intrepid can't start with via vn896 vga

Please attach the output of 'lspci -vvnn'

Hy there!

now I tested it again with the "current" iso what is compiled ad sept. 23.
Here is the output of lspci and the version of the openchrome package!

Tell me if you need any more information or testing!

Thanks!

Could you please retry with upstream's latest svn? (latest revision as of the time of writing is 599).
And please post your /var/log/Xorg.0.log too

Hy!

I sucked with this shit since today morning, and now I have an Xorg.0.log what I got with the current svn/trunk from openchrome.org... It's a bit better than the included one, but it still unusable... the cursor looks like an 50x50 pixels memory dump, and everything looks scambled ecept the background image, and to make me absolutley happy when the gnome panels appear (but scambled) my cursos stops and I can't get any response from my computer anymore...

I got the log from the single user mode... asked ubuntu to drop me into a root terminal, downloaded, compiled the driver and fired xinit instead of let it continue booting...

It isn't freezes, the default background pattern and the xterm window (what I can't use because the keyboard and mouse configuration runs in the boot process after the single terminal) but the mouse is a memory dump again...

One word to another, this way my computer keeps responsive, I can move the mouse if I configure it in the xorg.conf manually, I went back into my terminal, and saved the Xorg.log, so here you ar! :)

And finally it has a little piece of positive result... it detects my lcd panel size correctly, what earlier versions can't do :)

Thanks for your help!
u-foka

Could you please apply the following patch to upstream's code and see if it fixes the mouse cursor problem?
After you build and install the driver from svn, could you please let the boot process finish so hal & friends are loaded to test the whole thing?

I'm anyway forwarding this bug to upstream.

Hy!

I discovered funny things :)

1. Your patch fixes the cursor when I start X with xinit from the single terminal
2. When I let it to continue booting, it chrashes the same way as the original live cd without any modifications, except that I can see the real cursor below the random froozen white pixels...
         (I dont known but if I'm right, it was better with the svn openchrome without the patch, because that way it frozes about 10 second latter :D:D LOL)
3. Just because my curiosity, i tryed to compile and install the current svn of openchrome on my hardy, and I found it doesn't work at all, after I fired xinit, I can't see anything on my screen :S It's interesting because before hardy, I always compiled that driver to get my card alive, and on hardy, I still can compile and use an old svn version of openchrome, youst I don't use it because I get the same result with the included openchrome driver
4. I tryed to install the old svn driver on intrepid to see if it works, but it doesn't :S I can't see anything on my screen when I try to use it... but it compiles too without any error... :S

So, I'm out of ideas :S
I tryed these things in any "may be work" combinations...

Thanks!
u-foka

P.S. Here is my Xorg.0.log with the patched driver (if it has any differences)

Ok, could you please provide the output of 'ls /etc/rc2.d' and 'apt-cache policy xserver-xorg libdrm2'?

And could you also please try by adding the following line to the Device section in /etc/x11/xorg.conf?

Option "XaaNoImageWriteRect"

Option "XaaNoImageWriteRect" does indeed allow X to run without lockup when using the OpenChrome driver with this chipset. Although, the desktop seems no faster than the VESA driver. Attaching the output of the commands

----- Original Message ----
From: Raphael Geissert <email address hidden>
To: <email address hidden>
Sent: Thursday, October 23, 2008 3:21:44 PM
Subject: [Bug 273775] Re: intrepid can't start with via vn896 vga

Ok, could you please provide the output of 'ls /etc/rc2.d' and 'apt-
cache policy xserver-xorg libdrm2'?

And could you also please try by adding the following line to the Device
section in /etc/x11/xorg.conf?

Option "XaaNoImageWriteRect"

--
intrepid can't start with via vn896 vga
https://bugs.launchpad.net/bugs/273775
You received this bug notification because you are a direct subscriber
of the bug.

From the attachment:
> xserver-xorg:
> Installed: 1:7.4~5ubuntu2

That seems to be the source of the problem (will have to wait until
u-foka confirms he is also using 7.4): there are some issues (aka
bugs) in xorg 7.4 with XAA. The other bug reported against -openchrome
related to gdm has more information about the problem (and even more
at openchrome's trac).

Thanks for the assistance and for taking a look. Always appreciate the hard work of everyone!

Hy!

I have tested it, booted the last livecd added only that xaa option, and voila :) it works like in hardy... detects my panel as 1600x1200, but I can set it to 1280x800 and that's it :)

Many thanks for your help!

I've tested the packages from https://launchpad.net/~atomo64/+archive, but the pointer problem is there too. X starts only with XaaNoImageWriteRect and proper mode set in xorg.conf. It's a VIA P4M900 (Chrome 9 HC).

Rafał, could you please tell me what version of xserver-xorg-core you have?

'apt-cache policy xserver-xorg-core' will tell you. If you are still
using a version below 2:1.5.2-1 update it, try again, and report back
please.

