Re: [PATCH] matroxfb big-endian update (was Re: [PATCH] ppc64: Fix __raw_* IO accessors)

[Petr Vandrovec]

On Fri, Sep 24, 2004 at 04:25:37PM +1000, Benjamin Herrenschmidt wrote:
> On Fri, 2004-09-24 at 06:26, Petr Vandrovec wrote:
> > Hi Andrew & Linus,
> > 
> > This change disconnects matroxfb accelerator endianess from processor endianess, plus
> > ports big-endian accessors from __raw_xxx to xxx + appropriate byte swaps.
> 
> Applied to current bk, make oldconfig (FB_MATROX_BIG_ENDIAN_ACCEL is not set),
> works like a charm on the ppc POWER3 I have here, haven't had a chance to
> test X though.

Thanks. 

XFree 3.x did not touch BE mode bit and accessed MMIO directly with pointer 
dereference (expecting firmware to put card into BE mode?), while XFree 4.x (if
I understand code properly) does not touch BE bit on primary device
(while it clears it on secondary devices) while expecting hardware to be
in LE mode...

So I'm either confused, or XF3 needs BE_ACCEL set while XF4 needs BE_ACCEL
disabled.  Does anybody actually use matroxfb with XFree server on PPC (or any
other BE machine) at all?
							Petr Vandrovec

P.S.: Trimmed down CC list a bit. 
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/