjjohansen: ah ok.. any help to do that?10:04
jjohansenfairuz: no files at all or no initrd10:04
fairuzjjohansen: what this thing do? it's ubuntu thing or kernel thing?10:04
jjohansenfairuz: hrmmm well sort of both10:04
jjohansenthe kernel doesn't need one but one is often used for flexibility10:04
jjohansenubuntu uses one10:05
fairuzjjohansen: i have for example initrd.img-2.6.35-980-omap4 but no initrd.img-2.6.35-980-omap4-fairuz10:05
jjohansenfairuz: I am going to assume its an initramfs10:05
fairuzand also i have vmlinuz-2.6.35-980-omap4 but no vmlinuz-2.6.35-980-omap4-fairuz10:05
fairuzjjohansen: same thing applies for System.map and abi files10:06
jjohansensudo update-initramfs -ck 2.6.35-980-omap4-fairuz10:06
fairuzjjohansen: ok10:06
jjohansenfairuz: hrmm, do you do a make install?10:07
jjohansenfairuz: or did you just copy the vmlinuz10:07
fairuzjjohansen: nope, just make xxx_defconfig and make uImage10:07
fairuzjjohansen: then i just copy the uImage10:07
jjohansenfairuz: generally make install will copy those as well as the kernel10:07
jjohansenyou could use the same sshfs trick to do make install as make modules install10:08
jjohansenfairuz: also update-initramfs isn't the only way to make a ramdisk but it is the way ubuntu uses, I am not up to date on other methods10:10
fairuzjjohansen: i have now initrd.img for my kernel10:14
fairuzjjohansen: it still not working. i will try the make install now10:16
jjohansenfairuz: make install and then remake the ramdisk10:16
jjohansenfairuz: instead of -ck use -uk10:17
fairuzjjohansen: actually it generate the initrd but it gives this grep: /boot/config-2.6.35-980-omap4-fairuz: No such file or directory10:18
fairuzi should copy my config to /boot folder first?10:18
fairuzor make install will do that10:18
jjohansenhrmm, configs shouldn't be needed but system.map and abi I believe are10:19
jjohansenfairuz: but if its complaining about the config missing if make install doesn't put it there copy it yourself10:20
fairuzjjohansen: ok understood10:22
fairuzjjohansen: sudo install will not recompile the kernel i suppose?10:28
jjohansenfairuz: sudo make install should not recompile the kernel10:29
jjohansenwell at least as long as you haven't changed any timestamps10:29
fairuzjjohansen: if both machines have different time? :D10:31
fairuzjjohansen: because right now i'm seeing a lot of CC10:31
fairuzjjohansen: which i think it's in the process of the compiling10:31
jjohansenfairuz: its the files time stamps that should be important, I'm not sure why you would be seeing lots of CC10:31
jjohansenwhich is indeed compiling10:32
fairuzjjohansen: well i will let it compile then ( which will take forever since it's compiling on the target machine )10:32
fairuzjjohansen: So if I understand well, I should do make install that will install necessary files to /boot10:33
jjohansenfairuz: it should10:34
fairuzjjohansen: and this file is used during the boot i suppose?10:34
jjohansenfairuz: or for setting up the module dependencies, building the initramfs (which is used during boot)10:35
=== smb` is now known as smb
=== jjohansen is now known as jj-afk
* jj-afk is going to be off for a bit10:38
DrDetroitjj-afk: up  8:16,  2 users,  load average: 0.58, 0.79, 0.7410:38
fairuzjj-afk: ok thanks for your help10:38
DrDetroitjust for info purposes i think 27 is ok10:38
DrDetroitsee you later10:38
jj-afkDrDetroit: okay, 27 looks okay but not 28, I'll build a kernel for you to test, while I am a way10:39
DrDetroitthanks, see you later on today10:41
jj-afkDrDetroit: I am give me 20 min, and I'll have the kernel for you10:42
jj-afkDrDetroit: kernel.ubuntu.com/~jj/linux-image-2.6.32-28-generic_2.6.32-28.50~lp740549.13417a_i386.deb11:10
jj-afkthese ones are going to be slightly different in that I am tagging them with part of the commit sha11:10
jj-afkso if you don't want all the test kernels installed at the same time, then you will need to dpkg -r the package when done testing with it and moving on11:12
fairuzjj-afk: there?12:10
=== yofel_ is now known as yofel
DrDetroitjj=afk: so if i understand correctly, I should 1) download the package 2) install it the same way we did the others with dpkg -i <packagename>12:40
