=== jono is now known as Guest52708 | ||
mlankhorst | marvin24: ah.. I forgot to check if drivers used the platform id | 06:59 |
---|---|---|
mlankhorst | I thought busid was only used in the xserver | 06:59 |
mlankhorst | I'll determine platform id from sysfs too, should be safe | 07:01 |
marvin24 | mlankhorst: yes, opentegra also used busid from xf86_get_platform_device_attrib | 11:17 |
marvin24 | but I guess it won't be NULL, because it's not a pci device | 11:18 |
marvin24 | mlankhorst: notify me if you have a fix I can test | 11:18 |
mdeslaur | tjaalton: I think precise is going to need a nvidia-settings update too, does that make sense? | 12:50 |
mdeslaur | whoops tab fail | 12:52 |
mdeslaur | tjaalton: sorry, not you | 12:52 |
mdeslaur | tseliot: I think precise is going to need a nvidia-settings update too, does that make sense? | 12:52 |
tjaalton | :) | 12:55 |
tseliot | mdeslaur: ok, I can do that, is this only for nvidia-current-updates? | 13:45 |
mdeslaur | tseliot: I only tested nvidia-current in precise so far, and the nvidia-settings that goes with it doesn't allow me to change resolutions with the 304 driver | 13:45 |
mdeslaur | tseliot: precise's nvidia-settings-updates has a more recent 304.43, so it may work there | 13:46 |
tseliot | mdeslaur: ok, but are you going to upload both nvidia-current and nvidia-current-updates in precise? | 13:46 |
mdeslaur | tseliot: but, would it make sense to push 304.88 of settings everywhere? | 13:47 |
mdeslaur | tseliot: yes, that was my intention | 13:47 |
mdeslaur | tseliot: that's what you prepared for me, no? | 13:47 |
tseliot | mdeslaur: yes, it is entirely your choice though whether to update nvidia-current in Precise or not | 13:48 |
tseliot | mdeslaur: and yes, it's best if nvidia-settings matches the driver version it's supposed to be used with | 13:48 |
mdeslaur | tseliot: I don't think we have much of a choice...do you know if there are any issues in 304.x that 295.x doesn't have? | 13:49 |
mdeslaur | I'm sure we're going to break _some_ people | 13:49 |
tseliot | mdeslaur: I'm not sure but 304 is the only supported legacy branch, whereas the 295 branch was short-lived and is not supported any more | 13:51 |
mdeslaur | It's a damned if we do, damned if we don't scenario | 13:51 |
mdeslaur | does anyone else in here have an opinion on pushing 304.88 to precise as a security update? | 13:52 |
tseliot | mdeslaur: right. It still supports GeForce 6 series (which is pretty old) | 13:52 |
tseliot | mdeslaur: here's the list of supported cards for 295.40: http://it.download.nvidia.com/XFree86/Linux-x86/295.40/README/supportedchips.html | 13:54 |
tjaalton | mdeslaur: +1 | 13:54 |
tseliot | and here's the list for 304.88: http://it.download.nvidia.com/XFree86/Linux-x86/304.88/README/supportedchips.html | 13:55 |
mdeslaur | if we push 304.88, do we need to get stuff re-tested on precise? like steam or such? or is everything pretty much better with 304.88? | 13:55 |
tseliot | mdeslaur: well, with Steam, you might really wanna use the experimental driver | 13:55 |
tseliot | mdeslaur: let me package nvidia-settings for you... | 14:01 |
mdeslaur | tseliot: awesome, thanks! | 14:01 |
bjsnider | mdeslaur, i think 304.88 should be pushed | 14:05 |
bjsnider | in fact i didn't know there would be any question of it | 14:05 |
mdeslaur | ok, I feel a lot better being able to blame #ubuntu-x when I get my usual security team death threats email :) | 14:06 |
bjsnider | blame nvidia | 14:06 |
bjsnider | should be pushed further back than precise too | 14:06 |
mdeslaur | oh! yeah, I keep forgetting I can just blame nvidia :) | 14:07 |
bjsnider | i'm just trying to get that done so i don't have to put it in the ppa | 14:07 |
tseliot | :) | 14:29 |
mlankhorst | marvin24: fixed, I think.. | 15:31 |
mlankhorst | attached to the bug report | 15:31 |
mlankhorst | oops | 15:34 |
