aboutsummaryrefslogtreecommitdiff
path: root/arch/arc/include/asm/atomic.h
blob: c614857eb20917cd9a89b30b9a8f2436200c849d (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
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
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
/* SPDX-License-Identifier: GPL-2.0-only */
/*
 * Copyright (C) 2004, 2007-2010, 2011-2012 Synopsys, Inc. (www.synopsys.com)
 */

#ifndef _ASM_ARC_ATOMIC_H
#define _ASM_ARC_ATOMIC_H

#ifndef __ASSEMBLY__

#include <linux/types.h>
#include <linux/compiler.h>
#include <asm/cmpxchg.h>
#include <asm/barrier.h>
#include <asm/smp.h>

#ifndef CONFIG_ARC_PLAT_EZNPS

#define atomic_read(v)  READ_ONCE((v)->counter)

#ifdef CONFIG_ARC_HAS_LLSC

#define atomic_set(v, i) WRITE_ONCE(((v)->counter), (i))

#define ATOMIC_OP(op, c_op, asm_op)					\
static inline void atomic_##op(int i, atomic_t *v)			\
{									\
	unsigned int val;						\
									\
	__asm__ __volatile__(						\
	"1:	llock   %[val], [%[ctr]]		\n"		\
	"	" #asm_op " %[val], %[val], %[i]	\n"		\
	"	scond   %[val], [%[ctr]]		\n"		\
	"	bnz     1b				\n"		\
	: [val]	"=&r"	(val) /* Early clobber to prevent reg reuse */	\
	: [ctr]	"r"	(&v->counter), /* Not "m": llock only supports reg direct addr mode */	\
	  [i]	"ir"	(i)						\
	: "cc");							\
}									\

#define ATOMIC_OP_RETURN(op, c_op, asm_op)				\
static inline int atomic_##op##_return(int i, atomic_t *v)		\
{									\
	unsigned int val;						\
									\
	/*								\
	 * Explicit full memory barrier needed before/after as		\
	 * LLOCK/SCOND thmeselves don't provide any such semantics	\
	 */								\
	smp_mb();							\
									\
	__asm__ __volatile__(						\
	"1:	llock   %[val], [%[ctr]]		\n"		\
	"	" #asm_op " %[val], %[val], %[i]	\n"		\
	"	scond   %[val], [%[ctr]]		\n"		\
	"	bnz     1b				\n"		\
	: [val]	"=&r"	(val)						\
	: [ctr]	"r"	(&v->counter),					\
	  [i]	"ir"	(i)						\
	: "cc");							\
									\
	smp_mb();							\
									\
	return val;							\
}

#define ATOMIC_FETCH_OP(op, c_op, asm_op)				\
static inline int atomic_fetch_##op(int i, atomic_t *v)			\
{									\
	unsigned int val, orig;						\
									\
	/*								\
	 * Explicit full memory barrier needed before/after as		\
	 * LLOCK/SCOND thmeselves don't provide any such semantics	\
	 */								\
	smp_mb();							\
									\
	__asm__ __volatile__(						\
	"1:	llock   %[orig], [%[ctr]]		\n"		\
	"	" #asm_op " %[val], %[orig], %[i]	\n"		\
	"	scond   %[val], [%[ctr]]		\n"		\
	"	bnz     1b				\n"		\
	: [val]	"=&r"	(val),						\
	  [orig] "=&r" (orig)						\
	: [ctr]	"r"	(&v->counter),					\
	  [i]	"ir"	(i)						\
	: "cc");							\
									\
	smp_mb();							\
									\
	return orig;							\
}

#else	/* !CONFIG_ARC_HAS_LLSC */

#ifndef CONFIG_SMP

 /* violating atomic_xxx API locking protocol in UP for optimization sake */
#define atomic_set(v, i) WRITE_ONCE(((v)->counter), (i))

#else

static inline void atomic_set(atomic_t *v, int i)
{
	/*
	 * Independent of hardware support, all of the atomic_xxx() APIs need
	 * to follow the same locking rules to make sure that a "hardware"
	 * atomic insn (e.g. LD) doesn't clobber an "emulated" atomic insn
	 * sequence
	 *
	 * Thus atomic_set() despite being 1 insn (and seemingly atomic)
	 * requires the locking.
	 */
	unsigned long flags;

	atomic_ops_lock(flags);
	WRITE_ONCE(v->counter, i);
	atomic_ops_unlock(flags);
}

#define atomic_set_release(v, i)	atomic_set((v), (i))

#endif

/*
 * Non hardware assisted Atomic-R-M-W
 * Locking would change to irq-disabling only (UP) and spinlocks (SMP)
 */

