Intel graphics driver -- thanks!

Jin, Gordon gordon.jin at
Tue Feb 13 21:29:53 PST 2007

Nick Lidakis wrote:
> Keith Packard wrote:
>> I use it on notebook machines extensively; my 915GMS has seen
>> uptimes on the order of months while being suspended/resumed almost
>> every day. Most of the credit here goes to the kernel folks who have
>> been working out the hard problems for several years. The graphics
>> piece is simple, by comparison, as it only requires that we be able
>> to reinitialize the card from power-on state. 
> I should have been more specific. My question was in regards to your
> experience with "desktop" mother boards and their chip sets. My
> Thinkpad X40 suspends reliably as well, but I have yet to get any of
> my Linux desktop machines (via ATi -closed or open driver- or Nvidia
> graphics) to do the same. Are the Intel desktop boards, with their
> integrated 
> graphics, better supported under Linux for S3 suspend?
> Forgive me if this thread strays a bit off topic; from reading various
> lists, I can't tell if S3 suspend was troublesome with desktop
> machines because of the closed nature of the Nvidia and ATI drivers,
> or because 
> of difference between mobile and desktop (ACPI?) chip sets. I just
> want 
> to be sure that an Intel desktop chip set will perform as similar as
> an Intel mobile chip set as far as suspend states are concerned
> --before I spend money on a new CPU, board, and RAM.
> If anyone else on the list has had good results using S3 suspend with
> an Intel "desktop" chip sets & integrated graphics,  pleased drop me
> an email. 

In most cases, video S3 suspend failure is caused by BIOS. And
unfortunately upgrading to the latest BIOS doesn't always help.
Different system has different behavior and workaround. This is
explained in Kernel-src/Documentation/power/video.txt.

I have some experience for S3 suspend on those models listed in
For the IBM 945GM and 915GM laptop, I could use boot option
"acpi_sleep=s3_bios" to workaround.
But for the DELL 965/945/915, I have to use vbetool to workaround.

Hope this helps.


More information about the xorg mailing list