Sarvatt | http://git.gnome.org/browse/gnome-settings-daemon/log/?h=randr-rotate-wacom-tablet | 00:00 |
---|---|---|
Sarvatt | thats the rotate stuff you're looking for I assume, automatically rotates the tablet with the display rotation settings | 00:02 |
Alan | that is definitely a help | 00:02 |
Alan | the other part is actually hooking the ACPI events for the lid rotation | 00:03 |
Alan | so that it does something in the first place... | 00:03 |
Alan | also, that code is a little scary | 00:03 |
Sarvatt | yeah thats a kernel problem, I think hp touchpads are the only ones that do that now | 00:03 |
Sarvatt | err hp touchbooks? whatever they are called | 00:03 |
Alan | but mainly because snippets of your xorg configuration become part of the command line.... | 00:03 |
Alan | you can do it on lenovo thinkpad x200t | 00:04 |
Alan | which is what I have | 00:04 |
Alan | but you need to add the ACPI event and handler scripts yourself.... | 00:04 |
Alan | oh | 00:04 |
Alan | you're kidding.... | 00:04 |
Alan | looks like it's stopped working in Lucid | 00:04 |
Alan | why's that... | 00:04 |
Sarvatt | needs udevification? | 00:05 |
Alan | hmm | 00:07 |
Alan | probably.... | 00:07 |
VladimirBG | can someone help me with xorg.conf in Ubuntu 10.04 | 00:07 |
Sarvatt | VladimirBG: if you're using KMS that option has no effect, you need to use a kernel module option instead (which should already be defaulted on if your card supports it) | 00:08 |
Sarvatt | if that question was in reference to dynclks that you asked about earlier | 00:08 |
Alan | interesting, not firing the ACPI events :( | 00:09 |
Sarvatt | VladimirBG: echo "options radeon dynclks=1" | sudo tee /etc/modprobe.d/radeon.conf | 00:09 |
VladimirBG | Sarvatt, how do I implement that command, just copy/paste in terminal, or? | 00:10 |
Sarvatt | Alan: acpi_listen doesn't show it? | 00:11 |
Sarvatt | VladimirBG: yeah just copy/paste that into a terminal | 00:11 |
Sarvatt | VladimirBG: first though, check if its already enabled | 00:11 |
VladimirBG | I apologize for my ignorance, I'm a long time windows user... | 00:11 |
Sarvatt | cat /sys/module/radeon/parameters/dynclks | 00:12 |
VladimirBG | I pasted it, and got "-1" | 00:12 |
Alan | Sarvatt: the events happen, acpid doesn't appear to be firing them... | 00:13 |
Alan | also, i just tried to run acpid in the foreground to see what's happening, and Xorg crashed when i put the lid down... | 00:13 |
Alan | weird stuff.. | 00:13 |
Alan | WOW | 00:14 |
Alan | nice | 00:14 |
Alan | ay 3 23:55:43 scribus kernel: [ 200.579077] Xorg[1273]: segfault at 18 ip 00007f8bf27a17ce sp 00007fffc6581940 error 4 in wacom_drv.so[7f8bf2799000+12000] | 00:15 |
Alan | In an unrelated note, Virtualbox needs to catch up and support a non-HAL'd Ubuntu :( | 00:17 |
Sarvatt | Alan: try wacomrotate from this PPA, this guy has the same tablet and set it all up as a deb - https://edge.launchpad.net/~thjaeger/+archive/tabletpc | 00:17 |
VladimirBG | Sarvatt, when I used the cat command, I got -1, does it mean it is already enabled? Also, http://manpages.ubuntu.com/manpages/lucid/en/man4/radeon.4.html shows an option called "ForceLowPowerMode", will that work? | 00:17 |
Sarvatt | VladimirBG: no the powermanagement options in xorg.conf don't work with KMS, if you really want to use them then just echo "options radeon modeset=0" | sudo tee /etc/modprobe.d/radeon.conf | 00:18 |
VladimirBG | ok | 00:19 |
VladimirBG | thanks | 00:19 |
Sarvatt | if you do that one then they'll all work and ya can go to town with all the stuff in the wiki | 00:19 |
Sarvatt | only thing you'll really notice is the boot will be uglier and your terminals ugly again, but things will be faster :D | 00:20 |
VladimirBG | it that command permanent? | 00:20 |
VladimirBG | meaning I don't need to use it after? | 00:20 |
Sarvatt | yeah, to remove it just sudo rm /etc/modprobe.d/radeon.conf | 00:20 |
Alan | Sarvatt: will try that, thanks - such a thing didn't exist when i bought this laptop! | 00:21 |
VladimirBG | Sarvatt, done. so, after I reboot, I'll have power management enabled on my gpu and have longer battery life? | 00:21 |
Sarvatt | depends on if your GPU supports it :) you still have to add the xorg.conf options too | 00:22 |
Sarvatt | power management sucks on hd2000+ at the moment but they are working on it heavily upstream | 00:22 |
VladimirBG | so I need to make xorg.conf, add some generic stuff like this: http://pastebin.com/cJ2ynTRi and add Option "ForceLowPowerMode" "enable" ? | 00:24 |
Sarvatt | yup, add it to the device section underneath driver "radeon" and before EndSection | 00:25 |
Sarvatt | try adding Option "ClockGating" "1" Option "DynamicPM" "1" Option "ForceLowPowerMode" "1" | 00:26 |
Sarvatt | (all on seperate lines) | 00:26 |
VladimirBG | can you give me a pastebin of that? | 00:27 |
VladimirBG | please? | 00:27 |
Alan | Sarvatt: that wacomrotate is doing exactly nothing | 00:27 |
Alan | :( | 00:27 |
Sarvatt | http://pastebin.com/w3Tzg1ts | 00:27 |
Sarvatt | Alan: you rebooted? | 00:28 |
VladimirBG | thanks! | 00:28 |
Alan | Sarvatt: yup | 00:28 |
Sarvatt | maybe he hasn't updated it for the latest stuff yet :( | 00:28 |
VladimirBG | by the way, what is KMS? | 00:28 |
Sarvatt | it puts the kernel in charge of low level things like managing the memory on the GPU and setting the resolution | 00:30 |
Sarvatt | (and power management in this case) | 00:30 |
Sarvatt | sad as it is i'd really recommend just using the crappy fglrx if you can and you're worried about power management | 00:30 |
VladimirBG | so, I need to also turn that off? | 00:31 |
Sarvatt | VladimirBG: thats what "options radeon modeset=0" | sudo tee /etc/modprobe.d/radeon.conf did | 00:31 |
Sarvatt | it's enabled by default for the pretty boot stuff | 00:32 |
Sarvatt | not that it makes much of a difference getting loaded so late in the boot process as it is now :) | 00:32 |
VladimirBG | I love Ubuntu... but getting little thing to work puts a lot of people off... But I guess that is the price of free... | 00:34 |
Sarvatt | VladimirBG: just curious, why don't you use the proprietary drivers on your system? | 00:35 |
VladimirBG | ah... used to be too much hasle a few years back | 00:36 |
VladimirBG | and video playback was bad | 00:36 |
VladimirBG | aside from power management, radeon is really nice | 00:36 |
Sarvatt | yeah its still crappy but its pretty much required if you care about battery life on a recent GPU, that will definitely change by 10.10 though | 00:37 |
Alan | Sarvatt: hmm, methinks there's a bug here... :( | 00:37 |
VladimirBG | so, the lesson is next laptop has either nvidia or intel? | 00:38 |
Alan | May 4 00:37:09 scribus kernel: [ 823.134541] Xorg[1241] general protection ip:7f3e7c4947ce sp:7fff89fe3540 error:0 in wacom_drv.so[7f3e7c48c000+12000] | 00:38 |
