summaryrefslogtreecommitdiffstats
path: root/bsps/shared/cache/cacheimpl.h
blob: 16fe59d8040bbd1c147ab96137787f962313e179 (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
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
/*
 *  Cache Manager
 *
 *  Copyright (C) 2014, 2018 embedded brains GmbH
 *
 *  COPYRIGHT (c) 1989-1999.
 *  On-Line Applications Research Corporation (OAR).
 *
 *  The license and distribution terms for this file may be
 *  found in the file LICENSE in this distribution or at
 *  http://www.rtems.org/license/LICENSE.
 */

/*
 * The functions in this file implement the API to the RTEMS Cache Manager.
 * This file is intended to be included in a cache implemention source file
 * provided by the architecture or BSP, e.g.
 *
 *  - bsps/${RTEMS_CPU}/shared/cache/cache.c
 *  - bsps/${RTEMS_CPU}/${RTEMS_BSP_FAMILY}/start/cache.c
 *
 * In this file a couple of defines and inline functions may be provided and
 * afterwards this file is included, e.g.
 *
 *  #define CPU_DATA_CACHE_ALIGNMENT XYZ
 *  ...
 *  #include "../../../bsps/shared/cache/cacheimpl.h"
 *
 * The cache implementation source file shall define
 *
 *  #define CPU_DATA_CACHE_ALIGNMENT <POSITIVE INTEGER>
 *
 * to enable the data cache support.
 *
 * The cache implementation source file shall define
 *
 *  #define CPU_INSTRUCTION_CACHE_ALIGNMENT <POSITIVE INTEGER>
 *
 * to enable the instruction cache support.
 *
 * The cache implementation source file shall define
 *
 *  #define CPU_CACHE_SUPPORT_PROVIDES_RANGE_FUNCTIONS
 *
 * if it provides cache maintenance functions which operate on multiple lines.
 * Otherwise a generic loop with single line operations will be used.  It is
 * strongly recommended to provide the implementation in terms of static inline
 * functions for performance reasons.
 *
 * The cache implementation source file shall define
 *
 *  #define CPU_CACHE_SUPPORT_PROVIDES_CACHE_SIZE_FUNCTIONS
 *
 * if it provides functions to get the data and instruction cache sizes by
 * level.
 *
 * The cache implementation source file shall define
 *
 *  #define CPU_CACHE_SUPPORT_PROVIDES_INSTRUCTION_SYNC_FUNCTION
 *
 * if special instructions must be used to synchronize the instruction caches
 * after a code change.
 *
 * The cache implementation source file shall define
 *
 *  #define CPU_CACHE_SUPPORT_PROVIDES_DISABLE_DATA
 *
 * if an external implementation of rtems_cache_disable_data() is provided,
 * e.g. as an implementation in assembly code.
 *
 * The cache implementation source file shall define
 *
 *  #define CPU_CACHE_NO_INSTRUCTION_CACHE_SNOOPING
 *
 * if the hardware provides no instruction cache snooping and the instruction
 * cache invalidation needs software support.
 *
 * The functions below are implemented with inline routines found in the cache
 * implementation source file for each architecture or BSP.  In the event that
 * not support for a specific function for a cache is provided, the API routine
 * does nothing (but does exist).
 */

#include <rtems.h>

#include <sys/param.h>

#if defined(RTEMS_SMP) && defined(CPU_CACHE_NO_INSTRUCTION_CACHE_SNOOPING)
#include <rtems/score/smpimpl.h>
#include <rtems/score/threaddispatch.h>
#endif

#if CPU_DATA_CACHE_ALIGNMENT > CPU_CACHE_LINE_BYTES
#error "CPU_DATA_CACHE_ALIGNMENT is greater than CPU_CACHE_LINE_BYTES"
#endif

#if CPU_INSTRUCTION_CACHE_ALIGNMENT > CPU_CACHE_LINE_BYTES
#error "CPU_INSTRUCTION_CACHE_ALIGNMENT is greater than CPU_CACHE_LINE_BYTES"
#endif

