[00:17] sigh, cs46xx hardware is largely broken :( [00:18] I suppose we can reprobe the register until it appears to return valid data [00:37] anyone around who can answer a question about filing a bug ? [00:45] bladernr_, don't ask if you can ask. bad irc etiquette [00:45] better to just ask your question straight off, then people can answer or ignore as appropriate ;-) [00:46] bladernr_, also you probably should ask on #ubuntu or #ubuntu+1, as this is more a developer-oriented channel [00:52] bryceh: sorry... was talking in other channels too ... I know better :/ === lexical is now known as kengyu [04:23] Hey! Is there a working PPA? [04:24] 2.6.31-20-generic-pae is killing me. [04:24] Slowest kernel I've seen in a long time. [04:24] Memory consumption is off the charts and I am actually swapping... [04:26] If not, I am going to give a little script I wrote a try to see if I can successfully make synaptic think kernel.org and its change log is a valid ppa [05:35] Out of 93 people nobody is awake? [05:45] * hyperair compiles his own kernels [05:46] isn't PAE supposed to have some pretty bad performance, though? [05:55] well it depends on what you are doing [05:55] PAE does have a penalty [05:56] memory consumption off the charts just seems wrong though [06:04] Well, I am building a 2.6.33 right now on the laptop that crashed and see if its any better, but my ubuntu box, I want to stay with something known not to mess with the tool chain, I am building my own tool chain and all on the crashed laptop [06:06] I don't want to have to do that for my main box and frankly I don't even know how pae got on here, I only have 512MB of ran and I didn't go looking for pae so it should be the vanilla generic kernel, but it is not [06:06] The only added repositories I have in my source.list is for the ubuntu debugging symbols. [06:07] slackguru: https://launchpad.net/~kernel-ppa/+archive/ppa has builds of mainline kernels [06:07] and you should be able to just install the linux-image-*-generic to replace the pae one [06:08] Number of packages: [06:08] 0 source packages (0 bytes) [06:08] 0 binary packages (0 bytes) [06:08] Repository size: [06:08] 0 bytes (0.00%) of 8.0 GiB [06:08] i'd try the generic before installing a mainline one from that PPA. [06:09] that repo is dry [06:09] yeah, odd [06:09] nothing there. [06:09] anyhow, try -generic [06:09] I am actually thinking of switching to -rt [06:09] and if that doesn't work, file a bug [06:10] -rt that may improve latency, but not performance [06:10] http://kernel.ubuntu.com/~kernel-ppa/mainline/ for the builds [06:10] but still, try -generic first. [06:11] thanks jk- [06:12] np [06:12] I am actually looking for a ppa that has stable development releases, kernel.org has 2.6.33 and ubuntu has only caught up to 2.6.32 [06:13] you're probably best off using the default kernel, it's had the most testing with the ubuntu userspace. [06:13] with 2.6.31-20 I feel like I am the testing ground [06:14] let me know if their is a daily ppa if you hear anything. [06:15] thanks [06:17] There is an upstream kernel build, but it's not a great idea to switch kernels. [06:18] There's a lot of GPU work going on in the kernels, and you'll hit a poorly-tested version combination quite quickly with different kernel versions. [09:28] hie! i am looking at bug 537465 [09:28] Malone bug 537465 in ttf-sazanami "package ttf-sazanami-mincho 20040629-2ubuntu2 failed to install/upgrade: (dup-of: 533609)" [Undecided,Incomplete] https://launchpad.net/bugs/537465 [09:28] Malone bug 533609 in mono "package mono-runtime 2.0.1-4ubuntu0.1 failed to install/upgrade: " [Undecided,Incomplete] https://launchpad.net/bugs/533609 [09:28] and in the lspci output attached by apport [09:28] i cannot seem to see any driver module against the VGA compatible controller [09:28] any guesses why this might be happening? [09:31] will this be an in kernel compiled driver? [09:32] amitk, psurbhi, http://paste.ubuntu.com/397697/ has the first two panics i get with the omap kernel [09:32] ogra: dmesg? [09:33] psurbhi, the lspci -vvvnn output would list a driver attached to the device even if it is build-in [09:33] amitk, the board hangs hard, no way to get to dmesg at that point [09:34] ogra, any snap possible? [09:34] amitk, but i assume you should see the same on your board if you have other ways to get data [10:26] apw, have you delved into acpi debugging your dell after s/r w.r.t the temp sensor issue? [10:40] cking, nope not had a second to look at it other than to be annoyed by it [10:40] you can teach me when i get back === matti_ is now known as matti [11:48] a netxen (10 Gb) network adapter (NC522SFP) is not having it's module (netxen_nic) loaded automatically on hardy. what to do? [12:00] * cking appreciates apw's daily kernel builds - makes bisecting so much easier [12:02] cking, cool [12:03] apw, at least smb will have something to do whilst we're gone. I see a bunch more .32 stable updates in the pipe [12:03] tgardner, yeah wouldn't want him to be bored now would we [12:49] and for extended self flagellation he even tries to add to them [12:51] smb, IIRC