| cprofitt | anyone home | 00:01 |
|---|---|---|
| cprofitt | https://bugs.launchpad.net/ubuntu/+source/nvidia-common/+bug/981019 | 00:01 |
| ubottu | Launchpad bug 981019 in nvidia-common (Ubuntu) "3D desktops crash or are unusably slow" [Undecided,Confirmed] | 00:01 |
| Sarvatt | cprofitt: reporting it on http://www.nvnews.net/vbulletin/forumdisplay.php?f=14 is going to be a lot more effective, we just ship their binary drivers, its not like we can fix it | 00:05 |
| cprofitt | Sarvatt: not sure it is 100% their driver though... | 00:05 |
| cprofitt | I reverted back to 295.20 and still had issues... the problem may be with kernel or xorg | 00:05 |
| Sarvatt | if its fine with nouveau like the bug says its pretty clear cut, but that bug is very vague | 00:05 |
| cprofitt | just not sure how to proceed to triage | 00:05 |
| cprofitt | I can use Unity 2D with the driver just fine as well... it appears to be something with 3D | 00:06 |
| cprofitt | when I rever to 295.20 Unity 3D works better, but 3D games still end up looking like slideshows | 00:07 |
| cprofitt | Is there any thing I can do to help sort that out? | 00:09 |
| Sarvatt | 295.xx is a longterm release and will be updated many times in the future in 12.04, might be worth switching to nvidia-current-updates so you get future updates automatically now. but yeah that bug is basically "i have problems with 295.xx" so of course people will click it too if they do and its not going to go anywhere, a distinct bug for your issue will be more useful with actual apport info listing what gpu you have and stuff. its certainly not | 00:10 |
| Sarvatt | a widespread problem and might be specific to your gpu | 00:10 |
| Sarvatt | but if you want it fixed for real reporting it on that forum directly to nvidia is the way to go | 00:11 |
| cprofitt | alright... I am going to reinstall -- fresh install... then update... then if the problem persists I will follow the steps in the nvidia forums | 00:11 |
| cprofitt | thanks | 00:12 |
| Sarvatt | i didnt help, but no worries, its the sad reality with the binary drivers :( | 00:12 |
| Sarvatt | I did see that 295.40 has issues on old IGP nvidias from the reports, you have one of those? | 00:13 |
| Sarvatt | like 6150 | 00:13 |
| Sarvatt | or 7100 | 00:14 |
| Sarvatt | oh hmm 7050 i guess it is | 00:15 |
| cnd | Sarvatt, yay! | 00:15 |
| cprofitt | I have an 8800 GTS 320 | 00:15 |
| cprofitt | https://bugs.launchpad.net/ubuntu/+source/unity/+bug/980672 | 00:16 |
| ubottu | Launchpad bug 980672 in unity (Ubuntu) "compiz crashed with SIGSEGV" [Undecided,Incomplete] | 00:16 |
| cprofitt | https://bugs.launchpad.net/ubuntu/+source/unity/+bug/980879 | 00:16 |
| ubottu | Launchpad bug 980879 in unity (Ubuntu) "compiz crashed with SIGSEGV" [Undecided,Invalid] | 00:16 |
| cprofitt | finding some more bugs with this | 00:16 |
| cprofitt | https://bugs.launchpad.net/compiz-core/+bug/980521 | 00:16 |
| ubottu | Launchpad bug 980672 in unity (Ubuntu) "duplicate for #980521 compiz crashed with SIGSEGV" [Undecided,Incomplete] | 00:16 |
| Sarvatt | yeah i'm actually seeing lots of bugs about 295.40 now that i'm looking, i think they rushed out the update for the security fix too fast :) | 00:17 |
| cprofitt | yeah... looking like it... | 00:17 |
| cprofitt | not sure why reverting to 295.20 still caused me issues, but a fresh install will help me troubleshoot | 00:18 |
| cprofitt | the text box had been an upgrade from 11.10 | 00:18 |
