patch-1.3.94 linux/include/asm-m68k/delay.h
Next file: linux/include/asm-m68k/dma.h
Previous file: linux/include/asm-m68k/checksum.h
Back to the patch index
Back to the overall index
- Lines: 48
- Date:
Sat Feb 24 19:42:43 1996
- Orig file:
v1.3.93/linux/include/asm-m68k/delay.h
- Orig date:
Mon Nov 28 20:41:29 1994
diff -u --recursive --new-file v1.3.93/linux/include/asm-m68k/delay.h linux/include/asm-m68k/delay.h
@@ -9,16 +9,13 @@
extern __inline__ void __delay(int loops)
{
- __asm__("\n\tmovel %0,d0\n1:\tsubql #1,d0\n\tbpls 1b\n"
+ __asm__("\n\tmovel %0,%/d0\n1:\tsubql #1,%/d0\n\tbpls 1b\n"
: /* no outputs */
: "g" (loops)
: "d0");
}
/*
- * division by multiplication: you don't have to worry about
- * loss of precision.
- *
* 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
@@ -27,14 +24,24 @@
*/
extern __inline__ void udelay(unsigned long usecs)
{
- asm ("mulul %1,d0,%0\n\t"
- "divul %2,d0,%0"
+ usecs *= 0x000010c6; /* 2**32 / 1000000 */
+
+ asm ("mulul %1,%0:%2"
: "=d" (usecs)
: "d" (usecs),
- "i" (1000000),
- "0" (loops_per_sec)
- : "d0");
+ "d" (loops_per_sec));
__delay(usecs);
+}
+
+extern __inline__ unsigned long muldiv(unsigned long a, unsigned long b, unsigned long c)
+{
+ __asm__("mulul %1,%/d0:%0\n\tdivul %2,%/d0:%0"
+ :"=d" (a)
+ :"d" (b),
+ "d" (c),
+ "0" (a)
+ :"d0");
+ return a;
}
#endif /* defined(_M68K_DELAY_H) */
FUNET's LINUX-ADM group, [email protected]
TCL-scripts by Sam Shen, [email protected]
with Sam's (original) version of this