[01:23] <penguin42> I'm seeing black screen off a boot of the maverick 2.6.35-7 on a machine just upgraded from Lucid (i7 with Radeon 4350 graphics), if I remove the load_video and set gfxpayload=keep  from the grub text I at least see a panic (related to interrupt mapping) - should removing the load_video and set gfxpayload=keep be safe?
[01:25] <penguin42> failsafe is just as broke
[01:58] <penguin42> works with nointremap
[02:08] <achiang> https://lists.ubuntu.com/archives/ubuntu-devel/2010-July/030995.html
[02:08] <achiang> https://bugs.launchpad.net/ubuntu/+source/grub2/+bug/605614
[02:08] <ubot2> Launchpad bug 605614 in grub2 (Ubuntu) "Maverick's grub-pc package (1.98+20100710-1ubuntu1) causes unbootable system (affects: 2) (heat: 12)" [Undecided,New]
[02:09] <penguin42> achiang: It's not just the gfxpayload though, taking those out gets me a panic about interrupt remapping
[02:10]  * penguin42 is just writing a bug report
[02:10] <achiang> penguin42: a stack trace would help
[02:10] <penguin42> achiang: Just give me 3 mins to finish the bug report - it includes a lot of detail
[02:11] <achiang> heh, ok. /me was just drive-by triaging, i'm actually in the middle of something else.
[02:13]  * penguin42 taps flippers as he waits for launchpad
[02:14] <penguin42> ok, what have I missed on bug 605686
[02:14] <ubot2> Launchpad bug 605686 in linux (Ubuntu) "[maverick] noioremap needed - Blocked an interrupt request due to source-id verificiation failure (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/605686
[02:20] <achiang> penguin42: can you reproduce that with upstream?
[02:21] <penguin42> achiang: Havne't tried yet, and I'm about to go to bed (it's 2:20am :-) )   - but will upstream be much different from a maverick 2.6.35 ?
[02:21] <achiang> penguin42: not that much different, but i was going to suggest a bisect. the kernel shouldn't be panicing like that
[02:22] <penguin42> achiang: Well, a bit of googling suggests it could always be broken bios data structures - although I thought this was a pretty popular mbd
[02:23] <achiang> penguin42: are you running some unreleased bios?
[02:23] <penguin42> no
[02:23] <achiang> the stuff in your DMI table looks a little funky
[02:23] <achiang> dmi.chassis.vendor: To Be Filled By O.E.M.
[02:23] <penguin42> achiang: That's normal ; that's what you get if you buy a motherboard rather than buying a whole machine
[02:24] <achiang> penguin42: i mean, even if your bios is broken, 2.6.32 worked, right? so that means 2.6.35 shouldn't be choking
[02:24] <penguin42> yeh except they've turned on a new feature in .35 that wasn't previously used
[02:25] <penguin42> see http://old.nabble.com/-PATCH-1-2---Maverick--UBUNTU%3A--Config--Enable-CONFIG_INTR_REMAP%3Dy-to28966711.html#a28971597
[02:26] <achiang> right, so maybe the kernel needs a quirk for your bios
[02:26] <penguin42> yeh maybe; of course the interesting question is what % of bios actually work!
[02:27] <achiang> i'd expect that to work well on any server class machine. on a home machine... not so much
[02:27] <penguin42> why?
[02:28] <achiang> because as you've discovered, it needs the bios to be written correctly
[02:29] <penguin42> nod; you're hoping that the server ones are better written - that's being optimistic!
