[06:32] how the hell does evdev work? [06:55] im getting pushback from scottk on closing a bug about requiring root prives to use uinput; im wondering if evdev solves anything [07:20] pwnguin: which bug [07:22] bryce: so is the pot file in the binary now or just the source package? [07:22] bryce: xkeyboard-config.. [07:29] bug #312957 [07:29] Launchpad bug 312957 in cwiid "CWiid Version 0.6.00 unable to open uinput " [Undecided,Invalid] https://launchpad.net/bugs/312957 [07:32] does it work in jaunty? [07:32] in intrepid evdev grabbed the input devices [07:32] but doesn't do that anymore in jaunty [07:33] it works as well in jaunty as it ever has [07:33] the problem is that the program uses uinput [07:33] which is a security black hole [07:33] ok [07:33] well, I don't know about that :) [07:33] the conclusion ive come to is to just use sudo and be done with it [07:34] user beware [07:34] well, the package could use policykit to use the plugdev group for the device [07:34] erm [07:35] no policykit needed there [07:35] PK adds the group for the active user [07:35] when i spoke with scott remnant a year or longer ago [07:35] or something like that [07:36] This implies to me that any user could insert input events to your X server, and thus click on any part of the screen, etc. [07:36] These permissions are definitely inappropriate. [07:36] heh, ok then [07:36] im not about to overrule the udev maintainer ;) [07:36] so the driver is faulty [07:36] there's no way from here to there [07:36] im wondering if evdev works substantially differently [07:38] doesn't use uinput [07:38] anyway, afk-> [07:38] the uinput.h does say it's substantially copied from evdev [07:38] so they cant be all that different in practice ;) [07:39] g'nite or whatnot [07:43] night [09:34] tjaalton: just the source package [09:35] ok [09:42] I trust that's sufficient for rosetta [09:42] probably is [09:42] fwiw, it wasn't as simple as just running make .pot; it seems this part of the makefile has bit rotted some [09:42] * bryce -> bed [13:12] tjaalton: hi. could you have a look at bug #352335, please? [13:12] Launchpad bug 352335 in xserver-xorg-input-acecad "FTBFS due to xserver-1.6" [High,Confirmed] https://launchpad.net/bugs/352335 [13:12] I added a patch and would like you to check if it's ok. [13:18] Ampelbein: the patch itself does, but it's a lot of changes for just a simple patch [13:18] I'll update them in debian so we can sync them [13:18] there are six input drivers that FTBFS, minus aiptek [13:18] tjaalton: i had a little chat with bryce yesterday about using or not using quilt [13:19] i've done the change for aiptek using quilt and he said it was ok. [13:19] sure [13:19] Ampelbein: the fix should be as simple as updating to the new upstream [13:19] all those drivers got a new version because of xserver 1.6 [13:21] so do we wait until debian has the changes and sync? [13:22] it's the least effort (for you anyway ;) [13:22] ok, just thought it would be helping. [13:23] np, it was just delayed on my part [13:27] heh, I already had the git repos cloned === seb128_ is now known as seb128 [17:45] ah - maybe you can help me here: Is there a kernel including the openchrome-driver so VIA-chips work out of the box? [17:45] hello, i have a problem getting 3D acceleration with my Intel 945 gfx in Jaunty. i've asked in #ubuntu+1 and was directed here [17:48] PolitikerNEU: there is xserver-xorg-video-openchrome in jaunty, isn't it enough? [17:48] laanan: /var/log/Xorg.0.log plese [17:48] pleAse [17:48] ok, just a sec [17:49] tjaalton, (the Driver "intel" was added by me, no change from the default of nothing): http://pastebin.com/m6f83c26b [17:50] laanan: that's xorg.conf [17:50] tjaalton, oh sorry [17:51] tjaalton, http://pastebin.com/m29001fee [17:52] laanan: dri is used, so what doesn't work? [17:53] if it helps, i seemed to have the same problem with a fresh install of 8.10, and changing `Driver "vesa"` to `Driver "intel"` in xorg.conf fixed it. the problem is all 3D accel. apps seem software rendered [17:54] that is, it's more than just sluggish, frame rates are around 1-2hz for most gl screensavers and games [17:54] what does 'glxinfo| grep direct' say? [17:55] it says `direct rendering: Yes`, but it said that in 8.10 too and obviously it wasn't (again, adding Driver "intel" fixed it in 