On Wed, Dec 10, 2003 at 08:48:33PM -0500, David Dawes wrote:
> On Thu, Dec 11, 2003 at 12:54:10AM +0000, Alan Hourihane wrote:
> >Actually,
> >
> >I've just examined the code for reading on alpha, and regardless of
> >a Jensen or other type of alpha it's always the same for a 32bit read.
> >
> >So can you reverse your patch and try this....
> >
> >Alan.
> >
> >Index: r200_screen.c
> >===================================================================
> >RCS file: /X11R6/x-cvs/xc/lib/GL/mesa/src/drv/r200/r200_screen.c,v
> >retrieving revision 1.6
> >diff -u -r1.6 r200_screen.c
> >--- r200_screen.c 2 Dec 2003 13:02:39 -0000 1.6
> >+++ r200_screen.c 11 Dec 2003 00:53:47 -0000
> >@@ -65,6 +65,18 @@
> > #define PCI_CHIP_RV200_QW 0x5157 /* Radeon 7500 - not an R200 at all */
> > #endif
> >
> >+#ifdef __alpha__
> >+
> >+#define mem_barrier() __asm__ __volatile__("mb" : : : "memory")
> >+
> >+int
> >+xf86ReadMmio32(pointer Base, register unsigned long Offset)
> >+{
> >+ mem_barrier();
> >+ return *(volatile CARD32*)((unsigned long)Base+(Offset));
> >+}
> >+#endif
> >+
> > static r200ScreenPtr __r200Screen;
> >
> > static int getSwapInfo( __DRIdrawablePrivate *dPriv, __DRIswapInfo * sInfo );
> >
>
> xf86ReadMmio32 is declared in compiler.h as a function pointer,
> not a function.
Gack. It was late here. I'll plug in a Radeon into my Alpha and come
up with something today.
Alan.
_______________________________________________
XFree86 mailing list
XFree86@xxxxxxxxxxx
http://XFree86.Org/mailman/listinfo/xfree86
