summaryrefslogtreecommitdiffstats
path: root/security/nss/lib/freebl/mpi/mpi_sse2.s
blob: 16a47019c3eafc9ba135069c990ffb57543a5a19 (plain)
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
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
# This Source Code Form is subject to the terms of the Mozilla Public
# License, v. 2.0. If a copy of the MPL was not distributed with this
# file, You can obtain one at http://mozilla.org/MPL/2.0/.

#ifdef DARWIN
#define s_mpv_mul_d          _s_mpv_mul_d
#define s_mpv_mul_d_add      _s_mpv_mul_d_add
#define s_mpv_mul_d_add_prop _s_mpv_mul_d_add_prop
#define s_mpv_sqr_add_prop   _s_mpv_sqr_add_prop
#define s_mpv_div_2dx1d      _s_mpv_div_2dx1d
#define TYPE_FUNCTION(x)
#else
#define TYPE_FUNCTION(x) .type x, @function
#endif

.text

 #  ebp - 8:    caller's esi
 #  ebp - 4:    caller's edi
 #  ebp + 0:    caller's ebp
 #  ebp + 4:    return address
 #  ebp + 8:    a       argument
 #  ebp + 12:   a_len   argument
 #  ebp + 16:   b       argument
 #  ebp + 20:   c       argument
 #  registers:
 #      ebx:
 #      ecx:    a_len
 #      esi:    a ptr
 #      edi:    c ptr
.globl s_mpv_mul_d
.private_extern s_mpv_mul_d
TYPE_FUNCTION(s_mpv_mul_d)
s_mpv_mul_d:
    push   %ebp
    mov    %esp, %ebp
    push   %edi
    push   %esi
    psubq  %mm2, %mm2           # carry = 0
    mov    12(%ebp), %ecx       # ecx = a_len
    movd   16(%ebp), %mm1       # mm1 = b
    mov    20(%ebp), %edi
    cmp    $0, %ecx
    je     2f                   # jmp if a_len == 0
    mov    8(%ebp), %esi        # esi = a
    cld
1:
    movd   0(%esi), %mm0        # mm0 = *a++
    add    $4, %esi
    pmuludq %mm1, %mm0          # mm0 = b * *a++
    paddq  %mm0, %mm2           # add the carry
    movd   %mm2, 0(%edi)        # store the 32bit result
    add    $4, %edi
    psrlq  $32, %mm2            # save the carry
    dec    %ecx                 # --a_len
    jnz    1b                   # jmp if a_len != 0
2:
    movd   %mm2, 0(%edi)        # *c = carry
    emms
    pop    %esi
    pop    %edi
    leave  
    ret    
    nop

 #  ebp - 8:    caller's esi
 #  ebp - 4:    caller's edi
 #  ebp + 0:    caller's ebp
 #  ebp + 4:    return address
 #  ebp + 8:    a       argument
 #  ebp + 12:   a_len   argument
 #  ebp + 16:   b       argument
 #  ebp + 20:   c       argument
 #  registers:
 #      ebx:
 #      ecx:    a_len
 #      esi:    a ptr
 #      edi:    c ptr
.globl s_mpv_mul_d_add
.private_extern s_mpv_mul_d_add
TYPE_FUNCTION(s_mpv_mul_d_add)
s_mpv_mul_d_add:
    push   %ebp
    mov    %esp, %ebp
    push   %edi
    push   %esi
    psubq  %mm2, %mm2           # carry = 0
    mov    12(%ebp), %ecx       # ecx = a_len
    movd   16(%ebp), %mm1       # mm1 = b
    mov    20(%ebp), %edi
    cmp    $0, %ecx
    je     2f                   # jmp if a_len == 0
    mov    8(%ebp), %esi        # esi = a
    cld
1:
    movd   0(%esi), %mm0        # mm0 = *a++
    add    $4, %esi
    pmuludq %mm1, %mm0          # mm0 = b * *a++
    paddq  %mm0, %mm2           # add the carry
    movd   0(%edi), %mm0
    paddq  %mm0, %mm2           # add the carry
    movd   %mm2, 0(%edi)        # store the 32bit result
    add    $4, %edi
    psrlq  $32, %mm2            # save the carry
    dec    %ecx                 # --a_len
    jnz    1b                   # jmp if a_len != 0
2:
    movd   %mm2, 0(%edi)        # *c = carry
    emms
    pop    %esi
    pop    %edi
    leave  
    ret    
    nop

