/srv/irclogs.ubuntu.com/2011/06/03/#ubuntu-kernel.txt

ppisatimorning05:35
jjohansenmorning ppisati05:54
ppisatihey06:01
=== _LibertyZero is now known as LibertyZero
ppisaticooloney: ping07:42
cooloneyppisati: hey, man07:42
ppisaticooloney: you have an imx51, right? care to see if a kernel boots?07:43
cooloneyppisati: oh, sorry, i don't07:43
cooloneyppisati: ericm|ubuntu has one, he can helep07:43
ppisaticooloney: ah, perhaps was eric then...07:43
ppisatiericm|ubuntu: ping07:43
ericm|ubuntuppisati, pong07:43
cooloneycooloney: yeah07:43
ppisaticooloney: ok, 5min, it's still compiling07:44
* ppisati gets a barrel of coffee...07:44
ppisatiericm|ubuntu: cooloney: btw, did you go to computex?07:47
ericm|ubuntuppisati, nope07:47
cooloneyppisati: no chance for us07:47
ppisatitoo bad07:47
ppisatiis it so difficult to go there?07:47
cooloneyppisati: heh, yeah, it's not easy for us to visit TW07:47
ppisatii tought you could get a visa, somehow07:47
cooloneyppisati: yeah, visa issu07:48
cooloneyissue07:48
ppisatiouch07:48
ppisatiwhat a pity07:48
ppisatiericm|ubuntu: http://people.canonical.com/~ppisati/imx51/linux-image-2.6.31-608-imx51_2.6.31-608.26_armel.deb07:58
ppisatiericm|ubuntu: give it a boot test07:58
ericm|ubuntuppisati, could you make a uImage/uInitrd?07:58
ppisatiericm|ubuntu: i'm cross compiling07:59
ppisatiericm|ubuntu: why do you need a uImage?07:59
ericm|ubuntuppisati, cuz that's easier for me to test07:59
ericm|ubuntuppisati, I'm downloading - I'll let you know if I need a uImage08:00
* apw yawns08:05
cooloneyapw: morning, man08:09
ericm|ubuntuppisati, installing ....08:13
ericm|ubuntuppisati, I need uImage08:24
ericm|ubuntuppisati, you just want to test boot or?08:24
ericm|ubuntuppisati, kernel boot or boot into a full system?08:24
ppisatiwell, boot it and see if the system work08:28
ericm|ubuntuppisati, so you want a boot into full system08:28
ericm|ubuntuppisati, that means I need to setup ubuntu08:29
ericm|ubuntuppisati, it's almost bare metal here - with u-boot08:29
ericm|ubuntuppisati, it's going to take time to install a full system here :-/, is it lucid/maverick/natty/onerick?08:30
ppisatiericm|ubuntu: ah, i didn't know08:32
ppisatiericm|ubuntu: no ok, give it a boot them08:32
ppisatithen08:32
ericm|ubuntuppisati, wait08:34
jjohansenapw: have you played with https://github.com/termie/git-bzr-ng?  its quite interesting08:35
apwjjohansen, no i've not, sounds interesting thought08:37
apwjjohansen, any idea if its packaged 08:37
jjohansenapw: not that I found08:37
* jjohansen is playing with it to do some bzr to git mirroring atm08:38
jjohansenhaven't gotten as far as trying to do dev with it08:38
apwi see we have the opposite packaged08:38
jjohansenyeah08:38
apwit seems to be a single file, did you drop it into your git lib directory for testing ?08:39
ericm|ubuntuppisati, http://paste.ubuntu.com/617335/08:40
ericm|ubuntuppisati, seems to be an old known issue - the parport08:40
ericm|ubuntuppisati, ain't this been fixed already?08:40
ppisatiericm|ubuntu: dunno, never had an imx51 before08:41
apwyeah that rings a bell, something like we had that parport module disabled on the imx51 configuration08:41
ppisatibut i didn't enable it08:41
apwericm|ubuntu, i think you can disable it via a bad argument right, parport.broken=108:43
ppisatiCONFIG_PARPORT_PC=m08:44
apwlag, was it you who originally hit the parport loading blew up ARM ?08:45
ppisatii guess it has always been there, and people blakclisted it08:45
apwericm|ubuntu, we may have been lucky before, life is like that sometimes08:45
