]>
Commit | Line | Data |
---|---|---|
1 | ============================ | |
2 | LINUX KERNEL MEMORY BARRIERS | |
3 | ============================ | |
4 | ||
5 | By: David Howells <dhowells@redhat.com> | |
6 | Paul E. McKenney <paulmck@linux.vnet.ibm.com> | |
7 | ||
8 | Contents: | |
9 | ||
10 | (*) Abstract memory access model. | |
11 | ||
12 | - Device operations. | |
13 | - Guarantees. | |
14 | ||
15 | (*) What are memory barriers? | |
16 | ||
17 | - Varieties of memory barrier. | |
18 | - What may not be assumed about memory barriers? | |
19 | - Data dependency barriers. | |
20 | - Control dependencies. | |
21 | - SMP barrier pairing. | |
22 | - Examples of memory barrier sequences. | |
23 | - Read memory barriers vs load speculation. | |
24 | - Transitivity | |
25 | ||
26 | (*) Explicit kernel barriers. | |
27 | ||
28 | - Compiler barrier. | |
29 | - CPU memory barriers. | |
30 | - MMIO write barrier. | |
31 | ||
32 | (*) Implicit kernel memory barriers. | |
33 | ||
34 | - Locking functions. | |
35 | - Interrupt disabling functions. | |
36 | - Sleep and wake-up functions. | |
37 | - Miscellaneous functions. | |
38 | ||
39 | (*) Inter-CPU locking barrier effects. | |
40 | ||
41 | - Locks vs memory accesses. | |
42 | - Locks vs I/O accesses. | |
43 | ||
44 | (*) Where are memory barriers needed? | |
45 | ||
46 | - Interprocessor interaction. | |
47 | - Atomic operations. | |
48 | - Accessing devices. | |
49 | - Interrupts. | |
50 | ||
51 | (*) Kernel I/O barrier effects. | |
52 | ||
53 | (*) Assumed minimum execution ordering model. | |
54 | ||
55 | (*) The effects of the cpu cache. | |
56 | ||
57 | - Cache coherency. | |
58 | - Cache coherency vs DMA. | |
59 | - Cache coherency vs MMIO. | |
60 | ||
61 | (*) The things CPUs get up to. | |
62 | ||
63 | - And then there's the Alpha. | |
64 | ||
65 | (*) Example uses. | |
66 | ||
67 | - Circular buffers. | |
68 | ||
69 | (*) References. | |
70 | ||
71 | ||
72 | ============================ | |
73 | ABSTRACT MEMORY ACCESS MODEL | |
74 | ============================ | |
75 | ||
76 | Consider the following abstract model of the system: | |
77 | ||
78 | : : | |
79 | : : | |
80 | : : | |
81 | +-------+ : +--------+ : +-------+ | |
82 | | | : | | : | | | |
83 | | | : | | : | | | |
84 | | CPU 1 |<----->| Memory |<----->| CPU 2 | | |
85 | | | : | | : | | | |
86 | | | : | | : | | | |
87 | +-------+ : +--------+ : +-------+ | |
88 | ^ : ^ : ^ | |
89 | | : | : | | |
90 | | : | : | | |
91 | | : v : | | |
92 | | : +--------+ : | | |
93 | | : | | : | | |
94 | | : | | : | | |
95 | +---------->| Device |<----------+ | |
96 | : | | : | |
97 | : | | : | |
98 | : +--------+ : | |
99 | : : | |
100 | ||
101 | Each CPU executes a program that generates memory access operations. In the | |
102 | abstract CPU, memory operation ordering is very relaxed, and a CPU may actually | |
103 | perform the memory operations in any order it likes, provided program causality | |
104 | appears to be maintained. Similarly, the compiler may also arrange the | |
105 | instructions it emits in any order it likes, provided it doesn't affect the | |
106 | apparent operation of the program. | |
107 | ||
108 | So in the above diagram, the effects of the memory operations performed by a | |
109 | CPU are perceived by the rest of the system as the operations cross the | |
110 | interface between the CPU and rest of the system (the dotted lines). | |
111 | ||
112 | ||
113 | For example, consider the following sequence of events: | |
114 | ||
115 | CPU 1 CPU 2 | |
116 | =============== =============== | |
117 | { A == 1; B == 2 } | |
118 | A = 3; x = A; | |
119 | B = 4; y = B; | |
120 | ||
121 | The set of accesses as seen by the memory system in the middle can be arranged | |
122 | in 24 different combinations: | |
123 | ||
124 | STORE A=3, STORE B=4, x=LOAD A->3, y=LOAD B->4 | |
125 | STORE A=3, STORE B=4, y=LOAD B->4, x=LOAD A->3 | |
126 | STORE A=3, x=LOAD A->3, STORE B=4, y=LOAD B->4 | |
127 | STORE A=3, x=LOAD A->3, y=LOAD B->2, STORE B=4 | |
128 | STORE A=3, y=LOAD B->2, STORE B=4, x=LOAD A->3 | |
129 | STORE A=3, y=LOAD B->2, x=LOAD A->3, STORE B=4 | |
130 | STORE B=4, STORE A=3, x=LOAD A->3, y=LOAD B->4 | |
131 | STORE B=4, ... | |
132 | ... | |
133 | ||
134 | and can thus result in four different combinations of values: | |
135 | ||
136 | x == 1, y == 2 | |
137 | x == 1, y == 4 | |
138 | x == 3, y == 2 | |
139 | x == 3, y == 4 | |
140 | ||
141 | ||
142 | Furthermore, the stores committed by a CPU to the memory system may not be | |
143 | perceived by the loads made by another CPU in the same order as the stores were | |
144 | committed. | |
145 | ||
146 | ||
147 | As a further example, consider this sequence of events: | |
148 | ||
149 | CPU 1 CPU 2 | |
150 | =============== =============== | |
151 | { A == 1, B == 2, C = 3, P == &A, Q == &C } | |
152 | B = 4; Q = P; | |
153 | P = &B D = *Q; | |
154 | ||
155 | There is an obvious data dependency here, as the value loaded into D depends on | |
156 | the address retrieved from P by CPU 2. At the end of the sequence, any of the | |
