1 |
3 |
xianfeng |
Semantics and Behavior of Atomic and
|
2 |
|
|
Bitmask Operations
|
3 |
|
|
|
4 |
|
|
David S. Miller
|
5 |
|
|
|
6 |
|
|
This document is intended to serve as a guide to Linux port
|
7 |
|
|
maintainers on how to implement atomic counter, bitops, and spinlock
|
8 |
|
|
interfaces properly.
|
9 |
|
|
|
10 |
|
|
The atomic_t type should be defined as a signed integer.
|
11 |
|
|
Also, it should be made opaque such that any kind of cast to a normal
|
12 |
|
|
C integer type will fail. Something like the following should
|
13 |
|
|
suffice:
|
14 |
|
|
|
15 |
|
|
typedef struct { volatile int counter; } atomic_t;
|
16 |
|
|
|
17 |
|
|
Historically, counter has been declared volatile. This is now discouraged.
|
18 |
|
|
See Documentation/volatile-considered-harmful.txt for the complete rationale.
|
19 |
|
|
|
20 |
|
|
local_t is very similar to atomic_t. If the counter is per CPU and only
|
21 |
|
|
updated by one CPU, local_t is probably more appropriate. Please see
|
22 |
|
|
Documentation/local_ops.txt for the semantics of local_t.
|
23 |
|
|
|
24 |
|
|
The first operations to implement for atomic_t's are the initializers and
|
25 |
|
|
plain reads.
|
26 |
|
|
|
27 |
|
|
#define ATOMIC_INIT(i) { (i) }
|
28 |
|
|
#define atomic_set(v, i) ((v)->counter = (i))
|
29 |
|
|
|
30 |
|
|
The first macro is used in definitions, such as:
|
31 |
|
|
|
32 |
|
|
static atomic_t my_counter = ATOMIC_INIT(1);
|
33 |
|
|
|
34 |
|
|
The initializer is atomic in that the return values of the atomic operations
|
35 |
|
|
are guaranteed to be correct reflecting the initialized value if the
|
36 |
|
|
initializer is used before runtime. If the initializer is used at runtime, a
|
37 |
|
|
proper implicit or explicit read memory barrier is needed before reading the
|
38 |
|
|
value with atomic_read from another thread.
|
39 |
|
|
|
40 |
|
|
The second interface can be used at runtime, as in:
|
41 |
|
|
|
42 |
|
|
struct foo { atomic_t counter; };
|
43 |
|
|
...
|
44 |
|
|
|
45 |
|
|
struct foo *k;
|
46 |
|
|
|
47 |
|
|
k = kmalloc(sizeof(*k), GFP_KERNEL);
|
48 |
|
|
if (!k)
|
49 |
|
|
return -ENOMEM;
|
50 |
|
|
atomic_set(&k->counter, 0);
|
51 |
|
|
|
52 |
|
|
The setting is atomic in that the return values of the atomic operations by
|
53 |
|
|
all threads are guaranteed to be correct reflecting either the value that has
|
54 |
|
|
been set with this operation or set with another operation. A proper implicit
|
55 |
|
|
or explicit memory barrier is needed before the value set with the operation
|
56 |
|
|
is guaranteed to be readable with atomic_read from another thread.
|
57 |
|
|
|
58 |
|
|
Next, we have:
|
59 |
|
|
|
60 |
|
|
#define atomic_read(v) ((v)->counter)
|
61 |
|
|
|
62 |
|
|
which simply reads the counter value currently visible to the calling thread.
|
63 |
|
|
The read is atomic in that the return value is guaranteed to be one of the
|
64 |
|
|
values initialized or modified with the interface operations if a proper
|
65 |
|
|
implicit or explicit memory barrier is used after possible runtime
|
66 |
|
|
initialization by any other thread and the value is modified only with the
|
67 |
|
|
interface operations. atomic_read does not guarantee that the runtime
|
68 |
|
|
initialization by any other thread is visible yet, so the user of the
|
69 |
|
|
interface must take care of that with a proper implicit or explicit memory
|
70 |
|
|
barrier.
