summaryrefslogtreecommitdiffstats
path: root/kernel/arch/arc/include/asm/bitops.h
blob: 624a9d048ca9ce2ddab51683b2c9c98e5ea7e6aa (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
/*
 * Copyright (C) 2004, 2007-2010, 2011-2012 Synopsys, Inc. (www.synopsys.com)
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 */

#ifndef _ASM_BITOPS_H
#define _ASM_BITOPS_H

#ifndef _LINUX_BITOPS_H
#error only <linux/bitops.h> can be included directly
#endif

#ifndef __ASSEMBLY__

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

/*
 * Hardware assisted read-modify-write using ARC700 LLOCK/SCOND insns.
 * The Kconfig glue ensures that in SMP, this is only set if the container
 * SoC/platform has cross-core coherent LLOCK/SCOND
 */
#if defined(CONFIG_ARC_HAS_LLSC)

static inline void set_bit(unsigned long nr, volatile unsigned long *m)
{
	unsigned int temp;

	m += nr >> 5;

	/*
	 * ARC ISA micro-optimization:
	 *
	 * Instructions dealing with bitpos only consider lower 5 bits (0-31)
	 * e.g (x << 33) is handled like (x << 1) by ASL instruction
	 *  (mem pointer still needs adjustment to point to next word)
	 *
	 * Hence the masking to clamp @nr arg can be elided in general.
	 *
	 * However if @nr is a constant (above assumed it in a register),
	 * and greater than 31, gcc can optimize away (x << 33) to 0,
	 * as overflow, given the 32-bit ISA. Thus masking needs to be done
	 * for constant @nr, but no code is generated due to const prop.
	 */
	if (__builtin_constant_p(nr))
		nr &= 0x1f;

	__asm__ __volatile__(
	"1:	llock   %0, [%1]	\n"
	"	bset    %0, %0, %2	\n"
	"	scond   %0, [%1]	\n"
	"	bnz     1b	\n"
	: "=&r"(temp)
	: "r"(m), "ir"(nr)
	: "cc");
}

static inline void clear_bit(unsigned long nr, volatile unsigned long *m)
{
	unsigned int temp;

	m += nr >> 5;

	if (__builtin_constant_p(nr))
		nr &= 0x1f;

	__asm__ __volatile__(
	"1:	llock   %0, [%1]	\n"
	"	bclr    %0, %0, %2	\n"
	"	scond   %0, [%1]	\n"
	"	bnz     1b	\n"
	: "=&r"(temp)
	: "r"(m), "ir"(nr)
	: "cc");
}

static inline void change_bit(unsigned long nr, volatile unsigned long *m)
{
	unsigned int temp;

	m += nr >> 5;

	if (__builtin_constant_p(nr))
		nr &= 0x1f;

	__asm__ __volatile__(
	"1:	llock   %0, [%1]	\n"
	"	bxor    %0, %0, %2	\n"
	"	scond   %0, [%1]	\n"
	"	bnz     1b		\n"
	: "=&r"(temp)
	: "r"(m), "ir"(nr)
	: "cc");
}

/*
 * Semantically:
 *    Test the bit
 *    if clear
 *        set it and return 0 (old value)
 *    else
 *        return 1 (old value).
 *
 * Since ARC lacks a equivalent h/w primitive, the bit is set unconditionally
 * and the old value of bit is returned
 */
static inline int test_and_set_bit(unsigned long nr, volatile unsigned long *m)
{
	unsigned long old, temp;

	m += nr >> 5;

	if (__builtin_constant_p(nr))
		nr &= 0x1f;

	/*
	 * Explicit full memory barrier needed before/after as
	 * LLOCK/SCOND themselves don't provide any such semantics
	 */
	smp_mb();

	__asm__ __volatile__(
	"1:	llock   %0, [%2]	\n"
	"	bset    %1, %0, %3	\n"
	"	scond   %1, [%2]	\n"
	"	bnz     1b		\n"
	: "=&r"(old), "=&r"(temp)
	: "r"(m), "ir"(nr)
	: "cc");

	smp_mb();

	return (old & (1 << nr)) != 0;
}

static inline int
test_and_clear_bit(unsigned long nr, volatile unsigned long *m)
{
	unsigned int old, temp;

	m += nr >> 5;

	if (__builtin_constant_p(nr))
		nr &= 0x1f;

	smp_mb();