157 | following results are possible: | |
158 | ||
159 | (Q == &A) and (D == 1) | |
160 | (Q == &B) and (D == 2) | |
161 | (Q == &B) and (D == 4) | |
162 | ||
163 | Note that CPU 2 will never try and load C into D because the CPU will load P | |
164 | into Q before issuing the load of *Q. | |
165 | ||
166 | ||
167 | DEVICE OPERATIONS | |
168 | ----------------- | |
169 | ||
170 | Some devices present their control interfaces as collections of memory | |
171 | locations, but the order in which the control registers are accessed is very | |
172 | important. For instance, imagine an ethernet card with a set of internal | |
173 | registers that are accessed through an address port register (A) and a data | |
174 | port register (D). To read internal register 5, the following code might then | |
175 | be used: | |
176 | ||
177 | *A = 5; | |
178 | x = *D; | |
179 | ||
180 | but this might show up as either of the following two sequences: | |
181 | ||
182 | STORE *A = 5, x = LOAD *D | |
183 | x = LOAD *D, STORE *A = 5 | |
184 | ||
185 | the second of which will almost certainly result in a malfunction, since it set | |
186 | the address _after_ attempting to read the register. | |
187 | ||
188 | ||
189 | GUARANTEES | |
190 | ---------- | |
191 | ||
192 | There are some minimal guarantees that may be expected of a CPU: | |
193 | ||
194 | (*) On any given CPU, dependent memory accesses will be issued in order, with | |
195 | respect to itself. This means that for: | |
196 | ||
197 | ACCESS_ONCE(Q) = P; smp_read_barrier_depends(); D = ACCESS_ONCE(*Q); | |
198 | ||
199 | the CPU will issue the following memory operations: | |
200 | ||
201 | Q = LOAD P, D = LOAD *Q | |
202 | ||
203 | and always in that order. On most systems, smp_read_barrier_depends() | |
204 | does nothing, but it is required for DEC Alpha. The ACCESS_ONCE() | |
205 | is required to prevent compiler mischief. Please note that you | |
206 | should normally use something like rcu_dereference() instead of | |
207 | open-coding smp_read_barrier_depends(). | |
208 | ||
209 | (*) Overlapping loads and stores within a particular CPU will appear to be | |
210 | ordered within that CPU. This means that for: | |
211 | ||
212 | a = ACCESS_ONCE(*X); ACCESS_ONCE(*X) = b; | |
213 | ||
214 | the CPU will only issue the following sequence of memory operations: | |
215 | ||
216 | a = LOAD *X, STORE *X = b | |
217 | ||
218 | And for: | |
219 | ||
220 | ACCESS_ONCE(*X) = c; d = ACCESS_ONCE(*X); | |
221 | ||
222 | the CPU will only issue: | |
223 | ||
224 | STORE *X = c, d = LOAD *X | |
225 | ||
226 | (Loads and stores overlap if they are targeted at overlapping pieces of | |
227 | memory). | |
228 | ||
229 | And there are a number of things that _must_ or _must_not_ be assumed: | |
230 | ||
231 | (*) It _must_not_ be assumed that the compiler will do what you want with | |
232 | memory references that are not protected by ACCESS_ONCE(). Without | |
233 | ACCESS_ONCE(), the compiler is within its rights to do all sorts | |
234 | of "creative" transformations, which are covered in the Compiler | |
235 | Barrier section. | |
236 | ||
237 | (*) It _must_not_ be assumed that independent loads and stores will be issued | |
238 | in the order given. This means that for: | |
239 | ||
240 | X = *A; Y = *B; *D = Z; | |
241 | ||
242 | we may get any of the following sequences: | |
243 | ||
244 | X = LOAD *A, Y = LOAD *B, STORE *D = Z | |
245 | X = LOAD *A, STORE *D = Z, Y = LOAD *B | |
246 | Y = LOAD *B, X = LOAD *A, STORE *D = Z | |
247 | Y = LOAD *B, STORE *D = Z, X = LOAD *A | |
248 | STORE *D = Z, X = LOAD *A, Y = LOAD *B | |
249 | STORE *D = Z, Y = LOAD *B, X = LOAD *A | |
250 | ||
251 | (*) It _must_ be assumed that overlapping memory accesses may be merged or | |
252 | discarded. This means that for: | |
253 | ||
254 | X = *A; Y = *(A + 4); | |
255 | ||
256 | we may get any one of the following sequences: | |
257 | ||
258 | X = LOAD *A; Y = LOAD *(A + 4); | |
259 | Y = LOAD *(A + 4); X = LOAD *A; | |
260 | {X, Y} = LOAD {*A, *(A + 4) }; | |
261 | ||
262 | And for: | |
263 | ||
264 | *A = X; *(A + 4) = Y; | |
265 | ||
266 | we may get any of: | |
267 | ||
268 | STORE *A = X; STORE *(A + 4) = Y; | |
269 | STORE *(A + 4) = Y; STORE *A = X; | |
270 | STORE {*A, *(A + 4) } = {X, Y}; | |
271 | ||
272 | ||
273 | ========================= | |
274 | WHAT ARE MEMORY BARRIERS? | |
275 | ========================= | |
276 | ||
277 | As can be seen above, independent memory operations are effectively performed | |
278 | in random order, but this can be a problem for CPU-CPU interaction and for I/O. | |
279 | What is required is some way of intervening to instruct the compiler and the | |
280 | CPU to restrict the order. | |
281 | ||
282 | Memory barriers are such interventions. They impose a perceived partial | |
283 | ordering over the memory operations on either side of the barrier. | |
284 | ||
285 | Such enforcement is important because the CPUs and other devices in a system | |
286 | can use a variety of tricks to improve performance, including reordering, | |
287 | deferral and combination of memory operations; speculative loads; speculative | |
288 | branch prediction and various types of caching. Memory barriers are used to | |
289 | override or suppress these tricks, allowing the code to sanely control the | |
290 | interaction of multiple CPUs and/or devices. | |
291 | ||
292 | ||
293 | VARIETIES OF MEMORY BARRIER | |
294 | --------------------------- | |
295 | ||
296 | Memory barriers come in four basic varieties: | |
297 | ||
298 | (1) Write (or store) memory barriers. | |
299 | ||
300 | A write memory barrier gives a guarantee that all the STORE operations | |
301 | specified before the barrier will appear to happen before all the STORE | |
302 | operations specified after the barrier with respect to the other | |
303 | components of the system. | |
304 | ||
305 | A write barrier is a partial ordering on stores only; it is not required | |
306 | to have any effect on loads. | |
307 | ||
308 | A CPU can be viewed as committing a sequence of store operations to the | |
309 | memory system as time progresses. All stores before a write barrier will | |
310 | occur in the sequence _before_ all the stores after the write barrier. | |
311 | ||
312 | [!] Note that write barriers should normally be paired with read or data | |
313 | dependency barriers; see the "SMP barrier pairing" subsection. | |
314 | ||
315 | ||
316 | (2) Data dependency barriers. | |
317 | ||
318 | A data dependency barrier is a weaker form of read barrier. In the case | |
319 | where two loads are performed such that the second depends on the result | |
320 | of the first (eg: the first load retrieves the address to which the second | |
321 | load will be directed), a data dependency barrier would be required to | |
322 | make sure that the target of the second load is updated before the address | |
323 | obtained by the first load is accessed. | |
324 | ||
325 | A data dependency barrier is a partial ordering on interdependent loads | |
326 | only; it is not required to have any effect on stores, independent loads | |
327 | or overlapping loads. | |
328 | ||
329 | As mentioned in (1), the other CPUs in the system can be viewed as | |
330 | committing sequences of stores to the memory system that the CPU being | |
331 | considered can then perceive. A data dependency barrier issued by the CPU | |
332 | under consideration guarantees that for any load preceding it, if that | |
333 | load touches one of a sequence of stores from another CPU, then by the | |
334 | time the barrier completes, the effects of all the stores prior to that | |
335 | touched by the load will be perceptible to any loads issued after the data | |
336 | dependency barrier. | |
337 | ||
338 | See the "Examples of memory barrier sequences" subsection for diagrams | |
339 | showing the ordering constraints. | |
340 | ||
341 | [!] Note that the first load really has to have a _data_ dependency and | |
342 | not a control dependency. If the address for the second load is dependent | |
343 | on the first load, but the dependency is through a conditional rather than | |
344 | actually loading the address itself, then it's a _control_ dependency and | |
345 | a full read barrier or better is required. See the "Control dependencies" | |
346 | subsection for more information. | |
347 | ||
348 | [!] Note that data dependency barriers should normally be paired with | |
349 | write barriers; see the "SMP barrier pairing" subsection. | |
350 | ||
351 | ||
352 | (3) Read (or load) memory barriers. | |
353 | ||
354 | A read barrier is a data dependency barrier plus a guarantee that all the | |
355 | LOAD operations specified before the barrier will appear to happen before | |
356 | all the LOAD operations specified after the barrier with respect to the | |
357 | other components of the system. | |
358 | ||
359 | A read barrier is a partial ordering on loads only; it is not required to | |
360 | have any effect on stores. | |
361 | ||
362 | Read memory barriers imply data dependency barriers, and so can substitute | |
363 | for them. | |
364 | ||
365 | [!] Note that read barriers should normally be paired with write barriers; | |
366 | see the "SMP barrier pairing" subsection. | |
367 | ||
368 | ||
369 | (4) General memory barriers. | |
370 | ||
371 | A general memory barrier gives a guarantee that all the LOAD and STORE | |
372 | operations specified before the barrier will appear to happen before all | |
373 | the LOAD and STORE operations specified after the barrier with respect to | |
374 | the other components of the system. | |
375 | ||
376 | A general memory barrier is a partial ordering over both loads and stores. | |
377 | ||
378 | General memory barriers imply both read and write memory barriers, and so | |
379 | can substitute for either. | |
380 | ||
381 | ||
382 | And a couple of implicit varieties: | |
383 | ||
384 | (5) LOCK operations. | |
385 | ||
386 | This acts as a one-way permeable barrier. It guarantees that all memory | |
