mov prsave=pr };;
.body
-{ .mib; and r40=3,r32
- mov pr.rot=7<<16
- brp.exit.imp .Le_rounds_cexit,.Le_cexit_insn
- };;
+{ .mmi; and r40=3,r32
+ ADDP r32=0,r32
+ mov pr.rot=7<<16 };;
#if defined(_HPUX_SOURCE) // HPUX is big-endian, cut 15 cycles...
{ .mib; cmp.ne p6,p0=r40,r0
- ADDP r41=4,r32 // 1st arg, borrow teN
+ add r41=4,r32 // 1st arg, borrow teN
(p6) br.dpnt.many .Le_unaligned };;
{ .mmi; ld4 r19=[r32],8
add te2=2048,te0 }
.Le_common:
{ .mib; add te3=3072,te0
- add r35=-3,r35 };;
-
+ add r35=-3,r35
+ brp.exit.imp .Le_rounds_cexit,.Le_cexit_insn
+ };;
{ .mii; mov ar.lc=r35 // borrowed s1
mov ar.ec=3 };;
mov prsave=pr };;
.body
-{ .mib; and r40=3,r32
- mov pr.rot=7<<16
- brp.exit.imp .Ld_rounds_cexit,.Ld_cexit_insn
- };;
+{ .mmi; and r40=3,r32
+ ADDP r32=0,r32
+ mov pr.rot=7<<16 };;
#if defined(_HPUX_SOURCE) // HPUX is big-endian, cut 15 cycles...
{ .mib; cmp.ne p6,p0=r40,r0
- ADDP r41=4,r32 // 1st arg, borrow teN
+ add r41=4,r32 // 1st arg, borrow teN
(p6) br.dpnt.many .Ld_unaligned };;
{ .mmi; ld4 r19=[r32],8
add te2=2048,te0 }
.Ld_common:
{ .mib; add te3=3072,te0
- add r35=-3,r35 };;
-
+ add r35=-3,r35
+ brp.exit.imp .Ld_rounds_cexit,.Ld_cexit_insn
+ };;
{ .mii; mov ar.lc=r35 // borrowed s1
mov ar.ec=3 };;