[02:29] <achiang> in my experience, server BIOSes are written by actual engineers. PC BIOSes seem to be a collection of random bits gathered by Arecibo and shoved into the motherboard
[02:29] <penguin42> hehe
[02:30] <achiang> in my past life, i was a kernel developer on server class machines, and had quite a lot of interface with the firmware teams
[02:31] <achiang> suffice to say, when i occasionally talked to the laptop guys, i didn't get the same level of confidence from their bios guys
[02:31] <penguin42> nod
[02:32] <achiang> penguin42: anyhow, if you have the time to do a bisect, i think that'd be great. otherwise, your bug report looks fine
[02:32] <penguin42> achiang: I'm fairly sure it's the intremap - it was turned on 23rd June according to that series of posts
[02:33] <achiang> penguin42: right, i mean to say, there's probably something wrong in the intremap code; a bisect will help figure out what it is
[02:34] <penguin42> that assumes that there is a working version in the remap case though - I'm not sure there is
[02:35] <jjohansen-afk> be back on later
[02:36] <penguin42> right, anyway, it's 2:36am and time for bed, I can look forward to figuring out all the other alpha fun tomorrow
[02:37] <penguin42> oops - just spotted typo in title of that bug - did say noioremap, changed to nointremap
[02:38] <penguin42> right, bed
[08:48] <smb> morning
[08:48] <cking_> morning
[08:49] <cooloney> smb: cking_ morning
[08:49] <cooloney> smb: GrueMaster tested security updates for fsl-imx51. boots fine
[08:50] <smb> cooloney, Yes, saw his email this morning
[08:50] <smb> cooloney, thanks to all guys testing
[08:51] <cooloney> smb: no problem, man
[08:52] <kraut> moin
[09:24] <insider> please help,  i use Ubuntu 9.10 with splash screen off and when i boot i see this message "udevd [1815] CONFIG_SYSF_DEPRECATED option udev ..." how to fix it?
[09:25] <insider> without reinstalling system because it works fine
[09:25] <insider> Linux 2.6.34 #1 SMP Thu Jul 8 19:41:21 EDT 2010 i686 GNU/Linux
[09:30] <apw> insider, those are just warnings arn't they
[09:30] <insider> yes but they slow down my booting speed
[09:30] <apw> they indicate your kernel and userspace are out of sync, which they are as you are not running the matched kernel
[09:30] <smb> I would think those are. And probably unavoidable with a newer kernel
[09:31] <insider> how can i disable CONFIG_SYSFS_DEPRECATED option?
[09:31] <insider> would it cause more problems?
[09:32] <apw> insider, you could try that, no idea if it would work not something i've tried
[09:32] <apw> how much time are you losing here, about .5s totoal ?
[09:32] <insider> about 15 seconds
[09:33] <apw> how many of those messages are you getting printed in totoal ?
[09:33] <apw> it would have to be a heck of a lot to account for 15s of boot time
[09:33] <insider> my laptop is Fujitsu V5535 with 2Gb Ram, Intel core 2 Duo 2ghz and boot time is 1.5-2 minutes
[09:34] <apw> how many of those messages are you getting printed in totoal ?
[09:34] <insider> i think they appear 2 or 3 times not less then 2
[09:34] <apw> printing 3 messages is not adding 15s to your boot
[09:34] <insider> previous version of Ubuntu 9.10 that i used booted at 45 seconds
[09:36] <apw> insider, so can you pastbin the dmesg of the slower boot
[09:36] <insider> can you explain how to do it or where to find logs?
[09:36] <insider> i'm newbie
[09:37] <apw> type 'dmesg > FILE' to collect the data (in a terminal window)
[09:37] <insider> ok
[09:37] <apw> and then use pastebinit to send it up to the pastebin
[09:40] <insider> wht is pastebinit?
[09:40] <insider> i haven't it installed
[09:40] <insider> i've already read
[09:41] <apw> pastebinit is a command to send a file to a pastebin and tell you the url to paste in here
[09:41] <apw> so that we can see the file contents
[09:41] <smb> or use http://pastebin.com/ ?
[09:42] <insider> thanks i'm installing it
[09:43] <insider> http://pastebin.com/R2LUWbBK
[09:49] <apw> insider, there is exactly one of those messages reported by the kernel
[09:50] <apw> insider, i assume this is your own configuration as the dmesg is lacking the normal timestamps which might allow one to find out whether it is taking a long time and which bit
[09:50] <apw> [    0.000000] Initializing cgroup subsys cpuset
[09:50] <apw> from my dmesg ^^
[09:51] <insider> strange
[09:52] <apw> why are you running a 2.6.34 kernel anyhow?
