/srv/irclogs.ubuntu.com/2012/11/28/#ubuntu-kernel.txt

=== smb` is now known as smb
=== yofel_ is now known as yofel
=== stefano is now known as Guest75762
henrixrtg: have you reached my email about the build failures?12:09
rtghenrix, just responded12:21
henrixrtg: ack, thanks. i'll check the 3 git trees in a minute12:22
henrixrtg: ok, they look fine. thanks12:25
=== amitk is now known as amitk-afk
ogra_rtg, i belibe that kexec hack stuff is for making the kernel work as second stage kernel for some android kexec boot kernel, i think the guy tries to achieve dual boot (without taking into account that the userspace isnt capable of supporting it and the next kernel upgrade will overwrite the kexec loader)13:04
ogra_*belive13:04
rtgogra_, perhaps you should respond thusly on the mailing list.13:06
ogra_well, lets see what the submitter says13:06
ogra_i'm not thrilled by having to carry an extra patch for a feature we dont support13:07
rtgnor am I, hence my questioning of the intent for his patches13:07
ogra_the only proper way of dual booting i see is to use the recovery partition for the ubuntu bootimg and properly adjust flash-kernel for that, sadly nobody wants to work on that and poeple coming up witgh dual boot stuff come up with something insane like that one ... 13:09
ogra_well, shrug13:09
hertonrtg, I got a problem with missing firmware when testing the 3.5 stable candidate. The issue is the firmware got removed from linux-firmware, because of being already on our linux-image packages. Not sure if this warrants some change or not for mainline kernels...13:17
rtgherton, which one?13:17
hertonsome firmware was missing for bnx2 on "gonzo" (our lab):13:17
hertonW: Possible missing firmware /lib/firmware/bnx2/bnx2-mips-09-6.2.1b.fw for module bnx213:17
hertonW: Possible missing firmware /lib/firmware/bnx2/bnx2-mips-06-6.2.3.fw for module bnx213:17
hertonrtg ^13:17
hertonin my case the missing one was bnx2-mips-09-6.2.1b.fw13:18
rtgherton, did one of the stable patches update the firmware version ?13:18
hertonrtg, nope, kernel doesn't carry this firmware13:18
apwherton, but our kernel for Q does ?13:18
hertonapw, yes13:18
apwcommit ?13:18
rtgapw, herton: its a PXE boot essential file13:19
hertonapw, 23bae1788f8c6f216f0261adb6260f57debbf41913:19
apwrtg, makes sense, does that mean we have a sauce patch pulling it into the kernle13:19
rtgapw, likely. I'll have to check13:19
hertonrtg, yeah, I thought it could be the case, so didn't bother very much13:20
hertonbut just reporting13:20
pgranerapw, rtg: any idea about this? http://paste.ubuntu.com/1394421/   hint look around line 123313:20
rtgherton, this is vanille 3.5 stable ?13:20
rtgvanilla*13:20
apwrtg, seems to be a firware update patch in fact13:20
hertonrtg, yep, we noticed it on the lab because the machine didn't came up (not network), when testing my 3.5 stable update13:21
herton*no13:21
hertonthe dell one which uses bnx213:21
rtgherton, this is _our_ quantal kernel then ? WHat changed, or did it never work ?13:21
apwpgraner, looks like we lost the pcie bus ...13:21
rtgpgraner, looks kind of bad13:22
pgranerapw, not quite the drives are still online just readonly13:22
hertonrtg, no, it's not the quantal kernel, it is a stock 3.5 kernel, with my stable update on top (3.5.7.113:22
pgranerapw, its a storage cab with 4x 3 TB drives13:22
rtgherton, ok, that makes sense. there is a reason we carry bnx2 (and bnx2x) in the kernel package.13:22
pgranerapw, and it suddenly flipped to ro13:22
pgranerapw, digging around I found that, at the end of the dmesg you'll see where ext4 flipps em13:23
rtgpgraner, consistent after power cycle?13:23
