diff options
Diffstat (limited to 'libkern/vlop-arm.s')
| -rw-r--r-- | libkern/vlop-arm.s | 34 |
1 files changed, 34 insertions, 0 deletions
diff --git a/libkern/vlop-arm.s b/libkern/vlop-arm.s new file mode 100644 index 00000000..d972ac1b --- /dev/null +++ b/libkern/vlop-arm.s @@ -0,0 +1,34 @@ +#define UMULL(Rs,Rm,Rhi,Rlo,S) WORD $((14<<28)|(4<<21)|(S<<20)|(Rhi<<16)|(Rlo<<12)|(Rs<<8)|(9<<4)|Rm) +#define UMLAL(Rs,Rm,Rhi,Rlo,S) WORD $((14<<28)|(5<<21)|(S<<20)|(Rhi<<16)|(Rlo<<12)|(Rs<<8)|(9<<4)|Rm) +#define MUL(Rs,Rm,Rd,S) WORD $((14<<28)|(0<<21)|(S<<20)|(Rd<<16)|(Rs<<8)|(9<<4)|Rm) +arg=0 + +/* replaced use of R10 by R11 because the former can be the data segment base register */ + +TEXT _mulv(SB), $0 + MOVW 8(FP), R9 /* l0 */ + MOVW 4(FP), R11 /* h0 */ + MOVW 16(FP), R4 /* l1 */ + MOVW 12(FP), R5 /* h1 */ + UMULL(4, 9, 7, 6, 0) + MUL(11, 4, 8, 0) + ADD R8, R7 + MUL(9, 5, 8, 0) + ADD R8, R7 + MOVW R6, 4(R(arg)) + MOVW R7, 0(R(arg)) + RET + +/* multiply, add, and right-shift, yielding a 32-bit result, while + using 64-bit accuracy for the multiply -- for fast fixed-point math */ +TEXT _mularsv(SB), $0 + MOVW 4(FP), R11 /* m1 */ + MOVW 8(FP), R8 /* a */ + MOVW 12(FP), R4 /* rs */ + MOVW $0, R9 + UMLAL(0, 11, 9, 8, 0) + MOVW R8>>R4, R8 + RSB $32, R4, R4 + ORR R9<<R4, R8, R0 + RET + |