| Sarvatt | hmm, about radeontool, I dont see it in a VCS anywhere in debian to merge with? can we just make a release based off the upstream 1.6.0 tarball? | 00:58 |
|---|---|---|
| Sarvatt | ah I should have read Bug #490052 better, guess tormod's going to try to get it in a vcs in debian. I was just packaging one up for edgers | 01:00 |
| ubottu | Launchpad bug 490052 in radeontool "please sync radeontool 1.5+git20091120-1 from Debian main unstable" [Wishlist,In progress] https://launchpad.net/bugs/490052 | 01:00 |
| Sarvatt | looks like theres been some important updates post 1.6.0 too even though its like a week old, darn | 01:02 |
| Sarvatt | i'll wait to talk to him about it, lessee about intel-gpu-tools updates, just noticed lucid still has the old git checkout from june when I was getting some dumps | 01:06 |
| M25 | Hey, I'm on Karmic with ATI free drivers, I have a problem with suspending/resuming X, from VT switches, logouts, or suspend/hibernate. Everything in the logs points to memory corruption. Are there any settings I can change or should I just file a bug report and hope? | 01:08 |
| Sarvatt | M25: you might want to try the xorg-edgers PPA to see if things are any different, if your problem is fixed there it would help narrow down where your problem is. I would recommend trying out a newer kernel to see if that fixes things first though because that's most likely where the fix will be | 01:12 |
| Sarvatt | what ATI card is it? | 01:14 |
| Sarvatt | http://kernel.ubuntu.com/~kernel-ppa/mainline/ | 01:15 |
| Sarvatt | 2.6.32.2 would be a good one to try from there, unless its an r600 or newer card because there's a bug with those in 2.6.32.2 | 01:15 |
| Sarvatt | (aka HD2xxx or newer) | 01:16 |
| Sarvatt | https://edge.launchpad.net/~xorg-edgers/+archive/ppa -- thats the xorg-edgers PPA, note it says latest crack but we tend to be somewhat conservative outside of the development releases so it's not a huge risk of breaking things :D I would try the kernel first though, and if that doesn't fix things try that PPA with both the new and older kernels to see if any of the problems are fixed. xserver-xorg-video-ati had a few fixes post karmic that mig | 01:22 |
| Sarvatt | ht be relevant.. there is a package called ppa-purge in there that you can install to remove all the packages in the PPA and go back to stock karmic ones after | 01:22 |
| Sarvatt | shouldnt intel-gpu-tools have been synced from debian testing automatically? | 02:09 |
| M25 | Sarvatt: it's a RS4XX series | 02:27 |
| M25 | the bugs I'm getting aren't related to the drivers, they happen on my desktop with an Intel card | 02:28 |
| Sarvatt | ah yeah 200M, I see a fix in xserver-xorg-video-ati regarding VT problems in the past few days | 02:28 |
| M25 | :D how soon would that be pushed to Ubuntu? | 02:29 |
| Sarvatt | does it freeze when you switch to a VT? | 02:29 |
| M25 | Sarvatt: no, it doesn't even always freeze when switching back. It'll do it the second or third or somethingth try, rarely ever the first, and only when switching back. It also has by far the weirdest resume-from-suspend bug ever | 02:30 |
| Sarvatt | is it a constant problem or just sometimes? http://cgit.freedesktop.org/xorg/driver/xf86-video-ati/commit/?id=3a30210d50b27f8772fc5045133940246764fce9 | 02:30 |
| Sarvatt | yeah that sounds like what that commit fixes | 02:30 |
| Sarvatt | theres a VT switch on suspend/resume too | 02:30 |
| M25 | Sarvatt: if I suspend/resume X enough times, it always happens... oh really now | 02:31 |
| M25 | on suspend-resume I get a different crash. It's not immediate, it can be anywhere from 3 minutes to half an hour, but eventually one of my programs will stop responding, either taking 0 or 100% cpu. then every program I touch crashes instantly, one after the other. When they're all crashed, X crashes, and even a REISUB won't reset it, I have to hard boot | 02:32 |
