Lines Matching +full:11 +full:a
11 * Delay routines, using a pre-computed "loops_per_jiffy" value.
19 * The 0x4a8e is of course a 'tstl %fp' instruction. This is better
20 * than using a NOP (0x4e71) instruction because it executes in one
22 * for bus cycles to finish. Also fp/a6 isn't likely to cause a
49 * The simpler m68k and ColdFire processors do not have a 32*32->64
50 * multiply instruction. So we need to handle them a little differently.
51 * We use a bit of shifting and a single 32*32->32 multiply to get close.
56 __delay(((((u) * HZSCALE) >> 11) * (loops_per_jiffy >> 11)) >> 6)
71 * The definition of __const_udelay is specifically made a macro so that
73 * the delay is a const.
85 * Use only for very small delays ( < 1 msec). Should probably use a
87 * short delays. This is a "reasonable" implementation, though (and the
89 * a constant)
97 * ((((HZSCALE) >> 11) * (loops_per_jiffy >> 11)) >> 6) is the number of loops
100 * 1000 / ((((HZSCALE) >> 11) * (loops_per_jiffy >> 11)) >> 6) is the number of
103 * So n / ( 1000 / ((((HZSCALE) >> 11) * (loops_per_jiffy >> 11)) >> 6) ) would
108 * The simpler m68k and ColdFire processors do not have a 32*32->64
109 * multiply instruction. So we need to handle them a little differently.
110 * We use a bit of shifting and a single 32*32->32 multiply to get close.
111 * This is a macro so that the const version can factor out the first
119 ((((HZSCALE) >> 11) * in ndelay()
120 (loops_per_jiffy >> 11)) >> 6), in ndelay()