1 /* SPDX-License-Identifier: GPL-2.0 */
2 #ifndef _ASM_POWERPC_ATOMIC_H_
3 #define _ASM_POWERPC_ATOMIC_H_
4 
5 /*
6  * PowerPC atomic operations
7  */
8 
9 #ifdef __KERNEL__
10 #include <linux/types.h>
11 #include <asm/cmpxchg.h>
12 #include <asm/barrier.h>
13 #include <asm/asm-const.h>
14 
15 /*
16  * Since *_return_relaxed and {cmp}xchg_relaxed are implemented with
17  * a "bne-" instruction at the end, so an isync is enough as a acquire barrier
18  * on the platform without lwsync.
19  */
20 #define __atomic_acquire_fence()					\
21 	__asm__ __volatile__(PPC_ACQUIRE_BARRIER "" : : : "memory")
22 
23 #define __atomic_release_fence()					\
24 	__asm__ __volatile__(PPC_RELEASE_BARRIER "" : : : "memory")
25 
arch_atomic_read(const atomic_t * v)26 static __inline__ int arch_atomic_read(const atomic_t *v)
27 {
28 	int t;
29 
30 	__asm__ __volatile__("lwz%U1%X1 %0,%1" : "=r"(t) : "m<>"(v->counter));
31 
32 	return t;
33 }
34 
arch_atomic_set(atomic_t * v,int i)35 static __inline__ void arch_atomic_set(atomic_t *v, int i)
36 {
37 	__asm__ __volatile__("stw%U0%X0 %1,%0" : "=m<>"(v->counter) : "r"(i));
38 }
39 
40 #define ATOMIC_OP(op, asm_op, suffix, sign, ...)			\
41 static __inline__ void arch_atomic_##op(int a, atomic_t *v)		\
42 {									\
43 	int t;								\
44 									\
45 	__asm__ __volatile__(						\
46 "1:	lwarx	%0,0,%3		# atomic_" #op "\n"			\
47 	#asm_op "%I2" suffix " %0,%0,%2\n"				\
48 "	stwcx.	%0,0,%3 \n"						\
49 "	bne-	1b\n"							\
50 	: "=&r" (t), "+m" (v->counter)					\
51 	: "r"#sign (a), "r" (&v->counter)				\
52 	: "cc", ##__VA_ARGS__);						\
53 }									\
54 
55 #define ATOMIC_OP_RETURN_RELAXED(op, asm_op, suffix, sign, ...)		\
56 static inline int arch_atomic_##op##_return_relaxed(int a, atomic_t *v)	\
57 {									\
58 	int t;								\
59 									\
60 	__asm__ __volatile__(						\
61 "1:	lwarx	%0,0,%3		# atomic_" #op "_return_relaxed\n"	\
62 	#asm_op "%I2" suffix " %0,%0,%2\n"				\
63 "	stwcx.	%0,0,%3\n"						\
64 "	bne-	1b\n"							\
65 	: "=&r" (t), "+m" (v->counter)					\
66 	: "r"#sign (a), "r" (&v->counter)				\
67 	: "cc", ##__VA_ARGS__);						\
68 									\
69 	return t;							\
70 }
71 
72 #define ATOMIC_FETCH_OP_RELAXED(op, asm_op, suffix, sign, ...)		\
73 static inline int arch_atomic_fetch_##op##_relaxed(int a, atomic_t *v)	\
74 {									\
75 	int res, t;							\
76 									\
77 	__asm__ __volatile__(						\
78 "1:	lwarx	%0,0,%4		# atomic_fetch_" #op "_relaxed\n"	\
79 	#asm_op "%I3" suffix " %1,%0,%3\n"				\
80 "	stwcx.	%1,0,%4\n"						\
81 "	bne-	1b\n"							\
82 	: "=&r" (res), "=&r" (t), "+m" (v->counter)			\
83 	: "r"#sign (a), "r" (&v->counter)				\
84 	: "cc", ##__VA_ARGS__);						\
85 									\
86 	return res;							\
87 }
88 
89 #define ATOMIC_OPS(op, asm_op, suffix, sign, ...)			\
90 	ATOMIC_OP(op, asm_op, suffix, sign, ##__VA_ARGS__)		\
91 	ATOMIC_OP_RETURN_RELAXED(op, asm_op, suffix, sign, ##__VA_ARGS__)\
92 	ATOMIC_FETCH_OP_RELAXED(op, asm_op, suffix, sign, ##__VA_ARGS__)
93 
94 ATOMIC_OPS(add, add, "c", I, "xer")
95 ATOMIC_OPS(sub, sub, "c", I, "xer")
96 
97 #define arch_atomic_add_return_relaxed arch_atomic_add_return_relaxed
98 #define arch_atomic_sub_return_relaxed arch_atomic_sub_return_relaxed
99 
100 #define arch_atomic_fetch_add_relaxed arch_atomic_fetch_add_relaxed
101 #define arch_atomic_fetch_sub_relaxed arch_atomic_fetch_sub_relaxed
102 
103 #undef ATOMIC_OPS
104 #define ATOMIC_OPS(op, asm_op, suffix, sign)				\
105 	ATOMIC_OP(op, asm_op, suffix, sign)				\
106 	ATOMIC_FETCH_OP_RELAXED(op, asm_op, suffix, sign)
107 
108 ATOMIC_OPS(and, and, ".", K)
109 ATOMIC_OPS(or, or, "", K)
110 ATOMIC_OPS(xor, xor, "", K)
111 
112 #define arch_atomic_fetch_and_relaxed arch_atomic_fetch_and_relaxed
113 #define arch_atomic_fetch_or_relaxed  arch_atomic_fetch_or_relaxed
114 #define arch_atomic_fetch_xor_relaxed arch_atomic_fetch_xor_relaxed
115 
116 #undef ATOMIC_OPS
117 #undef ATOMIC_FETCH_OP_RELAXED
118 #undef ATOMIC_OP_RETURN_RELAXED
119 #undef ATOMIC_OP
120 
121 #define arch_atomic_cmpxchg(v, o, n) \
122 	(arch_cmpxchg(&((v)->counter), (o), (n)))
123 #define arch_atomic_cmpxchg_relaxed(v, o, n) \
124 	arch_cmpxchg_relaxed(&((v)->counter), (o), (n))
125 #define arch_atomic_cmpxchg_acquire(v, o, n) \
126 	arch_cmpxchg_acquire(&((v)->counter), (o), (n))
127 
128 #define arch_atomic_xchg(v, new) \
129 	(arch_xchg(&((v)->counter), new))
130 #define arch_atomic_xchg_relaxed(v, new) \
131 	arch_xchg_relaxed(&((v)->counter), (new))
132 
133 /*
134  * Don't want to override the generic atomic_try_cmpxchg_acquire, because
135  * we add a lock hint to the lwarx, which may not be wanted for the
136  * _acquire case (and is not used by the other _acquire variants so it
137  * would be a surprise).
138  */
139 static __always_inline bool
arch_atomic_try_cmpxchg_lock(atomic_t * v,int * old,int new)140 arch_atomic_try_cmpxchg_lock(atomic_t *v, int *old, int new)
141 {
142 	int r, o = *old;
143 	unsigned int eh = IS_ENABLED(CONFIG_PPC64);
144 
145 	__asm__ __volatile__ (
146 "1:	lwarx	%0,0,%2,%[eh]	# atomic_try_cmpxchg_acquire		\n"
147 "	cmpw	0,%0,%3							\n"
148 "	bne-	2f							\n"
149 "	stwcx.	%4,0,%2							\n"
150 "	bne-	1b							\n"
151 "\t"	PPC_ACQUIRE_BARRIER "						\n"
152 "2:									\n"
153 	: "=&r" (r), "+m" (v->counter)
154 	: "r" (&v->counter), "r" (o), "r" (new), [eh] "n" (eh)
155 	: "cr0", "memory");
156 
157 	if (unlikely(r != o))
158 		*old = r;
159 	return likely(r == o);
160 }
161 
162 /**
163  * atomic_fetch_add_unless - add unless the number is a given value
164  * @v: pointer of type atomic_t
165  * @a: the amount to add to v...
166  * @u: ...unless v is equal to u.
167  *
168  * Atomically adds @a to @v, so long as it was not @u.
169  * Returns the old value of @v.
170  */
arch_atomic_fetch_add_unless(atomic_t * v,int a,int u)171 static __inline__ int arch_atomic_fetch_add_unless(atomic_t *v, int a, int u)
172 {
173 	int t;
174 
175 	__asm__ __volatile__ (
176 	PPC_ATOMIC_ENTRY_BARRIER
177 "1:	lwarx	%0,0,%1		# atomic_fetch_add_unless\n\
178 	cmpw	0,%0,%3 \n\
179 	beq	2f \n\
180 	add%I2c	%0,%0,%2 \n"
181 "	stwcx.	%0,0,%1 \n\
182 	bne-	1b \n"
183 	PPC_ATOMIC_EXIT_BARRIER
184 "	sub%I2c	%0,%0,%2 \n\
185 2:"
186 	: "=&r" (t)
187 	: "r" (&v->counter), "rI" (a), "r" (u)
188 	: "cc", "memory", "xer");
189 
190 	return t;
191 }
192 #define arch_atomic_fetch_add_unless arch_atomic_fetch_add_unless
193 
194 /*
195  * Atomically test *v and decrement if it is greater than 0.
