[kaffe] Re: kaffe on pxa255 runs now

[Holger Schurig]

--nextPart1924836.tpCUGyLtag
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8Bit

> thanks for the patch. I've added an explicit --enable-pxa option to the
> configure script to ease building with gcc-2.95.3. Could you try that
> out, and report back if it works?

Oh, one note: yesterday I was running with the geri's dflush-Patch. Now this
patch is in CVS and I get this error:

/root/tmp/ccGWflE4.s: Assembler messages:
/root/tmp/ccGWflE4.s:179: Error: Internal_relocation (type 208) not fixed up
(IMMEDIATE)

Here is a patch to make this working.

-- 
Try Linux 2.6 from BitKeeper for PXA2x0 CPUs at
http://www.mn-logistik.de/unsupported/linux-2.6/
--nextPart1924836.tpCUGyLtag
Content-Type: text/x-diff; name="pxajit.patch"
Content-Transfer-Encoding: 8Bit
Content-Disposition: attachment; filename="pxajit.patch"


#
# Patch managed by http://www.mn-logistik.de/unsupported/pxa250/patcher
#

--- head/config/arm/linux/md.c~pxajit
+++ head/config/arm/linux/md.c
@@ -14,13 +14,6 @@
 #include <sched.h>
 #include <asm/unistd.h>
 
-/* If VM_EXEC is not defined, use definition from 
- * linux/mm.h.
- */
-#if !defined(VM_EXEC)
-long VM_EXEC =       0x00000004;
-#endif /* !defined(VM_EXEC) */
-
 void            
 init_md(void)
 {       
@@ -32,12 +25,13 @@
 /**
  * Shamelessly stolen from parrot... ([perl]/parrot/jit/arm/jit_emit.h arm_sync_d_i_cache)
  *
- * r2 should be zero for 2.4 (but it's ignored) so passing VM_EXEC (needed for 2.6) should be okay.
+ * r2 should be zero for 2.4 (but it's ignored) so passing VM_EXEC (needed
+ * for 2.6) should be okay.
  */
 void flush_dcache(void *start, void *end) {
   __asm __volatile ("mov r0, %0\n"
 		    "mov r1, %1\n"
-		    "mov r2, #VM_EXEC\n"
+		    "mov r2, #4\n"		/* this is VM_EXEC from linux/mm.h, needed for Kernel 2.6 */
 		    "swi " __sys1(__ARM_NR_cacheflush) "\n"
 		    : /* no return value */
 		    : "r" ((long)start), "r" ((long)end)

--nextPart1924836.tpCUGyLtag--