Alan | more segfaults :( | 00:38 |
bjsnider | VladimirBG, the linux graphics situation across the board is a disaster. you can't go right no matter what hardware you buy | 00:39 |
VladimirBG | what is the least bad solution? | 00:40 |
Alan | There we go, "xsetwacom set stylus rotate cw" is enough to segfault Xorg :( | 00:40 |
bjsnider | VladimirBG, it involves firearms | 00:40 |
VladimirBG | bjsnider, any light at the end of tunnel? Any hope for a good driver set? | 00:45 |
VladimirBG | for any manufacturer? | 00:46 |
Sarvatt | VladimirBG: you'll probably find something annoying no matter which you pick, ATI is good but power management with KMS is a sore spot with it at the moment, it getting a lot of work done on it and should be better with 2.6.35. nvidia has great binary drivers but you're stuck with problems like being unable to suspend/resume for months until they fix them. intel gpu's just suck in general :) | 00:46 |
RAOF | VladimirBG: *Modern* intel chips tend to be well supported, but they're pretty sucky chips. | 00:46 |
bjsnider | well, we all thought gallium3d would solve most problems, but now the developers don't want it because there'd be too much pain and it's only a vmware thing. so mesa's crappiness remains | 00:47 |
* RAOF concurs with Sarvatt: the open ATI drivers are most likely to get better fastest. | 00:47 | |
cozzimoto | hey guys.. anyone having issues installing nvidia drivers from nvidia on lucid and also setting fast writes and sba on agp cards? | 00:47 |
bjsnider | you cannot use the nvidia-installer on lucid | 00:48 |
Alan | going to live dangerously and try xorg-edgers.... | 00:48 |
cozzimoto | bjsnider, i noticed :) | 00:49 |
cozzimoto | bjsnider, why is that/ | 00:49 |
VladimirBG | If I install fglrx from repo, it will disable radeon, and power management will work? | 00:50 |
bjsnider | i'm tired of explaining why | 00:50 |
bjsnider | cozzimoto, use jockey and do whatever it wants you to | 00:50 |
cozzimoto | bjsnider, wow i am tired of doing support myself with things | 00:50 |
RAOF | VladimirBG: Yup. Use System→Preferences→Hardware Drivers for the most-tested route. | 00:50 |
cozzimoto | bjsnider, well apparenlty i have no choice unfortunately | 00:50 |
VladimirBG | RAOF, I tried that, but fglrx isn't offered there, but I did find it in ubuntu software center | 00:51 |
cozzimoto | bjsnider, may add the reason in the topic :0 | 00:51 |
bjsnider | cozzimoto, i'm for that | 00:51 |
RAOF | VladimirBG: Hm. If fglrx isn't offered there it's entirely possible that fglrx no longer supports your card. What *is* your card? | 00:52 |
cozzimoto | i meant to say "maybe add it" | 00:52 |
bjsnider | that's certainly why. it scans for the pciid | 00:52 |
bjsnider | it doesn't match a supported card | 00:52 |
cozzimoto | what about enabling fast writes and SBA on agp nvidia cards...nothing seems to work on lucid | 00:53 |
VladimirBG | RAOF, mobility radeon x1300 | 00:53 |
bjsnider | yep, not supported | 00:54 |
bjsnider | cozzimoto, how did you do that before lucid? | 00:54 |
cozzimoto | bjsnider, never on lucid just on pevious ubuntu versions on lucid it doesn work | 00:55 |
bjsnider | i said before lucid | 00:55 |
bjsnider | previous distros | 00:55 |
cozzimoto | bjsnider, i would edit /etc/modprobe.d/nvidia-kernel-nkc | 00:56 |
cozzimoto | alias char-major-195* nvidia | 00:57 |
cozzimoto | options nvidia NVreg_EnableAGPSBA=1 NVreg_EnableAGPFW=1 | 00:57 |
cozzimoto | reboot and magic happened :) | 00:57 |
RAOF | You'll probably need to s/nvidia/nvidia-current/g; the kernel module got renamed to allow parallel-installability. | 00:58 |
bjsnider | if he even has it installed. it sounds like it isn't | 00:59 |
cozzimoto | RAOF, I tried that once no luck | 00:59 |
cozzimoto | bjsnider, the nvidia_current is installed I had no choice | 00:59 |
bjsnider | what? | 01:00 |
cozzimoto | although on the one system i reinstalled karmic so I didnt have to deal with this :) | 01:00 |
bjsnider | what do you mean you had no choice? | 01:00 |
cozzimoto | bjsnider, I cannot installe the current nvidia driver from nvidia,,which is what i started with | 01:01 |
bjsnider | yes, because it will overwrite system files. that's always beent he problem with it | 01:01 |
cozzimoto | bjsnider, no I have never had issues with the official nvidia driver | 01:01 |
cozzimoto | only on lucid | 01:02 |
bjsnider | but it has always overwritten and destroyed mesa files, you just didn't notice | 01:02 |
cozzimoto | bjsnider, mm that may have affected compiz at some point no? | 01:03 |
bjsnider | no, it replaces mesa's libglx files with its own | 01:04 |
cozzimoto | sorry | 01:04 |
bjsnider | i don't know why, since mesa works so smashingly well, and always has | 01:04 |
cozzimoto | bjsnider, well still I always preferred the official driver but with lucid "i have no choice" | 01:05 |
bjsnider | oh, you can override the block if you really want to. i'm sure google has the instructions somewhere | 01:05 |
cozzimoto | ok I will try to figure out the fast writes and SBA at some point | 01:05 |
bjsnider | RAOF, the intel chips may suck, but their windows driver destroys the linux driver in opengl performance in the recent phoronix tests | 01:06 |
bjsnider | although he did note amusingly that they locked up the windows kernel on numerous occasions | 01:07 |
RAOF | bjsnider: Heh. | 01:07 |
RAOF | That's entirely possible; our graphics stack still hasn't had a bunch of optimisation work done. | 01:07 |
RAOF | Sarvatt: Thinking of which… are you building r300g in xorg-edgers alongside nouveau? My ati box isn't here yet, but I'm interested to see how it's going; as far as I've gathered it seems to be roughly equivalent to the classic driver ATM. | 01:09 |
Sarvatt | RAOF: yeah libgl1-mesa-dri-gallium | 01:12 |
RAOF | I suspect that we might want to be shipping that in Maverick. We probably want to have nouveau gallium in there, at least. | 01:13 |
RAOF | Discussions can be had in a week or so! :) | 01:14 |
jg | Sarvatt: OK, I get basically the same results with both 32 and 64 bit images, with and without modesetting: the panel does not appear to be on once X has started. | 01:14 |
Sarvatt | the intel drivers in windows are such an abomination compared to the linux ones, the 945g ones were lucky to get yearly updates the past few years and i had to do allll kinds of faking support for features to get games to work (which is probably where some of the performance losses are from since at least you get software fallbacks for those features in linux) | 01:14 |
Sarvatt | i had the 965gm drivers hacked up to work on 945 so anything recent that needed vertex shader support could work | 01:15 |
jg | Sarvatt: the system does seem to mostly boot up; sometimes I get the cheery ubuntu audio sound to make me think so. | 01:16 |
