/srv/irclogs.ubuntu.com/2010/10/19/#ubuntu-kernel.txt

=== rsalveti` is now known as rsalveti
=== jjohansen is now known as jj-afk
=== smb` is now known as smb
* smb yawns08:46
amitksmb: \o08:53
smbamitk, Dude!08:53
smbStill alive? :) I saw you ride the same plane to Orlandi08:54
smbo08:54
amitksmb: cool, will see you in frankfurt on sunday then...08:57
smbYep08:58
amitkI guess that will be one of the "ubuntu" flights08:59
smbWell I think I only saw ogra on the same flight. But searching the wiki is kind of error prone. And it does not show the non-canonicals09:00
smbAt least it reaches cab allowance mass. :)09:00
amitkheh09:07
apwamitk, hey09:13
amitkapw: howdy?09:15
apwi hear you are linaro power management dude, checking you have the abilty to sort out getting your blueprints accepted09:16
apwyou have a lot sitting on the queue for uds-n, but not yet waved through to the agenda.  i can help if not09:17
apwamitk, looks like you have 25 blueprints proposed, at the moment09:18
smbapw, speaking of those, what was the sort of target we wanted the proposed testing one be? We spoke about that but I already forgot09:19
amitkapw: I'll have to remove all the proposals except the 4 that are correctly named09:19
apwsmb, i think we thoguht it should be a qa one really, maybe we need to look at marjo's ones to see if there already is one to go to or see if he wants to take it on09:20
amitkapw: or are you talking about all of linaro?09:20
apwamitk, i am talking about just the ones with your name on ... i assume most are just tracking prints to hold work items and don't need sessions?  i can accept/reject them for uds-N if you can't; if you tell me the four that need sessions i can sort out getting those in and the rest off the agenda09:21
smbapw, Ok. Want me to sent him an email?09:21
apwamitk, to get them on your work items list you need to get them series-goal of natty rather than sprint of uds-n ... if that makes sense09:22
apwamitk, overall i've been asked to help linaro clean up their blueprints so we can see the wood for the trees on the blueprint queue as you all seem to have proposed all of your blueprints rather than just the ones needing sessions09:23
amitkapw: my 4 blueprints that I care about start with "hardware-linaro-n-meta-"09:23
amitkapw: everything else doesn't require a session09:24
apwamitk, excellent ... happy for me to run through them and sort the sessions out then ?09:24
amitkapw: and for those 4, I want back to back double sessions, one per day09:24
amitkapw: yes please09:24
apwamitk, sadly i am not able to target things in linaro projects to natty series-goal, you need a linaro 'driver' for that09:25
amitkapw: I can do that (only for the 4, right?)09:25
apwamitk, if you want the blueprint to carry work items which show up on the burn-down stuff then that blueprint needs a series-goal, likely thats all of them ... presume you using blueprints as 'jobs to do' prints09:26
amitkapw: yes, but we have our own "project" - linaro-pm-wg (https://edge.launchpad.net/linaro-pm-wg) with our own series and milestones09:27
amitkand natty doesn't show up in that series09:27
* apw realises there really should be a like #ubuntu-uds channel like #ubuntu-release where you can find help with things and discuss things like this09:27
amitkwe also have our own burn-down infrastructure09:27
apwamitk, ah good point09:27
* apw looks where the are looked for in your burndown09:28
amitkapw: http://people.canonical.com/~pitti/workitems/linaro-pm-wg/all.html09:28
apwamitk, oh you may be ok actually it may not be tied to a release09:29
apwthen you'd not need to target them all, and you save loads of effort!09:29
amitkapw: we've had to shoehorn of lot of stuff to fit into the ubuntu-way for uds though we don't strictly follow the milestones and processes09:32
apwyeah i bet, the process doesn't fit snugly for us where it was designed for09:33
amitkapw: so will you reject everything but the 4?09:37
apwamitk, working on it now09:37
amitkapw: many thanks and a beer pledge :)09:38
apwamitk, nice :)09:39
apwamitk, do you have the names of the four you wanted accepted, i only saw two on the way through09:42
apwso i can double check they are right09:42
amitkapw: https://blueprints.edge.launchpad.net/linaro-pm-wg (The first four in there marked essential)09:43
apwamitk, so you personally only have 1 session (2 hours)09:44
apwok all of those first 4 are accepted, i'll see about getting them 2 hours next09:45
amitkapw: yeah, others have been assigned to those metas. And if you could, please spread around the double slots so that one is in the morning session, one in afternoon, etc. And only 1 session per day. And start the monday session in the afternoon.09:47
