[ltp] T22 sometimes doesn't display X-Server
FeRD
linux-thinkpad@linux-thinkpad.org
Sat, 10 Feb 2007 01:18:08 -0500
This is a multi-part message in MIME format.
--------------080109020009070606010009
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Alex Deucher wrote:
> On 2/1/07, Gerhard Brauer <gerhard.brauer@web.de> wrote:
>> My Device-Section in xorg.conf is very minimalistic at the moment:
>> Section "Device"
>> Identifier "S3 Inc. 86C270-294 Savage/IX-MV"
>> Driver "savage"
>> BusID "PCI:1:0:0"
>> Option "UseFBDev" "false"
>> Option "BiosHotKeys" "true"
>
> neither of the above options are valid for savage. you can remove them.
Indeed -- it sounds, Gerhard, like you've been working with
Thinkpad-graphics-problems advice written for Thinkpads that contain ATI
graphics chips. (MOST advice you'll find out there is specific to ATI
graphics chips. Whether that's because they're more powerful, more
popular, or more problematic is a question for another day...)
The savage(4) man page runs through the (very limited) options for the
display chip in a T22. It's worth reading through. Specifically, the T22
contains this type of graphics chip:
Savage/MX (8c10 and 8c11) (2D, 3D, Dualhead)
(However, don't ge t too excited about the 3D support -- on my T22, the
8MB of video RAM was nowhere near enough to support 3D at the LCD
panel's native 1440x1050 resolution. And since resolution switching is
death (see below), I personally never even bothered to test accelerated 3D.)
With my T22, I remember exactly two issues that made the difference
between the laptop video being really flaky with lots of
hangs/corruptions, and its display support being rock-solid.
Option "ShadowStatus" "on"
You should definitely enable this in your xorg.conf. The savage(4)
man page goes into detail on what this means, and I can report from
personal experience that the bug it works around is definitely
present on the T22. It will make a huge difference in the stability
of X on your laptop.
Boot with the console framebuffer set to the same resolution you run X
at, and NEVER SWITCH RESOLUTION
Resolution-switching is death to the savage chipset, especially on
laptops. I found that when I booted in text mode, or tried to run X
at a different resolution than the console framebuffer, 9 times out
of 13 I ended up with a corrupted display or a frozen computer.
Switching from X to a text console left me with a screen smeared
with random purple noise, which wasn't very useful even if the
computer was still running.
On my T22, running the LCD at 1440x1050 native resolution requires
adding VGA=0x33e (or VGA=830, in decimal) to the kernel command
line. Different resolutions/color depths require different VGA mode
numbers, I forget if I used "vga=ask" on boot or the vbetool(1)
command from a console to obtain the list. One of those methods
should work... if you get a list with lots of mode numbers in the
0x100's range, just add 0x200 to those to get the proper BIOS modes.
Now, my second piece of advice above may be a problem for you, since you
mentioned using a port replicator on your laptop... I'm not sure if
there's an easy solution to being able to switch resolutions for an
external display, or run a display on the VGA port at any resolution
other than the LCD's native settings. But at least for running the LCD
itself at native resolution, or an external monitor at the same
resolution as the LCD, I found that by following the two rules above, my
T22's video under Fedora Core 3 thru Fedora Core 5 was perfectly dependable.
Hope this helps,
-FeRD
--------------080109020009070606010009
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: 8bit
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=UTF-8" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
<br>
<br>
Alex Deucher wrote:
<blockquote
cite="mida728f9f90702011224t3fb07752s1d4044b2b32c4a1@mail.gmail.com"
type="cite">On 2/1/07, Gerhard Brauer <a class="moz-txt-link-rfc2396E" href="mailto:gerhard.brauer@web.de"><gerhard.brauer@web.de></a>
wrote:
<br>
<blockquote type="cite">My Device-Section in xorg.conf is very
minimalistic at the moment:
<br>
Section "Device"
<br>
Identifier "S3 Inc. 86C270-294 Savage/IX-MV"
<br>
Driver "savage"
<br>
BusID "PCI:1:0:0"
<br>
Option "UseFBDev" "false"
<br>
Option "BiosHotKeys" "true"
<br>
</blockquote>
<br>
neither of the above options are valid for savage. you can remove
them.
<br>
</blockquote>
Indeed -- it sounds, Gerhard, like you've been working with
Thinkpad-graphics-problems advice written for Thinkpads that contain
ATI graphics chips. (MOST advice you'll find out there is specific to
ATI graphics chips. Whether that's because they're more powerful, more
popular, or more problematic is a question for another day...)<br>
<br>
The savage(4) man page runs through the (very limited) options for the
display chip in a T22. It's worth reading through. Specifically, the
T22 contains this type of graphics chip:<br>
<br>
Savage/MX (8c10 and 8c11) (2D, 3D, Dualhead)<br>
<br>
(However, don't ge t too excited about the 3D support -- on my T22, the
8MB of video RAM was nowhere near enough to support 3D at the LCD
panel's native 1440x1050 resolution. And since resolution switching is
death (see below), I personally never even bothered to test accelerated
3D.)<br>
<br>
<br>
With my T22, I remember exactly two issues that made the difference
between the laptop video being really flaky with lots of
hangs/corruptions, and its display support being rock-solid.<br>
<br>
<dl>
<dt>Option "ShadowStatus" "on"</dt>
<dd>You should definitely enable this in your xorg.conf. The
savage(4) man
page goes into detail on what this means, and I can report from
personal experience that the bug it works around is definitely present
on the T22. It will make a huge difference in the stability of X on
your laptop.<br>
<br>
</dd>
<dt>Boot with the console framebuffer set to the same resolution you
run X at, and NEVER SWITCH RESOLUTION</dt>
<dd>Resolution-switching is death to the savage chipset, especially
on laptops. I found that when I booted in text mode, or tried to run X
at a different resolution than the console framebuffer, 9 times out of
13 I ended up with a corrupted display or a frozen computer. Switching
from X to a text console left me with a screen smeared with random
purple noise, which wasn't very useful even if the computer was still
running.<br>
<br>
On my T22, running the LCD at 1440x1050 native resolution requires
adding VGA=0x33e (or VGA=830, in decimal) to the kernel command line.
Different resolutions/color depths require different VGA mode numbers,
I forget if I used "vga=ask" on boot or the vbetool(1) command from a
console to obtain the list. One of those methods should work... if you
get a list with lots of mode numbers in the 0x100's range, just add
0x200 to those to get the proper BIOS modes.</dd>
</dl>
Now, my second piece of advice above may be a problem for you, since
you mentioned using a port replicator on your laptop... I'm not sure if
there's an easy solution to being able to switch resolutions for an
external display, or run a display on the VGA port at any resolution
other than the LCD's native settings. But at least for running the LCD
itself at native resolution, or an external monitor at the same
resolution as the LCD, I found that by following the two rules above,
my T22's video under Fedora Core 3 thru Fedora Core 5 was perfectly
dependable.<br>
<br>
Hope this helps,<br>
-FeRD<br>
</body>
</html>
--------------080109020009070606010009--