 #  ebp - 12:   caller's ebx
 #  ebp - 8:    caller's esi
 #  ebp - 4:    caller's edi
 #  ebp + 0:    caller's ebp
 #  ebp + 4:    return address
 #  ebp + 8:    a       argument
 #  ebp + 12:   a_len   argument
 #  ebp + 16:   b       argument
 #  ebp + 20:   c       argument
 #  registers:
 #      eax:
 #      ebx:    carry
 #      ecx:    a_len
 #      esi:    a ptr
 #      edi:    c ptr
.globl s_mpv_mul_d_add_prop
.private_extern s_mpv_mul_d_add_prop
TYPE_FUNCTION(s_mpv_mul_d_add_prop)
s_mpv_mul_d_add_prop:
    push   %ebp
    mov    %esp, %ebp
    push   %edi
    push   %esi
    push   %ebx
    psubq  %mm2, %mm2           # carry = 0
    mov    12(%ebp), %ecx       # ecx = a_len
    movd   16(%ebp), %mm1       # mm1 = b
    mov    20(%ebp), %edi
    cmp    $0, %ecx
    je     2f                   # jmp if a_len == 0
    mov    8(%ebp), %esi        # esi = a
    cld
1:
    movd   0(%esi), %mm0        # mm0 = *a++
    movd   0(%edi), %mm3        # fetch the sum
    add    $4, %esi
    pmuludq %mm1, %mm0          # mm0 = b * *a++
    paddq  %mm0, %mm2           # add the carry
    paddq  %mm3, %mm2           # add *c++
    movd   %mm2, 0(%edi)        # store the 32bit result
    add    $4, %edi
    psrlq  $32, %mm2            # save the carry
    dec    %ecx                 # --a_len
    jnz    1b                   # jmp if a_len != 0
2:
    movd   %mm2, %ebx
    cmp    $0, %ebx             # is carry zero?
    jz     4f
    mov    0(%edi), %eax
    add    %ebx, %eax
    stosl
    jnc    4f
3:
    mov    0(%edi), %eax        # add in current word from *c
    adc    $0, %eax
    stosl                       # [es:edi] = ax; edi += 4;
    jc     3b
4:
    emms
    pop    %ebx
    pop    %esi
    pop    %edi
    leave  
    ret    
    nop

 #  ebp - 12:   caller's ebx
 #  ebp - 8:    caller's esi
 #  ebp - 4:    caller's edi
 #  ebp + 0:    caller's ebp
 #  ebp + 4:    return address
 #  ebp + 8:    pa      argument
 #  ebp + 12:   a_len   argument
 #  ebp + 16:   ps      argument
 #  registers:
 #      eax:
 #      ebx:    carry
 #      ecx:    a_len
 #      esi:    a ptr
 #      edi:    c ptr
.globl s_mpv_sqr_add_prop
.private_extern s_mpv_sqr_add_prop
TYPE_FUNCTION(s_mpv_sqr_add_prop)
s_mpv_sqr_add_prop:
    push   %ebp
    mov    %esp, %ebp
    push   %edi
    push   %esi
    push   %ebx
    psubq  %mm2, %mm2           # carry = 0
    mov    12(%ebp), %ecx       # ecx = a_len
    mov    16(%ebp), %edi
    cmp    $0, %ecx
    je     2f                   # jmp if a_len == 0
    mov    8(%ebp), %esi        # esi = a
    cld
1:
    movd   0(%esi), %mm0        # mm0 = *a
    movd   0(%edi), %mm3        # fetch the sum
    add    $4, %esi
    pmuludq %mm0, %mm0          # mm0 = sqr(a)
    paddq  %mm0, %mm2           # add the carry
    paddq  %mm3, %mm2           # add the low word
    movd   4(%edi), %mm3
    movd   %mm2, 0(%edi)        # store the 32bit result
    psrlq  $32, %mm2
    paddq  %mm3, %mm2           # add the high word
    movd   %mm2, 4(%edi)        # store the 32bit result
    psrlq  $32, %mm2            # save the carry.
    add    $8, %edi
    dec    %ecx                 # --a_len
    jnz    1b                   # jmp if a_len != 0
2:
    movd   %mm2, %ebx
    cmp    $0, %ebx             # is carry zero?
    jz     4f
    mov    0(%edi), %eax
    add    %ebx, %eax
    stosl
    jnc    4f
3:
    mov    0(%edi), %eax        # add in current word from *c
    adc    $0, %eax
    stosl                       #  [es:edi] = ax; edi += 4;
    jc     3b
4:
    emms
    pop    %ebx
    pop    %esi
    pop    %edi
    leave  
    ret    
    nop

 #
 # Divide 64-bit (Nhi,Nlo) by 32-bit divisor, which must be normalized
 # so its high bit is 1.   This code is from NSPR.
 #
 # mp_err s_mpv_div_2dx1d(mp_digit Nhi, mp_digit Nlo, mp_digit divisor,
 #                        mp_digit *qp, mp_digit *rp)

 #  esp +  0:   Caller's ebx
 #  esp +  4:   return address
 #  esp +  8:   Nhi     argument
 #  esp + 12:   Nlo     argument
 #  esp + 16:   divisor argument
 #  esp + 20:   qp      argument
 #  esp + 24:   rp      argument
 #  registers:
 #      eax:
 #      ebx:    carry
 #      ecx:    a_len
 #      edx:
 #      esi:    a ptr
 #      edi:    c ptr
 # 
.globl s_mpv_div_2dx1d
.private_extern s_mpv_div_2dx1d
TYPE_FUNCTION(s_mpv_div_2dx1d)
s_mpv_div_2dx1d:
       push   %ebx
       mov    8(%esp), %edx
       mov    12(%esp), %eax
       mov    16(%esp), %ebx
       div    %ebx
       mov    20(%esp), %ebx
       mov    %eax, 0(%ebx)
       mov    24(%esp), %ebx
       mov    %edx, 0(%ebx)
       xor    %eax, %eax        # return zero
       pop    %ebx
       ret    
       nop

#ifndef DARWIN
 # Magic indicating no need for an executable stack
.section .note.GNU-stack, "", @progbits
.previous
#endif