bugtracker i855GM DualHead

David Bronaugh dbronaugh at linuxboxen.org
Thu Feb 3 20:00:39 PST 2005


Alex Deucher wrote:

>On Thu, 3 Feb 2005 22:18:37 +0100, hephaistos6 at sms.at
><hephaistos6 at sms.at> wrote:
>  
>
>>Hello,
>>
>>I would like to ask why the bug https://bugs.freedesktop.org/show_bug.cgi?id=1064 opened on 2004-08-12 is still "NEW".
>>Is there a special matter with i855GM (i810, i915) and dualhead?
>>
>>I would really enjoy if some could please tell me what the problem is.
>>
>>thanks very much in advance,
>>stefan walkner
>>
>>PS: sorry for sounding "impolite" >> I don't mean it impolite - I'm just interested in the reason why the bug is still unattended
>>
>>    
>>
>apparently a few users are having problems with the code, but none of
>the developers are able to reproduce the bug, or no one's had a chance
>to look into it.  The code seems to work for most people.  I guess
>lack of hardware, documentation, and time...
>
>Alex
>
I can't help with the lack of documentation and time, but I do have the 
hardware.

Here's what I've observed with the DRI CVS Xorg server, i830:
 
The Xorg configuration is very finicky; if you don't have everything 
exactly right in the config, it doesn't work.

If you start the machine with a monitor connected to the external 
output, things work pretty well -- when you fire up with dualhead, the 
second screen is enabled

However, if you don't start up with a monitor connected, the external 
monitor is disabled when you start X -- I have to run i810switch crt on

Switching to text mode causes bizarre corruption of the accelerator; 
text never appears to be cleared, and is often never draw. I'm not sure 
-what- this is. It also persists over restarts of the X server. I end up 
rebooting my laptop when I make this mistake.

Occasionally, and for unknown reasons, text drawing in gtk1 apps and 
xemacs, and text position cursors (not mouse cursors) in at least gtk2 
apps, seem to get corrupted so that only part of it draws. It -seems- 
like some kind of thing where random horizontal lines go missing - I 
imagine these aren't random, but that's what it looks like.

The error log contains wonderful information like:
 - (WW) I810(0): Bad V_BIOS checksum
 - (WW) I810(1): Extended BIOS function 0x5f11 not supported.

At the end of the log, I see things like this (this may be recent):

(WW) I810(1): Extended BIOS function 0x5f1c not supported.
(EE) I810(0): unknown reason for exception
(II) I810(0): EAX=0x00004f10, EBX=0x00000000, ECX=0x00000000, EDX=0x00000098
(II) I810(0): ESP=0x00000fac, EBP=0x00000fdc, ESI=0xffff1415, EDI=0x00002000
(II) I810(0): CS=0xc000, SS=0x0100, DS=0x0000, ES=0x0000, FS=0x0000, 
GS=0x0000
(II) I810(0): EIP=0x000040c6, EFLAGS=0x00033046
(II) stack at 0x00001fac:
 01 00 00 00 42 41 93 32 98 00 01 00 00 00 15 14
 07 00 00 00 26 3e 10 4f 1e 14 07 00 66 33 00 00
 ff ff ee 18 10 4f 45 80 ff ff ff ff ff ff ff ff
 40 00 00 00 dc 0f 00 20 00 00 15 14 ff ff 98 00
 00 00 00 00 00 00 01 00 00 00 10 4f 00 00 00 06
 00 00 00 32
(II) I810(0): code at 0x000c40c6:
 8a 06 eb 0e 80 fd 01 75 05 67 8b 06 eb 04 67 66
 8b 06 5b 1f c3 52 eb 11 52 8b 16 63 04 eb 0a 52
(EE) I810(0): cannot continue

However, it still seems to be working; cross fingers.

None of these problems -appear- to happen when not running dualhead.

FWIW, here's what the VBE BIOS info says:

(II) I810(0): VESA VBE OEM: Almador Graphics Chip Accelerated VGA BIOS
(II) I810(0): VESA VBE OEM Software Rev: 1.0
(II) I810(0): VESA VBE OEM Vendor: Intel Corporation
(II) I810(0): VESA VBE OEM Product: Almador Graphics Controller
(II) I810(0): VESA VBE OEM Product Rev: Hardware Version 0.0

I've attached my config, so that others may learn from it or something 
like that.

David Bronaugh
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: xorg.conf
URL: <http://lists.x.org/archives/xorg/attachments/20050203/2d0e5fca/attachment.ksh>


More information about the xorg mailing list