3: std r8,8(r3)
beq 3f
addi r3,r3,16
- ld r9,8(r4)
.Ldo_tail:
bf cr7*4+1,1f
- rotldi r9,r9,32
+ lwz r9,8(r4)
+ addi r4,r4,4
stw r9,0(r3)
addi r3,r3,4
1: bf cr7*4+2,2f
- rotldi r9,r9,16
+ lhz r9,8(r4)
+ addi r4,r4,2
sth r9,0(r3)
addi r3,r3,2
2: bf cr7*4+3,3f
- rotldi r9,r9,8
+ lbz r9,8(r4)
stb r9,0(r3)
3: ld r3,48(r1) /* return dest pointer */
blr
cmpwi cr1,r5,8
addi r3,r3,32
sld r9,r9,r10
- ble cr1,.Ldo_tail
+ ble cr1,6f
ld r0,8(r4)
srd r7,r0,r11
or r9,r7,r9
- b .Ldo_tail
+6:
+ bf cr7*4+1,1f
+ rotldi r9,r9,32
+ stw r9,0(r3)
+ addi r3,r3,4
+1: bf cr7*4+2,2f
+ rotldi r9,r9,16
+ sth r9,0(r3)
+ addi r3,r3,2
+2: bf cr7*4+3,3f
+ rotldi r9,r9,8
+ stb r9,0(r3)
+3: ld r3,48(r1) /* return dest pointer */
+ blr
.Ldst_unaligned:
PPC_MTOCRF 0x01,r6 # put #bytes to 8B bdry into cr7