387 | operations after the LOCK operation will appear to happen after the LOCK | |
388 | operation with respect to the other components of the system. | |
389 | ||
390 | Memory operations that occur before a LOCK operation may appear to happen | |
391 | after it completes. | |
392 | ||
393 | A LOCK operation should almost always be paired with an UNLOCK operation. | |
394 | ||
395 | ||
396 | (6) UNLOCK operations. | |
397 | ||
398 | This also acts as a one-way permeable barrier. It guarantees that all | |
399 | memory operations before the UNLOCK operation will appear to happen before | |
400 | the UNLOCK operation with respect to the other components of the system. | |
401 | ||
402 | Memory operations that occur after an UNLOCK operation may appear to | |
403 | happen before it completes. | |
404 | ||
405 | LOCK and UNLOCK operations are guaranteed to appear with respect to each | |
406 | other strictly in the order specified. | |
407 | ||
408 | The use of LOCK and UNLOCK operations generally precludes the need for | |
409 | other sorts of memory barrier (but note the exceptions mentioned in the | |
410 | subsection "MMIO write barrier"). | |
411 | ||
412 | ||
413 | Memory barriers are only required where there's a possibility of interaction | |
414 | between two CPUs or between a CPU and a device. If it can be guaranteed that | |
415 | there won't be any such interaction in any particular piece of code, then | |
416 | memory barriers are unnecessary in that piece of code. | |
417 | ||
418 | ||
419 | Note that these are the _minimum_ guarantees. Different architectures may give | |
420 | more substantial guarantees, but they may _not_ be relied upon outside of arch | |
421 | specific code. | |
422 | ||
423 | ||
424 | WHAT MAY NOT BE ASSUMED ABOUT MEMORY BARRIERS? | |
425 | ---------------------------------------------- | |
426 | ||
427 | There are certain things that the Linux kernel memory barriers do not guarantee: | |
428 | ||
429 | (*) There is no guarantee that any of the memory accesses specified before a | |
430 | memory barrier will be _complete_ by the completion of a memory barrier | |
431 | instruction; the barrier can be considered to draw a line in that CPU's | |
432 | access queue that accesses of the appropriate type may not cross. | |
433 | ||
434 | (*) There is no guarantee that issuing a memory barrier on one CPU will have | |
435 | any direct effect on another CPU or any other hardware in the system. The | |
436 | indirect effect will be the order in which the second CPU sees the effects | |
437 | of the first CPU's accesses occur, but see the next point: | |
438 | ||
439 | (*) There is no guarantee that a CPU will see the correct order of effects | |
440 | from a second CPU's accesses, even _if_ the second CPU uses a memory | |
441 | barrier, unless the first CPU _also_ uses a matching memory barrier (see | |
442 | the subsection on "SMP Barrier Pairing"). | |
443 | ||
444 | (*) There is no guarantee that some intervening piece of off-the-CPU | |
445 | hardware[*] will not reorder the memory accesses. CPU cache coherency | |
446 | mechanisms should propagate the indirect effects of a memory barrier | |
447 | between CPUs, but might not do so in order. | |
448 | ||
449 | [*] For information on bus mastering DMA and coherency please read: | |
450 | ||
451 | Documentation/PCI/pci.txt | |
452 | Documentation/DMA-API-HOWTO.txt | |
453 | Documentation/DMA-API.txt | |
454 | ||
455 | ||
456 | DATA DEPENDENCY BARRIERS | |
457 | ------------------------ | |
458 | ||
459 | The usage requirements of data dependency barriers are a little subtle, and | |
460 | it's not always obvious that they're needed. To illustrate, consider the | |
461 | following sequence of events: | |
462 | ||
463 | CPU 1 CPU 2 | |
464 | =============== =============== | |
465 | { A == 1, B == 2, C = 3, P == &A, Q == &C } | |
466 | B = 4; | |
467 | <write barrier> | |
468 | ACCESS_ONCE(P) = &B | |
469 | Q = ACCESS_ONCE(P); | |
470 | D = *Q; | |
471 | ||
472 | There's a clear data dependency here, and it would seem that by the end of the | |
473 | sequence, Q must be either &A or &B, and that: | |
474 | ||
475 | (Q == &A) implies (D == 1) | |
476 | (Q == &B) implies (D == 4) | |
477 | ||
478 | But! CPU 2's perception of P may be updated _before_ its perception of B, thus | |
479 | leading to the following situation: | |
480 | ||
481 | (Q == &B) and (D == 2) ???? | |
482 | ||
483 | Whilst this may seem like a failure of coherency or causality maintenance, it | |
484 | isn't, and this behaviour can be observed on certain real CPUs (such as the DEC | |
485 | Alpha). | |
486 | ||
487 | To deal with this, a data dependency barrier or better must be inserted | |
488 | between the address load and the data load: | |
489 | ||
490 | CPU 1 CPU 2 | |
491 | =============== =============== | |
492 | { A == 1, B == 2, C = 3, P == &A, Q == &C } | |
493 | B = 4; | |
494 | <write barrier> | |
495 | ACCESS_ONCE(P) = &B | |
496 | Q = ACCESS_ONCE(P); | |