| M25 | this is new with Ubuntu Karmic, never seen anything like it before. only ever happens after a suspend (this is computer suspend, although I assume there's an X suspend associated with it) | 02:33 |
| Sarvatt | does it happen without KMS enabled as well? | 02:33 |
| M25 | I don't think I have KMS enabled | 02:33 |
| M25 | what's the default? I've heard enough bug reports about it not to mess with it yet | 02:33 |
| Sarvatt | not positive but I thought it was enabled by default, you can edit the grub boot line to have radeon.modeset=0 after the quiet splash part if you want to try disabled | 02:34 |
| M25 | I'll try that, what do I lose by disabling it? | 02:34 |
| Sarvatt | dmesg | grep radeondrmfb -- if theres anything there its enabled | 02:34 |
| M25 | nope, nothing back from that | 02:34 |
| M25 | yeah, says it's disabled already | 02:35 |
| M25 | so, before I go installing the edgers ppa, exactly how would I remove the packages it installs, preferably from console. With my luck, X will refuse to run entirely :p | 02:37 |
| Sarvatt | sudo apt-get install ppa-purge after you enable the repo, then you would sudo ppa-purge xorg-edgers to switch back to karmic | 02:38 |
| Sarvatt | i was going to throw a newer ati driver in the drivers-only ppa but i just remembered it needs libdrm 2.4.17+ to build now :( | 02:39 |
| M25 | thanks, I'll see how this goes, would the VT thing that's just been committed to the git above be in the PPA yet? | 02:39 |
| Sarvatt | yeah it is | 02:39 |
| Sarvatt | I would actually give KMS a try and see if it works any better | 02:43 |
| Sarvatt | M25: the problem after suspend/resume sounds different for sure though, sounds more like a kernel issue there | 02:51 |
| M25 | Sarvatt: it's been duplicated by somebody else, on Intel drivers, he said the new kernel didn't fix it, neither did disabling compiz | 02:51 |
| M25 | alright, rebooting, hope this fixes something and breaks nothing | 02:55 |
| M25 | Sarvatt: appears to work, I still get glitchy screens of flickering lines on occasion when switching back to an X session, but I can keep switching around and they disappear | 03:02 |
| M25 | that's been the case for ages, thanks for the help! | 03:02 |
| M25 | Sarvatt: so where would I try the latest kernel to see if it fixes my suspend/resume issues? | 03:03 |
| Sarvatt | things should be better when ya upgrade to lucid if that's the case, if you want I can try adding that one commit as a patch to karmic's ati ddx so you can see if it fixes it and try for an SRU if so. | 03:04 |
| Sarvatt | http://kernel.ubuntu.com/~kernel-ppa/mainline/ | 03:04 |
| Sarvatt | i'd recommend trying 2.6.32.2 from there | 03:04 |
| M25 | Ok, I'm a linux user of two years, what's ddx and what's an SRU? | 03:05 |
| M25 | Sarvatt: that's the one the other guy tried with no effect, rc6 if I remember correctly | 03:05 |
| Sarvatt | xserver-xorg-video-ati is the ddx, SRU is a stable release update which is how you get fixes like that into already released series | 03:06 |
| Sarvatt | well 2.6.32.2 is 5 versions newer than rc6, might be worth trying out :D | 03:07 |
| M25 | right, I'd say go for it, I can do a ppa-purge once it's in and see if the fix stays. I like being on cutting edge of development, but not for core utilities, I'd like to get off the PPA if possible | 03:07 |
| Sarvatt | his problem could be different even though it seems similar too, never know | 03:08 |
| M25 | Sarvatt: that link takes me to a list of directories... I found the one I want, should I just install the .deb manually? | 03:09 |
| Sarvatt | yep, linux-image-2.6.32-02063202-generic_2.6.32-02063202_(whatever your architecture is).deb | 03:10 |
| M25 | I won't need the headers? | 03:10 |
| Sarvatt | not unless you have something else installed that needs them, like virtualbox or something | 03:10 |
| M25 | nope | 03:10 |
| M25 | wait | 03:11 |
| M25 | I've got them installed, methinks it'd be a bad idea to have headers for the wrong version of my kernel | 03:12 |
| Sarvatt | theres no linux-libc-dev where the important headers are for those anyhow, should only need those headers if you need to build an external kernel module or something | 03:13 |
| M25 | I'll take your word for it | 03:14 |
| Sarvatt | i only install the headers on my box i use nvidia binary drivers on | 03:15 |
| M25 | right, new kernel installed, rebooting | 03:15 |
| M25 | Sarvatt: new kernel is weird, the first suspend I tried left me with no X, had to REISUB, but when I rebooted and tried it again, it worked fine | 03:45 |
| M25 | except now my panel is using 100% cpu and isn't being killed by kill-9 | 03:49 |
| M25 | Sarvatt: well, I'm never trying THAT kernel again | 04:01 |
| M25 | 3 boots, 3 completely separate things breaking, all after suspend | 04:01 |
| Sarvatt | suspend/resume doesnt work for me on 2.6.32 either if I ever mount an SD card, kind of stinks because thats whats going into lucid | 04:03 |
| M25 | they've got 4 months, they'll fix it | 04:13 |
| superm1 | no, it's been broke for a few releases | 04:15 |
| superm1 | doesnt necessarily mean it will be fixed... | 04:15 |
| hyperair | Sarvatt: what's going into lucid? | 04:22 |
| Sarvatt | hmm, 2.6.33 needs external firmware to work right for r600 KMS with those mainline kernel builds | 04:23 |
| hyperair | O_o | 04:24 |
| Sarvatt | whatcha mean hyperair? | 04:24 |
| hyperair | nevermind | 04:24 |
| Sarvatt | oh I was saying 2.6.32 was going into lucid and has some problems for me | 04:24 |
| hyperair | i see | 04:24 |
| hyperair | but i haven't heard of a single kernel that supported susend/resume with an SD card mounted | 04:25 |
| Sarvatt | 2.6.31 and under always worked fine here, i dont mean anything important on an SD but they all refuse to suspend if I have ever mounted an SD card now, the controller isnt shutting down right or something | 04:26 |
| Sarvatt | maybe if I compile my own with unsafe suspend enabled it might work, gotta try that | 04:26 |
| Sarvatt | oh wait no it doesnt, i did that at 2.6.32-rc6 | 04:27 |
| Sarvatt | its not just having one mounted that blocks suspend here though, even if i mount one then unmount it it'll freeze trying to suspend | 04:28 |
| Sarvatt | or mount/unmount/remove all modules completely | 04:29 |
| Sarvatt | wonder why R600_rlc.bin isnt getting included with 2.6.33-999 | 04:34 |
| hyperair | O_o | 04:34 |
| hyperair | Sarvatt: by the way, is xserver-xorg-video-intel safe to upgrade yet? | 04:36 |
| Sarvatt | I dunno, it's not on my system but could be fine for you :D | 04:37 |
| hyperair | heh | 04:37 |
| hyperair | last i recall, it sucked on my system as bad as yours | 04:37 |
| Sarvatt | havent seen any i965 people having the problem i'm having | 04:37 |
| hyperair | hmmm | 04:37 |
| Sarvatt | http://bugs.freedesktop.org/show_bug.cgi?id=25475 | 04:38 |
| ubottu | Freedesktop bug 25475 in Driver/intel "[i915] Xorg crash / Execbuf while wedged" [Critical,New] | 04:38 |
| Sarvatt | i think 965 can recover from that hang anyway | 04:38 |
| hyperair | hmmm | 04:39 |
| hyperair | dunno about that | 04:39 |
| hyperair | i was having a lot of artifacts | 04:39 |
| Sarvatt | yah i had lots of those problems from 11-11 till 12-06 or so | 04:42 |
| Sarvatt | well i think the last artifact problem (corrupted glyphs after resume) was fixed about a week after that | 04:42 |
| Sarvatt | oh ok theres no R600_rlc firmware even in the upstream kernel yet, no wonder | 04:44 |
| Sarvatt | hyperair: if you do try it and it works i'd like to hear about it :D | 05:01 |
| hyperair | Sarvatt: heh okay, i'll give it a go | 05:39 |
| === mac_v is now known as \vish | ||
| === mac_v is now known as \vish | ||
| unggnu | hi all | 12:13 |
| unggnu | I tried 10.04 on an old laptop with a Geforce 440 Go an nv was used. | 12:14 |
