Kanohi, it seems aufs2 crashes with 2.6.34, could somebody update it00:34
Kanoseems like it was updated 10h ago00:36
=== MTeck is now known as MTecknology
* abogani2 waves07:45
=== hrw|gone is now known as hrw
smbGood $TOD08:19
* apw waves to smb08:19
* smb waves back08:19
ericm|ubuntusmb, tried a trivial merge of LSP 5.1.1 into mvl-dove, it's OK, so the merge conflict happens when rebasing mvl-dove onto master?08:20
* ikepanhc waves08:20
lagMorning :)08:20
* ericm|ubuntu waves to all08:20
* amitk waves east and west08:21
smbericm|ubuntu, It happened when I tried to either merge your branch into my security branch of mvl-dove and the same happened when I tried to export and import the patches08:21
ericm|ubuntusmb, where's your security branch of mvl-dove?08:22
ikepanhceh, I remember I do not identify for my nick, why I can speak here?08:22
smbericm|ubuntu, Thats the repo I gave you in my mail08:22
apwikepanhc, we had that fixed08:22
ericm|ubuntusmb, OK - let me try again08:22
ikepanhcyeah, I think so08:23
ericm|ubuntuapw, that possibly means I can use ericm again, let me try08:23
=== ericm|ubuntu is now known as ericm
smbericm|ubuntu, I have not officially pushed because I always fear last second changes. So I wait until its out08:23
ericmas long as that 'ericm' isn't here08:23
ericmsmb, no problem08:23
smbapw, Speaking of changes. Do I need to point you to the important mails I sent you? :)08:24
apwsmb, always08:25
* smb tries to remember08:25
smbapw,  One was about the sync problem with ext4 (or better block layer in general)08:26
smbapw, Oh, one thing was not really as mail to you, but at some point we probably want to virtually sit together with abogani2 to speak about -rt and -preemt08:28
apwsmb, yeah though i am tooo sleepy at the mo08:29
* smb knows that feeling08:29
smbI was up at 6am today to make an early visit to the tax agency to complete my declaration08:29
cooloneyapw: i met a question about our kernel module size.08:30
apwsmb gah that sounds vile08:30
apwcooloney, s'up?08:30
smbcooloney, Have you greeted it nicely?08:30
cooloneyapw: the size of module which was generated by 'make modules' is much bigger than the one in kernel package08:31
cooloneywhat's up apw and smb, hehe08:31
smbcooloney, You tried to strip it? 08:31
cooloneyapw and smb good morning08:31
smbcooloney, mornin man08:31
cooloneysmb, no08:31
cooloneyhold on a sec08:32
apwcooloney, yep it has full debugging symbols turned on; we make the debugging version and make -dbgsyms, then strip everything and make the normal .deb08:32
smbcooloney, Default options are set to include debug info. which is stripped upon packaging08:32
cooloneyroot@roc-desktop:/opt/git/Ubuntu/roc/ubuntu-lucid# ls ./debian/build/build-omap4/fs/nfs/nfs.ko -lh08:32
cooloney-rw-r--r-- 1 root root 5.7M 2010-06-01 13:34 ./debian/build/build-omap4/fs/nfs/nfs.ko08:32
cooloneyroot@roc-desktop:/opt/git/Ubuntu/roc/ubuntu-lucid# ls -lh ./debian/linux-image-2.6.33-900-omap4/lib/modules/2.6.33-900-omap4/kernel/fs/nfs/nfs.ko08:32
cooloney-rw-r--r-- 1 root root 383K 2010-06-01 13:34 ./debian/linux-image-2.6.33-900-omap4/lib/modules/2.6.33-900-omap4/kernel/fs/nfs/nfs.ko08:32
cooloneyapw and smb, yeah, i guess so. which script in our 'debian' directory is responsible for that?08:33
smbcooloney, I'd think the debian/rules.d/2-binary...08:33
smbcooloney, You can just say "strip nfs.ko" and should get the same size08:34
amitkhrw: the cross-compile fails because when we try to cross-build perf, the perf Makefile tries to test for capabilities (line 495 in http://paste.ubuntu.com/442658/)08:34
cooloneysmb: i checked that, but failed to find something related to that. 08:34
apwcooloney, its in debian/rules/2-*08:35
apwcooloney, search for INSTALL_MOD_STRIP08:35
smbIf the bcf does not censor that word...08:35
hrwamitk: thx08:35
cooloneyapw: oh, sh*t, i missed that. thanks man08:36
apwhrw you can skip making the tools package, do_tools=false something like that08:36
hrwapw: thx too08:36
amitkapw: perhaps we need a flag (CROSS_BUILD?) that skips that so that one can do _almost_ the entire build as a cross before upload?08:38
cooloneysmb: so interesting. i striped the .ko manually, but got a smaller one than the one in our package08:38
cooloneyroot@roc-desktop:/opt/git/Ubuntu/roc/ubuntu-lucid# arm-none-linux-gnueabi-strip ./debian/build/build-omap4/fs/nfs/nfs.ko08:38
cooloneyroot@roc-desktop:/opt/git/Ubuntu/roc/ubuntu-lucid# ls ./debian/build/build-omap4/fs/nfs/nfs.ko -lh08:38
cooloney-rw-r--r-- 1 root root 249K 2010-06-01 15:37 ./debian/build/build-omap4/fs/nfs/nfs.ko08:38
