[PATCH] Dead code in i386/kernel/process.c

[PATCH] Dead code in i386/kernel/process.c

[Brian Gerst]

[-- Attachment #1: Type: text/plain, Size: 110 bytes --]

This patch removes an artifact of code left over from the 2.0 days when 
the kernel didn't use flat segments.

[-- Attachment #2: pg0-1 --]
[-- Type: text/plain, Size: 597 bytes --]

diff -urN linux-2.5.30/arch/i386/kernel/process.c linux/arch/i386/kernel/process.c
--- linux-2.5.30/arch/i386/kernel/process.c	Fri Aug  2 10:15:28 2002
+++ linux/arch/i386/kernel/process.c	Sat Aug  3 16:41:03 2002
@@ -313,11 +313,6 @@
 	memcpy (swapper_pg_dir, swapper_pg_dir + USER_PGD_PTRS,
 		sizeof (swapper_pg_dir [0]) * KERNEL_PGD_PTRS);
 
-	/* Make sure the first page is mapped to the start of physical memory.
-	   It is normally not mapped, to trap kernel NULL pointer dereferences. */
-
-	pg0[0] = _PAGE_RW | _PAGE_PRESENT;
-
 	/*
 	 * Use `swapper_pg_dir' as our page directory.
 	 */

Re: [PATCH] Dead code in i386/kernel/process.c

[Brian Gerst]

Brian Gerst wrote:
> This patch removes an artifact of code left over from the 2.0 days when 
> the kernel didn't use flat segments.
> 
> 
> ------------------------------------------------------------------------
> 
> diff -urN linux-2.5.30/arch/i386/kernel/process.c linux/arch/i386/kernel/process.c
> --- linux-2.5.30/arch/i386/kernel/process.c	Fri Aug  2 10:15:28 2002
> +++ linux/arch/i386/kernel/process.c	Sat Aug  3 16:41:03 2002
> @@ -313,11 +313,6 @@
>  	memcpy (swapper_pg_dir, swapper_pg_dir + USER_PGD_PTRS,
>  		sizeof (swapper_pg_dir [0]) * KERNEL_PGD_PTRS);
>  
> -	/* Make sure the first page is mapped to the start of physical memory.
> -	   It is normally not mapped, to trap kernel NULL pointer dereferences. */
> -
> -	pg0[0] = _PAGE_RW | _PAGE_PRESENT;
> -
>  	/*
>  	 * Use `swapper_pg_dir' as our page directory.
>  	 */

I should clarify that the dead code does the same thing as the preceding 
memcpy.

Re: [PATCH] Dead code in i386/kernel/process.c

[Jamie Lokier]

Brian Gerst wrote:
> This patch removes an artifact of code left over from the 2.0 days when 
> the kernel didn't use flat segments.

This code dates back to a patch for 1.2.8 :-)
The comment is quite wrong now.

Can we trust that arch/i386/mm/init.c will continue to map the page at
0xc0000000 (PAGE_OFFSET) to physical address 0?  I guess so, hence the
patch is fine.

-- Jamie