|
71 |
|
|
|
72 |
|
|
*** WARNING: atomic_read() and atomic_set() DO NOT IMPLY BARRIERS! ***
|
73 |
|
|
|
74 |
|
|
Some architectures may choose to use the volatile keyword, barriers, or inline
|
75 |
|
|
assembly to guarantee some degree of immediacy for atomic_read() and
|
76 |
|
|
atomic_set(). This is not uniformly guaranteed, and may change in the future,
|
77 |
|
|
so all users of atomic_t should treat atomic_read() and atomic_set() as simple
|
78 |
|
|
C statements that may be reordered or optimized away entirely by the compiler
|
79 |
|
|
or processor, and explicitly invoke the appropriate compiler and/or memory
|
80 |
|
|
barrier for each use case. Failure to do so will result in code that may
|
81 |
|
|
suddenly break when used with different architectures or compiler
|
82 |
|
|
optimizations, or even changes in unrelated code which changes how the
|
83 |
|
|
compiler optimizes the section accessing atomic_t variables.
|
84 |
|
|
|
85 |
|
|
*** YOU HAVE BEEN WARNED! ***
|
86 |
|
|
|
87 |
|
|
Now, we move onto the atomic operation interfaces typically implemented with
|
88 |
|
|
the help of assembly code.
|
89 |
|
|
|
90 |
|
|
void atomic_add(int i, atomic_t *v);
|
91 |
|
|
void atomic_sub(int i, atomic_t *v);
|
92 |
|
|
void atomic_inc(atomic_t *v);
|
93 |
|
|
void atomic_dec(atomic_t *v);
|
94 |
|
|
|
95 |
|
|
These four routines add and subtract integral values to/from the given
|
96 |
|
|
atomic_t value. The first two routines pass explicit integers by
|
97 |
|
|
which to make the adjustment, whereas the latter two use an implicit
|
98 |
|
|
adjustment value of "1".
|
99 |
|
|
|
100 |
|
|
One very important aspect of these two routines is that they DO NOT
|
101 |
|
|
require any explicit memory barriers. They need only perform the
|
102 |
|
|
atomic_t counter update in an SMP safe manner.
|
103 |
|
|
|
104 |
|
|
Next, we have:
|
105 |
|
|
|
106 |
|
|
int atomic_inc_return(atomic_t *v);
|
107 |
|
|
int atomic_dec_return(atomic_t *v);
|
108 |
|
|
|
109 |
|
|
These routines add 1 and subtract 1, respectively, from the given
|
110 |
|
|
atomic_t and return the new counter value after the operation is
|
111 |
|
|
performed.
|
112 |
|
|
|
113 |
|
|
Unlike the above routines, it is required that explicit memory
|
114 |
|
|
barriers are performed before and after the operation. It must be
|
115 |
|
|
done such that all memory operations before and after the atomic
|
116 |
|
|
operation calls are strongly ordered with respect to the atomic
|
117 |
|
|
operation itself.
|
118 |
|
|
|
119 |
|
|
For example, it should behave as if a smp_mb() call existed both
|
120 |
|
|
before and after the atomic operation.
|
121 |
|
|
|
122 |
|
|
If the atomic instructions used in an implementation provide explicit
|
123 |
|
|
memory barrier semantics which satisfy the above requirements, that is
|
124 |
|
|
fine as well.
|
125 |
|
|
|
126 |
|
|
Let's move on:
|
127 |
|
|
|
128 |
|
|
int atomic_add_return(int i, atomic_t *v);
|
129 |
|
|
int atomic_sub_return(int i, atomic_t *v);
|
130 |
|
|
|
131 |
|
|
These behave just like atomic_{inc,dec}_return() except that an
|
132 |
|
|
explicit counter adjustment is given instead of the implicit "1".
