/srv/irclogs.ubuntu.com/2011/12/02/#ubuntu-kernel.txt

ogasawarajsalisbury: hey, just saw your email.  if you don't mind, I00:25
ogasawara'll respond in the morning00:25
=== bladernr_ is now known as bladernr_afk
=== TeTeT_ is now known as TeTeT
=== ericm-Zzz is now known as ericm-afk
=== jibel_ is now known as jibel
=== ericm-afk is now known as ericm|ubuntu
=== gema_afk is now known as gema
ppisatimoin08:57
alkisgHi, is there a backported package that contains /lib/firmware/htc_9271.fw in 10.04? So that linux-image-generic-lts-backport-natty will be able to use that wireless card...09:51
alkisgIf not, can I install linux-firmware from Oneiric to Lucid?09:52
alkisgapt-file search /lib/firmware/htc_9271.fw in Lucid returns nothing... can I install Oneiric's linux-firmware to Lucid?10:06
ppisatiuhmmm11:25
ppisatianyway to workaround the maint-getabis script that tries to download a "non-existing" package?11:26
ppisatiherton: maint-startnewrelease uses Kernel Team <kernel-team@lists.ubuntu.com> instead of my user when creating a new release11:59
ppisatiherton: have you heard of it?11:59
hertonppisati, did you have DEBEMAIL and DEBFULLNAME set in your invironment, I think it uses that, let me check12:00
herton*do you12:00
ppisatithat's my laptop, i guess i don't have these set12:01
hertonppisati, yes, you must have DEBEMAIL="your email" DEBFULLNAME="your name"12:01
hertonexported12:01
herton*environment, sigh12:02
`-`#ubuntu ops are nazi fags. please remember to use your brain not that other bit of the anatomy the #ubuntu team appears to think is best.12:38
hertonppisati, you forgot to tag the ti-omap4 natty branch14:01
ppisatinope14:04
ppisatii forgo to push the tags14:05
ppisatiwait14:05
ppisatiherton: ok, all done14:10
hertonppisati, thanks14:10
ppisatibtw, since we don't pubblish anymore lucid/dove but i need it for maverick14:10
ppisatiis there a way to make maint-startnewrelease happy even with a non-existing kernel/package?14:11
hertonppisati, lucid is being pushed/uploaded later, so I think you can wait that to happen to prepare the maverick mvl-dove14:11
ppisatiactually the problem is maint-getabi that tries to download from a repository the package14:11
ppisatiah14:11
ppisatino wait14:11
ppisatii mean lucid/dove, i thought we don't pubblish it anymore14:11
dholbachhiya14:11
dholbachwhere is the current linux upstream bug tracker? is http://bugzilla.kernel.org/ still down because of the security issues?14:12
ppisatiherton: e.g. no maint-getabis look for Ubuntu-2.6.32-220.38 package14:12
hertonppisati, yep, but we still push lucid/dove to the repositories, I pushed your last rebase14:12
ppisatiah14:12
ppisatiwhen?14:12
ogasawaratgardner: thanks for doing the rebase14:12
ppisatiso the .deb should be ready soon14:12
hertonppetraki, ah ok, well about the abi, indeed we don't build the kernel now14:13
hertonppisati ^14:13
ppisatiyep14:13
tgardnerogasawara, had some idle cycles while waiting for lucid builds14:13
ppisatiso the only way to make getabis happy, is to hack it i guess14:13
=== bladernr_afk is now known as bladernr_
hertonppisati, hmm... yes, may be you can build manually somewhere the package, and copy the abi files?14:14
ppisatiyes14:14
ppisatithat was the plan14:14
ppisatiok14:14
dholbachdid anyone of you hear anything about some usb modules not being loadable with 3.2 due to 28 - ENOSPC (uvcvideo and snd-usb-caiaq in my case)? sorry for the vague question - but this the only bits I was able to find out up until now14:21
Razecgood morning14:21
brendanddholbach - hey, i got the same problem just yesterday14:22
brendanddholbach - well, sounds similar anyway14:23
dholbachbrendand, I filed bug 899165 and bug 899099 - which modules were it in your case? (usb-storage and my usb scanner, not sure which module that is, work though)14:23
ubot2Launchpad bug 899165 in linux "uvcvideo: Failed to submit URB 0 (-28)." [Undecided,Confirmed] https://launchpad.net/bugs/89916514:23