#define ATOMIC_OP(op, c_op, asm_op)					\
static inline void atomic_##op(int i, atomic_t *v)			\
{									\
	unsigned long flags;						\
									\
	atomic_ops_lock(flags);						\
	v->counter c_op i;						\
	atomic_ops_unlock(flags);					\
}

#define ATOMIC_OP_RETURN(op, c_op, asm_op)				\
static inline int atomic_##op##_return(int i, atomic_t *v)		\
{									\
	unsigned long flags;						\
	unsigned long temp;						\
									\
	/*								\
	 * spin lock/unlock provides the needed smp_mb() before/after	\
	 */								\
	atomic_ops_lock(flags);						\
	temp = v->counter;						\
	temp c_op i;							\
	v->counter = temp;						\
	atomic_ops_unlock(flags);					\
									\
	return temp;							\
}

#define ATOMIC_FETCH_OP(op, c_op, asm_op)				\
static inline int atomic_fetch_##op(int i, atomic_t *v)			\
{									\
	unsigned long flags;						\
	unsigned long orig;						\
									\
	/*								\
	 * spin lock/unlock provides the needed smp_mb() before/after	\
	 */								\
	atomic_ops_lock(flags);						\
	orig = v->counter;						\
	v->counter c_op i;						\
	atomic_ops_unlock(flags);					\
									\
	return orig;							\
}

#endif /* !CONFIG_ARC_HAS_LLSC */

#define ATOMIC_OPS(op, c_op, asm_op)					\
	ATOMIC_OP(op, c_op, asm_op)					\
	ATOMIC_OP_RETURN(op, c_op, asm_op)				\
	ATOMIC_FETCH_OP(op, c_op, asm_op)

ATOMIC_OPS(add, +=, add)
ATOMIC_OPS(sub, -=, sub)

#define atomic_andnot		atomic_andnot
#define atomic_fetch_andnot	atomic_fetch_andnot

#undef ATOMIC_OPS
#define ATOMIC_OPS(op, c_op, asm_op)					\
	ATOMIC_OP(op, c_op, asm_op)					\
	ATOMIC_FETCH_OP(op, c_op, asm_op)

ATOMIC_OPS(and, &=, and)
ATOMIC_OPS(andnot, &= ~, bic)
ATOMIC_OPS(or, |=, or)
ATOMIC_OPS(xor, ^=, xor)

#else /* CONFIG_ARC_PLAT_EZNPS */

static inline int atomic_read(const atomic_t *v)
{
	int temp;

	__asm__ __volatile__(
	"	ld.di %0, [%1]"
	: "=r"(temp)
	: "r"(&v->counter)
	: "memory");
	return temp;
}

static inline void atomic_set(atomic_t *v, int i)
{
	__asm__ __volatile__(
	"	st.di %0,[%1]"
	:
	: "r"(i), "r"(&v->counter)
	: "memory");
}

#define ATOMIC_OP(op, c_op, asm_op)					\
static inline void atomic_##op(int i, atomic_t *v)			\
{									\
	__asm__ __volatile__(						\
	"	mov r2, %0\n"						\
	"	mov r3, %1\n"						\
	"       .word %2\n"						\
	:								\
	: "r"(i), "r"(&v->counter), "i"(asm_op)				\
	: "r2", "r3", "memory");					\
}									\

#define ATOMIC_OP_RETURN(op, c_op, asm_op)				\
static inline int atomic_##op##_return(int i, atomic_t *v)		\
{									\
	unsigned int temp = i;						\
									\
	/* Explicit full memory barrier needed before/after */		\
	smp_mb();							\
									\
	__asm__ __volatile__(						\
	"	mov r2, %0\n"						\
	"	mov r3, %1\n"						\
	"       .word %2\n"						\
	"	mov %0, r2"						\
	: "+r"(temp)							\
	: "r"(&v->counter), "i"(asm_op)					\
	: "r2", "r3", "memory");					\
									\
	smp_mb();							\
									\
	temp c_op i;							\
									\
	return temp;							\
}

#define ATOMIC_FETCH_OP(op, c_op, asm_op)				\
static inline int atomic_fetch_##op(int i, atomic_t *v)			\
{									\
	unsigned int temp = i;						\
									\
	/* Explicit full memory barrier needed before/after */		\
	smp_mb();							\
									\
	__asm__ __volatile__(						\
	"	mov r2, %0\n"						\
	"	mov r3, %1\n"						\
	"       .word %2\n"						\
	"	mov %0, r2"						\
	: "+r"(temp)							\
	: "r"(&v->counter), "i"(asm_op)					\
	: "r2", "r3", "memory");					\
									\
	smp_mb();							\
									\
	return temp;							\
}

#define ATOMIC_OPS(op, c_op, asm_op)					\
	ATOMIC_OP(op, c_op, asm_op)					\
	ATOMIC_OP_RETURN(op, c_op, asm_op)				\
	ATOMIC_FETCH_OP(op, c_op, asm_op)