|
133 |
|
|
This means that like atomic_{inc,dec}_return(), the memory barrier
|
134 |
|
|
semantics are required.
|
135 |
|
|
|
136 |
|
|
Next:
|
137 |
|
|
|
138 |
|
|
int atomic_inc_and_test(atomic_t *v);
|
139 |
|
|
int atomic_dec_and_test(atomic_t *v);
|
140 |
|
|
|
141 |
|
|
These two routines increment and decrement by 1, respectively, the
|
142 |
|
|
given atomic counter. They return a boolean indicating whether the
|
143 |
|
|
resulting counter value was zero or not.
|
144 |
|
|
|
145 |
|
|
It requires explicit memory barrier semantics around the operation as
|
146 |
|
|
above.
|
147 |
|
|
|
148 |
|
|
int atomic_sub_and_test(int i, atomic_t *v);
|
149 |
|
|
|
150 |
|
|
This is identical to atomic_dec_and_test() except that an explicit
|
151 |
|
|
decrement is given instead of the implicit "1". It requires explicit
|
152 |
|
|
memory barrier semantics around the operation.
|
153 |
|
|
|
154 |
|
|
int atomic_add_negative(int i, atomic_t *v);
|
155 |
|
|
|
156 |
|
|
The given increment is added to the given atomic counter value. A
|
157 |
|
|
boolean is return which indicates whether the resulting counter value
|
158 |
|
|
is negative. It requires explicit memory barrier semantics around the
|
159 |
|
|
operation.
|
160 |
|
|
|
161 |
|
|
Then:
|
162 |
|
|
|
163 |
|
|
int atomic_xchg(atomic_t *v, int new);
|
164 |
|
|
|
165 |
|
|
This performs an atomic exchange operation on the atomic variable v, setting
|
166 |
|
|
the given new value. It returns the old value that the atomic variable v had
|
167 |
|
|
just before the operation.
|
168 |
|
|
|
169 |
|
|
int atomic_cmpxchg(atomic_t *v, int old, int new);
|
170 |
|
|
|
171 |
|
|
This performs an atomic compare exchange operation on the atomic value v,
|
172 |
|
|
with the given old and new values. Like all atomic_xxx operations,
|
173 |
|
|
atomic_cmpxchg will only satisfy its atomicity semantics as long as all
|
174 |
|
|
other accesses of *v are performed through atomic_xxx operations.
|
175 |
|
|
|
176 |
|
|
atomic_cmpxchg requires explicit memory barriers around the operation.
|
177 |
|
|
|
178 |
|
|
The semantics for atomic_cmpxchg are the same as those defined for 'cas'
|
179 |
|
|
below.
|
180 |
|
|
|
181 |
|
|
Finally:
|
182 |
|
|
|
183 |
|
|
int atomic_add_unless(atomic_t *v, int a, int u);
|
184 |
|
|
|
185 |
|
|
If the atomic value v is not equal to u, this function adds a to v, and
|
186 |
|
|
returns non zero. If v is equal to u then it returns zero. This is done as
|
187 |
|
|
an atomic operation.
|
188 |
|
|
|
189 |
|
|
atomic_add_unless requires explicit memory barriers around the operation.
|
190 |
|
|
|
191 |
|
|
atomic_inc_not_zero, equivalent to atomic_add_unless(v, 1, 0)
|
192 |
|
|
|
193 |
|
|
|
194 |
|
|
If a caller requires memory barrier semantics around an atomic_t
|
195 |
|
|
operation which does not return a value, a set of interfaces are
|
196 |
|
|
defined which accomplish this:
|
197 |
|
|
|
198 |
|
|
void smp_mb__before_atomic_dec(void);
|
199 |
|
|
void smp_mb__after_atomic_dec(void);
|
200 |
|
|
void smp_mb__before_atomic_inc(void);
|
201 |
|
|
void smp_mb__after_atomic_inc(void);
|
202 |
|
|
|
203 |
|
|
For example, smp_mb__before_atomic_dec() can be used like so:
|
204 |
|
|
|
205 |
|
|
obj->dead = 1;
|
206 |
|
|
smp_mb__before_atomic_dec();
|
207 |
|
|
atomic_dec(&obj->ref_count);
|
208 |
|
|
|
209 |
|
|
It makes sure that all memory operations preceding the atomic_dec()
|
210 |
|
|
call are strongly ordered with respect to the atomic counter
|
211 |
|
|
operation. In the above example, it guarantees that the assignment of
|
212 |
|
|
"1" to obj->dead will be globally visible to other cpus before the
|
213 |
|
|
atomic counter decrement.
