/srv/irclogs.ubuntu.com/2012/02/23/#ubuntu-arm.txt

GrueMasterTheMuso: Ping.  Have you seen this issue with alsa 1.0.25? http://paste.ubuntu.com/853433/ Lines 2-4.00:19
GrueMasterNot sure if it is a pulse issue or an alsa issue.00:20
GrueMasterdefinitely new to alsa-lib 1.0.25.  I just reverted to libasound2_1.0.24.1-4ubuntu1 and I don't get that error.00:38
Nekomaybe it's old or corrupted? :D00:38
GrueMasterNeko: New image.  Probably just need to kick pulse to work with some change in upstream alsa.00:39
GrueMasterGuess I'll need to see if I can reproduce this on x86, since it is pulse-alsa.conf.00:43
GrueMasterYea, I can reproduce it on an x86 VM instance with "sudo alsactl restore".00:47
GrueMasterWell, I have a solution to the pandaes audio issue.  Not perfect, but works within the current confines of alsaucm.02:28
GrueMasterWill upload tomorrow, when my eyes are no longer crossed.02:28
TheMusoGrueMaster: I just fixed some of those alsa-lib errors you showed me the other day, seems 1.0.25 changed some syntax for its conf files.05:07
NCommanderaramadxp images now building momentary07:45
=== LetoTheII is now known as LetoThe2nd
ppisatiogra_: why is it called the FINAL arm meeting? do you plan to commit mass suicide after it (e.g. Lemmings)? :)08:51
janimo`ppisati, the ARM team within Canonical is being dissolved and its memebers assigned to other teams in the company09:09
rbasakNCommander: around? Daviey asked me about http://people.canonical.com/~ubuntu-archive/testing/precise_probs.html - are you aware of the "linux-armadaxp-tools-3.0.0-1500 has no installation candidate" in there, or is that resolved now?09:17
NCommander!@#$#!@!09:18
NCommanderwe shouldn't be building linux-tools09:18
NCommander^- cooloney09:18
cooloneylinux-tools built failed under armhf09:19
cooloneyso i disable it in out linux-armadaxp package due to urgent upload request09:20
cooloneyit can be built under armel, i think09:20
NCommandercooloney: I don't think we even need it09:23
cooloneyNCommander: perf is very useful tool even for arm, we use it a lot for ti-omap409:24
cooloneyNCommander: and I saw some new tools was added from Marvell LSP09:24
NCommandercooloney: what would it take to fix it ?09:24
cooloneyNCommander: i guess it's related to compiler, but don't have much time to take a look09:25
infinitycooloney: Oh, can you merge my latest armadaxp upload into git?10:18
infinity(I really should request zinc access)10:18
infinitycooloney: Oh, I see in another channel that you're already on top of it.  Nevermind. :)10:20
=== Jack87|Away is now known as Jack87
=== juan is now known as Guest54640
=== Guest54640 is now known as jsol
GrueMasterogra_: Did you get my email on the pandaES audio situation?15:50
ogra_GrueMaster, yes, looks ok, lets upload it after someone (i.e. infinity) also eyeballed it quickly15:50
ogra_that udev rule is our hack anyway, adding another few lines for PandaES wont do any harm15:51
infinityogra_: I'm sure you and Tobin are enough of a review.  And I'm happy to not take the blame on Panda audio this cycle. ;)15:55
ogra_haha, ok, then i'll just upload after the call15:55
=== Ursinha is now known as Ursinha-nom
ogra_WHEEE!16:37
ogra_so chromium 17 seems to work just fine on armel16:37
ogra_looks shiny, but takes twice as long as FF to start16:37
=== Ursinha-nom is now known as Ursinha
infinityogra_: Yeah, no big shock there.  It's just as bloated, just differently. :P16:45
ogra_yep, but its great to have it after we had to live with 14 for ages16:46