196  * The function returns the old value of *v minus 1, even if
197  * the atomic variable, v, was not decremented.
198  */
arch_atomic_dec_if_positive(atomic_t * v)199 static __inline__ int arch_atomic_dec_if_positive(atomic_t *v)
200 {
201 	int t;
202 
203 	__asm__ __volatile__(
204 	PPC_ATOMIC_ENTRY_BARRIER
205 "1:	lwarx	%0,0,%1		# atomic_dec_if_positive\n\
206 	cmpwi	%0,1\n\
207 	addi	%0,%0,-1\n\
208 	blt-	2f\n"
209 "	stwcx.	%0,0,%1\n\
210 	bne-	1b"
211 	PPC_ATOMIC_EXIT_BARRIER
212 	"\n\
213 2:"	: "=&b" (t)
214 	: "r" (&v->counter)
215 	: "cc", "memory");
216 
217 	return t;
218 }
219 #define arch_atomic_dec_if_positive arch_atomic_dec_if_positive
220 
221 #ifdef __powerpc64__
222 
223 #define ATOMIC64_INIT(i)	{ (i) }
224 
arch_atomic64_read(const atomic64_t * v)225 static __inline__ s64 arch_atomic64_read(const atomic64_t *v)
226 {
227 	s64 t;
228 
229 	__asm__ __volatile__("ld%U1%X1 %0,%1" : "=r"(t) : "m<>"(v->counter));
230 
231 	return t;
232 }
233 
arch_atomic64_set(atomic64_t * v,s64 i)234 static __inline__ void arch_atomic64_set(atomic64_t *v, s64 i)
235 {
236 	__asm__ __volatile__("std%U0%X0 %1,%0" : "=m<>"(v->counter) : "r"(i));
237 }
238 
239 #define ATOMIC64_OP(op, asm_op)						\
240 static __inline__ void arch_atomic64_##op(s64 a, atomic64_t *v)		\
241 {									\
242 	s64 t;								\
243 									\
244 	__asm__ __volatile__(						\
245 "1:	ldarx	%0,0,%3		# atomic64_" #op "\n"			\
246 	#asm_op " %0,%2,%0\n"						\
247 "	stdcx.	%0,0,%3 \n"						\
248 "	bne-	1b\n"							\
249 	: "=&r" (t), "+m" (v->counter)					\
250 	: "r" (a), "r" (&v->counter)					\
251 	: "cc");							\
252 }
253 
254 #define ATOMIC64_OP_RETURN_RELAXED(op, asm_op)				\
255 static inline s64							\
256 arch_atomic64_##op##_return_relaxed(s64 a, atomic64_t *v)		\
257 {									\
258 	s64 t;								\
259 									\
260 	__asm__ __volatile__(						\
261 "1:	ldarx	%0,0,%3		# atomic64_" #op "_return_relaxed\n"	\
262 	#asm_op " %0,%2,%0\n"						\
263 "	stdcx.	%0,0,%3\n"						\
264 "	bne-	1b\n"							\
265 	: "=&r" (t), "+m" (v->counter)					\
266 	: "r" (a), "r" (&v->counter)					\
267 	: "cc");							\
268 									\
269 	return t;							\
270 }
271 
272 #define ATOMIC64_FETCH_OP_RELAXED(op, asm_op)				\
273 static inline s64							\
274 arch_atomic64_fetch_##op##_relaxed(s64 a, atomic64_t *v)		\
275 {									\
276 	s64 res, t;							\
277 									\
278 	__asm__ __volatile__(						\
279 "1:	ldarx	%0,0,%4		# atomic64_fetch_" #op "_relaxed\n"	\
280 	#asm_op " %1,%3,%0\n"						\
281 "	stdcx.	%1,0,%4\n"						\
282 "	bne-	1b\n"							\
283 	: "=&r" (res), "=&r" (t), "+m" (v->counter)			\
284 	: "r" (a), "r" (&v->counter)					\
285 	: "cc");							\
286 									\
287 	return res;							\
288 }
289 
290 #define ATOMIC64_OPS(op, asm_op)					\
291 	ATOMIC64_OP(op, asm_op)						\
292 	ATOMIC64_OP_RETURN_RELAXED(op, asm_op)				\