	__asm__ __volatile__(
	"1:	llock   %0, [%2]	\n"
	"	bclr    %1, %0, %3	\n"
	"	scond   %1, [%2]	\n"
	"	bnz     1b		\n"
	: "=&r"(old), "=&r"(temp)
	: "r"(m), "ir"(nr)
	: "cc");

	smp_mb();

	return (old & (1 << nr)) != 0;
}

static inline int
test_and_change_bit(unsigned long nr, volatile unsigned long *m)
{
	unsigned int old, temp;

	m += nr >> 5;

	if (__builtin_constant_p(nr))
		nr &= 0x1f;

	smp_mb();

	__asm__ __volatile__(
	"1:	llock   %0, [%2]	\n"
	"	bxor    %1, %0, %3	\n"
	"	scond   %1, [%2]	\n"
	"	bnz     1b		\n"
	: "=&r"(old), "=&r"(temp)
	: "r"(m), "ir"(nr)
	: "cc");

	smp_mb();

	return (old & (1 << nr)) != 0;
}

#else	/* !CONFIG_ARC_HAS_LLSC */

#include <asm/smp.h>

/*
 * Non hardware assisted Atomic-R-M-W
 * Locking would change to irq-disabling only (UP) and spinlocks (SMP)
 *
 * There's "significant" micro-optimization in writing our own variants of
 * bitops (over generic variants)
 *
 * (1) The generic APIs have "signed" @nr while we have it "unsigned"
 *     This avoids extra code to be generated for pointer arithmatic, since
 *     is "not sure" that index is NOT -ve
 * (2) Utilize the fact that ARCompact bit fidding insn (BSET/BCLR/ASL) etc
 *     only consider bottom 5 bits of @nr, so NO need to mask them off.
 *     (GCC Quirk: however for constant @nr we still need to do the masking
 *             at compile time)
 */

static inline void set_bit(unsigned long nr, volatile unsigned long *m)
{
	unsigned long temp, flags;
	m += nr >> 5;

	if (__builtin_constant_p(nr))
		nr &= 0x1f;

	bitops_lock(flags);

	temp = *m;
	*m = temp | (1UL << nr);

	bitops_unlock(flags);
}

static inline void clear_bit(unsigned long nr, volatile unsigned long *m)
{
	unsigned long temp, flags;
	m += nr >> 5;

	if (__builtin_constant_p(nr))
		nr &= 0x1f;

	bitops_lock(flags);

	temp = *m;
	*m = temp & ~(1UL << nr);

	bitops_unlock(flags);
}

static inline void change_bit(unsigned long nr, volatile unsigned long *m)
{
	unsigned long temp, flags;
	m += nr >> 5;

	if (__builtin_constant_p(nr))
		nr &= 0x1f;

	bitops_lock(flags);

	temp = *m;
	*m = temp ^ (1UL << nr);

	bitops_unlock(flags);
}

static inline int test_and_set_bit(unsigned long nr, volatile unsigned long *m)
{
	unsigned long old, flags;
	m += nr >> 5;

	if (__builtin_constant_p(nr))
		nr &= 0x1f;

	/*
	 * spin lock/unlock provide the needed smp_mb() before/after
	 */
	bitops_lock(flags);

	old = *m;
	*m = old | (1 << nr);

	bitops_unlock(flags);

	return (old & (1 << nr)) != 0;
}

static inline int
test_and_clear_bit(unsigned long nr, volatile unsigned long *m)
{
	unsigned long old, flags;
	m += nr >> 5;

	if (__builtin_constant_p(nr))
		nr &= 0x1f;

	bitops_lock(flags);

	old = *m;
	*m = old & ~(1 << nr);

	bitops_unlock(flags);

	return (old & (1 << nr)) != 0;
}

static inline int
test_and_change_bit(unsigned long nr, volatile unsigned long *m)
{
	unsigned long old, flags;
	m += nr >> 5;

	if (__builtin_constant_p(nr))
		nr &= 0x1f;

	bitops_lock(flags);

	old = *m;
	*m = old ^ (1 << nr);

	bitops_unlock(flags);

	return (old & (1 << nr)) != 0;
}

#endif /* CONFIG_ARC_HAS_LLSC */

/***************************************
 * Non atomic variants
 **************************************/

static inline void __set_bit(unsigned long nr, volatile unsigned long *m)
{
	unsigned long temp;
	m += nr >> 5;

	if (__builtin_constant_p(nr))
		nr &= 0x1f;

	temp = *m;
	*m = temp | (1UL << nr);
}

static inline void __clear_bit(unsigned long nr, volatile unsigned long *m)
{
	unsigned long temp;
	m += nr >> 5;

	if (__builtin_constant_p(nr))
		nr &= 0x1f;