| Sarvatt | you might have to purge nvidia, then install 295.20 to work right, but im not sure about that | 00:19 |
| cprofitt | yeah... reading that now... | 00:19 |
| Sarvatt | not sure downgrades work properly, the packaging for nvidia is nuts and lots of room for things to go wrong | 00:19 |
| cprofitt | the install is already in progress though... | 00:19 |
| cprofitt | https://bugs.launchpad.net/ubuntu/+source/unity/+bug/980298 | 00:21 |
| ubottu | Launchpad bug 980298 in unity (Ubuntu) "compiz crashed with SIGSEGV in GLWindow::glDraw() from UnityMTGrabHandlesWindow::glDraw()" [Undecided,Confirmed] | 00:21 |
| Sarvatt | https://launchpad.net/ubuntu/+source/nvidia-graphics-drivers/295.20-0ubuntu1 the debs are there if you click the amd64/i386 links | 00:21 |
| cprofitt | there are times it becomes a mess just to find related things in LP | 00:21 |
| Sarvatt | i use google to find things on launchpad, thats pretty sad :) | 00:21 |
| cprofitt | that is why I don't file bugs until I have a great deal of certainty as to what it is | 00:21 |
| cprofitt | lol -- me too | 00:21 |
| cprofitt | at least some of these folks are getting errors... I was just getting 'slow' | 00:22 |
| Sarvatt | well slow is a different issue so safe to assume the crashes arent related to your problem | 00:23 |
| * cprofitt nods | 00:24 | |
| Sarvatt | honestly if you reported it on nvnews forums it will get the highest possible chance of getting fixed in the next driver release, we only engage nvidia over issues about new platforms not working, something like 8800gts specifically being abnormally slow might not get noticed on launchpad | 00:25 |
| * cprofitt nods | 00:26 | |
| Sarvatt | but it would still be worth reporting on launchpad, maybe theres an issue with how it was installed causing it that we could notice | 00:26 |
| Sarvatt | ubuntu-bug nvidia-graphics-drivers only takes 30 seconds to do | 00:27 |
| cprofitt | I will do that after the reinstall -- if the issue is still present... and will report via the nvidia forums as well. | 00:27 |
| bjsnider | Sarvatt, what are the nature of the bugs? | 01:15 |
| bjsnider | not having any trouble with my gt520 yet | 01:18 |
| cprofitt | I can't speak to the others... | 01:18 |
| cprofitt | my issue was the update that I did today caused Unity 3D to be non-functional with Nvidia 295.40 | 01:19 |
| Sarvatt | bjsnider: sounds like if you hit whatever cprofitt is hitting you would notice, super slow 3D | 01:19 |
| cprofitt | I just reinstalled and updated only the nvidia driver | 01:19 |
| cprofitt | Unity was useable -- testing a game now | 01:19 |
| cprofitt | then will update the rest to see if the issue comes back | 01:19 |
| cprofitt | there is a chance something was latent due to it being an upgrade from 11.10 | 01:19 |
| bjsnider | well, maybe it's a compiz issue, because i'm using mutter | 01:20 |
| bjsnider | is all opengl slow? | 01:20 |
| cprofitt | yes, it seems to be... games were slow as well | 01:23 |
| cprofitt | I used Unity 2D and launched games they had issues | 01:23 |
| cprofitt | interesting... no issues on initial install of 12.04 Beta 2... and upgrade of 295.40 drivers only | 01:29 |
| cprofitt | exiting the game caused some issues though... took about 20 seconds to redraw the desktop | 01:34 |
| RAOF | hyperair: Yo! Did you end up sending something to xorg-devel@ about the coasting scroll thing? | 01:37 |
| Sarvatt | the person who wrote the patch did | 01:38 |
| Sarvatt | RAOF: guessing you're referring to the coasting thing? | 01:38 |
