[06:23] <ppisati> moin
[07:30] <smb> morning and ciao ppisati ;)
[08:02] <ppisati> smb: ciao Stefan :)
[08:02] <ppisati> brb
[08:02]  * ppisati goes to preparare some more ICED coffee...
[08:02] <ppisati> it's already 29C here...
[08:09] <cooloney> ppisati and smb morning
[08:27]  * henrix is having probs with isp... again
[08:32] <hrw> hi people
[08:33] <hrw> I bought Toshiba harddrive with usb 3.0. Connected to desktop (running quantal) and got this in dmesg: http://paste.ubuntu.com/1158604/ - disk is unusable as it connects/disconnects all the time
[08:35] <hrw> http://paste.ubuntu.com/1158610/ is lsusb -vvv output
[08:36] <hrw> ideas?
[08:46] <cooloney> ppisati: nice trip for vacation. heh
[08:47] <xnox> hrw: I am all ears, since I am experiencing similar. But for me it might be lvm related.
[08:48]  * xnox is not kernel dev, just a lurker.
[08:48] <ppisati> cooloney: thank you :)
[08:48] <hrw> xnox: http://www.digipedia.pl/usenet/thread/19505/33113/ is what I want to check
[08:49] <hrw> xnox: disk will be mostly used with usb2.0 laptop but for future use I prefer usb3.0 ;)
[08:50] <xnox> hmmm... for me it just spins down and gets "ejected" and then all of my LVs and VGs are borked =/ giving i/o errors. not sure it's the same.
[08:52] <cooloney> ppisati: if you need me take care of some ARM stuff during your vacation, please let me know
[08:59] <ppisati> cooloney: feel free to chime in, the ti-omap4 lp bugs page is always looking for some love :)
[09:49] <cooloney> ogra_: hey, i'm looking at the bug https://bugs.launchpad.net/ubuntu/+source/linux-ti-omap4/+bug/746137
[09:49] <ubot2`> Ubuntu bug 746137 in jasper-initramfs "Page allocation failure on Pandaboard and Beagle XM" [Undecided,Fix released]
[09:49] <ogra_> cooloney, ah, cool, note that this isnt arm specific 
[09:50] <ogra_> seems to be an issue with many USB based wlan cards
[09:50] <cooloney> ogra_: you mentioned to use GFP_ATOMIC instead of GFP_KERNEL, is that any bug about that?
[09:50] <ogra_> cooloney, not sure, marvin24 pointed me in #ubuntu-arm to it 
[09:50] <ogra_> ask him, i guess
[09:51] <cooloney> ogra_: oh, can we reproduce on x86? actually i got this on my pandaES frequently
[09:51] <ogra_> i think you can if you boot with a really low mem= value and have the same highmem split in the kernel 
[09:52] <ogra_> but ask marvin24, he was really into that bug for a while
[09:52] <cooloney> ogra_: ok, cool
[09:55] <cooloney> ogra_: also from infinity's comments, looks like if we change vm.min_free_kbytes = 32768, the issue is gone
[09:55] <ogra_> right, its just that we dont want to set that var for everyone...
[09:55] <ogra_> and setting a sysctl variable conditionally isnt easy
[09:56] <ogra_> also its somnething that looks like the kernel should set it right
[09:56] <ogra_> instead of userspace with hackish detection scripts
[09:57] <hrw> II: Checking modules for generic...previous or current modules file missing!
[09:57] <hrw>    /home/hrw/HDD/devel/ubuntu/linux-3.5.0/debian.master/abi/3.5.0-10.10ubuntu1/amd64/generic.modules
[09:57] <hrw>    /home/hrw/HDD/devel/ubuntu/linux-3.5.0/debian.master/abi/3.5.0-10.10/amd64/generic.modules
[09:57] <hrw> make: *** [module-check-generic] Błąd 1
[09:57] <hrw> argh...
[10:00] <hrw> how to disable all abi checks etc? I want to test some usb3.0 patches and kernel compilation drives me mad
[10:07] <cooloney> ogra_: looks like marvin24 is not online
[10:08] <ogra_> patient :) its noon here, he is probably at lunch
[10:08] <cooloney> ogra_: oh, got it, i assume he is in US
[10:08] <ogra_> germany 
[10:09] <ogra_> about 80km south of me 
[10:09] <cooloney> ogra_: that's close. how's you imx6 board.
