| [07:36] <ubotu> New bug: #179797 in xserver-xorg-video-intel (main) "intel GM965 (X3100) no TV-Out (S-Video)" [Undecided,New] https://launchpad.net/bugs/179797 | |
| [10:11] <ubotu> New bug: #179822 in linux-restricted-modules-2.6.24 (restricted) "[hardy] Enabling xinerama crashes X" [Undecided,New] https://launchpad.net/bugs/179822 | |
| [11:11] <ubotu> New bug: #57338 in xorg-server (main) "Graphic mode is broken (dup-of: 57153)" [Medium,Incomplete] https://launchpad.net/bugs/57338 | |
| [11:46] <ubotu> New bug: #179839 in xorg (main) "shift key overflow locks keyboard" [Undecided,New] https://launchpad.net/bugs/179839 | |
| [14:40] <tjaalton> whee, soon all the xorg-server bugs fit on one page (ie. less than 75 bugs) \o/ | |
| [17:04] <ubotu> New bug: #174156 in xorg (main) "mouse pointer disappears in hardy" [Undecided,Incomplete] https://launchpad.net/bugs/174156 | |
| [17:07] <ubotu> New bug: #158482 in xorg (main) "Wrong Live CD resolution" [Undecided,Incomplete] https://launchpad.net/bugs/158482 | |
| [17:08] <ubotu> New bug: #159764 in xorg (main) "kubuntu / xubuntu 7.10 cannot start X, find wrong PCI slot" [Undecided,Incomplete] https://launchpad.net/bugs/159764 | |
| [17:13] <ubotu> New bug: #73606 in xorg (main) "Ubuntu automatically X server configuration" [Wishlist,In progress] https://launchpad.net/bugs/73606 | |
| [18:47] <bryce> tjaalton: whoa, you've been busy!! | |
| [18:54] <bryce> tjaalton: calc says he's getting a backtrace when enabling xinerama with nvidia which he didn't have pre-break - was there an upload of -nv or lgm recently? | |
| [19:34] <tjaalton> bryce: lgm? you mean lrm? I've prepared a new lrm with new nvidia & ati, but haven't uploaded it | |
| [19:36] <tjaalton> there is a bug report about xinerama & nvidia crashing on gutsy, and today someone reported that against hardy | |
| [19:39] <bryce> yeah sorry lrm | |
| [19:39] <bryce> ah, bet that's it | |
| [19:40] <bryce> I've asked calc to file a bug, I'll let you know when he tells me the bug id# | |
| [19:40] <bryce> I also notice there's a new -ati 6.7.197-1? | |
| [19:41] <tjaalton> and a sync request, yes | |
| [19:41] <tjaalton> bug 144777 | |
| [19:41] <ubotu> Launchpad bug 144777 in xorg-server "xinerama xserver crash" [Undecided,New] https://launchpad.net/bugs/144777 | |
| [19:42] <tjaalton> that reporter claims that some patch breaks it | |
| [19:42] <bryce> ah | |
| [19:42] <tjaalton> bug 179822 | |
| [19:42] <ubotu> Launchpad bug 179822 in linux-restricted-modules-2.6.24 "[hardy] Enabling xinerama crashes X" [Undecided,New] https://launchpad.net/bugs/179822 | |
| [19:42] <tjaalton> that's the hardy version | |
| [19:44] * bryce passes along | |
| [19:44] <bryce> hmm, however the backtrace calc posted looks different | |
| [19:44] <bryce> <calc> Backtrace: | |
| [19:44] <bryce> 0: /usr/bin/X(xf86SigHandler+0x6a) [0x483d5a] | |
| [19:44] <bryce> 1: /lib/libc.so.6 [0x2b099b4e6100] | |
| [19:44] <bryce> 2: /lib/libc.so.6(realloc+0x79) [0x2b099b52bd79] | |
| [19:44] <bryce> 3: /usr/bin/X(Xrealloc+0x1b) [0x56753b] | |
| [19:44] <bryce> 4: /usr/bin/X [0x4bdde9] | |
| [19:44] <bryce> 5: /usr/lib/xorg/modules/drivers//nvidia_drv.so [0x2b099e1ae7a3] | |
| [19:44] <bryce> that was the backtrace it gives when enabling xinerama | |
| [19:46] <tjaalton> ok | |
| [19:46] <tjaalton> hum, ice cream :) -> | |
| [19:47] <tjaalton> (and Heroes.. S01E10 or so :P ) | |
| [20:16] <bryce> tjaalton: did you see the dec22 comment here about fglrx incompatibiltiy on hardy? http://lwn.net/Articles/263048/#Comments | |
| [20:17] <bryce> I'm assuming that one's fixed already (not enough info to say) | |
| [20:32] <Q-FUNK> re | |
| [20:32] <Q-FUNK> bryce: hiya! it seems that what makes DDC fail with -amd is some new x86 emulator code that crept in since X core 1.3 | |
| [20:33] <bryce> heya Q-FUNK | |
| [20:33] <Q-FUNK> bryce: various people have been investigating this and noticed some weir redundant bits in the X core code. | |
| [20:34] <bryce> what needs to be done to fix it? | |
| [20:34] <Q-FUNK> I'm starting to guess that more and more functionalities that were previously implemented in each driver are slowly being meatballed into more generic features and moved to X core. | |
| [20:34] <bryce> sounds likely | |
| [20:35] <Q-FUNK> some BIOS apparently don't like being tempered with twice when polling VBE registers | |
| [20:35] <Q-FUNK> this is what's been causing the freezes that -amd users on platforms that use General Software BIOS have been experimenting. | |
| [20:36] <Q-FUNK> the driver probably need to have large parts of it #ifdef'ed to skip them and #include newer X core features instead. | |
| [20:37] <Q-FUNK> then again, since this used to work mostly OK until Edgy (X core 1.2), there could be some nasty bugs in that new generic x86 emulator inside X core. | |
