lkml.org 
[lkml]   [2008]   [Jun]   [17]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
SubjectRe: Bug in arch/i386/lib/delay.c file, delay_loop function

* Jiri Hladky <hladky.jiri@googlemail.com> wrote:

> Hi Ingo, hi Martin!
>
> As Martin correctly pointed out, my code will fail when loops=0. So I
> have added a fix for this.

applied to tip/x86/delay, thanks Jiri! If it passes testing it will show
up in the v2.6.27 kernel.

[ A few small patch format nits: the first hunk of the patch didnt
apply, it was whitespace-damaged - i hand-merged that. The assembly
had non-standard formatting - nowadays we try to format the assembly
so that it looks nice in the .c file, not i the .s file. I fixed that
too. Also, the best way is to send fixes in the format below, with a
nice commit description, followed by a signed-off-by line. I fixed
that as well. ]

please double-check the end result. You can pick up tip/master via:

http://people.redhat.com/mingo/tip.git/README

do "git-log -1 e01b70ef" or "git-log arch/x86/lib/delay_32.c" to see
your commit.

Ingo

------------->
commit e01b70ef3eb3080fecc35e15f68cd274c0a48163
Author: Jiri Hladky <hladky.jiri@googlemail.com>
Date: Mon Jun 2 12:00:19 2008 +0200

x86: fix bug in arch/i386/lib/delay.c file, delay_loop function

when trying to understand how Bogomips are implemented I have found a
bug in arch/i386/lib/delay.c file, delay_loop function.

The function fails for loops > 2^31+1. It because SF is set when dec
returns numbers > 2^31.

The fix is to use jnz instruction instead of jns (and add one decl
instruction to the end to have exactly the same number of loops as in
original version).

Martin Mares observed:

> It is a long time since I have hacked that file, but you should definitely
> make sure that the function is never called with a zero argument. In such
> case, the original version made just a single pass, but your version
> makes 2^32 of them.

fixed that.

Signed-off-by: Ingo Molnar <mingo@elte.hu>

diff --git a/arch/x86/lib/delay_32.c b/arch/x86/lib/delay_32.c
index d710f2d..ef69131 100644
--- a/arch/x86/lib/delay_32.c
+++ b/arch/x86/lib/delay_32.c
@@ -3,6 +3,7 @@
*
* Copyright (C) 1993 Linus Torvalds
* Copyright (C) 1997 Martin Mares <mj@atrey.karlin.mff.cuni.cz>
+ * Copyright (C) 2008 Jiri Hladky <hladky _dot_ jiri _at_ gmail _dot_ com>
*
* The __delay function must _NOT_ be inlined as its execution time
* depends wildly on alignment on many x86 processors. The additional
@@ -28,16 +29,22 @@
/* simple loop based delay: */
static void delay_loop(unsigned long loops)
{
- int d0;
-
__asm__ __volatile__(
- "\tjmp 1f\n"
- ".align 16\n"
- "1:\tjmp 2f\n"
- ".align 16\n"
- "2:\tdecl %0\n\tjns 2b"
- :"=&a" (d0)
- :"0" (loops));
+ " test %0,%0 \n"
+ " jz 3f \n"
+ " jmp 1f \n"
+
+ ".align 16 \n"
+ "1: jmp 2f \n"
+
+ ".align 16 \n"
+ "2: decl %0 \n"
+ " jnz 2b \n"
+ "3: decl %0 \n"
+
+ : /* we don't need output */
+ :"a" (loops)
+ );
}

/* TSC based delay: */

\
 
 \ /
  Last update: 2008-06-17 11:07    [W:0.208 / U:0.016 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site