[10:10] <ogra_> fine, idling on the shelf 
[10:10] <ogra_> i'm working on an automated qemu based testsuite atm ... so i havent looked much at real HW the last days
[10:15] <hrw> ok, time to reboot and check will usb3 drive work
[10:16] <cooloney> ogra_: i think ppisati probably have already sloved this issue in Ubuntu-3.5.0-208.14
[10:16] <ogra_> dunno, i havent seen the issue on that kernel ... but then i didnt do many RAM hungry tasks yet
[10:21] <cooloney> commit e3a095d141d53651dd9f406502cbdb87f456c002
[10:21] <cooloney> Author: Andy Green <andy.green@linaro.org>
[10:21] <cooloney> Date:   Wed Jun 27 15:10:23 2012 +0800
[10:21] <cooloney> HACK force min_free_kbytes set by proc to min 32K to workaround smsc problems 
[10:21] <cooloney> Signed-off-by: Andy Green <andy.green@linaro.org> Signed-off-by: Paolo Pisati <paolo.pisati@canonical.com>
[10:21] <cooloney> commit a8fe4ec3a2207dfc7b18a9875569864a68e6cb87
[10:21] <cooloney> Author: Andy Green <andy.green@linaro.org>
[10:21] <cooloney> Date:   Wed Jun 27 12:16:28 2012 +0800
[10:21] <cooloney> HACK force min_free_kbytes to 32K to workaround smsc problems 
[10:21] <cooloney> Signed-off-by: Andy Green <andy.green@linaro.org> Signed-off-by: Paolo Pisati <paolo.pisati@canonical.com>
[10:22] <cooloney> ogra_: i saw it set that min_free_kbytes to 32K as minium value
[10:24] <ogra_> ugh, ugly 
[10:25] <ogra_> we cant do that for omap i suppose
[10:25] <ogra_> might work for omap4 indeed
[10:26] <cooloney> ogra_: yeah, only on ti-omap4 branch
[10:27] <cooloney> actually there is no kernel config option for that default value. i think we can introduce one
[10:27] <ogra_> in mainline ? 
[10:28] <ogra_> on the pandaboard that bug isnt hitting as hard as on the beagle ... so that fix would have to go into master too
[10:28] <cooloney> if it is a right method to solve the issue, i think it is good for mainline
[10:28] <ogra_> k
[10:29] <cooloney> ogra_: right, but master also includes x86.
[10:30] <cooloney> probably have an option named CONFIG_MIN_FREE_KBYTES, if the min_free_kbytes < CONFIG_MIN_FREE_KBYTES, then let min_free_kbytes = CONFIG_MIN_FREE_KBYTES
[10:31] <cooloney> so we can set CONFIG_MIN_FREE_KBYTES as 32K in omap and omap4
[10:32] <ogra_> yeah
[10:32] <ogra_> sounds good
[10:32] <cooloney> ogra_: but maybe kernel maintainers will push that back as why not change that in userspace. heh
[10:34] <ppisati> ogra_: cooloney IMO that issue was never really resolved
[10:35] <cooloney> ppisati: ok, so even with the patches you applied from linaro, we still have this issue, right?
[10:36] <cooloney> ppisati: i will try this kernel soon
[10:42]  * cooloney out for dinner
[11:08] <apw> hrw, i assume you figured out the abi disable
[11:11] <hrw> apw: yes and got my usb 3.0 device working
[11:14] <hrw> will open bug against quantal kernel
[11:16] <apw> hrw, lets us know the bug number here
[11:17] <hrw> sure
[11:29] <hrw> bug 1039478
[11:29] <ubot2`> Launchpad bug 1039478 in linux "WARN Successful completion on short TX: needs XHCI_TRUST_TX_LENGTH quirk?" [Undecided,New] https://launchpad.net/bugs/1039478
[11:29] <hrw> patches attached
[11:49]  * henrix -> lunch
[11:51] <ppisati> rsalveti: http://people.canonical.com/~ppisati/linux-image-3.5.0-208-omap4_3.5.0-208.14~plus1_armhf.deb
[11:51] <ppisati> rsalveti: http://people.canonical.com/~ppisati/linux-headers-3.5.0-208-omap4_3.5.0-208.14~plus1_armhf.deb
[11:52] <ppisati> rsalveti: i rebased all the missing tilt-3.4 patches on our Q/omap4 kernel
[11:52] <ogra_> plus1 !
[11:52] <ppisati> rsalveti: check if it makes any diff wrt your pvr hangs
[11:52] <ppisati> ogra_: plus4! :)
[11:52] <ogra_> with such a name it *must* be better :)
[11:54]  * ppisati crosses fingers...
[11:54] <rsalveti> ppisati: sure, will test it
[11:55]  * ppisati goes to enable -DDEBUG and -Wall in arm/arm/mach-omap2 again...