pgranerrtg, after alot of disk I/O  i.e. moving/deleting files13:23
pgranerrtg, its the 2nd time its done it to me13:24
rtgpgraner, this is sort of what happened on tyler, e.g., going RO after lots of disk errors.13:24
sjanc_hi,  would you consider enabling CONFIG_NFC_LLCP in ubuntu kernel 3.7 ?13:24
rtgpgraner, wiggle some cables ?13:24
apwpgraner, that seems to show a bunch or recovered error on the pci bus i think, then it stops and times out13:25
rtgsjanc_, email it to the kteam list13:25
pgranerrtg, I'm going to boot single and fsck the drives and see if the issues go away, if not I guess I'll upgrade to Precise and see how it goes13:25
sjanc_rtg: ok13:25
rtgpgraner, this is your everyday RAID set, right ?13:26
pgranerrtg, yea, it gets pounded heavy, is a md raid0 set13:26
rtgpgraner, since its external I'd have a close look at cabling13:26
hertonrtg, ok, just wanted to confirm that (needed on main kernel due to pxe). I know that also linux-firmware was stripped to save space, and the firmware needed was removed from it. It makes testing not possible on the machine on the lab needing it, since we don't carry anymore the firmware also on linux-firmware. But I can live with one machine not being able to do testing for mainline/stock kernels.13:26
pgranerrtg, ack13:26
rtgherton, ack13:27
rtgogra_, well, there you have it: https://lists.ubuntu.com/archives/kernel-team/2012-November/023407.html13:41
ogra_yeah13:41
* ogra_ is already looking at the xda thread13:41
ogra_"So what does the MultiROM does?13:42
ogra_It acts as the boot manager, it manages status of device - what ROM is currently in /data, which boot.img is actually in boot partition. It changes android *.rc files so that they do not mount /system, /data and /cache by themselves."13:42
ogra_oh my, i guess they would try to do the same in ubuntu then too13:42
rtgogra_, I assume we'd only do this for raring ? the kernel and user space mods I mean.13:43
ogra_its a 14.04 thing13:43
* henrix_ -> late lunch13:43
ogra_but i'm not sure we'll stay with nexus7 for 13.1013:43
rtgogra_, you mean _after_ 13.10, right ?13:44
ogra_essentially i would like to leave the images alive as long as we're not short on image builder resources even beyond 14.0413:44
ogra_rtg, nope, i mean *for* :)13:44
ogra_we might switch to another device for 13.10 13:45
rtgogra_, um, I thought we were committed  to N7 for at _least_ 13.10, but now I realize I'm think 13.04.13:45
rtgok, I get it13:45
ogra_we will release 13.04 for the nexus13:45
rtgminor dislexia13:46
ogra_we might switch to new hw after that13:46
rtgagreed13:46
rtgogra_, so the user space hacks to support dual boot will take a bit of work ?13:47
ogra_surely a change in how flash-kernel handles kernel updates13:47
ogra_(or initrd)13:47
ogra_in any case i wouldnt go for that kexec method since it changes userspace from initrd apparently and if they do that on android i would expect that this will have to happen on ubuntu too13:48
ogra_"2. Multi-booting ubuntu13:49
ogra_This is a bit different. First, ubuntu's boot.img is flashed and device is rebooted. Then, I have to move all folders in /data except "media" somewhere else, and move Ubuntu's root into /data. This is very dangerous and hacky, but I cannot just mount -o bind the ubuntu folder, because it has it's own init and would remount it again."13:49
ogra_lovely13:49
rtgegads13:49
ogra_so lets just skip upstart altogether :P13:50
rtg:)13:50
rtgogra_, I might be inclined to carry the kernel patch if it is truly benign, but not until after I get janimo's review. I don't think the user space mods will ever make it.13:51
