| === gnomefre4k is now known as gnomefreak | ||
| === gnomefre2k is now known as gnomefreak | ||
| === gnomefre1k is now known as gnomefreak | ||
| _roanth_ | why are the NUMA config settings enabled by default on my amd x86_64 dual core machine (ubuntu 7.10) ?? | 06:29 |
|---|---|---|
| === asac_ is now known as asac | ||
| benh | BenC: ping | 06:56 |
| duncanm | hello amitk | 07:03 |
| duncanm | i'm having trouble with suspend on my eee because my /home is mounted on an SDHC card, and it seems like USB persist is not working | 07:07 |
| amitk | duncanm: sorry... got distracted | 07:32 |
| duncanm | no problem | 07:33 |
| amitk | duncanm: hardy does have USB persist enabled, but apparently we need another patch on top to get it to work correctly. I have that patch applied and working with Classmate. | 07:33 |
| duncanm | ah | 07:34 |
| duncanm | amitk: do you have packages available? | 07:34 |
| duncanm | a ppa or something like that? | 07:34 |
| amitk | unfortunately, due to a conflict with another patch, this patch couldn't make it to Hardy before it was released. I am working to make it available through a ppa - https://edge.launchpad.net/~cmpc-developers/+archive | 07:36 |
| duncanm | hmm, but eventually, this patch will be merged into the mainline kernel, right? | 07:37 |
| duncanm | hmm | 07:37 |
| amitk | duncanm: that is the idea | 07:38 |
| duncanm | i'm not so keen on building a kernel on an eee pc, i wish my desktop didn't died earlier this week ;-( | 07:38 |
| duncanm | amitk: oh, also, it'll be great if you could update that bug report with a note about this | 07:38 |
| duncanm | i think as more eee pc users upgrade to hardy, more people will run into this problem | 07:38 |
| amitk | duncanm: what is the bug id, if you have it handy? | 07:39 |
| duncanm | i think it's 197166 | 07:40 |
| duncanm | https://bugs.launchpad.net/ubuntu/+source/linux/+bug/197166 | 07:40 |
| ubotu | Launchpad bug 197166 in linux "[hardy] kernel should have usb persist mode built in" [Wishlist,Fix committed] | 07:40 |
| duncanm | i can post on the eeeuser forum too | 07:40 |
| amitk | duncanm: there are two things - enabling CONFIG_MMC_RESUME_UNSAFE and the USB persist patch | 07:41 |
| duncanm | ah | 07:41 |
| amitk | duncanm: I don't have access to an eeePC, so I am not 100% positive, but the CONFIG option should be required for a mounted fs to be restored on resume | 07:43 |
| duncanm | amitk: do you have a pre-built package? i could test it out for you right now | 07:46 |
| duncanm | amitk: could you check on my message before i post it to the eeeuser forum? | 07:47 |
| duncanm | amitk: i PM'ed you my message, i'll post it once you tell me it's good to go | 07:49 |
| amitk | duncanm: I should have one early next week. Just add the above PPA to your /etc/apt/sources.list and upgrade next week. | 07:50 |
| duncanm | hmm, okay | 07:51 |
| duncanm | amitk: can you check on my PM? | 07:51 |
| duncanm | oh, i changed the wording a bit so people won't think there's a pre-built kernel | 07:52 |
| duncanm | amitk: yeah? | 07:56 |
| amitk | duncanm: I'll comment in detail on the bug, than you can reword your message. There is a bug there :) | 07:58 |
| duncanm | oh | 08:02 |
| duncanm | okay | 08:02 |
| duncanm | amitk: i shorted my message and posted it, http://forum.eeeuser.com/viewtopic.php?pid=232684#p232684 | 08:09 |
| amitk | duncanm: check bug | 08:17 |
| duncanm | asac: hey, i found out about this mozilla bug (https://bugzilla.mozilla.org/show_bug.cgi?id=424915) | 08:26 |
| ubotu | Mozilla bug 424915 in Layout: View Rendering "Scrolling on FF3/gmail/linux much slower than on FF2" [Normal,Resolved: fixed] | 08:26 |
| duncanm | which is probably related to the slowdown that people have been seeing in beta5 | 08:26 |
| duncanm | the package on my SuSE box was rebuilt with that patch, and performance is a lot better | 08:27 |
| duncanm | amitk: thanks for updating the bug | 08:46 |
| duncanm | amitk: good night | 08:46 |
| james_w | Hi all. What's the easiest way to find out what version of something is in l-u-m? | 09:18 |
| james_w | I wan't to know whether it's worth fixing the ov511 package. | 09:18 |
| james_w | #define DRIVER_VERSION "v1.64 for Linux 2.5" <- Is that likely to be correct? | 09:27 |
| amitk | james_w: "grep VERSION" in the ov511 directory will show the version of the decompressor modules too. But the version above looks right. | 09:57 |
| james_w | amitk: ok, thanks, it's probably worth fixing the ov511 package then as that is 2.32. | 10:04 |