[09:53] <insider> i upgraded it from 2.6.30
[09:54] <insider> how to configure dmesg to show time?
[09:56] <apw> CONFIG_PRINTK_TIME=y
[09:57] <insider> thanks
[09:58] <apw> insider, why are you needing to run a custom kernel at all ?
[09:58] <smb> 2.6.30 was never a release
[09:58] <smb> Jaunty: 2.6.28
[09:59] <smb> Karmic: 2.6.31
[09:59] <apw> what userspace are you running all these on, as 2.6.30 was never a release ...
[09:59] <apw> heh jinz
[09:59] <apw> jinx
[09:59] <insider> 2.6.30.9
[10:00] <insider> intrepid
[10:00] <smb> That was 2.6.27
[10:00] <apw> you are running a kernel which is newer than lucid on an intrepid userspace ... no wonder its having compatibility issues
[10:00] <apw> insider, is there some reason you cannot just update the whole thing ?
[10:00] <insider> now i'm running karmic
[10:01] <insider> apw: i updated system
[10:01] <insider> apt-get update && apt-get upgrade won't help
[10:01] <apw> upgrade-manager ?
[10:02] <insider> didn't try
[10:03] <insider> apw: where to past it? 
[10:03] <apw> past ?
[10:04] <insider> paste
[10:04] <apw> to paste what ?
[10:05] <insider> this command upgrade-manager or it is not a command i misunderstood you
[10:07] <apw> sorry its update-manager ... normally use the menu when i use it
[10:10] <insider> only skype updates)))
[10:10] <apw> insider, what are you trying to do by running these kernels, are you just trying to get later stuff so a full update is acceptable
[10:10] <apw> if so you can use update-manager -d which will offer you the later releases as upgrade destinations
[10:10] <apw> without -d it wil only update your current release
[10:10] <insider> will try now
[10:10] <apw> if you need to keep the older release userspace but want thekernel i would suggest taking the kernels from the later releases directly instrad of making your own
[10:32] <insider> thanks for help going to reboot
[11:45] <amitk> any way to exclude a certain path from a git-diff ? I'm doing 'git diff --dirstat  v2.6.34..' and want to exclude the ubuntu/ directory
[11:57] <ikepanhc> want to know too.... listening
[12:00] <apw> amitk, don't know of any way to do that inversion no
[12:01] <amitk> bummer
[12:02] <apw> as file lists can be directories would something like
[12:02] <apw> git diff --dirstat  v2.6.34.. -- `ls -d | grep -v ubuntu`
[12:02] <apw> approximate what you wanted
[12:03] <apw> ls -1d strictly
[12:03] <apw> (thats a one)
[12:03] <apw> amitk, ^^
[12:15] <amitk> apw: 'grep -v ubuntu' takes away the offending directories from display but not from the end result (% of code in various directories among changes I care about).
[12:16] <apw> amitk, not from the result, i am using the directories as file limits so it should onlly report the diff for the directories not ubuntu and same with the overall results
[12:18] <amitk> apw: doesn't seem to work. http://paste.ubuntu.com/463985/
[12:20] <apw> git diff --dirstat v2.6.34.. -- `ls -1d * | grep -v ubuntu` 
[12:20] <apw> try that instead
[12:21] <apw> amitk, ^^
[12:21] <amitk> apw: yeah, that seems more sensible.
[12:21] <apw> git diff --dirstat v2.6.34.. -- `ls -1 | grep -v ubuntu` 
[12:21] <apw> or indeed that probabally works too
[12:22] <apw> yeah the last is the simplest
[12:22] <amitk> apw: yup, that seems to DTRT
[14:02] <amitk> tgardner: are we there yet? :) (IGEPv2)
[14:04] <tgardner> amitk, not quite. I at least git it to do something different on boot but I don't have the right serial cable. Loic is bringing one to Prague
[14:05] <tgardner> amitk, so, what else do we need for omap4 in mainline Maverick other then the config options?
