[MIPS] Optimize flow of csum_partial
Delete dead codes at end of the function and move small_csumcopy there. This makes some labels (maybe_end_cruft, small_memcpy, end_bytes, out) needless and eliminates some branches. Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp> Signed-off-by: Ralf Baechle <ralf@linux-mips.org>hifive-unleashed-5.1
Atsushi Nemoto
Ralf Baechle
parent
52ffe760ea
commit
773ff78838
|
|
@ -65,64 +65,6 @@
|
|||
|
||||
.text
|
||||
.set noreorder
|
||||
|
||||
/* unknown src alignment and < 8 bytes to go */
|
||||
small_csumcpy:
|
||||
move a1, t2
|
||||
|
||||
andi t0, a1, 4
|
||||
beqz t0, 1f
|
||||
andi t0, a1, 2
|
||||
|
||||
/* Still a full word to go */
|
||||
ulw t1, (src)
|
||||
PTR_ADDIU src, 4
|
||||
ADDC(sum, t1)
|
||||
|
||||
1: move t1, zero
|
||||
beqz t0, 1f
|
||||
andi t0, a1, 1
|
||||
|
||||
/* Still a halfword to go */
|
||||
ulhu t1, (src)
|
||||
PTR_ADDIU src, 2
|
||||
|
||||
1: beqz t0, 1f
|
||||
sll t1, t1, 16
|
||||
|
||||
lbu t2, (src)
|
||||
nop
|
||||
|
||||
#ifdef __MIPSEB__
|
||||
sll t2, t2, 8
|
||||
#endif
|
||||
or t1, t2
|
||||
|
||||
1: ADDC(sum, t1)
|
||||
|
||||
/* fold checksum */
|
||||
sll v1, sum, 16
|
||||
addu sum, v1
|
||||
sltu v1, sum, v1
|
||||
srl sum, sum, 16
|
||||
addu sum, v1
|
||||
|
||||
/* odd buffer alignment? */
|
||||
beqz t7, 1f
|
||||
nop
|
||||
sll v1, sum, 8
|
||||
srl sum, sum, 8
|
||||
or sum, v1
|
||||
andi sum, 0xffff
|
||||
1:
|
||||
.set reorder
|
||||
/* Add the passed partial csum. */
|
||||
ADDC(sum, a2)
|
||||
jr ra
|
||||
.set noreorder
|
||||
|
||||
/* ------------------------------------------------------------------------- */
|
||||
|
||||
.align 5
|
||||
LEAF(csum_partial)
|
||||
move sum, zero
|
||||
|
|
@ -132,8 +74,7 @@ LEAF(csum_partial)
|
|||
bnez t8, small_csumcpy /* < 8 bytes to copy */
|
||||
move t2, a1
|
||||
|
||||
beqz a1, out
|
||||
andi t7, src, 0x1 /* odd buffer? */
|
||||
andi t7, src, 0x1 /* odd buffer? */
|
||||
|
||||
hword_align:
|
||||
beqz t7, word_align
|
||||
|
|
@ -232,8 +173,9 @@ move_32bytes:
|
|||
PTR_ADDU src, src, 0x20
|
||||
|
||||
do_end_words:
|
||||
beqz t8, maybe_end_cruft
|
||||
LONG_SRL t8, t8, 0x2
|
||||
beqz t8, small_csumcpy
|
||||
andi t2, a1, 0x3
|
||||
LONG_SRL t8, t8, 0x2
|
||||
|
||||
end_words:
|
||||
lw t0, (src)
|
||||
|
|
@ -242,21 +184,58 @@ end_words:
|
|||
bnez t8, end_words
|
||||
PTR_ADDU src, src, 0x4
|
||||
|
||||
maybe_end_cruft:
|
||||
andi t2, a1, 0x3
|
||||
/* unknown src alignment and < 8 bytes to go */
|
||||
small_csumcpy:
|
||||
move a1, t2
|
||||
|
||||
small_memcpy:
|
||||
j small_csumcpy; move a1, t2 /* XXX ??? */
|
||||
beqz t2, out
|
||||
move a1, t2
|
||||
andi t0, a1, 4
|
||||
beqz t0, 1f
|
||||
andi t0, a1, 2
|
||||
|
||||
end_bytes:
|
||||
lb t0, (src)
|
||||
LONG_SUBU a1, a1, 0x1
|
||||
bnez a2, end_bytes
|
||||
PTR_ADDU src, src, 0x1
|
||||
/* Still a full word to go */
|
||||
ulw t1, (src)
|
||||
PTR_ADDIU src, 4
|
||||
ADDC(sum, t1)
|
||||
|
||||
out:
|
||||
1: move t1, zero
|
||||
beqz t0, 1f
|
||||
andi t0, a1, 1
|
||||
|
||||
/* Still a halfword to go */
|
||||
ulhu t1, (src)
|
||||
PTR_ADDIU src, 2
|
||||
|
||||
1: beqz t0, 1f
|
||||
sll t1, t1, 16
|
||||
|
||||
lbu t2, (src)
|
||||
nop
|
||||
|
||||
#ifdef __MIPSEB__
|
||||
sll t2, t2, 8
|
||||
#endif
|
||||
or t1, t2
|
||||
|
||||
1: ADDC(sum, t1)
|
||||
|
||||
/* fold checksum */
|
||||
sll v1, sum, 16
|
||||
addu sum, v1
|
||||
sltu v1, sum, v1
|
||||
srl sum, sum, 16
|
||||
addu sum, v1
|
||||
|
||||
/* odd buffer alignment? */
|
||||
beqz t7, 1f
|
||||
nop
|
||||
sll v1, sum, 8
|
||||
srl sum, sum, 8
|
||||
or sum, v1
|
||||
andi sum, 0xffff
|
||||
1:
|
||||
.set reorder
|
||||
/* Add the passed partial csum. */
|
||||
ADDC(sum, a2)
|
||||
jr ra
|
||||
move v0, sum
|
||||
.set noreorder
|
||||
END(csum_partial)
|
||||
|
|
|
|||
Loading…
Reference in New Issue