kernel-aes67/arch/mips/lib/csum_partial.S
Atsushi Nemoto 773ff78838 [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>
2006-12-09 01:03:59 +00:00

242 lines
4.2 KiB
ArmAsm

/*
* This file is subject to the terms and conditions of the GNU General Public
* License. See the file "COPYING" in the main directory of this archive
* for more details.
*
* Quick'n'dirty IP checksum ...
*
* Copyright (C) 1998, 1999 Ralf Baechle
* Copyright (C) 1999 Silicon Graphics, Inc.
*/
#include <asm/asm.h>
#include <asm/regdef.h>
#ifdef CONFIG_64BIT
/*
* As we are sharing code base with the mips32 tree (which use the o32 ABI
* register definitions). We need to redefine the register definitions from
* the n64 ABI register naming to the o32 ABI register naming.
*/
#undef t0
#undef t1
#undef t2
#undef t3
#define t0 $8
#define t1 $9
#define t2 $10
#define t3 $11
#define t4 $12
#define t5 $13
#define t6 $14
#define t7 $15
#endif
#define ADDC(sum,reg) \
addu sum, reg; \
sltu v1, sum, reg; \
addu sum, v1
#define CSUM_BIGCHUNK(src, offset, sum, _t0, _t1, _t2, _t3) \
lw _t0, (offset + 0x00)(src); \
lw _t1, (offset + 0x04)(src); \
lw _t2, (offset + 0x08)(src); \
lw _t3, (offset + 0x0c)(src); \
ADDC(sum, _t0); \
ADDC(sum, _t1); \
ADDC(sum, _t2); \
ADDC(sum, _t3); \
lw _t0, (offset + 0x10)(src); \
lw _t1, (offset + 0x14)(src); \
lw _t2, (offset + 0x18)(src); \
lw _t3, (offset + 0x1c)(src); \
ADDC(sum, _t0); \
ADDC(sum, _t1); \
ADDC(sum, _t2); \
ADDC(sum, _t3); \
/*
* a0: source address
* a1: length of the area to checksum
* a2: partial checksum
*/
#define src a0
#define sum v0
.text
.set noreorder
.align 5
LEAF(csum_partial)
move sum, zero
move t7, zero
sltiu t8, a1, 0x8
bnez t8, small_csumcpy /* < 8 bytes to copy */
move t2, a1
andi t7, src, 0x1 /* odd buffer? */
hword_align:
beqz t7, word_align
andi t8, src, 0x2
lbu t0, (src)
LONG_SUBU a1, a1, 0x1
#ifdef __MIPSEL__
sll t0, t0, 8
#endif
ADDC(sum, t0)
PTR_ADDU src, src, 0x1
andi t8, src, 0x2
word_align:
beqz t8, dword_align
sltiu t8, a1, 56
lhu t0, (src)
LONG_SUBU a1, a1, 0x2
ADDC(sum, t0)
sltiu t8, a1, 56
PTR_ADDU src, src, 0x2
dword_align:
bnez t8, do_end_words
move t8, a1
andi t8, src, 0x4
beqz t8, qword_align
andi t8, src, 0x8
lw t0, 0x00(src)
LONG_SUBU a1, a1, 0x4
ADDC(sum, t0)
PTR_ADDU src, src, 0x4
andi t8, src, 0x8
qword_align:
beqz t8, oword_align
andi t8, src, 0x10
lw t0, 0x00(src)
lw t1, 0x04(src)
LONG_SUBU a1, a1, 0x8
ADDC(sum, t0)
ADDC(sum, t1)
PTR_ADDU src, src, 0x8
andi t8, src, 0x10
oword_align:
beqz t8, begin_movement
LONG_SRL t8, a1, 0x7
lw t3, 0x08(src)
lw t4, 0x0c(src)
lw t0, 0x00(src)
lw t1, 0x04(src)
ADDC(sum, t3)
ADDC(sum, t4)
ADDC(sum, t0)
ADDC(sum, t1)
LONG_SUBU a1, a1, 0x10
PTR_ADDU src, src, 0x10
LONG_SRL t8, a1, 0x7
begin_movement:
beqz t8, 1f
andi t2, a1, 0x40
move_128bytes:
CSUM_BIGCHUNK(src, 0x00, sum, t0, t1, t3, t4)
CSUM_BIGCHUNK(src, 0x20, sum, t0, t1, t3, t4)
CSUM_BIGCHUNK(src, 0x40, sum, t0, t1, t3, t4)
CSUM_BIGCHUNK(src, 0x60, sum, t0, t1, t3, t4)
LONG_SUBU t8, t8, 0x01
bnez t8, move_128bytes
PTR_ADDU src, src, 0x80
1:
beqz t2, 1f
andi t2, a1, 0x20
move_64bytes:
CSUM_BIGCHUNK(src, 0x00, sum, t0, t1, t3, t4)
CSUM_BIGCHUNK(src, 0x20, sum, t0, t1, t3, t4)
PTR_ADDU src, src, 0x40
1:
beqz t2, do_end_words
andi t8, a1, 0x1c
move_32bytes:
CSUM_BIGCHUNK(src, 0x00, sum, t0, t1, t3, t4)
andi t8, a1, 0x1c
PTR_ADDU src, src, 0x20
do_end_words:
beqz t8, small_csumcpy
andi t2, a1, 0x3
LONG_SRL t8, t8, 0x2
end_words:
lw t0, (src)
LONG_SUBU t8, t8, 0x1
ADDC(sum, t0)
bnez t8, end_words
PTR_ADDU src, src, 0x4
/* 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
END(csum_partial)