| bryceh_ | RAOF, happen to know what the procedure is for switching between fglrx and ati? The only howtos I ran across looked rather hairy (and a bit oldish) | 00:20 |
|---|---|---|
| RAOF | bryceh_: IIUC the ati control panel has a button that you press. | 00:21 |
| bryceh_ | hmm, ok | 00:21 |
| bryceh_ | I suppose if people have problems we just point them at AMD... | 00:22 |
| bryceh_ | RAOF, btw sarvatt didn't think it'd be worth having people test that -intel patch quite yet | 00:25 |
| RAOF | Ah, ok. | 00:25 |
| bryceh_ | <bryceh_> Sarvatt, did that -intel patch get taken care of by anyone already? Should I roll up a ppa and solicit testers? | 00:25 |
| bryceh_ | <Sarvatt> bryceh_: nope not yet, it'll need a bit of cleaning up for 2.14.0 because of has_kernel_flush | 00:25 |
| bjsnider | bryceh_, wouldn't jockey allow you to deactivate fglrx at which time you'd be using ati by default? | 00:30 |
| === yofel_ is now known as yofel | ||
| Sarvatt | bryceh_: oh I just meant we're gonna need to backport the commit and cant just cherrypick it, we definitely do need to get it in I was just busy at the time | 01:10 |
| Sarvatt | or we could just revert http://cgit.freedesktop.org/xorg/driver/xf86-video-intel/commit/?id=cc930a37612341a1f2457adb339523c215879d82 :P | 01:11 |
| Sarvatt | bryceh_: you mean like for the hybrid stuff in fglrx now? that runs on scripts we dont ship that rename libs. | 01:12 |
| Sarvatt | (so it doesnt do anything) | 01:15 |
| Sarvatt | its really incompatible with how we package fglrx, not sure whats going to happen with it but tseliot was discussing it with them and might have some info on it. they back up and replace the system libs when you pick ati and move things back when you want to switch to intel | 01:20 |
| Sarvatt | it also doesnt work for switching between fglrx and ati, only for fglrx and intel hybrids | 01:20 |
| bryceh_ | Sarvatt, ok yeah I suspected this as the case | 01:25 |
| === soreau_ is now known as soreau | ||
| mdeslaur | hmm...is it a known issue that with today's iso, jockey shows the nvidia driver as being installed, even though I am clearly using nouveau? | 14:27 |
| tseliot | mdeslaur: what's the output of "jockey-text -l" ? | 14:30 |
| mdeslaur | tseliot: it says "xorg:nvidia_current - blah blah (Proprietary, Enabled, In use)" | 14:30 |
| Sarvatt | have you rebooted since installing it? | 14:30 |
| mdeslaur | and underneath "xorg:nouveau - blah blah (Free, Disabled, Not in use)" | 14:31 |
| Sarvatt | (stupid question but have to ask, sorry) | 14:31 |
| mdeslaur | Sarvatt: yes | 14:31 |
| tseliot | mdeslaur: what about the output of "update-alternatives --display gl_conf" ? (please put it on pastebin) | 14:31 |
| mdeslaur | Sarvatt: oh, I didn't change _anything_ in jockey...that's what it came up as by default | 14:31 |
| tseliot | mdeslaur: also your jockey.log in /var/log/ might help | 14:32 |
| mdeslaur | tseliot: ok., hold on...I'll file a bug and put all those in it | 14:33 |
| tseliot | thanks | 14:34 |
| mdeslaur | tseliot, Sarvatt: bug 759804 | 14:37 |
| ubot4 | Launchpad bug 759804 in jockey (Ubuntu) "after fresh install, jockey thinks nvidia binary driver is in use (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/759804 | 14:37 |
| tseliot | mdeslaur: you're still using nvidia, at least its GL libraries | 14:38 |
| tseliot | you can't be using nouveau as it's blacklisted when nvidia is enabled | 14:39 |
| tseliot | so, in short, Jockey is right ;) | 14:40 |
| Sarvatt | the nvidia kernel module isn't built | 14:40 |
| mdeslaur | tseliot: did you look at the xorg log file? | 14:40 |
| tseliot | mdeslaur: it seems that the kernel module wasn't loaded. What does "dkms status" say? | 14:42 |
| mdeslaur | tseliot: it doesn't say anything | 14:43 |
| tseliot | mdeslaur: please try "sudo dpkg-reconfigure nvidia-current" and see what happens. I bet that somehow dkms failed | 14:44 |
