/srv/irclogs.ubuntu.com/2011/11/23/#ubuntu-kernel.txt

=== ericm-Zzz is now known as ericm
=== Edgan_ is now known as Edgan
smbmorning08:16
sakthihi, can anyone point to right resources in order to form a good foundation in kernel programming..08:46
=== smb` is now known as smb
=== tkamppeter_ is now known as tkamppeter
ppisatito whom shall i ask for the qa-regression-testing?10:52
smbppisati, If I am not confusing people, maybe hggdh. And I guess he'll scream if I did confuse people... 10:55
ppisatiok, let me try10:56
ppisatihggdh: ping10:56
gemappisati: https://blueprints.launchpad.net/ubuntu/+spec/other-p-qa-qa-regression-testing11:44
gemappisati: nuclearbob is the right person to ask11:44
ppisatigema: k, thanks11:46
gemanp11:47
hggdhppisati: hello14:31
hggdhhum, I see all is resolved14:31
hertonppisati, just confirming, ti-omap4 rebased on oneiric doesn't change the abi?14:38
ppisatiherton: compilation didn't complain about it so...14:39
hertonppisati, ack14:39
apwppisati, that does assume its enabled !15:07
ppisatippisati: uh?15:11
* ogasawara back in 2015:31
=== ericm|ubuntu is now known as ericm-Zzz
lifeboyIf I have a display driver problem on an ltsp client and the screen keeps on initialising every 5-10 seconds, which process to I stop in the console (screen 0) to inspect what's going on?15:37
lifeboyIt's not ldm, startx or xdmcp15:37
apwbah15:39
apwlifeboy, i thought it stopped after 5 or so tried15:40
apwi assume its lightdm which is respawning X if its that15:40
ogra_lifeboy, better go to #ltsp to debug that ;)15:51
lifeboysoory, that was posted in the wrong channel! 15:55
lifeboyI meant to go to #ltsp :-)15:55
=== yofel_ is now known as yofel
hertonis anybody being able to clone from zinc? (I'm getting "fatal: The remote end hung up unexpectedly")16:32
smbit seems exceptionally crawling16:34
ppisatiyep, really slow today16:34
smbload of 7616:35
smbseems to be swapping to death16:35
hertoncloning through ssh:// seems to be going now16:36
* apw bets its those web-crawlers again16:45
smbapw is right16:45
smbor has been looking on #is :)16:46
apwheh not looking on #is16:46
apwam just talking on #is16:46
smbapw, I just had been completed to hassle him :)16:47
apwheh so you had, /me leans on them too16:49
ppisatiGrueMaster: ping17:03
GrueMasterppisati: Sup?17:04
GrueMasterppisati: You rang?17:08
ppisatiGrueMaster: yes17:15
ppisatiGrueMaster: http://people.canonical.com/~ppisati/linux-image-2.6.35-903-omap4_2.6.35-903.27~lp893190_armel.deb17:16
ppisatiGrueMaster: can you tell me if you still see the ptrace FAILures?17:16
ppisatiGrueMaster: wrt https://launchpadlibrarian.net/85419190/test-kernel-security.log17:16
GrueMasterAnd you want me to test it?  I'm flattered.  :P17:16
GrueMasterWill do.  Give me a few minutes.17:16
l3onHi all... does some ppa  exist backporting kernel 3.2.0 on 11.10 ?17:18
GrueMasterppisati: Ok, all ptrace errors now resolved.17:39
GrueMasterI'll have to talk with smb and kees about the /dev/mem and seccomp errors.  The script indicates they were available in armel kernels beginning with 2.6.35.17:41
GrueMasterMaybe they were in the omap kernel?  I don't have access to my beaglexm this week, but can check next week when I get home.17:42
ppisatiGrueMaster: for sure these were not available in maverick/omap*17:43
ppisatiGrueMaster: the exact kernel release after that (.36), had them17:43
ppisatianyway thanks17:43
GrueMasterbtw:  Tomorrow is a national holiday, andI will also be offline Friday.17:43
GrueMasterppisati: Ok, I will update the script then.17:44
ppisatiok17:49
GrueMasterGrrr.   When making changes to the test-kernel-security script for armel, kees also turned on a blanket test for SECCOMP (which didn't exist in <2.6.36 armel kernels). 18:16
* GrueMaster facepalms18:16
GrueMastertest_070_config_seccomp fails because not only was it not enabled until 2.6.36 on armel, it didn't exist as an option.  New false failure introduced on all <natty armel platforms (babbage3, dove, omap4).  Wheee.18:20
GrueMasterppisati: Was SECCOMP something that was backported to the x86 kernels?  I'm trying to figure out why it doesn't even exist on any of my armel kernels.18:23
* ogasawara bails. see ya'll next week.18:27
tgardnerGrueMaster, you might be a bit  late for Paolo. its 7:30P in Italy.18:28
