=== ripps_ is now known as ripps === Ng_ is now known as Ng [09:55] tjaalton: did nvidia -96 fail to build again (for the same reason)? I've reuploaded the source to my website again. [09:58] tjaalton: please reupload it when you have the time [10:21] tseliot: i guess it doesn't try to rebuild it if the version didn't change [10:22] but I've requested a rebuild [10:22] tjaalton: thanks [10:23] nope, didn't help [10:23] tseliot: you need to bump the version [10:23] tjaalton: ok, let me do it and reupload the source [10:26] tjaalton: ok, uploaded. http://albertomilone.com/ubuntu/karmic/to_upload.txt [10:32] tseliot: and uploaded [10:33] tjaalton: thanks again [16:48] bryce: what is missing to get http://launchpadlibrarian.net/31292587/xserver_1.6.x_add_Geode_GX1_GX2_LX_to_xf86AutoConfig.patch merged into Karmic? [17:12] Someone at ubuntu+1 told me this would be a good place to discuss some issues I have with Xorg Intel drivers for Karmic. [18:25] good morning tseliot [18:26] bryce: good morning to you (or good evening in Italy) [18:34] tseliot, I'll work on the libdrm update [18:35] bryce: ok, I'll work on the other merges then (I haven't started yet) [18:39] yeah looks like a lot of stuff got done upstream recently, probably prepping for their next release [18:43] good [20:17] bryce, i am seeing some strange behaviour with pretty recent karmic bits on ati ... specificially it is showing me the shutdown screen very briefly on X start [20:18] hmm [20:18] apw, I've upped us to pretty recent git versions on -ati [20:18] i can see he window i ran reboot in, for a couple of seconds [20:18] and its not the best either in use [20:18] apw, is this with stock karmic kernel or one with backported ati drm bits? [20:19] this is with a stock -9 kernel [20:19] indeed all stock bits [20:19] ok, there's been some problems I've heard with -ati when using the stock kernel [20:19] bbiab (breakfast) [20:20] bryce, it almost feels like every other refresh is not on the real screen, i get junk instead [20:23] it is definatly not clearing out its video memory on init, which is potentially a security issue [20:25] has bryce moved to japan? [20:26] even if he did, 4:26 is a little early for breakfast isn't it? [20:27] either that or he got up reaaaaaaaaaaaaaaaaally late in USA [20:35] or he is soooo dedicated he only just got round to breaky [20:38] it's 12:38 in portland.. not that late for a hacker ;) [20:38] PM [20:39] ah, yes. west coast [20:45] apw, hmm weird, it works okay on my ati system [20:46] its utterly corrupting my display most of the time [20:46] yeah apw is right, I neglected too long to take a break. blood sugar was getting pretty low ;-) [20:47] morning got kinda screwed up due to meetings and wife having dual flat tires [20:47] dual! what did she do, drive over a milk float? [20:48] she had a "bolt" in one, and nfi on the other [20:48] (that metaphore probabally doesn't translate, they used to use glass bottles) [20:48] (a nail or screw I could understand, but a bolt?) [20:49] yeah thats sooo very pointy, not [20:49] I suspect the one was due to a stuck valve from after I checked all her tires. They've replaced the valve as they didn't see anything else wrong [20:50] 2 flat tires + wife 1-day from baby due date == trouble [20:50] 10x trouble [20:59] does anyone here happen to know anything about Matrox Parhelia on Ubuntu? [20:59] is there some nice PPA to add, or at least a guide to get better than VESA-graphics? [21:01] yikes [21:02] *ahh* found https://help.ubuntu.com/community/BinaryDriverHowto/MatroxParhelia :) [21:02] of course :) [21:02] help.ubuntu.com should be my new google.. :) [21:03] h.u.c is waaay out of date on X stuff [21:04] but then, matrox is so ancient even out of date info is probably still of use ;-) [21:04] well.. perhaps.. except all changes in core X breaks all that good ol' stuff.. ;) [21:04] and, it's not exactly.. parhelia.. it's P650, but same driver I