Linus Torvalds wrote:
> Note that the i386 has _long_ been a "stepchild", though: because of the
> lack of WP, the kernel simply doesn't do threaded MM correctly on a 386.
That reminds me of a patch to gut verify_write when CONFIG_X86_WP_WORKS_OK=y.
Saves a few bytes. Patch is against 2.5.7pre2.
We could bin the whole thing and stop exporting it, but there may be users
who try to load binary only modules that were built for 386 and hence expect
verify_write to be exported regardless of CPU type the kernel was built for.
Paul.
--- ../linux/arch/i386/mm/fault.c Sun Mar 17 07:14:09 2002
+++ arch/i386/mm/fault.c Sun Mar 17 07:02:21 2002
@@ -4,6 +4,7 @@
* Copyright (C) 1995 Linus Torvalds
*/
+#include <linux/config.h>
#include <linux/signal.h>
#include <linux/sched.h>
#include <linux/kernel.h>
@@ -29,6 +30,9 @@
extern int console_loglevel;
+#ifdef CONFIG_X86_WP_WORKS_OK
+int __verify_write(const void * addr, unsigned long size) {return 1;}
+#else
/*
* Ugly, ugly, but the goto's result in better assembly..
*/
@@ -93,6 +97,7 @@
}
goto bad_area;
}
+#endif
extern spinlock_t timerlist_lock;
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sat Mar 23 2002 - 22:00:13 EST