/*
 * THESE FUNCTIONS ONLY HAVE BODIES IF WE HAVE A DATA CACHE
 */

/*
 * This function is called to flush the data cache by performing cache
 * copybacks. It must determine how many cache lines need to be copied
 * back and then perform the copybacks.
 */
void
rtems_cache_flush_multiple_data_lines( const void * d_addr, size_t n_bytes )
{
#if defined(CPU_DATA_CACHE_ALIGNMENT)
#if defined(CPU_CACHE_SUPPORT_PROVIDES_RANGE_FUNCTIONS)
  _CPU_cache_flush_data_range( d_addr, n_bytes );
#else
  const void * final_address;

 /*
  * Set d_addr to the beginning of the cache line; final_address indicates
  * the last address_t which needs to be pushed. Increment d_addr and push
  * the resulting line until final_address is passed.
  */

  if( n_bytes == 0 )
    /* Do nothing if number of bytes to flush is zero */
    return;

  final_address = (void *)((size_t)d_addr + n_bytes - 1);
  d_addr = (void *)((size_t)d_addr & ~(CPU_DATA_CACHE_ALIGNMENT - 1));
  while( d_addr <= final_address )  {
    _CPU_cache_flush_1_data_line( d_addr );
    d_addr = (void *)((size_t)d_addr + CPU_DATA_CACHE_ALIGNMENT);
  }
#endif
#endif
}

/*
 * This function is responsible for performing a data cache invalidate.
 * It must determine how many cache lines need to be invalidated and then
 * perform the invalidations.
 */
void
rtems_cache_invalidate_multiple_data_lines( const void * d_addr, size_t n_bytes )
{
#if defined(CPU_DATA_CACHE_ALIGNMENT)
#if defined(CPU_CACHE_SUPPORT_PROVIDES_RANGE_FUNCTIONS)
  _CPU_cache_invalidate_data_range( d_addr, n_bytes );
#else
  const void * final_address;

 /*
  * Set d_addr to the beginning of the cache line; final_address indicates
  * the last address_t which needs to be invalidated. Increment d_addr and
  * invalidate the resulting line until final_address is passed.
  */

  if( n_bytes == 0 )
    /* Do nothing if number of bytes to invalidate is zero */
    return;

  final_address = (void *)((size_t)d_addr + n_bytes - 1);
  d_addr = (void *)((size_t)d_addr & ~(CPU_DATA_CACHE_ALIGNMENT - 1));
  while( final_address >= d_addr ) {
    _CPU_cache_invalidate_1_data_line( d_addr );
    d_addr = (void *)((size_t)d_addr + CPU_DATA_CACHE_ALIGNMENT);
  }
#endif
#endif
}

/*
 * This function is responsible for performing a data cache flush.
 * It flushes the entire cache.
 */
void
rtems_cache_flush_entire_data( void )
{
#if defined(CPU_DATA_CACHE_ALIGNMENT)
   /*
    * Call the CPU-specific routine
    */
   _CPU_cache_flush_entire_data();
#endif
}

/*
 * This function is responsible for performing a data cache
 * invalidate. It invalidates the entire cache.
 */
void
rtems_cache_invalidate_entire_data( void )
{
#if defined(CPU_DATA_CACHE_ALIGNMENT)
 /*
  * Call the CPU-specific routine
  */

 _CPU_cache_invalidate_entire_data();
#endif
}

/*
 * This function returns the data cache granularity.
 */
size_t
rtems_cache_get_data_line_size( void )
{
#if defined(CPU_DATA_CACHE_ALIGNMENT)
  return CPU_DATA_CACHE_ALIGNMENT;
#else
  return 0;
#endif
}

size_t
rtems_cache_get_data_cache_size( uint32_t level )
{
#if defined(CPU_CACHE_SUPPORT_PROVIDES_CACHE_SIZE_FUNCTIONS)
  return _CPU_cache_get_data_cache_size( level );
#else
  return 0;
#endif
}

