Re: [MiNT] New MiNT distribution is [long]

[Michael Schwingen <rincewind@discworld.dascon.de>]

On Mon, Feb 21, 2000 at 11:31:19AM -0500, Katherine Ellis wrote:
> 
> hmm driver is architecture dependent, you don't access the graphics, the
> same way on a falcon or hades, so I really don't see the relation. 

No, they are not.

A lot of drivers for PCI cards in Linux are architecture independent, for
example the LSI Logic SCSI driver works at least on i386, alpha and sparc,
the same is true for many ethernet card drivers (eg. DEC tulip) and I am
sure a lot more (I happen to use the driver for PC-style serial, parallel,
kbd/mouse and IDE ports on my Alpha machine, as well as XFree86).

And those drivers that are not yet architecture-independant are usually easy
to fix, much easier than porting them to some other platform.

> Linux is not as generic as you seem to make us believe.
> Heck, TC mode on falcon is not even supported with linux68k.

Well? Linux (as in the kernel) does not have to do much with graphics. If
the falcon happens to be the only machine with such a braindamaged video
memory layout in highcolor modes (the falcon does not have 24-bit TC), then
obviously someone has to modify Xfree86 adequately. What was your point? We
are talking about *drivers*.

> Linux68k is in the same situation as MiNT/TOS, If the manufacturer doesn't
> want to give specs, you're screwed. 
> While this is not true with linuxi86, as binaries are florishing.

Not in the kernel. *All* drivers that come with the kernel are in source and
distributed under GPL - Linus will not accept binary-only drivers for
inclusion with the kernel.

cu
Michael
-- 
Michael Schwingen, Ahornstrasse 36, 52074 Aachen