| RAOF | Sarvatt: Yes. | 01:39 |
| Sarvatt | yah check out the "[PATCH synaptics] Fix coasting speed" thread, whot merged it but was waiting on the signoff before pushing from what i saw | 01:40 |
| RAOF | Yeah, we've pulled that into precise. | 01:42 |
| RAOF | But the initial patch seemed wrong, and hyperair fixed it. I'm not sure if that problem was in the patch on the list, though. | 01:43 |
| Sarvatt | he certainly didnt mark it as such in the actual patch if so | 01:47 |
| cprofitt | Sarvatt: https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers/+bug/982710 | 01:48 |
| ubottu | Launchpad bug 982710 in nvidia-graphics-drivers (Ubuntu) "Nvidia 295.40 - Ubuntu 12.04 Beta 2 Fresh Install (slow performance in Unity 3D)" [Undecided,New] | 01:48 |
| Sarvatt | cprofitt: long shot but maybe try adding Option "UseEvents" "True" to the device section of /etc/X11/xorg.conf? | 01:53 |
| RAOF | cprofitt: Does the same problem happen with the 295.33 drivers? There's at least one known performance hit (in rare cases) to the security fix that went in to .40 | 01:55 |
| cprofitt | RAOF: I have not reverted this install yet, but I did my previous one and the performance issue got slightly better, but there was still an issue | 01:57 |
| cprofitt | Sarvatt: will try that | 01:57 |
| Sarvatt | i dont see any NVIDIA(1): WAIT's in the logs that i had when that helped me with on my last nvidia 8xxx machine so its really a random guess | 01:59 |
| cprofitt | random guesses are worthy of a try | 02:00 |
| cprofitt | seems like a rock and a hard place right now | 02:00 |
| cprofitt | do not want a security hole... but do not want people to find fault with the new release based on Nvidia drivers | 02:01 |
| RAOF | To be clear: I don't think it's particularly likely that the security fix is the cause of performance problems; AFAIK it just interfered with some infrequently used 3D paths on some cards. | 02:04 |
| cprofitt | Sarvatt: that setting resulted in a black screen and no login | 02:05 |
| cprofitt | removed it and got to login screen | 02:05 |
| cprofitt | gah... can't register on the NVnews forums | 02:16 |
| cprofitt | page not found error | 02:16 |
| cprofitt | :-) | 02:16 |
| bjsnider | cprofitt, how does nouveau work? | 02:40 |
| cprofitt | bjsnider: it worked fine for Unity | 02:43 |
| cprofitt | I did not test it with games -- not sure it is supposed to work there | 02:43 |
| bjsnider | itmight | 02:46 |
| cprofitt | I will test that as well | 02:46 |
| hyperair | RAOF: er whoops, i forgot to. | 03:20 |
| cprofitt | thanks everyone... gotta get some sleep so I can work in the AM | 03:20 |
| cprofitt | have a great day, night, morning | 03:20 |
| RAOF | hyperair: :) | 03:20 |
| RAOF | hyperair: Just to be clear, the changes you made were in HandleScrolling, right? | 03:21 |
| hyperair | RAOF: i'm not subscribed to the list. could you help me forward it please? :-) | 03:21 |
| RAOF | hyperair: Of course. | 03:21 |
| hyperair | RAOF: yeah, you could just do an interdiff to see the changes | 03:22 |
| hyperair | there's only one extra hunk | 03:22 |
| hyperair | i just amended the way the coasting_friction was deducted from the speed. | 03:22 |
| Sarvatt | it got applied and pushed now like it was sent upstream, if theres a problem it needs a new patch :) | 03:22 |
| hyperair | i see | 03:22 |
| Sarvatt | http://cgit.freedesktop.org/xorg/driver/xf86-input-synaptics/commit/?id=5a1612d4496b51682c9043aa064025c545249de6 | 03:23 |