jjohansenapw: yeah I dropped it in /usr/lib/git-core/08:47
cooloneyapw: yeah, i remember lag send an patch to upstream about parport08:47
cooloneyapw: but RMK has some very old ARM machine with parport08:47
* apw struggles to find the commit somewhere08:47
apwi am sure we have something applied to do ti08:47
apwsomehwere08:48
ppisatiactually the previous kernel was tested by GrueMaster, but i don't think he's around now08:48
ppisatiperhaps he disabled it on boot08:49
apwppisati, maybe so, seems unfortuanate though08:49
apwmanybe the regualr install does it08:49
ppisaticould be08:49
ppisatii pull req08:49
apw<apw> ericm|ubuntu, i think you can disable it via a bad argument right, parport.broken=108:52
=== chrisccoulson_ is now known as chr1sccoulson
apwseems it was CUPS starting which triggers this08:55
ppisatiapw: so it has always been there'08:56
apwppisati, still trying to work out what we did08:56
apwi remember there was a patch, which was NAKd upstream, and then we did something08:56
apweit08:56
apweither applied the patch locally or just changed the config to turn it off on our build08:57
apwon one of the 5 arm branches08:57
apwcommit abbbcd0c482d0fb3b2d39501f4debc956bea250608:57
apwAuthor: Lee Jones <lee.jones@canonical.com>08:57
apwDate:   Fri Jul 16 09:37:31 2010 +010008:57
apw    UBUNTU: Stop ARM boards crashing when CUPS is loaded08:57
apwwe have that applied on the _maverick_ master branch08:58
apwso i think this was an omap3 fix, not one which was o08:58
apwon fsl-imx51, so i am a little confused as to how its ever worked if its breaking now08:58
apwppisati, but the effect is just to turn off PARPORT_PC on omap, so you could try just turning PARPORT_PC off and see if that fixes things08:59
apwfor eric ... if so i can't see why we wouldn't have that switched over09:00
ppisatiuhm09:00
ppisatiok09:00
ppisatii'll do09:00
apwand no i cannot see why we haven't seen this before at all09:00
apwit could be a consequence of tightening something up in one of the security fixes though09:01
ppisatibut the crash is happening in drivers/parport/parport_pc.c, and that has always been =m09:01
ppisatiand the cve didn't touch that part09:02
ppisatiwell, anyway, it's not a big deal09:02
apwno did any affect io remapping or anything, i agree it is somewhat unexpected09:02
ppisatii'll turn off the partport and ask ericm|ubuntu to test another one09:02
apwsounds good09:02
ppisatibtw, besides ericm|ubuntu and GrueMaster, who has an imx51 board?09:02
apwi don't think i know the answer, ask ogra they have a list somewhere09:03
ppisatiok09:03
* apw suspects we should get you one09:03
ppisatiapw: it's scarcer than an albin tiger09:04
ppisatii already aksed around, and besides them it seems no one has one09:05
apwhrm09:06
ppisatiapw: what do i have to change to rename the kernel? i.e. i would like to add a postfix "parport-off"09:07
apwif you edit the debian.imx51/changelog, and add it in the version field09:07
apwso if its UNRELEASED you can add like ~parportoff1 to the version number09:08
apwnote that a - is not permitted09:08
apwthe ~ makes the version "older" than the offiicial version we release later09:09
ppisatiyou mean i change the numbering in parentheses, right?09:09
* ericm|ubuntu network sucks here much09:10
apwppisati, yep add it at the end there,  2.6.31-200.10 -> 2.6.31-200.10~parportoff109:11
ericm|ubuntuapw, let me try the command line09:11
ppisatibuilding...09:12
ericm|ubuntuapw, parport: Unknown parameter `broken'09:13