[11:55] <ogra_> so that the rigbuffer eats all our ram again !
[11:55] <ogra_> *ring
[11:57] <ppisati> ogra_: just in my build, to catch as much crap as possible
[11:58] <ogra_> yeah, i wasnt actually complaining, i just tried to sound like :)
[14:13] <verwilst> hi
[14:13] <verwilst> i'm currently debugging a gfs2 kernel oops with the 12.04 kernel
[14:15] <rtg> ogasawara, I uploaded the LTS meta package that you had in the can.
[14:30] <verwilst> https://bugs.launchpad.net/ubuntu/+source/gfs2-utils/+bug/1020207 this bug is currently attached to gfs2-utils, but i think it's a kernel issue
[14:30] <ubot2`> Ubuntu bug 1020207 in gfs2-utils "gfs2 kernel oops when deleting file on other cluster node" [Undecided,Confirmed]
[14:33] <ppisati> diwic: i've an atom board with nvidia chipset, and i'm running the nvidia driver
[14:33] <ppisati> diwic: /proc/asound/cards shows both the intel and nvidia card
[14:33] <ppisati> diwic: but aplay shows only the intel side
[14:34] <ppisati> diwic: what could be wrong?
[14:34] <diwic> ppisati, hmm
[14:34] <diwic> ppisati, can you give alsa-info?
[14:37] <diwic> ppisati, or check in dmesg for strange errors
[14:38] <bjf> cking: bug 1039561
[14:38] <ubot2`> Launchpad bug 1039561 in ecryptfs "The ecryptfs test file-concurrent.sh hangs on both natty and lucid" [Undecided,New] https://launchpad.net/bugs/1039561
[14:38] <Worf> I hope here is the right place for questions like this: Maybe i'm missing something obvious, but i tried a minimal ubuntu install with debootstrap and with  linux-image-3.5.0-11-generic, and i have all kinds of hardware support issues. Like no mouse device, harddisk devices missing, network card missing, ... Did i miss something? I admit i'm not familiar how kernel stuff is packaged in Ubuntu ...
[14:39] <cking> bjf, can you add some more context? which kernel perhaps? ;-)
[14:39] <ppisati> diwic: it's an old board i've it around
[14:39] <Worf> 3.5.0-11.11 for amd64 to be precise
[14:40] <ppisati> diwic: and i was trying to getr video (and audio) playback via hdmi
[14:40] <ppisati> diwic: it's not connected now
[14:40] <ppisati> diwic: e.g. does aplay go through the sound server? or does it check the hw device?
[14:41] <ogasawara> Worf: did you make sure to also install the linux-image-extra package?
[14:41]  * ogasawara back in 20
[14:41] <diwic> ppisati, aplay -l would list the hardware devices only
[14:42] <bjf> cking: kernel versions added
[14:42] <ppisati> diwic: ok so if it's no there it's an hw/kernel problem
[14:42] <Worf> ogasawara: ok, it looks like that would be what i missed ... sorry for bothering ...
[14:43] <diwic> ppisati, yep. I would look in dmesg for errors
[14:43] <cking> bjf, /me has a look
[14:43] <diwic> ppisati, (note that aplay -L lists something different)
[14:43] <ppisati> diwic: it's disconnected as i said, i'll give it a second look tonight
[14:43] <diwic> ppisati, ok
[14:44] <ppisati> rsalveti: any news on plus1 kernel?
[15:51]  * henrix will be out for a bit
[15:53] <hrw> jsalisbury: ping
[15:54] <hrw> jsalisbury: bug 1039478 is what I would like to talk about case today I will have to give disk back to the owner
[15:54] <ubot2`> Launchpad bug 1039478 in linux "WARN Successful completion on short TX: needs XHCI_TRUST_TX_LENGTH quirk?" [Medium,Triaged] https://launchpad.net/bugs/1039478
[16:09]  * apw curses overlayfs ...
[16:10] <ogra_> switch back to unionfs :)
[16:10] <smb> overlayfs withers and dies...
[16:12] <cwillu_at_work> the cool kids use btrfs with compression :p
[16:12]  * hrw curses xhci
[16:12] <xnox> cwillu_at_work: the cool kids also do not notice their data vanishing into the clouds
[16:13] <cwillu_at_work> sure they do
[16:13] <cwillu_at_work> the checksums would fail!
[16:13] <xnox> then what?