293 	ATOMIC64_FETCH_OP_RELAXED(op, asm_op)
294 
ATOMIC64_OPS(add,add)295 ATOMIC64_OPS(add, add)
296 ATOMIC64_OPS(sub, subf)
297 
298 #define arch_atomic64_add_return_relaxed arch_atomic64_add_return_relaxed
299 #define arch_atomic64_sub_return_relaxed arch_atomic64_sub_return_relaxed
300 
301 #define arch_atomic64_fetch_add_relaxed arch_atomic64_fetch_add_relaxed
302 #define arch_atomic64_fetch_sub_relaxed arch_atomic64_fetch_sub_relaxed
303 
304 #undef ATOMIC64_OPS
305 #define ATOMIC64_OPS(op, asm_op)					\
306 	ATOMIC64_OP(op, asm_op)						\
307 	ATOMIC64_FETCH_OP_RELAXED(op, asm_op)
308 
309 ATOMIC64_OPS(and, and)
310 ATOMIC64_OPS(or, or)
311 ATOMIC64_OPS(xor, xor)
312 
313 #define arch_atomic64_fetch_and_relaxed arch_atomic64_fetch_and_relaxed
314 #define arch_atomic64_fetch_or_relaxed  arch_atomic64_fetch_or_relaxed
315 #define arch_atomic64_fetch_xor_relaxed arch_atomic64_fetch_xor_relaxed
316 
317 #undef ATOPIC64_OPS
318 #undef ATOMIC64_FETCH_OP_RELAXED
319 #undef ATOMIC64_OP_RETURN_RELAXED
320 #undef ATOMIC64_OP
321 
322 static __inline__ void arch_atomic64_inc(atomic64_t *v)
323 {
324 	s64 t;
325 
326 	__asm__ __volatile__(
327 "1:	ldarx	%0,0,%2		# atomic64_inc\n\
328 	addic	%0,%0,1\n\
329 	stdcx.	%0,0,%2 \n\
330 	bne-	1b"
331 	: "=&r" (t), "+m" (v->counter)
332 	: "r" (&v->counter)
333 	: "cc", "xer");
334 }
335 #define arch_atomic64_inc arch_atomic64_inc
336 
arch_atomic64_inc_return_relaxed(atomic64_t * v)337 static __inline__ s64 arch_atomic64_inc_return_relaxed(atomic64_t *v)
338 {
339 	s64 t;
340 
341 	__asm__ __volatile__(
342 "1:	ldarx	%0,0,%2		# atomic64_inc_return_relaxed\n"
343 "	addic	%0,%0,1\n"
344 "	stdcx.	%0,0,%2\n"
345 "	bne-	1b"
346 	: "=&r" (t), "+m" (v->counter)
347 	: "r" (&v->counter)
348 	: "cc", "xer");
349 
350 	return t;
351 }
352 
arch_atomic64_dec(atomic64_t * v)353 static __inline__ void arch_atomic64_dec(atomic64_t *v)
354 {
355 	s64 t;
356 
357 	__asm__ __volatile__(
358 "1:	ldarx	%0,0,%2		# atomic64_dec\n\
359 	addic	%0,%0,-1\n\
360 	stdcx.	%0,0,%2\n\
361 	bne-	1b"
362 	: "=&r" (t), "+m" (v->counter)
363 	: "r" (&v->counter)
364 	: "cc", "xer");
365 }
366 #define arch_atomic64_dec arch_atomic64_dec
367 
arch_atomic64_dec_return_relaxed(atomic64_t * v)368 static __inline__ s64 arch_atomic64_dec_return_relaxed(atomic64_t *v)
369 {
370 	s64 t;
371 
372 	__asm__ __volatile__(
373 "1:	ldarx	%0,0,%2		# atomic64_dec_return_relaxed\n"
374 "	addic	%0,%0,-1\n"
375 "	stdcx.	%0,0,%2\n"
376 "	bne-	1b"
377 	: "=&r" (t), "+m" (v->counter)
378 	: "r" (&v->counter)
379 	: "cc", "xer");
380 
381 	return t;
382 }
383 
384 #define arch_atomic64_inc_return_relaxed arch_atomic64_inc_return_relaxed
385 #define arch_atomic64_dec_return_relaxed arch_atomic64_dec_return_relaxed
386 
387 /*
388  * Atomically test *v and decrement if it is greater than 0.
