[10:23] is this the channel for the maintainers of the x-swat ppa? [10:26] yes it is. [10:26] not that i'm one. [10:26] Sarvatt maintains it afaik. [10:29] Oh ok [10:29] I'm just wondering if there is an issue currently with the latest nvidia driver in lucid, from the stable ppa [10:30] I was experiencing poor-ish performance with the shipped version of the nvidia driver, so I grabbed this ppa, and updated the nvidia-current package, yet it doesn't seem to worked :S [10:30] Perhaps I'll try the beta release [10:31] * hyperair has given up on nvidia and ati. [10:31] intel ftw! [10:31] lol [10:31] besides, my nvidia-based desktop doesn't work with drivers newer than 96.xx [10:31] stupid nvidia, deprecating my card just like that [10:31] it's only a few years old.. [10:32] I just recently got back into the *nix game from windows [10:32] It's been awhile [10:32] hrmm why did you even switch back to Windows in the first place? [10:32] Haven't used ubuntu since 6.XX === freeflyi1g is now known as freeflying [10:32] haha. [10:32] Work reasons [10:32] that was the first Ubuntu I used =p [10:32] 6.06 [10:32] I found that I spent way too much time playing with the OS [10:33] i screwed my xorg.conf, couldn't figure out hwo to fix it, and purged the system thrice before i got it working. [10:33] Can't exactly do that as easily in windows [10:33] oh [10:33] heh [10:33] that sounds familiar =p [10:33] you need to let the novelty wear off [10:33] I think it was dapper [10:33] I then used arch for a few years [10:33] but then again, from time to time, i still end up playing too much with the OS.. like how i spent a week trying to get my kernel compiling.... [10:34] oh Archlinux eh.. [10:34] i used that for a grand total of 4 months. [10:34] at the end of 4 months, the only difference between my Arch and Ubuntu installations were the package manager. [10:34] I liked it, it was fun knowing I built it from the ground up...ish [10:34] so i just gave up using Arch and went back to Ubuntu. [10:34] go LFS ;-) [10:34] Linux From Scratch [10:34] Heh, I just installed 10.04, and I'm extremely impressed... [10:35] It's nicer than my mac! [10:35] \o/ [10:35] yes, 10.04 is a very impressive release. [10:35] unfortunately my nvidia drivers are being poo [10:35] unfortunately. [10:35] what card is that anyway? [10:35] 130m in an hp notebook [10:35] geforce* [10:36] works fine in mac and windows for me, I was just getting a few oddities with Lucid's standard driver [10:36] which is odd, considering it's LTS [10:37] not at all. [10:37] there are limits to how much QA we can do [10:37] haha [10:37] and nvidia is a great shining example of that. [10:37] it's a closed source blob, nobody knows wtf's going on in there [10:37] I know, I know, I'm just giving people a hard time :) [10:38] hmm, giving who a hard time? ;-) [10:38] i'm enjoying myself putting down nvidia here. [10:38] It's probably my fault anyway [10:38] I'll try the beta drivers and let you know [10:38] good luck [11:18] *sigh [11:18] I must have found a bug [11:19] or whatever it is that the hardware drivers app does, I couldn't figure out. [11:19] I got the beta drivers working, and performance is much smoother [11:21] well that's nice =) [11:23] yeah, its just irritating how I couldn't get it to work by just installing it. I had to install the old ones via hardware drivers, then add the ppa for the beta ones, then restart, then upgrade, then restart :S [11:25] yay nvidia === albert231 is now known as albert23 [15:26] * hyperair eyes the new x-x-v-i upload suspiciously [17:06] I'm confused, the ones in the ppa *are* the beta drivers, he must have been installing from nvidia.com? [17:28] crashing hyperair? [17:28] is it any worse than before? [17:32] * bryceh waves [17:37] Sarvatt: page flipping support is enabled =O [17:37] Sarvatt: which explains why i keep getting hangs within 10 minutes of each other >_> [17:39] Sarvatt: oh hey i managed to switch TTYs [17:49] hyperair: new one uploaded without it [17:49] sorry :( [17:49] i will warn you before i do it again, i passed out last night after uploading it [17:50] heyo bryceh! [17:50] Sarvatt: no problem. [17:50] i think i'll write a script that greps the current Xorg log and checks if page-flipping is enabled [17:51] and warn me [17:51] =p [17:56] lol [18:02] #!