[16:13]  * cwillu_at_work had daily backups long before btrfs was a twinkle in cmason's eye
[16:13]  * xnox ext4 doesn't need checksums cause it doesn't loose dat
[16:13] <cwillu_at_work> in other news:  ext4 adds checksums
[16:14] <xnox> =)))))
[16:18] <hrw> drivers/video/backlight/progear_bl.ko should be disabled for all ubuntu kernels
[16:19] <hrw> the device which uses it has transmeta cpu so 586 and 128MB ram
[16:19] <jwi> hrw: the xhci patches should be trickling down through -stable to quantal fairly soon
[16:19] <hrw> jwi: yep
[16:20] <hrw> jwi: good part is that disk owner lacks usb3 so will not have a problem. and she is using winxp 
[16:22] <jsalisbury> hrw, ok.  I'll take a look at those patches today.  It sounds like v3.6-rc2 has issues, so you can't test it properly?
[16:23] <hrw> jsalisbury: looks like usb is not working at all
[16:23] <hrw> jsalisbury: I do not have ps/2 keyboards anymore
[16:24] <hrw> jsalisbury: rc1 did not even brought 'cpus' online (i7-2600K) - only #0 and #7 were up
[16:25] <jsalisbury> hrw, ok.  we'll have to take a look at the patches and see if they can be easily backported to v3.5
[16:28] <hrw> jsalisbury: I applied them to 3.5.0-10.10 source without fuzz
[16:29] <jsalisbury> hrw, cool, so it should be an easy cherrypick.  I can build a test kernel this afternoon.  We can submit the patches to the kernel-team mailing list for consideration in to Quantal
[16:31] <hrw> cool
[16:34] <rtg> ogasawara, linux-meta-lts-quantal is still waiting to build.
[16:34] <ogasawara> rtg: sheesh, 9hrs!
[16:39] <hrw> reboot...
[16:39] <hggdh> bjf: quantal-on-quantal, VM, show failure on ecryptfs stress test, see https://jenkins.qa.ubuntu.com/view/SRU%20Kernel/job/sru_kernel-quantal-virtual_amd64-kvm-virtual/ 
[16:41] <hrw> re
[16:42] <hrw> jsalisbury: mainline kernel rev '9160338de92c0305329be5163a76f849806e83de' with 3.5.0-10.10 config booted fine
[16:42] <bjf> hggdh: that's "stress" not ecryptfs and you can ignore it though i'll make a note of it
[16:43] <hrw> jsalisbury: and usb3.0 disk works after plugging
[16:43] <jsalisbury> hrw, thanks for the info. Is that the only commit that would need to be cherry-picked?
[16:44] <hrw> jsalisbury: no, its today's HEAD of mainline ;D
[16:44] <hrw> jsalisbury: those 3 patches which I added to bug were extracted from mainline
[16:44] <hggdh> bjf: sorry for the mixup :-)
[16:44] <bjf> hggdh: not a problem
[16:44] <jsalisbury> hrw, ahh, ok.  
[16:44] <hrw> HEAD + http://paste.ubuntu.com/1159318/ config == http://paste.ubuntu.com/1159319/ dmesg
[16:44] <hrw> and disk works
[16:52] <jsalisbury> **
[16:52] <jsalisbury> ** Ubuntu Kernel Team Meeting - Today @ 17:00 UTC - #ubuntu-meeting
[16:52] <jsalisbury> **
[16:56] <bjf> hggdh: why don't you file that as a bug against the tests so i don't forget :-)
[17:00] <ppisati> isn't it meeting time?
[17:09] <smb> cking, apw I just did an upgrade of my quantal 64bit alternate kvm and I think there is still an issue with the default cirrus gfx and the X driver. Has one if you a similar vm to cross check?
[17:09] <cking> smb, not one handy
[17:10]  * ppisati -> EOD
[17:11] <smb> ok, maybe I just file a bug and let it be ignored like the other bugs ...
[17:12] <cking> smb, lemme know the bug number and I can see if I can reproduce it tomorrow
[17:12] <smb> cking, sure
[17:14] <cking> smb, aren't you supposed to be packing for your travels soon?
[17:14] <smb> cking, Did that already, but trying to pass on some other pointers to things I may forget
[17:15] <smb> Gosh, cannot even get a graphical desktop now after the latest upgrade
[17:15] <smb> Maybe its the great plan of dprecating unity2d before really having a replacement 
[17:17] <cking> smb,  I guess that's it
[17:18]  * cking uses server images for his VMs
[17:18] <smb> At least the part of now not getting a desktop, the clash between the cirrus drm module and the x driver was before
[17:27] <smb> cking, bug 1039648
[17:27] <ubot2`> Launchpad bug 1039648 in ubuntu "kvm/cirrus: X driver will not load when kernel module is loaded" [Undecided,New] https://launchpad.net/bugs/1039648
[17:27] <cking> smb, I'm nealy finished in installing, will try to repro
[17:28] <smb> cking, wow that was quick. I suppose you already had the iso
[17:28] <cking> smb, no I have a fat pipe to download the images
[17:29]  * smb forgot cking 's big ass pipe
[17:32] <cking> ..and speedy laptop too
[17:33] <rtg_> cking, are you actually getting 50mbit ?
[17:49] <rtg> apw, what iwlwifi errors are you seeing? 'fail to flush all tx fifo queues' ?
[17:50] <apw> rtg, http://paste.ubuntu.com/1159446/
[17:52] <apw> rtg, the end of the trace occurs after i rmmod and modprobe it again to get it working
[17:53] <cking> rtg, yep, I'm getting 50mbit 
[17:53] <rtg> apw, I'm only seeing 'fail to flush all tx fifo queues' once in awhile
[17:53] <apw> rtg, i think i have seen that in the past, but not recently
[17:54] <rtg> apw, you're getting some pretty ugly warnings even after the firmware restart
[17:54] <cking> smb, yep, I can reproduce that bug
[17:54] <smb> cking, Ah good, I mean bad, but good that it is the code broken and not me
[17:55] <cking> smb, alpha3 worked, updated it and it breaks
[17:56] <smb> cking, Right it as working at some point. And before todays upgrade the desktop came up when I blacklisted the kernel module
[17:59] <cking> smb, desktop is so heavyweight for my kind of testing, I normally just install server 
[17:59] <cking> ..so I would not normally encounter this bug :-/
[18:00]  * rtg -> lunch
[18:00] <smb> cking, Me too, but sometimes I want to do some desktop in box stuff. And some other people seem neither... 
[18:38]  * henrix -> EOD
[18:50]  * smb -> away
[19:12] <jsalisbury> hrw, re. bug 1039478 I have a test kernel building now with those three patches applied.  I'll post a link to the bug shortly.
[19:12] <ubot2`> Launchpad bug 1039478 in linux "WARN Successful completion on short TX: needs XHCI_TRUST_TX_LENGTH quirk?" [Medium,In progress] https://launchpad.net/bugs/1039478
[19:55]  * cking -> EOD
[20:18] <hrw> jsalisbury: I do not have hardware now - gave it back to the owner
[20:43] <verwilst> hi, i'm following up on a bugreport that handles a gfs2 kernel oops, very reproducible
[20:43] <verwilst> if there is a kernel developer/maintainer here who could have a look?
[20:45] <verwilst> https://bugs.launchpad.net/ubuntu/+source/gfs2-utils/+bug/1020207 for those interested
[20:45] <ubot2`> Ubuntu bug 1020207 in gfs2-utils "gfs2 kernel oops when deleting file on other cluster node" [Undecided,Confirmed]
[20:46] <verwilst> makes redhat cluster suite on ubuntu unusable when using gfs2
[20:50] <dupondje> verwilst: assigned it to 'linux' package, cause this is a kernel bug :)
[20:51] <verwilst> yeah, didnt really know who are what to assign :)
[20:51] <verwilst> so i thought, let's ask here ;)
[20:51] <verwilst> i've put my fair share of hours into this bug, so i hope i get some response :)
[20:52] <dupondje> is the bug already reported upstream ?
[20:53] <verwilst> it's fixed upstream
[20:53] <verwilst> as seen in one of my many comments :P
[20:54] <verwilst> it's fixed in 3.3-rc6 even
[20:54] <dupondje> but not in the stable 3.2.x updates?
[20:56] <verwilst> rah
[20:56] <verwilst> i messed up the bugreport
[20:56] <verwilst> can somebody set it back to confirmed? :)
[20:56] <verwilst> no
[20:56] <verwilst> not in 3.2
[20:56] <dupondje> you could spank upstream about that
[20:56] <dupondje> Kernel OOPS'es should get backported imo
[20:56] <dupondje> then they get automaticly in ubuntu also
[20:57] <verwilst> that's what i'm doing :)
[20:58] <verwilst> just applying pressure to both parties :)
[20:58] <dupondje> :)
[20:59] <verwilst> plz revert the fix released :) since i with my pleb powers cannot
[20:59]  * dupondje trying to contact mozilla developpers to fix a critical bug, now thats a no-go :(
[21:00] <dupondje> me neither ... i'm sorry :)