DrDetroitthen once I am done testing do a dpkg -r <packagename> before moving on to the next one12:41
DrDetroitjj-afk: so if i understand correctly, I should 1) download the package 2) install it the same way we did the others with dpkg -i <packagename>12:42
DrDetroitthen once I am done testing do a dpkg -r <packagename> before moving on to the next one12:42
=== sconklin-gone is now known as sconklin
DrDetroitjj-afk: downloaded and testing first kernel13:00
DrDetroitl13:28
tgardnerogasawara, are you planning to rebase natty master-next for 2.6.38.1 today?13:33
ogasawaratgardner: yep, was gonna do it right after I finish up this Maverick to Natty config delta review13:34
tgardnerogasawara, cool13:34
ppisatiogasawara: lp#737073 - why did you mark it invalid for lucid/maverick? the description says it affects kernels before 2.6.3713:45
ogasawarabug 73707313:45
ubot2Launchpad bug 737073 in linux "CVE-2010-4527" [Medium,Fix committed] https://launchpad.net/bugs/73707313:45
ogasawarappisati: I want to recall that it's because the patch is already in maverick/lucid via stable updates, but I'll have to double check13:46
tgardnerogasawara, if so, shouldn't it be 'fix released' ?13:47
ppisatiogasawara: ah ok13:47
ppisatippisati: i didn't check the tree before asking, ok13:47
ogasawaratgardner: yah, I suppose it could be marked Fix Released rather than Invalid13:48
tgardnerogasawara, that more closely matches the CVE tracker state, doesn't it?13:48
ogasawaratgardner: the CVE tracker state is marked as not-affected so I guess it just depends how we want to interpret it13:50
tgardnerogasawara, well, the security team has been dinging us for not noticing when stable updates fix a CVE, even though its not noted in the upstream commit log.13:51
ogasawaratgardner: ah, in that case Fix Released is probably a better status to use13:51
* ogasawara back in 20min14:03
ppisatibug 73639414:04
ubot2Launchpad bug 736394 in linux-source-2.6.15 "CVE-2010-4342" [Medium,In progress] https://launchpad.net/bugs/73639414:04
ppisatii nominated it for maverick14:05
ppisatican anyone pass the nomination?14:05
tgardnerppisati, LP keeps timing out14:10
smbtgardner, ppisati Did it (was closer)14:11
ogasawarahttps://launchpad.net/ubuntu/+spec/hardware-kernel-n-version-and-flavours14:47
ogasawara^^ anyone else get an error if they try to edit the whiteboard?14:47
tgardnerogasawara, I just changed your name in the Whiteboard14:48
ogasawaratgardner: cool, thanks14:48
tgardnerogasawara, maybe your LP privs are eroding.14:49
sconklintgardner: so the ABI problems with the packages were not the root cause of the FTBFS? (just trying to stay current)14:50
tgardnersconklin, I can't say for sure since LUM had the same issues, and is in the dependency chain.14:51
sconklinok14:51
tgardnersconklin, thats gotta get fixed first before I can tell if its the root of the HPPA FTBS. its just about gotta be that 'cause nothing else has changed.14:51
sconklintgardner: yeah, and it makes sense that it could be the cause14:52
tgardner sconklin: so, in order to avoid this ever happening again, I'm ripping out all generated files from dapper/hardy and subsidiary packages.14:53
sconklintgardner: great! Brad and I had the "OMG how many other packages could have this problem" conversation last night, but you got there first14:54
sconklinEven more scary is that potentially it could have happened and not resulted in a build failure14:54
tgardnersconklin, it likely is an indicator of how many folks aren't using Hardy LUM 'cause I don't think the modules would load, bot nobody has complained.14:55
smbtgardner, Was Hardy also still carrying the abi number asclear text or was it just re-creating those depending files?14:55
sconklinBrad and I have also talked about trying to get download stats for dapper and hardy, I'm curious how many consumers they have, and I think the stats are available now14:56
smbtgardner, I usually would notice when booting the test laptop. But it is more something like "hm, it is very silent on boot"14:56
tgardnersmb, Hardy LUM/LBM had debian/control committed in the repo, which makes it easy to _not_ correctly generate the ABI info.14:56
