Re: Some user feedback on recent Intel chipset and 2.6.29 KMS

[Dave Airlie]

On Wed, Feb 25, 2009 at 8:19 PM, Eric Valette <Eric.Valette@xxxxxxx> wrote:
> Hi,
>
> Thanks for the nice work: for the first time (starting with
> 2.6.29-rc6-git1), I managed to remove the vesa framebuffer in favor of
> intel's ones.  Recent X server (1.5.99.902 + git intel driver and git
> drm) also comes up correctly and detects the KMS.
>
> This amilo xi 3650 is a transportable SLI enabled laptop with Nvidia
> GeForce 9600M GT & Mobile Intel® GM45 Express Chipset. By defaut the
> Intel card is active. I can force using one or other in the bios but
> expects one day I will be able to be able to dynamically switch like on
> vista.
>
> However there are still some problem with drm, Xorg, suspend/resume:
>
>   1. The X server needs interrupts (mouse or keyboard) to redraw things
>      (e.g if I do apt-get update in a konsole, I see it freeeze but if
>      I move the mouse I get the complete output instantly. Same thing
>      when I start a new application, I have to hit shift or move mouse.
>      The same user land setting with 2.6.28.7 behaves correctly,
>   2. I have a lot of problem with suspend to ram: when console comes up
>      but with a cyan background and then X comes up completely garbled
>      although I can distinguish some shadow of the windows and a
>      konsole is still active if I enter init 0. Sometimes it purely
>      freeze. If I switch back to tty (when I can), and restart kdm,
>      then it starts only in 800x600 mode. Using krandrtray to put it
>      back at 1920x1080, it simply freezes.
>

See if the attached patch helps resume, there are also some other patches
queued for Linus in my drm-fixes tree. I sent the pull req earlier.

Dave.

> If I can help testing anything, providing traces (dmesg, config, lspci,
> xorg config) let me know.
>
> PS: CC me if anwering to the list I'm not suscribed.
>
> -- eric
>

Attachment: mypatch
Description: Binary data