GrueMasterNot seeing chromium for armhf (at least not on my mirror yet).16:48
ogra_GrueMaster, thats what we taljked about in the meeting :)16:52
ogra_only armel yet16:52
ogra_but it builds there for the first time in years16:53
GrueMasterah, ok.16:54
mahmohNCommander: armadaxp netinstall uImage resets, bug 93964517:07
mahmoh(which doesn't exist yet)17:07
NCommandermahmoh: the image fails to boot?17:07
NCommandergreaaaaaaaaaaaat17:07
mahmohright17:07
NCommanderGrueMaster: ^, can you take a look?17:08
mahmohubot2`: ?17:08
GrueMasterNCommander: Can I nuke the armada here or do you have data that you want to keep?17:08
mahmohfor me at least, I'm using the 1500.3 uImage and the net installer uInitrd now17:08
NCommanderGrueMaster: nuke and path.17:08
NCommander*pave17:08
GrueMasterok17:09
GrueMasterI'm in a meeting now, but will be soon.17:09
NCommandermahmoh: so it doesn't boot at all? You got the bootargs right?17:09
mahmohNCommander: it loads and tries to boot but resets immediately17:10
NCommandermahmoh: ugh, might be an issue with the mkimage commands17:10
infinitymahmoh: Gets into userspace, or dies in kernel init?17:10
mahmohdisclaimer: I did tftpboot it though but the checksum was fine so ...17:10
mahmohinfinity: dies before kernel init, right after load - you should be able to get to the bug17:11
infinityOh, shiny.17:11
infinity(And no, I can't get the bug, I don't have the hardware)17:11
mahmohinfinity: I meant see the bug (maybe) to see the details of where it fails exactly17:12
mahmohStarting kernel ...  interrupt request pc : [<0000803c>]          lr : [<006505cc>] sp : 005ffde0  ip : fffeffff     fp : 006d6ce8 r10: 005fff98  r9 : 00000bdc     r8 : 005fffcc r7 : 00000018  r6 : 005ffdec     r5 : 00000000  r4 : 005ffef7 r3 : 00008010  r2 : 000000f8     r1 : 00000bdc  r0 : 0040ff14 Flags: Nzcv  IRQs off  FIQs off  Mode SVC_32 Resetting CPU ...17:13
mahmohNCommander: the good news is the uInitrd looks like it's fine so far ...17:13
NCommandermahmoh: god thats special17:30
NCommandermahmoh: try it with a differentuImage17:31
GrueMasterYea, I see the same thing.  Will look at the kernel a little.17:42
GrueMasterDid the linux-armadaxp meta get bumped?  It seems to be pulling linux-image-3.0.0-1500-armadaxp 3.0.0-1500.2 instead of 3.0.0-1500.317:48
GrueMasterapt-get dist-upgrade is pulling the kernel in, but usually I should be able to just do apt-get update && apt-get install linux-armadaxp to update the kernel.17:50
GrueMastergrrr.  resolvconf install error.17:53
infinityGrueMaster: That trick only works in the case of ABI bumps.17:56
GrueMasterok.17:56
infinityGrueMaster: No ABI bump means no new linux-meta, means you need to upgrade the kernel image itself, not count on linux-meta to do it.17:56
infinity(But, as you note, apt-get upgrade gets it right)17:56
GrueMasterGuess I'm just used to more that a minor change with new kernels.18:01
GrueMasterOk, the kernel in ubuntu-ports/dists/precise/main/installer-armhf/20101020ubuntu113/images/armadaxp/ is busted.  Not even sure of it's origins.18:03
micahginfinity: sorry, I started to look at it, but lost my session (forgot to run in screen)18:08
micahginfinity: for chromium on armhf, if someone else has time to look, feel free18:08
infinitymicahg: Oops.18:08
infinityI'm pretty craptacularly busy today, but remind me tomorrow to kick off a build, so I can look at it Monday. :P18:09
infinityGrueMaster: Yeah, I think the SRU 2-week cadence with an ABI bump almost every time has conditioned people to assume that new kernel == new ABI.18:09
micahginfinity: that's about the same timeframe I have :)18:12
GrueMasterppisati: The mmap patch seems to have fallen out of the 2.6.35 omap4 kernel.18:18
ppisatiGrueMaster: uh? it was there18:21
ppisatiGrueMaster: wait18:21
ppisatiGrueMaster: gitweb on kernel.u.com is sloooooooowwww...18:23
ogra_use bzr !18:24
ogra_:P18:24
ppisatiogra_: i'll send a pull req to convert from git to bzr :)18:25
ppisatiGrueMaster: patches are still there18:25
ppisatihttp://kernel.ubuntu.com/git?p=ubuntu/ubuntu-maverick.git;a=commit;h=2956dd26b949343aca5581356bff6c1cf18a22c218:25
ppisatihttp://kernel.ubuntu.com/git?p=ubuntu/ubuntu-maverick.git;a=commit;h=ddc72fa76d3965ca3576cfbc7a1de6e1e4b4a68118:25
ppisatiM/omap418:25
GrueMasterMaybe in the tree, but the tests are failing.18:26
GrueMastersudo ./mmap-test18:26
GrueMasterCouldn't allocate the heap: 1902Mb18:26
ppisatidoh!18:26
ppisatiSRU kernel?18:26
GrueMasterThese are the tests included from the earlier bug (can't remember the bug number.18:26
GrueMasterYes, latest SRU kernel.18:26
ppisatiok, i'll check it out18:26
GrueMaster2.6.35-903-omap4 #31-Ubuntu18:27
ppisatiremind me where the test are located18:27
ppisati*these tests18:27
GrueMasterI'm trying to find the original bug.  It was fix released a while ago.18:27
=== calculu5 is now known as calculus
GrueMaster(and of course the bug isn't in the test source.18:28
GrueMasterhttps://bugs.launchpad.net/bugs/86129618:32
ubot2`Launchpad bug 861296 in linux-ti-omap4 "mmap fails to allocate 2030Mb heap on ARM" [Undecided,Fix committed]18:32
xranby_ac100GrueMaster: the original bug was that eclipse failed to compile on arm18:32
xranby_ac100since the eclipse build passed some herejava please use insane amounts of memory plx18:33
xranby_ac100and this made x86 builds pass while arm where failing due to this bug18:33
GrueMasterxranby_ac100: I think it was renamed or a new bug with specific info was created (see above).18:33
xranby_ac100im not sure there exist an original bug.. try ask doko18:35
xranby_ac100since it was clear mmap did not behave identical on x86 and arm i filed the above bug18:35
GrueMasterxranby_ac100: I already posted the link (see backscroll).18:35
* xranby_ac100 scrolls back18:35
xranby_ac100GrueMaster: before i rejoined?18:36
GrueMasterbug 861296 is what I was referring to.18:36
ubot2`Launchpad bug 861296 in linux-ti-omap4 "mmap fails to allocate 2030Mb heap on ARM" [Undecided,Fix committed] https://launchpad.net/bugs/86129618:36
GrueMasterIt has the mmap-test source.18:36
infinityxranby_ac100: haskell-src-exts and qtwebkit-source are still FTBFS with OOMing issues.18:38
xranby_ac100(07:26:30 PM) GrueMaster: Couldn't allocate the heap: 1902Mb  <---- how much swap are in use on your test system?18:38
infinityI'm not convinced that it's this particular issue, though.  I can get the mmap test to pass on some kernels where the builds will still fail. :/18:39
GrueMasterSwapTotal:      33554428 kB18:39
=== Jack87 is now known as Jack87|Away
xranby_ac100ok18:39
infinityI think I probably need to spend some solid time revisiting that next week and hunt down the root cause before we get too close to release.18:40
GrueMasterxranby_ac100: I run the same SRU test suite on all platforms and kernels.  This passed on the previous kernel for this platform.  Every image is installed from netboot & preseed.18:40