ATOMIC_OPS(add, +=, CTOP_INST_AADD_DI_R2_R2_R3)
#define atomic_sub(i, v) atomic_add(-(i), (v))
#define atomic_sub_return(i, v) atomic_add_return(-(i), (v))
#define atomic_fetch_sub(i, v) atomic_fetch_add(-(i), (v))

#undef ATOMIC_OPS
#define ATOMIC_OPS(op, c_op, asm_op)					\
	ATOMIC_OP(op, c_op, asm_op)					\
	ATOMIC_FETCH_OP(op, c_op, asm_op)

ATOMIC_OPS(and, &=, CTOP_INST_AAND_DI_R2_R2_R3)
ATOMIC_OPS(or, |=, CTOP_INST_AOR_DI_R2_R2_R3)
ATOMIC_OPS(xor, ^=, CTOP_INST_AXOR_DI_R2_R2_R3)

#endif /* CONFIG_ARC_PLAT_EZNPS */

#undef ATOMIC_OPS
#undef ATOMIC_FETCH_OP
#undef ATOMIC_OP_RETURN
#undef ATOMIC_OP

#ifdef CONFIG_GENERIC_ATOMIC64

#include <asm-generic/atomic64.h>

#else	/* Kconfig ensures this is only enabled with needed h/w assist */

/*
 * ARCv2 supports 64-bit exclusive load (LLOCKD) / store (SCONDD)
 *  - The address HAS to be 64-bit aligned
 *  - There are 2 semantics involved here:
 *    = exclusive implies no interim update between load/store to same addr
 *    = both words are observed/updated together: this is guaranteed even
 *      for regular 64-bit load (LDD) / store (STD). Thus atomic64_set()
 *      is NOT required to use LLOCKD+SCONDD, STD suffices
 */

typedef struct {
	s64 __aligned(8) counter;
} atomic64_t;

#define ATOMIC64_INIT(a) { (a) }

static inline s64 atomic64_read(const atomic64_t *v)
{
	s64 val;

	__asm__ __volatile__(
	"	ldd   %0, [%1]	\n"
	: "=r"(val)
	: "r"(&v->counter));

	return val;
}

static inline void atomic64_set(atomic64_t *v, s64 a)
{
	/*
	 * This could have been a simple assignment in "C" but would need
	 * explicit volatile. Otherwise gcc optimizers could elide the store
	 * which borked atomic64 self-test
	 * In the inline asm version, memory clobber needed for exact same
	 * reason, to tell gcc about the store.
	 *
	 * This however is not needed for sibling atomic64_add() etc since both
	 * load/store are explicitly done in inline asm. As long as API is used
	 * for each access, gcc has no way to optimize away any load/store
	 */
	__asm__ __volatile__(
	"	std   %0, [%1]	\n"
	:
	: "r"(a), "r"(&v->counter)
	: "memory");
}

#define ATOMIC64_OP(op, op1, op2)					\
static inline void atomic64_##op(s64 a, atomic64_t *v)			\
{									\
	s64 val;							\
									\
	__asm__ __volatile__(						\
	"1:				\n"				\
	"	llockd  %0, [%1]	\n"				\
	"	" #op1 " %L0, %L0, %L2	\n"				\
	"	" #op2 " %H0, %H0, %H2	\n"				\
	"	scondd   %0, [%1]	\n"				\
	"	bnz     1b		\n"				\
	: "=&r"(val)							\
	: "r"(&v->counter), "ir"(a)					\
	: "cc");							\
}									\

#define ATOMIC64_OP_RETURN(op, op1, op2)		        	\
static inline s64 atomic64_##op##_return(s64 a, atomic64_t *v)		\
{									\
	s64 val;							\
									\
	smp_mb();							\
									\
	__asm__ __volatile__(						\
	"1:				\n"				\
	"	llockd   %0, [%1]	\n"				\
	"	" #op1 " %L0, %L0, %L2	\n"				\
	"	" #op2 " %H0, %H0, %H2	\n"				\
	"	scondd   %0, [%1]	\n"				\
	"	bnz     1b		\n"				\
	: [val] "=&r"(val)						\
	: "r"(&v->counter), "ir"(a)					\
	: "cc");	/* memory clobber comes from smp_mb() */	\
									\
	smp_mb();							\
									\
	return val;							\
}

#define ATOMIC64_FETCH_OP(op, op1, op2)		        		\
static inline s64 atomic64_fetch_##op(s64 a, atomic64_t *v)		\
{									\
	s64 val, orig;							\
									\
	smp_mb();							\
									\
	__asm__ __volatile__(						\
	"1:				\n"				\
	"	llockd   %0, [%2]	\n"				\
	"	" #op1 " %L1, %L0, %L3	\n"				\
	"	" #op2 " %H1, %H0, %H3	\n"				\
	"	scondd   %1, [%2]	\n"				\
	"	bnz     1b		\n"				\
	: "=&r"(orig), "=&r"(val)					\
	: "r"(&v->counter), "ir"(a)					\
	: "cc");	/* memory clobber comes from smp_mb() */	\
									\
	smp_mb();							\
									\
	return orig;							\
}