Sarvatt | I cant find any bugs with the same symptoms :( do you see the splash screen during the boot? | 01:25 |
Sarvatt | did you put the livecd's on a USB stick with persistant storage space so you can see the logs by any chance? | 01:26 |
jg | Sarvatt: I've not tried usb sticks. I can try to dig one up. | 01:33 |
jg | I get the little keyboardish icon on the bottom of the screen, and press a space bar so I can interact with the menu. | 01:34 |
jg | If KMS is off, I get the textish version of Ubuntu with dots that animate as it spins the CD a long time. | 01:34 |
jg | If KMS is on, I don't get that; the screen goes pretty close to black, while the CD spins a long time. | 01:35 |
jg | Sarvatt: I'll try to grab a usb stick tomorrow. do I have to do anything magic to get persistent storage? | 01:38 |
ilmari | choose it in usb-creator | 01:46 |
Bernardo | good morning | 04:59 |
furan | hio | 05:46 |
furan | anyone messed with lucid & gma500 recently? | 05:46 |
RAOF | furan: Yes indeed. Both Sarvatt and Bernardo have been looking at it. | 05:52 |
furan | I saw a blip on an updated ppa(https://launchpad.net/~lucazade/+archive/ppa) but it's still building 32 bit | 05:53 |
furan | I was curious if I could just download and build myself | 05:53 |
RAOF | Is there anyone here crazy enough to be using Xinerama? If so, how's your lucid experience? I'm noticing a number of bugs with Xinerama being the tying theme. | 06:02 |
Bernardo | hi all | 06:08 |
furan | hi | 06:09 |
furan | trying to get gma500 up on lucid, not sure where to start | 06:09 |
Bernardo | lol | 06:09 |
Bernardo | Lucazade set up a repository with me | 06:09 |
Bernardo | all we need now is to fix dri | 06:09 |
furan | the googlecode one? | 06:10 |
Bernardo | yes | 06:10 |
furan | ok | 06:10 |
Bernardo | I've ported all Mandriva patches, including the renaming of drm to psb-drm to that repository | 06:10 |
furan | cool | 06:11 |
furan | what's left to do on dri? | 06:11 |
Bernardo | Sarvatt here suggested the problem might be with udev, but even after renaming lbm-drm to psb-drm in the udev scripts it doens't work | 06:11 |
slytherin | I am getting very slow video playback with KMS enabled on Radeon 7000. Should I file a separate bug for this or reopen bug #544496? | 06:11 |
ubottu | Launchpad bug 544496 in xserver-xorg-video-ati "[RS100] desktop became slow after upgrading to Lucid" [Critical,Fix released] https://launchpad.net/bugs/544496 | 06:11 |
Bernardo | I don't know... The problem is that the modules get loaded (psb and psb-drm), with debug enabled there is no error, but still /dev/dri never gets created, and the xorg psb driver fails because of that | 06:12 |
Bernardo | I've already put my xorg log on the ubuntuforum thread, and my dmesg in dropbox, if you want to take a look | 06:15 |
Bernardo | furan: the repository in googlecode is http://code.google.com/p/gma500/ | 06:27 |
slytherin | What cab the reason that for a particular resolution part of the UI goes off-screen? | 08:16 |
RAOF | Because you're connecting your display over a VGA cable, which sucks? | 08:17 |
RAOF | Or because you have a thrice-damned LCD TV that refuses to believe that digital inputs shouldn't be overscanned? | 08:17 |
slytherin | RAOF: Looks like either or both are possible. With KMS this resolution does not work at all. Without KMS some UI goes off screen. | 08:18 |
RAOF | Are you actually using a VGA cable? | 08:19 |
slytherin | yes. and connected to an year old LCD TV (32"). | 08:19 |
RAOF | Mmm, double the fun! | 08:20 |
RAOF | So, if the top and bottom go offscreen then the answer will be “televisions do that deliberately”, and if it's shifted left/right the answer should be “fiddle with the display settings until the TV has scaled the image correctly”. | 08:21 |
RAOF | Mostly. | 08:21 |
slytherin | the display is shifted left. But that happens only for 1360x768 resolution. So I am inclines to think that the frequencies used by TV are not getting detected properly. | 08:22 |
RAOF | Actually making what comes out of a VGA cable centred on a display is somewhat of a hack. Which is why displays which accept VGA input generally let you shift the output until it matches what it needs to be. | 08:24 |
slytherin | hmm | 08:24 |
slytherin | by the way, should reopen bug 544496 if the symptoms are still present? Or should I file a new bug. | 08:25 |
ubottu | Launchpad bug 544496 in xserver-xorg-video-ati "[RS100] desktop became slow after upgrading to Lucid" [Critical,Fix released] https://launchpad.net/bugs/544496 | 08:25 |
RAOF | Other answers include “you can fiddle with the ModeLine (xrandr --newmode is your friend) until it's shifted nicely” which is basically doing what fiddling with the TV's alignemnt thingy would do, but without useful guidance. | 08:25 |
slytherin | I will try that | 08:26 |
RAOF | I'd only try that if the TV really doesn't have any way to adjust the picture; fiddling with the modeline will be annoying and slow. | 08:29 |
RAOF | I guess another option would be to buy a DVI or HDMI cable (and potentially a new laptop to plug those in :)) | 08:30 |
RAOF | slytherin: I'd be happy for you to reopen that bug; the upstream bug is still open, and Bryce wasn't sure that commit would help. | 08:33 |
RAOF | vline seems to be pretty much the same as vsync, which is a serious performance impediment if your card isn't powerful enough to draw at or above the screen's refresh rate. | 08:35 |
slytherin | RAOF: After setting resolution on PC and using 'Auto Adjust' on TV it looks almost perfect. And looks like I will need to buy a new graphics card soon. | 08:36 |
RAOF | slytherin: Turning off kms doesn't resolve the performance problem? | 08:36 |
slytherin | It does. | 08:36 |
slytherin | That's why I added comment in bug first. | 08:37 |
RAOF | Right. My understanding is that vline is only enabled with kms, which makes your performance problems make sense. | 08:37 |
RAOF | Possibly the other reporter's card got a kms blacklist ;) | 08:37 |
RAOF | There is hope from the upstream bug; there's a suggestion that things will improve once DRI2 swapbuffers have been implemented. | 08:38 |
slytherin | also with KMS enabled EXA is being used which is known not to work with cards with low memory. With KMS disabled, XAA is used. | 08:38 |
slytherin | I am also wondering if I should first try "RenderAccel off" with KMS, if that is going to make any difference. | 08:40 |
* RAOF is off to dinner | 08:41 | |
Bernardo | hi | 09:42 |
lucazade | hi | 10:02 |
=== JanC_ is now known as JanC | ||
Sarvatt | RAOF: it looks to be our friend vga16fb causing the problem this time | 15:00 |
Sarvatt | vgaarb is screwing up the blob with multiple cards | 15:01 |
tseliot | Sarvatt: I uploaded a patch about the vga arbiter and multiple cards | 15:10 |