tgardnerHardy LRM as well14:57
fairuzhow to generate vmlinuz kernel file? I just have uImage, zimage and compressed/vmlinux. Thanks14:57
smbtgardner, Ah yes, I was unsure whether Hardy also still have the abi number in the rules file instead of using the number from the changelog14:58
smbThat is what caught me when doing dapper lrm14:59
tgardnersmb, dapper is worse. you _have_ to run the bumpabi target in order to get the udeb stuff henerated correctly.14:59
smbtgardner, Even worse that is the kernel package only. There is no bumpabi target in lrm15:00
tgardnersmb, right, but we only have to live with that for a couple of months more15:00
smbThat is the hope one can hold on. :)15:01
tgardnerogasawara, there is no '-' in your LP name. I see you've corrected it already in https://launchpad.net/ubuntu/+spec/hardware-kernel-n-version-and-flavours15:02
ogasawaratgardner: yep, no -15:03
=== sconklin is now known as sconklin-afk
bjfGrueMaster, have you signed off on the mvl-dove kernels now in -proposed ?15:51
GrueMasterI have tested the 216/416 kernels, yes.  I had noted that in the release tracking bug.15:52
GrueMasteror one of the bugs.15:52
bjfGrueMaster, i'm looking at what I think is the tracking bug, can you point me at the one you added a comment to? I'll also hunt for it.15:55
GrueMasterWell, looking at http://kernel.ubuntu.com/~kernel-ppa/reports/sru-report.html I see 4 proposed tracker bugs listed under linux-mvl-dove.  I'm not sure at this point as there were 3 kernels in proposed at one point.15:57
bjfGrueMaster, ok, thanks, we are working on the "tracking bug" issue15:58
GrueMasterThe only bug I cannot verify is bug 732700.15:58
ubot2Launchpad bug 732700 in linux-mvl-dove "apparmor_parser triggers a kernel panic" [Undecided,Fix committed] https://launchpad.net/bugs/73270015:58
DrDetroitl15:58
GrueMasterBut I believe ppisati was going to test it (since he filed it).15:58
ppisatiGrueMaster: did it, and it's ok15:59
ppisatiGrueMaster: i mean, it doesn't happer anymore15:59
ppisati*happen15:59
GrueMasterright.15:59
GrueMastercool15:59
GrueMasterbjf: I just added my test notes to tracking bug 73495016:00
ubot2Launchpad bug 734950 in linux-mvl-dove "linux: 2.6.32-31.60 / linux-ec2 2.6.32-315.28 -proposed tracker" [Undecided,Fix committed] https://launchpad.net/bugs/73495016:00
GrueMasterAlthough it doesn't list the specific kernel versions for dove.16:01
=== jj-afk is now known as jjohansne
=== jjohansne is now known as jjohansen
DrDetroitjjohansen: been running the 1st kernel for about 3hrs now, no problems yet16:07
DrDetroiti forgot how to remove my away status, hehe16:08
jjohansenDrDetroit: okay, I kick off the next one16:10
=== herton is now known as herton_lunch
=== diwic is now known as diwic_afk
tayali am willing to build a kernel.. can i do this in my maverick if it will have no harm?16:20
jjohansenDrDetroit: kernel.ubuntu.com/~jj/linux-image-2.6.32-28-generic_2.6.32-28.50~lp740549.4636a8_i386.deb16:22
jjohansentayal: you can build your own kernel if you want https://wiki.ubuntu.com/Kernel/BuildYourOwnKernel16:23
bjftayal, yes you can16:23
bjftayal, is there a reason you want to build your own kernel ?16:23
aakshayjjohansen: hey.. :) thanks.. 16:23
aakshaybjf,16:24
aakshaybjf: hi. :).. yes16:24
aakshayi am learning device driver design16:24
aakshaybjf,: so need to hav kernel builded to run my modules16:24
aakshaybjf: but i am confused exactly how to build the kernel? How to start? :(16:25
bjfaakshay, see the url that jjohansen just posted ^^16:25
tayalbjf: are these the only steps? 16:27
bjftayal, yes, that all you need to get a kernel built and installed16:28
tayalbjf: what is the meaning of Root of Kernel Source?16:28
bjftayal, the topmost directory in the kernel source tree16:29
tayalbjf: ok.. i am downloading the source. thanks for the help. :)16:29
bjftayal, np16:29
tayaljjohansen: thanks :)16:29
DrDetroitjjohansen: dl now, to remove the one I just tested it is dpkg -r linux-image-2.6.32-28-generic?16:30
jjohansenDrDetroit: dpkg -r linux-image-2.6.32-28-generic_2.6.32-28.50~lp740549.13417a_i386.deb16:32
jjohansenah, stupid paste16:32