| mdeslaur | Error! There are no instances of module nvidia-current 270.30 located in the DKMS tree | 14:44 |
| mdeslaur | ok, it seems to have built it now | 14:45 |
| mdeslaur | I'll try rebooting | 14:45 |
| tseliot | ok | 14:46 |
| mdeslaur | ok, looks like I'm running -nvidia now | 14:47 |
| mdeslaur | hmm...not a nice bug | 14:47 |
| janimo | RAOF, I see you commented on this one https://bugs.freedesktop.org/show_bug.cgi?id=32254 . Is the fix supposed to be in natty, the comment does not mention which branch the fix went into | 14:48 |
| ubot4 | Freedesktop bug 32254 in Other "EGL+OpenGL API failed to work" [Normal,Resolved: fixed] | 14:48 |
| tseliot | mdeslaur: can you close your bug report, please? I don't think we can figure out what happened before... | 14:49 |
| mdeslaur | now jockey says "This driver is activated but not currently in use" | 14:49 |
| mdeslaur | tseliot: I can reinstall and reproduce at will I bet you | 14:49 |
| Sarvatt | janimo: because it went into master, it'll be in 7.11, no it wont be in natty | 14:50 |
| janimo | Sarvatt, ah thanks. Too bad though | 14:50 |
| tseliot | mdeslaur: if you can reproduce it, please attach your /var/lib/dkms/nvidia-current/270.30/build/make.log | 14:51 |
| mdeslaur | tseliot: ok, I'm reinstalling now, and will attach that log | 14:51 |
| tseliot | mdeslaur: as regards the fact that Jockey says that the driver is not in use, I'd like to see your new jockey.log | 14:52 |
| mdeslaur | tseliot: sure, one sec | 14:53 |
| mdeslaur | tseliot: done | 14:56 |
| * tseliot has a look | 14:57 | |
| mdeslaur | tseliot: I have reinstalled, and hit the issue again | 15:18 |
| mdeslaur | tseliot: there is no log file in /var/lib/dkms | 15:18 |
| tseliot | mdeslaur: you must have something in /var/lib/dkms/nvidia-current/270.30/build/ | 15:19 |
| mdeslaur | the only thing I have in /var/lib/dkms is a file called dkms_dbversion | 15:19 |
| mdeslaur | that's it | 15:19 |
| mdeslaur | no other subdirectory | 15:20 |
| tseliot | mdeslaur: your last jockey.log should contain some errors | 15:20 |
| mdeslaur | tseliot: huh? | 15:21 |
| mdeslaur | tseliot: do you want my new jockey.log file? | 15:21 |
| tseliot | mdeslaur: yes, if you installed the driver using jockey | 15:21 |
| mdeslaur | tseliot: I don't touch _anything_ in jockey. I just open the window, and click close | 15:21 |
| mdeslaur | jockey already thinks nvidia is installed | 15:21 |
| Sarvatt | did you check the enable 3rd party software box during the install? | 15:22 |
| mdeslaur | Sarvatt: yes, I did | 15:22 |
| mdeslaur | Sarvatt: would you like me to reinstall without checking it? | 15:22 |
| tseliot | mdeslaur: ok, if you do this manually, please capture the output of the operation so that I can see what dkms says | 15:23 |
| Sarvatt | thats ok, wont help figure out why the heck its getting in the state where it thinks its installed and activated but isnt really, it wont install if you dont check that box | 15:24 |
| mdeslaur | tseliot: you want me to trigger dkms and capture the output? | 15:24 |
| tseliot | mdeslaur: I'd like you to reproduce the issue and capture the output | 15:25 |
| Sarvatt | he just did | 15:25 |
| tseliot | let me check again | 15:25 |
| mdeslaur | tseliot: I'm not sure what you mean....the output of _what_? | 15:25 |
| tseliot | mdeslaur: I guess you remove the driver and then re-install it with apt, right? | 15:25 |
| Sarvatt | tseliot: the problem is when you check the activate the 3rd party drivers box during the install it tries to install the blobs and for some reason its not actually installing it but jockey thinks its installed and activated | 15:26 |
| mdeslaur | tseliot: not at all | 15:26 |
| tseliot | mdeslaur: how do you do it? | 15:26 |
| mdeslaur | tseliot: I don't do _anything_. I install, and then open jockey. That's it. | 15:26 |