ericm|ubuntuapw, but interestingly - the oops didn't come09:14
apwericm|ubuntu, yep, thats whats meant to happen, it breaks the module init so it doesn't load09:14
ericm|ubuntuapw, clever :-)09:14
apwericm|ubuntu, now you can tell if everything else is working09:14
ericm|ubuntuapw, it hanged - I'll boot again09:14
apwericm|ubuntu, not sounding so good09:15
ericm|ubuntuapw, hanged again - stopped somewhere09:18
apwppisati, ^^ doesn't sound good09:18
ericm|ubuntuapw, ppisati, I'm using Linaro's natty ubuntu image - it could be rootfs issue09:18
ericm|ubuntuapw, ppisati, after a long long long hang - now it dropped me into the shell09:19
ericm|ubuntuSkipping profile in /etc/apparmor.d/disable: usr.bin.firefox [ok]09:19
apwericm|ubuntu, can you test the kernel which is in the -updates pocket on that userspace to see if thats a new issue or 'expected'09:19
ericm|ubuntusystem hanged for a long time before the above showed09:19
ericm|ubuntuapw, ok09:19
ppisatiyep, try the shipped kernel09:19
lagapw: ericm|ubuntu: ppisati: http://paste.ubuntu.com/617351/09:19
lagIt was CUPS, trying to speak to ALL of the serial ports09:20
lagNaturally, you need to change it to ARCH_<YOUR_ARCH>09:20
ericm|ubuntuapw, which release? lucid/maverick/natty/onerick?09:21
ppisatiericm|ubuntu: lucid09:21
ericm|ubuntuapw, it's in ports.ubuntu.com already? or somewhere else? the -update one09:23
apwericm|ubuntu, yeah whatever the latest official one is, as that is tested already09:27
ppisatiericm|ubuntu: http://people.canonical.com/~ppisati/imx51/linux-image-2.6.31-608-imx51_2.6.31-608.26~parportoff_armel.deb09:35
ppisatiericm|ubuntu: that's the parport off version09:36
* ericm|ubuntu finally get on again09:36
ericm|ubuntuppisati, downloading09:36
ppisatiericm|ubuntu: but i think it's better if you try first the -update version09:37
apwppisati, reviewed and pulled09:37
ppisatiericm|ubuntu: since it's pretty obvious turning off parport will fix it09:37
ppisatioooook09:37
ppisati:)09:37
ericm|ubuntuppisati, 'k - save me a few boots09:37
apwppisati, now lets hope there isn't a problem for you to find in the heap :)09:37
ppisatiapw: that would be fun :)09:38
apwfun for ericm|ubuntu doing the boots for the bisect :)09:38
ppisatiyep :)09:38
ericm|ubuntuapw, ppisati, I'll pretend to be having network issue then09:38
ericm|ubuntu:-)09:39
apwericm|ubuntu, :)09:39
apwppisati, is it mvl-dove which is still rebasable ?09:44
ericm|ubuntuapw, same - 608.25 paused a long time before dropping me into the shell09:44
ericm|ubuntuapw, so it's not a regression09:44
ericm|ubuntuppisati, I'll test your kernel next09:44
apwericm|ubuntu, ok cool, so ppisati we are likely good with whats on the branch, phew09:45
ppisatiericm|ubuntu: but did you get the parport trace?09:45
ppisatiapw: yes, i'll wait till next week and the do a rebase09:45
ericm|ubuntuppisati, yes - same09:45
apwppisati, and i think mvl-dove is programatically copied over into maverick from lucid right09:46
ppisatiapw: yep, exactly09:46
apwexcellent, so those two should look better soon too ... great09:46
ppisatithe other branch that need work is maverick/ti-omap409:46
apwyeah she is the worst looking at the moment09:47
ppisatibut first i've this:09:48
ppisatihttps://bugs.launchpad.net/ubuntu/+source/linux/+bug/79155209:48
ubot2Ubuntu bug 791552 in linux "No USB support on beagle/beagleXM" [Undecided,New]09:48
apwppisati, but you have a mountain to climb here, and the rock is heavy .... progress is good, we've applied more cves to the arm branches in the last week than in the last two months09:48