* amitk believes a #ubuntu-uds helpdesk is definitely called for...09:48
apwamitk, heh if they were even on the agenda at all we'd be looking better, looks like the importer is stuck09:48
amitkbi-annual circle of pain09:50
apwi know ... so much stuff is other- i think it broke09:51
smbbig badaboom?10:07
amitkbadaboom10:11
smbSo we probably need someone who knows how to negotiate... ;-P10:12
apwamitk, ok ... currently your 'prints arn't making it into the scheduling system, something todo with a filter ... they are working on it and hope to have it fixed today10:20
amitkapw: ack, thx10:22
apwamitk, now who can i talk to about the toolchain linaro prints10:22
amitkmichaelh1 on #linaro10:22
apwwhat timezone is he ?10:23
amitkNZ, I'm asking if he's around10:23
amitklikely not10:23
apwamitk, seems late there for him10:40
apwamitk, i see that michael has actually named a subset of his blueprints foo-linaro-n-*, so i think i can just accept the ones in that form and reject the rest 10:51
apwamitk, trying to debug this issue with the blueprints, could you move https://blueprints.edge.launchpad.net/linaro-pm-wg/+spec/hardware-linaro-n-meta-pm-tools to 'Drafting' or 'Discussion'10:57
amitkapw: i remember him mentioning doing the renames11:01
amitkapw: marked all four as discussion11:02
=== amitk is now known as amitk-afk
=== diwic is now known as diwic_afk
=== amitk-afk is now known as amitk
apwamitk, does http://summit.ubuntu.com/uds-n/track/hardware/ look like what you intended12:15
amitkapw: almost - can you interchange cpufreq with multicore? (thursday and monday)12:18
apwamitk, you bugger ... 12:19
amitkapw: I hope it doesn't involve hacking assembly code :-p12:19
apwamitk, nope ... which on monday ... i've got them both off now12:20
apwand forgotten which was which already ... doh12:20
amitkapw: lol, so cpufreq on monday and multicore on thursday12:20
apwamitk, how about that12:22
apwsadly you have to use the day views to move things at the mo (due to a bug) so three windows are involved12:22
apwamitk, its a good job we are doing these long ones now, its not easy to get them in12:22
apweven this early12:22
Ampelbeinhi there! For a lcd-display driver I need the usbhid_modify_dquirk symbol address, but I can't seem to find it in /boot/System.map. Can you point me in the right direction on where to look?12:23
apwAmpelbein, which release are you trying to find it in12:23
amitkapw: looks good!12:23
Ampelbeinmaverick, 2.6.35-22-generic12:23
apwAmpelbein, in maverick it is a static function so its not exported at all12:24
apwamitk, keep your fingers crossed people leave them alone12:25
apwi don't think i can lock them or anything12:25
apwAmpelbein, also was static in lucid12:25
Ampelbeinapw, ok, that means kernel patching for me then :-( thanks!12:26
apwamitk, as it is so hard to get things scheduled i am going to assume that the toolchain ones which are named correctly are the ones he wants in and the others not, and he can correct them if not afterwards12:27
apwelse there will be no slots for them left12:27
apwsmb, have you marked your participation (particularly essential ones) on your blueprints, being the owner or assignee or anything else has no effect on the attendance thing12:29
smbsmb, No; I guess I should do so now12:35
smbapw, And probably I should stop talking to myself12:41
JFolag, morning... reading the e-mails and thread on the subject now :)12:53
JFomy only concern is that I will be in Cambridge for Plumbers that week12:54
JFowill you be there too?12:55
apwJFo, i think he wandered off12:58
JFoah12:58
JFomorning apw 12:58
JFo:)12:58
apwmorning JFo 12:58
lagI'm back12:58
JFolag, will you be at Linux Plumbers?12:59
lagYeah :(12:59
JFoso yeah, we'll both during that meeting :)12:59
lagThat's gonna suck12:59
JFoshrug* :)13:00
JFoso yeah, 4AM to 4:45AM then13:01
JFosince they are exactly 12 hrs off from this TZ13:01
lagOuch13:01
JFoyours is 3:15 to 4 AM13:02
JFoif we keep that schedule13:02
JFomeaning, if we are present for the whole thing, we will have been up all night13:03
JFo:)13:03
JFoyay!13:03
lagGet the coffee on the go13:06
JFoyup13:06
JFoand get in bed early that next night :)13:06
JFoI've responded to his e-mail13:07
JFoso we shall see13:07
JFoonly thing I see possibly impacting us is if we need to call in to them13:07
JFobut I am sure we can work around that13:07
lagSkype13:08
JFoyep :)13:08
persiaapw, My opinion (which I think I've stated before) is that we ought just have powerpc be in the main kernel, to avoid config issues and scheduling issues.13:11