tseliot | (with nvidia) | 15:10 |
Sarvatt | oh nice! i haven't had the time to dig into it more but that seemed to be what the problem was | 15:10 |
tseliot | Sarvatt: bug #573557 | 15:21 |
ubottu | Launchpad bug 573557 in nvidia-graphics-drivers "nvidia-current 195.36.15 + kernel 2.6.32 + dual cards crashes system" [High,Confirmed] https://launchpad.net/bugs/573557 | 15:21 |
Sarvatt | yeah I saw in your changelog message, uploaded it with 195.36.24 to x-updates too | 15:23 |
Sarvatt | phew i'm lucky dch doesn't pick up maverick yet :) | 15:24 |
tseliot | I uploaded that to lucid-proposed too but it will have to be accepted | 15:27 |
tseliot | Sarvatt: do we have this patch in lucid? http://cvs.fedoraproject.org/viewvc/rpms/xorg-x11-server/F-13/xserver-1.8.0-dri2-fix-handling-of-redirected-pixmaps.patch?view=markup | 15:30 |
Sarvatt | nope because its a fix for another patch series that only applies to xserver 1.8 (and was reverted to be done a different way upstream even) and we dropped the glx 1.3+ stuff so it shouldn't be needed | 15:36 |
Sarvatt | that was fixing a problem where compiz was showing the same image for different windows sometimes on xserver 1.8.0 | 15:36 |
VladimirBG | hello | 16:04 |
VladimirBG | how to get fglrx insted of radeon? Tried ubuntu software center, but ubuntu still uses radeon | 16:04 |
tseliot | VladimirBG: try Jockey | 16:05 |
VladimirBG | nothing is offered there | 16:05 |
tseliot | VladimirBG: what card are you using? | 16:06 |
tseliot | graphics card | 16:06 |
VladimirBG | ati mobility x1300 | 16:07 |
tseliot | Sarvatt: what's the fix that upstream adopted? | 16:07 |
tseliot | VladimirBG: I don't think your card is still supported by the fglrx driver | 16:08 |
VladimirBG | :( | 16:08 |
Sarvatt | VladimirBG: fglrx doesn't support that card, the power management options you can do in xorg.conf is as good as it's going to get for those because they dont have all of the power options that newer cards do anyway | 16:08 |
VladimirBG | thanks | 16:09 |
Sarvatt | ugh libdrm-psb needs a complete packaging overhaul now | 16:10 |
Sarvatt | tseliot: they reverted the commit that needed that patch for now :D | 16:11 |
tseliot | Sarvatt: and what are we going to do in Ubuntu? The netbook remix causes compiz to crash (as it uses clutter) | 16:12 |
Sarvatt | that patch didn't fix any crashes, it just fixed a display problem afaik.. | 16:12 |
Sarvatt | what is crashing? | 16:13 |
tseliot | Sarvatt: "application calling glxCreatePixmap when glx 1.3 is not supported" | 16:14 |
Sarvatt | its still crashing? | 16:14 |
tseliot | I have just installed Ubuntu on my netbook | 16:14 |
Sarvatt | its valid for compiz to call glxCreatePixmap because it needs to to use GLX_EXT_texture_from_pixmap and compiz has always done that (except when we had the glx 1.4 enablement) | 16:15 |
Sarvatt | its just a harmless message in that case, if its crashing its something else probably.. | 16:16 |
tseliot | you're right, the gnome decorator was not installed | 16:17 |
tseliot | never mind | 16:17 |
Sarvatt | ah yeah i've seen a few reports of that too, any idea how you got into the situation of not having compiz-gnome installed? | 16:17 |
tseliot | yes, I installed some compiz packages manually (as they are not installed in UNE) and forgot to install the decorator | 16:18 |
tseliot | I guess ubuntu-desktop installs all of the required packages | 16:18 |
Sarvatt | hmm wonder why mandriva is setting up the alternatives for the standard libdri.so and libglx.so in libdrm-psb | 16:30 |
Sarvatt | darn Bernardo is gone, need to see what the contents of /proc/ are for him when booting with psb | 16:34 |
outlaw45 | anyone else working on the gma500 on lucid at the moment? | 16:41 |
tseliot | Sarvatt: because poulsbo works with a rather old version libdrm | 16:43 |
tseliot | was it libdrm 2.3 ? | 16:47 |
tseliot | s/works/used to work/ | 16:47 |
outlaw45 | I believe its 2.3 indeed | 16:47 |
outlaw45 | but poulsbo doesn't work on lucid, myself and a couple of other guys are trying to get it working for a few days now | 16:48 |
outlaw45 | most credit goes to the other guys :p | 16:48 |
jcristau | credit goes to intel | 16:50 |
outlaw45 | not a big fan of intel at the moment | 16:50 |
outlaw45 | no working driver for x-server 1.7 from them! | 16:50 |
tseliot | I think it was powervr + intel | 16:50 |
tseliot | but yes... | 16:50 |
outlaw45 | yeah you're right powervr are the real bastards | 16:51 |
Sarvatt | i'm working on it here too but its rough without access to any actual hardware to see whats happening - https://edge.launchpad.net/~sarvatt/+archive/psb/+packages | 16:51 |
tseliot | do they packages use alternatives? | 16:52 |
Sarvatt | haven't gotten to the parts that need changing to alternatives yet :) | 16:52 |
outlaw45 | some guys think the problem is in drm | 16:53 |
outlaw45 | http://code.google.com/p/gma500/ | 16:54 |
outlaw45 | that's the repo we're using, maybe you can check if you're on the same path? | 16:54 |
lucazade | jbernardo found the problem - libdrm2 now installs to /lib instead of /usr/lib | 16:55 |
Sarvatt | yay svn | 16:55 |
lucazade | I removed /lib/libdrm.so.*, and the driver loaded! | 16:55 |
jcristau | svn + psb. all the fail in one place. | 16:55 |
jcristau | ;) | 16:55 |
outlaw45 | @lucazade: do you need to compile libdrm manual or did anyone edit the debian files? | 16:56 |
outlaw45 | back in a minute... | 16:56 |
Sarvatt | change the install rule to usr/lib/lib*.so.* /lib to fix that | 16:57 |
lucazade | ok... going to try it | 16:57 |
tseliot | lucazade: does it mean that the driver works or just that the module can be loaded? | 16:57 |
lucazade | the drivers works | 16:57 |
lucazade | *driver | 16:57 |
tseliot | interesting | 16:58 |
lucazade | jbernardo had anyway a segfault, and was dumped out of X | 16:59 |
Sarvatt | tseliot: yeah mandriva got the driver working with xserver 1.7 | 16:59 |
tseliot | ouch | 16:59 |
jcristau | so, err, rather than introducing yet another set of fucked up alternatives for psb, i don't suppose it's possible to rename the psb libdrm to libdrm_psb or so, and rename its symbols.. | 16:59 |
jcristau | but i guess "blob" means no | 17:01 |
tseliot | I guess it's possible but I don't remember why we didn't do it when we maintained the driver | 17:01 |
lucazade | i believe jbernardo already renamed the module to work with the mandriva patches (haven't checked myself) | 17:01 |
lucazade | "Rename of the module to contemplate the changes (drm to psb-drm, psb_ added to various prefixes, added mandriva patches), and most changes applied" revision6 | 17:03 |
=== |Bernardo| is now known as Bernardo | ||
Bernardo | HI | 17:06 |
Bernardo | I only changed the kernel module (drm to psb-drm) | 17:07 |
Bernardo | I was thinking of doing the same to the lindrm | 17:07 |