ogra_yeah13:51
ogra_adding kexec support surely isnt bad if it doesnt influence any nrmal features13:52
janimortg, ogra-cb I blame Pan for not sending out my email yesterday as an answer to that patch13:54
janimoit was not a review, more like, 'does it depend on user space patches, and why should this be a nexus7 only patch as opposed to other-ARM too or Ubuntu SAUCE"13:54
ogra_well, it seems to heavily depend on userspace changes according to the forum13:55
ogra_(to use it)13:56
janimoogra_, right that's why I asked13:56
janimobut my mail is not sent out13:56
ogra_but as long as it not breaks anything 13:56
ogra_(when not using it i mean)13:57
janimoit's not the asm that looks intimidating, but I don't know the details of very early boot to know what this could endanger13:57
janimobut still it looks just fine13:57
janimosmall non-intrusive13:57
rtgjanimo, shall I expect to see your email  response soon ?14:06
janimortg, yes, I'll resend it14:08
apwrtg, could it be caught in the approval queue ?14:17
janimoapw, if it's the mail you're asking it is likely on my side. I switched to Pan to read newsgroups months ago, but only now had to send one14:18
rtgapw, it wasn't there first thing. I'll check again.14:18
janimoit briefly popped up some errorish msg which I ignored then but apparently the mail is gone 14:19
rtgapw, nope, only one email in the moderator queue14:19
=== amitk-afk is now known as amitk
* ppisati goes out for a bit15:30
joaquinHello everyone! I have an issue with my Laptop when I use Ubuntu, it gets too hot!16:07
joaquinI think it's a Kernel issue, but I don't know where to ask, here? or to the Kernel developers?16:08
apwjoaquin, have you filed a bug with all the details?16:08
apwand how hot is "too"16:08
joaquinThere is a bug a launchpad https://bugs.launchpad.net/ubuntu/+source/linux/+bug/763477 but the bug at bugzila.kernel.org is closed due to inactivity16:09
ubot2Launchpad bug 763477 in linux (Ubuntu) "lenovo y560 overheat. no /proc/acpi/fan" [Undecided,Confirmed]16:09
apwseems to be open to me16:09
joaquinapw, at launchpad is open, but at bugzilla.kernel.org is closed16:10
apwjoaquin, so the fan interface not appearing does not mean much as the bios is responsible for contorlling it anyhow16:11
apwjoaquin, just how hot does the machine get16:11
rtgapw, Fetching omap(armhf)...extracting linux-image...GCC: (Ubuntu/Linaro 4.7.2-11ubuntu1) 4.7.2...done.16:11
rtgFetching generic(amd64)...extracting linux-image linux-image-extra...GCC: (Ubuntu/Linaro 4.7.2-12ubuntu1) 4.7.2...done.16:11
rtgFetching generic(i386)...extracting linux-image linux-image-extra...GCC: (Ubuntu/Linaro 4.7.2-12ubuntu1) 4.7.2...done.16:11
rtgWARNING: inconsistant compiler versions detected16:11
joaquinthe fun work at full speed all the time, it reaced 78-80 C degrees16:12
rtgapw, did we just straddle a compiler change ?16:12
apwrtg, oh well, its during devel and not a milestone so ... *shrug*16:12
apwlooks like it indeed -11 -> -1216:12
apwjoaquin, so with the fan at full it overheats ... so the bios is controlling the fan16:13
apwvanhoof, have we had a lenovo y560 pass through our hands at all ?16:13
joaquinapw, what does it mean that the bios is controlling the fun?16:14
apwjanimo, on which release/kernel version does this behaviour exhibit and which (if any) was it ok16:14
apwjoaquin, normally linux has no controll over fans, if the fans are changing then the bios is doing something, it is seeing the heat and increasing fan speed to compensate16:15
joaquinapw, OK, but the reason it's getting to hot, could it be a Kernel issue?16:16
apwjanimo, well it could be possible, we could be upsetting something in the bios during initialisation16:17