cooloneyroot@roc-desktop:/opt/git/Ubuntu/roc/ubuntu-lucid# ls -lh ./debian/linux-image-2.6.33-900-omap4/lib/modules/2.6.33-900-omap4/kernel/fs/nfs/nfs.ko08:38
cooloney-rw-r--r-- 1 root root 383K 2010-06-01 13:34 ./debian/linux-image-2.6.33-900-omap4/lib/modules/2.6.33-900-omap4/kernel/fs/nfs/nfs.ko08:38
apwcooloney, so look in the kernel itself and find out what that INSTALL_MOD_STRIP option actuall does08:38
smbOr it could be different compile options or optimizations08:39
apwamitk, possible, though i tend to use an sbuild environment to do the final upload build tests08:39
apwamitk, which can actually do the tools package08:39
amitkfair enough08:40
ericmapw, is sbuild capable of doing a cross-build on a PC?08:42
apwericm, yes, it uses qemu to do the execution in the chroot08:43
ericmapw, ah I see08:43
apwericm, its not very fast, but rather faster than the H/W08:44
ericmapw, btw - do you know who is maintaining the package libusb-0.1?08:45
ericmor to whom to ask this?08:45
ericmit looks like it's not been maintained though it's in main/ repo,08:46
amitkno autosync from debian?08:46
cooloneyapw and smb, you guys are right, INSTALL_MOD_STRIP=1 will call 'strip --strip-debug' instead of 'strip'08:46
apwericm, there does appear to be a libusb-1.0 as well08:46
cooloneyso it the size is litter bigger than without '--strip-debug'08:47
ericmI talked with smb yesterday, and so far got not much response to ubuntu-devel-discuss@lists.ubuntu.com08:47
ericmapw, they are two different things, apparently libusb-1.0 is better but a lot apps still linked with libusb-0.108:47
smbThe thing is that its likely a package which is only synced with debian08:47
ericmsmb, but it's in main/ repo, so this is true even for packages maintained in main/ (coming from debian?)08:48
ericmthere is a LP project page: https://launchpad.net/libusb08:49
smbThere is probably someone who got that package sort of on his list. But who and when that person looks...08:49
ericmjust curious if there is a team like us responsible for all the patches related08:49
ericmand I've no idea who's maintaining those branches on https://launchpad.net/libusb08:50
ericmthough it does look like progressing08:50
apwericm, finally found the source package, seems to be being updated from debian08:51
apwericm, so its belongs to 'all coredevs' in ubuntu08:51
smbapw, The fun patch is that two similar source packages one for libusb and another for libusb-1.0. Though both reference the same maintainer strings08:52
ericmcking, morning08:53
cooloneycking: morning08:53
smbHello cking 08:53
lagMorning cking 08:53
ericmsmb, yeah that's nasty and there is a bug being discussed to deprecate libusb-0.1 with libusb-1.0 (the latter is believed to be better), but it seems to be a lot work08:54
smbcking, I took the liberty of proposing you as my roomie for July. Hope this is ok with you08:54
ckingsmb,  that's good with me - thanks!08:54
* cking needs to get ear-plugs - the gas pipes outside are being dug up and pneumatic drills are being noisy today08:55
smbcking, No mumbling for you today08:56
ckingsmb, probably not 08:57
* apw notes the 2.6.35 merge window is now closed09:12
hrwapw: and according to phoronix tests 2.6.35 got performance regressions09:14
apwhrw nice, how bad09:14
smbProbably not too surprising for an rc109:14
hrwericm: most of apps which use libusb 0.1 can be built against libusb 1.0 or libusb-compat09:15
ericmhrw, let me do some experiment09:16
hrwericm: we did such transision in openembedded09:16
ericmhrw, ok09:17
ericmhrw, I don't seem to find libusb-compat in lucid, is this the correct package name?09:22
ericmhrw, or do I need to add some repo?09:22
hrwericm: I do not know is it at all in ubuntu09:27
hrwericm: libusb.sf.net basically09:28
hrwericm: http://www.libusb.org/wiki/LibusbCompat0.1 to be exact09:30
ericmhrw, not even in universe09:40
hrw167 packages rdepends on libusb009:43
ericmhrw, looks like it's a huge task09:47
hrwericm: libusb-compat is drop-in replacement09:47
hrw"As the compatibility layer implements the exact same ABI and API, no modifications to existing libusb-0.1-based applications are needed. You do not even have to recompile them. This compatibility layer is a drop-in replacement. "09:47
ericmhrw, en - looks like no debian package existing yet, need to first debianize it first09:47
hrwit uses autotools so easy task09:49
hrwdepends only on libusb-1.0-0, has pkg-config and libusb-config09:50
abogani2apw: Is ubuntu/ubuntu-lucid.git the official tree for Abstracted Debian (https://wiki.ubuntu.com/KernelTeam/AbstractedDebian) ?09:56
