[01:37] RAOF, hey how are things looking with nouveau at the moment? got anything worth testing? [06:35] Hm. Testing the nouveau->nv->vesa fallback patch would work better if I had xserver-xorg-video-nv or xserver-xorg-video-vesa installed. [07:01] Strange. It'll fallback to VESA, but not to nv. [07:04] logfile? [07:13] Bah, it got overwritten. I'll get a new one. [07:21] tjaalton: http://pastebin.com/f3d8d0b6c [07:34] RAOF: huh, that's funny [07:38] Yes. [07:41] tjaalton: http://pastebin.com/f2ab1e6fc is a log with the nouveau kernel module available, if that's interesting to compare. [07:42] that looks normal [07:43] If the vesa DDX isn't available & the nouveau kernel module isn't available, then X just fails to load. [07:56] RAOF: doesn't it use fbdev instead of vesa? [07:56] tjaalton: ^^ [07:57] tseliot: Not in my log; X can't load that module (because I don't have the package installed for some reason). [07:59] I think fbdev should be our first fallback [08:00] A can change the patch if that's the case. Will fbdev actually work well, though? I'm not familiar with it, but isn't our framebuffer going to be vga16fb? [08:11] vesa doesn't work at all with kms [08:16] But if we're falling back to vesa in this case then that means that nouveau has failed, which probably means that kms isn't available. [08:26] it can fail for other reasons too [08:26] like intel [08:26] or ati [08:41] I've got wayland built [08:42] Does that have any input yet? [08:45] haven't booted it yet. maybe not [08:45] Heh. [08:45] just getting it to build is a challenge [08:46] There's now something worth testing in ppa:xorg-edgers/nouveau [08:47] awesome, I'll todo that for tomorrow [08:47] have you had a chance to test it? any tips? [08:47] bryceh: are we going to have nvidia cards at the sprint? [08:48] tseliot, I can bring some, but it'd be for a desktop system [08:48] right [08:48] bryceh: Yeah, I've tested it with my laptop; it works :) [08:50] Well, apart from suspend, because my laptop still doesn't quite manage to actually suspend with the Lucid kernel. [08:51] tseliot, I'm going to be coming from home each day so if it turns out we want a desktop box we can slap cards into, I can bring one. and an lcd monitor, kbd, mice, etc. too [08:52] bryceh: right, I forgot that you live near Portland [08:52] * RAOF goes to check on the status of his kernel bisect for that. [08:53] bryceh: I was asking as I'm not sure that I'll be able to test Jockey today and I'll leave tomorrow [09:16] gahh buildd's too slow [09:17] "Start in 49 minutes (2505)" [09:18] why not test it locally? [09:22] tjaalton, it builds fine locally [09:22] so autoreconf helped? [09:22] I didn't have the original problem locally [09:22] hmm ok [09:23] well, but I'm not building in a clean pbuilder environment [09:29] tseliot: its happening again :( gdm log is logging X errors :( [09:30] * vish reboots again without splash [09:40] same error repeats http://paste.ubuntu.com/365081/ , not sure how it starts :s [09:41] not sure why gdm logs it , but i'v noticed one problem , conky stops being displayed for some reason , even if i restart it it wont be displayed [09:46] tjaalton, rats. make: autoreconf: Command not found [09:47] guess tseliot was right, probably needs explicit automate build-dep [09:48] autoconf [10:01] anyone? any suggestions on how to debug the above error ^ also KMS doesnt seems to setup from xorg.1.log > http://paste.ubuntu.com/365090/ [10:01] seem* === ripps|sleep is now known as ripps [10:04] vish: I think I missed the 1st error. I can only see the one about KMS [10:05] tseliot: the same error i showed you couple of days ago> http://paste.ubuntu.com/365081/ [10:05] not sure why gdm logs it , but i'v noticed one problem , conky stops being displayed for some reason , even if i restart it it wont be displayed [10:06] it seems to start after a few hours ... I dont notice any errors when the system is started [10:07] the gdm logs just eat my / space :s [10:08] * tseliot looks for hints in the -ati driver [10:09] thanks :) [10:55] vish: do videoclips work? [10:56] tseliot: yeah , i can watch videos [11:03] vish: maybe you need this patch? http://cgit.freedesktop.org/xorg/driver/xf86-video-ati/commit/?id=e1200cb89218930d01330ba0114e013438655cce [11:03] which reverts another commit that we have in Lucid [11:05] yeah , maybe , tseliot when was the earlier commit? [11:05] this is something very recent , has been happening since less than 10days [11:07] tseliot: umm, that's old [11:07] vish: check the kernel changelog then [11:07] tjaalton: that commit, yes, but we still don't have it [11:09] tseliot: how come? the snapshot is from master [11:18] kl;' [11:24] kl;' [11:31] * vish hmm... still trying to figure out what starts it :s [11:32] vish: how much ram does your card have? [11:34] 2GB [11:36] tseliot, btw got cairo-drm built [11:36] bryceh: ah, nice, what was it? [11:36] vish: really? What does the bios say? [11:37] you put me on the right track. just kept throwing deps at it. libtool, automake, autoconf... [11:38] tseliot: ah.. the radeon card is 512mb [i guess] let me check.. how do i check BIOS from session? [11:38] the easiest way is to just enter the bios [11:39] * vish brb [11:39] bryceh: which explains why it was building in your environment and failed in buildds [11:40] exactly [11:41] tjaalton: I guess airlied fixed that in master (after that revert) [11:42] bryceh: it looks like a few macros in -ati trigger these errors: http://paste.ubuntu.com/365081/ [11:43] e.g. BEGIN_ACCEL, FINISH_ACCEL, etc. [11:44] I'm wondering if we're missing something is mesa [11:46] weird [11:46] anyway I'm going to bed [11:46] good night [11:46] looks like I've got wayland in the bag, but a few odds and ends to finish up tomorrow [11:47] sounds good [11:47] apw: I think we'll be on the same flight to Portland (from Washington DC) [11:52] tseliot: I'm confused.. fixed in master means we have it in lucid [11:52] tjaalton: right, so I was taking back what I said before ;) [11:53] tseliot: hehe... 128 MB .. [with some humbug about being upto 512mb HyperMemory] so 128 i guess ;D [11:53] Okay, not cool. I've lost my glx extensions in X. I can't start compiz [11:55] tseliot: ah :) [11:56] glxinfo just spits out: Xlib: extension "GLX" missing on display ":0.0" a bunch of times [11:56] ripps: log? [11:56] tseliot: which one? [11:56] bryceh: the same build-deps as on every X module :) [11:56] which run autoreconf on build [11:56] ripps: Xorg.0.log [11:57] automake, libtool should be enough [11:58] tseliot: http://pastebin.com/f666e8286 [11:59] ookkaayy,,nnooww mmyy kkeeyybbooaarrdd hhaass ggoonnee ssccrreewwyy!! [12:01] there, undid an option in keyboard prefences, but I can hold down backspace to delete anymore, I have to press it a dozen times. Wth is going on? [12:01] ^can^can't [12:01] ripps: that was something wrong with the > keybindings reported an issue with the keybinding code not loading yes [12:01] ops bad quote :/ [12:03] ripps: just set the keypress setting again ;) and all will return [12:03] ripps: try with sudo update-alternatives --config gl_conf and select mesa (if you're using ati), then sudo ldconfig [12:03] and reboot [12:04] tseliot: ah that was it, it was trying to load it from nvidia-current [12:05] vish: turning on repeated keypressess does this: hheelllloo wwoorrlldd [12:05] and backspace still doesn't work correctly [12:06] ripps: why do you have nvidia installed? [12:06] ripps: weird , well for me and a few others the slider was all the way to the left... switching the setting back on and adjusting the slider fixed it... [12:06] tjaalton: I don't know, it installed itself with the recent lucid updates [12:06] * tseliot -> lunch [12:07] ripps: ok, that's a bug then [12:07] should I remove it? [12:07] yes [12:08] tjaalton: oh, i see what happened, my latest mplayer ppa package is pulling it in. Why would it do that? [12:09] probably broken libvdpau depends [12:09] vdpau is supposed to be optional, not forced [12:09] it's never forced an nvidia install before, just the one I built an hour ago... [12:10] show the deps [12:11] http://pastebin.com/f320e864b [12:11] Depends: nvidia-current [12:12] the package is automatically making nvidia-current a depend, how do I change it to a recommend? [12:12] or rather, a suggest [12:12] fix the source instead [12:12] it shouldn't need to do that [12:13] there's libvdpau1 you know.. [12:15] tjaalton: but libdpau1 is only available for lucid, what do i do with my hardy->karmic backports? [12:15] ripps: your problem :) [12:17] vish: yeah that worked. I wonder how they got set all the way to the left. [12:57] whew... everything worked after I uninstalled nvidia-current. I had to downgrade my mplayer though. I'll have to test some builds to see if I can fix the nvidia-current dependency. === albert231 is now known as albert23 [14:40] Hello. [14:41] Sarvatt: I tried the mainline kernel. It does not load the ath5k driver which I require for wifi connection; modprobe loads it but it does not find the appropriate interfaces, and without that, could