ubot2Launchpad bug 899099 in linux "snd-usb-caiaq fails to load" [Undecided,Confirmed] https://launchpad.net/bugs/89909914:23
brendanddholbach - my webcam stopped working14:26
brendandi saw 14:26
brendand~/.xsession-errors14:26
brendand  libv4l2: error turning on stream: No space left on device14:26
brendand  ** (cheese:26114): WARNING **: Error starting streaming on device '/dev/video0'.14:26
brendand  ** (cheese:26114): WARNING **: Could not negotiate format14:26
brendandbut didn't check syslog actually14:26
brendanddholbach - which model webcam have you got?14:27
* brendand can check the bug probably14:27
dholbachbrendand, Lenovo x220 - the camera should be "Bus 001 Device 005: ID 04f2:b217 Chicony Electronics Co., Ltd"14:27
brendanddholbach - yeah, saw that just now in lsusb. i have a z-star one, so it must be something more general14:28
dholbachapart from these two bugs I'm actually quite happy with 12.04, considering we're at 12.04 :)14:29
brendanddholbach - i guess you mean at alpha114:30
brendand:/14:30
dholbacherrr14:30
dholbachyes14:30
brendanddholbach - upgrades as usual are a different story though. big regression in Firefox for me, but only with the system i upgraded14:32
dholbachit seems I didn't run into this one14:34
brendanddholbach - i can't access bug pages in launchpad with javascript turned on14:35
dholbachmaybe an addon?14:36
=== ericm|ubuntu is now known as ericm-Zzz
* herton -> lunch15:08
* ogasawara back in 2015:20
dholbachjsalisbury, I tested the new kernel and replied to the two bugs - if there's anything else I can do to test, please let me know15:21
jsalisburydholbach, cool, thanks15:25
jsalisburydholbach, they both appear to be upstream bugs.  Unfortunately bugzilla.kernel.org is still offline, so we can't report it upstream yet :-(15:28
dholbachjsalisbury, and I guess there's nothing I can do to further help debug it? could you imagine that the bugs have anything to do with each other? both usb-related, both having error code -28 - or could that be a red herring?15:31
dholbachalso I don't know if you saw it earlier, but brendand seems to have a similar problem with a z-star webcam (not sure if he filed a bug)15:32
jsalisburydholbach, they could be related.  brendand, it would be great if you could review bugs 899165 and 899099 to see if you have the same or similar issue.15:34
ubot2Launchpad bug 899165 in linux "uvcvideo: Failed to submit URB 0 (-28)." [Medium,Triaged] https://launchpad.net/bugs/89916515:34
ubot2Launchpad bug 899099 in linux "snd-usb-caiaq fails to load" [Medium,Triaged] https://launchpad.net/bugs/89909915:34
dholbachjsalisbury, would it be useful to try older precise kernels?15:34
jsalisburydholbach, yes, that would be.  You could try back to 3.2rc1 or even the later 3.0 kernels.  15:34
* dholbach nods15:35
dholbachI'll let you know15:35
jsalisburydholbach, thanks.  I also noticed this message in your report:  libv4l2: error turning on stream: No space left on device15:35
dholbachjsalisbury, that was the user space message in .xsession-errors which came up after "uvcvideo: Failed to submit URB 0 (-28)"15:36
dholbachdiwic told me earlier:15:36
dholbach<diwic> #define ENOSPC      28  /* No space left on device */15:36
jsalisburydholbach, ahh, ok.  misleading message :-)15:37
dholbach(as you can imagine, I had no idea)15:37
dholbach:)15:37
jsalisburydholbach, It would be great if you could narrow down at what version the bugs happen, and don't happen.15:37
dholbachI'll do my best15:37
jsalisburydholbach, thanks!15:38
brendandjsalisbury - i definitely have 89916515:40
jsalisburybrendand, ok, can you mark that as "affects me too"15:40
brendandjsalisbury - think i did15:40
jsalisburybrendand, if we can narrow down at what version the kernel breaks, we may be able to do a bisect and create some test kernels.15:40
