Re: PROBLEM: Oops in __is_prefetch with 2.6.0-test9-bk4 at boottime with Athlon XP 1800+

[Andi Kleen]

On Sat, 1 Nov 2003 07:01:38 +0100 (MET)
<matze@xxxxxxxxxxx> wrote:

> Andi, it seems that you are the right person for this. It looks like my athlon
> dies in the __get_user call in __is_prefetch at boot up when bash is used as
> init process (look in [2.] for more details).
> 
> Just ask for any additional information.

Does it go away when you apply the following patch?

I need your config, but the complete config (without the comments dropped) 
and without the attachment converted to CR-LF line endings. I just spent several minutes
debugging the extremly cryptic error message from Kconfig you get for that.

Most likely something is reordering the exception table, breaking __get_user.

-Andi

diff -u linux-test9bk4/include/linux/init.h-o linux-test9bk4/include/linux/init.h
--- linux-test9bk4/include/linux/init.h-o	2003-09-28 10:53:23.000000000 +0200
+++ linux-test9bk4/include/linux/init.h	2003-11-01 22:10:21.000000000 +0100
@@ -40,15 +40,15 @@
 
 /* These are for everybody (although not all archs will actually
    discard it in modules) */
-#define __init		__attribute__ ((__section__ (".init.text")))
+#define __init		
 #define __initdata	__attribute__ ((__section__ (".init.data")))
 #define __exitdata	__attribute__ ((__section__(".exit.data")))
-#define __exit_call	__attribute_used__ __attribute__ ((__section__ (".exitcall.exit")))
+#define __exit_call	
 
 #ifdef MODULE
 #define __exit		__attribute__ ((__section__(".exit.text")))
 #else
-#define __exit		__attribute_used__ __attribute__ ((__section__(".exit.text")))
+#define __exit		
 #endif
 
 /* For assembly routines */


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/