[POWERPC] Use mtocrf instruction in asm when CONFIG_POWER4_ONLY=y
mtocrf is a faster single-field mtcrf (move to condition register fields) instruction available in POWER4 and later processors. It can make quite a difference in performance on some implementations, so use it for CONFIG_POWER4_ONLY builds. Signed-off-by: Olof Johansson <olof@lixom.net> Signed-off-by: Paul Mackerras <paulus@samba.org>
This commit is contained in:
committed by
Paul Mackerras
parent
569975591c
commit
3467bfd340
@ -12,7 +12,7 @@
|
||||
.align 7
|
||||
_GLOBAL(memcpy)
|
||||
std r3,48(r1) /* save destination pointer for return value */
|
||||
mtcrf 0x01,r5
|
||||
PPC_MTOCRF 0x01,r5
|
||||
cmpldi cr1,r5,16
|
||||
neg r6,r3 # LS 3 bits = # bytes to 8-byte dest bdry
|
||||
andi. r6,r6,7
|
||||
@ -128,7 +128,7 @@ _GLOBAL(memcpy)
|
||||
b .Ldo_tail
|
||||
|
||||
.Ldst_unaligned:
|
||||
mtcrf 0x01,r6 # put #bytes to 8B bdry into cr7
|
||||
PPC_MTOCRF 0x01,r6 # put #bytes to 8B bdry into cr7
|
||||
subf r5,r6,r5
|
||||
li r7,0
|
||||
cmpldi r1,r5,16
|
||||
@ -143,7 +143,7 @@ _GLOBAL(memcpy)
|
||||
2: bf cr7*4+1,3f
|
||||
lwzx r0,r7,r4
|
||||
stwx r0,r7,r3
|
||||
3: mtcrf 0x01,r5
|
||||
3: PPC_MTOCRF 0x01,r5
|
||||
add r4,r6,r4
|
||||
add r3,r6,r3
|
||||
b .Ldst_aligned
|
||||
|
Reference in New Issue
Block a user