diff options
-rw-r--r-- | arch/mn10300/lib/do_csum.S | 49 |
1 files changed, 22 insertions, 27 deletions
diff --git a/arch/mn10300/lib/do_csum.S b/arch/mn10300/lib/do_csum.S index e138994e1667..1d27bba0cd8f 100644 --- a/arch/mn10300/lib/do_csum.S +++ b/arch/mn10300/lib/do_csum.S | |||
@@ -10,26 +10,25 @@ | |||
10 | */ | 10 | */ |
11 | #include <asm/cache.h> | 11 | #include <asm/cache.h> |
12 | 12 | ||
13 | .section .text | 13 | .section .text |
14 | .balign L1_CACHE_BYTES | 14 | .balign L1_CACHE_BYTES |
15 | 15 | ||
16 | ############################################################################### | 16 | ############################################################################### |
17 | # | 17 | # |
18 | # unsigned int do_csum(const unsigned char *buff, size_t len) | 18 | # unsigned int do_csum(const unsigned char *buff, int len) |
19 | # | 19 | # |
20 | ############################################################################### | 20 | ############################################################################### |
21 | .globl do_csum | 21 | .globl do_csum |
22 | .type do_csum,@function | 22 | .type do_csum,@function |
23 | do_csum: | 23 | do_csum: |
24 | movm [d2,d3],(sp) | 24 | movm [d2,d3],(sp) |
25 | mov d0,(12,sp) | ||
26 | mov d1,(16,sp) | ||
27 | mov d1,d2 # count | 25 | mov d1,d2 # count |
28 | mov d0,a0 # buff | 26 | mov d0,a0 # buff |
27 | mov a0,a1 | ||
29 | clr d1 # accumulator | 28 | clr d1 # accumulator |
30 | 29 | ||
31 | cmp +0,d2 | 30 | cmp +0,d2 |
32 | beq do_csum_done # return if zero-length buffer | 31 | ble do_csum_done # check for zero length or negative |
33 | 32 | ||
34 | # 4-byte align the buffer pointer | 33 | # 4-byte align the buffer pointer |
35 | btst +3,a0 | 34 | btst +3,a0 |
@@ -41,17 +40,15 @@ do_csum: | |||
41 | inc a0 | 40 | inc a0 |
42 | asl +8,d0 | 41 | asl +8,d0 |
43 | add d0,d1 | 42 | add d0,d1 |
44 | addc +0,d1 | ||
45 | add -1,d2 | 43 | add -1,d2 |
46 | do_csum_addr_not_odd: | ||
47 | 44 | ||
45 | do_csum_addr_not_odd: | ||
48 | cmp +2,d2 | 46 | cmp +2,d2 |
49 | bcs do_csum_fewer_than_4 | 47 | bcs do_csum_fewer_than_4 |
50 | btst +2,a0 | 48 | btst +2,a0 |
51 | beq do_csum_now_4b_aligned | 49 | beq do_csum_now_4b_aligned |
52 | movhu (a0+),d0 | 50 | movhu (a0+),d0 |
53 | add d0,d1 | 51 | add d0,d1 |
54 | addc +0,d1 | ||
55 | add -2,d2 | 52 | add -2,d2 |
56 | cmp +4,d2 | 53 | cmp +4,d2 |
57 | bcs do_csum_fewer_than_4 | 54 | bcs do_csum_fewer_than_4 |
@@ -66,20 +63,20 @@ do_csum_now_4b_aligned: | |||
66 | 63 | ||
67 | do_csum_loop: | 64 | do_csum_loop: |
68 | mov (a0+),d0 | 65 | mov (a0+),d0 |
69 | add d0,d1 | ||
70 | mov (a0+),e0 | 66 | mov (a0+),e0 |
71 | addc e0,d1 | ||
72 | mov (a0+),e1 | 67 | mov (a0+),e1 |
