[regression fix] really dumb fuckup in sparc64 __csum_partial_copy() changes

~0U is -1, not 1

Reported-by: Anatoly Pugachev <matorola@gmail.com>
Tested-by: Anatoly Pugachev <matorola@gmail.com>
Fixes: fdf8bee96f "sparc64: propagate the calling convention changes down to __csum_partial_copy_...()"
X-brown-paperbag: yes
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>

arch/sparc/lib/csum_copy.S

@@ -71,7 +71,7 @@
 FUNC_NAME:		/* %o0=src, %o1=dst, %o2=len */
 	LOAD(prefetch, %o0 + 0x000, #n_reads)
 	xor		%o0, %o1, %g1
-	mov		1, %o3
+	mov		-1, %o3
 	clr		%o4
 	andcc		%g1, 0x3, %g0
 	bne,pn		%icc, 95f