assume? [21:06] btw, out of curiosity, what are the odds of seeing old drivers, like matrox and friends up-ported to KMS-glory etc? [21:06] (from a guy that know ~0 about driver-creation) [21:07] ~0 [21:07] allright.. :) I thought so.. [21:07] requires someone to care enough about the driver to work on it [21:07] by definition old drivers are usually old because no such person exists [21:07] as with all software, of course.. :) [21:08] too bad for KMS though.. it's a bit flaky to rely on it unless also old and exotic hardware can support it.. :S [21:08] anyway, not something we'd put resources into for ubuntu, since we're sufficiently tied up dealing with bugs in the !old drivers [21:09] hehe.. yep.. :) [21:09] speaking of which, I haven't quite understood how hardware without KMS will work under the new scheme. [21:09] just my question.. :) [21:09] thinking about Nvidia in particular.. [21:09] UMS still works [21:10] for now anyway [21:10] e.g. Geode is not likely to ever get KMS, now that OLPC switched to VIA [21:10] I'm still waiting for xrandr on some drivers [21:11] then again, kernel 2.6.31 is still no go on Geode anyhow. bug filed, camera snapshot added, LKML informed. no action. [21:13] ogasawara has been trying to help me narrow down the exact commit that broke things. we sort-of spotted it, but LKML has not acted upon it. instead Ingo went postal and called our bisect botched. [21:13] gotta love upstream [21:13] yup [21:14] :) [21:14] I suspect some hw corporations paying for X development that ends up breaking hw produced by corporations not paying for X development is not a stable configuration [21:14] but basically, older Geodes already broke around 2.6.23 and now recent Geodes broke with 2.6.31, which basically kills the whole LTSP hardware base. [21:31] Q-FUNK: you can thank AMD for that.. [21:32] jg: how so? [21:32] though I wouldn't say it is "the whole LTSP hardware base".... [21:32] Q-FUNK: they bought the Geode, and ended up laying off a very good guy who was maintaining the X (and other) drivers. [21:33] yup. I know him. I've been involved in xf86-video-geode. [21:33] OLPC could have just as easily hired him, but didn't. [21:33] heh. [21:34] I would have, if I could have... [21:34] jg: that you could not says a lot about the viability of the whole project, TBH. [21:34] Q-FUNK: or priorities of certain people.... [21:35] I wouldn't draw that conclusion.... [21:35] or lack of [21:40] Jordan had actually planned to get around implementing KMS and was looking forward to it. [21:40] now, the only one I can think of who might pull it off would be dilinger. [21:42] here, I'm mostly worried of how many nice surprises like the devPrivates breakage we're gonna encounter before -geode becomes hopelessly broken. [21:46] yup. [22:16] or cjb, but he keeps on having less and less time for -geode. [22:34] bryce: every so often, my screen goes off, and then back on again, when it is idle. Seems Koala's screen saver or blanking is busted.... [22:34] jg, yeah everyone is seeing that [22:35] jg, and everyone is claiming "not our fault" [22:35] ok, fine... I'll let you help throw rotten eggs.... [22:36] heh [22:37] jg, let me retrieve the blog about it [22:38] http://blogs.gnome.org/hughsie/2009/08/17/ [22:39] jg, I've pulled in the various patches and fixes described up through that blog post, but it seems people still have problems (as do I) [22:40] bryce: wasnt the gpm side fixed? what is delaying the xorg fixes? upstream or hasnt landed for Ubuntu? [22:40] mac_v, afaik everything on the X side is there [22:41] * mac_v constantly getting the icon in the panel , pointing to that blog :( [22:42] yeah, I've seen the icon in the panel, but it wasn't kind enough to let me cut and paste the URL. [22:43] that was what asac was complaining too ;) [22:43] yep, both patches mentioned there went into our xserver, on aug 14 and 19th [22:43] icon? [22:43] last month! o.0 [22:43] notification