/*
 * This function freezes the data cache; cache lines
 * are not replaced.
 */
void
rtems_cache_freeze_data( void )
{
#if defined(CPU_DATA_CACHE_ALIGNMENT)
  _CPU_cache_freeze_data();
#endif
}

void rtems_cache_unfreeze_data( void )
{
#if defined(CPU_DATA_CACHE_ALIGNMENT)
  _CPU_cache_unfreeze_data();
#endif
}

void
rtems_cache_enable_data( void )
{
#if defined(CPU_DATA_CACHE_ALIGNMENT)
  _CPU_cache_enable_data();
#endif
}

#if !defined(CPU_CACHE_SUPPORT_PROVIDES_DISABLE_DATA)
void
rtems_cache_disable_data( void )
{
#if defined(CPU_DATA_CACHE_ALIGNMENT)
  _CPU_cache_disable_data();
#endif
}
#endif

/*
 * THESE FUNCTIONS ONLY HAVE BODIES IF WE HAVE AN INSTRUCTION CACHE
 */

#if defined(CPU_INSTRUCTION_CACHE_ALIGNMENT) \
  && defined(RTEMS_SMP) \
  && defined(CPU_CACHE_NO_INSTRUCTION_CACHE_SNOOPING)

typedef struct {
  const void *addr;
  size_t size;
} smp_cache_area;

static void smp_cache_inst_inv(void *arg)
{
  smp_cache_area *area = arg;

  _CPU_cache_invalidate_instruction_range(area->addr, area->size);
}

static void smp_cache_inst_inv_all(void *arg)
{
  _CPU_cache_invalidate_entire_instruction();
}

static void smp_cache_broadcast( SMP_Action_handler handler, void *arg )
{
  uint32_t         isr_level;
  Per_CPU_Control *cpu_self;

  isr_level = _ISR_Get_level();

  if ( isr_level == 0 ) {
    cpu_self = _Thread_Dispatch_disable();
  } else {
    cpu_self = _Per_CPU_Get();
  }

  ( *handler )( arg );
  _SMP_Othercast_action( handler, arg );

  if ( isr_level == 0 ) {
    _Thread_Dispatch_enable( cpu_self );
  }
}

#endif

/*
 * This function is responsible for performing an instruction cache
 * invalidate. It must determine how many cache lines need to be invalidated
 * and then perform the invalidations.
 */
#if defined(CPU_INSTRUCTION_CACHE_ALIGNMENT) \
  && !defined(CPU_CACHE_SUPPORT_PROVIDES_RANGE_FUNCTIONS)
static void
_CPU_cache_invalidate_instruction_range(
  const void * i_addr,
  size_t n_bytes
)
{
  const void * final_address;

 /*
  * Set i_addr to the beginning of the cache line; final_address indicates
  * the last address_t which needs to be invalidated. Increment i_addr and
  * invalidate the resulting line until final_address is passed.
  */

  if( n_bytes == 0 )
    /* Do nothing if number of bytes to invalidate is zero */
    return;

  final_address = (void *)((size_t)i_addr + n_bytes - 1);
  i_addr = (void *)((size_t)i_addr & ~(CPU_INSTRUCTION_CACHE_ALIGNMENT - 1));
  while( final_address >= i_addr ) {
    _CPU_cache_invalidate_1_instruction_line( i_addr );
    i_addr = (void *)((size_t)i_addr + CPU_INSTRUCTION_CACHE_ALIGNMENT);
  }
}
#endif

void
rtems_cache_invalidate_multiple_instruction_lines(
  const void * i_addr,
  size_t n_bytes
)
{
#if defined(CPU_INSTRUCTION_CACHE_ALIGNMENT)
#if defined(RTEMS_SMP) && defined(CPU_CACHE_NO_INSTRUCTION_CACHE_SNOOPING)
  smp_cache_area area = { i_addr, n_bytes };

  smp_cache_broadcast( smp_cache_inst_inv, &area );
#else
  _CPU_cache_invalidate_instruction_range( i_addr, n_bytes );
#endif
#endif
}

