[MPlayer-dev-eng] [PATCH] vo_vesa fixes for some cards

[Sascha Sommer]

On Friday 01 October 2004 13:59, Aurelien Jacobs wrote:
> Hi,
>
> Here are some improvements for vo_vesa.
>
> Currently vbeGetProtModeInfo call the 0x4f0a function of int 10h the get a
> simple 32 bits protected mode interface to some VESA functions. This
> protected mode interface is interesting because it's quicker than the raw
> int 10h interface. Unfortunatly, begining with VBE 3.0, the 0x4f0a function
> is optional, and some video cards don't implement it (3dfx, intel
> 845/855/865...). This protected mode interface is then only used in
> vbeSetWindow and vbeSetDisplayStart : - vbeSetWindow already implement an
> alternative methode if protected mode interface is not available. -
> vbeSetDisplayStart also contain an alternative implementation, but this one
> is disabled with a #if 0. I don't exactly know why because it works well !
>
> So currently, cards which don't have the 0x4f0a function are not
> supported. This patch correct this.
>  - vbeGetProtModeInfo failure is not fatal.
>  - vbeSetDisplayStart has it's alternative implementation reenabled.
>    it's used only with cards which don't have the 0x4f0a function
>    so this won't make any difference for cards which were already
>    working.
>
> This patch also make the failure of vbeGetModeInfo not fatal. The
> VBE 3.0 standard state that GetModeInfo can fail with some mode
> which are listed as supported if the mode can't be used in the
> current situation (not enough video memory for example). So a
> failure of vbeGetModeInfo don't mean that other modes won't work
> and should really not be fatal.
>
> This patch make vo_vesa compatible with intel 845/855/865 and with
> 3dfx video cards.
>

Anyone against this one?
Otherwise I'm goint to commit. 
It looks like Geexbox users are the only ones who are using vo vesa anyway, 
therefore I would have no problem with the geexbox devels maintaing vo vesa.

Sascha