DrDetroitjjohansen: dpkg: you must specify packages by their own names, not by quoting the names of the files they come in16:34
jjohansenDrDetroit: yes, sorry I was looking up what the exact package was that you installed and ended up pasting that in16:35
DrDetroitit removed it when i did  -r linux-image-2.6.32-28-generic?16:35
DrDetroiti am ok16:35
jjohansenyeah16:35
DrDetroitinstalling now16:35
DrDetroitok ready to test it for a couple hours16:35
DrDetroitback shortly16:35
DrDetroitand back16:47
DrDetroitjjohansen: will give this one a few hours at least 16:48
jjohansenDrDetroit: yep16:48
DrDetroitI guess that means I will have to go mow the yard16:53
* DrDetroit pouts16:53
=== herton_lunch is now known as herton
DrDetroitl17:10
=== sforshee is now known as sforshee-lunch
hertonogasawara: bug 735640 can be marked as fixed in changelog with latest 2.6.38.1 update in natty17:39
ubot2Launchpad bug 735640 in linux "Request support for Ortek PKB1700 wireless kb" [Undecided,In progress] https://launchpad.net/bugs/73564017:39
hertonand may be someone should nominate it for natty? I don't know17:39
ogasawaraherton: cool, I'll add it17:39
hertonalso bug 735450 should be fixed, but no reporter confirmed it the testing kernel I uploaded17:39
ubot2Launchpad bug 735450 in linux "BUG: unable to handle kernel NULL pointer dereference at 000001e4" [Undecided,Incomplete] https://launchpad.net/bugs/73545017:40
hertons/confirmed it/confirmed17:40
ogasawaraherton: yah, I suspect they filed it and then forgot about it.  I'll just add it so it gets closed.  They can always reopen it if necessary.17:41
hertonok, thanks17:42
=== sconklin-afk is now known as sconklin
cndogasawara, do you know if the canonical hwe or platform services guys read the ubuntu-kernel mailing list?17:52
cndI want to send something that will reach everyone, but specifically them17:53
=== sforshee-lunch is now known as sforshee
cndthe kernel-team mail list is pretty verbose now...17:53
ogasawaracnd: hrm, hard to say.  I'd say they probably don't follow it very closely.17:54
cndok17:54
cndI'll probably send two emails then17:55
cndI don't like to add private list CC's to a public email17:55
ogasawaracnd: makes sense17:55
* tgardner --> lunch18:16
vanhoofcnd: we read it :)18:21
vanhoofcnd: just fwd if you'd like18:21
cndvanhoof, ok18:22
sconklinjjohansen: what's the status of the performance regression that you were bisecting? Just curious . . .18:43
jjohansensconklin: haven't found it yet, turns out it show up in the entire range we bisected, the is -28 wasn't good18:44
jjohansensconklin: so we have expanded the range and each bisect is being tested for the 3-4 hours necessary for the problem to surface18:44
sconklinjjohansen: so the possibility exists that it really is from the scheduler updates18:44
jjohansensconklin: yeah18:45
sconklinjjohansen: could you please keep the bug updated and point to the bisection branch so we can tag-team this?18:45
jjohansensconklin: sure18:45
sconklinthanks. I'm curious to see what this one is and why it takes hours to show up18:46
jjohansensconklin: me too18:46
sconklinDrDetroit: is it possible for you to provide a very detailed set of steps for how to reproduce the issue and tell when it happens? Ideally we could get it added as a unit test for QA18:47
bjfsconklin, jjohansen, DrDetroit, it would be nice if we could duplicate the problem as well18:51
jjohansenbjf: yep, I have started trying to do that, on actual hardware vs vm18:53
* jjohansen -> lunch19:06
DrDetroitsconklin bjf jjohansen: sorry i was outside mowin. I am not sure how to tell you to duplicate it19:11
DrDetroitI just run my normal stuff and after awhile the machine just turns to mush, but nothing really shows up in top or ps aux19:11
DrDetroitit might be 20 min, it might be several hours19:12
DrDetroitbut I don't mind doing this, since I have plenty of extra time on my hands19:13
sconklinDrDetroit: one more tool, it may not show anything, but you can try latencytop19:15
DrDetroitlet me try that19:16
DrDetroitinstalling19:16
jjohansenDrDetroit: right, but we can't be sure whether a bisect is good for several hours, its easy when its bad19:17
