--vkogqOf2sHV7VnPd
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

The attached patch is the straight port of a change I did to sparc64 to fix
a SIGSEGV due to access just above the allocated stack in the cloned process
when running the regress/lib/libc/clone test.

I'm not sure why sparc does not seem to hit this segfault, but IMHO the
same fix should be needed.

Am I overlooking something?

Martin

--vkogqOf2sHV7VnPd
Content-Type: text/plain; charset=us-ascii
Content-Disposition: attachment; filename=patch

Index: __clone.S
===================================================================
RCS file: /cvsroot/src/lib/libc/arch/sparc/sys/__clone.S,v
retrieving revision 1.3
diff -u -r1.3 __clone.S
--- __clone.S	23 Jan 2004 13:45:43 -0000	1.3
+++ __clone.S	25 Dec 2006 21:47:43 -0000
@@ -62,7 +62,7 @@
 	/*
 	 * Store the function and the argument on our child's stack
 	 */
-	sub	%o1, 8, %o1	! make space on the child's stack
+	sub	%o1, 8+96, %o1	! make space on the child's stack
 	st	%i0, [%o1]	! save func
 	st	%i3, [%o1+4]	! save arg
 

--vkogqOf2sHV7VnPd--