hertonapw, I think you can commit that fakeroot fix to mainline-build-one16:17
apwjoaquin, then again it cuold just be a bios bug, and windows never could drive the machine hard enough to trigger the issue16:17
apwherton, did i not ?16:17
hertonapw, not yet :)16:17
apwherton, done16:19
hertonapw, pulled it's here now, thanks16:20
joaquinOK, so what can I do?16:20
apwjoaquin, well we need to try and eliminate some things.  if the machine has windows on you could see if windows has the same issue16:21
apwif so its BIOS and we can do little16:22
apwjoaquin, if it used to work on an older Ubuntu, working out the latest one it was ok on/first one it is not ok on would help16:22
apwjoaquin, as we have had lenovos where the heatpipes have dried up and the issue was h/w just age16:22
joaquinapw, with windows it works fine, no problem at all16:24
apwjanimo, and what temperature does it go up to if you run it hard16:25
apw(in windows)16:25
apwand what temperature when idle16:25
apwand the same for ubuntu, and you still have not told me which release you are running even16:25
joaquinSorry, Ubuntu 12.10 64bits16:25
apwhave you run any older Ubuntu's on this machine16:26
* apw orders up some teeth to save the effort of pulling them16:26
joaquinapw, no, I only used ubuntu 12.1016:27
apwjsalisbury, yo ... i think i have a candidate for a 'rough bisect' for you :)16:27
joaquinapw, in responde to a previous question, with some games Windows also ges hot, but I do not know the temperature, cos I never meassured16:28
jsalisburyapw, cool, I like it rough16:28
jsalisburyapw, bisects that is ;-)16:28
apwjsalisbury, joaquin here could do with testing out like a P kernel or somethign to see if it this heat problem is new16:28
apwjoaquin, so is this "getting hot" when the machine is idle in Ubuntu or only when busy16:29
joaquinapw, iddle16:29
apwjsalisbury, might you be able to help joaquin test an older kernel for me16:30
ckingapw, that lenovo has Nvidia graphics, I suspect that's why it's getting hot16:32
joaquincking, apw, no NVIDIA, it's an older model with ATI RADEON16:32
jsalisburyjoaquin, are you the original reporter of bug 763477 ?16:33
ubot2Launchpad bug 763477 in linux (Ubuntu) "lenovo y560 overheat. no /proc/acpi/fan" [Undecided,Confirmed] https://launchpad.net/bugs/76347716:33
joaquinjsalisbury, no16:33
apwcking, ahh could indeed be the GPU not being clocked well .. sigh16:34
jsalisburyjoaquin, ok.  It's probably best for you to open a new bug, so we have all your hardware details16:34
sforsheethe open source radeon driver is pretty terrible at power management, so it would likely contribute16:34
jsalisburyjoaquin, that bug could have similar symptoms, but be completely different if you have different hardware16:35
jsalisburyjoaquin, if you open a new bug, I can work through this with you in the bug, that way no information is lost and others can search for it.16:35
joaquinjsalisbury, OK, I'll open a new bug tonight (I'm at work right now)16:36
ckinghrm, that bug lists VGA compatible controller [0300]: nVidia Corporation GT216 [GeForce GT 240M] [10de:0a34] (rev a2) (prog-if 00 [VGA controller])16:36
jsalisburyjoaquin, great.  I review all the new bugs, so I'll keep an eye out for it. 16:36
joaquinGREAT thanks!16:36
jsalisburyjoaquin, np16:38
apwjoaquin, use 'ubuntu-bug linux' from a terminal window to make sure we get all the basic info16:41
apwcking, yeah hense a new bug is appropriate16:41
joaquinapw, that command will create a new bug at launchpad?16:43
apwjoaquin, yep thats the one16:43
apwand attack lots of nice h/w info etc16:44