| unggnu | It seems that nouveau isn't installed at all. Is there a change of plan? | 12:14 |
| unggnu | Ok, I guess I ask another time. :) | 12:17 |
| unggnu | Merry Christmas or whatever you celebrate :D | 12:17 |
| hyperair | Sarvatt: i've had no issues with the new xserver-xorg-video-intel so far, except for one compiz crash a few seconds ago | 14:56 |
| hyperair | xserver-xorg-video-intel: Installed: 2:2.9.99.902~git20091222.2c142e42-0ubuntu0sarvatt2~karmic | 14:56 |
| Sarvatt | well, down to just these 2 bugs with libdrm 2.4.16+ on my aspire one - https://bugs.freedesktop.org/show_bug.cgi?id=25598 https://bugs.freedesktop.org/show_bug.cgi?id=25475 | 19:05 |
| ubottu | Freedesktop bug 25598 in Driver/intel "[965GM] Corruption on resume from hibernation with xf86-video-intel-git" [Critical,New] | 19:05 |
| Sarvatt | still have this bug with 2.6.32+ kernels but at least thats fixed already in drm-intel-next https://bugs.edge.launchpad.net/ubuntu/+source/linux/+bug/492392 | 19:06 |
| ubottu | Ubuntu bug 492392 in linux "[lucid, intel] After suspend, flickering screen and then blank screen." [Medium,Triaged] | 19:06 |
| Sarvatt | ahh just noticed that bug reporter is on my exact netbook even | 19:07 |
| Sarvatt | hmm maybe intel_reg_dumper moving to intel-gpu-tools instead of the ddx might be blocking syncing the new version. should we be installing the udev rule with intel-gpu-tools instead of xserver-xorg-video-intel because intel-gpu-tools is only a recommends? | 19:16 |
| Duke` | are you saying that in 2.6.33 (or 2.6.34?) the execbuf bug is fixed? or is it just the suspend/resume issue? | 19:39 |
| Duke`_ | I just got this bug :/ | 19:47 |
| Sarvatt | suspend/resume | 19:56 |
| Sarvatt | and just in drm-intel-next, they havent pushed it to linus in a long time | 19:57 |
| jcristau | Sarvatt: fwiw the submitter for fd.o 24928 says he still gets the issue with drm-intel after resume | 19:59 |
| Sarvatt | hmmmm | 20:03 |
| Sarvatt | i've got 3 days uptime post resume here since updating the kernel to drm-intel-next with that commit with nothing out of the ordinary | 20:04 |
| Sarvatt | his picture looks very different than what happens to me | 20:05 |
| Sarvatt | oh i'm looking at old posts, I did get the corruption he got from the intel ddx back then, hes still getting the solid color freezes | 20:06 |
| jcristau | not sure when i'll get a chance to try a newer kernel on the 945 at the office... | 20:06 |
| Sarvatt | sounds like theres another commit in drm-intel-next above and beyond that then if it really didnt fix it for him :( | 20:09 |
| Sarvatt | actually, I'm using the powersave=0 module parameter on top of that commit | 20:11 |
| Sarvatt | going to try without that now | 20:11 |
| === mac_v is now known as \vosh | ||
| === \vosh is now known as \vish | ||
| Sarvatt | jcristau: thanks for getting me to look into that more, powersave=0 is indeed required as well to fix it on top of that patch :( | 21:03 |
| jcristau | damn | 21:03 |
| Sarvatt | i froze up just starting xchat after a resume without it | 21:03 |
| jcristau | well at least you get consistent results.. | 21:04 |
| Sarvatt | got a bunch of gpu and reg dumps and logs from it at least, that one only liked to crash me when i was out of the house and actually needed suspend :D | 21:04 |
| Sarvatt | RAOF: hmm nouveau-firmware didnt get removed when i removed nouveau-kernel-source? dunno if you care about that | 21:08 |
| Sarvatt | its too unstable for me right now, every time i open the lid after having it closed for a day or two the screens corrupted and gpu hung | 21:09 |
| Sarvatt | oh scratch that, didnt show up in the autoremove list after but now it is | 21:12 |
| Sarvatt | gotta figure out what bug to put that powersave=0 needed comment on now | 21:13 |
| Sarvatt | when it hangs after resume intel_gpu_top says its spending 100% of the time in framebuffer compression | 21:38 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!