Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit 4e653e04 authored by Michael Hennerich's avatar Michael Hennerich Committed by Bryan Wu
Browse files

Blackfin arch: Fix udelay implementation 



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

Reported-by: default avatarMarco Reppenhagen <marco.reppenhagen@auerswald.de>
Signed-off-by: default avatarMichael Hennerich <michael.hennerich@analog.com>
Signed-off-by: default avatarBryan Wu <cooloney@kernel.org>
parent 972de7d9

arch/blackfin/include/asm/delay.h

+5 −6
Original line number Diff line number Diff line
@@ -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