abogani2apw: Is ubuntu-lucid.git the official tree for Abstracted Debian ( wiki.ubuntu.com/KernelTeam/AbstractedDebian ) ?09:57
apwabogani2, the current versions in lucid and maverick are basically the same09:59
abogani2apw: These instructions (https://wiki.ubuntu.com/KernelTeam/AbstractedDebian section "How to Make a New Branch") said that we need to execute "sed -i -e 's/debian.master/debian.mybranch' debian/rules".10:00
abogani2apw: Why not simple replace "debian.master" in debian/rules (at lines 59 and 60) with $(DEBIAN) ?10:00
* apw looks10:01
* abogani2 shudder every times see someone using the sed's -i option...10:02
=== ericm is now known as ericm-afk
apwabogani2, ok that sed is just not needed any more, or wouldn't be if those masters wern't there which is just a bug10:06
abogani2apw: Ok. Sorry for disturb.10:09
apwabogani2, not at all, i'll get it sorted.10:13
stentenWhich mainline kernel branch should people use for testing kernel bugs? /current?10:13
apwstenten, /current represents the very latest build from linus' tip10:21
apwstenten, i might suggest the latest release tag, 2.6.34 first as that is more likely to be stable10:21
apwthan the first -rc of 3510:21
stentenok, thank you.10:22
* cking wonders how good Czech airlines are...10:26
amitkcking: the planes won't fall out of the sky10:27
ckingamitk, that's comforting to know10:29
* cking sees that amitk does his twice yearly blog post yesterday.. ;-)10:41
amitkcking: :) I don't get peeved easily, this was one such incident10:44
ckingi can understand that10:44
tseliotapw, smb: hi, do you know how to generate a directory in debian.master/abi/ for my customised kernel?11:10
tseliotinstead of passing skipabi=true11:11
apwtseliot, you can just create echo "1" >debian.master/abi/<version>/<arch>/ignore and ignore.modules11:12
apwfor each build arch11:12
tseliotapw: ah, thanks11:13
=== smb is now known as smb-afk
tseliotapw: it looks like the "clean" target in debian/rules removes that directory11:17
tseliotrm -rf debian.master/abi/2.6.32-22.33ppa111:17
apwtseliot, remember you need to fill in the _previous_ version number11:17
apwfdr printenv should tell you the version it thinks is previous11:17
tseliotapw: where do I fill it in?11:18
apwtseliot, no i meant you need to add the abi information to the previous version not the current version11:20
apwthe abi comparison is with the previous ABI, so it deletes and recreates the new one as you noted.  but the check is between the generated new one and the previous one11:20
apwand its there you need the ignores11:20
tseliotapw: aah, ok. Let me try again11:20
mdzI've got 50%+ CPU going to the "events/1" kernel thread, and many short periods of unresponsiveness12:02
mdzthis just started today12:02
mdzany advice for how I could track this down?12:02
mdz(this is lucid + proposed updates)12:02
=== oubiwann is now known as oubiwann_
apwevents/1 is a general purpose thread, you might be able to get an idea of what it is doing with sysrq-l12:19
apwmdz ^^12:20
ckingapw, that means any misbehaving driver or H/W could be the culprit?12:20
apwor indeed anything yes12:20
ckingmaybe worth looking at /proc/interrupts to see if any H/W is producing spurious interrupts then?12:21
=== smb-afk is now known as smb
mdzapw, thanks. call trace -> http://pastebin.com/6DYtVDpf12:22
mdzhas e1000e in it12:22
mdzwhich has nothing plugged into it at the moment12:23
mdz(and hasn't for over a week)12:23
mdzcking, nothing untoward in /proc/interrupts that I see; eth0 is only firing once every 5 seconds or so12:24
apwmdz, is it consistantly doing the same thing ?12:25
mdzapw, a few sysrq-Ls in a row all show similar call stacks12:27
=== oubiwann_ is now known as oubiwann
ckingmaybe also sanity checking this with ftrace function call tracing, http://www.mjmwired.net/kernel/Documentation/trace/ftrace.txt12:27
ckingecho 1 > /proc/sys/kernel/ftrace_enabled12:29
ckingecho 'function_graph' > /sys/kernel/debug/tracing/current_tracer12:29
ckingcat /sys/kernel/debug/tracing/trace > trace.log12:30
=== ericm-afk is now known as ericm
apwmdz, there does appear to be up to a second long loop in there for h/w interlock, though i might expect some diagnostics in dmesg abuot it12:36
mdzcking, that gives me a 2M log that I'm not sure how to interpret12:36
mdzapw, dmesg is pretty clean apart from the sysrq spam12:37
mdzdmesg |grep e1000 is all normal stuff12:37
mdzcking, it's mostly browser and Flash plugin wakeups, i915 activity and the like, though I do see similar e1000 activity in there as well12:40
mdzcking, should I send it to you?12:42
ckingmdz,  I may be able to spot something, so, yes, please send it to me12:44
mdzcking, you should have it now12:45
vishhi , i was asked to report a bug upstream , https://bugzilla.kernel.org/show_bug.cgi?id=16077 , not sure if i reported in the right component , could someone check?12:47
ubot2bugzilla.kernel.org bug 16077 in webcam "Drop is video frame rate in kernel .34" [Normal,New]12:47
ckingmdz, I cannot see anything glaringly obvious from that :-(12:51
ckingjust a load of chrome + flash activity really12:54
abogani2Is there a way to turn off temporarily the kernel's bugs mail notifications?13:04
smbvish, It looks reasonable to me13:06
vishsmb: thanks13:06
smbabogani2, I don't think so. Why? I guess most of them you get automatically though the team subscription...13:08
Kanohi, when will aufs2 refreshed in the .34 kernel? it does not work at all13:53
* ogra doubts that will happen at all for .3413:54
apwi had a look and there doesn't seem to be any big changes from .32 to .3413:54
ograapw, wasnt there a spec to fix that once and for all in .35 ?13:54
apwas there are changes for .35 compatibility i don't see it changing before A113:54
apwogra, union-mounts you mean, that depends on functionality testing13:55
apwi presume aufs works enough for the live-cds as noone has mentioned them not working13:55
Kanoi created a live image, it did not boot, but crashed aufs213:55
apwKano, bah, seems that we only made the first one last night, and hit the same issue14:01
Kanoi created my own yesterday too and therefore i know that it does not work14:02
ograKano, known, use union-fuse14:02
Kanoogra: thats too slow14:02
ograthats what we'll use in A114:03
Kanolike your first images without aufs, i patched it back, even modified an u image, was more than twice as fast14:04
Kanoaufs2 for .34 does not build externally - maybe some changes are needed14:05
bmfCheers everyone. I'm not sure if this is the right place but I've been searching and can't seem to get  to a conclusion... Does anyone know of a known bug with the suspend/resume on Core i7 (arrendale) + intel GMA HD platforms?14:11
ckingbmf, I've only seen a weird hibernate/resume bug on i7 CPUs, but that's fixed now14:14
bmfWell... I have a HP EliteBook 8440p and it seems to suspend just fine... but when I try to resume it just sits there with the fan on and the leds on... but the screen stays all black... I've had more problems with this laptop than I've had in the last 10 years :P14:16
ckingbmf, have you tried kernel boot parameter acpi_sleep=sci_force_enable?14:18
bmfno, I have not... let me try it give me a minute14:18
mdzcking, apw, is there anything else I can try to track down this events/1 activity? my palms are sweating from the heat :-)14:22
bmfcking: the behaviour is the same... black screen on resume14:23
ckingmdz, was this just from a kernel update and nothing else?14:23
mdzcking, not even a kernel update. it's been two days since I rebooted, and the problem only surfaced this morning.14:23
mdzI'm running 2.6.32-22-generic #33-Ubuntu14:24
* cking is a stumped - apw any ideas?14:24
apwhrm, its so hard to be sure, the routines showing up look like a wait is ongoing, which is kinda a cpu spin waiting on the TSC value14:26
apwmdz did you try inserting a cable (if thats an option)14:26
bmfcking: suspend to hard drive appears to work properly tho... Do you think I should post a bug report?14:30
ckingbmf, yep, I would14:33
JFobmf, drop me the bug number once you have it14:34
* cking reflashes a laptop - wish me luck14:34
bmferm... I have to say I have never reported a bug before... what info should I include apart from the problem description?14:35
apwbmf use 'ubuntu-bug linux' if its a kernel bug ... it knows the basics to collect14:36
bmfI have a patched kernel (for other, known, kernel problems with my hardware... that'll teach me not to buy bleeding edge) is that a problem?14:40
ckingbmf, makes it possibly harder to diagnose and debug14:41
bmfcking: aha! just noticed something14:55
bmfif I remove the module ricoh_mmc the resume works "almost" fine14:55
ckingbmf, what's that?14:55
bmfexcept for the fact that the laptop's screen brightness is 0 after resume14:55
smblag, git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-2.6.33.y.git14:55
smblag, That would be the git repo for 2.6.33.y. replace the 33 with 32 or whatnot for the other stable trees14:56
smbmsg lag That is an example of a random patch in the 2.6.33.y tree: commit 2b5a5d1697b2d4a96428ac6439b1d81660de379c14:57
smbAuthor: Herbert Xu <herbert@gondor.apana.org.au>14:57
smbDate:   Mon Apr 26 09:14:05 2010 +080014:57
smb    crypto: authenc - Add EINPROGRESS check14:57
smb    14:57
smb    commit 180ce7e81030e1ef763d58f97f9ab840ff57d848 upstream.14:57
bmfcking: then I have to do an echo 100 > /proc/acpi/video/GFX0/DD02/brightness14:58
ckingbjf, well, it may be worth adding that module into the MODULES=  list in /etc/default/acpi-support14:58
ckingbmf, ^^14:58
bmfyeah I was just reading about that same file14:59
bmfany idea about what I can do about the brightness? =P14:59
ckingbmf, not from anything I can recall at the moment14:59
bmfcking: thanks anyway :) i'm happy just knowing that it DOES resume... setting the brightness seems to be a smaller problem15:00
tgardnerlag, have a look at Documentation/stable_kernel_rules.txt in the kernel repo.15:01
manjogood morning all15:08
amitkmorning dude15:09
manjohow are you amitk 15:09
amitkhanging there15:09
manjomoin bjf15:10
manjonetwork manager icon dissapeared for me after an update 15:11
ericm_smb, rebased on top of your security tree, please try repull15:23
smbericm_, will do15:23
tseliotdoes anybody know why I'm getting this error when building a kernel? http://pastebin.ubuntu.com/442829/15:26
tseliotI can't find class_attr_version anywhere. Maybe it's something that's created by the preprocessor?15:27
smbtseliot, I'd more concentrate on the previous error15:33
* cnd has to go retrieve a handbag my sister left at a restaurant when she visited yesterday, be back in an hour or so... *sigh*15:34
=== cnd is now known as cnd-afk
smallfoot-why there is no 2.6.35-rc1 in ppa?15:34
tseliotsmb: that error makes little sense to me. Here's the full file: http://pastebin.ubuntu.com/442834/15:36
smallfoot-hey you assholes15:36
smallfoot-why the hell didnt you put 2.6.35-rc1 in the kernel ppa?15:37
tseliotKeybuk: troll ^^15:39
smbtseliot, Seems like one of these things that make only sense when you make it leave the precompiled versions15:39
tseliotsmb: what I did is replace drivers/gpu and include/drm with the ones from a more recent git branch15:40
Keybuktseliot: what do you want me to do about it? :)15:41
tseliotKeybuk: a kick in the butt? ;)15:41
Keybuktseliot: you want an IRC Op for that15:41
smbtseliot, The CLASS_ATTR_STRING likely expands into some struct probably containing the other part it complains about later.15:41
tseliotKeybuk: ok15:42
smbtseliot, Maybe that did not exist in the older sysfs definitions15:42
tseliotsmb: ok, I'll check that, thanks15:42
tseliotaah, they used CLASS_ATTR instead of CLASS_ATTR_STRING15:44
tseliotsmb: thanks a lot15:48
smbtseliot, No worries, I was just pointing a bit. :)15:48
JFoapw, you have ops in here yes?15:52
bjfJFo, i do, what you need?15:52
JFoplease read back comments from smallfoot- 15:53
JFothx :)15:53
apwhe seems to have gone quiet15:53
apwhardly worth your energy15:53
JFoyes, but there are ways to behave15:54
JFoand that isn't it15:54
cking'twas an interesting contribution to the channel15:54
ckingit only applied to ***holes, so it got ignored I suppose15:55
JFo:) which is of course why I noticed it ;015:55
bjf## Kernel team meeting in two hours16:00
apwJFo, indeed ...16:03
apwcking, its an irony that there was a locking problem and builds not occuring16:05
smallfoot-kernel meeting in 2 hours?16:05
smallfoot-then put 2.6.35-rc1 in ppa16:05
=== apw changed the topic of #ubuntu-kernel to: Home: https://wiki.ubuntu.com/Kernel/ || Maverick Kernel Version: 2.6.35 || Ubuntu Kernel Team Meeting - June-01 - 17:00 UTC
=== ogra_ is now known as ogra
lagsmb -^16:18
achianghow does one reset a file in bzr?16:25
achiangin git, i just say "git checkout foo.c"16:25
achiangah, bzr revert foo.c seems to work16:28
ckingyay, my SPAM filters are now behaving themselves after 4 days of training16:28
Keybuksomeone's been doing spam runs with my e-mail address16:28
KeybukI had thousands of "Undelivered mail" type mails in my INBOX today16:29
ckingmy filters marking 98% of my mail as junk last week - very annoying16:29
mjg59Keybuk: As long as you always use the same smarthost, that's a solved problem16:30
Keybukhow is it solved?16:30
mjg59BATV or similar16:31
Keybukshall have to look into that16:31
mjg59You tag each outgoing envelope from with a hash16:31
mjg59And then drop any mail with a null envelope from unless it's to a valid hashed address16:31
KeybukI guess it's easy to add to exim?16:32
mjg59There's a couple of tutorials16:32
mjg59I've been using it for a couple of years - the only proble is ezmlm, which looks at your envelope from rather than your From: to decide whether you're a list subscriber16:32