8.10) [17:55] adding that shouldn't do anything [17:56] ok [17:56] just pastebin 'LIBGL_DEBUG=verbose glxinfo' [17:56] tjaalton: At least in 8.10 (was openchrome there?) it was not enough - has the situation changed in 9.04? [17:56] PolitikerNEU: try the beta livecd [17:57] tjaalton: I will - the problem it is not my laptop [17:57] tjaalton, http://pastebin.com/m1b3276ca [17:58] so it's accelerated alright [17:58] yep [18:01] the advice i got for 8.10 did fix the problem, and from what i remember glxinfo was pretty much the same before and after the xorg.conf change [18:01] it's not clear to me what the problem is. you said you wanted 3d accel, you've got that already. [18:02] laanan: gnome or kde? [18:02] gnome [18:03] try with a new user or purge your compiz configuration [18:03] a livecd would do as well [18:03] well, for example, glxgears gives me around 300fps now, it was the same with 8.10 before the change. after the change glxgears in 8.10 runs at over 5 times the speed [18:04] glxgears is not a benchmark [18:04] pick another example [18:05] it's just an indicator, the skyrocket screensaver gets around 1-2hz, GtkRadiant is almost unusable as well [18:07] tjaalton, could you explain purging my clmpiz configuration ? [18:08] laanan: probably in too many places I don't know.. just create a new user and try there [18:08] or try without the effects (ie. with metacity) [18:09] ok, well i have None checked in Appearance/Visual Effects.. i will try to create a new user [18:10] you need to log out and log back in as that user, quick switch to it basically means you got no DRI [18:10] in that session [18:10] ok [18:15] hmm, no change. this problem is not a large issue for me, as 8.10 works fine for the most part. i just wanted to try Jaunty since it seems there was a bug fixed in Mesa affecting one of my apps [18:17] thank you for the help though, i appreciate it [18:18] you could try the mesa packages at https://edge.launchpad.net/~tjaalton/+archive/ppa [18:19] i can use those with Intrepid ? [18:19] no [18:21] oh, well from what i gather the bug was unrelated to this current issue [18:22] (and i am a Linux newbie, so i'm a little over my head here anyway :)) [18:24] try a livecd, it should work there [18:24] if not, then it's a bug [18:25] that's what i tried first (the daily build), same problem [18:25] ok [18:28] it was explained to me that my previous 8.10 install was using the generic vesa driver, and was slow even tho direct rendering was enabled. i figured it was the same situation here but apparently not [18:28] vesa doesn't support dri [18:29] yeah, that was the reason was everything crawled there [18:31] for me, apparently i need `Driver "intel"` in 8.10 for everything to work as it should.. are you saying that (if everything is goes well) i don't need that in 9.04 ? [18:32] you shoudn't need that in 8.10 either [18:32] just remove the line [18:33] ok, i will try that the next time i am in 8.10. (but the problem still remains with or without that line in 9.04) [18:34] anyway, if i have time and can get it sorted out, i will come back and let you know. thanks again [18:34] np [18:38] hmmmm [18:38] tjaalton: interesting data point, I just booted a live CD... compiz stuff seems fine :/ [18:39] so something about my setup which has no appreciable xorg.conf, or .drirc or interesting compiz settings, is very much worse [18:39] Ng: there you go, upgrades ftl :) [18:39] this could also explain why the 965 belonging to the guy next to me is nowhere near as badly affected as mine [18:47] tjaalton: well, I'm not sure it's 100% conclusive, there will be a lot more happening on my install than a live CD [18:47] but I am now very tempted to do a full re-install ;) [18:59] heh [19:03] aha! a more comparable test from the live CD still shows the same juddering [19:03] * Ng feels vindicated and re-assured ;) [20:13] bryce: ping [20:13] how are we with support for intel 8x5 for jaunty? [20:27] BUGabundo: piss poor at the moment [20:27] BUGabundo: do you have feedback on it? I asked on the list for input from any 8xx owners but no one responded [20:34] thanks bryce [20:34] i changed xorg to use DRI false [20:34] and i can use this test PC [20:34] i have a bunch of then to install until saturday [20:34] BUGabundo: what chipset? [20:34] bryce: 865 [20:35] i comment on the bug [20:35] do