brendandjsalisbury - i asked yesterday if the kernel had changed in precise recently. the response was that it hadn't15:41
brendandjsalisbury - nearly certain it was working just a few days ago15:41
jsalisburybrendand, did you apply updates recently?  You could always look at your /var/log/dpkg logs to see if your kernel was updated.15:42
brendandjsalisbury - well, the laptop in question i use for iso testing. so it's fresh images15:43
jsalisburybrendand, ahh, ok.15:43
brendandjsalisbury - pretty sure 3.1 worked15:43
jsalisburybrendand, Do you have time, or a system to test the 3.1.0-2.3 kernel, and confirm the bug is not there?15:45
brendandjsalisbury - point me at the .deb15:45
jsalisburybrendand, https://launchpad.net/ubuntu/+source/linux/3.1.0-2.315:45
jsalisburybrendand, from there select your desired arch15:45
jsalisburybrendand, Under the "Builds" title15:46
brendandjsalisbury - got it now. should just need the generic i386 .deb file, right?15:47
jsalisburybrendand, yes15:47
jsalisburybrendand, Can you post your results to the bug, so we can keep track of them?15:49
Sarvattjsalisbury: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/899165/comments/415:49
ubot2Launchpad bug 899165 in linux "uvcvideo: Failed to submit URB 0 (-28)." [Medium,Triaged]15:49
jsalisburySarvatt, your the man ;-)15:50
jsalisburybrendand, ^^^^15:50
ogasawarajsalisbury: lets tag that 'rls-p-tracking' and milestone it for Alpha-215:52
jsalisburyogasawara, sounds good. 15:53
ogasawarajsalisbury: feel free to assign it to me so I keep track of it15:53
jsalisburyogasawara, will do.  Almost got another chance to bisect ;-)15:53
ogasawarajsalisbury: if you want, if there is a patch available, you can try building them a test kernel to use in the mean time15:54
jsalisburyogasawara, not sure if there is a patch, but I can check.15:54
Sarvattjsalisbury: /home/sarvatt/uvcvideo.patch on tangerine.buildd15:54
jsalisburySarvatt, great, thanks15:55
brendandyah, camera works again :)16:00
cyphermoxanyone know why applying some sysctls don't appear to affect all interfaces? for instance, if I try to change net.ipv4.conf.all.log_martians to 1; only that particular entry is changed, not the values for eth0.log_martians, wlan0.log_martians, etc. the one I'm particularly interested in is net.ipv6.conf.all.use_tempaddr (ipv6 privacy extensions)16:02
dholbachSarvatt, thanks a lot for digging the patch out - I hope it will fix bug 899099 as well16:04
ubot2Launchpad bug 899099 in linux "snd-usb-caiaq fails to load" [Medium,Triaged] https://launchpad.net/bugs/89909916:04
jsalisburybrendand, dholbach, I'm going to build a test kernel with the patch.  I'll ping you when its ready.16:08
=== dholbach_ is now known as dholbach
dholbachbrendand, did you build the test kernel yourself? :)16:36
brendanddholbach - i just downloaded the previous one that jsalisbury pointed me to16:38
jsalisburydholbach, brendand tested the  3.1.0-2.3 kernel, which did not have the bug.16:38
jsalisbury:-)16:38
jsalisburydholbach, brendand I have a test kernel building now.  Should be ready shortly.16:39
dholbachahhh ok16:39
dholbachthank jsalisbury, Sarvatt, ogasawara16:39
brendandjsalisbury - ok. let me know16:39
brendandthanks kernel folks16:39
dholbachif the patch fixes both bugs, I'll mark the other one as dup16:39
jsalisburydholbach, cool, thanks.16:39
* vanhoof waves16:40
ckingachooo16:43
ppisaticking: got a cold?16:44
ckingit's the weekend, which means yes. I seem to get one from the kids during the week and it hits me over the weekends16:45
bjfapw, about?17:00
jsalisburySarvatt, Just curious, where did you get the uvcvideo.patch file from?  I see the patch on LKML in the body of am email, but I don't see it in real patch format anywhere like you have.17:04
dholbachjsalisbury, http://www.spinics.net/lists/linux-usb/msg54992.html maybe?17:05
jsalisburydholbach, ahh ok. 17:06