	temp = *m;
	*m = temp & ~(1UL << nr);
}

static inline void __change_bit(unsigned long nr, volatile unsigned long *m)
{
	unsigned long temp;
	m += nr >> 5;

	if (__builtin_constant_p(nr))
		nr &= 0x1f;

	temp = *m;
	*m = temp ^ (1UL << nr);
}

static inline int
__test_and_set_bit(unsigned long nr, volatile unsigned long *m)
{
	unsigned long old;
	m += nr >> 5;

	if (__builtin_constant_p(nr))
		nr &= 0x1f;

	old = *m;
	*m = old | (1 << nr);

	return (old & (1 << nr)) != 0;
}

static inline int
__test_and_clear_bit(unsigned long nr, volatile unsigned long *m)
{
	unsigned long old;
	m += nr >> 5;

	if (__builtin_constant_p(nr))
		nr &= 0x1f;

	old = *m;
	*m = old & ~(1 << nr);

	return (old & (1 << nr)) != 0;
}

static inline int
__test_and_change_bit(unsigned long nr, volatile unsigned long *m)
{
	unsigned long old;
	m += nr >> 5;

	if (__builtin_constant_p(nr))
		nr &= 0x1f;

	old = *m;
	*m = old ^ (1 << nr);

	return (old & (1 << nr)) != 0;
}

/*
 * This routine doesn't need to be atomic.
 */
static inline int
test_bit(unsigned int nr, const volatile unsigned long *addr)
{
	unsigned long mask;

	addr += nr >> 5;

	if (__builtin_constant_p(nr))
		nr &= 0x1f;

	mask = 1 << nr;

	return ((mask & *addr) != 0);
}

/*
 * Count the number of zeros, starting from MSB
 * Helper for fls( ) friends
 * This is a pure count, so (1-32) or (0-31) doesn't apply
 * It could be 0 to 32, based on num of 0's in there
 * clz(0x8000_0000) = 0, clz(0xFFFF_FFFF)=0, clz(0) = 32, clz(1) = 31
 */
static inline __attribute__ ((const)) int clz(unsigned int x)
{
	unsigned int res;

	__asm__ __volatile__(
	"	norm.f  %0, %1		\n"
	"	mov.n   %0, 0		\n"
	"	add.p   %0, %0, 1	\n"
	: "=r"(res)
	: "r"(x)
	: "cc");

	return res;
}

static inline int constant_fls(int x)
{
	int r = 32;

	if (!x)
		return 0;
	if (!(x & 0xffff0000u)) {
		x <<= 16;
		r -= 16;
	}
	if (!(x & 0xff000000u)) {
		x <<= 8;
		r -= 8;
	}
	if (!(x & 0xf0000000u)) {
		x <<= 4;
		r -= 4;
	}
	if (!(x & 0xc0000000u)) {
		x <<= 2;
		r -= 2;
	}
	if (!(x & 0x80000000u)) {
		x <<= 1;
		r -= 1;
	}
	return r;
}

/*
 * fls = Find Last Set in word
 * @result: [1-32]
 * fls(1) = 1, fls(0x80000000) = 32, fls(0) = 0
 */
static inline __attribute__ ((const)) int fls(unsigned long x)
{
	if (__builtin_constant_p(x))
	       return constant_fls(x);

	return 32 - clz(x);
}

/*
 * __fls: Similar to fls, but zero based (0-31)
 */
static inline __attribute__ ((const)) int __fls(unsigned long x)
{
	if (!x)
		return 0;
	else
		return fls(x) - 1;
}

/*
 * ffs = Find First Set in word (LSB to MSB)
 * @result: [1-32], 0 if all 0's
 */
#define ffs(x)	({ unsigned long __t = (x); fls(__t & -__t); })

/*
 * __ffs: Similar to ffs, but zero based (0-31)
 */
static inline __attribute__ ((const)) int __ffs(unsigned long word)
{
	if (!word)
		return word;

	return ffs(word) - 1;
}

/*
 * ffz = Find First Zero in word.
 * @return:[0-31], 32 if all 1's
 */
#define ffz(x)	__ffs(~(x))

#include <asm-generic/bitops/hweight.h>
#include <asm-generic/bitops/fls64.h>
#include <asm-generic/bitops/sched.h>
#include <asm-generic/bitops/lock.h>

#include <asm-generic/bitops/find.h>
#include <asm-generic/bitops/le.h>
#include <asm-generic/bitops/ext2-atomic-setbit.h>

#endif /* !__ASSEMBLY__ */

#endif