|
214 |
|
|
|
215 |
|
|
Without the explicit smp_mb__before_atomic_dec() call, the
|
216 |
|
|
implementation could legally allow the atomic counter update visible
|
217 |
|
|
to other cpus before the "obj->dead = 1;" assignment.
|
218 |
|
|
|
219 |
|
|
The other three interfaces listed are used to provide explicit
|
220 |
|
|
ordering with respect to memory operations after an atomic_dec() call
|
221 |
|
|
(smp_mb__after_atomic_dec()) and around atomic_inc() calls
|
222 |
|
|
(smp_mb__{before,after}_atomic_inc()).
|
223 |
|
|
|
224 |
|
|
A missing memory barrier in the cases where they are required by the
|
225 |
|
|
atomic_t implementation above can have disastrous results. Here is
|
226 |
|
|
an example, which follows a pattern occurring frequently in the Linux
|
227 |
|
|
kernel. It is the use of atomic counters to implement reference
|
228 |
|
|
counting, and it works such that once the counter falls to zero it can
|
229 |
|
|
be guaranteed that no other entity can be accessing the object:
|
230 |
|
|
|
231 |
|
|
static void obj_list_add(struct obj *obj)
|
232 |
|
|
{
|
233 |
|
|
obj->active = 1;
|
234 |
|
|
list_add(&obj->list);
|
235 |
|
|
}
|
236 |
|
|
|
237 |
|
|
static void obj_list_del(struct obj *obj)
|
238 |
|
|
{
|
239 |
|
|
list_del(&obj->list);
|
240 |
|
|
obj->active = 0;
|
241 |
|
|
}
|
242 |
|
|
|
243 |
|
|
static void obj_destroy(struct obj *obj)
|
244 |
|
|
{
|
245 |
|
|
BUG_ON(obj->active);
|
246 |
|
|
kfree(obj);
|
247 |
|
|
}
|
248 |
|
|
|
249 |
|
|
struct obj *obj_list_peek(struct list_head *head)
|
250 |
|
|
{
|
251 |
|
|
if (!list_empty(head)) {
|
252 |
|
|
struct obj *obj;
|
253 |
|
|
|
254 |
|
|
obj = list_entry(head->next, struct obj, list);
|
255 |
|
|
atomic_inc(&obj->refcnt);
|
256 |
|
|
return obj;
|
257 |
|
|
}
|
258 |
|
|
return NULL;
|
259 |
|
|
}
|
260 |
|
|
|
261 |
|
|
void obj_poke(void)
|
262 |
|
|
{
|
263 |
|
|
struct obj *obj;
|
264 |
|
|
|
265 |
|
|
spin_lock(&global_list_lock);
|
266 |
|
|
obj = obj_list_peek(&global_list);
|
267 |
|
|
spin_unlock(&global_list_lock);
|
268 |
|
|
|
269 |
|
|
if (obj) {
|
270 |
|
|
obj->ops->poke(obj);
|
271 |
|
|
if (atomic_dec_and_test(&obj->refcnt))
|
272 |
|
|
obj_destroy(obj);
|
273 |
|
|
}
|
274 |
|
|
}
|
275 |
|
|
|
276 |
|
|
void obj_timeout(struct obj *obj)
|
277 |
|
|
{
|
278 |
|
|
spin_lock(&global_list_lock);
|
279 |
|
|
obj_list_del(obj);
|
280 |
|
|
spin_unlock(&global_list_lock);
|
281 |
|
|
|
282 |
|
|
if (atomic_dec_and_test(&obj->refcnt))
|
283 |
|
|
obj_destroy(obj);
|
284 |
|
|
}
|
285 |
|
|
|
286 |
|
|
(This is a simplification of the ARP queue management in the
|
287 |
|
|
generic neighbour discover code of the networking. Olaf Kirch
|
288 |
|
|
found a bug wrt. memory barriers in kfree_skb() that exposed
|
289 |
|
|
the atomic_t memory barrier requirements quite clearly.)
