Bug 57527

Can you please provide the X log ?

Created attachment 70624 [details]
Xorg.log when started with ModeSwitch="legacy"

This is the Xorg.log when X is started with Option "ModeSwitchMethod" "legacy".  I am pretty sure the display resolution is in fact 1280 x 800.  There is a strange output line containing to modes, maybe this is a bug?  After this, the screen is black and X seems to be crashed.  There is no reaction to Ctrl-Alt-F1, but i can login via ssh and shutdown the notebook with the power button.

Created attachment 70626 [details]
Xorg.log with some features switched off, cf. subsequent xorg.conf

Created attachment 70628 [details]
xorg.conf reproducing red and black stripes

When looking closely, there are a few lines with red switched off completely, except for a few stray pixels, then a horizontal stripe (~ 200 px high) where red is switched on, then a dark stripe with red switched off, then again a red and then a black bar.  I could tage a photograph, if that helps.

The kernel in use is 3.5.7, the machine is started with acpi=noirq (otherwise the graphics had serious lags with the previous Ubuntu release), the stripes occur whether or not via_agp is blacklisted (and thus not loaded).

Try to add "text" option to linux kernel. This line could be found during boot, by pressing "e" key, in grub.

Please provide also photo (by external camera) of your monitor screen.

Try to provide /var/log/Xorg.0.log without xorg.conf configuration file (just move or delete it).

In Bug #53307 i saw some other information provided that i hope to be helpful:

root@amilo:/home/thomas# lspci -v -s 1:0
01:00.0 VGA compatible controller: VIA Technologies, Inc. CN700/P4M800 Pro/P4M800 CE/VN800 Graphics [S3 UniChrome Pro] (rev 01) (prog-if 00 [VGA controller])
	Subsystem: Fujitsu Technology Solutions Device 10cd
	Flags: bus master, 66MHz, medium devsel, latency 64, IRQ 10
	Memory at f0000000 (32-bit, prefetchable) [size=64M]
	Memory at d1000000 (32-bit, non-prefetchable) [size=16M]
	Expansion ROM at <unassigned> [disabled]
	Capabilities: [60] Power Management version 2
	Capabilities: [70] AGP version 2.0

root@amilo:~$ dmesg | grep -iE '(agp|drm)'
Linux agpgart interface v0.103
agpgart: Detected VIA VT3314 chipset
agpgart-via 0000:00:00.0: AGP aperture is 256M @ 0xe0000000
[drm] Initialized drm 1.1.0 20060810
[drm] Supports vblank timestamp caching Rev 1 (10.10.2010).
[drm] No driver support for vblank timestamp query.
[drm] Initialized via 2.11.1 20070202 for 0000:01:00.0 on minor 0
agpgart-via 0000:00:00.0: AGP 3.5 bridge
agpgart-via 0000:00:00.0: bridge is in legacy mode, falling back to 2.x
agpgart-via 0000:00:00.0: putting AGP V2 device into 4x mode
pci 0000:01:00.0: putting AGP V2 device into 4x mode

And this:

root@amilo:~$ grep '(EE)' /var/log/Xorg.0.log
	(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
[     9.999] (EE) AIGLX error: dlopen of /usr/lib/i386-linux-gnu/dri/unichrome_dri.so failed (/usr/lib/i386-linux-gnu/dri/unichrome_dri.so: cannot open shared object file: No such file or directory)
[     9.999] (EE) AIGLX: reverting to software rendering
root@amilo:~$ locate unichrome
/usr/lib/i386-linux-gnu/directfb-1.2-9/gfxdrivers/libdirectfb_unichrome.so

Should i use unichrome_dri?  (This would probably affect performance only, but not mere working or not, i guess.)  It doesn't seem to be offered by the ubuntu repository, anyway.

The "text" parameter leaves me with a text console after boot.  I have started X with startx, with no difference, however.  Starting X without xorg.conf shows no difference, either -- except for the mouse cursor being accompanied by a red square.

Created attachment 70762 [details]
Login screen

Created attachment 70763 [details]
Desktop

Created attachment 70764 [details]
Xorg.log when started without xorg.conf

BTW: i switched off VIA framebuffer in the kernel, should i enable it?
I switched off the whole of plymouth (the sugery boot screen hiding away anything important during bootup), at least i hope so.
I switched off the console font change, too.

Created attachment 70835 [details] [review]
Fix line compare register programming.

This bug looks a lot like 53307. We seen a similar bug the DRM/KMS driver for the P4M800 chipset. Can you try the attached patch to see if that fixes your problem.

Yes please turn off viafb.

Thanks a lot for your support.  Alas, as it seems, the patch did not have any effect at all.
This is what i did (i am not familiar with package management on debian-like platforms):

Install required build tools and lib headers:
# apt-get build-dep xserver-xorg-video-openchrome

Get and unpack the ubuntu source of the driver:
$ apt-get source xserver-xorg-video-openchrome

Patch it:
$ patch -p1 < (your patch, succeeded)

Build a deb package:
$ dpkg-buildpackage -us -uc -nc

Install it:
# dpkg -i ./xserver-xorg-video-openchrome_0.3.1-0ubuntu1_i386.deb 

(The md5sum of /usr/lib/xorg/modules/drivers/openchrome_drv.so has changed afterwards.)