I have xserver-xorg-core 2:1.5.2-2ubuntu3 (and I had it before I reported, only hal was upgraded today)

Just to make sure, I made the following: moved all xorg.conf and xorg log files, dpkg-reconfigured xserver-xorg-core and rebooted*, introducing changes one at a time (all the time using your packages with openchrome 599 revision):

1. Without xorg.conf: screen blurred, broken cursor. System hangs after login (before gnome is fully loaded).

2. Xorg.conf with XaaNoImageWriteRect only: normal quality, broken cursor.

3. Xorg.conf with XaaNoImageWriteRect and SWCursor (haven't tried that one before): normal quality, cursor working.

* I was trying with restarting X at first, but each time I log out and restarted X (ctrl+alt+backspace) - no matter the setting (1, 2 or 3) - resolution breaks. I get 1024x576 instead of 1024x600 and wrong viewport size (only 2/3 screen visible). Proper resolution works for the first login only. This is the reason I tried with some modes set last time - I logged out with XaaNoImageWriteRect to see proper quality but wrong resolution...

I have problems with this package too. I've added the Xaa option to the xorg.conf I was using with Hardy and it still locks when going to open the GDM. In order to boot properly I have to choose the recovery mode on GRUB, then when it gives me the list of options (shell as root, fix Xserv, continue, etc) I choose to continue the boot normally and the graphic server comes fine.
Any solution?
I have:
xserver-xorg-core 2:1.5.2-2ubuntu3
xserver-xorg 1:7.4~5ubuntu3

Hy there!

I had this problem before in hardy, if I booted with vga=952 option what gives me an 1280x800 framebuffer with 24 bit color... and as soon I boot with vga=951, what is the same except that using 16 bit color in framebuffer, then my X works fine, so I think the problem was the 24 bit color mode in the framebuffer.

Thanks... it could be something related to that. Actually I had a problem with the splash screen I was using. After the update I had lost it, when booting a had something like a default splash to development, with some patterns for testing, I guess. I changed it to the default splash in a different resolution and all became nice and smooth again ^^

Woops, VIA has just released a beta version of their driver to Intrepid.
http://linux.via.com.tw/support/downloadFiles.action

I was using this driver with Hardy and it was working preety well. I got a few problems with Compiz (there is a 3d driver for Hardy, but not for Intrepid yet).
Can't wait to get home and test it, hehe

2008/10/26 Rafał Próchniak <email address hidden>:
[...]
>
> 2. Xorg.conf with XaaNoImageWriteRect only: normal quality, broken
> cursor.
>
> 3. Xorg.conf with XaaNoImageWriteRect and SWCursor (haven't tried that
> one before): normal quality, cursor working.

Could you please try with the new package from my PPA?
It includes some register changes that might fix the hardware cursor
problem (i.e. SWCursor disabled).

If that still doesn't fix the cursor problem I'll write a patch to
disable ARGB hardware cursor for the VN896

Thanks for your time.

The package in Mr. Geissert's PPA partially fixes the cursor problem. There is some artifacts (several pixes below the cursor) that are visible, as well as the cursor being partly cut off in WINE windows, when option "SWcursor" "true" is commented out in xorg.conf. I'm not sure what problems others are experiencing with the standard intrepid OpenChrome driver, but on my system, without the SWCursor option enabled, the only problem I have is the cursor is completely invisible when in WINE windows. Mr. Geissert's package at least makes the cursor visible in these windows on my system, so that it is usable, even if it is partly invisible as well.

Hy there!

This problem with the hwcursor is very mature in the openchrome driver... I ever used my card with swcursor because I had problems with hw... How I can see, the real problem is how the driver can load only some number of cursors. So for ex I haven't got the HAND cursor in the browser because I think that exceeds that limit... I can't change my cursor theme, because gdm/kdm still loads the default theme for itself, and that limit is exceeded by them.. And I think this is because wine cant load cursors too...

I hope I can help with this!

I've tested the newer packages and both XaaNoImageWriteRect and SWcursor options are still necessary to run openchrome on Intrepid.

New documentation is out from Via for the hardware, I hope when updates to the driver come out based on the new information, Ubuntu acts quickly to put them in the repositories. Unfortunately, at this point there is no telling when that might be.

I won't repeat myself over and over again: the issue has its origins
in Xorg itself in version 7.5

I don't know much about X.org and EXA and don't have time to learn
about them. I'll prod Xavier to see if he can suggest some other
change in X.org.

An alternative is to downgrade to X.org 7.4 and build the latest
openchrome drivers against that version. IOW: downgrade to the X.org
version from hardy and wait for me to upload the latest openchrome
package to my ppa for hardy (or compile it yourself, as you wish, it
is not rocket science).

Providing more config or log files is useless for this issue (unless
it does happen with X.org < 7.5).

If somebody knows how to clone this report and reassign it to the xorg
packages please do it, as I've found launchpad very counter-intuitive
and have no idea (nor found by looking at the web frontend) how to do
it at all.

This bug should now be fixed as of svn Rev 696 :
http://www.openchrome.org/trac/ticket/263

The build instructions is available at:
https://help.ubuntu.com/community/OpenChrome

After build/install/run openchrome driver rev696, my VIA K8M890 graphic not freezes.

Unfortunately the problem with mouse pointer still exists (the mouse cursor is not displayed).

Only after add to xorg.conf file option:
    Option "SWcursor"
my graphic card work properly.

May I please someone to create xserver-xorg-video-openchrome_0.2.903+svn696 package?

The X freeze and cursor problem was fixed in openchrome Revision 707.
On Ubuntu Intrepid and chipset VIA K8M890 its works perfeclty !

Build instruction I described at:
https://help.ubuntu.com/community/OpenChrome