there was an igb patch you can skip [12:52] tgardner, I think I saw some drop thing. But I usually check with whats really in the queue anyway [12:52] tgardner, Btw, have we been having problems with tg3 based cards? [12:53] smb, nothing thats attracted my attention [12:54] Ok, might just be my imagination by half seeing things passing by [12:56] smb, the tigon driver is quite popular, and generally stable [12:56] Just because that is in the queue "tg3: Fix 5906 transmit hangs" [12:57] and another on on that [12:57] tg3: Fix tg3_poll_controller() passing wrong pointer to tg3_interrupt() [12:57] smb, perhaps its a rare circumstance [12:58] Quite possible. Usually the commit message does not say how high the probability of an issue is [13:30] hey, does anyone know what the bcdDeviceMin, bcdDeviceMax parameters in the UNUSUAL_DEV macro in drivers/usb/storage/usb.c signify? [13:30] psurbhi, IIRC those were version numbers [13:31] ok, so how do you aquire those? [13:31] Like device version/firmware version x to y [13:31] lsusb -v [13:31] aah, ok :) thanks [13:32] psurbhi, Of course you only get one number for the device and version you actually have [13:32] * psurbhi checks [13:32] psurbhi, so the limits usually grow over time or were gathered by asking around [13:33] so then is it safe to put the min as 0x0000 and max as 0xffff? [13:33] as i see for other entries? [13:34] psurbhi, Only if you assume all are broken. So not safe [13:34] ok.. [13:34] The safe approach is only to use the one you have (or a range if you have that) [13:34] so then safer is to be specific and put it for the specific version that lsusb gives [13:34] right [13:34] smb, ack! thanks [13:35] psurbhi, btw its the bcdDevice entry [13:35] smb, thanks a lot! just when i was searching :) [13:35] And I believe it was a bcd encoding used for the integer [13:36] so a bcdDevice = 6.01 woule be 0x0601, but better check that [13:36] which is why the name has a bcd om ot [13:36] ok.. [13:37] s/om ot/in it [13:37] psurbhi, Yeah, actually true :) [13:40] smb, i have the hal output.. and am hoping the lshal entry "revision_bcd" is the same [13:41] psurbhi, You could check with a usb device you got? [13:42] yeah [13:46] got that :) thanks smb! [13:46] np :) === sconklin-gone is now known as sconklin [15:17] Found a bug that affects me that I'd like to pitch in on. https://bugs.launchpad.net/ubuntu/+source/linux/+bug/532374 -- comment 21 and 22 shows where we're stalled, and I suspect it's something I can overcome with a bit of guidance. [15:17] Malone bug 532374 in linux "[Lenovo 2537AD3, 2537AB8] one successful suspend/resume per boot" [Undecided,Confirmed] [15:17] ubot3: You just scared the crap out of me. Don't do that. :-) [15:17] KB1JWQ: Error: I am only a bot, please don't think I'm intelligent :) [15:19] Specifically, the question becomes "how do I compile a kernel 'The Ubuntu way' with ACPI_PROCESSOR as a module. [15:24] The docs for kernel recompilation on Karmic seem a bit outdated for Lucid; is there a pointer to fresher documentation on this than https://help.ubuntu.com/community/Kernel/Compile ? [15:25] KB1JWQ, https://wiki.ubuntu.com/KernelTeam/KernelMaintenance [15:29] RAOF, how can I find it? I wouldn't be asking if I didn't understand what I was getting myself into. [15:30] tgardner: Thanks! [16:03] Hello everybody, I'm currently an intern @ Canonical's office in Montreal and I have some questions about the kernel's ACPI management.I want to be able to wake up the computer after a period of time, and I use the following script to do it: http://pastebin.ubuntu.com/397848/, with pm-hibernate of poweroff on the last line. We have over 60 laptops in the lab and I've tested most of them. Last week, a machine, woke up just from poweroff stat [16:03] e and not from pm-hibernate. This week, it wakes up from both states. It happens alos with several others machines. My question is Were there some changes in the kernel in the last week that are related to this issue? If yes, please point me into the right direction to further investigate. Thanks! [16:09] alexmoldovan, It might help to say which kernel (series) you are using. Lucid? [16:10] yes [16:10] I'm testing lucid in the lab [16:10] alexmoldovan, kernel freeze was March 11, so if your results have changed since then, it's related to userspace changes [16:10] alexmoldovan: IIRC, the wakeup is mostly hardware induced (IOW, no kernel changes affect whether it wakes up or not) [16:11] it's a case of the ACPI alarm working or not (if it's software related, then it's a case of access to the ACPI alarm, not so much kernel really) [16:12] alexmoldovan: if it's a case of the kernel not working on wakup, then it's unrelated to the alarm itself, and just a suspend/resume issue [16:12] alexmoldovan: do your tests show that a forced wakeup works, and just that the alarm isn't waking the system up? [16:13] the wake from the power button works fine [16:13] but not with the alarm [16:14] Ok, then it sounds like that the ACPI alarm just isn't triggering, which is usually a bug in BIOS more times than userspace/kernel [16:14] The power button is a