>> > And what about video cards that implement in hardware functions
traditionally
>> > found in the software driver? They certainly give better
performance but they
>> > are tied to one particular OS interface - such cards exist, so what
other OSes
>>
>> Not really. At a drawing level there is really no difference between
X11
>> and Win32 other than the cursor format. Cursor format conversion is
not
>> exactly hard and some chips like the S3 ones actually support both
natively!
>And what about cards that would have on-board implementation of the
TRUE Win32
>video driver APIs (similarily to those cards that have on-board OpenGL
APIs
>implemented) - then you'd have to use the Win32 routines under any OS
that
>wants to support that cards, right?
Then we create a dumb framebuffer driver. The Win32 GDI API has simple
operations -- put a pixel, fill a rectangle. It's not any different
than using other accelerator functions -- just a different way of
calling them. Depending on how it's implemented, it might not be fast,
but it could be done.
Actually, this might be an improvement if it's done right... don't know
if it's feasible or not, but if they implement the Win32 API in a common
(Micro$oft endorsed) way, then we might be able to get away with writing
one driver for Win32-API graphics cards. At least, if we used the Win32
routines, the cards will have a "standard" enforced -- even if it's not
one we like. The API isn't the OS. ;)
But, it's late and I might be completely wrong due to lack of caffeine.
Josh
buyssej@coffman.umn.edu
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu