summaryrefslogtreecommitdiffstats
path: root/firmware/target/arm/ata-as-arm.S
blob: c1475c0a5a77c8c5835cddbf2427e0ab59db1b51 (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
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
/***************************************************************************
 *             __________               __   ___.
 *   Open      \______   \ ____   ____ |  | _\_ |__   _______  ___
 *   Source     |       _//  _ \_/ ___\|  |/ /| __ \ /  _ \  \/  /
 *   Jukebox    |    |   (  <_> )  \___|    < | \_\ (  <_> > <  <
 *   Firmware   |____|_  /\____/ \___  >__|_ \|___  /\____/__/\_ \
 *                     \/            \/     \/    \/            \/
 * $Id$
 *
 * Copyright (C) 2007 by Jens Arnold
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License
 * as published by the Free Software Foundation; either version 2
 * of the License, or (at your option) any later version.
 *
 * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
 * KIND, either express or implied.
 *
 ****************************************************************************/

#include "config.h"


#if CONFIG_CPU == PP5002
    /* Causes ATA retries on iPod G3 probably related to improper controller
     * setup. Needs investigation. */
    .section    .icode,"ax",%progbits
    .equ    .ata_port, 0xc00031e0
#elif defined CPU_PP502x
    /* Verified working on (PP5020, PP5022) targets */
    .section    .icode,"ax",%progbits
    .equ    .ata_port, 0xc30001e0
#elif CONFIG_CPU == S3C2440
    /* Untested */
    .text
    .equ    .ata_port, 0x18000000
#elif defined(CREATIVE_ZVx) /* Zen Vision could have an other address */
    .text
    .equ    .ata_port, 0x50FEE000
#endif

    .align  2
    .global copy_read_sectors
    .type   copy_read_sectors,%function
    
/* Read a number of words from the ATA data port
 *
 * Optimised for speed; assumes wordcount >= 10
 *
 * Arguments:
 *   r0 - buffer address
 *   r1 - word count
 *
 * Register usage:
 *   r0 - current address
 *   r1 - word count
 *   r2 - ata port
 *   r3..r5, r12, lr - read buffers
 */

copy_read_sectors:
    stmfd   sp!, {r4, r5, lr}
    ldr     r2, =.ata_port
    tst     r0, #1              /* 16 bit aligned? */
    beq     .r_aligned

    /* not 16-bit aligned */
    sub     r1, r1, #1          /* one halfword is handled unconditionally */
    ldrh    r3, [r2]            /* read first halfword */
    strb    r3, [r0], #1        /* store low byte */
    mov     r3, r3, lsr #8

    tst     r0, #2              /* 32 bit aligned? */
    beq     .r_noword_u
    ldrh    r4, [r2]            /* read second halfword */
    orr     r3, r3, r4, lsl #8  /* combine with old byte */
    strh    r3, [r0], #2        /* store */
    mov     r3, r4, lsr #8
    sub     r1, r1, #1          /* another halfword taken */
.r_noword_u:

    sub     r1, r1, #8          /* adjust for zero-check and doing 8 halfwords/loop */
.r_loop_u:
    ldrh    r4, [r2]            /* Read 8 halfwords and combine them into */
    orr     r3, r3, r4, lsl #8  /*  4 words so that they're properly aligned */
    ldrh    r4, [r2]            /*  in memory. Bottom byte of first word is */
    orr     r3, r3, r4, lsl #24 /*  the top byte from the last round. Write */
    mov     r4, r4, lsr #8      /*  all 4 words at once. */
    ldrh    r5, [r2]
    orr     r4, r4, r5, lsl #8
    ldrh    r5, [r2]
    orr     r4, r4, r5, lsl #24
    mov     r5, r5, lsr #8
    ldrh    r12, [r2]
    orr     r5, r5, r12, lsl #8
    ldrh    r12, [r2]
    orr     r5, r5, r12, lsl #24
    mov     r12, r12, lsr #8
    ldrh    lr, [r2]
    orr     r12, r12, lr, lsl #8
    ldrh    lr, [r2]
    orr     r12, r12, lr, lsl #24
    stmia   r0!, {r3, r4, r5, r12}
    mov     r3, lr, lsr #8
    subs    r1, r1, #8          /* 8 or more halfwords left? */
    bge     .r_loop_u

    /* No need to adjust the count, only checking bits from now on. */
    tst     r1, #4              /* 4 or more halfwords left? */
    beq     .r_end4_u
    ldrh    r4, [r2]
    orr     r3, r3, r4, lsl #8
    ldrh    r4, [r2]
    orr     r3, r3, r4, lsl #24
    mov     r4, r4, lsr #8
    ldrh    r5, [r2]
    orr     r4, r4, r5, lsl #8
    ldrh    r5, [r2]
    orr     r4, r4, r5, lsl #24
    stmia   r0!, {r3, r4}
    mov     r3, r5, lsr #8
.r_end4_u:

    tst     r1, #2              /* 2 or more halfwords left? */
    beq     .r_end2_u
    ldrh    r4, [r2]
    orr     r3, r3, r4, lsl #8
    ldrh    r4, [r2]
    orr     r3, r3, r4, lsl #24
    str     r3, [r0], #4
    mov     r3, r4, lsr #8
.r_end2_u:

    tst     r1, #1              /* one halfword left? */
    ldrneh  r4, [r2]
    orrne   r3, r3, r4, lsl #8
    strneh  r3, [r0], #2
    movne   r3, r4, lsr #8

    strb    r3, [r0], #1        /* store final byte */

    ldmfd   sp!, {r4, r5, pc}

    /* 16-bit aligned */
.r_aligned:
    tst     r0, #2              /* 32 bit aligned? */
    ldrneh  r3, [r2]            /* no: read first halfword */
    strneh  r3, [r0], #2        /*     store */
    subne   r1, r1, #1          /*     one halfword taken */

    sub     r1, r1, #8          /* adjust for zero-check and doing 8 halfwords/loop */
.r_loop_a:
    ldrh    r3, [r2]            /* Read 8 halfwords and combine each pair */
    ldrh    r4, [r2]            /*  into a word, then store all at once. */
    orr     r3, r3, r4, lsl #16
    ldrh    r4, [r2]
    ldrh    r5, [r2]
    orr     r4, r4, r5, lsl #16
    ldrh    r5, [r2]
    ldrh    r12, [r2]
    orr     r5, r5, r12, lsl #16
    ldrh    r12, [r2]
    ldrh    lr, [r2]
    orr     r12, r12, lr, lsl #16
    stmia   r0!, {r3, r4, r5, r12}
    subs    r1, r1, #8          /* 8 or more halfwords left? */
    bge     .r_loop_a
    
    /* No need to adjust the count, only checking bits from now on. */
    tst     r1, #4              /* 4 or more halfwords left? */
    beq     .r_end4_a
    ldrh    r3, [r2]
    ldrh    r4, [r2]
    orr     r3, r3, r4, lsl #16
    ldrh    r4, [r2]
    ldrh    r5, [r2]
    orr     r4, r4, r5, lsl #16
    stmia   r0!, {r3, r4}
.r_end4_a:

    tst     r1, #2              /* 2 or more halfwords left? */
    ldrneh  r3, [r2]
    ldrneh  r4, [r2]
    orrne   r3, r3, r4, lsl #16
    strne   r3, [r0], #4

    tst     r1, #1              /* one halfword left? */
    ldrneh  r3, [r2]
    strneh  r3, [r0], #2

    ldmfd   sp!, {r4, r5, pc}

.r_end:
    .size   copy_read_sectors,.r_end-copy_read_sectors

    .align  2
    .global copy_write_sectors
    .type   copy_write_sectors,%function
    
/* Write a number of words to the ATA data port
 *
 * Optimised for speed; assumes wordcount >= 10
 *
 * Arguments:
 *   r0 - buffer address
 *   r1 - word count
 *
 * Register usage:
 *   r0 - current address
 *   r1 - word count
 *   r2 - ata port
 *   r3..r5, r12, lr - read buffers
 */

copy_write_sectors:
    stmfd   sp!, {r4, r5, lr}
    ldr     r2, =.ata_port
    tst     r0, #1              /* 16 bit aligned? */
    beq     .w_aligned
    
    /* not 16-bit aligned */
    sub     r1, r1, #1          /* one halfword is done unconditionally */
    ldrb    r3, [r0], #1        /* load 1st byte, now halfword aligned. */

    tst     r0, #2              /* 32 bit aligned? */
    beq     .w_noword_u
    ldrh    r4, [r0], #2        /* load a halfword */
    orr     r3, r3, r4, lsl #8  /* combine with old byte */
    strh    r3, [r2]            /* write halfword */
    mov     r3, r4, lsr #8
    sub     r1, r1, #1          /* another halfword taken */
.w_noword_u:

