[01:36] RAOF, I've committed the glx revert to git; shall I go ahead and upload for pitti to review? [01:37] Is it possible to set a conditional umask based on a regex? So that if a file is created with a name matching pattern A it gets permissions X but if it matches pattern B it gets permissions Y? [01:37] bryceh: Yes please. [01:37] skimj, wrong channel, this is for X.org discussion [01:37] oops sorry. [01:37] wrong tab [01:37] RAOF, ok xserver uploading [01:39] \o/ [01:42] will be nice having that one put to bed. hope there's no serious side effects [01:43] [ubuntu/lucid] xorg-server 2:1.7.6-2ubuntu7 (Waiting for approval) [01:44] We've got a contingency plan in the first upstream commit of the glx-as-resource refactor. [01:44] right :-) [02:46] bryceh: fixed up that intel 2.8 you put in the retro ppa, backported all of the commits it needed to build against xserver 1.7. but I think you got the versions confused because EXA was ripped out before 2.8 :) [02:48] it actually builds against xserver 1.8 too, going to try it out for the heck of it [02:51] I forgot alllll about the cursor flicker :) [03:06] I have a ubuntu ppc problem anyone up for the challenge [03:09] i believe #ubuntu-ports is the channel you want? [03:09] No one in those rooms [03:09] I guess I will try again [03:09] is it X related? [03:10] in that my display is limited to 256 in x [03:11] any attempts to change the xorg fle causes the system to hang at boot [03:12] You're limited to an 8 bit display? [03:12] yes and the display option show no other listing in the drop downs [03:13] What version of Ubuntu? [03:13] 9.04 [03:13] 9.04 ppc [03:14] ryan__: ati? [03:14] yes i have a imac g3 [03:15] an [03:15] the jaunty kernel was screwed up for ppc with ati [03:15] can you upgrade to karmic? [03:15] hmm i had to use a no-video command to install [03:17] i can't remember the specifics but I had the same problem, the powerpc kernel for jaunty was screwed up though due to some patches brought in just before release [03:18] I guess I can try installing 9.1 with the update manager and see wha happens [03:18] they helped out x86 but screwed up powerpc :( [03:18] ah you have a rage 128 in that thing [03:18] ryan__: http://tjworld.net/wiki/Linux/Ubuntu/PowerPC/AppleiMacG3 [03:19] (first hit on google...) [03:19] the "Bug: Xorg doesn't use high-resolution true-colour mode" section specifically [03:22] i have attempted to change the xorg before [03:22] any changes cause the boot to freeze [03:22] openfirmware/yaboot are pains in the butt to use with linux, i have a crazy boot stanza to fix consoles [03:22] i then have to escape to a shell [03:22] (on my ibook) [03:23] are you booting with nosplash? [03:23] so would it be advised to update to 9.1 before attempting changes [03:23] well the problems I was talking about were limited to radeon I believe, I have a radeon 9200 in my ibook [03:24] it boots with splah i suppose [03:30] ok weird i was attempting to modify my yaboot file and i show no files in mnt directory [03:30] oh [03:30] reading up a paragraph [03:30] doh [03:32] imac g3 is a bit old [03:35] hell yes it is [03:35] i was running it for my kids with panther and panther ate it [03:38] ok sarvatt i canged the yah boot for no splash screen and odified the xorg i am going to reboot we'll see [03:39] osx 10.3 ate it? i'm scared how slow even ubuntu is going to run on it :) [04:51] Sarvatt: the xup2 seems stable , but as Conn mentions it seems to grow much slower.. a lot slower :s [04:52] Sarvatt: btw , new_pll=0 still causes jumping screen in new session and guest sessions [04:52] with xup2^ [04:52] that persons problems are more likely specific to ati and low memory systems and not that bug [04:54] ah k.. but the jumping/dancing screen is present with -xup2 [04:55] dont know what you're saying, it was fixed without it and now it isn't? [04:56] or new_pll=0 didnt really fix it? [04:57] Sarvatt: new_pll=0 did fix the problem .... it is not a