| hyperair | Sarvatt: http://paste.debian.net/163369/ | 03:40 |
| hyperair | RAOF: ^ | 03:40 |
| hyperair | so that's the amended patch =) | 03:40 |
| hyperair | hmm since it now actually has my name on it, i guess i should send it to xorg-devel myself.. | 03:42 |
| RAOF | Would be easier, yes. | 03:43 |
| RAOF | You'll need to add a signed-off-by tag, too. | 03:43 |
| hyperair | ah. shit. | 03:44 |
| hyperair | it's already sent. | 03:44 |
| hyperair | ¬_¬" | 03:44 |
| hyperair | now what? | 03:44 |
| Sarvatt | hyperair: reply with Signed-off-by: Chow Loong Jin <hyperair@debian.org> | 03:48 |
| Sarvatt | or whatever | 03:48 |
| hyperair | nevermind that, thunderbird broke my patch >_> | 03:49 |
| hyperair | i'll just resend it as an attachment | 03:49 |
| Sarvatt | looks fine to me | 03:50 |
| hyperair | yeah, it looks fine, until you try to git am | 03:50 |
| hyperair | then git complains of a broken patch | 03:50 |
| Sarvatt | oh yeah malformed | 03:52 |
| Sarvatt | git send-email works really well with gmail now easily | 03:52 |
| hyperair | hmm really | 03:53 |
| hyperair | i should try that next time | 03:53 |
| Sarvatt | http://paste.ubuntu.com/932043/ in ~/.gitconfig | 03:53 |
| hyperair | hmmm can i not specify my password there? | 03:54 |
| hyperair | i transplant the .gitconfig around | 03:54 |
| RAOF | Yeah, if you don't specify the password it'll prompt for it. | 03:55 |
| hyperair | okay | 03:55 |
| Sarvatt | oh cool | 03:55 |
| Sarvatt | it used to be a PITA setting up mstmp for gmail access in git send-email but its super easy now | 03:56 |
| Sarvatt | not sure what you use for @debian.org but heard ya moaning about gmail in evolution earlier so figured it might be that :) | 03:57 |
| hyperair | actually come to think of it, i have postfix installed... | 04:03 |
| hyperair | which directly hooks up to gmail | 04:03 |
| hyperair | and yeah, i use gmail for my @debian.org. | 04:04 |
| Sarvatt | you didn't pick no config at the debconf prompt when you installed devscripts and inherited postfix?! | 04:04 |
| * Sarvatt hates doing that on every system :) | 04:05 | |
| Sarvatt | i have >500k emails in my gmail so i totally feel your pain wrt evolution | 04:07 |
| hyperair | haha | 04:07 |
| hyperair | well i installed postfix out of curiosity | 04:07 |
| hyperair | and then found out it could send via gmail using sasl | 04:07 |
| hyperair | so there we go =) | 04:07 |
| Sarvatt | oh so you already had an MTA, lucky. i could care less about having a MTA installed just to use debuild :) | 04:08 |
| hyperair | debuild? | 04:27 |
| hyperair | did debuild need an MTA? | 04:27 |
| hyperair | i use sbuild though | 04:27 |
| hyperair | and there were a crapton of other things that needed an MTA, which i couldn't remember | 04:28 |
| hyperair | er right, like my cron jobs | 04:28 |
| === yofel_ is now known as yofel | ||
| cprofitt | reported the Nvidia driver problem on the nvnews.net forums now | 12:05 |
| cprofitt | http://www.nvnews.net/vbulletin/showthread.php?p=2546106 | 12:05 |
| cprofitt | https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers/+bug/982762 | 12:06 |
| ubottu | Launchpad bug 982485 in nvidia-graphics-drivers-updates (Ubuntu) "duplicate for #982762 nvidia 295.40 breaks unity 3d" [Undecided,Triaged] | 12:06 |
| cprofitt | https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-updates/+bug/982485 | 12:06 |
| ubottu | Launchpad bug 982485 in nvidia-graphics-drivers-updates (Ubuntu) "nvidia 295.40 breaks unity 3d" [Undecided,Triaged] | 12:06 |