Keybukand I guess it doesn't matter that the replies are deliverered via canonical's mail server, whereas mails from me are sent through my own?16:32
mjg59Ha. That kind of matters, yes16:33
mjg59You need the outgoing server and the incoming server to agree on the hash salt, otherwise it won't work16:33
mjg59If you have localpart suffixes or prefixes, you could probably put the hash in there and filter locally16:34
mjg59The C mailserver should just pass them through in that case16:34
ograamitk, http://paste.ubuntu.com/442850/ so where's my kernel ? 16:35
ograamitk, (we have bootloader packages for XM now)16:35
kassahhow long does it ussually take for a kernel patch to go from proposed to production? days/weeks/months? ( https://bugs.launchpad.net/ubuntu/+source/linux/+bug/511066 )16:36
ubot2Launchpad bug 511066 in linux (Ubuntu Lucid) (and 1 other project) "ModemManager: HP ev2210 Sierra MC5725 not detected (affects: 2) (heat: 16)" [Low,In progress]16:36
=== ericm_ is now known as ericm-Zzz
kassahto stable I guess would be the better word for it16:36
ckingurgh, the gas board have dumped a load of soil on my path/lawn. back in 1016:37
smbkassah, Thats still queued to be uploaded after the security release, which unfortunately took longer than expected16:38
kassahsmb, got an estimate? I'm about to go on a trip... trying to guage if I should wait since it'll be out anyway... or if I need to drop in the proposed kernel on my laptop for the trip.16:39
kassahsmb, honestly I'm still amazed at how fast the bug has progressed =) I'm not in a rush to push it out... just an information finding mission =)16:40
kassahI just need it on the 5th =)16:40
smbI *might* get uploaded next week, if this is the patch I am thinking is in. Then likely needs good testing in proposed. But I cannot say for sure16:41
kassahsounds like I should just hunt down the proposed kernel and do that =)... thanks16:42
kassahthis patch is just for adding a usb id16:42
kassahso not really sure how much more testing I could give it16:42
smbIts not only about that one. You would test that the modem works. But there are a bulkload of other patches pending and it needs some time after uploading that to porposed to get a feeling whether there are probably regressiosn in it16:44
kassahah! =) makes sense =)16:44
=== hrw is now known as hrw|gone
JFoogasawara, you around?17:03
ogasawaraJFo: yep17:03
JFoplease see pw :)17:03
=== cnd-afk is now known as cnd
bjf## Kernel team meeting in 55 minutes17:06
mdzapw, cking, should I give up and reboot it then?17:07
ckingapw, ^^ any further ideas?17:09
mdzapw, oh, missed your earlier message17:09
mdzI can try connecting a cable, but it's not trivial and involves installing a battery (and waiting until the end of my work day)17:10
apwhrm, that sounds annoying.17:10
jjohansenbjf: did I send status to you for the meeting, I won't be there17:18
bjfjjohansen, nope (at least i didn't see it)17:18
jjohansenbjf: hehe, nope let me rephrase should I send to you :)17:19
bjfjjohansen, please do :-)17:19
manjobjf, meeting in 30mts ? 17:21
jjohansenmanjo: more like 3817:22
bjfmanjo, yes, I've already announced it twice this a.m.17:22
* manjo scrolls back17:22
manjoI was on my laptop earlier17:23
lapionthink there is definitively something wrong with the hangcheck timer code of the i915 driver. Seeing as a hangcheck can be provoked by simply overstressing other functions of the chipset, such as the hdd17:31
keesogasawara: yeah, let me know what I can do to help with bug 587888.17:33
ubot2Launchpad bug 587888 in linux (Ubuntu Maverick) (and 1 other project) "aufs oops in au_do_open() on maverick live system boot (affects: 3) (dups: 1) (heat: 24)" [Critical,Triaged] https://launchpad.net/bugs/58788817:33
apwogasawara, i am lookign at that now17:34
apwthough i suspect we are not not going to be usinf aufs for a117:34
=== kamalmostafa is now known as kamal
ogasawaraapw: ack. thanks for taking ownership of that.  is union mounts looking like a proper alternative?17:35
apwogasawara, too early to say at the moment17:37
apwthough the effort i am putting in now is teaching me how to do the testing needed for union-mounts17:37
bjf## Kernel team meeting in 5 minutes17:57
jjohansenbjf: you get my mail17:57
=== jjohansen is now known as jjohansen-afk
jjohansen-afksee you tomorrow17:58
ogasawaramanjo: 796df74 UBUNTU: SAUCE: Added quirk to recognize GE0301 3G modem as an interface.18:11
ogasawaraSeems this should be submitted upstream18:11
manjook will send it to stable18:11
manjoogasawara, had not realized I had done that patch :) 18:13