ppisatiwe totally lost usb on beagle*09:48
apwppisati, that sounds like a more interesting problem indeed09:49
ppisatialpha1 image has that problem09:49
apwppisati, if it was me, i'd be doing like a CVE (maybe 2) for each branch every day and doing interesting stuff the rest of the time09:49
apwpushing out the accumulation weekly09:49
ppisatibut when i tried an oneiric kernel on a maverick userland it worked09:50
apw(not that i am complaining to have 20 in a week or anything)09:50
apwinteresting ...09:50
apwppisati, i assume you have a beagle, so you can bisect using the kernels in launchpad09:50
ppisatiactually version_signature are the same09:51
ppisatii suspect it's something else09:51
apwppisati, huh ...09:51
ppisatibut yes, i'have 2 beagles09:51
apwthe version in oneiric is the _same_ version number?  then its a literal copy from natty to there09:51
apwand if thats the case it may be the act of reinstalling which fixed you09:52
ogra_ppisati, ggez, why do you care, i promise you there will never be an imx51 board with a parallel port, just disable parport and be done :)09:52
apwogra_, well we do wonder why he is suddenly hitting it, yes disabling it is sane anyhow, but why did this kernel suddenly exhibit the issue!09:53
ppisatiogra_: i'll push a config change then09:53
ppisatiapw: actually it's the omap3 kernel, so get it from mainline09:53
ppisatiapw: and the version on the alpha1 image and my own compiled kernel are the same09:54
ppisatii'll try some kernel/bootloader swapping09:54
apwppisati, i'd accept a config change for sure, its clearly not needed on the boards we care about there not being one09:54
ogra_wht do yopu hit exactly ?  bug 603062 or bug 601226 ? 09:54
ubot2Launchpad bug 603062 in linux-ti-omap4 "oops in parport_pc_probe_port function of parport_pc module (dup-of: 601226)" [Medium,New] https://launchpad.net/bugs/60306209:55
ubot2Launchpad bug 601226 in linux-ti-omap4 "Unable to handle kernel NULL pointer dereference in ppdev module" [Medium,Fix released] https://launchpad.net/bugs/60122609:55
ppisatido i disable parport_pc or the entire parport driver?09:55
ogra_i'm üpretty sure we disabled at least the loading of the module in the past 09:55
apwogra_, probabally an image issue then, as in your image has blacklist and the one ericm|ubuntu is using does not09:55
ogra_iirc there is a cups udev rule or the upstart job that have a modprobe parport hardcoded 09:55
apwif it blows chunks on loading, it likely should be removed09:55
ogra_one of the bugs above should have info for it 09:56
apwppisati, i think its just the PC part which is bust09:56
ogra_but fixing it is really not worth the effort given the usefullness of the fix (beyond the fact of having it fixed there is no other advantage anywhere)09:57
apwogra_, who in your team has one of these boards ?09:58
ogra_oh, and i have an imx51 board but that sits on the shelf and preparing it would take me half the day 09:58
ogra_tobin too 09:58
ogra_i think the rest went back to london 09:58
apwogra_, did those all become buildds or is there spares there we could get to ppisati 09:59
ogra_hmm, i can ask around09:59
apwogra_, thanks09:59
ogra_the panda buidcluster should be up within the next two weeks or so09:59
ogra_then we have all buildds spare 09:59
ogra_so you can definitely get one i think10:00
apwogra_, who do i poke to make sure ppisati is at the top of the list to receieve one10:00
ogra_i will take care, no worries10:01
=== ericm_ is now known as ericm|ubuntu