497 | <data dependency barrier> | |
498 | D = *Q; | |
499 | ||
500 | This enforces the occurrence of one of the two implications, and prevents the | |
501 | third possibility from arising. | |
502 | ||
503 | [!] Note that this extremely counterintuitive situation arises most easily on | |
504 | machines with split caches, so that, for example, one cache bank processes | |
505 | even-numbered cache lines and the other bank processes odd-numbered cache | |
506 | lines. The pointer P might be stored in an odd-numbered cache line, and the | |
507 | variable B might be stored in an even-numbered cache line. Then, if the | |
508 | even-numbered bank of the reading CPU's cache is extremely busy while the | |
509 | odd-numbered bank is idle, one can see the new value of the pointer P (&B), | |
510 | but the old value of the variable B (2). | |
511 | ||
512 | ||
513 | Another example of where data dependency barriers might be required is where a | |
514 | number is read from memory and then used to calculate the index for an array | |
515 | access: | |
516 | ||
517 | CPU 1 CPU 2 | |
518 | =============== =============== | |
519 | { M[0] == 1, M[1] == 2, M[3] = 3, P == 0, Q == 3 } | |
520 | M[1] = 4; | |
521 | <write barrier> | |
522 | ACCESS_ONCE(P) = 1 | |
523 | Q = ACCESS_ONCE(P); | |
524 | <data dependency barrier> | |
525 | D = M[Q]; | |
526 | ||
527 | ||
528 | The data dependency barrier is very important to the RCU system, | |
529 | for example. See rcu_assign_pointer() and rcu_dereference() in | |
530 | include/linux/rcupdate.h. This permits the current target of an RCU'd | |
531 | pointer to be replaced with a new modified target, without the replacement | |
532 | target appearing to be incompletely initialised. | |
533 | ||
534 | See also the subsection on "Cache Coherency" for a more thorough example. | |
535 | ||
536 | ||
537 | CONTROL DEPENDENCIES | |
538 | -------------------- | |
539 | ||
540 | A control dependency requires a full read memory barrier, not simply a data | |
541 | dependency barrier to make it work correctly. Consider the following bit of | |
542 | code: | |
543 | ||
544 | q = ACCESS_ONCE(a); | |
545 | if (q) { | |
546 | <data dependency barrier> /* BUG: No data dependency!!! */ | |
547 | p = ACCESS_ONCE(b); | |
548 | } | |
549 | ||
550 | This will not have the desired effect because there is no actual data | |
551 | dependency, but rather a control dependency that the CPU may short-circuit | |
552 | by attempting to predict the outcome in advance, so that other CPUs see | |
553 | the load from b as having happened before the load from a. In such a | |
554 | case what's actually required is: | |
555 | ||
556 | q = ACCESS_ONCE(a); | |
557 | if (q) { | |
558 | <read barrier> | |
559 | p = ACCESS_ONCE(b); | |
560 | } | |
561 | ||
562 | However, stores are not speculated. This means that ordering -is- provided | |
563 | in the following example: | |
564 | ||
565 | q = ACCESS_ONCE(a); | |
566 | if (ACCESS_ONCE(q)) { | |
567 | ACCESS_ONCE(b) = p; | |
568 | } | |
569 | ||
570 | Please note that ACCESS_ONCE() is not optional! Without the ACCESS_ONCE(), | |
571 | the compiler is within its rights to transform this example: | |
572 | ||
573 | q = a; | |
574 | if (q) { | |
575 | b = p; /* BUG: Compiler can reorder!!! */ | |
576 | do_something(); | |
577 | } else { | |
578 | b = p; /* BUG: Compiler can reorder!!! */ | |
579 | do_something_else(); | |
580 | } | |
581 | ||
582 | into this, which of course defeats the ordering: | |
583 | ||
584 | b = p; | |
585 | q = a; | |
586 | if (q) | |
587 | do_something(); | |
588 | else | |
589 | do_something_else(); | |
590 | ||
591 | Worse yet, if the compiler is able to prove (say) that the value of | |
592 | variable 'a' is always non-zero, it would be well within its rights | |
593 | to optimize the original example by eliminating the "if" statement | |
594 | as follows: | |
595 | ||
596 | q = a; | |
597 | b = p; /* BUG: Compiler can reorder!!! */ | |
598 | do_something(); | |
599 | ||
600 | The solution is again ACCESS_ONCE(), which preserves the ordering between | |
601 | the load from variable 'a' and the store to variable 'b': | |
602 | ||
603 | q = ACCESS_ONCE(a); | |
604 | if (q) { | |
605 | ACCESS_ONCE(b) = p; | |
606 | do_something(); | |
607 | } else { | |
608 | ACCESS_ONCE(b) = p; | |
609 | do_something_else(); | |
610 | } | |
611 | ||
612 | You could also use barrier() to prevent the compiler from moving | |
613 | the stores to variable 'b', but barrier() would not prevent the | |
614 | compiler from proving to itself that a==1 always, so ACCESS_ONCE() | |
615 | is also needed. | |
616 | ||
617 | It is important to note that control dependencies absolutely require a | |
618 | a conditional. For example, the following "optimized" version of | |
619 | the above example breaks ordering: | |
620 | ||
621 | q = ACCESS_ONCE(a); | |
622 | ACCESS_ONCE(b) = p; /* BUG: No ordering vs. load from a!!! */ | |
623 | if (q) { | |
624 | /* ACCESS_ONCE(b) = p; -- moved up, BUG!!! */ | |
625 | do_something(); | |
