1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
| /* SPDX-License-Identifier: GPL-2.0 */
| /*
| * arch/alpha/lib/csum_ipv6_magic.S
| * Contributed by Richard Henderson <rth@tamu.edu>
| *
| * unsigned short csum_ipv6_magic(struct in6_addr *saddr,
| * struct in6_addr *daddr,
| * __u32 len,
| * unsigned short proto,
| * unsigned int csum);
| *
| * Misalignment handling (which costs 16 instructions / 8 cycles)
| * added by Ivan Kokshaysky <ink@jurassic.park.msu.ru>
| */
|
| #include <asm/export.h>
| .globl csum_ipv6_magic
| .align 4
| .ent csum_ipv6_magic
| .frame $30,0,$26,0
| csum_ipv6_magic:
| .prologue 0
|
| ldq_u $0,0($16) # e0 : load src & dst addr words
| zapnot $20,15,$20 # .. e1 : zero extend incoming csum
| extqh $18,1,$4 # e0 : byte swap len & proto while we wait
| ldq_u $21,7($16) # .. e1 : handle misalignment
|
| extbl $18,1,$5 # e0 :
| ldq_u $1,8($16) # .. e1 :
| extbl $18,2,$6 # e0 :
| ldq_u $22,15($16) # .. e1 :
|
| extbl $18,3,$18 # e0 :
| ldq_u $2,0($17) # .. e1 :
| sra $4,32,$4 # e0 :
| ldq_u $23,7($17) # .. e1 :
|
| extql $0,$16,$0 # e0 :
| ldq_u $3,8($17) # .. e1 :
| extqh $21,$16,$21 # e0 :
| ldq_u $24,15($17) # .. e1 :
|
| sll $5,16,$5 # e0 :
| or $0,$21,$0 # .. e1 : 1st src word complete
| extql $1,$16,$1 # e0 :
| addq $20,$0,$20 # .. e1 : begin summing the words
|
| extqh $22,$16,$22 # e0 :
| cmpult $20,$0,$0 # .. e1 :
| sll $6,8,$6 # e0 :
| or $1,$22,$1 # .. e1 : 2nd src word complete
|
| extql $2,$17,$2 # e0 :
| or $4,$18,$18 # .. e1 :
| extqh $23,$17,$23 # e0 :
| or $5,$6,$5 # .. e1 :
|
| extql $3,$17,$3 # e0 :
| or $2,$23,$2 # .. e1 : 1st dst word complete
| extqh $24,$17,$24 # e0 :
| or $18,$5,$18 # .. e1 : len complete
|
| extwh $19,7,$7 # e0 :
| or $3,$24,$3 # .. e1 : 2nd dst word complete
| extbl $19,1,$19 # e0 :
| addq $20,$1,$20 # .. e1 :
|
| or $19,$7,$19 # e0 :
| cmpult $20,$1,$1 # .. e1 :
| sll $19,48,$19 # e0 :
| nop # .. e0 :
|
| sra $19,32,$19 # e0 : proto complete
| addq $20,$2,$20 # .. e1 :
| cmpult $20,$2,$2 # e0 :
| addq $20,$3,$20 # .. e1 :
|
| cmpult $20,$3,$3 # e0 :
| addq $20,$18,$20 # .. e1 :
| cmpult $20,$18,$18 # e0 :
| addq $20,$19,$20 # .. e1 :
|
| cmpult $20,$19,$19 # e0 :
| addq $0,$1,$0 # .. e1 : merge the carries back into the csum
| addq $2,$3,$2 # e0 :
| addq $18,$19,$18 # .. e1 :
|
| addq $0,$2,$0 # e0 :
| addq $20,$18,$20 # .. e1 :
| addq $0,$20,$0 # e0 :
| unop # :
|
| extwl $0,2,$2 # e0 : begin folding the 64-bit value
| zapnot $0,3,$3 # .. e1 :
| extwl $0,4,$1 # e0 :
| addq $2,$3,$3 # .. e1 :
|
| extwl $0,6,$0 # e0 :
| addq $3,$1,$3 # .. e1 :
| addq $0,$3,$0 # e0 :
| unop # :
|
| extwl $0,2,$1 # e0 : fold 18-bit value
| zapnot $0,3,$0 # .. e1 :
| addq $0,$1,$0 # e0 :
| unop # :
|
| extwl $0,2,$1 # e0 : fold 17-bit value
| zapnot $0,3,$0 # .. e1 :
| addq $0,$1,$0 # e0 :
| not $0,$0 # .. e1 : and complement.
|
| zapnot $0,3,$0 # e0 :
| ret # .. e1 :
|
| .end csum_ipv6_magic
| EXPORT_SYMBOL(csum_ipv6_magic)
|
|