[14:07] <amitk> tgardner: that itself will enable omap4 + one board based on omap4 (sdp). But Bryan will pull 1-2 patches from the tree slated for 2.6.36 merge to add support for the panda board that ARM team already has access to
[14:07] <amitk> that = applying my config patch
[14:08] <tgardner> amitk, cool. will that be sufficient such that we can drop the ti-omap4 branch altogether?
[14:08] <amitk> tgardner: :-D
[14:08] <amitk> you're kidding right?
[14:08] <tgardner> amitk, well, no.
[14:08] <amitk> tgardner: that branch has over 700 non-mainlined patches.
[14:09] <tgardner> we'll have 2 packages that both support omap4 ?
[14:09] <tgardner> guess I was too naive
[14:09] <ogra> FSVO *support*
[14:09] <amitk> tgardner: yes, one that is the whole works (-omap4 only) and one that is mainline-only with multi-omap support (omap3 and 4) in a single binary
[14:10] <apw> amitk, waht use if the multi-omap one if it doesn't work ?
[14:10] <tgardner> amitk, *sigh* - you've just dashed my hopes. maybe next release?
[14:10] <amitk> apw: it gets you the serial port on omap4
[14:10] <apw> amitk, ahh i see
[14:10] <amitk> and the same kernel gets you almost everything on omap3
[14:11] <amitk> so one small step to show kernel consolidation
[14:12] <amitk> tgardner: next release - hopefully a big chunk but not everything
[14:55] <apw> akgraner, hey you about ?
[14:55] <akgraner> apw, I am - what's up?
[14:56] <apw> akgraner, looks like the wireless link into your house might be sick
[14:57] <akgraner> hmmm - let me investigate - brb
[14:57] <apw> akgraner, your other link must be up
[14:57] <akgraner> it is  - I'm online  ;-)
[14:58] <apw> akgraner, could you try and telnet to hellhawk.shadowen.org for me from you rmachine
[14:58] <apw> then we can find out the IP addy of the link and see if its avail both ways
[15:10] <penguin42> does anyone understand the check_timer interrupt checking code?
[15:13]  * penguin42 is wondering if it could just disable interrupt remapping rather than panicing, there appears to be an iommu_disable_intr_remapping that sounds promissing
[15:17] <tgardner> penguin42, have you seen https://launchpad.net/bugs/605686 ? Same problem?
[15:17] <ubot2> Launchpad bug 605686 in linux (Ubuntu) "[maverick] nointremap needed - Blocked an interrupt request due to source-id verificiation failure (affects: 1) (heat: 6)" [Undecided,New]
[15:17] <penguin42> tgardner: Yeh, that's my bug
[15:17] <penguin42> and my mail :-)
[15:18] <penguin42> tgardner: But I was just wondering if it could actually fall back rather than panicing
[15:18] <tgardner> penguin42, ah, you're cloaked so its hard to tell.
[15:18] <penguin42> I am?
[15:18] <penguin42> hmm I should fix that
[15:19] <tgardner> penguin42, I'll bug Intel about this driver. Its quite complex.
[15:19] <penguin42> tgardner: Yeh, the interrupt handling on modern machines is scarily complex
[16:01] <cnd> http://git.kernel.org/?p=linux/kernel/git/acme/linux-2.6.git;a=shortlog;h=refs/heads/perf/core
[16:09] <akgraner> apw, you around?  
[16:11] <apw> akgraner, yep talking to pete
[16:12] <akgraner> ok the server is back online now
[16:13] <apw> akgraner, ok thanks ... will check it out
[16:29] <ogasawara> cking_: your work item to investigate intel graphics drivers on EFI, can I mark that done?
[16:29] <ogasawara> cking_: I saw no response from the email you sent
[16:30] <lag> akgraner: Do you know what the issue was?