Bernardo | libdrm | 17:07 |
lucazade | hi | 17:08 |
Bernardo | seems like I lost tseliot by a couple of minutes | 17:09 |
lucazade | Sarvatt said: change the install rule to usr/lib/lib*.so.* /lib to fix that | 17:09 |
Bernardo | It is not enough | 17:09 |
Sarvatt | http://sarvatt.com/downloads/patches/02_fix_kms_detection_with_psb.diff | 17:10 |
Bernardo | I've already changed the preinstall and postrm scripts, just need to change the rules file to install in /lib | 17:10 |
lucazade | @Bernardo have you seen my pm? got some issue compiling xpsb-glx | 17:11 |
Bernardo | lucazade: I'm booting my netbook now, will check that asap | 17:11 |
lucazade | ok thanks | 17:11 |
Bernardo | Sarvatt: that is to apply to the "standard" libdrm2, or to the hacked libdrm-poulsbo? | 17:12 |
Sarvatt | all of the mesa stuff is going to need a huge amount of changes for the way things work in lucid | 17:13 |
Bernardo | ok, I guessed as much, when the driver started crashing like crazy | 17:18 |
Bernardo | lucazade: I cleaned xpsb-glx, I'm uploading it to svn asap | 17:19 |
Bernardo | just need to check if my tree is "clean" | 17:19 |
lucazade | Bernardo: ok | 17:20 |
Bernardo | outlaw45: give lucazade or me your email so we can add you to the project | 17:21 |
outlaw45 | quach.c@gmail.com | 17:21 |
lucazade | adding | 17:21 |
Bernardo | I've also updated libdrm-poulsbo, just need to add this patch from Sarvatt, what did you do? | 17:21 |
outlaw45 | gonna install psb-kernel-source package from sarvatt and then build libdrm | 17:21 |
Bernardo | are there many diffs between sarvatt's and ours? | 17:22 |
lucazade | outlaw45: added to the project | 17:22 |
outlaw45 | :p | 17:23 |
outlaw45 | Error! Could not find module source directory | 17:24 |
Bernardo | outlaw45: did you check the diff's between sarvatt's source and ours? | 17:24 |
outlaw45 | that's what I get if I try to install the package.. | 17:24 |
Bernardo | which are you building? | 17:24 |
outlaw45 | sarvatt's | 17:24 |
Bernardo | Try ours... :) | 17:24 |
Bernardo | and try purging the previous version before installing the new one, that usually helps with that error | 17:25 |
outlaw45 | purging did the trick :) thx | 17:26 |
outlaw45 | installing the one from svn at the moment | 17:27 |
Bernardo | lucazade: I am building xpsb-glx, if it finishes without problems I'll upload it to svn | 17:28 |
lucazade | Bernardo: perfect... there also some redudant .patch inside xorg-xserver-video-psb | 17:30 |
Bernardo | lucazade: I think I left one or two that have been superseeded, I'll check if I can remove them also | 17:31 |
Bernardo | rev 19 has the libdrm changes for /lib | 17:33 |
Bernardo | but still needs to be tested | 17:33 |
=== apachelogger is now known as huntlogger | ||
Bernardo | Sarvatt's patch doesn't apply, I'll check why before adding it | 17:36 |
outlaw45 | crappy system hear, multiple versions of psb.ko and drm-psb.ko | 17:37 |
outlaw45 | try to clean and retry | 17:37 |
outlaw45 | removed psb-kernel-source but still have | 17:39 |
outlaw45 | /lib/modules/2.6.32-21-generic/extra/psb.ko | 17:39 |
outlaw45 | /lib/modules/2.6.32-21-generic/updates/char/drm/psb.ko | 17:39 |
outlaw45 | anyone one know where they're from? | 17:40 |
Bernardo | psb-kernel-modules | 17:40 |
Bernardo | remove the package and those two | 17:40 |
outlaw45 | just rm? | 17:41 |
Bernardo | yes | 17:43 |
* Bernardo kicks svn, seems like my tree is borked (again) | 17:44 | |
outlaw45 | restart.. | 17:48 |
Bernardo | lucazade: rev20 has the correct xpsb-glx | 17:48 |
lucazade | ok checking | 17:49 |
outlaw45 | dh_install --sourcedir=debian/tmp -Xlibdrm.la --list-missing | 17:56 |
outlaw45 | dh_install: libdrm-poulsbo-dev missing files (usr/lib/lib*.a), aborting | 17:56 |
outlaw45 | that's what I get if I compile libdrm | 17:56 |
Bernardo | outlaw45: just fixed that on rev 21 :) | 17:56 |
outlaw45 | :p | 17:57 |
Bernardo | I had made the changes as I was compiling xpsb-glx and didn't check them :( | 17:57 |
outlaw45 | building again ;) | 17:57 |
lucazade | xpsb-glx gives me same error | 17:57 |
lucazade | trying also libdrm latest rev | 17:58 |
Bernardo | strange... Let me check again | 17:58 |
lucazade | libdrm same error of outlaw45 | 17:59 |
Bernardo | yep. I am going to delete it from svn and re-add it | 17:59 |
lucazade | ok | 17:59 |
Bernardo | lucazade: which svn rev? | 17:59 |
lucazade | 20 | 17:59 |
Bernardo | I've fixed libdrm on rev 21 | 17:59 |
lucazade | haven't seen | 18:00 |
outlaw45 | build libdrm and installing now.. | 18:01 |
outlaw45 | installed.. I'll try xpsb-glx now | 18:01 |
Bernardo | wait | 18:02 |
Bernardo | I'm uploading the fixed one in a few secs | 18:02 |
Bernardo | commiting now the working sources from lucazade's tar, will tell you when it finishes uploading | 18:04 |
lucazade | ok i'll wait | 18:05 |
outlaw45 | dpkg-checkbuilddeps: error: syntax error in debian/control at line 23: block lacks a package field | 18:08 |
outlaw45 | if I build xpsb | 18:08 |
lucazade | yes xpsb-glx is broken.. bernardo is uploading a new one | 18:09 |
outlaw45 | ow ok.. | 18:09 |
Bernardo | still uploading... it is big | 18:09 |
outlaw45 | gives me some time to finish my pizza :D | 18:09 |
lucazade | lol | 18:09 |
Bernardo | and xpsb-glx is one of the huge problems with the psb driver - it basically installs a older version of mesa | 18:10 |
lucazade | another thing: echo "blacklist i915" | sudo tee -a /etc/modprobe.d/blacklist.conf | 18:11 |
lucazade | is still necessary? | 18:11 |
outlaw45 | maybe intel releases an IEGD version which works with xserver 1.7 in the future.. | 18:11 |
Bernardo | lucazade: no idea | 18:11 |
Bernardo | but psb still corrupts the screen if you try to load it in fb mode | 18:12 |
Bernardo | outlaw45: they have already made some noises to the effect that they will jump 1.7 and work on 1.8 | 18:12 |
outlaw45 | yeah I read.. | 18:13 |
Bernardo | and the mythical driver using gallium3d tech seems to have vanished | 18:13 |
lucazade | like duke nukem 3d | 18:14 |
Bernardo | ok, rev 23 commited | 18:26 |
Bernardo | still need to change dependencies, it is looking for libdrm | 18:28 |
Bernardo | somewhere inside configure | 18:29 |
Bernardo | time to check my backups again | 18:29 |
lucazade | ;) | 18:30 |
Bernardo | lucazade: had you managed to build it in lucid? | 18:30 |
lucazade | no.. haven't been able to compile on lucid | 18:31 |
Bernardo | the strangest thing is that it used to build on my tree, but svn ultimately didn't like it | 18:32 |
lucazade | yes still looking for libdrm | 18:33 |
Bernardo | I'm trying to hack it now | 18:36 |
outlaw45 | building xpsb-glx takes a loooong time :p | 18:43 |
Bernardo | ok, it seems like pkgconfig isn't finding the file in /lib/pkgconfig/ (libdrm.pc) | 18:43 |
Bernardo | outlaw45: it is building for you? | 18:43 |
lucazade | come back soon | 18:47 |
Bernardo | temporary solution - copy /lib/pkgconfig/libdrm.pc to /usr/lib/pkgconfig, but obviously libdm-poulsbo install needs fixing | 18:50 |
outlaw45 | still building without errors... | 18:52 |
outlaw45 | dh_install: xpsb-glx missing files (drivers/*), aborting | 18:58 |
outlaw45 | only thing I haven't installed is xserver-xorg-video-psb | 18:58 |
lucazade | i'm back | 19:05 |
outlaw45 | got a file named libdrm.pc.libdrmdev in /usr/lib/pkgconfig.. | 19:06 |
outlaw45 | gonna copy libdrm.pc to /usr/lib/pkgconfig and try building again.. | 19:07 |
=== Rusty is now known as Guest20947 | ||
Rusty73 | lucazade, are you building deb only for amd64? | 19:18 |
lucazade | no.. also i386 but launchpad is going slow. | 19:19 |
lucazade | i'll also update those package as soon as they are working | 19:19 |
Rusty73 | launchpad made debs alone??? | 19:19 |
lucazade | i386 packages are in queue | 19:20 |
Rusty73 | Excuse my ignorance but i don't know that, that will help me for my project | 19:21 |
lucazade | np :9 | 19:22 |
lucazade | :) | 19:22 |
Rusty73 | you and bernardo are gods of gma500 | 19:22 |
lucazade | we should thanks intel | 19:24 |
Bernardo | Rusty73: I wish... :) | 19:24 |
=== huntlogger is now known as apachelogger | ||
leed__ | my respect for all guys working on the gma500 problem | 19:28 |
lucazade | Bernardo: i'm not able to compile with the workaround (copy /lib/pkgconfig/libdrm.pc to /usr/lib/pkgconfig) | 19:31 |
awoodland | lucazade: I was just abouts to point out that very problem | 19:32 |
lucazade | sudo cp /usr/lib/pkgconfig/libdrm.pc /lib/pkgconfig/libdrm.pc | 19:34 |
awoodland | the obvious workaround built for me just now, by adding -Llib in libdrm.pc as well as doing that | 19:34 |
Bernardo | strange, for me it built | 19:35 |
Rusty73 | I don't follow all things about your workaround, but if you make only a symbolic link to the correct folder don't solve the problem? | 19:35 |
Bernardo | should work, also copying the file to the right place should work | 19:35 |
Bernardo | lucazade: what error do you get? | 19:36 |
lucazade | no package libdrm found >= 2.3.1 | 19:36 |
awoodland | anything in config.log? | 19:38 |
Bernardo | what does pkgconfig say if you call it from the command line "pkg-config --exists --print-errors libdrm"? | 19:38 |
awoodland | pkgconfig tends to say a lot more in config.log than it does out loud | 19:38 |
lucazade | package libdrm was not found in the pkg-config search path | 19:39 |
Bernardo | ok, so it isn't finding the libdrm.pc file | 19:40 |
lucazade | i've copied it with "sudo cp /usr/lib/pkgconfig/libdrm.pc /lib/pkgconfig/libdrm.pc" | 19:40 |
Bernardo | should have been the other way around | 19:40 |
Bernardo | we're now placing it (erroneously) in /lib/pkgconfig, and it should be in /usr/lib/pkgconfig/ | 19:41 |
Bernardo | check if it is in /usr/lib/pkgconfig/ | 19:41 |
Bernardo | awoodland: what is in your /lib/libdrm.pc? | 19:42 |
lucazade | in /usr/lib/pkgconfig there is a "libdrm.pc.libdrmdev" | 19:42 |
Bernardo | yep, that is the one libdrm-poulsbo-dev redirects | 19:43 |
Bernardo | so it isn't detected by pkgconfig | 19:43 |
awoodland | nothing special in mine | 19:43 |
Bernardo | awoodland: mine already has libdir=/lib | 19:43 |
lucazade | yes pkgconfig seems to not detect it | 19:44 |
awoodland | so does mine actually | 19:44 |
awoodland | I just assumed that libdir was formed usign ${prefix} | 19:44 |
awoodland | and was keen to avoid building things for the 7th time | 19:44 |
Bernardo | well, things seem stable on my pc now | 19:45 |
outlaw45 | mine does compile, but fails with the error | 19:45 |
outlaw45 | dh_install: xpsb-glx missing files (drivers/*), aborting | 19:45 |
outlaw45 | make: *** [binary-install/xpsb-glx] Error 2 | 19:45 |
Bernardo | with the rebuilt xpsb-glx | 19:45 |
awoodland | hmm mine is decididly unstable | 19:46 |
awoodland | i.e. I'm lucky if I get past login | 19:46 |
Bernardo | outlaw45: try doing a "svn revert drivers/*" | 19:46 |
outlaw45 | still empty.. | 19:47 |
awoodland | they seemed to be missing in r23 | 19:48 |
awoodland | so I pulled them from r19 | 19:48 |
Bernardo | ooops... My svn seems to be in a very bad mood indeed. Time to re-upload them | 19:48 |
outlaw45 | ok.. :) | 19:48 |
Bernardo | ok, x crashed again | 19:50 |
Bernardo | and svn is (once again) complaining of a changed special status | 19:51 |
Bernardo | lucazade: can you please add the missing dirs from your original tarball and upload them? Seems like svn on my machine goes crazy when I add them | 19:52 |
lucazade | ok | 19:52 |
outlaw45 | can I download them? | 19:52 |
outlaw45 | bypass the svn thing :) | 19:53 |
Bernardo | download lucazade's tarball from the thread, it's there | 19:54 |
Bernardo | :) | 19:54 |
awoodland | hmm I have a backtrace in a log now | 19:55 |
awoodland | which is more than I had this morning | 19:55 |
awoodland | http://pastebin.org/201625 trying to make it useful now | 19:56 |
Bernardo | we need to edit the debian/rules in xserver-xorg-video-psb to build debug packages also | 19:58 |
outlaw45 | copied content from dri/ and drivers/ folder and building again.. | 20:01 |
awoodland | ccache is your friend :) | 20:03 |
Bernardo | I knew I had forgotten to install something... :) | 20:03 |
awoodland | btw there were a couple of missing build deps I noticed today as well | 20:04 |
awoodland | on libtool | 20:04 |
awoodland | and autoconf | 20:04 |
gang65 | Hi | 20:04 |
awoodland | but I've rather briliantly forgotten which package that was in | 20:05 |
Bernardo | lol | 20:05 |
Bernardo | I know I added those for one of the packages, probably xpsb-glx | 20:05 |
Bernardo | but since svn is acting up on my pc, I'm no longer sure | 20:05 |
awoodland | hmm I can't make it break in gdb | 20:07 |
awoodland | it seems to hang gdb | 20:07 |
awoodland | rather than actually get to the point where a backtrack would be insightful | 20:07 |
awoodland | and I think I'd probably better go home | 20:08 |
lucazade | Bernardo i'm fighting with svn authentication...grrrr | 20:09 |
Bernardo | lucazade: you have to go to the google code site and check the password they generated for you | 20:09 |
outlaw45 | still have anonymous svn access so if this build finishes I checkout again and commit the missing files.. | 20:10 |
gang65 | Hi | 20:23 |
outlaw45 | build xpsb-glx succesfully | 20:24 |
lucazade | committing missing files | 20:24 |
gang65 | I'm xserver-xorg-video-openchrome developer and I would like to start creating PPA | 20:24 |
DRK | YEAH GO ON GUYS, YOUL BE MY HERO OF THE DAY WHEN MY NETBOOK RUNS AGAIN! :D | 20:25 |
lucazade | gang65: interesting... | 20:25 |
gang65 | sorry | 20:26 |
lucazade | Bernardo outlaw45 committed 24 xpsb-glx | 20:26 |