/bin/bash [18:02] if [ -z "$(grep forcibly /var/log/Xorg.0.log)" ]; then [18:02] notify-send -t 30000 -u critical -i gtk-dialog-info "Page flipping is enabled." "YOUR X IS BUSTED!" [18:02] fi [18:03] Sarvatt: thanks. =p [18:03] i call it lolintel.sh :) [18:03] hehehe [18:06] maybe a zenity warning in the notification area would be better :) [18:06] so its persistant [18:06] Sarvatt: if ! grep -q forcibly /var/log/Xorg.0.log; then [18:08] Sarvatt: grep forcibly /var/log/Xorg.$(echo ${DISPLAY} | grep -o '[0-9]' | head -n1).log)" &> /dev/null && notify-send -t 30000 -u critical -i gtk-dialog-info "Page flipping is enabled. "YOUR X IS BUSTED!" [18:14] hah, I wasn't crazy! [18:14] 3.6.35-rc1 is nasty [18:14] my machine no longer goes into C3 states [18:15] battery life sucked yesterday, guess thats why [18:15] lol [18:15] rc1 is nasty indeed. [18:15] i spent a whole week attempting to compile it [18:15] C0 (cpu running) (80.6%) 1.60 Ghz 1.7% [18:15] polling 1.5ms ( 0.0%) 1333 Mhz 0.0% [18:15] C1 halt 2.6ms ( 3.3%) 1066 Mhz 0.7% [18:15] C2 0.6ms (16.0%) 800 Mhz 97.7% [18:15] C3 0.0ms ( 0.0%) [18:16] aha. [18:16] that makes sense. [18:16] nasty! 13 watts idle power usage vs 7 [18:16] i was wondering why my temperature was hovering 10 degrees celcius higher than usual [18:16] its fixed in -rc2 [18:16] cool [18:16] saw the commit then looked at powertop lol [18:16] i'll compile rc2 as soon as i get my btrfs fixes [18:17] Sarvatt: do you know of this patch: http://lists.freedesktop.org/archives/intel-gfx/2010-April/006463.html ? [18:17] well whenever rc2 comes out i mean, fixed in git though [18:17] Sarvatt: rc2's out. [18:17] commit e44a21b7268a022c7749f521c06214145bd161e4 [18:17] Author: Linus Torvalds [18:17] Date: Sat Jun 5 20:43:24 2010 -0700 [18:17] Linux 2.6.35-rc2 [18:18] oh dang, i opened the log last night and didnt refresh it :) [18:18] that patch doesnt affect you though? [18:18] Sarvatt: http://www.pubbs.net/201005/fedora/1529-page-flipping-on-intel-211-driver.html <-- this says that the patch i linked above fixes the hangs with page flipping. [18:19] for gen3 yeah you have gen4 though [18:20] oh isit? [18:20] =\ [18:20] how do you tell? [18:21] 965+ = gen4, 915-945 is gen3 [18:21] weird. [18:22] it said GMA3100 [18:22] 3100 is the one i have.. [18:24] whats the pci id? [18:25] those 3xxx ones are oddballs [18:27] you use 965_dri so though i'm sure yours is gen4 [18:27] drivers/gpu/drm/i915/i915_drv.h has the list of what each pci id is [18:31] Sarvatt: 2a02 [18:32] 00:02.0 VGA compatible controller [0300]: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller [8086:2a02] (rev 03) [18:33] yep thats gen4 [18:33] heh oh well. [18:33] #define IS_GEN4(dev) ((dev)->pci_device == 0x2972 || \ [18:33] ..snip.. [18:33] (dev)->pci_device == 0x2A02 || \ [18:34] ah [18:34] thats an X3100, different beast than a 3100 [18:34] 3100 is a desktop G31/G33 [18:35] @_@ [18:35] damn them, can't they just stick to one numbering scheme? [18:35] http://en.wikipedia.org/wiki/Intel_GMA [18:35] hyperair: that's called marketing [18:36] yeah make people think its a better gpu than it really is :) [18:43] Amaranth, RAOF: compiz works again on intel when resolution == max texture size in mesa 7.9 [18:44] so