jjohansensconklin: yeah thats another good one to use, I have been asking for vmstat, sar, top, and a few other things19:17
DrDetroitok how do i read this19:20
DrDetroitlet me copy the global here so you guys can see it19:22
DrDetroit'fudge it wont let me copy19:22
DrDetroitwell scheduler seems to have the most pecentage19:22
DrDetroit36-40%19:23
DrDetroitwhen i have it set on global one of the items i see is jbd2/sdb1-8, I never heard of that19:26
DrDetroitits at 46%19:26
skaetDFo, ogasawara - based on scans so far (it changes every minute after all..  :P ) only new high/critical bug I've found to add to kernel's list this week is Bug:63448719:26
skaetthe searching/scanning will continue though this afternoon.... ;)19:27
skaetrest of the bugs are already on the list from last week.19:27
ogasawaraskaet: cool, thanks19:30
ogasawarabug 63448719:30
ubot2Launchpad bug 634487 in linux "t1.micro instance hangs when installing java" [High,Confirmed] https://launchpad.net/bugs/63448719:30
DrDetroitsconklin: latencytop installed and running19:31
DrDetroitwhat should I look for?19:31
sconklinheh, it's not helpful to say "anything that looks unusual", is it.19:33
DrDetroitwell19:33
DrDetroitscheduler stays above 35%19:33
sconklinJust see if anything jumps to the top when the system starts going south19:33
DrDetroitok19:34
sconklinAlso look for any of the latency numbers in the "maximum" column which get really huge19:35
DrDetroitoh19:36
DrDetroithow big is huge?19:37
DrDetroitkswapd0 is 9719:38
tgardnerbjf, wtf does a version look like as an option to create-stable-tracker ?20:22
bjftgardner, it's just a string which will be used in the title, and description so for example: 2.6.32.35+drm33.1520:24
tgardnerbjf, so in theory it could be anything?20:24
bjftgardner, yes, i don't check it for any kind of "correct-ness"20:25
tgardnerhmm, it keeps blowing up. create-stable-tracker --version 2.6.32.35:20:25
tgardnerversion: 2.6.32.3520:25
tgardnerTraceback (most recent call last):20:25
tgardner  File "/usr3/ubuntu/kteam-tools/stable/create-stable-tracker", line 228, in <module>20:25
tgardner    app.main()20:25
tgardner  File "/usr3/ubuntu/kteam-tools/stable/create-stable-tracker", line 144, in main20:25
tgardner    raise AppError("The version (%s) is not a recognized version string." % (self.cfg['version']))20:25
tgardner__main__.AppError20:25
JFothanks skaet 20:25
tgardnerr20:25
JFoaparently I have been running around all day set to away :-/20:26
bjftgardner, --version=2.6.32.35  though i expect it to behave better than that20:26
tgardnerbjf, that was the next thing I tried: create-stable-tracker --version=2.6.32.3520:27
tgardnerversion: 2.6.32.3520:27
tgardnerTraceback (most recent call last):20:27
tgardner  File "/usr3/ubuntu/kteam-tools/stable/create-stable-tracker", line 228, in <module>20:27
tgardner    app.main()20:27
tgardner  File "/usr3/ubuntu/kteam-tools/stable/create-stable-tracker", line 144, in main20:27
tgardner    raise AppError("The version (%s) is not a recognized version string." % (self.cfg['version']))20:27
tgardner__main__.AppError20:27
tgardnerbjf, these error messages lead me to believe that there  is a required format.20:28
bjftgardner, looking20:28
bjftgardner, interesting, it checks against the debian version, one sec20:29
bjftgardner, i'm debugging it20:32
ogasawaraskaet: hrmph, it looks like the gcc armel build was restarted for some reason which is going to delay our no-change kernel upload from today until tomorrow.  The kernel then won't finish being built until Saturday now.20:38
bjftgardner, pull it and see if that works better, it's just a string now20:44
DrDetroitjjohansen: Been running this one for just over 4hrs, downloaded a ubuntu server instal cd, did some browsing etc, it seems ok for now20:45
jjohansenDrDetroit: okay, I'll kick off the next one20:45
tgardnerbjf, seems to have worked.20:47
bjftgardner, double check the bug for content20:47
tgardnerbjf, bug #74205620:48
ubot2Launchpad bug 742056 in linux "Lucid update to v2.6.32.35.15 stable release" [Undecided,New] https://launchpad.net/bugs/74205620:48
bjftgardner, looking20:48