marvin24 | mlankhorst: compiling ... | 16:08 |
marvin24 | mlankhorst: no change | 17:12 |
marvin24 | but removing your patch alltogther works :-( | 17:26 |
mlankhorst | ugh | 17:42 |
mlankhorst | whats the sysfs path? | 17:42 |
mdeslaur | tseliot: so, the issue I was seeing with nvidia-settings on precise is still there, even after upgrading to nvidia-settings 304.88...I can't change screen resolution on my laptop anymore | 17:47 |
mdeslaur | tseliot: all I have is "Auto" and "1680x1050" in the dropdown instead of the list of resolutions that was available with the 295 driver | 17:48 |
mdeslaur | tseliot: any ideas what it could be? | 17:48 |
tseliot | mdeslaur: I wouldn't know. Some failure to read the EDID? I'm not sure | 17:49 |
Sarvatt | mdeslaur: it uses system settings>display now instead of nvidia-settings | 17:49 |
mlankhorst | marvin24: are you sure it should be platform:omapdrm:0 ? I figured it would be platform:omapdrm:00 | 17:49 |
mlankhorst | oh duh! | 17:51 |
mdeslaur | Sarvatt: system settings->display only has 1680x1050 in the dropdown also | 17:51 |
mlankhorst | marvin24: remove the \n on the asprintf line | 17:51 |
mlankhorst | that should make the patch work | 17:51 |
marvin24 | mlankhorst: I'm on tegra ... | 17:53 |
marvin24 | will check | 17:53 |
mlankhorst | that's fine, format should be the same | 17:53 |
mlankhorst | I only tested on x86 by copying that logic into a separate c program anyway, and hardcoded a test string, didn't notice the newline there :P | 17:54 |
Sarvatt | is there any kind of scaling options anywhere in nvidia-settings that might make it show more resolutions? | 17:54 |
mlankhorst | like scaling on a lcd is a good idea anyway.... | 17:54 |
* Sarvatt looks over the readme | 17:54 | |
marvin24 | mlankhorst: nope | 18:03 |
marvin24 | on tegra, there is no platorm device related to graphics I guess | 18:06 |
marvin24 | why can't we just check if busid is NULL? | 18:07 |
mlankhorst | marvin24: $ udevadm info --query=all --name=/dev/dri/card0 ? | 18:09 |
mlankhorst | on the tegra | 18:22 |
marvin24 | mlankhorst: /devices/host1x/drm/card0 | 18:43 |
mlankhorst | oh right, I forgot that tegra reinvents the wheel | 18:43 |
marvin24 | yes, that may change in the future maybe | 18:44 |
marvin24 | should be platfrom/host1x ... | 18:44 |
marvin24 | there was a long discussion about this | 18:44 |
mlankhorst | still wouldn't help | 18:44 |
mdeslaur | ah, seems nvidia 302.xx introduced this change: "Removed Flat Panel Scaling configurability in nvidia-settings."...so I guess it's normal when going from 295 to 304 | 18:44 |
mlankhorst | where's the tegra git tree? | 18:45 |
marvin24 | mlankhorst: mainlined | 18:48 |
mlankhorst | i mean for xf86-video-*tegra | 18:48 |
marvin24 | ah | 18:48 |
mlankhorst | oh they're just using busid, sigh | 18:48 |
mlankhorst | marvin24: what's the busid on tegra? | 18:49 |
marvin24 | git://gitorious.org/thierryreding/xf86-video-opentegra.git | 18:49 |
marvin24 | mmh, no idea | 18:49 |
mlankhorst | blergh I guess I'll fall back if parsing fails | 18:50 |
* mlankhorst grumbls at busid being used to open device, rather than raw fd | 18:53 | |
marvin24 | mlankhorst: busid:platform:host1x:00 | 18:56 |
mlankhorst | exactly what I expected it to be, I guess I'll special case host1x after adding a fallback probe :P | 18:58 |
mlankhorst | marvin24: but yeah, it should be /devices/platform/host1x.0 I suppose.. | 19:03 |
marvin24 | mlankhorst: you can ask tagr on #tegra why it isn't like this | 19:04 |
mlankhorst | sigh | 19:08 |
mlankhorst | drmfreebusid takes a const char * | 19:08 |
mlankhorst | marvin24: http://paste.ubuntu.com/5680811/ last attempt for now | 19:41 |
marvin24 | mlankhorst: yup - this worked | 20:36 |
marvin24 | thanks! | 20:37 |
=== ubott2 is now known as ubottu |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!