ogasawaratgardner: I'm gonna upload -rc4 and stefan's kunmap_atomic patch he sent out yesterday.  anything else you want included?17:08
tgardnerogasawara, other ten the non-pae patch :) nothing I can think of.17:09
cyphermoxapw, tgardner: know why applying some sysctls don't appear to affect all interfaces? for instance, if I try to change net.ipv4.conf.all.log_martians to 1; only that particular entry is changed, not the values for eth0.log_martians, wlan0.log_martians, etc. the one I'm particularly interested in is net.ipv6.conf.all.use_tempaddr (ipv6 privacy extensions)17:10
cyphermoxI was expecting "all" to change the settings for the interfaces too; otherwise it's unclear to me what its use is... I can always open a bug though :)17:11
jsalisburydholbach, brendand, Build is done.  A 64 bit test kernel is available at: http://people.canonical.com/~jsalisbury/lp899165/17:11
dholbachthanks jsalisbury17:11
tgardnercyphermox, dunno, but I remember something like that plaguing me  in the 2.4 days17:11
dholbachI'll report back in a bit17:11
jsalisburybrendand, do you need the 32bit kernel?17:11
brendandjsalisbury - yes please17:11
dholbachjsalisbury, if it works, maybe that should be included in the -rc4 kernel as well? ;-)17:11
jsalisburybrendand, ok, I'll put the 32bit kernel in that same directory.17:12
dholbachbrb17:12
cyphermoxtgardner: fun. because right now there's this which could help working around the fact that ipv6 privacy extensions aren't turned on by .default. because eth0 is discovered before the sysctls are applied... otherwise maybe it would make sense to make use_tempaddr the default globally everywhere... as a patch?17:13
tgardnercyphermox, maybe you should ask the crew on netdev@vger.kernel.org . I haven't worked on that stuff for years. who knows what has changed.17:14
cyphermoxtgardner: ah, cool, thanks17:15
dholbachjsalisbury, no, the situation is unchanged17:15
jsalisburydholbach, so the bug is still there with this test kernel?  Can you run uname -a to ensure its the patched kernel?17:16
dholbachdaniel@daydream:~$ uname -a; dpkg -l | grep linux-image-3.2.0-317:17
dholbachLinux daydream 3.2.0-3-generic #7 SMP Fri Dec 2 16:44:58 UTC 2011 x86_64 x86_64 x86_64 GNU/Linux17:17
dholbachii  linux-image-3.2.0-3-generic            3.2.0-3.7                                  Linux kernel image for version 3.2.0 on x86_6417:17
dholbachdaniel@daydream:~$ 17:17
dholbachjsalisbury, ^17:17
apwcyphermox, i think thats a tricky one, if the interface differes from the default is that cause you changed it, or because its the default.  it would be detectable, but i am not supprised if an interface is set before you change it it has the old value.  it wuold not be hard to update all existing interfaces too i presume17:18
dholbacheither the patch does not fix it, or maybe the patch was not applied during the build?17:18
* cking kicks of some more tests, takes a break for 4 hrs17:18
apwjsalisbury, you should add clear identifiers to the version when you make a test kernel, so its more obvious if they have it17:19
apwI typically append ~lpNNNNNvNNNN to mine17:19
cyphermoxapw: for default, it's indeed what I expect, it would keep the old value.. but I was also expecting "conf.all.*" to propagate the change to all interfaces ;)17:19
cyphermoxwhich doesn't appear to work at all, for ipv4 or ipv617:19
dholbachbrendand, maybe you can test your i386 kernel: http://people.canonical.com/~jsalisbury/lp899165/ - it seems to be uploaded17:20
apwahh missread, that does sound like something that should override17:20
jsalisburyapw, dholbach, right, I added this to the changelog: ~lp899165v1 trying to figure out why its not applied17:20
brendanddholbach, jsalisbury - i'll try now17:20
cyphermoxapw: the only exception I've see as discussed somewhere is all.forwarding which might be working, I haven't tried17:20
jsalisburybrendand, hold off one minute.17:20