| cprofitt | https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers/+bug/982710 | 12:06 |
| ubottu | Launchpad bug 982485 in nvidia-graphics-drivers-updates (Ubuntu) "duplicate for #982710 nvidia 295.40 breaks unity 3d" [Undecided,Triaged] | 12:06 |
| tseliot | cprofitt: I've replied to your email about it | 13:17 |
| cprofitt | tseliot: thanks | 13:17 |
| cprofitt | I will try to get that additional file for you when I get home | 13:18 |
| tseliot | thanks | 13:19 |
| tjaalton | bryceh: back to the drawing board with the wacom touchscreen patch :) upstream committed this http://git.gnome.org/browse/gnome-settings-daemon/commit/?id=e40cb71a910164d1cb2c366d3c001eeaac79aaaf but if that alone doesn't fix it we'd need debug logs to figure out why gsd_wacom_device_is_screen_tablet was zero there | 13:54 |
| tjaalton | bryceh: so, since you have an actual device, could you post debug logs from g-s-d to the bug? | 13:59 |
| tjaalton | if you have time | 13:59 |
| tjaalton | I'll reopen it | 13:59 |
| seb128 | bug #943880 | 13:59 |
| ubottu | Launchpad bug 943880 in xorg-server (Ubuntu Precise) "Xorg crashed with SIGABRT in __libc_message() from malloc_printerr() via XIDestroyDeviceProperty" [High,Confirmed] https://launchpad.net/bugs/943880 | 13:59 |
| seb128 | is somebody working on that issue? | 13:59 |
| tjaalton | I think cnd was | 14:00 |
| seb128 | it's showing up in the top retraced bugs | 14:00 |
| seb128 | I've added a precise milestone to it | 14:00 |
| cnd | seb128, no one knows what it is | 14:03 |
| cnd | it looks to be some corruption, but I can't reproduce it | 14:03 |
| cnd | valgrind doesn't show anything either | 14:03 |
| seb128 | cnd, ok, well I guess the precise milestone still makes sense in any case, that's a bug we will want to SRU fix if we get a fix at some point | 14:03 |
| cnd | yeah | 14:04 |
| bryceh | Sarvatt, regarding regressions from 295.40, we don't seem to have very many new bugs filed against nvidia-graphics-drivers | 16:33 |
| bryceh | wherre does nvidia-current-updates_hybrid.conf come from? | 16:33 |
| Sarvatt | nvidia-common | 16:34 |
| bryceh | well, I can see that :-) but I mean why is it there? is this from alberto's hybrid graphics work? | 16:34 |
| Sarvatt | yeah, i dont know what its doing :) | 16:35 |
| bryceh | bug 981048 implies it isn't getting removed on nvidia disable, making the system unbootable | 16:37 |
| ubottu | Launchpad bug 981048 in jockey (Ubuntu) "nvidia-current-updates_hybrid.conf persists after nvidia drivers are disabled" [Undecided,Confirmed] https://launchpad.net/bugs/981048 | 16:37 |
| * mlankhorst has a hacky setup that basically does the same :x | 16:37 | |
| bryceh | since we're not yet advertising hybrid support, wondering if this should be disabled. if it ends up being flimsy, then is going to be tough to support post-release | 16:38 |
| Sarvatt | huh? i thought nvidia-current-updates_hybrid.conf would be part of the hybrid-detect part but its installing the blacklist as nvidia-current-updates_hybrid.conf? | 16:38 |
| Sarvatt | i guess hybrid-detect is writing out a new blacklist outside of the package | 16:39 |
| bryceh | eek | 16:40 |
| Sarvatt | dont see it doing that looking at it though, its just calling update-alternatives to switch | 16:42 |
| bryceh | hmm | 16:42 |
| Sarvatt | what the heck is nvidia-current-updates_hybrid.conf | 16:42 |
| kklimonda | what does hybrid-detect do? | 16:43 |
| Sarvatt | https://bugs.launchpad.net/ubuntu/+source/jockey/+bug/956091 is the same bug | 16:43 |
