summaryrefslogtreecommitdiffstats
path: root/c/src/lib/libcpu/powerpc/mpc860/vectors/align_h.S
blob: 9a785e347ddc3a3c6babeab8ff2ecb56f1d110c2 (plain) (blame)
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
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
/*  align_h.s	1.1 - 95/12/04
 *
 *  This file contains the assembly code for the PowerPC 403
 *  alignment exception handler for RTEMS.
 *
 *  Based upon IBM provided code with the following release:
 *
 *  This source code has been made available to you by IBM on an AS-IS
 *  basis.  Anyone receiving this source is licensed under IBM
 *  copyrights to use it in any way he or she deems fit, including
 *  copying it, modifying it, compiling it, and redistributing it either
 *  with or without modifications.  No license under IBM patents or
 *  patent applications is to be implied by the copyright license.
 *
 *  Any user of this software should understand that IBM cannot provide
 *  technical support for this software and will not be responsible for
 *  any consequences resulting from the use of this software.
 *
 *  Any person who transfers this source code or any derivative work
 *  must include the IBM copyright notice, this paragraph, and the
 *  preceding two paragraphs in the transferred software.
 *
 *      COPYRIGHT   I B M   CORPORATION 1995
 *      LICENSED MATERIAL  -  PROGRAM PROPERTY OF I B M
 *
 * Modifications:
 *
 *  Author:	Andrew Bray <andy@i-cubed.co.uk>
 *
 *  COPYRIGHT (c) 1995 by i-cubed ltd.
 *
 *  To anyone who acknowledges that this file is provided "AS IS"
 *  without any express or implied warranty:
 *      permission to use, copy, modify, and distribute this file
 *      for any purpose is hereby granted without fee, provided that
 *      the above copyright notice and this notice appears in all
 *      copies, and that the name of i-cubed limited not be used in
 *      advertising or publicity pertaining to distribution of the
 *      software without specific, written prior permission.
 *      i-cubed limited makes no representations about the suitability
 *      of this software for any purpose.
 *
 *  $Id$
 */

#include "asm.h"
/*#include "bsp.h"*/
#define ALIGN_REGS 0x0140

.set    CACHE_SIZE,16           # cache line size of 32 bytes
.set    CACHE_SIZE_L2,4         # cache line size, log 2

.set	Open_gpr0,0
.set	Open_gpr1,4
.set	Open_gpr2,8
.set	Open_gpr3,12
.set	Open_gpr4,16
.set	Open_gpr5,20
.set	Open_gpr6,24
.set	Open_gpr7,28
.set	Open_gpr8,32
.set	Open_gpr9,36
.set	Open_gpr10,40
.set	Open_gpr11,44
.set	Open_gpr12,48
.set	Open_gpr13,52
.set	Open_gpr14,56
.set	Open_gpr15,60
.set	Open_gpr16,64
.set	Open_gpr17,68
.set	Open_gpr18,72
.set	Open_gpr19,76
.set	Open_gpr20,80
.set	Open_gpr21,84
.set	Open_gpr22,88
.set	Open_gpr23,92
.set	Open_gpr24,96
.set	Open_gpr25,100
.set	Open_gpr26,104
.set	Open_gpr27,108
.set	Open_gpr28,112
.set	Open_gpr29,116
.set	Open_gpr30,120
.set	Open_gpr31,124
.set	Open_xer,128
.set	Open_lr,132
.set	Open_ctr,136
.set	Open_cr,140
.set	Open_srr2,144
.set	Open_srr3,148
.set	Open_srr0,152
.set	Open_srr1,156


/*
 *  This code makes several assumptions for processing efficiency
 *  *  General purpose registers are continuous in the image, beginning with
 *     Open_gpr0
 *  *  Hash table is highly dependent on opcodes - opcode changes *will*
 *     require rework of the instruction decode mechanism.
 */

        .text
        .globl  align_h

	.align	CACHE_SIZE_L2
align_h:
        /*-----------------------------------------------------------------------
         * Store GPRs in Open Reg save area
         * Set up r2 as base reg, r1 pointing to Open Reg save area
         *----------------------------------------------------------------------*/
        stmw    r0,ALIGN_REGS(r0)
	li	r1,ALIGN_REGS
        /*-----------------------------------------------------------------------
         * Store special purpose registers in reg save area
         *----------------------------------------------------------------------*/
        mfxer   r7
        mflr    r8
        mfcr    r9
        mfctr   r10
        stw     r7,Open_xer(r1)
        stw     r8,Open_lr(r1)
        stw     r9,Open_cr(r1)
        stw     r10,Open_ctr(r1)
        mfspr	r7, srr2		/* SRR 2 */
        mfspr	r8, srr3		/* SRR 3 */
        mfspr	r9, srr0		/* SRR 0 */
        mfspr	r10, srr1		/* SRR 1 */
        stw     r7,Open_srr2(r1)
        stw     r8,Open_srr3(r1)
        stw     r9,Open_srr0(r1)
        stw     r10,Open_srr1(r1)

/*      Set up common registers */
        mfspr	r5, dear		/* DEAR: R5 is data exception address */
        lwz     r9,Open_srr0(r1)        /* get faulting instruction */
        addi    r7,r9,4                 /* bump instruction */
        stw     r7,Open_srr0(r1)        /* restore to image */
        lwz     r9, 0(r9)               /* retrieve actual instruction */
        rlwinm  r6,r9,18,25,29          /* r6 is RA * 4 field from instruction */
        rlwinm  r7,r9,6,26,31           /* r7 is primary opcode */
        bl      ref_point               /* establish addressibility */
ref_point:
        mflr    r11                     /* r11 is the anchor point for ref_point */
        addi    r10, r7, -31            /* r10 = r7 - 31 */
        rlwinm  r10,r10,2,2,31          /* r10 *= 4 */
        add     r10, r10, r11           /* r10 += anchor point */
        lwz     r10, primary_jt-ref_point(r10)
        mtlr    r10
        rlwinm  r8,r9,13,25,29          /* r8 is RD * 4 */
        la      r7,Open_gpr0(r1)        /* r7 is address of GPR 0 in list */
        blr
primary_jt:
        .long   xform
        .long   lwz
        .long   lwzu
        .long   0
        .long   0
        .long   stw
        .long   stwu
        .long   0
        .long   0
        .long   lhz
        .long   lhzu
        .long   lha
        .long   lhau
        .long   sth
        .long   sthu
        .long   lmw
        .long   stmw
/*
 *   handlers
 */
/*
 * xform instructions require an additional decode.  Fortunately, a relatively
 * simple hash step breaks the instructions out with no collisions
 */
xform:
        rlwinm  r7,r9,31,22,31          /* r7 is secondary opcode */
        rlwinm  r10,r7,27,5,31          /* r10 = r7 >> 5 */
        add     r10,r7,r10              /* r10 = r7 + r10 */
        rlwinm  r10,r10,2,25,29         /* r10 = (r10 & 0x1F) * 4 */
        add     r10,r10,r11             /* r10 += anchor point */
        lwz     r10, secondary_ht-ref_point(r10)
        mtlr    r10
        la      r7,Open_gpr0(r1)        /* r7 is address of GPR 0 in list */
        rlwinm  r8,r9,13,25,29          /* r8 is RD * 4 */
        blrl

secondary_ht:
        .long   lhzux                   /* b 0  0x137 */
        .long   lhax                    /* b 1  0x157 */
        .long   lhaux                   /* b 2  0x177 */
        .long   sthx                    /* b 3  0x197 */
        .long   sthux                   /* b 4  0x1b7 */
        .long   0                       /* b 5 */
        .long   lwbrx                   /* b 6  0x216 */
        .long   0                       /* b 7 */
        .long   0                       /* b 8 */
        .long   0                       /* b 9 */
        .long   stwbrx                  /* b A  0x296 */
        .long   0                       /* b B */
        .long   0                       /* b C */
        .long   0                       /* b D */
        .long   lhbrx                   /* b E   0x316 */
        .long   0                       /* b F */
        .long   0                       /* b 10 */
        .long   0                       /* b 11 */
        .long   sthbrx                  /* b 12  0x396 */
        .long   0                       /* b 13 */
        .long   lwarx                   /* b 14  0x014 */
        .long   dcbz                    /* b 15  0x3f6 */
        .long   0                       /* b 16 */
        .long   lwzx                    /* b 17  0x017 */
        .long   lwzux                   /* b 18  0x037 */
        .long   0                       /* b 19 */
        .long   stwcx                   /* b 1A  0x096 */
        .long   stwx                    /* b 1B  0x097 */
        .long   stwux                   /* b 1C  0x0B7 */
        .long   0                       /* b 1D */
        .long   0                       /* b 1E */
        .long   lhzx                    /* b 1F 0x117 */