joaquinapw, OK, great, I like the simplicity16:44
* ppisati -> gym18:13
* rtg skulks off for lunch...18:14
=== kamal2 is now known as kamal
bjfrtg, bug 1084192 while testing your rc7 kernel18:38
ubot2Launchpad bug 1084192 in linux (Ubuntu) "[regression] symlink and hardlink restrictions default to off" [Undecided,New] https://launchpad.net/bugs/108419218:38
=== henrix_ is now known as henrix
* henrix -> EOD18:49
rtgbjf, ack18:50
rtgI think that was a SAUCE patch that got dropped18:50
jsalisburyrtg, bjf, I added that bug to the hot list.18:54
=== henrix is now known as henrix_
=== henrix_ is now known as henrix
rtgkees, so I've reverted 561ec64ae67ef25cac8d72bb9c4bfc955edfd415 in Raring in order to address bug #1084192. Are you good with that ?19:12
ubot2Launchpad bug 1084192 in linux (Ubuntu Raring) "[regression] symlink and hardlink restrictions default to off" [High,In progress] https://launchpad.net/bugs/108419219:12
keesrtg: yup! it's exactly the right solution. I hate that we'll have to carry that forever, but I lost a hated battle with linus over it.19:17
sbeattiertg: that's excellent.19:17
keesrtg: I tried to get CONFIGs for it :(19:17
rtgkees, yeah, I followed that thread, then went on vacation and kind of forgot about it.19:18
keesI'm pondering introducing a system upstream for kernel builders to be able to set the defaults for the entire sysctl tree at build-time.19:18
mdeslaurkees: oh! nice!19:18
keesit'd let us remove all kinds of CONFIGs and satisfy my desires for these kinds of things being definable at build time19:18
sbeattieooh. that would be groovy.19:18
rtgkees, it would certainly reduce the crust we carry in procps.19:19
keesI have no earthly clue how to actually get it done yet, but it's been bouncing around in my head19:19
rtgcruft*19:19
keesrtg: well, procps serves a few purposes -- the main things is for people with stock kernels to get the same important behaviors as an ubuntu kernel where possible.19:19
kees(and also serves as documentation for many of these settings too)19:20
rtgkees, I'd have fixed this in procps, but I also have to think about the LTS backport kernels. I'm too lazy to SRU Precise procps.19:20
keesI actually feel that one of the critical times to have link restrictions available is at early boot.19:20
kees(i.e. it should be fixed in both places)19:21
rtgkees, Linus does pint out that "if you have security problems due to link attacks during your early boot sequence, you have bigger problems"19:22
rtgpoint*19:22
keesrtg: I disagree strongly with that, especially when looking at some android vulnerabilities.19:27
keesrtg: he's a security absolutist, so can't handle the concept of layered security. why risk it, if it's trivially solvable, imo.19:27
rtgkees, well, I was thinking about procps in the initrd. How could anything get in front of that ?19:28
rtgobviously taht only applies if you actually _have_ an initrd19:29
keesprocps doesn't run in initrd and yeah, that too19:29
rtgwell anyway, its solved the Ubuntu way for now :)19:30
sbeattie\o/19:31
rtgsbeattie, do you think this is important enough to upload soon? otherwise we likely won't until the next -rc (or final).19:32
sbeattiertg: it can wait, that's fine.19:32
rtgsbeattie, ack19:32
* rtg slinks off to get some book learning on IPv619:35
=== rtg is now known as rtg-afk
=== henrix is now known as henrix_
=== lifeless_ is now known as lifeless
infinityhggdh: Daily ping, re: backlogged armadaxp verification.21:39
infinityhggdh: At this point, maybe we'd be better off skipping the current round and bringing in the next set to proposed?21:42
infinityhggdh: Since the ones in the PPA cover a CVE...21:42