626 | } else { | |
627 | /* ACCESS_ONCE(b) = p; -- moved up, BUG!!! */ | |
628 | do_something_else(); | |
629 | } | |
630 | ||
631 | It is of course legal for the prior load to be part of the conditional, | |
632 | for example, as follows: | |
633 | ||
634 | if (ACCESS_ONCE(a) > 0) { | |
635 | ACCESS_ONCE(b) = q / 2; | |
636 | do_something(); | |
637 | } else { | |
638 | ACCESS_ONCE(b) = q / 3; | |
639 | do_something_else(); | |
640 | } | |
641 | ||
642 | This will again ensure that the load from variable 'a' is ordered before the | |
643 | stores to variable 'b'. | |
644 | ||
645 | In addition, you need to be careful what you do with the local variable 'q', | |
646 | otherwise the compiler might be able to guess the value and again remove | |
647 | the needed conditional. For example: | |
648 | ||
649 | q = ACCESS_ONCE(a); | |
650 | if (q % MAX) { | |
651 | ACCESS_ONCE(b) = p; | |
652 | do_something(); | |
653 | } else { | |
654 | ACCESS_ONCE(b) = p; | |
655 | do_something_else(); | |
656 | } | |
657 | ||
658 | If MAX is defined to be 1, then the compiler knows that (q % MAX) is | |
659 | equal to zero, in which case the compiler is within its rights to | |
660 | transform the above code into the following: | |
661 | ||
662 | q = ACCESS_ONCE(a); | |
663 | ACCESS_ONCE(b) = p; | |
664 | do_something_else(); | |
665 | ||
666 | This transformation loses the ordering between the load from variable 'a' | |
667 | and the store to variable 'b'. If you are relying on this ordering, you | |
668 | should do something like the following: | |
669 | ||
670 | q = ACCESS_ONCE(a); | |
671 | BUILD_BUG_ON(MAX <= 1); /* Order load from a with store to b. */ | |
672 | if (q % MAX) { | |
673 | ACCESS_ONCE(b) = p; | |
674 | do_something(); | |
675 | } else { | |
676 | ACCESS_ONCE(b) = p; | |
677 | do_something_else(); | |
678 | } | |
679 | ||
680 | Finally, control dependencies do -not- provide transitivity. This is | |
681 | demonstrated by two related examples: | |
682 | ||
683 | CPU 0 CPU 1 | |
684 | ===================== ===================== | |
685 | r1 = ACCESS_ONCE(x); r2 = ACCESS_ONCE(y); | |
686 | if (r1 >= 0) if (r2 >= 0) | |
687 | ACCESS_ONCE(y) = 1; ACCESS_ONCE(x) = 1; | |
688 | ||
689 | assert(!(r1 == 1 && r2 == 1)); | |
690 | ||
691 | The above two-CPU example will never trigger the assert(). However, | |
692 | if control dependencies guaranteed transitivity (which they do not), | |
693 | then adding the following two CPUs would guarantee a related assertion: | |
694 | ||
695 | CPU 2 CPU 3 | |
696 | ===================== ===================== | |
697 | ACCESS_ONCE(x) = 2; ACCESS_ONCE(y) = 2; | |
698 | ||
699 | assert(!(r1 == 2 && r2 == 2 && x == 1 && y == 1)); /* FAILS!!! */ | |
700 | ||
701 | But because control dependencies do -not- provide transitivity, the | |
702 | above assertion can fail after the combined four-CPU example completes. | |
703 | If you need the four-CPU example to provide ordering, you will need | |
704 | smp_mb() between the loads and stores in the CPU 0 and CPU 1 code fragments. | |
705 | ||
706 | In summary: | |
707 | ||
708 | (*) Control dependencies can order prior loads against later stores. | |
709 | However, they do -not- guarantee any other sort of ordering: | |
710 | Not prior loads against later loads, nor prior stores against | |
711 | later anything. If you need these other forms of ordering, | |
712 | use smb_rmb(), smp_wmb(), or, in the case of prior stores and | |
713 | later loads, smp_mb(). | |
714 | ||
715 | (*) Control dependencies require at least one run-time conditional | |
716 | between the prior load and the subsequent store. If the compiler | |
717 | is able to optimize the conditional away, it will have also | |
718 | optimized away the ordering. Careful use of ACCESS_ONCE() can | |
719 | help to preserve the needed conditional. | |
720 | ||
721 | (*) Control dependencies require that the compiler avoid reordering the | |
722 | dependency into nonexistence. Careful use of ACCESS_ONCE() or | |
723 | barrier() can help to preserve your control dependency. Please | |
724 | see the Compiler Barrier section for more information. | |
725 | ||
726 | (*) Control dependencies do -not- provide transitivity. If you | |
727 | need transitivity, use smp_mb(). | |
728 | ||
729 | ||
730 | SMP BARRIER PAIRING | |
731 | ------------------- | |
732 | ||
733 | When dealing with CPU-CPU interactions, certain types of memory barrier should | |
734 | always be paired. A lack of appropriate pairing is almost certainly an error. | |
735 | ||
736 | A write barrier should always be paired with a data dependency barrier or read | |
737 | barrier, though a general barrier would also be viable. Similarly a read | |
738 | barrier or a data dependency barrier should always be paired with at least an | |
739 | write barrier, though, again, a general barrier is viable: | |
740 | ||
741 | CPU 1 CPU 2 | |
742 | =============== =============== | |
743 | ACCESS_ONCE(a) = 1; | |
744 | <write barrier> | |
745 | ACCESS_ONCE(b) = 2; x = ACCESS_ONCE(b); | |