/*
 * for all handlers
 *       r4 - Addressability to interrupt context
 *       r5 - DEAR address (faulting data address)
 *       r6 - RA field * 4
 *       r7 - Address of GPR 0 in image
 *       r8 - RD field * 4
 *       r9 - Failing instruction
 */

/*       Load halfword algebraic with update */
lhau:
/*       Load halfword algebraic with update indexed */
lhaux:
        stwx    r5,r7,r6                /* update RA with effective addr */

/*       Load halfword algebraic */
lha:
/*       Load halfword algebraic indexed */
lhax:
        lswi    r10,r5,2                /* load two bytes into r10 */
        srawi   r10,r10,16              /* shift right 2 bytes, extending sign */
        stwx    r10,r7,r8               /* update reg image */
        b       align_complete          /* return */

/*       Load Half Word Byte-Reversed Indexed */
lhbrx:
        lswi    r10,r5,2                /* load two bytes from DEAR into r10 */
        rlwinm  r10,r10,0,0,15          /* mask off lower 2 bytes */
        stwbrx  r10,r7,r8               /* store reversed in reg image */
        b       align_complete          /* return */

/*       Load Half Word and Zero with Update */
lhzu:
/*       Load Half Word and Zero with Update Indexed */
lhzux:
        stwx    r5,r7,r6                /* update RA with effective addr */

/*       Load Half Word and Zero */
lhz:
/*       Load Half Word and Zero Indexed */
lhzx:
        lswi    r10,r5,2                /* load two bytes from DEAR into r10 */
        rlwinm  r10,r10,16,16,31        /* shift right 2 bytes, with zero fill */
        stwx    r10,r7,r8               /* update reg image */
        b       align_complete          /* return */

/*
 *       Load Multiple Word
 */
lmw:
        lwzx    r9,r6,r7                /* R9 contains saved value of RA */
        addi    r10,r7,32*4             /* r10 points to r31 in image  + 4 */
        rlwinm  r8,r8,30,2,31           /* r8 >>= 2  (recovers RT) */
        subfic  r8,r8,32                /* r8 is reg count to load */
        mtctr   r8                      /* load counter */
        addi    r8,r8,-1                /* r8-- */
        rlwinm  r8,r8,2,2,31            /* r8 *= 4 */
        add     r5,r5,r8                /* update DEAR to point to last reg */
lwmloop:
        lswi    r11,r5,4                /* load r11 with 4 bytes from DEAR */
        stwu    r11,-4(r10)             /* load image and decrement pointer */
        addi    r5,r5,-4                /* decrement effective address */
        bdnz    lwmloop
        stwx    r9,r6,r7                /* restore RA (in case it was trashed) */
        b       align_complete          /* return */

/*
 *       Load Word and Reserve Indexed
 */
lwarx:
        lswi    r10,r5,4                /* load four bytes from DEAR into r10 */
        stwx    r10,r7,r8               /* update reg image */
        rlwinm  r5,r5,0,0,29            /* Word align address */
        lwarx   r10,0,r5                /* Set reservation */
        b       align_complete          /* return */

/*
 *       Load Word Byte-Reversed Indexed
 */
lwbrx:
        lswi    r10,r5,4                /* load four bytes from DEAR into r10 */
        stwbrx  r10,r7,r8               /* store reversed in reg image */
        b       align_complete          /* return */

/*       Load Word and Zero with Update */
lwzu:
/*       Load Word and Zero with Update Indexed */
lwzux:
        stwx    r5,r7,r6                /* update RA with effective addr */

/*       Load Word and Zero */
lwz:
/*       Load Word and Zero Indexed */
lwzx:
        lswi    r10,r5,4                /* load four bytes from DEAR into r10 */
        stwx    r10,r7,r8               /* update reg image */
        b       align_complete          /* return */

/*    Store instructions */

/* */
/*       Store Half Word and Update */
sthu:
/*       Store Half Word and Update Indexed */
sthux:
        stwx    r5,r7,r6                /* Update RA with effective address */