| ubottu | Launchpad bug 956091 in jockey (Ubuntu) "Jockey does not remove 'nouveau' driver blacklist when deactivating nvidia driver" [Undecided,New] | 16:43 |
| mlankhorst | yikes | 16:43 |
| Sarvatt | where is this _hybrid suffix coming from? | 16:43 |
| bryceh | https://wiki.ubuntu.com/X/HybridGraphics | 16:44 |
| bryceh | unfortunately that page doesn't document the _hybrid bits | 16:44 |
| bryceh | but I propose we add Q's there for alberto to fill in | 16:44 |
| Sarvatt | the alternatives setting up the modprobe.d conf file to /etc dont know about whatever this _hybrid.conf version is so its getting left there | 16:46 |
| kklimonda | ah, interesting - I've wrote something similar to switch betwee nvidia and virtualbox gfx, but put it into initramfs as it seemed easier to avoid potential race between changing gpu and launching lightdm | 16:47 |
| Sarvatt | i dont see hybrid-detect setting any thing special up there, its just calling update-alternatives and ldconfig to switch between mesa and nvidia if the gpu changes since the last boot | 16:47 |
| Sarvatt | tseliot: any clues whats going on? | 16:48 |
| kklimonda | ah, I can see how starting works now - clever little thing, I wonder if systemd has something similar | 16:49 |
| tseliot | Sarvatt: doesn't it work if you remove the package with --purge? | 16:49 |
| Sarvatt | oh ok nvidia changed a bit, ./debian/rules: $(CURDIR)/debian/$(PKG_driver)$(sysconfdir)/modprobe.d/$(PKG_driver)_hybrid.conf | 16:49 |
| tseliot | oh and in the postrm I have rm -f /etc/modprobe.d/#DRIVERNAME#_hybrid.conf | 16:50 |
| tseliot | this was introduced in 295.33-0ubuntu1 | 16:52 |
| tseliot | to fix bug #958848 | 16:52 |
| ubottu | Launchpad bug 958848 in NVIDIA Drivers Ubuntu "nouveau remains blacklisted after uninstall of nvidia drivers" [Undecided,New] https://launchpad.net/bugs/958848 | 16:52 |
| Sarvatt | i wonder if they are installing from a ppa without that change, the postrm looks right to me | 16:57 |
| Sarvatt | and i cant find any version info in the bug | 16:57 |
| bryceh | Sarvatt, yeah both the bug and dupe were filed against jockey which (ironically) doesn't include version info for nvidia | 16:59 |
| Sarvatt | but that doesnt make sense, 295.40 went into precise right away | 16:59 |
| Sarvatt | installing nvidia and removing it to see if i can reproduce | 16:59 |
| Sarvatt | both apt-get remove and apt-get purge removed it but i can't test jockey deactivate because no nvidias here | 17:02 |
| Sarvatt | err, i guess it only happens on actual hybrids even? lrwxrwxrwx 1 root root 47 Apr 16 13:00 nvidia-graphics-drivers.conf -> /etc/alternatives/i386-linux-gnu_nvidia_modconf | 17:03 |
| Sarvatt | bryceh: oh rereading it and the key word is disabled | 17:09 |
| Sarvatt | * debian/rules: | 17:09 |
| Sarvatt | - Make sure that nouveau is blacklisted even when nvidia | 17:09 |
| Sarvatt | is disabled. This is necessary in order to get hybrid | 17:09 |
| Sarvatt | graphics to work. | 17:09 |
| Sarvatt | sounds like tseliot intentionally did that | 17:09 |
| tseliot | Sarvatt: yes, otherwise nvidia couldn't be loaded | 17:11 |
| bryceh | ah | 17:11 |
| bryceh | hmm | 17:11 |
| tseliot | and I can't rebuild the initramfs on boot after blacklisting nouveau and require a reboot ;) | 17:12 |
| tseliot | Sarvatt: if the file is removed both with and without --purge I guess your problem is solved as Jockey simply calls apt-get remove | 17:13 |
| tseliot | through libapt | 17:13 |
