Video PM Methods

Matthew Garrett mjg59 at srcf.ucam.org
Wed May 17 03:20:30 PDT 2006


On Tue, May 16, 2006 at 05:33:04PM -0400, Pat Suwalski wrote:

> Does anyone know if it's just a rePOSTing issue or something? Could it 
> really not be rectified in the kernel?

You can work around it with vesafb to some extent:

1) Save the vesa mode number
2) Disable the framebuffer (echo 3 >/sys/class/graphics/fb0/state)
3) Suspend
4) Do whatever POST tricks you want to
5) Restore the vesa mode
6) Enable the framebuffer (echo 0 >/sys/class/graphics/fb0/state)

I added support for this in Ubuntu and it sort of works, but it doesn't 
seem to be as reliable as vga16fb. I suspect that the problem is that on 
some machines the card isn't sensibly reprogrammed until you switch back 
to X, meaning that enabling the framebuffer before then could lead to 
the kernel doing inappropriate things to a partially configured graphics 
chipset.

The enabling/disabling thing is a patch I pushed in around 2.6.15 - it 
just causes the framebuffer layer to discard all reads and writes before 
they hit the framebuffer driver, which then makes it less likely that 
the driver gets confused and takes down the hardware.
-- 
Matthew Garrett | mjg59 at srcf.ucam.org


More information about the hal mailing list