No register is being used in this code. The constraint is "m", which means
variable is in memory. %0 means, first argument.
Regards
Mohanlal
----- Original Message -----
From: "matrix reloaded" <matrix_reloaded18@xxxxxxxxxxxxxx>
To: <kernelnewbies@xxxxxxxxxxxx>
Sent: Monday, August 09, 2004 7:52 PM
Subject: Atomic Increment Query.
Hi,
I was going through atomic_inc() function for i386 platform and found the
following code :-
/**
99 * atomic_inc - increment atomic variable
100 * @v: pointer of type atomic_t
101 *
102 * Atomically increments @v by 1. Note that the guaranteed
103 * useful range of an atomic_t is only 24 bits.
104 */
105 static __inline__ void atomic_inc(atomic_t *v)
106
107 __asm__
olatile__(
108 LOCK "incl %0"
109 :"=m" (v->counter)
110 :"m" (v->counter));
}
This line of code is somewhat confusing me as according to me there should be
"incl %1" there. But I think that there is the same register being used for input and
output thing, so for this scenario actually %0 and %1 both refer to the same
register. Am I right or something else is there ?
Thanks in advane.
Sumit.
--
Kernelnewbies: Help each other learn about the Linux kernel.
Archive: http://mail.nl.linux.org/kernelnewbies/
FAQ: http://kernelnewbies.org/faq/