746 | <read barrier> | |
747 | y = ACCESS_ONCE(a); | |
748 | ||
749 | Or: | |
750 | ||
751 | CPU 1 CPU 2 | |
752 | =============== =============================== | |
753 | a = 1; | |
754 | <write barrier> | |
755 | ACCESS_ONCE(b) = &a; x = ACCESS_ONCE(b); | |
756 | <data dependency barrier> | |
757 | y = *x; | |
758 | ||
759 | Basically, the read barrier always has to be there, even though it can be of | |
760 | the "weaker" type. | |
761 | ||
762 | [!] Note that the stores before the write barrier would normally be expected to | |
763 | match the loads after the read barrier or the data dependency barrier, and vice | |
764 | versa: | |
765 | ||
766 | CPU 1 CPU 2 | |
767 | =================== =================== | |
768 | ACCESS_ONCE(a) = 1; }---- --->{ v = ACCESS_ONCE(c); | |
769 | ACCESS_ONCE(b) = 2; } \ / { w = ACCESS_ONCE(d); | |
770 | <write barrier> \ <read barrier> | |
771 | ACCESS_ONCE(c) = 3; } / \ { x = ACCESS_ONCE(a); | |
772 | ACCESS_ONCE(d) = 4; }---- --->{ y = ACCESS_ONCE(b); | |
773 | ||
774 | ||
775 | EXAMPLES OF MEMORY BARRIER SEQUENCES | |
776 | ------------------------------------ | |
777 | ||
778 | Firstly, write barriers act as partial orderings on store operations. | |
779 | Consider the following sequence of events: | |
780 | ||
781 | CPU 1 | |
782 | ======================= | |
783 | STORE A = 1 | |
784 | STORE B = 2 | |
785 | STORE C = 3 | |
786 | <write barrier> | |
787 | STORE D = 4 | |
788 | STORE E = 5 | |
789 | ||
790 | This sequence of events is committed to the memory coherence system in an order | |
791 | that the rest of the system might perceive as the unordered set of { STORE A, | |
792 | STORE B, STORE C } all occurring before the unordered set of { STORE D, STORE E | |
793 | }: | |
794 | ||
795 | +-------+ : : | |
796 | | | +------+ | |
797 | | |------>| C=3 | } /\ | |
798 | | | : +------+ }----- \ -----> Events perceptible to | |
799 | | | : | A=1 | } \/ the rest of the system | |
800 | | | : +------+ } | |
801 | | CPU 1 | : | B=2 | } | |
802 | | | +------+ } | |
803 | | | wwwwwwwwwwwwwwww } <--- At this point the write barrier | |
804 | | | +------+ } requires all stores prior to the | |
805 | | | : | E=5 | } barrier to be committed before | |
806 | | | : +------+ } further stores may take place | |
807 | | |------>| D=4 | } | |
808 | | | +------+ | |
809 | +-------+ : : | |
810 | | | |
811 | | Sequence in which stores are committed to the | |
812 | | memory system by CPU 1 | |
813 | V | |
814 | ||
815 | ||
816 | Secondly, data dependency barriers act as partial orderings on data-dependent | |
817 | loads. Consider the following sequence of events: | |
818 | ||
819 | CPU 1 CPU 2 | |
820 | ======================= ======================= | |
821 | { B = 7; X = 9; Y = 8; C = &Y } | |
822 | STORE A = 1 | |
823 | STORE B = 2 | |
824 | <write barrier> | |
825 | STORE C = &B LOAD X | |
826 | STORE D = 4 LOAD C (gets &B) | |
827 | LOAD *C (reads B) | |
828 | ||
829 | Without intervention, CPU 2 may perceive the events on CPU 1 in some | |
830 | effectively random order, despite the write barrier issued by CPU 1: | |
831 | ||
832 | +-------+ : : : : | |
833 | | | +------+ +-------+ | Sequence of update | |
834 | | |------>| B=2 |----- --->| Y->8 | | of perception on | |
835 | | | : +------+ \ +-------+ | CPU 2 | |
836 | | CPU 1 | : | A=1 | \ --->| C->&Y | V | |
837 | | | +------+ | +-------+ | |
838 | | | wwwwwwwwwwwwwwww | : : | |
839 | | | +------+ | : : | |
840 | | | : | C=&B |--- | : : +-------+ | |
841 | | | : +------+ \ | +-------+ | | | |
842 | | |------>| D=4 | ----------->| C->&B |------>| | | |
843 | | | +------+ | +-------+ | | | |
844 | +-------+ : : | : : | | | |
845 | | : : | | | |
846 | | : : | CPU 2 | | |
847 | | +-------+ | | | |
848 | Apparently incorrect ---> | | B->7 |------>| | | |
849 | perception of B (!) | +-------+ | | | |
850 | | : : | | | |
851 | | +-------+ | | | |
852 | The load of X holds ---> \ | X->9 |------>| | | |
853 | up the maintenance \ +-------+ | | | |
854 | of coherence of B ----->| B->2 | +-------+ | |
855 | +-------+ | |
856 | : : | |
857 | ||
858 | ||
859 | In the above example, CPU 2 perceives that B is 7, despite the load of *C | |
860 | (which would be B) coming after the LOAD of C. | |
861 | ||
862 | If, however, a data dependency barrier were to be placed between the load of C | |
863 | and the load of *C (ie: B) on CPU 2: | |
864 | ||
865 | CPU 1 CPU 2 | |
866 | ======================= ======================= | |
867 | { B = 7; X = 9; Y = 8; C = &Y } | |
868 | STORE A = 1 | |
869 | STORE B = 2 | |
870 | <write barrier> | |
871 | STORE C = &B LOAD X | |
872 | STORE D = 4 LOAD C (gets &B) | |
873 | <data dependency barrier> | |
874 | LOAD *C (reads B) | |
875 | ||
876 | then the following will occur: | |
877 | ||
878 | +-------+ : : : : | |
879 | | | +------+ +-------+ | |
880 | | |------>| B=2 |----- --->| Y->8 | | |
881 | | | : +------+ \ +-------+ | |
882 | | CPU 1 | : | A=1 | \ --->| C->&Y | | |
883 | | | +------+ | +-------+ | |
884 | | | wwwwwwwwwwwwwwww | : : | |
