[ltp] Savage Kernel Driver Problems (T21)
Kevin Locke
linux-thinkpad@linux-thinkpad.org
Sun, 20 Mar 2005 22:09:12 -0700
Hi all,
After recently installing Linux on a T21, I have been having a few
problems with the savage driver in the 2.6.10 kernel and I wanted to
know if anyone else on the list has encountered these problems or if
there are any known workarounds.
For the following discussion, "with savage" means the kernel was
compiled with:
CONFIG_FB_VESA=n
CONFIG_FB_SAVAGE=y
CONFIG_FB_SAVAGE_I2C=y
CONFIG_FB_SAVAGE_ACCEL=n (completely garbles the screen)
and "with vesafb" means:
CONFIG_FB_VESA=y
CONFIG_FB_SAVAGE=n
and both have
CONFIG_VIDEO_SELECT=y
CONFIG_FRAMEBUFFER_CONSOLE=y
First, with the savage driver, when I use a video mode with a resolution
of 1024x768 (e.g. vga=0x305) there is a large black border around the
screen text (it looks the same as using an 800x600 resolution and
turning off video-expansion in the BIOS). The problem only appears in
the console, inside X the black border disappears. With vesafb, video
modes with 1024x768 look great, no black border.
Second, with the savage driver, I am unable to resume the computer from
standby. After "echo standby > /sys/power/state" from single-user mode
the computer goes to sleep, but after pressing the Fn key to wake the
machine "state: 1" is displayed in the upper left corner of the screen
and the system becomes completely unresponsive, requiring a reboot.
With vesafb, the system enters and exits standby without problems (even
from within X).
Third, with both savage and vesafb, when the computer enters standby the
screen is filled with a wash of colors that fades out from the center.
There are also colorful vertical lines that appear in the center of the
screen (this is all very hard to describe, but rather beautiful to
watch ;) I am not overly concerned about this, since using vbetool to
turn off the monitor first should solve the problem, but I thought it
was worth a mention if only for the oddity of the effects (possibly
generated by random discharges?).
I do not know if there is any reason not to just switch to vesafb, but
if these problems are happening on other machines it would probably be
worth reporting to the kernel BTS.
Thanks in advance,
Kevin