389  * The function returns the old value of *v minus 1.
390  */
arch_atomic64_dec_if_positive(atomic64_t * v)391 static __inline__ s64 arch_atomic64_dec_if_positive(atomic64_t *v)
392 {
393 	s64 t;
394 
395 	__asm__ __volatile__(
396 	PPC_ATOMIC_ENTRY_BARRIER
397 "1:	ldarx	%0,0,%1		# atomic64_dec_if_positive\n\
398 	addic.	%0,%0,-1\n\
399 	blt-	2f\n\
400 	stdcx.	%0,0,%1\n\
401 	bne-	1b"
402 	PPC_ATOMIC_EXIT_BARRIER
403 	"\n\
404 2:"	: "=&r" (t)
405 	: "r" (&v->counter)
406 	: "cc", "xer", "memory");
407 
408 	return t;
409 }
410 #define arch_atomic64_dec_if_positive arch_atomic64_dec_if_positive
411 
412 #define arch_atomic64_cmpxchg(v, o, n) \
413 	(arch_cmpxchg(&((v)->counter), (o), (n)))
414 #define arch_atomic64_cmpxchg_relaxed(v, o, n) \
415 	arch_cmpxchg_relaxed(&((v)->counter), (o), (n))
416 #define arch_atomic64_cmpxchg_acquire(v, o, n) \
417 	arch_cmpxchg_acquire(&((v)->counter), (o), (n))
418 
419 #define arch_atomic64_xchg(v, new) \
420 	(arch_xchg(&((v)->counter), new))
421 #define arch_atomic64_xchg_relaxed(v, new) \
422 	arch_xchg_relaxed(&((v)->counter), (new))
423 
424 /**
425  * atomic64_fetch_add_unless - add unless the number is a given value
426  * @v: pointer of type atomic64_t
427  * @a: the amount to add to v...
428  * @u: ...unless v is equal to u.
429  *
430  * Atomically adds @a to @v, so long as it was not @u.
431  * Returns the old value of @v.
432  */
arch_atomic64_fetch_add_unless(atomic64_t * v,s64 a,s64 u)433 static __inline__ s64 arch_atomic64_fetch_add_unless(atomic64_t *v, s64 a, s64 u)
434 {
435 	s64 t;
436 
437 	__asm__ __volatile__ (
438 	PPC_ATOMIC_ENTRY_BARRIER
439 "1:	ldarx	%0,0,%1		# atomic64_fetch_add_unless\n\
440 	cmpd	0,%0,%3 \n\
441 	beq	2f \n\
442 	add	%0,%2,%0 \n"
443 "	stdcx.	%0,0,%1 \n\
444 	bne-	1b \n"
445 	PPC_ATOMIC_EXIT_BARRIER
446 "	subf	%0,%2,%0 \n\
447 2:"
448 	: "=&r" (t)
449 	: "r" (&v->counter), "r" (a), "r" (u)
450 	: "cc", "memory");
451 
452 	return t;
453 }
454 #define arch_atomic64_fetch_add_unless arch_atomic64_fetch_add_unless
455 
456 /**
457  * atomic_inc64_not_zero - increment unless the number is zero
458  * @v: pointer of type atomic64_t
459  *
460  * Atomically increments @v by 1, so long as @v is non-zero.
461  * Returns non-zero if @v was non-zero, and zero otherwise.
462  */
arch_atomic64_inc_not_zero(atomic64_t * v)463 static __inline__ int arch_atomic64_inc_not_zero(atomic64_t *v)
464 {
465 	s64 t1, t2;
466 
467 	__asm__ __volatile__ (
468 	PPC_ATOMIC_ENTRY_BARRIER
469 "1:	ldarx	%0,0,%2		# atomic64_inc_not_zero\n\
470 	cmpdi	0,%0,0\n\
471 	beq-	2f\n\
472 	addic	%1,%0,1\n\
473 	stdcx.	%1,0,%2\n\
474 	bne-	1b\n"
475 	PPC_ATOMIC_EXIT_BARRIER
476 	"\n\
477 2:"
478 	: "=&r" (t1), "=&r" (t2)
479 	: "r" (&v->counter)
480 	: "cc", "xer", "memory");
481 
482 	return t1 != 0;
483 }
484 #define arch_atomic64_inc_not_zero(v) arch_atomic64_inc_not_zero((v))
485 
486 #endif /* __powerpc64__ */
487 
488 #endif /* __KERNEL__ */
489 #endif /* _ASM_POWERPC_ATOMIC_H_ */
490