xranby_ac100GrueMaster: i am convinced18:41
xranby_ac100btw, thank you for running those tests18:41
GrueMasterI added them to my SRU regression testing specifically for this reason.18:41
infinitymicahg: Kicking off a chromium testbuild locally now, but yeah, my timeframe for looking at it is probably still Monday. ;)18:46
infinitymicahg: Kicking off a chromium testbuild locally now, but yeah, my timeframe for looking at it is probably still Monday. ;)18:48
* ppisati -> EOD18:53
GrueMasterppisati: Also, still missing headers from the dove headers package.18:54
GrueMasterlinux-headers-2.6.32-423 is virtually empty.18:55
GrueMaster(changelog & copyright info only).18:55
infinityGrueMaster: Yeah, didn't we agree that we just didn't deeply care anyway? :)18:55
GrueMasterinfinity: really should be a kernel team call on that, but in general yes.18:56
GrueMaster(and I had thought he was going to look at it).18:56
infinityGrueMaster: well, yes.  If they care, they can fix it.  But I'm betting they don't if we don't.18:56
infinityAnd I'd be shocked if we had any dove users.18:56
ogra_we do !18:57
ogra_GrueMaster !18:57
infinityogra_: I said users, not QA testers. :P18:57
ogra_heh18:57
GrueMasterI think there are commercial users, but maybe not for this specific kernel.18:58
infinityogra_: If GrueMaster's desk is representative of our average user, we're doing a pretty poor job of targetting, well, normal people.18:58
ogra_surely not on dove devboards18:58
GrueMasterI know HP has a little desktop system (I've seen one at a friends house).18:58
ogra_did they ever sell that armada ebox they showed in brussels ?18:59
infinityUbuntu: Linux for Crazy People?18:59
ogra_isnt it that ?18:59
GrueMasterAnd it has our kernel patches at least.18:59
ogra_from crazy people for crazy people18:59
infinityI'm pretty sure I have a t-shirt that says something about Human Beings.18:59
* GrueMaster is having issues with LP.18:59
infinityQA people don't qualify.18:59
GrueMasterThat's why my blog is titled "Terminal Insanity".19:00
infinity;)19:00
ogra_GrueMaster, about time you gain membership19:00
ogra_so it shows up on planet19:00
GrueMasterYea, I looked into that a while ago.  Lot of hassle.19:01
ogra_heh19:01
ogra_not really19:01
GrueMasterAnd I just haven't had the spare cycles to jump through the hoops.19:01
ogra_one wikipage that lists your contributions19:01
ogra_and one evening to attend the rmb meeting19:02
mahmohNCommander: GrueMaster: otherwise, the net-installer works fine as far as I can tell, need to preseed it now to enable nightly install verification testing19:02
NCommandermahmoh: I'll takea hammer to it19:02
ogra_you surely got enough bugs under your belt to qualify easily19:02
GrueMastermahmoh: I'll pass you my preseed.19:02
mahmohthx19:02
GrueMasterGuhhh.  What is going on with our web servers?  paste.ubuntu.com is slow, lp is almost non-existant.19:04
GrueMasterand email is also having issues.19:05
infinityGrueMaster: No issues here, perhaps you're suffering routing issues to the DC?19:06
GrueMasterMust be.19:06
GrueMasterping seems fine.19:07
GrueMastermahmoh: http://paste.ubuntu.com/854454/19:57
GrueMasterThat's the preseed I am currently working with.19:57
mahmohGrueMaster: thx!19:58
slangasekjanimo`: for this apr upload in the freeze queue, what's the definition of "recent enough" for the kernels?23:05
ogra_slangasek, hey, thanks for the welcome mail ! :)23:07
slangasek:-)23:07

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