|
290 |
|
|
|
291 |
|
|
Given the above scheme, it must be the case that the obj->active
|
292 |
|
|
update done by the obj list deletion be visible to other processors
|
293 |
|
|
before the atomic counter decrement is performed.
|
294 |
|
|
|
295 |
|
|
Otherwise, the counter could fall to zero, yet obj->active would still
|
296 |
|
|
be set, thus triggering the assertion in obj_destroy(). The error
|
297 |
|
|
sequence looks like this:
|
298 |
|
|
|
299 |
|
|
cpu 0 cpu 1
|
300 |
|
|
obj_poke() obj_timeout()
|
301 |
|
|
obj = obj_list_peek();
|
302 |
|
|
... gains ref to obj, refcnt=2
|
303 |
|
|
obj_list_del(obj);
|
304 |
|
|
obj->active = 0 ...
|
305 |
|
|
... visibility delayed ...
|
306 |
|
|
atomic_dec_and_test()
|
307 |
|
|
... refcnt drops to 1 ...
|
308 |
|
|
atomic_dec_and_test()
|
309 |
|
|
... refcount drops to 0 ...
|
310 |
|
|
obj_destroy()
|
311 |
|
|
BUG() triggers since obj->active
|
312 |
|
|
still seen as one
|
313 |
|
|
obj->active update visibility occurs
|
314 |
|
|
|
315 |
|
|
With the memory barrier semantics required of the atomic_t operations
|
316 |
|
|
which return values, the above sequence of memory visibility can never
|
317 |
|
|
happen. Specifically, in the above case the atomic_dec_and_test()
|
318 |
|
|
counter decrement would not become globally visible until the
|
319 |
|
|
obj->active update does.
|
320 |
|
|
|
321 |
|
|
As a historical note, 32-bit Sparc used to only allow usage of
|
322 |
|
|
24-bits of it's atomic_t type. This was because it used 8 bits
|
323 |
|
|
as a spinlock for SMP safety. Sparc32 lacked a "compare and swap"
|
324 |
|
|
type instruction. However, 32-bit Sparc has since been moved over
|
325 |
|
|
to a "hash table of spinlocks" scheme, that allows the full 32-bit
|
326 |
|
|
counter to be realized. Essentially, an array of spinlocks are
|
327 |
|
|
indexed into based upon the address of the atomic_t being operated
|
328 |
|
|
on, and that lock protects the atomic operation. Parisc uses the
|
329 |
|
|
same scheme.
|
330 |
|
|
|
331 |
|
|
Another note is that the atomic_t operations returning values are
|
332 |
|
|
extremely slow on an old 386.
|
333 |
|
|
|
334 |
|
|
We will now cover the atomic bitmask operations. You will find that
|
335 |
|
|
their SMP and memory barrier semantics are similar in shape and scope
|
336 |
|
|
to the atomic_t ops above.
|
337 |
|
|
|
338 |
|
|
Native atomic bit operations are defined to operate on objects aligned
|
339 |
|
|
to the size of an "unsigned long" C data type, and are least of that
|
340 |
|
|
size. The endianness of the bits within each "unsigned long" are the
|
341 |
|
|
native endianness of the cpu.