/*
 * This function is responsible for performing an instruction cache
 * invalidate. It invalidates the entire cache.
 */
void
rtems_cache_invalidate_entire_instruction( void )
{
#if defined(CPU_INSTRUCTION_CACHE_ALIGNMENT)
#if defined(RTEMS_SMP) && defined(CPU_CACHE_NO_INSTRUCTION_CACHE_SNOOPING)
  smp_cache_broadcast( smp_cache_inst_inv_all, NULL );
#else
 _CPU_cache_invalidate_entire_instruction();
#endif
#endif
}

/*
 * This function returns the instruction cache granularity.
 */
size_t
rtems_cache_get_instruction_line_size( void )
{
#if defined(CPU_INSTRUCTION_CACHE_ALIGNMENT)
  return CPU_INSTRUCTION_CACHE_ALIGNMENT;
#else
  return 0;
#endif
}

size_t
rtems_cache_get_instruction_cache_size( uint32_t level )
{
#if defined(CPU_CACHE_SUPPORT_PROVIDES_CACHE_SIZE_FUNCTIONS)
  return _CPU_cache_get_instruction_cache_size( level );
#else
  return 0;
#endif
}

/*
 * This function freezes the instruction cache; cache lines
 * are not replaced.
 */
void
rtems_cache_freeze_instruction( void )
{
#if defined(CPU_INSTRUCTION_CACHE_ALIGNMENT)
  _CPU_cache_freeze_instruction();
#endif
}

void rtems_cache_unfreeze_instruction( void )
{
#if defined(CPU_INSTRUCTION_CACHE_ALIGNMENT)
  _CPU_cache_unfreeze_instruction();
#endif
}

void
rtems_cache_enable_instruction( void )
{
#if defined(CPU_INSTRUCTION_CACHE_ALIGNMENT)
  _CPU_cache_enable_instruction();
#endif
}

void
rtems_cache_disable_instruction( void )
{
#if defined(CPU_INSTRUCTION_CACHE_ALIGNMENT)
  _CPU_cache_disable_instruction();
#endif
}

/* Returns the maximal cache line size of all cache kinds in bytes. */
size_t rtems_cache_get_maximal_line_size( void )
{
#if defined(CPU_MAXIMAL_CACHE_ALIGNMENT)
  return CPU_MAXIMAL_CACHE_ALIGNMENT;
#endif
  size_t data_line_size =
#if defined(CPU_DATA_CACHE_ALIGNMENT)
    CPU_DATA_CACHE_ALIGNMENT;
#else
    0;
#endif
  size_t instruction_line_size =
#if defined(CPU_INSTRUCTION_CACHE_ALIGNMENT)
    CPU_INSTRUCTION_CACHE_ALIGNMENT;
#else
    0;
#endif
  return MAX( data_line_size, instruction_line_size );
}

/*
 * Purpose is to synchronize caches after code has been loaded
 * or self modified. Actual implementation is simple only
 * but it can and should be repaced by optimized version
 * which does not need flush and invalidate all cache levels
 * when code is changed.
 */
void rtems_cache_instruction_sync_after_code_change(
  const void *code_addr,
  size_t      n_bytes
)
{
#if defined(CPU_CACHE_SUPPORT_PROVIDES_INSTRUCTION_SYNC_FUNCTION)
  _CPU_cache_instruction_sync_after_code_change( code_addr, n_bytes );
#else
  rtems_cache_flush_multiple_data_lines( code_addr, n_bytes );
  rtems_cache_invalidate_multiple_instruction_lines( code_addr, n_bytes );
#endif
}