hggdhinfinity: I am currently running (what I hope to be) the first real armada test22:11
hggdhinfinity: so your daily ping came on time :-)22:11
hggdhinfinity: now the only thing I do not know is how long it will take. May be from one to 6 hours per run22:12
infinityhggdh: Alright.  I'll hold off overwriting those kernels a bit longer, then.22:15
infinityhggdh: Just want to get all the flavours caught up, and those two are from the last cadence.22:16
hggdhinfinity: I know. Sorry for the delay22:17
hggdhinfinity: I have a problem on the armada. I have quantal-proposed enabled, and apt-get dist-upgrade does not find the kernel version 3.5.0-1605.7, only 3.5.0-1604.622:37
infinityhggdh: That was sort of my point...22:38
infinityhggdh: https://bugs.launchpad.net/ubuntu/+source/linux-armadaxp/+bug/106857322:38
ubot2Launchpad bug 1068573 in linux-armadaxp (Ubuntu) "linux-armadaxp: 3.5.0-1604.6 -proposed tracker" [Medium,In progress]22:38
infinityhggdh: That's for 3.5.0-1604.6, 3.5.0-1605.7 is still in the PPA because the previous one never got verified/released.22:38
infinityhggdh: So, should we just skip the two old ones and push forward with the PPA ones?22:39
hggdhinfinity: ah, this bug is not showing in the kernel SRU workflow22:40
hggdhbjf: ^22:40
infinityIt shows in http://status.qa.ubuntu.com/reports/kernel-bugs/reports/sru-report.html22:40
hggdhbut not in http://people.canonical.com/~kernel/reports/kernel-sru-workflow.html22:40
infinityAnd for the out-of-date precise kernel:22:41
infinityhttps://bugs.launchpad.net/ubuntu/+source/linux-armadaxp/+bug/106873322:41
ubot2Launchpad bug 1068733 in linux-armadaxp (Ubuntu Precise) "linux-armadaxp: 3.2.0-1610.15 -proposed tracker" [Undecided,In progress]22:41
bjfhggdh, ack22:41
infinityhggdh: Oh, the bug there is that ike/jani inverted the precise/quantal bugs by accident, so they're under the wrong heading in your report.22:42
hggdhactually, it does appear in the workflow, but it seems quantal and precise are mixed22:42
hggdhheh22:42
infinitybjf: Not your bug, really.22:42
bjfinfinity, hggdh lets just drop armadaxp support and that would solve this "problem"22:42
infinity:P22:43
hggdhbjf: now that I finally got it (almost) in shape? ;-)22:43
infinityIf we drop x86 support, we'd solve a lot more of our problems.22:43
bjfinfinity, hggdh we probably own 96% of all the armadaxp boards in existance22:43
bjfinfinity, but then i'd be out of a job and that would make my wife "sad"22:43
infinitybjf: Yeah, but on the bright side, no more tax returns.22:44
hggdhthere is that22:44
bjfinfinity, i didn't say I wouldn't be happy22:44
hggdhinfinity, bjf: well, it is running the correct kernel, then, let's wait a bit and see how it fares22:44
infinityhggdh: Fancy.  Do you get two machines to do precise and quantal in parallel, or did they only trust you with one? :)22:45
hggdhinfinity: I am absolutely untrusted, so there is only one in the lab22:46
hggdhinfinity: OTOH, I just gave you some kernels to promote (pandas andx86)22:49
infinityhggdh: The oneiric ones?  Or more still?22:51
* infinity is doing oneiric right now.22:51
hggdhinfinity: pandas -- oneiric, precise and quantal; x86 -- lucid, LTS-oneiric, and precise 22:54
infinityThe flavour ones won't get copied until master is ready.22:54
hggdhack22:54
infinity(But glad they're tested anyway)22:54
infinityOh, and x86/precise still needs cert testing.  But that's not your thing. :)22:56
hggdhinfinity: sorry, both oneiric and lts-oneiric are done22:56
infinityhggdh: Yeah, they're already copied. :)22:57
hggdhheh22:57

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