u need any further testing? [20:35] i guess i could give u reverse ssh to it or something [20:35] I'm going to limit 304871 to just 845 graphics [20:36] which I think is fixed now [20:36] 865 will need a separate bug report [20:36] ok [20:36] 317457 maybe [20:36] is it opened yet? [20:36] BUGabundo: have you filed one for this issue? If so I'd prefer making that the master since you're more active [20:37] i did not [20:37] otherwise I guess we can use 317457 [20:37] ok [20:37] thats why i came here 1st [20:37] ok, i'm subed to it now [20:37] 328528 is UXA lockup on 865 [20:38] I've not decided if that's a dupe, I think it may be a separate issue. You have that one too? [20:39] let ne see [20:40] bryce: i didnt mess much on this xorg [20:40] just added dri [21:18] bryce: is there any thing i can do to make this card run compiz? [21:20] BUGabundo: can you try the Legacy3D "false" option mentioned in the bug report? [21:20] BUGabundo: did you already try option "AccelMethod" "UXA"? [21:21] no i did not [21:21] let me see if i can do it [21:22] ok now where do i need to add that stuff? [21:22] device section of xorg.conf [21:23] but keep dri false? [21:23] not if you want compiz [21:23] oh [21:25] you can try them both together if you want but disabling dri sort of cancels out uxa [21:26] so i add legacy3d and uxa [21:26] is this ok? http://paste.ubuntu.com/141679/ [21:27] BUGabundo: can you also try each of those two options separately? [21:27] I'd like to know if Legacy3D will solve your issue without also having to go to UXA [21:27] sure [21:27] i'll just make dinner wait a bit more [21:28] no prob [21:28] killing X now [21:30] with both i can login [21:30] commenting 3d [21:30] spoke too soon [21:30] system froze [21:31] bryce: having both seems like a bad choice on 865 [21:31] :-/ [21:31] yeah not many good reports of UXA on 965 [21:32] er 965 [21:32] dah [21:32] 888865 [21:32] BUGabundo: try with only Legacy3d [21:32] rebooted [21:32] I don't expect that will work, but if it does we're golden since I have a patch for that already [21:32] and repeated the same thing [21:33] ok so, the only thing that has worked has been disabling DRI? [21:33] (which disables compiz) [21:33] bryce: now testing with only 3d [21:34] i dont use compiz, id be happy to test BUGabundo [21:34] after that i can try wit only uxa [21:34] yow|x2: what chipset do you have? [21:35] intel ich7 [21:35] bah [21:35] critical bug... Delete will not send to trash [21:36] 945gm / 945gms for video [21:36] yow|x2: we're only looking at 8xx for this, but thanks. [21:36] bryce: just 3d i have a black background [21:36] see top and bottom bars [21:36] mouse moves [21:36] ack. gotta figure out why im getting freezing so much [21:37] bryce: want me to test just UXA now? [21:37] cant even jump to TTY [21:37] BUGabundo: if you'd like; we already decided we're not going to do UXA, and I'm pretty sure it won't work for you [21:37] alt+sysqr+K just showed blue screen [21:38] then black [21:39] yeah that's because the kernel kills X but doesn't know how to reset the graphics card to a sane state [21:39] probably have to reboot [21:42] bryce: FAIL [21:43] UXA wont work on this system [21:43] just DRI [21:43] any more combos u need me to test? [21:44] nope, that should do it [21:44] ok, I'll go with the disable DRI patch for 865 then [21:44] shame to lose compiz and 3D on that chipset, but better that it not freeze up [21:44] so no compoiz? [21:45] ohh my students wont be able to play with it! [21:47] BUGabundo: there was one reporter that said he resolved it on 865 by booting a 2.6.27 kernel [21:47] so you could try that, if you just want to be able to show it off [21:47] er wait that reporter was on 855 [21:48] still, booting an earlier kernel might be worth trying if you have one handy [21:53] bryce: no change. this is a daily jaunty image [21:54] hrm, too bad [22:38] tjaalton, superm1: either of you know if anyone is going to up us to the latest -nvidia? Or should we be skipping it? [22:42] bryce, from the looks of it, it's bug fixes, but i'm not positive [22:42] I've seen several positive reports around it, and judging from the rate of new bugs against nvidia-180 so far, we could use bug fixes there [22:43] bryce, yeah especially KDE plasma fixes [23:24] good god, xorg is up to 35 bugs after just a couple weeks