| Sarvatt | tseliot: the file isnt even getting installed here | 17:13 |
| Sarvatt | thats why remove/purge works | 17:13 |
| tseliot | Sarvatt: can you try apt-get install --reinstall nvidia-current | 17:14 |
| tseliot | or, even better, purge it and then install it again | 17:14 |
| tseliot | if the file is removed when the package is upgraded then something's wrong | 17:15 |
| Sarvatt | i just did that but trying again | 17:15 |
| Sarvatt | now i have the _hybrid one, wth! | 17:17 |
| tseliot | :) | 17:17 |
| tseliot | the mysteries of apt... ;) | 17:17 |
| tseliot | I should probably make sure that the file is there, somehow... | 17:18 |
| tseliot | but I wouldn't know how to reproduce it | 17:18 |
| Sarvatt | well a lot of people are complaining because they can't use bumblebee now because of it | 17:20 |
| Sarvatt | by a lot i mean 1 that i've seen | 17:20 |
| tseliot | Sarvatt: :) | 17:40 |
| tseliot | Sarvatt: why? Does Bumblebee use nouveau? | 17:41 |
| tseliot | I thought it used intel + nvidia | 17:41 |
| jcristau | i thought it used duct tape | 17:42 |
| tseliot | that should work too :P | 17:42 |
| * mlankhorst doesn't know ;s | 17:42 | |
| * tseliot -> dinner | 17:44 | |
| cprofitt | Sarvatt: is there anything else I need on the bug https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers/+bug/982710 | 18:15 |
| ubottu | Launchpad bug 982710 in nvidia-graphics-drivers (Ubuntu) "Nvidia 295.40 - Ubuntu 12.04 Beta 2 Fresh Install (slow performance in Unity 3D)" [Medium,Triaged] | 18:15 |
| cprofitt | I looked at the triage documentation and did not see anything special for video cards... | 18:16 |
| cprofitt | so marked it triaged... but if you see something that is needed please let me know | 18:16 |
| cprofitt | thanks again for you assistance yesterday | 18:18 |
| bryceh | cprofitt, you should mention the gpu lockup on your report to the upstream forum | 18:26 |
| bryceh | cprofitt, I assume you did a purge and reinstall? | 18:27 |
| tjaalton | bryceh: forgot to git add the xserver patch? | 18:40 |
| tjaalton | ah no, it was added earlier | 18:42 |
| tjaalton | is that patch upstream? | 18:43 |
| tjaalton | assuming it isn't, added a note to the bug | 18:45 |
| bryceh | tjaalton, drive-by contrib presumably | 18:48 |
| tjaalton | right | 18:57 |
| bryceh | if he doesn't send it in, I'll try to remember to forward them later. there's a handful of other nullptr patches in our tree that should go up. | 18:58 |
| bryceh | tjaalton, I quit bothering sending them in when peter kept rejecting them as "just papering over problems" ;-) | 18:58 |
| tjaalton | bryceh: yeah, the next cycle would be great time to try again :) | 18:59 |
| bryceh | tjaalton, what I think we may want to do is once 12.10 opens, any of the nullptrs that look like they got fixed some other way upstream, drop the patches and see if the bugs come back. They'll be trivial enough to diagnose if they come back. | 18:59 |
| tjaalton | yeah | 18:59 |
| cprofitt | bryceh: I actually did a purge and reinstall... then a complete fresh install and only installed the video driver | 19:00 |
| bryceh | cprofitt, ok | 19:01 |
| bryceh | cprofitt, I think next step may be to wait on word from your upstreamed bug (thanks for doing that) | 19:02 |
| bryceh | cprofitt, sounds like workaround is to downgrade to .33 or .20? | 19:02 |
| cprofitt | bryceh: yep -- going to try that as well... though may not get to it tonight -- Cub Scouts (I am the lucky Den Leader) | 19:07 |