area icon [22:43] not following [22:43] apport crash icon? [22:44] bryce: you know who else has had to deal with this sort of stuff, who might have a valid opinion of what the "right thing" is, is cjb.... I'm not sure I believe hugsie' [22:45] claims, though. [22:45] I see the blanking problem even when I'm nowhere near the keyboard. [22:45] (like just now while watching video). [22:45] bryce: nope > http://launchpadlibrarian.net/30946483/Screenshot.png and Bug #423694 [22:45] Launchpad bug 423694 in gnome-power-manager "session active, not inhibited, screen idle message" [High,Invalid] https://launchpad.net/bugs/423694 [22:46] now messing with DPMS shouldn't be resetting any notion the X server has of idleness, I don't thin.... [22:46] so there may be multiple interacting bugs.... [22:47] yeah could be [22:47] hrmble [22:48] probably I should instrument all the dix screensaver stuff [22:49] maybe after all these freezes and crashes are solved ;-) [22:52] bryce: the low fps with radeon and KMS ? can we expect a solution before karmic? :) [22:52] from upstream [22:56] dunno [22:56] I'm about to go on leave for a month, though [23:03] have a fun vacation , when you come back , we'll have loads of bugs waiting for you ;p [23:08] it's paternity leave, hardly a vacation ;-) [23:21] bryce: first one? [23:23] yea [23:26] neat :) [23:37] 904 ships with xf86-video-intel 2.5.x right? [23:38] 2.6.3 actually - http://packages.ubuntu.com/jaunty/xserver-xorg-video-intel [23:38] ah ok thanks [23:39] bryce: was there anything left to do to get that geode autoconfig patch merged into karmic? [23:40] Q-FUNK, what patch? [23:40] I noticed that you confirmed the bug after I added the 1.4 and 1.6 patches, but nothing since then. [23:41] http://launchpadlibrarian.net/31292587/xserver_1.6.x_add_Geode_GX1_GX2_LX_to_xf86AutoConfig.patch [23:41] oh I have scripts that do that [23:42] bug #423866 [23:42] Launchpad bug 423866 in xorg-server "AutoConfigure support for older Geodes" [Undecided,Confirmed] https://launchpad.net/bugs/423866 [23:45] those patches [23:45] there's one for jaunty+karmic and one for hardy [23:46] submitted to debian as well [23:48] alright, I can do the karmic one. For the other you'll need an sru [23:48] indeed [23:49] hmm, patch didn't apply [23:49] no? [23:49] Applying patch 186_autoconfig_geode.patch [23:49] patching file hw/xfree86/common/xf86AutoConfig.c [23:49] Hunk #1 FAILED at 144. [23:49] Hunk #2 succeeded at 152 (offset 1 line). [23:49] 1 out of 2 hunks FAILED -- rejects in file hw/xfree86/common/xf86AutoConfig.c [23:49] Restoring hw/xfree86/common/xf86AutoConfig.c [23:49] Patch 186_autoconfig_geode.patch does not apply (enforce with -f) [23:49] Restoring hw/xfree86/common/xf86AutoConfig.c [23:49] failed! (check stampdir/log/patch for details) [23:49] make: *** [stampdir/patch] Error 1 [23:49] dpkg-buildpackage: error: debian/rules build gave error exit status 2 [23:49] E: Failed autobuilding of package [23:49] hm [23:50] you picked the 1.6 patch? [23:50] here, it worked when I built test packages for my PPA [23:50] right, commit 842b7a47d583a77 from the last comment [23:51] that commit might not apply as is [23:51] besides, I made minor adjustments in upstream git (mainly additional comments) [23:51] however, the attached patch is diffed against whatever is in karmic and jaunty [23:53] IIRC upstream git has a somewhat noticable diff between 1.6.3 and... was it 1.6.99 now? [23:54] ah well thanks for uploading it [23:55] er upstreaming it [23:55] uploaded other patch, hope that'll work for you [23:56] it should [23:56] in karmic, at least, the fixed gx2 support is already there, so the only part missing is that autoconfig [23:57] https://bugs.launchpad.net/debian/+source/xorg-server/+bug/423866/comments/9 [23:57] Ubuntu bug 423866 in xorg-server "AutoConfigure support for older Geodes" [Undecided,Confirmed] [23:57] this particular comment explains the whole situation