#define ATOMIC64_OPS(op, op1, op2)					\
	ATOMIC64_OP(op, op1, op2)					\
	ATOMIC64_OP_RETURN(op, op1, op2)				\
	ATOMIC64_FETCH_OP(op, op1, op2)

#define atomic64_andnot		atomic64_andnot
#define atomic64_fetch_andnot	atomic64_fetch_andnot

ATOMIC64_OPS(add, add.f, adc)
ATOMIC64_OPS(sub, sub.f, sbc)
ATOMIC64_OPS(and, and, and)
ATOMIC64_OPS(andnot, bic, bic)
ATOMIC64_OPS(or, or, or)
ATOMIC64_OPS(xor, xor, xor)

#undef ATOMIC64_OPS
#undef ATOMIC64_FETCH_OP
#undef ATOMIC64_OP_RETURN
#undef ATOMIC64_OP

static inline s64
atomic64_cmpxchg(atomic64_t *ptr, s64 expected, s64 new)
{
	s64 prev;

	smp_mb();

	__asm__ __volatile__(
	"1:	llockd  %0, [%1]	\n"
	"	brne    %L0, %L2, 2f	\n"
	"	brne    %H0, %H2, 2f	\n"
	"	scondd  %3, [%1]	\n"
	"	bnz     1b		\n"
	"2:				\n"
	: "=&r"(prev)
	: "r"(ptr), "ir"(expected), "r"(new)
	: "cc");	/* memory clobber comes from smp_mb() */

	smp_mb();

	return prev;
}

static inline s64 atomic64_xchg(atomic64_t *ptr, s64 new)
{
	s64 prev;

	smp_mb();

	__asm__ __volatile__(
	"1:	llockd  %0, [%1]	\n"
	"	scondd  %2, [%1]	\n"
	"	bnz     1b		\n"
	"2:				\n"
	: "=&r"(prev)
	: "r"(ptr), "r"(new)
	: "cc");	/* memory clobber comes from smp_mb() */

	smp_mb();

	return prev;
}

/**
 * atomic64_dec_if_positive - decrement by 1 if old value positive
 * @v: pointer of type atomic64_t
 *
 * The function returns the old value of *v minus 1, even if
 * the atomic variable, v, was not decremented.
 */

static inline s64 atomic64_dec_if_positive(atomic64_t *v)
{
	s64 val;

	smp_mb();

	__asm__ __volatile__(
	"1:	llockd  %0, [%1]	\n"
	"	sub.f   %L0, %L0, 1	# w0 - 1, set C on borrow\n"
	"	sub.c   %H0, %H0, 1	# if C set, w1 - 1\n"
	"	brlt    %H0, 0, 2f	\n"
	"	scondd  %0, [%1]	\n"
	"	bnz     1b		\n"
	"2:				\n"
	: "=&r"(val)
	: "r"(&v->counter)
	: "cc");	/* memory clobber comes from smp_mb() */

	smp_mb();

	return val;
}
#define atomic64_dec_if_positive atomic64_dec_if_positive

/**
 * atomic64_fetch_add_unless - add unless the number is a given value
 * @v: pointer of type atomic64_t
 * @a: the amount to add to v...
 * @u: ...unless v is equal to u.
 *
 * Atomically adds @a to @v, if it was not @u.
 * Returns the old value of @v
 */
static inline s64 atomic64_fetch_add_unless(atomic64_t *v, s64 a, s64 u)
{
	s64 old, temp;

	smp_mb();

	__asm__ __volatile__(
	"1:	llockd  %0, [%2]	\n"
	"	brne	%L0, %L4, 2f	# continue to add since v != u \n"
	"	breq.d	%H0, %H4, 3f	# return since v == u \n"
	"2:				\n"
	"	add.f   %L1, %L0, %L3	\n"
	"	adc     %H1, %H0, %H3	\n"
	"	scondd  %1, [%2]	\n"
	"	bnz     1b		\n"
	"3:				\n"
	: "=&r"(old), "=&r" (temp)
	: "r"(&v->counter), "r"(a), "r"(u)
	: "cc");	/* memory clobber comes from smp_mb() */

	smp_mb();

	return old;
}
#define atomic64_fetch_add_unless atomic64_fetch_add_unless

#endif	/* !CONFIG_GENERIC_ATOMIC64 */

#endif	/* !__ASSEMBLY__ */

#endif