[16:35] <akgraner> lag, I don't know exactly what happens with that particular server - I just know what to do when it hangs :-/   
[16:36] <lag> akgraner: But it happened to both servers? Unless there is a sever in between them?
[16:36] <akgraner> lag, you'll have to ask pgraner about that when he is online again... sorry :-/
[16:36] <lag> No worries :)
[16:36] <lag> Thanks for sorting it in any case
[16:36] <akgraner> lag, I'm the NTEU :-)  but happy to help when and where I can.
[16:37] <lag> :-)
[16:40] <tgardner> ogasawara, https://launchpad.net/bugs/605686
[16:40] <ubot2> Launchpad bug 605686 in linux (Ubuntu) "[maverick] nointremap needed - Blocked an interrupt request due to source-id verificiation failure (affects: 1) (heat: 6)" [Undecided,New]
[17:05] <penguin42> Sanity check; I'm following https://help.ubuntu.com/community/Kernel/Compile  ; if I've got the source archive and want to patch it then do I just patch the directory that's checked out there - what do I need to do to make sure it gets a name unique from the standard ubuntu kernel version?
[17:05] <penguin42> will the AUTOBUILD do that ?
[17:14] <tgardner> penguin42, edit the debian.master/changelog and add '.my_version_here' to the stuff in parens
[17:14] <penguin42> ah ok, thanks
[17:15] <tgardner> penguin42, don't forget to rerun the 'clean' target afterwards
[17:16] <penguin42> tgardner: OK, so do whatever edits I wanted, update the changelog, then do the clean thing and then the binary-whatever ?
[17:17] <tgardner> penguin42, sounds right
[17:17] <penguin42> ok, thanks
[17:38] <tgardner> mpoirier,  http://cdimage.ubuntu.com/ports/releases/10.04/release/ubuntu-10.04-server-armel+omap.img
[17:39]  * manjo getting lunch will be back soon
[18:01] <jjohansen> mpoirier: move ureadahead.conf to ureadahead.conf.disabled
[18:01] <mpoirier> jjohansen: fantastic thanks.
[20:10] <jjohansen> -> Lunch
[20:13] <pgraner> ogasawara, ping
[20:13] <ogasawara> pgraner: pong
[20:14] <pgraner> ogasawara, thought you might like to keep an eye on this since you just rebased
[20:14] <pgraner> ogasawara, https://bugzilla.kernel.org/show_bug.cgi?id=16401
[20:14] <ubot2> pgraner: Error: Could not parse XML returned by bugzilla.kernel.org: The read operation timed out (https://bugzilla.kernel.org/xml.cgi?id=16401)
[20:14] <pgraner> ogasawara, ext3 corruption at random 
[20:15] <ogasawara> ogasawara: thanks, I'll put it on my radar.  bugzilla seems to not want to load at the moment
[20:15] <pgraner> ogasawara, yea I caught it on the ext upstream list
[20:35]  * ogasawara lunch
[20:51] <akgraner> bjf, ping
[20:51] <bjf> akgraner, pong
[20:53]  * penguin42 wonders where dch put my changelog entry
[21:54]  * penguin42 does a 'wow - it boots'
[22:01]  * penguin42 attaches a gratuitous hack patch to bug 605686 - it seems to work, but I wouldn't put much more confidence in it
[22:01] <ubot2> Launchpad bug 605686 in linux (Ubuntu) "[maverick] nointremap needed - Blocked an interrupt request due to source-id verificiation failure (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/605686
[22:40] <jjohansen> running an errand
[23:19] <LLStarks> ogasawara, maverick kernel now has built-in ndiswrapper modules, right?
[23:24] <ogasawara> LLStarks: it's not built-in, enabled as a module
[23:24] <ogasawara> LLStarks: config.common.ubuntu:2750:CONFIG_NDISWRAPPER=m
[23:52] <LLStarks> thanks ogaswara, still need to pick juliank's brain about his dkms package