problem with the 2ubuntu1 but i get it with -xup2 , i recall it is also a problem with the 2ubuntu5[or the latest lucid stock ] I'll downgrade to the stock as retest [04:59] i'm confused, you weren't using 2ubuntu1 when you said it was fixed, you were using newer? [05:00] if i use new_pll=0 with 2ubuntu1 , i dont get the problem of the screen freaking out... but if i use the new_pll=0 with -xup2 i get the screen jumping [05:03] I just got a new(ish) MB with Intel GMA X4500. My xorg log says xvmc is disabled. I've got xorg 2.9, do I need 2.10 to get xvmc or is there something else? It looks like the video is using the i915 driver. [05:03] i'm confused because there's no way you were using 2ubuntu1 when you tried new_pll because 2ubuntu2-5 were released around the same time as ati 6.13 that you were testing downgrades on? [05:04] (when you said it was working) just trying to narrow down where it might be [05:04] skimj: xvmc is 100% useless but if you want to use it with lucid you need to disable KMS [05:05] or upgrade to 2.10+ to use it with kms yeah [05:07] OK. I'll try 2.10. Why do you say it's useless? I'm trying to get mythtv to run without stuttering. I just upgraded the MB (with video) trying to improve things. XvMC seems to be the only thing wrong in any of the logs. [05:07] Sarvatt: hrm... the 2ubuntu1 was because i was downgrading earlier to avoid the memory leak.. I believe it is a problem again after update 2ubuntu3 -> 2ubuntu5 ... I'll test each version and let you know which is the problematic version [05:08] Sarvatt: btw , why isnt new_pll=0 the default for the rv515 cards? I'm still having to add it to the kernel line [05:08] 2ubuntu5 is the only real change but i dont see how those could be causing it [05:09] ask the kernel team :) [05:09] hehe ;) [05:11] did everything i could short of writing the patches myself, identified what the problem was and what gpu's were affected and brought it up on the kernel list with a list of all reports that need the quirking.. [05:16] maybe you can poke someone since you work for canonical? :) [05:19] if its really a problem with 2ubunu5 and not with 2ubuntu1 (even though you need new_pll=0 still) thats interesting [05:27] Sarvatt: hmm , i dont work for canonical, but will poke folks :) [05:59] https://bugs.freedesktop.org/show_bug.cgi?id=27510 [05:59] Freedesktop bug 27510 in Acceleration/EXA "Xorg crashes when right-clicking on NoScript Icon in Firefox" [Major,Resolved: fixed] [05:59] seems thats hitting a lot of ati users [06:01] oh bryce already forwarded another report about that [06:01] Sarvatt: would you be willing to help debug a bit? I added the xorg-edgers ppa and did an upgrade. Now the intel module fails to load. Here's a snippet from the xorg log http://paste.ubuntu.com/420830/ [06:02] you didnt upgrade intel for some reason? [06:03] you need to update everything in that ppa, it's not compatible with lucid packages [06:04] it was probably because i didn't upload an xorg metapackage yet and you did a safe upgrade so it didn't remove xserver-xorg-video-all/etc [06:05] try a sudo apt-get dist-upgrade? [06:05] so I did an apt-get upgrade which got a lot of things. X failed to load after that so I did an apt-get update ...intel [06:06] OK I'll try it with the dist-upgrade. [06:06] did it say any packages were held back? [06:07] yes, I've got 4 packages held back [06:07] which? [06:07] knm-runtime kopete network-manager-kde plasma-widget-networkmanagement [06:07] ah not related [06:08] dist-upgrade offering intel? [06:08] hmm,.. what was the command to downgrade a package , which has the deb in /var/cache/apt/archives? [06:09] nope. It's offering libmsn0.3, knm-runtime kopete network-manager-kde [06:09] whats the version on your installed xserver-xorg-video-intel package? [06:10] 2:2.11.0+git20100422.72fd7d19-0ubuntu0sarvatt2 [06:10] huh [06:11] did you ever compile intel from source in /usr/local or anything? [06:12] oh sorry it says /usr/lib/xorg/modules/drivers/intel_drv.so there [06:12] nope never from source. I did have everything from the ppa installed, then I uninstalled with the ppa-purge tool, and now reinstalled [06:13] i'm stumped, can you try again and paste the full log? [06:13] try from reboot? [06:14] yeah [06:14] absolutely sure you had 2.11 installed when you booted there? [06:15] hmm wait [06:15] why isnt it 2.11.0+git20100422.72fd7d19-0ubuntu0sarvatt3 ? [06:16] yes, I did. So here's the weird thing. It's working now. The only change was dist-upgrade with knm-runtime kopete network-manager-kde plasma-widget-networkmanagement [06:16] nevermind thats a local build with pageflipping disabled... :) [06:18] thats really odd skimj, probably can find out why from your /var/log/apt/term.log [06:21] if you want to look at the term.log: http://paste.ubuntu.com/420830/ [06:22] Sarvatt: how do i downgrade a package to one which is in the /var/.../cache? [06:22] dpkg -i /path/to/file? [06:22] * vish tries [06:22] skimj: that was the xorg log snippet link [06:24] skimj: i've got to run, about to pass out. its all working now though? [06:25] oops sorry. http://paste.ubuntu.com/420837/ [06:25] yes, it's all working now. Thanks. [06:26] bah didnt work... /me downloads package .. is easier ;) [06:28] thats really odd, why didnt intel get upgraded there.. you're also missing synaptics [06:29] skimj: something definitely went funky there, it didn't pull in synaptics either [06:29] should I do something different? [06:30] do you have xserver-xorg-input-synaptics installed? [06:31] I don't see it [06:32] i just forced it with apt-get install xserver-xorg-input-synaptics [06:32] install xserver-xorg-video-all and xserver-xorg-input-all? [06:33] i'll upgrade a bunch of systems tomorrow and see if i can figure out what happened, got me stumped [06:33] FWIW, that's also pulling in video-ati [06:34] OK, let me know if I can get you any more info, logs, etc. [06:35] something tells me the ppa-purge might have messed up, and it fell back to aptitude and offered a solution involving removing some packages [06:36] must pass out though, sorry for the trouble and thanks for pointing out the problems :) [07:07] darnit, i just had to start updating another machine to edgers and now i'm back :D bjsnider have you seen any nvidia blob releases that work with xserver 1.8 without IgnoreABI by any chance? [08:42] Sarvatt, i think you can include the ia32-lib from https://launchpad.net/~ricotz/+archive/staging/+packages now? [08:47] RAOF, hello, or perhaps you can ^ [12:25] Sarvatt, which one is xserver 1.8? [12:26] eh? [12:32] is that the one in lucid? [12:34] no [12:34] it's the one in f13 [12:40] well, nvidia has to be given a bit of time to update their driver for it i suppose [12:43] first release candidate was over 2 months ago. considering the ABI bump is trivial this time, that should be more than enough... [12:44] i think the hdmi audio issue is a much bigger problem [12:44] it's a longstanding issue that is keeping some people on the 185 [12:44] they've had plenty of time to fix that one === Sinnerman is now known as Cobalt [14:30] ooh, I see the gem leak bug is fix committed [14:30] did we revert the glx backport or just drop in the extra patch RAOF found? [14:32] the former afaik [16:34] bjsnider: the 195 drivers supported xserver 1.8 for months but you need to use IgnoreABI, was just wondering if you had come across any newer betas or anything so I could put them on xorg-edgers but I dont see any [17:22] Hi, I'm having a problem setting my button map for my mouse [17:22] I used to set it with a HAL policy, now it looks like i should be seting it with a udev rule [17:23] I've got the udev rule set up correctly, and "udevadm info ..." shows that x11_options.ButtonMapping is set and correct on my mouse device, however X is completely ignoring this button mapping [17:30] on the machines I tested, fbc would get enabled, but actuall take more power [17:30] because