bjftgardner, did you have to do anything or did the script do all that for you ?20:48
tgardnerbjf, so, you don't target any other kernels? we should do this for every variant?20:48
tgardnerbjf, the script did it all20:49
bjftgardner, you mean fsl-imx51 and mvl-dove and ec2 for Lucid ?20:49
tgardnerbjf, yes20:49
bjftgardner, ok, sconklin and i were just going over our todo list and i'll get that on it20:50
* ogasawara back on later20:51
jjohansenDrDetroit: kernel.ubuntu.com/~jj/linux-headers-2.6.32-28-generic_2.6.32-28.50~lp740549.8d6842_i386.deb21:04
DrDetroitjjohansen: on my way21:05
DrDetroitjjohansen: that didnt take 1 sec to download21:05
DrDetroitis that coorect>?21:06
jjohansenhrmm let me check21:06
DrDetroitthat one says linux headers21:06
jjohansenDrDetroit: ha no, I gave you the header name not the image21:07
DrDetroitok21:07
tgardnerbjf, pushed lucid Linux 2.6.32.35, seems to build OK. did you empty out pre-proposed so that it'll build in the morning?21:07
DrDetroitlet me delete that one then21:07
jjohansenkernel.ubuntu.com/~jj/linux-image-2.6.32-28-generic_2.6.32-28.50~lp740549.8d6842_i386.deb21:07
bjftgardner, no21:07
bjftgardner, i can handle that if you want21:08
tgardnerbjf, um, not sure what you mean. I just pushed the stable updates to Lucid master-next.21:08
tgardneroh, you mean about emptying the PPA21:09
bjfyes21:09
tgardneryeah, it needs doing.21:09
tgardnergo ahead21:09
bjftgardner, i've deleted the email can you forward21:09
bjftgardner, and i'll take if from there21:09
tgardnerbjf, so have I. just delete all packages that have been superseded. I'll check back in a few hours.21:10
bjftgardner, what's the url to pre-proposed ?21:10
tgardnerhttps://launchpad.net/~kernel-ppa/+archive/pre-proposed21:11
bjftgardner, thanks21:11
tgardnerbjf: 6.8 GiB (84.65%) of 8.0 GiB21:12
tgardnerI'm outta here for the day.21:12
sbeattiebjf: ask on #soyuz on i.c.c to make that ppa larger, if you need.21:13
bjfsbeattie, thanks21:13
DrDetroitback in a few21:15
=== sconklin is now known as sconklin-gone
DrDetroitand back21:21
DrDetroitjjohansen: do you know about how many we will have to do this round?21:26
jjohansenDrDetroit: I think it was 4 more21:26
DrDetroitok cool, just curious21:27
jjohansenDrDetroit: yep 421:30
DrDetroitno problemo21:34
DrDetroitjust trying to understand ubuntu-server 10.04 LTS while I am waiting21:35
skaetogasawara, arm is a mess right now from the builds, so somehow not feeling surprised to hear it.  :P   thanks for the head's up.21:39
kristian-aalborghey jjohansen 23:35
jjohansenhi kristian-aalborg23:35
kristian-aalborgthanks a lot for building that kernel, I've not tested yet though23:36
kristian-aalborgI have a kernel-burnout today ;)23:36
kristian-aalborghow did compiling it go? anything fishy?23:36
jjohansenkristian-aalborg: np, hrmm I don't remember anything fishy I should try it again to be sure23:37
jjohansen:)23:37
kristian-aalborgnah, we'll know when I try to boot it23:37
kristian-aalborgbut, only one file?23:37
jjohansenkristian-aalborg: what were you looking for?  There are a few other pacakges related to it like headers, but they aren't needed to test if the kernel boots23:38
kristian-aalborgI thought at least two files were needed, image + headers23:39
DrDetroitjjohansen: I am going to take a break for a bit, I should be back before you log off for the night23:42
=== DrDetroit is now known as DrD_away
jjohansenDrD_away: that shouldn't be a problem, as I don't usually log off until after you wake up :)23:43
DrD_awayhaha23:43
DrD_awaythats cause i am old23:43
* DrD_away chuckles23:43
jjohansenkristian-aalborg: no the headers aren't needed, at least for straight boot testing, now if you want to build certain things like external modules, or actually many applications they are needed23:44
jjohansenDrD_away: nah, its because you are sane and value a good nights sleep23:44
kristian-aalborgjjohansen: ah, okay23:45
kristian-aalborgcan I have the link again?23:45
kristian_jjohansen, sorry, box died23:56
jjohansenkristian_: kernel.ubuntu.com/~jj/linux-image-2.6.32-31-generic_2.6.32-31.60~kristian_i386.deb23:57
kristian_thanks23:58

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!