|
342 |
|
|
|
343 |
|
|
void set_bit(unsigned long nr, volatile unsigned long *addr);
|
344 |
|
|
void clear_bit(unsigned long nr, volatile unsigned long *addr);
|
345 |
|
|
void change_bit(unsigned long nr, volatile unsigned long *addr);
|
346 |
|
|
|
347 |
|
|
These routines set, clear, and change, respectively, the bit number
|
348 |
|
|
indicated by "nr" on the bit mask pointed to by "ADDR".
|
349 |
|
|
|
350 |
|
|
They must execute atomically, yet there are no implicit memory barrier
|
351 |
|
|
semantics required of these interfaces.
|
352 |
|
|
|
353 |
|
|
int test_and_set_bit(unsigned long nr, volatile unsigned long *addr);
|
354 |
|
|
int test_and_clear_bit(unsigned long nr, volatile unsigned long *addr);
|
355 |
|
|
int test_and_change_bit(unsigned long nr, volatile unsigned long *addr);
|
356 |
|
|
|
357 |
|
|
Like the above, except that these routines return a boolean which
|
358 |
|
|
indicates whether the changed bit was set _BEFORE_ the atomic bit
|
359 |
|
|
operation.
|
360 |
|
|
|
361 |
|
|
WARNING! It is incredibly important that the value be a boolean,
|
362 |
|
|
ie. "0" or "1". Do not try to be fancy and save a few instructions by
|
363 |
|
|
declaring the above to return "long" and just returning something like
|
364 |
|
|
"old_val & mask" because that will not work.
|
365 |
|
|
|
366 |
|
|
For one thing, this return value gets truncated to int in many code
|
367 |
|
|
paths using these interfaces, so on 64-bit if the bit is set in the
|
368 |
|
|
upper 32-bits then testers will never see that.
|
369 |
|
|
|
370 |
|
|
One great example of where this problem crops up are the thread_info
|
371 |
|
|
flag operations. Routines such as test_and_set_ti_thread_flag() chop
|
372 |
|
|
the return value into an int. There are other places where things
|
373 |
|
|
like this occur as well.
|
374 |
|
|
|
375 |
|
|
These routines, like the atomic_t counter operations returning values,
|
376 |
|
|
require explicit memory barrier semantics around their execution. All
|
377 |
|
|
memory operations before the atomic bit operation call must be made
|
378 |
|
|
visible globally before the atomic bit operation is made visible.
|
379 |
|
|
Likewise, the atomic bit operation must be visible globally before any
|
380 |
|
|
subsequent memory operation is made visible. For example:
|
381 |
|
|
|
382 |
|
|
obj->dead = 1;
|
383 |
|
|
if (test_and_set_bit(0, &obj->flags))
|
384 |
|
|
/* ... */;
|
385 |
|
|
obj->killed = 1;
|
386 |
|
|
|
387 |
|
|
The implementation of test_and_set_bit() must guarantee that
|
388 |
|
|
"obj->dead = 1;" is visible to cpus before the atomic memory operation
|
389 |
|
|
done by test_and_set_bit() becomes visible. Likewise, the atomic
|
390 |
|
|
memory operation done by test_and_set_bit() must become visible before
|
391 |
|
|
"obj->killed = 1;" is visible.
|
392 |
|
|
|
393 |
|
|
Finally there is the basic operation:
|
394 |
|
|
|
395 |
|
|
int test_bit(unsigned long nr, __const__ volatile unsigned long *addr);
|
396 |
|
|
|
397 |
|
|
Which returns a boolean indicating if bit "nr" is set in the bitmask
|
398 |
|
|
pointed to by "addr".
|
399 |
|
|
|
400 |
|
|
If explicit memory barriers are required around clear_bit() (which
|
401 |
|
|
does not return a value, and thus does not need to provide memory
|
402 |
|
|
barrier semantics), two interfaces are provided:
|
403 |
|
|
|
404 |
|
|
void smp_mb__before_clear_bit(void);
|
405 |
|
|
void smp_mb__after_clear_bit(void);
|
406 |
|
|
|
407 |
|
|
They are used as follows, and are akin to their atomic_t operation
|
408 |
|
|
brothers:
|
409 |
|
|
|
410 |
|
|
/* All memory operations before this call will
|
411 |
|
|
* be globally visible before the clear_bit().