| Sarvatt | he only opens jockey to see what the status is and it says activated and in use | 15:26 |
| tseliot | mdeslaur: "install" how? | 15:27 |
| Sarvatt | from a livecd | 15:27 |
| mdeslaur | tseliot: from a livecd (well, from a usb key) | 15:27 |
| tseliot | aah, ok | 15:27 |
| tseliot | mdeslaur: that's not supposed to work | 15:27 |
| Sarvatt | he's installing onto a disk, that part is supposed to work | 15:27 |
| mdeslaur | tseliot: what's not supposed to work? | 15:27 |
| Sarvatt | not trying to install it in a live session where it's not supposed to work | 15:28 |
| mdeslaur | nono, I'm not trying to do anything in a live session....I'm installing a laptop from the livecd | 15:28 |
| tseliot | mdeslaur, Sarvatt: right but while this *should* work it never did. It's likely a bug in the installer which doesn't copy your compiled module | 15:29 |
| tseliot | then when you boot into your installed system you get no nvidia module | 15:30 |
| tseliot | *kernel module | 15:30 |
| Sarvatt | so this has always been a problem and didnt just start happening? | 15:30 |
| tseliot | yep | 15:30 |
| tseliot | (the former) | 15:30 |
| tseliot | this problem is something I can probably fix though: https://launchpadlibrarian.net/69240249/Screenshot-Additional%20Drivers-1.png | 15:31 |
| Sarvatt | cool, thats the main problem | 15:32 |
| Sarvatt | oh wait, thought that was the activated and in use pic | 15:32 |
| mdeslaur | so, for our 60% of users that have an nvidia card, and click the "3rd party" checkbox, they will have no way of activating the -nvidia drivers | 15:32 |
| Sarvatt | what he's saying is when you do a fresh install, nvidia thinks its installed and in use and you cant activate it tseliot | 15:32 |
| mdeslaur | Sarvatt: yes, there are two different issues | 15:32 |
| tseliot | Sarvatt: yes, it's that picture | 15:33 |
| tseliot | yep | 15:33 |
| mdeslaur | issue #1: dkms doesn't get triggered for -nvidia during install. Issue #2: once -nvidia dkms module is rebuilt, jockey thinks -nvidia is not in use | 15:33 |
| tseliot | actually | 15:34 |
| Sarvatt | mdeslaur: #2 might be intentional until you reboot and it'd show activated and in use | 15:34 |
| mdeslaur | Sarvatt: I did reboot between rebuilding the module and opening jockey again | 15:34 |
| Sarvatt | apologies if you're saying its still saying that after reboot, i didn't follow it all | 15:35 |
| Sarvatt | ahh sorry | 15:35 |
| tseliot | issue #1: modules built by dkms are not copied to the installed system. | 15:35 |
| tseliot | issue #2: Jockey doesn't detect the driver as activated and in use in some cases | 15:35 |
| mdeslaur | tseliot: yes, exactly | 15:36 |
| tseliot | mdeslaur: while I can fix issue #2, you should talk to cjwatson about issue #1 | 15:36 |
| tseliot | it might be useful to ping pitti too | 15:36 |
| mdeslaur | I'll add ubiquity to the bug | 15:37 |
| tseliot | mdeslaur: I think it would also be worth updating the bug description with these 2 issues | 15:38 |
| Sarvatt | its gotta be specific to nvidia, otherwise wl wouldn't work and it does | 15:39 |
| tseliot | ah | 15:40 |
| Sarvatt | i need to double check that, i might have been using brcm80211 in these recent installs instead and not have noticed | 15:41 |
| tseliot | it would be interesting to know | 15:45 |
| Sarvatt | bryceh_: think 40-xserver-xorg-video-intel.rules is triggering too much? http://paste.ubuntu.com/593618/ :) | 16:54 |
| bryceh_ | Sarvatt, heh | 18:15 |
| bryceh_ | Sarvatt, probably about time for us to turn it off anyway | 18:15 |
| Sarvatt | running the rule and dumping was triggering these http://paste.ubuntu.com/593614/ and they stopped when i got rid of the udev rule | 18:18 |
| Sarvatt | RAOF: hrm software-center started working without 113_fix_tls.diff again | 18:21 |