| [20:39] <Q-FUNK> bartman has been working on narrowing it down. there's been a new geode list created on x.org that is very active lately, with several players involved in tracking and fixing this. | |
| [20:39] <bryce> who will be trying out #ifdeffing stuff? | |
| [20:39] <Q-FUNK> bartman | |
| [20:39] <bryce> ah, bartman? | |
| [20:40] <Q-FUNK> a kernel developer | |
| [20:40] <bryce> good to hear about the activity | |
| [20:40] <bryce> is that bart massey? | |
| [20:40] <Q-FUNK> he's dabbling into X code simply because he has extensive knowledge of things like polling into BIOS registers | |
| [20:40] <Q-FUNK> Bart Trojanowksi | |
| [20:40] <jcristau> couldn't it be caused by using x86emu in 1.3 instead of vm86 in 1.2? | |
| [20:41] <Q-FUNK> http://www.x.org/wiki/AMDGeodeDriver | |
| [20:41] <bryce> do you need my help with anything in particular with it? I've been keeping an eye on things from the sidelines but don't know much about driver internals so haven't had much input to share | |
| [20:41] <Q-FUNK> jcristau: that is the current suspicion, yes | |
| [20:41] <Q-FUNK> x86emu sems at cause | |
| [20:42] <jcristau> not related to new code then | |
| [20:42] <Q-FUNK> http://lists.x.org/mailman/listinfo/xorg-driver-geode | |
| [20:42] <Q-FUNK> x86emu was already there in 1.2, but not used? | |
| [20:43] <Q-FUNK> here, the same -amd source code compiles and runs flawlessly against X 1.1, then bugs like unability to use ctrl+alt+Fx to change console creep in in X 1.2, even though it works mostly OK. | |
| [20:44] <Q-FUNK> by the time we try to build against 1.3, several competing hardware manufacturers reported complete system freeze | |
| [20:45] <jcristau> Q-FUNK: x86emu has been there forever | |
| [20:45] <Q-FUNK> bryce: we're mostly trying to figure out what x86emu is trying to do that freeze the system on some BIOS | |
| [20:45] <jcristau> just unused by default on x86_32 | |
| [20:46] <Q-FUNK> ah. ok | |
| [20:47] <bryce> tjaalton: seems the ati logo issue is back | |
| [20:47] <bryce> <brian> bryyce: I installed the fglrx driver and I have some weird watermark on my desktop say "AMD Testing use only Unsupported hardware" | |
| [20:47] <bryce> <bryyce> brian, was this the version from ubuntu, or downloaded from the ati site? | |
| [20:47] <bryce> <brian> bryyce: I used restricted-manager to install it so the Ubuntu version | |
| [20:47] <Q-FUNK> I wonder what x86emu does that make -siliconmotion and -amd fail with X 1.3 | |
| [20:50] <Q-FUNK> jcristau: any particular reason for emulating x86 on real x86 hardware instead of keeping with vm86? | |
| [20:51] <jcristau> Q-FUNK: it works on x86_64 kernels. vm86 doesn't. | |
| [20:51] <jcristau> plus, it makes it more likely that x86emu bugs get fixed, which benefits all archs | |
| [20:51] <Q-FUNK> true, but it leaves users in the dust until this is fixed. | |
| [20:52] <Q-FUNK> for all we know, -amd bugs are caused by x86emu and our -amd code is fine. | |
| [20:53] <jcristau> that's certainly possible | |
| [22:07] <tjaalton> bryce: well, the current lrm has fglrx 7.11 (aka 8.43), which is newer than the previous 8.42. maybe it's just a problem with 2.6.24 instead | |
| [22:08] <tjaalton> and I don't remember doing anything about the logo :) | |
| [22:55] <ubotu> New bug: #177829 in linux-restricted-modules-2.6.24 (restricted) "Please update fgrlx to 8.44 aka 7.12" [Undecided,New] https://launchpad.net/bugs/177829 | |
| [23:41] <tjaalton> bryce: I'll upload a new lrm with latest stable nvidia tomorrow. no fglrx update since it seems to have some known issues (as always) | |
| [23:42] <bryce> ok | |
| [23:50] <bryce> tjaalton: http://people.ubuntu.com/~bryce/plots-0.1.html | |
| [23:50] <bryce> been hacking this together today | |
| [23:51] <tjaalton> cool, please add xorg-server too :) | |
| [23:52] <bryce> unfortunately xorg-server data hasn't been collected (http://people.ubuntu.com/~brian/complete-graphs/) | |
| [23:52] <tjaalton> oh right | |
| [23:52] <bryce> there is also displayconfig-gtk, but graphs for them haven't been generated | |
| [23:52] <tjaalton> an now that it has less than 70 bugs.. | |
| [23:52] <bryce> I'll bug brian about it later | |
| [23:53] <tjaalton> I cleaned xorg&xorg-server bugs from crashers that seemed to be driver specific, so that's why -ati/-intel/lrm bug counts have soared :/ | |
| [23:54] <tjaalton> and -nv, obviously | |
| [23:54] <tjaalton> the launchpad janitor seems to be offline, btw | |
| [23:55] <tjaalton> there would be 2000 closed bugreports if it ran today | |
| [23:55] <bryce> ahh, I wondered about that | |
| [23:56] <bryce> I think I read in the last lp update that the janitor was showing what *would* expire, but wasn't actually doing the expiration | |
| [23:56] <bryce> I gather they want to implement code to allow projects to opt-in to the expiration stuff | |
| [23:57] <tjaalton> I guess that's a no-brainer for ubuntu at least.. | |