ppisatino, i'm still around18:28
ppisatiGrueMaster: natty/omap4 has it18:30
ppisatiGrueMaster: it has entered linus tree in 2.6.36-rc718:31
GrueMasterppisati: Thanks.  Now need to wrap this new code change.18:34
jsalisburyherton, bjf, regression from 3.0.0.12 to 3.0.0.13: bug 89387618:37
ubot2Launchpad bug 893876 in linux "wrong display-resolution and no compiz-unity-plugin working since Kernel 3.0.0.13" [Medium,Confirmed] https://launchpad.net/bugs/89387618:37
hertonjsalisbury, ok, .13 is already released so if having a solution a fix will land in a later update18:39
jsalisburyherton, thanks18:39
jsalisburytgardner, Is there a way to match up commits reported by "git log" with what is in the ~/debian.master/changelog for the same kernel tree?18:43
hertonjsalisbury, it seems 893876 is not a kernel regression. On the working previous kernel, the user has nvidia proprietary driver loaded, but not on new kernel for some reason, thus the vesafb which is always loaded seems to be used18:45
herton(checking both dmesgs on the report)18:45
jsalisburyherton, ahh, ok.  18:45
tgardnerjsalisbury, typically all commits in a version go into the changelog unless they are administrative (and therefore ignored)18:47
apwjsalisbury, yes the commit titles are the one liners18:47
jsalisburytgardner, apw, great, thanks for the info!18:47
jsalisburyapw, so I can grep for the one liners in the changelog18:48
tgardnerjsalisbury, if you know the commit SHA!, then you can 'git describe --contains <SHA1>' to get the most recent tag.18:49
apwjsalisbury, yep18:50
jsalisburytgardner, apw, cool, thanks for the help18:50
jsalisburyherton, for 893876, do you think they need to configure the new kernel to use the nvidia driver, it won't happen automatically?18:53
hertonjsalisbury, no idea, I don't know how nvidia is installed, or how the user installed, I know we have jockey to enable the binary drivers, don't know how that works out in detail.18:55
jsalisburyherton, ok.  I'll update the bug and report your findings.  Thanks for reviewing.18:56
hertonjsalisbury, ok, it's worth asking the reporter about how he installed, and checking if nvidia driver is built/enabled with the new kernel18:56
jsalisburyherton, ok, will do.  thanks19:00
* herton back in 30 min-1 hour (errand)19:25
cemchi. I'm trying to rebuild the Hardy kernel package. I've applied a one-liner to fix some bug and I'm trying to build it using pbuilder (or through the PPA), but I'm getting the following: https://launchpadlibrarian.net/85751961/buildlog_ubuntu-hardy-i386.linux_2.6.24-30.96%2Bcemc1_FAILEDTOBUILD.txt.gz . I'm not sure how the ABI stuff works. I've just renamed the directory in debian/abi/ and changed the abiname from 29 to 30. I also read the kernel b20:01
cemcuilding wiki page but I don't see what I need to do20:01
cemcdoes anybody have a minute to give me some pointers?20:01
hertoncemc, for a test build, just bumping the abi should be enough, no need to rename the directory in debian/abi, I guess it's what breaking your build21:05
cemcherton: you mean in debian/changelog?21:07
hertoncemc, yes21:07
cemcherton: the kernel in Hardy is 2.6.24-30.96, and I just added a +cemc1 to that, I didn't want to change the version so I know it's the same kernel just with my changes21:08
cemccuriously enough, the last one -29.95 did work like this21:08
hertoncemc, what would be the right way to do it, starting from a pristine 2.6.24-30.96, is something like "debian/rules startnewrelease", get the abis from the current release (you can use "./debian/scripts/misc/getabis 2.6.24 30.96"), then remove old abi dir in debian/abi, bump the abi, apply your patch(es)21:16
hertonbut I don't know if you need to do this for a test build, I would expect just applying the patch on top of .30 should work21:16
* herton -> eod21:20
=== BenC__ is now known as BenC
* tgardner -> EOD22:31
shabbledoes anyone have any experience with writing/mangling usb(-serial) drivers? I'm in a weird place where cdc-acm is needed to init the device, but I have to unload and replace it with usbserial (generic) otherwise it locks up the port.23:00
shabbleit's an MSP430 test board: RF2500 - there's plenty of threads about it being pretty horrible, but sadly nothing that seems to actually fix it.23:01

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