Re: [PATCH] Early init for security modules

Previous message: Stephen Smalley: "[RFC][PATCH] /proc/pid inode security labels 2.5.69-bk"
In reply to: Chris Wright: "[PATCH] Early init for security modules"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

samat_private

On Mon, May 12, 2003 at 08:03:09PM -0700, Chris Wright wrote:
> 
> --- 1.30/arch/i386/vmlinux.lds.S	Tue May  6 06:54:06 2003
> +++ edited/arch/i386/vmlinux.lds.S	Mon May 12 16:20:10 2003
> @@ -81,6 +81,9 @@
>    __con_initcall_start = .;
>    .con_initcall.init : { *(.con_initcall.init) }
>    __con_initcall_end = .;
> +  __security_initcall_start = .;
> +  .security_initcall.init : { *(.security_initcall.init) }
> +  __security_initcall_end = .;

I would much prefer to have only:

+ SECURITY_INIT

and moving the common stuff to include/asm-generic/vmlinux.lds.h.
Note that I moved definition of _start and _stop inside brackets.
Doing this makes sure the start address is always correct, independent
of the end address of last section.

Starting a new section will align to member with biggest alignment,
so we may see _start have a wrong value in some cases.

Using SECURITY_INIT will make changes to all architectures
even more trivial.

	Sam

===== include/asm-generic/vmlinux.lds.h 1.7 vs edited =====

--- 1.7/include/asm-generic/vmlinux.lds.h	Mon Feb  3 22:00:30 2003
+++ edited/include/asm-generic/vmlinux.lds.h	Tue May 13 20:02:45 2003
@@ -45,3 +45,9 @@
 		*(__ksymtab_strings)					\
 	}
 
+#define SECURITY_INIT							\
+	.security_initcall.init : {					\
+		__security_initcall_start = .;				\
+	       	*(.security_initcall.init) 				\
+		__security_initcall_end = .;				\
+	}

_______________________________________________
linux-security-module mailing list
linux-security-moduleat_private
http://mail.wirex.com/mailman/listinfo/linux-security-module