different source for wakeup. [16:14] alexmoldovan: do these systems have a BIOS setting where you can set the alarm in setup? [16:15] I'll check if some of them have, but the netbooks have a very limited BIOS options [16:15] if you test that, and it works, then it sounds like userspace just isn't setting the alarm properly [16:16] Here is some thing that may be helpful: http://smackerelofopinion.blogspot.com/2009/08/acpi-wake-alarm-bugs.html [16:16] alexmoldovan: it's been awhile since I messed with setting the wakeup alarm...does the script verify that the alarm was set by reading back the value? [16:16] cking, I think you also had some test script [16:16] chrisccoulson_, given your explanation, I am not entirely sure, it sounds like a network manager issue. [16:16] it references a test script in the blog [16:17] the one I use is taken from chackbox : http://pastebin.ubuntu.com/397848/ [16:17] http://kernel.ubuntu.com/git?p=cking/scripts/.git;a=blob;f=acpi-wakealarm-test/wakealarm.sh;h=a1674689dc49d0ffbea8f3bf76e529cfe445901c;hb=c88cec173caa995d3737437a82082152e34a46fc [16:17] chrisccoulson_, I'll look at the bug a bit deeper [16:17] JFo - yeah, possibly. although I can't recreate it today ;) [16:17] hmmm [16:17] fun :) [16:17] but, i'm sure the issue i saw yesterday was not kernel related [16:17] alexmoldovan, perhaps those links may be useful for resolving your bug [16:18] chrisccoulson_, yeah, I agree [16:18] We want to find a way to automatize the tests in the lab even further. So far, we have to manually start up each machine from the power button. [16:19] * BenC thought that's what interns were for ;) [16:19] Thanks for the quick response to everyone! [16:20] BenC, reminiscing your internships? :) [16:22] hehe, a little grunt work never hurt anyone [16:22] well, that's the theory, anyhow [16:23] I like it, though it would save time. So far we have ~70 laptops, but we'll have more in the future. [16:24] * JFo suggests a long stick with a pencil taped to the end :-P [16:24] j/k [16:24] alexmoldovan, well, checkout that blog article and see if it helps. If not, it's probably a ACPI/BIOS issue :-( [16:25] I thought we could buy a trained monkey from a circus. After that our only worry will be buying bananas [16:26] alexmoldovan: and banana peels on the lab floor... [16:26] not very appealing [16:26] alexmoldovan, that and cleaning up monkey poo that he slung at you :-) [16:27] nice one cking [16:27] I'd like to thank those who worked on the core i5 with graphics integrated in the CPU, we have one of those in the lad and in lucid it works flawlessly [16:30] JFo: are you running amd64 or i386 on your mini 10? [16:33] isn't the mini 10 an Atom? [16:35] cnd, i386 [16:36] cking, it is [16:36] which case it can only do i386 [16:36] yup === emma_ is now known as emma [17:17] x [17:29] apw: do you have a mmiotrace enabled kernel anywhere by any chance? [17:29] tseliot needs it to make a dump so his nva3 GPU can get supported by nouveau, they've been trying to find someone with one to do a dump for months [17:30] Sarvatt, its possible that one you are testing has it [17:30] oh let me check [17:31] sweet it does, thanks! [17:31] Sarvatt, yeah it should have ... yay [17:33] apw: what did you change in the disable FBC on i915GM/i945GM patch I sent you for this kernel you built? [17:36] Sarvatt, i didn't reformat the layout of the options at the top when removing the .has_fbc options [17:36] as that makes the end stay the same, where new options are likely to appear [17:36] reducing the likelyhood of collision with other patches [17:37] Sarvatt, second i changed the bit where it assigned the fbc functions instead of removing the bits down there, i added HAS_FBC to the surrounding if [17:37] the first change to reduce collissions if we have to carry it, the second cause it seemed safer in the future as it reminds us that bit is about FBC [17:41] Sarvatt, make sense? [17:41] i'm not sure about the change to the latter hunk, I think its the way it is to account for some other devices [17:42] though those functions are only for fbc [17:43] they are called _fbc and only called from places which check for fbc as far as i can see [17:43] all of the MOBILE things have FBC until we stop some of them [17:44] my main reason to change it was i felt we lost information, which fbc routines when with those cards [17:44] but i can be persuaded [17:44] ... [17:44] * apw has to run to another location, will check in later [17:45] Ooh, kernel bug! http://pastebin.com/DZm70xam Note the end, causes ndiswrapper to die [17:47] I'm open to suggestions besides "return the USB wireless card" [17:58] apw: guess it doesn't matter unless they add something later with has_fbc and is_mobile and doesn't use i8xx_fbc-* for setup and is checked after that part === sconklin is now known as sconklin-afk [17:58] ironlake-m FBC support that'll be coming in any time now is checked before if (IS_GM45(dev)) so thats safe [18:03] just about to hit the 4 hour mark of stability with disabled FBC post resume and powersave=1 on my 945 \o/ === maxb_ is now known as maxb === jepler_ is now known as jepler