it never finished a compression pass [17:30] a few others reported the same thing [17:30] that may be why the feature isn't enabled on windows [17:30] * Sarvatt doesn't feel bad about submitting a patch completely disabling FBC on 915-945 now [18:32] Ok, so i fixed my problem, it appears that Ubuntu Lucid has already deprecated configuring X with udev? [18:32] or maybe jsut some aspects? [18:36] Alan: yeah, udev is used transparently now, you configure things through xorg.conf or xorg.conf.d snippets and it is described pretty well in man xorg.conf [18:36] i have to run but i can point you to wiki's describing the new world order that is much easier to manage in about 45 minutes [18:39] Sarvatt: but is it right that it should ignore properties set in udev rules? [18:41] huh, i just found an article about the new configuration stuff... [18:59] Sarvatt, btw do you have much feel for the status of multi-card support (3+ displays) upstream? I know that's work being done but is it likely to be Meerkat material? [19:18] Alan: sorry for the delay, ended up running late. http://fedoraproject.org/wiki/Input_device_configuration is a really good guide, our xorg.conf.d snippets are in /usr/share/X11/xorg.conf.d/ if you want to see whats there currently. if you want to change options you just add an InputClass section to your xorg.conf with the options you want in it [19:18] err sorry, lucid's are in /usr/lib/X11/xorg.conf.d/ [19:20] i dont know what device you are setting up but for synaptics for instance you can see the full list of options in man synaptics (or synclient -l) [19:25] bryceh: need to look into it more but from what I know, GPU switching for systems with multiple GPU's yes, multiple GPU's at the same time no way [19:26] bryceh: you can do 3+ displays on ati 5xxx though with 1 GPU [19:27] know if anyone's tested that with an ati 5xxx on lucid? [19:29] pretty sure lucid can't handle it at all, we dont even have KMS support for them [19:29] its weirdly limited too, you have to use certain combinations of connectors to do it [19:29] http://www.botchco.com/agd5f/?p=51 is an *awesome* summary of it [19:30] aha thanks [19:30] i'm sure fglrx can handle it, no experience with it at all though [19:31] i'm seeing reports of odd problems on 5xxx with stock lucid though, one connector works and the other doesn't, stuff like that [19:36] pretty sure arrandale/clarkdale can handle 3+ displays too, know i saw a commit from jbarnes mentioning it [19:37] in the increase the stride limit on igdng commit [19:42] mm [19:46] Sarvatt, bryceh: so the packages that are in the repos now will go out on the final iso? [19:50] Dr_Jakob: your vmmouse fix got uploaded on the 19th, the archive was just frozen for the RC and its in the archive now [19:50] Yeah, I verified that the fix was in the stuff I pulled from the repo. [19:51] Dr_Jakob: nah there will be more updates that are just major bug fixes no doubt [19:51] the vmmouse update will be on the daily iso from today, it just isn't in the -rc release cd [19:52] Right, and the final iso will be based on one of the dailies? [19:54] whatever's in the archive at the time it's made, probably on tuesday? not sure of the exact date [19:55] bryceh: looks like i was wrong about that commit saying 3 displays was possible - " It can go up to 32k. Upping this lets me use my 2560x1600 and 1920x1200 monitors in an extended desktop configuration." [19:57] bryceh: i'm pretty sure you can do displays on multiple GPU's with seperate screens on nvidia with the blob? [19:57] yes [19:57] i've seen a bug about it at least and it needed lots of xorg.conf fiddling [19:58] right, with nvidia it's done via xorg.conf-ary [19:58] think i'll dig through the arrandale docs and see what it has to say :) [19:58] the nvidia config program takes care of that a bit [19:59] Ok, thanks. I got a bunch of managers breathing down my neck about this bug. [19:59] they pulled them off the intel website but i