ckingbjf, I forgot to prefix my bios test weekly report with "Added:" - can you insert that into the report later?18:14
bjfcking, email me with the exact text you want18:15
ckingbjf, will do18:16
mdzcking, apw, rmmod e1000e stopped the spinning18:22
mdz(that's the good news)18:22
mdzthe bad news is, a subsequent modprobe e1000e has given me a spinning modprobe18:23
mdzwhich I can't kill18:23
apwmdz that sounds bad18:23
mdzhey, it stopped18:24
* mdz rmmods again and leaves it that way18:24
manjoJFo, what time is our mumble on bugs ? 18:27
* cking goes for food after 9 hours debugging suspend/resume issues18:27
=== cking is now known as cking-afk
JFomanjo, nowish18:27
apwogasawara, will be off to make some dinner, if you urgently need me, ping my mobile18:50
ogasawaraapw: ack18:50
ogasawaraapw: just going to do a quick test build on i386 and amd64 and then will upload18:50
apwogasawara, i'd be pleased to see some touch testing that i didn't break anything else ... so that sounds ideal18:51
ogasawarakees: will you be around for the next hour? wouldn't mind having you do a quick test on the kernels I build18:53
ogasawarakees: this is in context to an aufs update we're doing18:55
keesogasawara: yup, should be.18:55
ogasawarakees:  you're on amd64?18:56
keesogasawara: yup19:00
* manjo off to get some lunch19:28
apwogasawara, that branch got to you ok ?19:48
cndkees: I'm curious about the ptrace changes and the fit over changing the behavior for devs20:03
cndwhat behavior exists today that this would prevent?20:03
cndmy naivete says to just prepend sudo to strace and gdb20:04
keescnd: prevent which part?20:04
keescnd: right, that's one solution, but requireseducation20:04
cndkees: I'm just interested in hearing of any behavior that people do today that they wouldn't be able to with the change20:05
mjg59Attach strace or gdb to running processes unless they have admin access20:05
cndis it *only* the need to run things as sudo that worries people?20:05
keeswell, one uncommon scenario would be a shared devel system where devs didn't have sudo20:05
keesthe concern is surprising people with the change20:06
elmocnd: err, so, I like to be able to attach strace/gdb to existing processes and I don't have root on *all* the boxes I have access to20:06
keesin the shared system case the admin could just flip the sysctl, though20:06
elmoit's not just a dev thing, it's also an SA thing20:06
cndkees: if I was on a shared system, could I execute a second shell, and use the top level shell to ptrace any processes in the sub shell?20:06
mjg59kees: I'm still not clear how your use case isn't adequately dealt with by dropping CAP_PTRACE by default and then restoring it on an app by app basis20:07
mjg59I mean, you do have this fine-grained security system and all :)20:08
keesmjg59: i wouldn't want to give out arbitrary ptrace to everything20:08
elmokees: I thought mjg59 was suggesting the opposite?20:08
elmokees: i.e. give it to strace + gdb but nothing else20:08
ogasawarakees: I think I may have dropped, http://people.canonical.com/~ogasawara/lp587888/amd64/  can you give that a quick test20:08
cndelmo: I think the problem is that gdb or strace could be used maliciously20:08
mjg59So restrict anything with network access from running strace or gdb20:09
keeselmo: giving strace/gdb cap_sys_ptrace means anyone who ran it would be able to ptrace anything on the system20:09
mjg59kees: Uh. Anything on the system running as them, surely?20:09
elmokees: err, caps survive process exit and onto the user who spawned them?20:09
keesmjg59: that only works when everything starts confined20:10
mjg59Pretty sure this would all be trivial under selinux20:10
keesmjg59: cap_sys_ptrace gives you the ability to ptrace _anything_20:10
elmooh, I see20:10
cndto me, I don't see why I *should* be able to ptrace any given process today20:11
cndso I think this approach makes the most sense20:11
cndbut I admittedly no next to nothing about security20:12
elmocnd: ptracing processes is SA-101 20:12
elmoeven in situations you don't have root20:12
kikocnd, yo20:13
cndelmo: shouldn't you have root if you're doing high level SA though?20:13
cndkiko: hi20:13
elmocnd: depends on the situation; I've used strace and gdb plenty of times without invoking or even having root20:13
kikocnd, what touchpad do the dell latitude's have?20:13
mjg59Hm. Yeah, the capabilities model here doesn't seem to map terribly usefully20:13
mjg59Is AA really incapable of providing a default policy?20:14
cndelmo: ok, but did you *need* to ptrace in those situations and/or should you have had sudo access in those situations?20:14
