Bug#397084: savage: a fix for respawn crash regression

["Alex Deucher" <alexdeucher@gmail.com>]

This is the right fix.  I'll check it in the next few days.

Thanks!

Alex

On 12/22/06, Mikko Rapeli <mikko.rapeli@iki.fi> wrote:
> Hello,
>
> The savage driver some time a go started crashing when respawned from
> kdm et al. When MapMMIO and MapFB functions were merged to MapMem[1], one
> MapFB call in SavageScreenInit was not replaced with MapMem:
>
> http://gitweb.freedesktop.org/?p=xorg/driver/xf86-video-savage.git;a=commitdiff_plain;h=2f8352df6488476b0c1a46798eca5dd38827444b;hp=6f9abbb972834561cd8494a1d4fb47402b285d7d
>
> Here's a patch which fixes the issue. It works on my Thinkpad T20,
> though I don't really understand the details.
>
> -Mikko
>
> --- xserver-xorg-video-savage-2.1.2.orig/src/savage_driver.c
> +++ xserver-xorg-video-savage-2.1.2/src/savage_driver.c
> @@ -3088,6 +3088,9 @@
>
>      SavageEnableMMIO(pScrn);
>
> +    if (!SavageMapMem(pScrn))
> +        return FALSE;
> +
>      psav->FBStart2nd = 0;
>
>      if (psav->overlayDepth) {
>
> _______________________________________________
> xorg mailing list
> xorg@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/xorg