|
412 |
|
|
*/
|
413 |
|
|
smp_mb__before_clear_bit();
|
414 |
|
|
clear_bit( ... );
|
415 |
|
|
|
416 |
|
|
/* The clear_bit() will be visible before all
|
417 |
|
|
* subsequent memory operations.
|
418 |
|
|
*/
|
419 |
|
|
smp_mb__after_clear_bit();
|
420 |
|
|
|
421 |
|
|
There are two special bitops with lock barrier semantics (acquire/release,
|
422 |
|
|
same as spinlocks). These operate in the same way as their non-_lock/unlock
|
423 |
|
|
postfixed variants, except that they are to provide acquire/release semantics,
|
424 |
|
|
respectively. This means they can be used for bit_spin_trylock and
|
425 |
|
|
bit_spin_unlock type operations without specifying any more barriers.
|
426 |
|
|
|
427 |
|
|
int test_and_set_bit_lock(unsigned long nr, unsigned long *addr);
|
428 |
|
|
void clear_bit_unlock(unsigned long nr, unsigned long *addr);
|
429 |
|
|
void __clear_bit_unlock(unsigned long nr, unsigned long *addr);
|
430 |
|
|
|
431 |
|
|
The __clear_bit_unlock version is non-atomic, however it still implements
|
432 |
|
|
unlock barrier semantics. This can be useful if the lock itself is protecting
|
433 |
|
|
the other bits in the word.
|
434 |
|
|
|
435 |
|
|
Finally, there are non-atomic versions of the bitmask operations
|
436 |
|
|
provided. They are used in contexts where some other higher-level SMP
|
437 |
|
|
locking scheme is being used to protect the bitmask, and thus less
|
438 |
|
|
expensive non-atomic operations may be used in the implementation.
|
439 |
|
|
They have names similar to the above bitmask operation interfaces,
|
440 |
|
|
except that two underscores are prefixed to the interface name.
|
441 |
|
|
|
442 |
|
|
void __set_bit(unsigned long nr, volatile unsigned long *addr);
|
443 |
|
|
void __clear_bit(unsigned long nr, volatile unsigned long *addr);
|
444 |
|
|
void __change_bit(unsigned long nr, volatile unsigned long *addr);
|
445 |
|
|
int __test_and_set_bit(unsigned long nr, volatile unsigned long *addr);
|
446 |
|
|
int __test_and_clear_bit(unsigned long nr, volatile unsigned long *addr);
|
447 |
|
|
int __test_and_change_bit(unsigned long nr, volatile unsigned long *addr);
|
448 |
|
|
|
449 |
|
|
These non-atomic variants also do not require any special memory
|
450 |
|
|
barrier semantics.
|
451 |
|
|
|
452 |
|
|
The routines xchg() and cmpxchg() need the same exact memory barriers
|
453 |
|
|
as the atomic and bit operations returning values.
|
454 |
|
|
|
455 |
|
|
Spinlocks and rwlocks have memory barrier expectations as well.
|
456 |
|
|
The rule to follow is simple:
|
457 |
|
|
|
458 |
|
|
1) When acquiring a lock, the implementation must make it globally
|
459 |
|
|
visible before any subsequent memory operation.
|
460 |
|
|
|
461 |
|
|
2) When releasing a lock, the implementation must make it such that
|
462 |
|
|
all previous memory operations are globally visible before the
|
463 |
|
|
lock release.
|
464 |
|
|
|
465 |
|
|
Which finally brings us to _atomic_dec_and_lock(). There is an
|
466 |
|
|
architecture-neutral version implemented in lib/dec_and_lock.c,
|
467 |
|
|
but most platforms will wish to optimize this in assembler.
