diff options
Diffstat (limited to 'sysdeps/powerpc/memset.S')
| -rw-r--r-- | sysdeps/powerpc/memset.S | 208 |
1 files changed, 104 insertions, 104 deletions
diff --git a/sysdeps/powerpc/memset.S b/sysdeps/powerpc/memset.S index 6ac32ddc99..63c4c14c64 100644 --- a/sysdeps/powerpc/memset.S +++ b/sysdeps/powerpc/memset.S @@ -1,5 +1,5 @@ /* Optimized memset implementation for PowerPC. - Copyright (C) 1997 Free Software Foundation, Inc. + Copyright (C) 1997, 1999 Free Software Foundation, Inc. This file is part of the GNU C Library. The GNU C Library is free software; you can redistribute it and/or @@ -30,170 +30,170 @@ EALIGN(memset,5,1) r7: number of bytes we are setting now (when aligning) */ /* take care of case for size <= 4 */ - cmplwi %cr1,%r5,4 - andi. %r7,%r3,3 - mr %r6,%r3 - ble- %cr1,L(small) + cmplwi cr1,r5,4 + andi. r7,r3,3 + mr r6,r3 + ble- cr1,L(small) /* align to word boundary */ - cmplwi %cr5,%r5,31 - rlwimi %r4,%r4,8,16,23 + cmplwi cr5,r5,31 + rlwimi r4,r4,8,16,23 beq+ L(aligned) # 8th instruction from .align - mtcrf 0x01,%r3 - subfic %r7,%r7,4 - add %r6,%r6,%r7 - sub %r5,%r5,%r7 + mtcrf 0x01,r3 + subfic r7,r7,4 + add r6,r6,r7 + sub r5,r5,r7 bf+ 31,0f - stb %r4,0(%r3) + stb r4,0(r3) bt 30,L(aligned) -0: sth %r4,-2(%r6) # 16th instruction from .align +0: sth r4,-2(r6) # 16th instruction from .align /* take care of case for size < 31 */ L(aligned): - mtcrf 0x01,%r5 - rlwimi %r4,%r4,16,0,15 - ble %cr5,L(medium) + mtcrf 0x01,r5 + rlwimi r4,r4,16,0,15 + ble cr5,L(medium) /* align to cache line boundary... */ - andi. %r7,%r6,0x1C - subfic %r7,%r7,0x20 + andi. r7,r6,0x1C + subfic r7,r7,0x20 beq L(caligned) - mtcrf 0x01,%r7 - add %r6,%r6,%r7 - sub %r5,%r5,%r7 - cmplwi %cr1,%r7,0x10 - mr %r8,%r6 + mtcrf 0x01,r7 + add r6,r6,r7 + sub r5,r5,r7 + cmplwi cr1,r7,0x10 + mr r8,r6 bf 28,1f - stw %r4,-4(%r8) - stwu %r4,-8(%r8) -1: blt %cr1,2f - stw %r4,-4(%r8) # 32nd instruction from .align - stw %r4,-8(%r8) - stw %r4,-12(%r8) - stwu %r4,-16(%r8) + stw r4,-4(r8) + stwu r4,-8(r8) +1: blt cr1,2f + stw r4,-4(r8) # 32nd instruction from .align + stw r4,-8(r8) + stw r4,-12(r8) + stwu r4,-16(r8) 2: bf 29,L(caligned) - stw %r4,-4(%r8) + stw r4,-4(r8) /* now aligned to a cache line. */ L(caligned): - cmplwi %cr1,%r4,0 - clrrwi. %r7,%r5,5 - mtcrf 0x01,%r5 # 40th instruction from .align - beq %cr1,L(zloopstart) # special case for clearing memory using dcbz - srwi %r0,%r7,5 - mtctr %r0 + cmplwi cr1,r4,0 + clrrwi. r7,r5,5 + mtcrf 0x01,r5 # 40th instruction from .align + beq cr1,L(zloopstart) # special case for clearing memory using dcbz + srwi r0,r7,5 + mtctr r0 beq L(medium) # we may not actually get to do a full line - clrlwi. %r5,%r5,27 - add %r6,%r6,%r7 -0: li %r8,-0x40 + clrlwi. r5,r5,27 + add r6,r6,r7 +0: li r8,-0x40 bdz L(cloopdone) # 48th instruction from .align -3: dcbz %r8,%r6 - stw %r4,-4(%r6) - stw %r4,-8(%r6) - stw %r4,-12(%r6) - stw %r4,-16(%r6) +3: dcbz r8,r6 + stw r4,-4(r6) + stw r4,-8(r6) + stw r4,-12(r6) + stw r4,-16(r6) nop # let 