keesogasawara: one sec20:14
cndkiko: no clue20:14
=== bjf changed the topic of #ubuntu-kernel to: Home: https://wiki.ubuntu.com/Kernel/ || Maverick Kernel Version: 2.6.35 || Ubuntu Kernel Team Meeting - June-08 - 17:00 UTC
keesmjg59: "user" caps would be niice here20:14
elmokiko: 20:14
elmokiko: synaptics ones, by and large, IME20:14
kikoheya elmo 20:14
mjg59kiko: Depends on the model20:14
mjg59Some synaptics, some alps20:14
elmokiko: but I'm working from a very small sampling20:14
keesmjg59: and yes, there isn't a concept of "default" for AA20:14
kikoI meant touchscreen20:15
mjg59kees: Huh. No wonder your life sucks :)20:15
kikothat was a pretty royal freudian20:15
elmocnd: yes, I 'needed' to ptrace to fix the problems I was having.  as to having sudo access, the people who run those servers would say no :)20:15
keesheh. AA is only a small part of my life20:15
mjg59kiko: Some Synaptics, some ntrig20:15
kikocnd, I'm asking because ubuntu is known to run on them, and so maybe if the synaptics driver is coming along then it might be worth it20:15
mjg59Uh. Actually probably no synaptics. Wacom or ntrig.20:16
kikomjg59, I've seen the n-trig ones, but I wanted to know if any were synaptics20:16
cndkiko: I'm guessing the MT driver you are referring to is the I2C driver20:16
kikocnd, I'm referring to your latest email on the subject ;-)20:16
cndthey *may* be using that interface, but I'd like to have proof before we do anything with the driver20:17
cndI would bet that they are actually wacom or ntrig as mjg59 saod20:17
mjg59synaptics are mostly common in embedded setups right now20:17
kikoI know for a fact that there are n-trig models as I have seen and used one20:17
cndkiko: yeah, we have ntrig drivers, and they work20:18
cndwe have wacom drivers, and they work for single touch20:18
kikocnd, we can investigate what samsung and TI use on their reference hardware if that helps20:18
cndbut not for MT20:18
cndkiko: the mobile team would probably be able to get that info faster I bet20:18
cndbut I can ask them20:18
kikothat doesn't mean that it'll apply to their actual OEMs but at least it's a step20:18
kikocnd, oh, I can ask samsung and TI myself!20:19
kikoI'm just asking because those are the only companies I've seen that have touch reference hardware20:19
kikowell, the only ARM SoCs anyway20:19
cndkiko: it wouldn't surprise me at all if such devices were using synaptics touchscreens20:19
kikowe will find out20:19
* bjf will be back in a bit20:23
=== jcrigby_ is now known as jcrigby
=== oubiwann is now known as oubiwann_
=== oubiwann_ is now known as oubiwann
* bjf is back21:00
=== MrPancake is now known as KB1JWQ
keesogasawara: that kernel fixes the aufs problem for me21:22
JFoogasawara, bug 588069 FYI21:30
ubot2Launchpad bug 588069 in linux (Ubuntu) "Lucid kernel is missing a large number of important ext4 bug fixes (affects: 3) (heat: 20)" [High,Triaged] https://launchpad.net/bugs/58806921:30
JFofrom Ted Ts'o21:31
apwkees, that is excellent news, thanks for that21:32
keesapw: thanks for the fix!  :)21:33
apwkees, i am just glad it worked21:34
keesmagic! :)21:34
bjfJFo, if it's lucid then it's smb21:51
ogasawarakees: sweet, thanks for testing.21:51
apwJFo, i've passed it through to the list for tommorow21:52
JFosaw your response apw21:52
apwit being kernel-fs21:52
* apw calls it a night21:53
ogasawaraJFo: for bug 288069, probably best to get it on smb's radar as it's for Lucid21:54
ubot2Launchpad bug 288069 in evolution (Ubuntu) "Calender seems to follow US DST, not local timezone's DST (dup-of: 281956)" [Low,Invalid] https://launchpad.net/bugs/28806921:54
ubot2Launchpad bug 281956 in tzdata (Ubuntu Hardy) (and 3 other projects) "evolution uses wrong date to switch to daylight saving timezone in timezone Europe/Brussels (affects: 3) (dups: 3) (heat: 54)" [Low,Invalid] https://launchpad.net/bugs/28195621:54
JFoogasawara, yep yep21:54
JFoyou mean 588... ;)21:55
ogasawaraJFo: oops yep, typo21:55
JFohee hee21:55
tgardnerogasawara, I'm getting close on the virtual flavour size shrink.  I'll try to post something later tonight.22:13
ogasawaratgardner: cool, thanks22:13
tgardnertime for a bike ride22:13
bjfpgraner, has something happened to voices.canonical.com? I can't seem to add a blog entry22:32
bjfogasawara, I took a quick look at those ext4 patches from the bug JFo mentioned, there are 57 patches in it and some are non-trival :-(, I'll talk to smb about them22:57
ogasawarabjf: yuck, sounds messy22:58
* ogasawara punched launchpad in the face22:59

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