we can drop that nasty hack that made it so everyone cant use it in MM eventually [18:46] jcristau: bah. [18:51] Depends: [18:51] xserver-xorg-core (>= 2:1.8), [18:51] xserver-xorg-video-all | xserver-xorg-video-7, [18:51] xserver-xorg-input-all | xserver-xorg-input-9, [18:51] that look right for xserver-xorg? [19:04] hyperair: btw I figured out how to get vblank_mode in drirc to be honored - http://pastebin.com/C37kfXkr [19:05] ah. [19:05] why that took me like 6 hours stepping through gdb i'll never know, at least it was a learning experience :) [19:05] lol [19:05] why did you need it disabled again? [19:06] because it's crap on 945, makes the mouse really laggy in mutter when its on even [19:06] vblank_mode=1 isn't bad either, just 2 and 3 suck [19:08] couldn't figure out how to fix mesa so it worked when you set it for i915 like its advertised though [19:14] i see. [19:20] i guess if we do go r300g by default for ati in MM we can just add a check for if KMS is in use to the DDX to use the gallium dri driver name otherwise use the classic one, the names are different so they wont interfere [19:23] heck, i'm going to just do that now in edgers so I can ship radeong_dri in /usr/lib/dri [19:25] oh wait, it's even easier than I thought, radeon dri2 is only used in KMS so just modifying the dri2 path is guaranteed to be gallium compatible [19:35] hmm, need to figure out how to check for gallium support in configure.ac [19:58] hmm after doing all that it seems like it'd be better as an xorg.conf option that only enables by default if you build with --enable-gallium so people could still use classic [20:01] http://sarvatt.com/downloads/patches/0001-Add-a-configure-option-at-build-time-to-use-the-gall.patch [20:59] at this point, i'm not seeing much reason to even have a /usr/lib/dri-gallium [21:02] working on a patch to the ati ddx that adds a gallium xorg.conf option that defaults to enabled that picks radeong instead of r300 so its possible to use both instead of that hacky rename. intel is the only one where the gallium versions are the same as the classic ones and we arent gonna be building that anyway [21:03] (and the xorg.conf option is only available with KMS, wont affect UMS) [21:08] ahh forgot about swrastg [22:21] this dri2 stuff is confusing.. can we specifiy more than one driver name in the driverNames array in DRI2InfoRec or is it hardcoded to use one name for each type? [22:26] i guess that doesn't make sense because it wouldnt need to use both at the same time, dont mind me just figuring crap out :) [22:27] it's one of each type. dri2 at offset DRI2DriverDRI, vdpau at offset DRI2DriverVDPAU, iirc [23:41] Sarvatt: Yeah, I ended up deciding against dri-gallium too. [23:41] Sarvatt: I've talked with airlied - it seems there's a good chance that radeong_dri.so will become r300_dri.so in 7.9 [23:42] doh [23:42] well crap [23:42] Doesn't make that patch less useful, though. [23:43] Just means you'd need to update the mesa packaging to copy r300_dri from glx/gallium to radeong_dri [23:44] was almost done making one that adds gallium as a xorg.conf option defaulted on to info->ChipFamily >= CHIP_FAMILY_R300 && info->ChipFamily < CHIP_FAMILY_R600 [23:44] but i cant for the life of me figure out a way to have a fallback [23:44] like if you have it enabled in xorg.conf with that option and it doesnt exist have it try the other [23:45] stat radeong in the setup codepath? [23:47] not sure the ddx driver knows the dri driver path [23:48] Well, if it's just a edgers hack patch, that's not too much of an issue :) [23:48] hehe, yeah [23:49] I don't _think_ upstream's likely to be interested in it - airlied seemed very much of the opinion that making it easy for the user to choose between the dri modules wasn't useful.