ericm|ubuntuppisati, apw, confirmed the ~parportoff kernel works10:01
apwericm|ubuntu, nice thanks10:02
ericm|ubuntuapw, np10:03
ppisatihttps://bugs.launchpad.net/ubuntu/+source/linux-fsl-imx51/+bug/60122610:07
ubot2Ubuntu bug 601226 in linux-ti-omap4 "Unable to handle kernel NULL pointer dereference in ppdev module" [Medium,Fix released]10:07
ppisatiapw: i nominated it for lucid/fsl-imx51, ack it10:08
apwppisati, done10:15
jussiSo, just wanted to follow up on a bug from a colleague in the office here, had a similar issue to me, on very different hardware. bug 79229112:00
ubot2Launchpad bug 792291 in linux "Machine hung, display frozen" [Undecided,New] https://launchpad.net/bugs/79229112:00
jussi(my bug was bug 791064 - although it maybe quite different, Im not sure).12:01
ubot2Launchpad bug 791064 in linux "Machine Hang, black screen with immobile mouse cursor only" [Undecided,Confirmed] https://launchpad.net/bugs/79106412:01
loolSince I upgraded to .39, I get regular wifi disconnects/reconnects; is this known?12:45
loolIntel(R) WiMAX/WiFi Link 5350 AGN, REV=0x2412:45
* apw lunches12:57
sconklinapw: there's a failed build of lts-backport-natty for Lucid in the PPA, what nneds to be done with that? and - does it need to go into this cycle?14:42
apwsconklin, is that on powerpc ? 14:55
sconklinlooking14:55
apwsconklin, if so it is not meant to build on there, and that is has is an artifiact of the fact its a new package i beleive, and nothing to worry about, it should be ignored14:56
sconklinlooks like armel, PPC, ia64 and sparc failed14:56
sconklinonly i386 and amd64 succeeded14:56
apwyeah, its only meant to build on i386 and amd64, its an lts backport kernel14:57
apwand as far as we know the packaging only says build there, but its thought to be a side effect of not being in the main archive already14:57
apwso just ignore those failures for now14:57
apwand only copy the amd64/i386 over14:58
sconklinok. Did you add a tracking bug?14:58
sconklinlazyweb14:58
apwno as originally it wasn't aimed to be a copy in in your team, it was for MIR, but then they decided we didn't need one14:58
apwso its ended up half baked, you could rebuild it if thats easier14:58
sconklinok, we'll manually track it this time and catch it in th enext cycle14:59
sconklinthanks!14:59
apwyeah my stupidity for not putting the MIR bug number in there, then you could have bastardised that for the job15:03
apwogasawara, do you know if we have oneiric capable PPAs yet ?15:16
ogasawaraapw: not that I'm aware15:17
apwhrm, it would be nice to test these changes in PPA first ... bah15:17
ogasawaraapw: indeed it would15:17
ogasawaraapw: also just fyi, I pushed linux-meta version fix to a master-next branch15:18
apwogasawara, sweet15:19
apwi am hopeful of getting this poo reviewed very early next week, if not then i'll plan on uploading just a simply tweak for versions to m-i-t moday15:19
apwmonday and get the resync in separatly15:19
ogasawaraack15:20
apwogasawara, i guess i could upload these both to a PPA but for natty and see happens15:23
apwogasawara, also did you say that you'd installed the m-i-t and kernel on a machine it booted ?15:23
apwpwd15:24
ogasawaraapw: yep, installed mit and the test kernel and all booted fine15:24
GrueMasterMorning kernel peoples.  15:46
GrueMasterppetraki: The bug on the parport-pc was mainly filed on dove, and iirc, the module was configed out.15:47
GrueMasterOops. I meant ppisati.15:47
vmlinuzTommeh: I've found a doc about i915 GPU lockups: https://wiki.ubuntu.com/X/Bugs/Lucidi8xxFreezes15:58