saved copies [19:59] I guess there's one thing we can be certain of [19:59] when this feature becomes available, in order to integrate it we'll probably have to break everything else [19:59] that's just how intel rolls ;-) [20:00] Dr_Jakob: have you seen the bug where a messed up xkb layout is getting applied from the vmware autoinstaller (?) and the keyboard isn't usable for login? [20:01] trying to dig it up, it's gotten alot of dupes [20:01] Sarvatt: can't say I have. Do you have a bug number. [20:01] digging for it [20:01] Sarvatt: the second session , screen jumping bug: .. it happens with 2ubuntu1 as well :/ it seems the new_pll=0 only works and keeps thing quite for some time ~6hrs of uptime, then it happens , everything is still jumping :( [20:03] so if you are testing the bug with a fresh boot , and try the guest session it wont have a problem , but after ~6hrs the problem arises even with new_pll=0 [20:03] darnit where is it hiding [20:06] Dr_Jakob: https://bugs.edge.launchpad.net/ubuntu/+source/console-setup/+bug/548891 [20:06] Ubuntu bug 548891 in console-setup "keyboard input broken due to invalid "SKIP" keyboard model" [High,Confirmed] [20:11] Sarvatt: hmm so we are not hitting this issue as far as I can tell in the beta of the Fusion/Workstation/Player. [20:13] ah nice, is it publically available? i'll mention that when the next person asks about the problem [20:13] http://communities.vmware.com/community/beta/ws [20:14] Tho looking at the forums, I can see that somebody has reported it.. hmm. [20:15] Oh, its after you update console-setup. [20:16] btw I plan on enabling svga in the xorg-edgers PPA here soon, it's already enabled in the 10.10 kernel config and I have that in xorg-edgers, just need to package up the libdrm/mesa sides [20:17] Its totaly unsupported [20:17] Just FYI [20:18] that's what xorg-edgers is all about :) [20:18] just make sure you have video-vmware version 11.x.y [20:20] it'll have git master updated whenever i see changes [20:20] right ok good. [20:21] changes there don't go to the xorg-commit mailing list do they? [20:21] Hmmm [20:21] they do. [20:22] http://lists.freedesktop.org/archives/xorg-commit/2010-February/025157.html [20:22] ah ok just haven't seen a commit in awhile [20:23] I have been stuck fixing weird input driver bugs ;-) [20:23] Also video-vmware only holds the old video driver and driver selector. [20:23] the 3D enabled driver comes from mesa [20:26] speaking of which i need to update that in edgers now anyway because of the new abi in there [20:28] Sarvatt: I like to in advance apologies for the circular dependancy between mesa and xorg for when added the X org driver to mesa. [20:29] yeah I've already been down that road when I was first enabling nouveau gallium and building the xorg state tracker by mistake :) [20:35] Sarvatt: do xorg-edgers ship r300g? [20:36] it's a work in progress, we've got it in another PPA at the moment trying to work out how to package it [20:36] ok [20:36] can you give me any tips on packaging gallium dri drivers at the same time as classic dri drivers? for instance i'm unsure what happens when both radeon_dri.so and radeong_dri.so exist at the same time [20:37] like for nouveau or svga it's no issue to package them together with classic dri [20:39] too bad tormod isn't here, he's the one most interested in shipping r300g and has been trying to work it out here - https://edge.launchpad.net/~xorg-edgers/+archive/radeon/+packages [20:40] the X driver tells libGL which driver to load, for radeon this is currently "radeon", libGL then adds /path/*_dri.so and dlopens it. [20:40] so just adding radeong_dri.so wont break anything. [20:42] ah ok so actually having the gallium driver be radeon_dri.so is going to be required [20:43] yeah, or adding a xorg.conf option to tell the driver to say radeong instead of radeon. [20:44] maybe even add it to xorg.conf.d if that works for video drivers. [20:53] hmm you can specify it to