601 fetch last 4 instructions of loop - stw %r4,-20(%r6) - stw %r4,-24(%r6) # 56th instruction from .align + stw r4,-20(r6) + stw r4,-24(r6) # 56th instruction from .align nop # let 601 fetch first 8 instructions of loop - stw %r4,-28(%r6) - stwu %r4,-32(%r6) + stw r4,-28(r6) + stwu r4,-32(r6) bdnz 3b L(cloopdone): - stw %r4,-4(%r6) - stw %r4,-8(%r6) - stw %r4,-12(%r6) - stw %r4,-16(%r6) # 64th instruction from .align - stw %r4,-20(%r6) - cmplwi %cr1,%r5,16 - stw %r4,-24(%r6) - stw %r4,-28(%r6) - stwu %r4,-32(%r6) + stw r4,-4(r6) + stw r4,-8(r6) + stw r4,-12(r6) + stw r4,-16(r6) # 64th instruction from .align + stw r4,-20(r6) + cmplwi cr1,r5,16 + stw r4,-24(r6) + stw r4,-28(r6) + stwu r4,-32(r6) beqlr - add %r6,%r6,%r7 + add r6,r6,r7 b L(medium_tail2) # 72nd instruction from .align .align 5 nop /* Clear lines of memory in 128-byte chunks. */ L(zloopstart): - clrlwi %r5,%r5,27 - mtcrf 0x02,%r7 - srwi. %r0,%r7,7 - mtctr %r0 - li %r7,0x20 - li %r8,-0x40 - cmplwi %cr1,%r5,16 # 8 + clrlwi r5,r5,27 + mtcrf 0x02,r7 + srwi. r0,r7,7 + mtctr r0 + li r7,0x20 + li r8,-0x40 + cmplwi cr1,r5,16 # 8 bf 26,0f - dcbz 0,%r6 - addi %r6,%r6,0x20 -0: li %r9,-0x20 + dcbz 0,r6 + addi r6,r6,0x20 +0: li r9,-0x20 bf 25,1f - dcbz 0,%r6 - dcbz %r7,%r6 - addi %r6,%r6,0x40 # 16 -1: cmplwi %cr5,%r5,0 + dcbz 0,r6 + dcbz r7,r6 + addi r6,r6,0x40 # 16 +1: cmplwi cr5,r5,0 beq L(medium) L(zloop): - dcbz 0,%r6 - dcbz %r7,%r6 - addi %r6,%r6,0x80 - dcbz %r8,%r6 - dcbz %r9,%r6 + dcbz 0,r6 + dcbz r7,r6 + addi r6,r6,0x80 + dcbz r8,r6 + dcbz r9,r6 bdnz L(zloop) - beqlr %cr5 + beqlr cr5 b L(medium_tail2) .align 5 L(small): /* Memset of 4 bytes or less. */ - cmplwi %cr5,%r5,1 - cmplwi %cr1,%r5,3 - bltlr %cr5 - stb %r4,0(%r6) - beqlr %cr5 + cmplwi cr5,r5,1 + cmplwi cr1,r5,3 + bltlr cr5 + stb r4,0(r6) + beqlr cr5 nop - stb %r4,1(%r6) - bltlr %cr1 - stb %r4,2(%r6) - beqlr %cr1 + stb r4,1(r6) + bltlr cr1 + stb r4,2(r6) + beqlr cr1 nop - stb %r4,3(%r6) + stb r4,3(r6) blr /* Memset of 0-31 bytes. */ .align 5 L(medium): - cmplwi %cr1,%r5,16 + cmplwi cr1,r5,16 L(medium_tail2): - add %r6,%r6,%r5 + add r6,r6,r5 L(medium_tail): bt- 31,L(medium_31t) bt- 30,L(medium_30t) L(medium_30f): bt- 29,L(medium_29t) L(medium_29f): - bge- %cr1,L(medium_27t) + bge- cr1,L(medium_27t) bflr- 28 - stw %r4,-4(%r6) # 8th instruction from .align - stw %r4,-8(%r6) + stw r4,-4(r6) # 8th instruction from .align + stw r4,-8(r6) blr L(medium_31t): - stbu %r4,-1(%r6) + stbu r4,-1(r6) bf- 30,L(medium_30f) L(medium_30t): - sthu %r4,-2(%r6) + sthu r4,-2(r6) bf- 29,L(medium_29f) L(medium_29t): - stwu %r4,-4(%r6) - blt- %cr1,L(medium_27f) # 16th instruction from .align + stwu r4,-4(r6) + blt- cr1,L(medium_27f) # 16th instruction from .align L(medium_27t): - stw %r4,-4(%r6) - stw %r4,-8(%r6) - stw %r4,-12(%r6) - stwu %r4,-16(%r6) + stw r4,-4(r6) + stw r4,-8(r6) + stw r4,-12(r6) + stwu r4,-16(r6) L(medium_27f): bflr- 28 L(medium_28t): - stw %r4,-4(%r6) - stw %r4,-8(%r6) + stw r4,-4(r6) + stw r4,-8(r6) blr END(memset) |