persiaI understand that this may not happen, for workflow reasons.13:11
persiaI'll continue to want it discussed each UDS until I don't notice differences in kernel function between architectures, but I don't have any strong expectations about conclusions.13:12
apwpersia, right now about the only thing thats not in the main kernel is the meta package, so yes that is on my agenda at the moment13:13
persiaI thought so, from the spec name.  Just figured I'd make sure when you raised that spec in another context.13:13
apwcool on the same page then13:14
persiaNever doubted it :)13:14
lyhana8please patch is to be done on fglrx kernel module to compile again on kernels with CVE-2010-3081 fixed13:23
ubot2lyhana8: The compat_alloc_user_space functions in include/asm/compat.h files in the Linux kernel before 2.6.36-rc4-git2 on 64-bit platforms do not properly allocate the userspace memory required for the 32-bit compatibility layer, which allows local users to gain privileges by leveraging the ability of the compat_mc_getsockopt function (aka the MCAST_MSFILTER getsockopt support) to control a certain length value, related to a "stack pointer13:23
lyhana8need for ubuntu 10.0413:23
dupondjecould somebody take a look @ https://bugs.launchpad.net/bugs/604122 ?13:37
ubot2Launchpad bug 604122 in linux (Ubuntu) "mmc0: Got command interrupt 0x00030000 even though no command operation was in progress. (affects: 2) (heat: 14)" [Undecided,Confirmed]13:37
JFodupondje, looking13:48
apwlyhana8, i thought that that was fixed already ... tseliot <--13:49
tseliotapw, lyhana8: yes, that was done some time ago13:50
apwtseliot, thought so ... lyhana8 do you have -updates enabled on this system13:51
dupondjeJFo: revert https://lists.ubuntu.com/archives/kernel-team/2010-May/010660.html, and it works again. So something must be broken in the sdhci driver ?13:56
JFoI was wondering about that. it is affected only by the config change yes?13:58
dupondjeseems so13:58
JFoodd13:58
JFoapw, if you get a sec, could you take a look at bug 60412213:59
dupondjeits quite annoying, as it makes SD cards useless on my computer now :s13:59
ubot2Launchpad bug 604122 in linux (Ubuntu) "mmc0: Got command interrupt 0x00030000 even though no command operation was in progress. (affects: 2) (heat: 14)" [Undecided,Confirmed] https://launchpad.net/bugs/60412213:59
JFodupondje, I can imagine13:59
JFo or lag ^^13:59
JFoif one of you are available13:59
dupondjethe weird thing is that it sometimes works, sometimes not ...13:59
JFothat is odd13:59
JFoso if you change the config it works all the time?13:59
JFojust wanting to have it clear in my head :)14:00
lagI can take a look14:00
smbJFo, apw dupondje Though what the option does is to activate some code do disable some controller which otherwise inteferes with the generic driver. Does this really work with the option disabled?14:01
dupondjewell I reverted that change, and then it worked without issues. Can try it again (was somewhere in the development stage of maverick) :)14:01
JFodupondje, that would be awesome if you don't mind14:01
dupondjewait :) i'll boot my quadcore ^^14:02
JFocool, thanks :)14:03
smblyhana8, For that compile failure all the ubuntu packages should be fixed by now.14:03
=== lan3y is now known as Laney
dupondjebtw, can I build kernel with pbuilder ?14:21
apwogra, does https://blueprints.edge.launchpad.net/ubuntu/+spec/hardware-arm-n-omap-edid-autodetection need a UDS session ?14:26
dupondjebuilding, gtg now, back in a hour :) then building will be done I guess also14:30
lyhana8smb: since when? cause I tried yeasterday and it wasn't working on Linux Mint 9 = Ubuntu 10.0414:31
smblyhana8, tseliot has uploaded packages the week after that security update went out14:32
smbtseliot, I forgot the dkms package name again display-drivers something?14:33
tseliotsmb: fglrx-installer14:34
lyhana8what is the package version I should look for?14:34
tseliotlyhana8: four weeks ago I uploaded 2:8.723.1-0ubuntu5 which contains the fix14:34
smbtseliot, Is that Lucid?14:35
tseliotlyhana8: if you don't have it, it's likely that you don't have -updates enabled14:35
tseliotsmb: yep14:35
tseliothere you can see the source for each version of Ubuntu: https://launchpad.net/ubuntu/+source/fglrx-installer14:36
tseliotit was bug 642518, BTW14:37
ubot2Launchpad bug 642518 in linux-restricted-modules-envy-2.6.24 (Ubuntu Karmic) (and 14 other projects) "[MASTER] package fglrx 2:8.723.1-0ubuntu4 failed to install/upgrade: Kernel fix for CVE-2010-3081 breaks fglrx (affects: 283) (dups: 240) (heat: 2098)" [Undecided,Invalid] https://launchpad.net/bugs/64251814:37