    sub     r1, r1, #8          /* adjust for zero-check and doing 8 halfwords/loop */
.w_loop_u:
    ldmia   r0!, {r4, r5, r12, lr}
    orr     r3, r3, r4, lsl #8  /* Load 4 words at once and decompose them */
    strh    r3, [r2]            /*  into 8 halfwords in a way that the words */
    mov     r3, r3, lsr #16     /*  are shifted by 8 bits, putting the high */
    strh    r3, [r2]            /*  byte of one word into the low byte of */
    mov     r4, r4, lsr #24     /*  the next. High byte of last word becomes */
    orr     r4, r4, r5, lsl #8  /*  low byte of next round. */
    strh    r4, [r2]
    mov     r4, r4, lsr #16
    strh    r4, [r2]
    mov     r5, r5, lsr #24
    orr     r5, r5, r12, lsl #8
    strh    r5, [r2]
    mov     r5, r5, lsr #16
    strh    r5, [r2]
    mov     r12, r12, lsr #24
    orr     r12, r12, lr, lsl #8
    strh    r12, [r2]
    mov     r12, r12, lsr #16
    strh    r12, [r2]
    mov     r3, lr, lsr #24
    subs    r1, r1, #8          /* 8 or more halfwords left? */
    bge     .w_loop_u
    
    /* No need to adjust the count, only checking bits from now on. */
    tst     r1, #4              /* 4 or more halfwords left? */
    beq     .w_end4_u
    ldmia   r0!, {r4, r5}
    orr     r3, r3, r4, lsl #8
    strh    r3, [r2]
    mov     r3, r3, lsr #16
    strh    r3, [r2]
    mov     r4, r4, lsr #24
    orr     r4, r4, r5, lsl #8
    strh    r4, [r2]
    mov     r4, r4, lsr #16
    strh    r4, [r2]
    mov     r3, r5, lsr #24
.w_end4_u:

    tst     r1, #2              /* 2 or more halfwords left? */
    beq     .w_end2_u
    ldr     r4, [r0], #4
    orr     r3, r3, r4, lsl #8
    strh    r3, [r2]
    mov     r3, r3, lsr #16
    strh    r3, [r2]
    mov     r3, r4, lsr #24
.w_end2_u:

    tst     r1, #1              /* one halfword left? */
    ldrneh  r4, [r0], #2
    orrne   r3, r3, r4, lsl #8
    strneh  r3, [r2]
    movne   r3, r3, lsr #16
    
    ldrb    r4, [r0], #1        /* load final byte */
    orr     r3, r3, r4, lsl #8
    strh    r3, [r2]            /* write final halfword */

    ldmfd   sp!, {r4, r5, pc}

    /* 16-bit aligned */
.w_aligned:
    tst     r0, #2              /* 32 bit aligned? */
    ldrneh  r3, [r0], #2        /* no: load first halfword */
    strneh  r3, [r2]            /*     write */
    subne   r1, r1, #1          /*     one halfword taken */

    sub     r1, r1, #8          /* adjust for zero-check and doing 8 halfwords/loop */
.w_loop_a:
    ldmia   r0!, {r3, r4, r5, r12}
    strh    r3, [r2]            /* Load 4 words and decompose them into */
    mov     r3, r3, lsr #16     /*  2 halfwords each, and write those. */
    strh    r3, [r2]
    strh    r4, [r2]
    mov     r4, r4, lsr #16
    strh    r4, [r2]
    strh    r5, [r2]
    mov     r5, r5, lsr #16
    strh    r5, [r2]
    strh    r12, [r2]
    mov     r12, r12, lsr #16
    strh    r12, [r2]
    subs    r1, r1, #8          /* 8 or more halfwords left? */
    bge     .w_loop_a

    /* No need to adjust the count, only checking bits from now on. */
    tst     r1, #4              /* 4 or more halfwords left? */
    beq     .w_end4_a
    ldmia   r0!, {r3, r4}
    strh    r3, [r2]
    mov     r3, r3, lsr #16
    strh    r3, [r2]
    strh    r4, [r2]
    mov     r4, r4, lsr #16
    strh    r4, [r2]
.w_end4_a:

    tst     r1, #2              /* 2 or more halfwords left? */
    ldrne   r3, [r0], #4
    strneh  r3, [r2]
    movne   r3, r3, lsr #16
    strneh  r3, [r2]

    tst     r1, #1              /* one halfword left? */
    ldrneh  r3, [r0], #2
    strneh  r3, [r2]

    ldmfd   sp!, {r4, r5, pc}

.w_end:
    .size   copy_write_sectors,.w_end-copy_write_sectors