vmlinuzTommeh: I tried re-enabling KMS in driver options, tried disabling DRI in xorg.conf - None of them worked15:59
vmlinuzTommeh: seems this is a known issue already been tracked by upstream16:00
sconklinapw: in there a natty ports meta?16:07
sconklins/in/is/16:07
apwsconklin, nope there is no ports meta in natty16:07
sconklinthx16:07
Tommehvmlinuz: bad times for you, but at least it's known.16:32
TommehYou might wish to try xorg-edgers for the latest versions.. I'd say it's not for the feint of heart, but it's been relatively stable for me. The only thing I will say is that make sure you're VERY comfortable with resolving dependency issues with aptitude, before installing it (as removing it  can be tricky)16:33
TommehI say 'latest' .. I'm fairly sure even Lucid/Maverick don't get the absolute latest changes from 'edgers, due to broken ABI stuff.16:33
TommehMost likely newer than what you have, however.16:34
apwthis is utterly hopeless, the upstream bzr branches we are merging also have the .pc stuff in them so there is no clean way to actually merge at all16:52
=== jon__ is now known as jonpry
* apw becomes vague ...17:47
ogasawarabjf: I've got a script that dumps the following for our weekly meetings - http://reports.qa.ubuntu.com/reports/ogasawara/kt-meeting.txt18:40
bjfogasawara, looks good18:40
ronin__hi, apologize in advance 18:41
ronin__I want change my gcc version from 4.5 to 3.4 18:42
ogasawarabjf: I figure I'll just run it weekly in a cron. I wonder if I should copy it to kernel.ubuntu.com/~kernel-ppa/reports/ so then anyone could pull the data and paste it in the meeting should one of us be unavailable18:42
ronin__could you help me?18:42
bjfogasawara, is your script in kteam-tools?18:42
ogasawarabjf: it's not, it's just an Arsenal hack so I wasn't sure if we'd want it in Arsenal of kteam-tools, I'm indifferent either way as long as I can put it somewhere18:43
ogasawaras/of/or/18:43
bjfogasawara, i'd like to eventually be in kteam-tools, but can live with it somewhere else for now18:44
ogasawarabjf: I'd prefer for it to be in kteam-tools too, for some reason that's where I'd look for it if I were wanting to find it18:44
bjfogasawara, ~kernel-ppa would probably be fine18:44
ogasawarabjf: I'll throw a note at the top of the script that it expects to be run within Arsenal18:44
bjfogasawara, wfm18:49
=== firewave is now known as [GP]FireWave
* jjohansen heads to lunch20:16
=== yofel_ is now known as yofel
bjfbryceh, what's a valid LP version other than 'devel'?21:26
vanhoof_bjf: like edge, staging, production?21:26
bjfvanhoof_, no, those are "services"21:26
vanhoof_ah nm then :)21:27
=== vanhoof_ is now known as vanhoof
bjfbryceh, nm, not the problem21:29
brycehok21:30
broderbjf: "beta" and "1.0"21:30
bjfbroder, many thanks21:30
broder(beta matched karmic's lifecycle, so it's dead now; 1.0 will match lucid's lifecycle)21:30
broderbjf: https://launchpad.net/+apidoc/21:31
bjfbroder, and 'devel' is current development21:31
bjfbroder, thanks, that's so obvious21:31
brodernp21:31
sforsheesconklin, I'm guessing it's too late to get a patch into the next proposed kernel for natty ?22:36
sconklinsforshee: yes, it's already built in the PPA22:37
sforsheesconklin, then it will be another 3 weeks before the next one, right ?22:37
sconklinyes22:38
sforsheethanks22:38
sconklinunless it's like this train wreck of a cycle and then who knows.22:38
* jjohansen steps away for a bit22:46
=== jjohansen is now known as jj-afk

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