|
468 |
|
|
|
469 |
|
|
int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock);
|
470 |
|
|
|
471 |
|
|
Atomically decrement the given counter, and if will drop to zero
|
472 |
|
|
atomically acquire the given spinlock and perform the decrement
|
473 |
|
|
of the counter to zero. If it does not drop to zero, do nothing
|
474 |
|
|
with the spinlock.
|
475 |
|
|
|
476 |
|
|
It is actually pretty simple to get the memory barrier correct.
|
477 |
|
|
Simply satisfy the spinlock grab requirements, which is make
|
478 |
|
|
sure the spinlock operation is globally visible before any
|
479 |
|
|
subsequent memory operation.
|
480 |
|
|
|
481 |
|
|
We can demonstrate this operation more clearly if we define
|
482 |
|
|
an abstract atomic operation:
|
483 |
|
|
|
484 |
|
|
long cas(long *mem, long old, long new);
|
485 |
|
|
|
486 |
|
|
"cas" stands for "compare and swap". It atomically:
|
487 |
|
|
|
488 |
|
|
1) Compares "old" with the value currently at "mem".
|
489 |
|
|
2) If they are equal, "new" is written to "mem".
|
490 |
|
|
3) Regardless, the current value at "mem" is returned.
|
491 |
|
|
|
492 |
|
|
As an example usage, here is what an atomic counter update
|
493 |
|
|
might look like:
|
494 |
|
|
|
495 |
|
|
void example_atomic_inc(long *counter)
|
496 |
|
|
{
|
497 |
|
|
long old, new, ret;
|
498 |
|
|
|
499 |
|
|
while (1) {
|
500 |
|
|
old = *counter;
|
501 |
|
|
new = old + 1;
|
502 |
|
|
|
503 |
|
|
ret = cas(counter, old, new);
|
504 |
|
|
if (ret == old)
|
505 |
|
|
break;
|
506 |
|
|
}
|
507 |
|
|
}
|
508 |
|
|
|
509 |
|
|
Let's use cas() in order to build a pseudo-C atomic_dec_and_lock():
|
510 |
|
|
|
511 |
|
|
int _atomic_dec_and_lock(atomic_t *atomic, spinlock_t *lock)
|
512 |
|
|
{
|
513 |
|
|
long old, new, ret;
|
514 |
|
|
int went_to_zero;
|
515 |
|
|
|
516 |
|
|
went_to_zero = 0;
|
517 |
|
|
while (1) {
|
518 |
|
|
old = atomic_read(atomic);
|
519 |
|
|
new = old - 1;
|
520 |
|
|
if (new == 0) {
|
521 |
|
|
went_to_zero = 1;
|
522 |
|
|
spin_lock(lock);
|
523 |
|
|
}
|
524 |
|
|
ret = cas(atomic, old, new);
|
525 |
|
|
if (ret == old)
|
526 |
|
|
break;
|
527 |
|
|
if (went_to_zero) {
|
528 |
|
|
spin_unlock(lock);
|
529 |
|
|
went_to_zero = 0;
|
530 |
|
|
}
|
531 |
|
|
}
|
532 |
|
|
|
533 |
|
|
return went_to_zero;
|
534 |
|
|
}
|
535 |
|
|
|
536 |
|
|
Now, as far as memory barriers go, as long as spin_lock()
|
537 |
|
|
strictly orders all subsequent memory operations (including
|
538 |
|
|
the cas()) with respect to itself, things will be fine.
|
539 |
|
|
|
540 |
|
|
Said another way, _atomic_dec_and_lock() must guarantee that
|
541 |
|
|
a counter dropping to zero is never made visible before the
|
542 |
|
|
spinlock being acquired.
|
543 |
|
|
|
544 |
|
|
Note that this also means that for the case where the counter
|
545 |
|
|
is not dropping to zero, there are no memory ordering
|
546 |
|
|
requirements.
|