Reboot (just in case).
No change. :-/

Where do i go next?

trying to connect up a external display and restarts

(In reply to comment #16)
> trying to connect up a external display and starts

I had this problem as well, and plugging in an external monitor while starting X was the only thing I could do to workaround the issue. Additionally, enabling VBEModes would trigger a segfault.

Working on getting some hardware to reproduce this bug on.

Should be recieving some hardware the middle of January to work on this.

I received the new hardware and could repeat the problem. I created a fix and have merged it into the master branch. Please do a checkout and try it.

*** Bug 58800 has been marked as a duplicate of this bug. ***

Plugging an external monitor is an interesting bit.
Could you please post a dump of the registers both when using only LVDS and when using LVDS and VGA. Comparing the 2 might give some insight on what's wrong.

Another thing to try would be to use Option "ModeSwitchMethod" "legacy" and see if that helps. Please provide the regs dump for this too.

In order to build the registers dumper tool, add --enable-viaregtool to the configure script. Then sudo /usr/sbin/via_regs_dump -dpm

Yet more things to try :
- What about the 0.2.906 release ? Does it work better ? Please provide a regs dump too.
- As a last resort, please try with kernel mode setting. Does it work ? Again, regs dump. Here's some directions for KMS : http://www.freedesktop.org/wiki/Openchrome/TtmGemKms

Created attachment 74056 [details]
not working git master

These are registers for not working git master. latest xorg, latest 3.7.5 kernel

Created attachment 74057 [details]
"working" registers for 0.2.906 of Debian wheezy

I run Debian wheezy in chroot, xserver-xorg-video-openchrome-1:0.2.906-2 i386
Keyboard and mouse didn't work, but screen was ok.

I'm using fresh versions of packages, it is not easy to go back.
Compilation of 0.2.906 with latest xorg failed. xaalocal.h, etc.
Using CRT monitor also didn't succed, because there are problems with reading EDID. (VGA-1 disconnected).

I did some git checkouts and compilations on wheezy and:
commit 9d0a25466 was last OK.
commit f77321419 was BAD. The screen is visible, but there are wrong colours and other glitches. Between those commits there were segfaults.
That was tested on Amilo Pro 2030.

Oh that is a old bug. This is just before the 0.2.905 release. Can you try that release to see igf it works for you.

0.2.906 is usable on Debian, but does not compile with latest xserver (1.14.0).

Created attachment 77497 [details]
Xorg log for Ubuntu 13.0

I'm having the same exact issue on my Gateway with Ubuntu 13.04 beta.

lspci:
01:00.0 VGA compatible controller: VIA Technologies, Inc. CN700/P4M800 Pro/P4M800 CE/VN800 Graphics [S3 UniChrome Pro] (rev 01)

Additionally, I don't see any hw cursor and (as it's using glx) whatever I see through the mess is update *extremely* slow.

Sorry for the late reply. I have installed the latest version of Ubuntu 13.04 on a flash drive and tested again. I can happily say that this problem has been resolved for my hardware; tested with the package supplied from the Ubuntu main repo (0.3.1-0).

Gateway MX3230
01:00.0 VGA compatible controller: VIA Technologies, Inc. CN700/P4M800 Pro/P4M800 CE/VN800 Graphics [S3 UniChrome Pro] (rev 01)

fdm, thanks for reporting back. Glad the issue is gone for you.

Thomas and other people, do you still have the issue ?

I just updated that notebook to 13.04.  I am not even sure, if the driver was updated at all, but the symptoms did not change.  I don't have that notebook within reach right now, but the version number seems to be unchanged.  Should i force a re-installation?  Or try 13.04 from scratch?

With my Amilo L7320GW anything newer than 0.2.906 (including git) results in a black screen. The 0.2.906 gives me usable desktop but xv does not work and after suspend to disk I need to restart X.

Created attachment 81268 [details]
Xorg0.log with git master

Created attachment 82098 [details] [review]
LVDS drm kernel fix.

You xorg log seems to be cut off. I have a theory why your system doesn't work. It looks like similar problems with the Amilo 2030 which was solved recently. Can you try a drm-openchrome kernel.

http://www.freedesktop.org/wiki/Openchrome/TtmGemKms

and apply the patch I attached here. I have a feeling this will fix your problem. The patch is a test work around and not the final fix.

Created attachment 82150 [details]
Xorg0.log with drm-openchrome + LVDS drm kernel fix and git master

With modeset=1 the screen got garbled and the system got unresponsive as soon as the via module was loaded. With modeset=0 the screen looks just like the original description and screenshots of this bug.

I am closing this bug report.
If the problem persists with the latest version, please reopen the bug.

*** Bug 99351 has been marked as a duplicate of this bug. ***