load radeong_dri.so in xorg.conf or just the path to radeon_dri.so? I didn't know that was possible but it would greatly simplify it if you could specify the name and not have to rename things [20:55] You can't now, but it wouldn't be hard to make the driver look obey a "DRIDriverName" option. [20:57] ah ok I get it now, it'd use radeong_dri if I went and used radeong_drv also [20:59] Hmm dunno, it might be hardcoded to "radeon" inside the driver or the name of the driver. [21:06] http://www.phoronix.com/vr.php?view=ODE3MA what's this I read about an X mem leak?? [21:07] BUGabundo: fixed already and didn't affect you on nvidia, it sure was good for ad revenue I'm sure though [21:08] I bet [21:09] oh hoh, looks like I might be able to abuse --with-dri-searchpath to have it look in the gallium directory first, then split the gallium drivers out to another package mangling the names that isn't installed by default === BUGabundo is now known as BUGabundo_diner [21:11] there's a newer xorg-server in main that in the x testing ppa [21:12] than* [21:16] virtuald: do you mean x-updates? because the one from there was uploaded to main, its the same thing [21:23] this --with-dri-searchpath hack just might work :) === BUGabundo_diner is now known as BUGabundo [21:42] so i'm thinking build everything at once in the config-dri rules target that has --with-dri-searchpath=/usr/lib/dri-gallium:/usr/lib/dri, only install mesa classic dri (and nouveau/svga) in /usr/lib/dri, copy the gallium dri drivers to /usr/lib/dri-gallium/ with the expected names, and have a seperate libgl1-mesa-dri-gallium (or whatever) package that's optionally installed with the /usr/lib/dri-gallium stuff [21:43] guess it doesn't matter where nouveau goes, it just works fine in /usr/lib/dri [21:44] sounds good [21:44] Could I request that the vmwgfx_dri.so ends up in the correct place? [21:45] That is /usr/lib/dri [21:45] ah yeah need to work out how to handle that right since it'll have the _drv.so too right? [21:46] just stuff that into /usr/lib/xorg/modules/drivers/ [21:46] vmware_drv.so will then check if your enviroment is sane and load it. [21:48] if not it will fallback to vmwlegacy_drv.so (renamed vmware_drv.so, the new vmware_drv.so is just a driver picker liker ati_drv.so). [21:49] yeah whenever I get that together it'd probably be better off in a seperate package installing both anyway [21:51] well the driver picker is quite paraniod about the enviroment (must be able to load vmwgfx_drv.so, must be able to open the drm device) otherwise it falls back to the old driver. [22:03] yeah i'll make sure it's all in /usr/lib/dri and /usr/lib/xorg/modules/drivers when i do it, nice that you have vmware picking the right one instead of having to patch xf86AutoConfig.c to add vmwgfx :) [22:15] jcristau: is there any way to give other people access to personal branches on git.debian.org? like if tormod created a mesa packaging branch on there and I wanted to commit to it [22:23] isn't bug 568605 a dup of bug 564181? [22:23] Launchpad bug 568605 in linux "ATI radeon KMS driver - gpu lockup" [High,New] https://launchpad.net/bugs/568605 [22:23] Launchpad bug 564181 in xserver-xorg-video-ati "[RV730] GPU soft reset infinite loop scrolling in firefox with compiz" [High,Triaged] https://launchpad.net/bugs/564181 [22:30] bdmurray, maybe [22:31] bdmurray, with X freezes I've learned from experience not to dupe them together too aggressively [22:32] it's better to forward each one upstream and let the experts tell us that they're dupes rather than guess wrongly and lose a bug in the cracks [22:33] also even if they are the same bug sometimes having several users investigating them in parallel can make it easier to find the solution since there's more eyes and hands [22:33] once a solution is found, often then it is pretty straightforward to identify dupes via ppa testing [22:34] also I hate the "X freezes" me-too bugs of death ;-) [23:05] Sarvatt: use acls?