gang65 | I'm openchrome developer | 20:26 |
Bernardo | ok, will pull it | 20:26 |
gang65 | I would like to add patch to xserver-xorg-video-openchrome package | 20:29 |
gang65 | to my PPA | 20:29 |
outlaw45 | thx, but already had it and compiled succesfully :) but also screwed up svn and was unable to commit.. | 20:29 |
outlaw45 | xserver-xorg-video-psb doesnt compile because psb_drm.h is missing | 20:41 |
fkefer | is the issue with the .h file in two packages fixed already? | 20:42 |
fkefer | i | 20:42 |
outlaw45 | can't find it in the sources package from lucazade either | 20:42 |
fkefer | it wasn't that one though, was it? | 20:42 |
outlaw45 | checking.. | 20:42 |
lucazade | i gotta go... i'll check when i'm back guys! | 20:43 |
lucazade | outlaw i was able to compile xserver-xorg-video-psb | 20:44 |
outlaw45 | @lucazade: cya :) | 20:44 |
lucazade | haven't seen that error | 20:44 |
outlaw45 | did you install psb-kernel-headers package? | 20:44 |
lucazade | no headers no | 20:45 |
outlaw45 | psb_drm.h is there.. | 20:45 |
lucazade | because they provide some files like the ones in libdrm-dev | 20:45 |
lucazade | i've never installed psb-kernel-headers in karmic | 20:46 |
outlaw45 | weird.. | 20:46 |
lucazade | see u | 20:46 |
outlaw45 | cya | 20:46 |
Bernardo | outlaw45: the missing headers are in libdrm-poulsbo-dev | 20:49 |
Bernardo | it should be a dependency to xpsb-glx | 20:50 |
outlaw45 | have got libdrm-poulsbo-dev installed and still can't compile the xorg driver | 20:52 |
outlaw45 | psb_drm.h does exist in /usr/include/drm | 20:53 |
mfonville | ah, so more people here fighting with the new psb progress? :) | 20:54 |
outlaw45 | the file doesn't exist in /usr/local/include/drm | 20:54 |
outlaw45 | gonna symlink it there and try again.. | 20:54 |
mfonville | @outlaw45 to solve the not found libdrm when compiling xpsb? | 20:54 |
Bernardo | it should never be in /usr/local/ | 20:55 |
Bernardo | mfonville: the librm.pc file is being installed to the wrong place (lib/pkgconfig), you need to copy it to /usr/lib/pkgconfig to build xpsb | 20:55 |
Bernardo | outlaw45: which package is looking for includes under /usr/local/ | 20:56 |
Bernardo | ? | 20:56 |
outlaw45 | I'm compiling xserver-xorg-video-psb at the moment | 20:56 |
outlaw45 | I believe the symlinks work.. just passed the file causing problems | 20:57 |
outlaw45 | maybe it should be configured with --prefix=/usr | 20:57 |
Bernardo | isn't it? | 20:57 |
mfonville | @outlaw, ok thanks. I will tweak the pkgconfig then now | 20:57 |
Bernardo | aren't you using the debian/rules file? | 20:58 |
outlaw45 | make is run with '-I/usr/local/include' | 20:58 |
Bernardo | shouldn't be | 20:58 |
Bernardo | and isn't on my machine | 20:59 |
outlaw45 | rules file does at the --prefix paramter | 20:59 |
outlaw45 | CFLAGS="$(CFLAGS)" ./configure --host=$(DEB_HOST_GNU_TYPE) --build=$(DEB_BUILD_GNU_TYPE) --prefix=/usr --mandir=\$${prefix}/share/man --infodir=\$${prefix}/share/info --enable-static=yes | 20:59 |
Bernardo | I know, I just checked | 20:59 |
Bernardo | how are you building it, using debuild? | 21:00 |
outlaw45 | yeah | 21:00 |
outlaw45 | debud -i -uc -us -b | 21:00 |
Bernardo | I used "debuild -i -us -uc -b" which should be the same, and it didn't look into /usr/local | 21:01 |
Bernardo | your autoconf setup is possibly with some problem | 21:02 |
outlaw45 | purging autoconf and trying again.. | 21:04 |
pazo | So... anyone got poulsbo working? | 21:06 |
outlaw45 | ok.. psb-kernel-headers screwed up everything... I'll try a reboot | 21:06 |
pazo | I see. | 21:06 |
DRK | pazo: they are still fighting | 21:06 |
pazo | :D | 21:06 |
mfonville | @pazo still busy with it | 21:06 |
mfonville | grmbl, compiling on a netbook takes ages :P | 21:07 |
Bernardo | pazo: I have it "working". I can boot into X, login, but after sometime it crashes | 21:12 |
Bernardo | awoodland seems to be at the same stage | 21:12 |
awoodland_ | I'm somewhat worried that this is another symptom of something bad in a part we don't have source for | 21:13 |
pazo | Bernardo, hmm... sounds like a hard problem to debug | 21:14 |
outlaw45 | ok.. fried my system.. | 21:14 |
awoodland_ | because there were a number of corruption bugs before anyway | 21:15 |
outlaw45 | gonna reinstall and try again :p | 21:15 |
awoodland_ | fried as in permanently? | 21:15 |
outlaw45 | no.. psb-kernel-headers screwed up apt | 21:15 |
outlaw45 | kan't purge it.. and that causes that I can't purge anything.. | 21:16 |
outlaw45 | @Bernardo: could you upload your packages? that would save a lot of time getting up to speed | 21:16 |
awoodland_ | is it prerm that's broken? | 21:16 |
outlaw45 | ? | 21:16 |
awoodland_ | with the purging | 21:17 |
awoodland_ | i.e. how does purge fail? | 21:17 |
awoodland_ | (lvm2 root fs+snapshot is pretty handy though) | 21:17 |
outlaw45 | @awoodland: if I purge any package it want's to purge psb-kernel-headers to and that fails | 21:18 |
Bernardo | outlaw45: ok, I'll add them to the repo | 21:19 |
outlaw45 | @Bernardo: thx | 21:19 |
Bernardo | doh | 21:20 |
mfonville | @outlaw45 | 21:21 |
Bernardo | svn failed (again) as it claims I have some unexpected special status under xpsb-glx. | 21:21 |
mfonville | about puring the psb-kernel-headers | 21:21 |
mfonville | *purging | 21:21 |
mfonville | you have to manually remove a directory that itself can not delete during the diversion changes | 21:21 |
mfonville | if you do that, it wil purge succesfully (at least, here it always did) | 21:21 |
mfonville | rmdir /usr/include/drm-linux-libc | 21:23 |
outlaw45 | @mfonville: the file causing the problem is provided by another package which I can't remove through the previous error | 21:23 |
awoodland_ | just delete it/move it manually? | 21:23 |
mfonville | yes | 21:23 |
outlaw45 | removing it manually breaks the other package.. | 21:23 |
mfonville | first start the purge | 21:23 |
awoodland_ | put it back again later? | 21:24 |
mfonville | then the purge will fail halfway | 21:24 |
mfonville | then remove that directory | 21:24 |
mfonville | then retry the purge | 21:24 |
mfonville | and it will succeed | 21:24 |
mfonville | you dont have to put it back later | 21:24 |
outlaw45 | I'll try always faster then reinstalling :p | 21:24 |
mfonville | it is a directory only for this poulsbo mess ;) | 21:24 |
mfonville | (and it should be empty at the point where you delte it) | 21:25 |
Bernardo | outlaw45: uploading now. I'm going, so keep checking, it should be there after a while | 21:26 |
outlaw45 | ok.. thx | 21:27 |
outlaw45 | I'll post the outcome on the forum.. | 21:27 |
outlaw45 | ok a lot more errors now.. so just gonna reinstall.. gives me a clean system to start with... | 21:27 |
mfonville | where is that repo of @Bernardo? | 21:27 |
awoodland_ | http://code.google.com/p/gma500/source | 21:28 |
