On Aug 9, 2006, at 12:37 PM, Keir Fraser wrote:
On 9/8/06 1:34 pm, "Jimi Xenidis" <jimix@xxxxxxxxxxxxxx> wrote:
2) nasty build break in libaio
Shouldn't you fix that instead of skirting the issue? :-)
I knew you would say that! :)
Will you take patches to libaio, or are you expecting that to stay
clean to the distro?
I may have missed this discussion, but why can't we just use the one
that comes with the Distro?
That's a question for Julian (cc'ed). I assume it *ought* to work on
powerpc, so perhaps at worst we add a patch and also send it upstream.
ok, I checked in the fix, you will get it in the next pull from our
merge tree.
Here it is for reference:
changeset: 11432:e60de2425ec47caf50301b6aa164c61ff1540eb7
user: Jimi Xenidis <jimix@xxxxxxxxxxxxxx>
date: Wed Aug 09 15:38:37 2006 -0400
files: tools/libaio/src/syscall-ppc.h
description:
[POWERPC][TOOLS] fix header so libaio builds
Signed-off-by: Jimi Xenidis <jimix@xxxxxxxxxxxxxx>
diff -r 0bb2e9a96c33777247319e6ff3428212bf5c1f17 -r
e60de2425ec47caf50301b6aa164c61ff1540eb7
tools/libaio/src/syscall-ppc.h
--- a/tools/libaio/src/syscall-ppc.h Wed Aug 09 15:37:42 2006 -0400
+++ b/tools/libaio/src/syscall-ppc.h Wed Aug 09 15:38:37 2006 -0400
@@ -1,3 +1,6 @@
+#include <asm/unistd.h>
+#include <errno.h>
+
#define __NR_io_setup 227
#define __NR_io_destroy 228
#define __NR_io_getevents 229
@@ -9,7 +12,7 @@
* "sc; bnslr" sequence) and CR (where only CR0.SO is clobbered to
signal
* an error return status).
*/
-
+#ifndef __syscall_nr
#define __syscall_nr(nr, type, name, args...) \
unsigned long __sc_ret, __sc_err; \
{ \
@@ -37,6 +40,7 @@
} \
if (__sc_err & 0x10000000) return -((int)__sc_ret); \
return (type) __sc_ret
+#endif
#define __sc_loadargs_0(name, dummy...) \
__sc_0 = __NR_##name
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
Home