/*       Store Half Word */
sth:
/*       Store Half Word Indexed */
sthx:
        lwzx    r10,r8,r7               /* retrieve source register value */
        rlwinm  r10,r10,16,0,15         /* move two bytes to high end of reg */
        stswi   r10,r5,2                /* store bytes to DEAR address */
        b       align_complete          /* return */

/* */
/*       Store Half Word Byte-Reversed Indexed */
sthbrx:
        lwbrx   r10,r8,r7               /* retrieve src reg value byte reversed */
        stswi   r10,r5,2                /* move two bytes to DEAR address */
        b       align_complete          /* return */

/* */
/*       Store Multiple Word */
stmw:
        addi    r10,r7,32*4             /* r10 points to r31 in image  + 4 */
        rlwinm  r8,r8,30,2,31           /* r8 >>= 2  (recovers RT) */
        subfic  r8,r8,32                /* r8 is reg count to load */
        mtctr   r8                      /* load counter */
        addi    r8,r8,-1                /* r8-- */
        rlwinm  r8,r8,2,2,31            /* r8 *= 4 */
        add     r5,r5,r8                /* update DEAR to point to last reg */
stmloop:
        lwzu    r11,-4(r10)             /* get register value */
        stswi   r11,r5,4                /* output to DEAR address */
        addi    r5,r5,-4                /* decrement effective address */
        bdnz    stmloop
        b       align_complete          /* return */

/* */
/*       Store Word and Update */
stwu:
/*       Store Word and Update Indexed */
stwux:
        stwx    r5,r7,r6                /* Update RA with effective address */

/*       Store Word */
stw:
/*       Store Word Indexed */
stwx:
        lwzx    r10,r8,r7               /* retrieve source register value */
        stswi   r10,r5,4                /* store bytes to DEAR address */
        b       align_complete          /* return */

/* */
/*       Store Word Byte-Reversed Indexed */
stwbrx:
        lwbrx   r10,r8,r7               /* retrieve src reg value byte reversed */
        stswi   r10,r5,4                /* move two bytes to DEAR address */
        b       align_complete          /* return */

/* */
/*       Store Word Conditional Indexed */
stwcx:
        rlwinm  r10,r5,0,0,29           /* r10 = word aligned DEAR */
        lwz     r11,0(r10)              /* save original value of store */
        stwcx.  r11,r0,r10              /* attempt store to address */
        bne     stwcx_moveon            /* store failed, move on */
        stw     r11,0(r10)              /* repair damage */
        lwzx    r9,r7,r8                /* get register value */
        stswi   r10,r5,4                /* store bytes to DEAR address */
stwcx_moveon:
        mfcr    r11                     /* get condition reg */
        lwz     r9,Open_cr(r1)          /* get condition reg image */
        rlwimi  r9,r11,0,0,2            /* insert 3 CR bits into cr image */
        lwz     r11,Open_xer(r1)        /* get XER reg */
        rlwimi  r9,r11,29,2,2           /* insert XER SO bit into cr image */
        stw     r9,Open_cr(r1)          /* store cr image */
        b       align_complete          /* return */

/* */
/*       Data Cache Block Zero */
dcbz:
        rlwinm  r5,r5,0,0,31-CACHE_SIZE_L2
                                        /* get address to nearest Cache line */
        addi    r5,r5,-4                /* adjust by a word */
        addi    r10,r0,CACHE_SIZE/4     /* set counter value */
        mtctr   r10
        addi    r11,r0,0                /* r11 = 0 */
dcbz_loop:
        stwu    r11,4(r5)               /* store a word and update EA */
        bdnz    dcbz_loop
        b       align_complete          /* return */

align_complete:
        /*-----------------------------------------------------------------------
         * Restore regs and return from the interrupt
         *----------------------------------------------------------------------*/
        lmw     r24,Open_xer+ALIGN_REGS(r0)
        mtxer   r24
        mtlr    r25
        mtctr   r26
        mtcrf   0xFF, r27
        mtspr	srr2, r28		/* SRR 2 */
        mtspr	srr3, r29		/* SRR 3 */
        mtspr	srr0, r30		/* SRR 0 */
        mtspr	srr1, r31		/* SRR 1 */
        lmw     r1,Open_gpr1+ALIGN_REGS(r0)
        lwz     r0,Open_gpr0+ALIGN_REGS(r0)
        rfi