From: Michael Hennerich <michael.hennerich@analog.com>
Date: Wed, 4 Feb 2009 08:49:45 +0000 (+0800)
Subject: Blackfin arch: Fix udelay implementation
X-Git-Url: http://pilppa.com/gitweb/?a=commitdiff_plain;h=4e653e04cc1c3553af539ffc81fb56d83d83c733;p=linux-2.6-omap-h63xx.git

Blackfin arch: Fix udelay implementation

Avoid possible overflow during 32*32->32 multiplies.

Reported-by: Marco Reppenhagen <marco.reppenhagen@auerswald.de>
Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>
---

diff --git a/arch/blackfin/include/asm/delay.h b/arch/blackfin/include/asm/delay.h
index 0889c3abb59..47f98c5067b 100644
--- a/arch/blackfin/include/asm/delay.h
+++ b/arch/blackfin/include/asm/delay.h
@@ -47,16 +47,15 @@ static inline void __delay(unsigned long loops)
 #include <linux/param.h>	/* needed for HZ */
 
 /*
- * Use only for very small delays ( < 1 msec).  Should probably use a
- * lookup table, really, as the multiplications take much too long with
- * short delays.  This is a "reasonable" implementation, though (and the
- * first constant multiplications gets optimized away if the delay is
- * a constant)
+ * close approximation borrowed from m68knommu to avoid 64-bit math
  */
+
+#define	HZSCALE		(268435456 / (1000000/HZ))
+
 static inline void udelay(unsigned long usecs)
 {
 	extern unsigned long loops_per_jiffy;
-	__delay(usecs * loops_per_jiffy / (1000000 / HZ));
+	__delay((((usecs * HZSCALE) >> 11) * (loops_per_jiffy >> 11)) >> 6);
 }
 
 #endif