73 | addc e1,d1 | ||
74 | mov (a0+),e3 | 68 | mov (a0+),e3 |
69 | add d0,d1 | ||
70 | addc e0,d1 | ||
71 | addc e1,d1 | ||
75 | addc e3,d1 | 72 | addc e3,d1 |
76 | mov (a0+),d0 | 73 | mov (a0+),d0 |
77 | addc d0,d1 | ||
78 | mov (a0+),e0 | 74 | mov (a0+),e0 |
79 | addc e0,d1 | ||
80 | mov (a0+),e1 | 75 | mov (a0+),e1 |
81 | addc e1,d1 | ||
82 | mov (a0+),e3 | 76 | mov (a0+),e3 |
77 | addc d0,d1 | ||
78 | addc e0,d1 | ||
79 | addc e1,d1 | ||
83 | addc e3,d1 | 80 | addc e3,d1 |
84 | addc +0,d1 | 81 | addc +0,d1 |
85 | 82 | ||
@@ -94,12 +91,12 @@ do_csum_remainder: | |||
94 | cmp +16,d2 | 91 | cmp +16,d2 |
95 | bcs do_csum_fewer_than_16 | 92 | bcs do_csum_fewer_than_16 |
96 | mov (a0+),d0 | 93 | mov (a0+),d0 |
97 | add d0,d1 | ||
98 | mov (a0+),e0 | 94 | mov (a0+),e0 |
99 | addc e0,d1 | ||
100 | mov (a0+),e1 | 95 | mov (a0+),e1 |
101 | addc e1,d1 | ||
102 | mov (a0+),e3 | 96 | mov (a0+),e3 |
97 | add d0,d1 | ||
98 | addc e0,d1 | ||
99 | addc e1,d1 | ||
103 | addc e3,d1 | 100 | addc e3,d1 |
104 | addc +0,d1 | 101 | addc +0,d1 |
105 | add -16,d2 | 102 | add -16,d2 |
@@ -131,9 +128,9 @@ do_csum_fewer_than_4: | |||
131 | xor_cmp d0,d0,+2,d2 | 128 | xor_cmp d0,d0,+2,d2 |
132 | bcs do_csum_fewer_than_2 | 129 | bcs do_csum_fewer_than_2 |
133 | movhu (a0+),d0 | 130 | movhu (a0+),d0 |
134 | do_csum_fewer_than_2: | ||
135 | and +1,d2 | 131 | and +1,d2 |
136 | beq do_csum_add_last_bit | 132 | beq do_csum_add_last_bit |
133 | do_csum_fewer_than_2: | ||
137 | movbu (a0),d3 | 134 | movbu (a0),d3 |
138 | add d3,d0 | 135 | add d3,d0 |
139 | do_csum_add_last_bit: | 136 | do_csum_add_last_bit: |
@@ -142,21 +139,19 @@ do_csum_add_last_bit: | |||
142 | 139 | ||
143 | do_csum_done: | 140 | do_csum_done: |
144 | # compress the checksum down to 16 bits | 141 | # compress the checksum down to 16 bits |
145 | mov +0xffff0000,d2 | 142 | mov +0xffff0000,d0 |
146 | and d1,d2 | 143 | and d1,d0 |
147 | asl +16,d1 | 144 | asl +16,d1 |
148 | add d2,d1,d0 | 145 | add d1,d0 |
149 | addc +0xffff,d0 | 146 | addc +0xffff,d0 |
150 | lsr +16,d0 | 147 | lsr +16,d0 |
151 | 148 | ||
152 | # flip the halves of the word result if the buffer was oddly aligned | 149 | # flip the halves of the word result if the buffer was oddly aligned |
153 | mov (12,sp),d1 | 150 | and +1,a1 |
154 | and +1,d1 | ||
155 | beq do_csum_not_oddly_aligned | 151 | beq do_csum_not_oddly_aligned |
156 | swaph d0,d0 # exchange bits 15:8 with 7:0 | 152 | swaph d0,d0 # exchange bits 15:8 with 7:0 |
157 | 153 | ||
158 | do_csum_not_oddly_aligned: | 154 | do_csum_not_oddly_aligned: |
159 | ret [d2,d3],8 | 155 | ret [d2,d3],8 |
160 | 156 | ||
161 | do_csum_end: | 157 | .size do_csum, .-do_csum |
162 | .size do_csum, do_csum_end-do_csum | ||