| Sarvatt | the tls patch was causing problems on amd64 with mesa master so i haven't been adding it in edgers | 18:22 |
| dbarth | bryceh_: hey bryce; hi | 18:27 |
| dbarth | i'm suddenly getting some x crashes since monday or so, and i can only get that kind of backtrace in the xorg log: http://pastebin.ubuntu.com/593615/ | 18:27 |
| dbarth | no real test case, but the info is that it's almost the first time in the cycle that i've had crashes with my server at all | 18:28 |
| bryceh_ | hmm interesting | 18:28 |
| bryceh_ | dbarth, RAOF posted a new (bugfix) release of mesa on Monday | 18:29 |
| dbarth | ah right, i heard about it | 18:29 |
| dbarth | so do you want me to downgrade and see for a day or two? | 18:29 |
| Sarvatt | not fixed in that, i'm digging up the bug number for that one | 18:29 |
| dbarth | i don't want to add another bug report to the pile with so little info | 18:30 |
| bryceh_ | dbarth, your /var/log/dpkg.log should have the details of what exactly got upgraded; I'd probably start by downgrading packages to isolate what package regressed | 18:30 |
| dbarth | and wait a little bit to see if the sudden crashes go away | 18:30 |
| dbarth | ok, i'll do that | 18:30 |
| dbarth | thanks | 18:30 |
| Sarvatt | dbarth: that backtrace is a known problem, i've only been able to reproduce it closing the lid for >30 minutes here | 18:30 |
| Sarvatt | trying to find the bug for it so you can track it still | 18:30 |
| Sarvatt | only happens in a unity session | 18:31 |
| dbarth | Sarvatt: ah thanks | 18:31 |
| dbarth | uh, /that/ doesn't sound good | 18:31 |
| Sarvatt | dbarth: hrm I take that back, got it confused with the other one where X doesn't crash, https://bugs.launchpad.net/ubuntu/+source/compiz/+bug/740126 | 18:35 |
| ubot4 | Launchpad bug 740126 in compiz (Ubuntu Natty) (and 1 other project) "compiz hangs randomly several times per day (affects: 4) (heat: 24)" [High,Confirmed] | 18:35 |
| Sarvatt | i've seen that backtrace a few times in the past few days though, robert_ancell had it http://paste.ubuntu.com/592987/ checking to see if he filed a bug | 18:35 |
| dbarth | Sarvatt: ah that one is the xcb hang | 18:36 |
| dbarth | (or let me check) | 18:36 |
| dbarth | well, it's an xcb hang | 18:37 |
| Sarvatt | that xcb hang was what I was saying I could reproduce and only had happen in unity, sorry about that :) | 18:37 |
| dbarth | nw | 18:37 |
| dbarth | but i'm aware of 2 of those | 18:37 |
| dbarth | one that was involving an xgetwindowproperty | 18:37 |
| dbarth | and that we workarounded | 18:37 |
| dbarth | and that other one | 18:37 |
| dbarth | but which may actually be triggered by another change we made to fight the stacking issues | 18:38 |
| dbarth | there is a daily build of compiz where you can turn the stackattack otpion off | 18:38 |
| dbarth | and that may help with this bug | 18:38 |
| dbarth | ie not trigger it that often | 18:38 |
| dbarth | Sarvatt: is there a bug that i could subscribe to for the backtrace? | 18:39 |
| Sarvatt | i'm not seeing one, it'd be appreciated if you could file one against xorg-server and we can dupe it later if we find it | 18:42 |
| Sarvatt | neither of the 2 people i've seen the backtrace from ended up filing a report and i'm not seeing a one filed against xorg xorg-server or compiz | 18:44 |
| Sarvatt | guess i should check intel, both of the other reporters were on intel | 18:44 |
| bryceh_ | I don't see one against -intel | 18:46 |
| bryceh_ | jeesh so many bugs | 18:47 |
| bryceh_ | Sarvatt, on https://bugs.freedesktop.org/show_bug.cgi?id=34014 check comment 28 :-/ | 18:49 |
| ubot4 | Freedesktop bug 34014 in Driver/intel "[i915gm] GPU lockup (ESR: 0x00000001 IPEHR: 0x02000004)" [Major,Resolved: fixed] | 18:49 |
| Sarvatt | bryceh_: ugh... | 18:50 |
| === Sarvatt_ is now known as Sarvatt | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!