| james_w | How do you feel generally about -source packages that are for modules present in l-u-m etc? | 10:05 |
| amitk | james_w: didn't quite understand that | 10:09 |
| james_w | so there is at least one (ov511) module that is in l-u-m, but also has (from Debian) a -source package for use with module-assistant to let people compile it for themselves | 10:11 |
| james_w | maintaining two is a bit of overhead, but allowing people to compile if they want could be a good thing. | 10:11 |
| james_w | in this case they would get a newer version of the module, so it could be really useful, but other times it might be the same version. | 10:11 |
| amitk | james_w: the Ubuntu kernel is completely independent of Debian. We have our own build system and compiling using Debian tools isn't really supported. | 10:13 |
| mjg59 | james_w: if ov511-source is still in Ubuntu, then it should be removed | 10:13 |
| mjg59 | It's an oversight | 10:13 |
| james_w | mjg59: even though it's a newer version? | 10:14 |
| mjg59 | Then it's also a bug that the newer version isn't in l-u-m | 10:15 |
| mjg59 | But yes, supplying two versions of the same code in different versions is obviously wrong | 10:16 |
| james_w | ok, thanks. I'll file that bug if it's not there, and then see about removing the -source packages that we ship. | 10:16 |
| james_w | just to confirm, we would want to remove all of the ones in -ubuntu-, -backports- and -restricted- (if any?) | 10:17 |
| cjwatson | we can't (well, I suppose won't) remove packages from hardy's release pocket now, though | 10:19 |
| amitk | james_w: yes, anything we carry in LUM, LBM and LRM shouldn't be available as separate source packages. And all other sources (if any) should be moved to one of these three. But this will happen only for Intrepid now, AFAICT. | 10:20 |
| james_w | cjwatson: *nod*, this will just be going forward | 10:21 |
| james_w | amitk: ah, ok, so you want to pull in the other -source packages as well? I couldn't find ov51x-jpeg anywhere, that would be a candidate, and the package that started me down this whole road in the first place. | 10:22 |
| james_w | I'll file bugs for those as well then. | 10:22 |
| james_w | what's the deciding factor between LRM and LBM? | 10:22 |
| mjg59 | l-b-m is for backported stuff from newer kernels | 10:23 |
| mjg59 | l-r-m is for non-free drivers | 10:24 |
| mjg59 | l-u-m is for everything else | 10:24 |
| amitk | james_w: R=restricted (binary blobs, firmware that are redistributable) | 10:24 |
| james_w | cool, thanks. I mistyped LRM instead of LUM. | 10:24 |
| === gnomefre1k is now known as gnomefreak | ||
| soren | I feel like an idiot for asking this, but if I get a "BUG: soft lockup - CPU blah" for *both* of my CPU's... How is it that I'm actually getting that error message? | 16:45 |
| soren | I mean... if it's in some sort of locked up state, how does it find the time to step out of the lock, tell me it's locked, and then wander back into the locked state? | 16:46 |
| rtg | it sounds like soft lockup threads get queued. I'd have to look for sure. | 16:47 |
| soren | There's just something about the state that I don't entirely understand. | 16:49 |
| soren | For some reason the CPU stops rescheduling? Is that it? So something disabled interrupts and never reenabled them? | 16:50 |
| rtg | kernel/softlockup.c: this code detects soft lockups: incidents in where on a CPU the kernel does not reschedule for 10 seconds or more. | 16:53 |
| soren | ...so if it doesn't reschedule.. | 16:54 |
| soren | No, still not getting it. | 16:54 |
| soren | If it doesn't reschedule, how does the thing that echoes the error get scheduled? | 16:55 |
| soren | (if all CPU's are in this state) | 16:55 |
| rtg | soren: I imagine the BUG warning is running off the timer interrupt. | 16:56 |
| soren | I suppose I'll understand when I grow up. | 16:58 |
| === gnomefre1k is now known as gnomefreak | ||
| === gnomefre1k is now known as gnomefreak | ||
| |DuReX| | https://bugs.launchpad.net/ubuntu/+source/linux/+bug/209971 | 22:20 |
| ubotu | Launchpad bug 209971 in linux "[Hardy Regression] cx22702 no longer works" [Undecided,Incomplete] | 22:20 |
| |DuReX| | any id when this will get fixxed ? | 22:20 |
| |DuReX| | the patch is there, and its working perfect | 22:20 |
| |DuReX| | would be nice if it was included into ubuntu :) | 22:20 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!