not use the most common test-case to reproduce the crash. === yofel_ is now known as yofel [18:05] Cobalt: Thats no good, do you have something installed that would interfere with it like madwifi? does grep -R "ath5k" /etc/modprobe.d/ show anything blacklisting it? you might want to try a newer kernel like 2.6.33-rc5 out or try using the lucid kernel instead of the kernel-ppa ones, I do have the same problem as you when I try to use the karmic .31 kernel with newer git libdrm and intel (invalid alignment 4096 errors) and I'm pretty sure it i [18:05] s kernel related. To be honest I would recommend just putting the intel driver on hold with 2.9.1 and still using edgers for mesa and such, thats what I'm doing currently because of all of the problems [18:06] Sarvatt: No, I don't. Just Wicd, but that wouldn't interfere. [18:06] Sarvatt: grepping shows nothing. [18:07] morning [18:07] oh thats another can of worms, i have no idea about that because I use network-manager [18:07] Sarvatt: The problem occurs before that. There are no wireless interfaces. Wicd is not that low-level. [18:08] Sarvatt: I've been withholding upgrading the driver as well, but as it appeared after one of the updates, I thought there was a regression introduced. [18:08] does iwconfig show the network? [18:09] Sarvatt: No, it does not. [18:09] was just thinking maybe it changed from ath0 to wlan0 recently [18:09] There wasn't _any_ wireless interfaces of any sort. :) [18:10] AAAAHHGGHGT. [18:10] * Cobalt headdesks. [18:10] Sarvatt: You know why no network interface was showing up? That's cos I don't have an Atheros NIC. I have Ralink. Damn, I need to reboot that thing and find out. [18:11] Sarvatt: Having said that, it also complained about /proc/mount not working and it was not able to load my Apparmor profiles. [18:11] ahh that'd explain it [18:11] try the lucid kernel [18:11] staging is disabled for the mainline kernel-ppa builds [18:11] yeah theres no apparmor in those [18:11] its a clean kernel, apparmor is a ubuntu patch and not upstream [18:12] You have an address handy for those later kernels? [18:12] try 2.6.32-12 from the lucid archives [18:12] What about /proc/mount, what is that, and do I need it? [18:12] should be on packages.ubuntu.com [18:13] Let me have a look at that. [18:14] I had like 15 kernels installed. I pruned it quite a bit, saved about 800MB. So much clutter since I had Hardy in this box. [18:14] /proc/mount? do you mean /proc/mounts? [18:14] packages.ubuntu.com isnt even loading for me, i have a really spotty connection where i am [18:15] RAOF: try adding an alias for lbm-nouveau to nouveau? [18:15] Something like that, I didn't grab the exact error message. What does /proc/mounts do? Is that a list of current mounts? [18:15] hmm drmOpen is looking for "nouveau" [18:15] yeah it is [18:16] the lucid kernel should work for you if its just the ralink problem, that would break because its a driver in staging in the kernel and thats not enabled [18:16] in those mainline kernel builds [18:16] I see. And Ubuntu includes it for greater hardware support? [18:18] RAOF: maybe something like this? http://sarvatt.com/downloads/0001-lbm.patch [18:19] yep, stuff in staging breaks constantly during the -rc cycle and gets fixed up before the final release usually, would be a major pain to track whats broken and rebuild a bunch of times for the daily kernel builds [18:20] I'm new to all that kernel stuff. Wouldn't have known about it. [18:21] alot of the changes for intel are in the kernel, if you're using edgers its usually a good idea to keep that up to date too :D [18:21] Do I need to mess with the linux-firmware package? [18:22] I couldn't NOT use Edgers either, on that Eee PC, the stock drivers offered like 5FPS at best, with Compiz. And that was as fluid as anything in Hardy, if you didn't go too crazy with effects. I just use a tiny amount of transparency and whatnot. [18:22] what kernel are you using now? -generic -generic-pae or x64? [18:23] intel on karmic shouldnt be bad at all, there shouldnt be much of a difference using edgers. jaunty on the other hand... [18:23] Lucid does not seem to have a headers package. I'm using -generic. My 1000H is 32-bit. [18:24] do you need headers? building external modules? [18:24] http://packages.ubuntu.com/lucid/i386/linux-image-2.6.32-12-generic/download [18:24] Sarvatt: Yeah, for VirtualBox, via dkms. [18:25] http://packages.ubuntu.com/lucid/i386/linux-headers-2.6.32-12-generic/download [18:25] I'm actually getting that package your link is pointing