mfonville | ah ok | 21:28 |
mfonville | that one :) | 21:29 |
awoodland_ | it's hidden somewhere in that massive thread | 21:29 |
mfonville | i thought maybe a binary repo ;) | 21:29 |
awoodland_ | unless there's some extra repository I don't know about | 21:29 |
mfonville | i just compiled xpsb btw | 21:29 |
mfonville | during install i get this warning, but not sure whether it will be critical: | 21:29 |
mfonville | dpkg-divert: rename involves overwriting `/usr/lib/libGL.so.1' with | 21:29 |
mfonville | different file `/usr/lib/psb/libGL.so.1.xlibmesa', not allowed | 21:29 |
mfonville | dpkg-divert: rename involves overwriting `/usr/lib/libGL.so.1.2' with | 21:29 |
mfonville | different file `/usr/lib/psb/libGL.so.1.2.xlibmesa', not allowed | 21:29 |
awoodland_ | that won't break anything early on | 21:29 |
awoodland_ | might be a problem when it comes to running opengl apps | 21:30 |
awoodland_ | but if you get that far it's working better than mine | 21:30 |
mfonville | k | 21:30 |
awoodland_ | and trivially hackable | 21:30 |
mfonville | true | 21:30 |
awoodland_ | mine reliably crashes just after login and the sound | 21:30 |
awoodland_ | but sometimes it breaks earlier | 21:31 |
awoodland_ | and so far I've not managed to get anything interesting from a backtrace | 21:31 |
outlaw45 | @awoodland: did you compile form the repository? | 21:31 |
awoodland_ | yes mostly | 21:31 |
awoodland_ | I had to work around a few problems with the version I checked out | 21:31 |
awoodland_ | but no new patching, only packaging things | 21:32 |
Bernardo | awoodland: we need to build debug packages for everything | 21:32 |
Bernardo | well, bye guys, see you tomorrow | 21:32 |
awoodland_ | ok | 21:32 |
Bernardo | the binaries just finished uploading, btw | 21:32 |
Bernardo | don't blame me if your pc burns when you install them | 21:32 |
=== Bernardo is now known as Bernardo|away | ||
mfonville | so there is a binary repo somewhere :P | 21:33 |
outlaw45 | @Bernardo: I'll hunt you down if it happens :D | 21:33 |
outlaw45 | he put the binary packages in the svn repo.. | 21:33 |
mfonville | ah | 21:34 |
mfonville | that is kinda evil :P | 21:34 |
mfonville | so then all the people who want the source will also get binaries :P | 21:34 |
outlaw45 | it is... but at least you could get up to speed fast | 21:35 |
mfonville | lol, he also uploaded skype :P | 21:36 |
DRK | psb*skype=build in remote desktop? ;) | 21:39 |
mfonville | ok, installed my psb mess | 21:43 |
mfonville | now time to activate it in the xorg and to reboot | 21:43 |
awoodland_ | you have a 2nd machine for remote debugging? | 21:44 |
mfonville | if necessary, yes. but my gf is using it atm :P | 21:44 |
mfonville | first i will have to see how good/bad the whole situation is ;) | 21:45 |
mfonville | ok, also my psb here crashed after the login | 21:53 |
mfonville | i did already see some windows and stuff | 21:54 |
mfonville | some things drawn 'bad' | 21:54 |
awoodland_ | mine was drawing the login incorrectly | 21:54 |
awoodland_ | if I started X manually I could run an xterm | 21:54 |
mfonville | but i think the autostart of some application (skype?) can trigger the bad opengl stuff | 21:54 |
awoodland_ | but making it scroll was enough to kill it for me | 21:54 |
awoodland_ | I can't do much more until I'm back at work tomorrow | 21:55 |
coz_ | well guys sorry to say i had to go back to karmic ...cannot enable fast writes and SBA on lucid plus a few more issues.... and cannot install official nvidia driver so I guess I must wait :) | 21:55 |
mfonville | and i will be hitchhiking tomorrow | 21:55 |
mfonville | so i also wont be able to work on it :P | 21:55 |
awoodland_ | take two laptops with you? :) | 22:00 |
mfonville | hehe :) | 22:01 |
mfonville | i have another laptop at my destination, so there i can continue the work | 22:01 |
mfonville | but i will have two nights in Berlin in between, where i wont spend the time debugging ;) | 22:01 |
bjsnider | coz_, i wonder what it is you're waiting for | 22:02 |
coz_ | bjsnider, waiting for well wallpaper rendering is buggy at the moment along with sba and fastwrites not being enabled...also when external drive with mulitple partitions...selecting all has no unmount availab;e in right click options and umounting one partitions of course gives errors ...all of which I have already reported ...<< is that what you meant ? :) | 22:04 |
Sarvatt | coz_: "options nvidia NVreg_EnableAGPSBA=1 NVreg_EnableAGPFW=1" | sudo tee /etc/modprobe.d/nvidia.conf | 22:06 |
coz_ | Sarvatt, didnt work on both agp machines here | 22:06 |
coz_ | and cannot install official nvidia driver which is not such a big deal but a little kink for me | 22:06 |
Sarvatt | then you're going to have to force it in the source package like you had to before because your card doesn't think it works with it | 22:07 |
Sarvatt | it might be "options nvidia-current NVreg_EnableAGPSBA=1 NVreg_EnableAGPFW=1" | sudo tee /etc/modprobe.d/nvidia.conf | 22:07 |
Sarvatt | actually :) | 22:07 |
AlanBell | hi all | 22:08 |
Sarvatt | not sure if its aliased to nvidia-current at the point it loads that | 22:08 |
bjsnider | or maybe nvidia_current | 22:08 |
coz_ | Sarvatt, yea I also tried that with no luck but combined with the other issues I can't suggest to any client to upgrade | 22:08 |
coz_ | bjsnider, I have tried both actually | 22:08 |
Sarvatt | or whatever driver you're using, guessing nvidia-96 instead if you care about SBA and fastwrites (that dont really do anything anyway) | 22:08 |
coz_ | Sarvatt, no the nvidia-current | 22:09 |
bjsnider | what card? | 22:09 |
AlanBell | I have a new laptop with a 1366x768 internal screen. It doesn't pick up 1024x768 as a valid mode, here is the Xorg.0.log http://pastebin.com/kWfbGcuG | 22:09 |
coz_ | be back have to restart x | 22:09 |
Sarvatt | AlanBell: tried adding the modeline manually? | 22:10 |
AlanBell | I can use xrandr to add a 1024x768 mode and it works fine. This is useful because when cloning the screen onto a projector on the VGA port it picks up 800x600 if it doesn't think the internal screen can do 1024x768 | 22:10 |
AlanBell | it would be great to have it working a bit better out of the box with projectors, this is on Lucid. | 22:11 |
bjsnider | well, lucid consumes almost twice as much power as windows 7 | 22:24 |
furan | heh | 22:25 |
bjsnider | on nvidia graphics anyway | 22:28 |
bjsnider | i think that's the issue. i don't really think their kernel is better at power consumption than linux | 22:29 |
* aarcane signs up to watch the poulsbo progress | 23:23 | |
lucazade | Bernardo: installed your debs (mainly because of xpbs-glx) on my acer751h, X starts but crash at gdm and/or at desktop.. will continue check it | 23:50 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!