cyphermoxapw: otherwise, and now I've just found a netdev email on the subject, I was wondering if it would be crack to make the default be that privacy extensions are enabled17:21
cyphermoxapw: http://marc.info/?l=linux-netdev&m=130518706601466&w=217:21
jsalisburydholbach, your kernel version from uname should have ~lp899165v1.  I'm looking into why it does'nt now.17:24
dholbachjsalisbury, is fb80a1f805bc9470733a3f830c036cd0 you have as md5sum for the amd64 .deb?17:25
cyphermoxapw: perhaps this needs a little more testing, which I'll do now. the *value* of eth0/use_tempaddr isn't changed, but I suspect it might still be enabled anyway, if I am to believe the documentation files17:27
dholbachjsalisbury, I think I'll call it a day now, but if you update the bug, I'll have a look at it later on again and reply back17:29
dholbachI have somebody here who's impatiently waiting for dinner ;-)17:29
jsalisburydholbach, ok, sorry.  I'll have a new build for you later today.  17:29
dholbachjsalisbury, you rock17:30
dholbachthanks a lot for looking into this17:30
jsalisburydholbach, brendand, I'll post a note to the bug when a new test kernel is ready.17:30
dholbachhave a great weekend everyone!17:30
jsalisburydholbach, you too17:30
jsalisburydholbach, can you also uninstall the test kernel you just installed.17:30
cyphermoxapw: totally not17:30
dholbachbye :)17:31
jsalisburytgardner, does the ~/kteam-tools/buildscripts/ukb-make-release script require a patch to be git committed to be built?  17:32
jsalisburytgardner, Will it ignore a change to the tree if only the patch command is run?17:33
apwcyphermox, indeed17:33
brendandjsalisbury - so what happened?17:33
brendandjsalisbury - something wrong with the patch?17:34
jsalisburybrendand, there was a build problem.  I'll have a new test kernel in a few.17:34
apwjsalisbury, most of teh build stuff builds teh top of the tree and not your working dir17:52
sconklinis there a list maintained of which drivers are built-in for our kernels?18:29
tgardnersconklin, I think apw has it attached to one of our config blueprints18:32
sconklinok I'll look18:32
Sarvattjsalisbury: yeah what dholbach pointed out, sorry I left and just got back. did you get the kernel built?19:20
Sarvatti can whip one up real quick if not19:26
jsalisburySarvatt, its built now.19:29
Sarvattjsalisbury: i do:  patch -p1 < /home/sarvatt/uvcvideo.patch, fakeroot debian/rules clean, add whatever to the version string in debian/changelog, schroot -c precise-i386, DEB_BUILD_OPTIONS=parallel=64 AUTOBUILD=1 NOEXTRAS=1 skipabi=true fakeroot debian/rules binary-headers binary-generic for these one-off patches19:29
Sarvattoh ok19:29
jsalisburySarvatt, thanks for finding that patch.19:29
brendandjsalisbury - is there a test kernel now?19:47
jsalisburybrendand, yes, I'll have it uploaded in just a minute.19:48
jsalisburybrendand, There is an amd64 kernel available at: http://people.canonical.com/~jsalisbury/lp899165/19:55
jsalisburybrendand, I'm building some 32 bit kernels now as well.19:55
bjfbrendand, did you get 11.10 installed on that server yesterday ?20:02
bjfbrendand, that PowerEdge 295020:02
brendandbjf - i never managed to. something wrong with dhcp config. either the nic is busted in Oneiric, or something wrong with the setup20:29
brendandbjf - though i tested it recently with Lucid so don't see why that would be the case20:29
bjfbrendand: ok, thanks for trying. will you try again?20:32
brendandbjf - doesn't work with Natty either. It was only certified up to Lucid so we don't usually test it with newer releases20:40
brendandbjf - it's been around for about 4 years. what a trooper20:41
bjfbrendand: it shouldn't have regressed20:42
brendandbjf - well, i don't think the NIC is totally busted, but definitely somethings different because DHCP config works in Lucid20:44
* ogasawara lunch20:45

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