| cprofitt | by the way you guys in here have been fantastic... it is much appreciated. | 19:07 |
| bryceh | cprofitt, Be Prepared! | 19:08 |
| * cprofitt smiles | 19:08 | |
| cprofitt | yep | 19:08 |
| jcristau | cnd: do you plan on sending the displayid patch to keithp? | 19:40 |
| jcristau | err displayfd | 19:40 |
| cnd | jcristau, he's Cc'd | 19:42 |
| cnd | do I need to do anything else? | 19:43 |
| jcristau | hmm. usually not. | 19:45 |
| Sarvatt | RAOF: http://article.gmane.org/gmane.comp.freedesktop.xorg.devel/29956 | 19:47 |
| FernandoMiguel | boua noute | 21:00 |
| bryceh | Sarvatt, bug #982889 is another "X starting faster than drm is ready" bug, but with -intel | 21:10 |
| ubottu | Launchpad bug 982889 in xserver-xorg-video-intel (Ubuntu) "xorg fails to start after boot on core i5" [Undecided,New] https://launchpad.net/bugs/982889 | 21:10 |
| Sarvatt | [ 2.441] | 21:11 |
| Sarvatt | X.Org X Server 1.11.3 | 21:11 |
| Sarvatt | are you freaking kidding me? | 21:11 |
| * Sarvatt drools on that person's SSD speed :) | 21:12 | |
| bryceh | and it's done by 2.555 | 21:13 |
| bryceh | 0.1sec X.org | 21:13 |
| Sarvatt | lightdm used to try starting x on 6 displays before giving up but it changed to just 1, sure would be nice to get that old functionality back to avoid these bugs since i have no clue how to fix it in the upstart scripts | 21:14 |
| bryceh | it's an Intel Z68 board | 21:18 |
| bryceh | wonder if he's OCing it | 21:19 |
| Sarvatt | i915 is ready 2.76 seconds in, x gives up 2.55, sheesh | 21:30 |
| bryceh | Sarvatt, having lightdm retry 6 times seems a bit hacky | 21:38 |
| bryceh | Sarvatt, I think upstart is the best way to do it, but if it can't be done that way, and if having lightdm do it is too hacky, could we just put the retry into X itself? | 21:38 |
| bryceh | whereever it's trying to load the drm driver, if it isn't there make it sleep 1 sec, try again, repeat 5-10 times before giving up? | 21:39 |
| bryceh | hmm, ok apparently ^that is the way to do it, as per slangasek and apw. | 22:17 |
| apw | bryceh, i am not sure how long the load can take, i'd probabally do a couple immediaitly before any sleeping | 22:20 |
| apw | bryceh, it would cirtainly be good to know it was that EAGAIN you were hitting either way | 22:20 |
| bryceh | apw, unfortunately I *think* the code that loads the kernel module is in the ddx drivers | 22:21 |
| bryceh | in which case, this'd be doable for -intel but no go on the blobs | 22:21 |
| bryceh | ok, looks like the logic in question is in libdrm. hrm. | 22:29 |
| Sarvatt | looks like cprofitt isnt the only one http://phoronix.com/forums/showthread.php?70452-Did-The-NVIDIA-295-40-Linux-Driver-Fall-Off-A-Cliff&p=258872#post258872 | 23:10 |
| bjsnider | just with the 8800gts? | 23:17 |
| Sarvatt | yeah looks like old IGP ones are busted, and 320mb versions of 8800gts are slow from all the reports | 23:36 |
| Sarvatt | cnd: any way to disable the 4 finger tap gesture? | 23:57 |
| cnd | Sarvatt, not that I know of | 23:57 |
| Sarvatt | i'm bringing up the dash just from 2 finger scrolling too much | 23:57 |
| cnd | why? | 23:57 |
| cnd | hmm... | 23:57 |
| cnd | that's odd | 23:57 |
| cnd | are you accidentally touches with four fingers? | 23:57 |
| cnd | touching* | 23:57 |
| Sarvatt | probably because i use the middle and ring fingers and the pinky and index are hovering over the pad | 23:57 |
| Sarvatt | yeah accidental 4 finger tap | 23:58 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!