to. [18:25] http://packages.ubuntu.com/lucid/i386/linux-headers-2.6.32-12/download [18:25] err well [18:25] theres a bit more of a problem there [18:25] What kinda? [18:25] i dont think the karmic virtualbox works with 2.6.32+ [18:26] It seemed to compile the modules without complaining, however, I didn't use it. Still, I've not used it much lately (I have it to run XP for Nokia PhoneManager to back stuff up). Been lazy. [18:26] the full featured ones from the virtualbox website work though [18:26] I could always boot to something lower just for that. [18:27] Ah I see. That's what I use. Not the -OSE. [18:27] gotta run, too much time sitting in the car and i'm late for the next job, good luck with it :) [18:27] Thank you again. Will drop in a line about what's going down. [18:44] morning bryce! missed you saying that in all the talk :D [18:44] apw: are your changes to add lbm-nouveau in git anywhere? [18:45] Sarvatt, https://edge.launchpad.net/~apw/+archive/blue/+packages [18:46] how can i see the driver of my graphic card? [18:46] your /var/log/Xorg.0.log [18:48] is there a specific line? [18:57] it should be really obvious in the first 2 or 3 pages of it, if you can put it on pastebin we can tell you [18:59] this lbm-nouveau stuff really looks like it will cause more problems than just building it with dkms, packing in a whole new drm and such with different names, is it really removing the old drm/ttm and such when it loads? i dont have my nvidia machine handy to mess with it to see [19:00] s [19:01] Sarvatt, i have 3 log's Xorg: Xorg.0.log Xorg.0.log.old Xorg.20.log. What do i put in pastebin? [19:01] 0 [19:02] first one [19:03] Sarvatt, here: http://pastebin.com/m76625838 [19:04] the i915 intel driver [19:05] version 2.6.3 [19:05] version i find [19:05] what line is i915? [19:06] around line 107, its the xserver-xorg-video-intel driver [19:08] i highly recommend upgrading to karmic though, there were quite alot of problems with intel on jaunty [19:10] Sarvatt, it's why i'm asking this to you [19:10] my ubuntu looks "strange" [19:10] maximizing and minimizing is slow [19:11] everything looks slow [19:11] yeah that should be *much* better on karmic [19:11] if upgrading is an option at all [19:11] i will try [19:12] you can try a karmic livecd out and see how much better it will be [19:12] because, this way is terrible [19:13] it was unfortunate but jaunty was released when intel was doing some major restructuring with their drivers and there were quite alot of problems, there are some things you can do to work around it if you want to stick with jaunty though [19:14] and the version 8.10, do you know about it? [19:16] sorry, gotta run for a few minutes but i'll give you some xorg.conf options that will help on jaunty when i get back [19:16] 8.10 is good but 9.10 fixes most of the problems from 9.04, i'd recommend upgrading instead of wiping and downgrading :D [19:17] i understading [19:17] because i work for a project, and we use 8.10 to develop [19:18] Sarvatt, do you have to go out? [19:42] herman_: the "Performance regressions on intel graphics" section on this page might help - http://www.ubuntu.com/getubuntu/releasenotes/904 [19:43] the Option "MigrationHeuristic" "greedy" part in particular helped alot in those days on my system with firefox scrolling speed [19:44] yeah I had to drive to my next job but I'm back now for a bit :) [19:46] from a terminal, type sudo gedit /etc/X11/xorg.conf and add Option "MigrationHeuristic" "greedy" at the end of the Device section (but before EndSection) and reboot [19:47] Shouldn't logging out and back in again do the trick? [19:48] need to restart the xserver [19:50] Sarvatt, i don't want to disrupt you [19:51] thanks for the help until here [19:51] i will try update my ubuntu to karmic koala [19:51] i'll find you here other days [19:52] Well, logging out and in seems to do that. [19:52] Gonna reboot to that new kernel. [20:02] Gah. Got lazy instead. [20:11] Sarvatt, hey, my ubuntu feels a bit better [20:11] i think === Sinnerman is now known as Cobalt [21:09] Sarvatt: Interestingly, it does load my rt2860sta module. But it does not automatically put my wlan0 interface up. Maybe because the scripts originally look for ra0. Once I do that, things go okay. I still need to test the latest drivers now. [22:44] my kid's first experience hacking on Ubuntu: http://www2.bryceharrington.org:8080/files/images/ubudutch.jpg [22:49] a future packager! [22:58] hack the planet! [23:07] bryceh: laptop on the duvet! think of the airflow! ;o