smbtseliot, That probably confused me because the binary is called differently14:37
tseliotlyhana8: or maybe you installed a more recent version of the driver and you're not getting the update because of that14:37
tseliotsmb: yes, the binary is just fglrx14:38
smbtseliot, Which is what lyhana8 has to look for in dpkg -l :)14:38
lyhana8I got the right version 2:8.723.1-0ubuntu514:38
lyhana8smb: I got:  fglrx-modaliases                      2:8.771-0ubuntu114:39
=== brot_away is now known as brot
tseliotlyhana8: 8.771? I'm wondering where you took that from14:40
=== brot is now known as Guest14723
lyhana8the binary package probably. I tried to compile from their .bin14:41
lyhana8s/.bin/.run/14:41
tseliotlyhana8: try with "sudo apt-get --purge remove fglrx" and then "sudo apt-get install fglrx" and let me know if there are any errors14:43
tseliotlyhana8: if there are errors, I'd like to see them14:43
apwslangasek, you have a number of other-linaro-* blueprints, just confirming that they all need sessions ?14:46
=== Guest14723 is now known as brot
lyhana8tseliot: my problem was that install the fglrx driver trigger this error: http://pastebin.com/3QbWsruq15:03
tseliotlyhana8: it's not the same problem then. The security update prevented the fglrx module from compiling.15:05
lyhana8tseliot: those log are from yesterday, I didn't get any error installing fglrx fglrx-amdcccle15:09
lyhana8(just minutes ago)15:09
tseliotlyhana8: that confirms my theory. It must be something else15:10
lyhana8tseliot: well it may be that the fglrx package in the repo isn't the good version15:13
tseliotlyhana8: no, it can't be that. It's a completely different problem15:13
tseliotlyhana8: please file a new bug report about it15:13
lyhana8my card is a Mobility radeon HD 5470, which is supported by ati driver 10.915:14
=== bjf[afk] is now known as bjf
slangasekapw: the ones that are targeted to uds-n should be the ones that need sessions, yes16:05
apwslangasek, cool, most of your fellow creators have made all of their blueprints targeted to uds-n16:06
apw(probabally due to the create link they used)16:06
slangasekmore likely due to it being their first time through the process and not having thought through whether each blueprint needs a session16:07
slangasekstrangely, I don't see any way to *un*target a blueprint to a sprint16:08
JFook, stepping away for a bit. I'll be back around in a bit.16:12
avinashhmhi, is there any way to remove windows line ending from patch .. [^M] .. i tried fromdos and :s/^M//g .. .but after this, my patch isn't applying .. any help ...16:25
smbavinashhm, to do s/<ctrl-v><ctrl-m>$//g should work. But maybe there is more broken than line ending? Like linebreak added by mailer16:28
avinashhmsmb, but number of lines still are same .. but some how something is getting disturbed .. not sure ..16:34
bjfavinashhm, meld is a good tool to look at the differences between two files, it might help16:36
smbavinashhm, Would you be 100% it was ok with the ^Ms. It could have been broken when sending, before you received it16:36
avinashhmbjf, ... hmmm.. let me try it now ... 16:37
avinashhmsmb, .. possible ... :-) ...16:38
cking_dos2unix works well in munging windows line endings to UNIX convention16:55
=== jj-afk is now known as jjohansen
=== cmagina is now known as cmagina-lunch
avinashhmcking_, yep ...  i suppose now a days its known as fromdos .. .[ package of tofrodos ] ...17:17
=== cmagina-lunch is now known as cmagina
dupondjeI tried to build kernel with pbuilder, but now I get: EE: Previous or current ABI file missing!18:17
dupondjeany ideas guys ? :)18:17
kamaldupondje: ah, the dreaded (but very very common) ABI check failure ... this will help:18:19
kamalhttps://wiki.ubuntu.com/KernelTeam/KernelMaintenance18:19
kamalor18:19
kamalhttps://wiki.ubuntu.com/KernelTeam/KernelForIdiots18:19
kamalfind "ABI" on those pages to read about how and why you need to disable the ABI check.18:19
kamaldupondje: specifically for pbuilder builds, the "skipabi=true" method won't work for you ... you'll need to touch the 'ignore' files as directed in https://wiki.ubuntu.com/KernelTeam/KernelMaintenance section "Overriding ABI check failures" item #2.18:22
manjoJFo, !!! you on mumble?18:55
JFomanjo, nope, do I need to be?.19:36
=== kancerman_ is now known as kancerman
* ogasawara lunch20:01
cking_time to call exit(EXIT_SUCCESS) on another day20:12
* jjohansen -> lunch20:59
EdganAnyone knowledgeable about realtime kernels in the channel?22:47
=== bjf is now known as bjf[afk]

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