885 | | | +------+ | : : | |
886 | | | : | C=&B |--- | : : +-------+ | |
887 | | | : +------+ \ | +-------+ | | | |
888 | | |------>| D=4 | ----------->| C->&B |------>| | | |
889 | | | +------+ | +-------+ | | | |
890 | +-------+ : : | : : | | | |
891 | | : : | | | |
892 | | : : | CPU 2 | | |
893 | | +-------+ | | | |
894 | | | X->9 |------>| | | |
895 | | +-------+ | | | |
896 | Makes sure all effects ---> \ ddddddddddddddddd | | | |
897 | prior to the store of C \ +-------+ | | | |
898 | are perceptible to ----->| B->2 |------>| | | |
899 | subsequent loads +-------+ | | | |
900 | : : +-------+ | |
901 | ||
902 | ||
903 | And thirdly, a read barrier acts as a partial order on loads. Consider the | |
904 | following sequence of events: | |
905 | ||
906 | CPU 1 CPU 2 | |
907 | ======================= ======================= | |
908 | { A = 0, B = 9 } | |
909 | STORE A=1 | |
910 | <write barrier> | |
911 | STORE B=2 | |
912 | LOAD B | |
913 | LOAD A | |
914 | ||
915 | Without intervention, CPU 2 may then choose to perceive the events on CPU 1 in | |
916 | some effectively random order, despite the write barrier issued by CPU 1: | |
917 | ||
918 | +-------+ : : : : | |
919 | | | +------+ +-------+ | |
920 | | |------>| A=1 |------ --->| A->0 | | |
921 | | | +------+ \ +-------+ | |
922 | | CPU 1 | wwwwwwwwwwwwwwww \ --->| B->9 | | |
923 | | | +------+ | +-------+ | |
924 | | |------>| B=2 |--- | : : | |
925 | | | +------+ \ | : : +-------+ | |
926 | +-------+ : : \ | +-------+ | | | |
927 | ---------->| B->2 |------>| | | |
928 | | +-------+ | CPU 2 | | |
929 | | | A->0 |------>| | | |
930 | | +-------+ | | | |
931 | | : : +-------+ | |
932 | \ : : | |
933 | \ +-------+ | |
934 | ---->| A->1 | | |
935 | +-------+ | |
936 | : : | |
937 | ||
938 | ||
939 | If, however, a read barrier were to be placed between the load of B and the | |
940 | load of A on CPU 2: | |
941 | ||
942 | CPU 1 CPU 2 | |
943 | ======================= ======================= | |
944 | { A = 0, B = 9 } | |
945 | STORE A=1 | |
946 | <write barrier> | |
947 | STORE B=2 | |
948 | LOAD B | |
949 | <read barrier> | |
950 | LOAD A | |
951 | ||
952 | then the partial ordering imposed by CPU 1 will be perceived correctly by CPU | |
953 | 2: | |
954 | ||
955 | +-------+ : : : : | |
956 | | | +------+ +-------+ | |
957 | | |------>| A=1 |------ --->| A->0 | | |
958 | | | +------+ \ +-------+ | |
959 | | CPU 1 | wwwwwwwwwwwwwwww \ --->| B->9 | | |
960 | | | +------+ | +-------+ | |
961 | | |------>| B=2 |--- | : : | |
962 | | | +------+ \ | : : +-------+ | |
963 | +-------+ : : \ | +-------+ | | | |
964 | ---------->| B->2 |------>| | | |
965 | | +-------+ | CPU 2 | | |
966 | | : : | | | |
967 | | : : | | | |
968 | At this point the read ----> \ rrrrrrrrrrrrrrrrr | | | |
969 | barrier causes all effects \ +-------+ | | | |
970 | prior to the storage of B ---->| A->1 |------>| | | |
971 | to be perceptible to CPU 2 +-------+ | | | |
972 | : : +-------+ | |
973 | ||
974 | ||
975 | To illustrate this more completely, consider what could happen if the code | |
976 | contained a load of A either side of the read barrier: | |
977 | ||
978 | CPU 1 CPU 2 | |
979 | ======================= ======================= | |
980 | { A = 0, B = 9 } | |
981 | STORE A=1 | |
982 | <write barrier> | |
983 | STORE B=2 | |
984 | LOAD B | |
985 | LOAD A [first load of A] | |
986 | <read barrier> | |
987 | LOAD A [second load of A] | |
988 | ||
989 | Even though the two loads of A both occur after the load of B, they may both | |
990 | come up with different values: | |
991 | ||
992 | +-------+ : : : : | |
993 | | | +------+ +-------+ | |
994 | | |------>| A=1 |------ --->| A->0 | | |
995 | | | +------+ \ +-------+ | |
996 | | CPU 1 | wwwwwwwwwwwwwwww \ --->| B->9 | | |
997 | | | +------+ | +-------+ | |
998 | | |------>| B=2 |--- | : : | |
999 | | | +------+ \ | : : +-------+ | |
1000 | +-------+ : : \ | +-------+ | | | |
1001 | ---------->| B->2 |------>| | | |
1002 | | +-------+ | CPU 2 | | |
1003 | | : : | | | |
1004 | | : : | | | |
1005 | | +-------+ | | | |
1006 | | | A->0 |------>| 1st | | |
1007 | | +-------+ | | | |
1008 | At this point the read ----> \ rrrrrrrrrrrrrrrrr | | | |
1009 | barrier causes all effects \ +-------+ | | | |
1010 | prior to the storage of B ---->| A->1 |------>| 2nd | | |
1011 | to be perceptible to CPU 2 +-------+ | | | |
1012 | : : +-------+ | |
1013 | ||
1014 | ||
1015 | But it may be that the update to A from CPU 1 becomes perceptible to CPU 2 | |
1016 | before the read barrier completes anyway: | |
1017 | ||
1018 | +-------+ : : : : | |
1019 | | | +------+ +-------+ | |
1020 | | |------>| A=1 |------ --->| A->0 | | |
1021 | | | +------+ \ +-------+ | |
1022 | | CPU 1 | wwwwwwwwwwwwwwww \ --->| B->9 | | |
1023 | | | +------+ | +-------+ | |
1024 | | |------>| B=2 |--- | : : | |
1025 | | | +------+ \ | : : +-------+ | |
1026 | +-------+ : : \ | +-------+ | | | |
1027 | ---------->| B->2 |------>| | | |
1028 | | +-------+ | CPU 2 | | |
1029 | | : : | | | |
1030 | \ : : | | | |
1031 | \ +-------+ | | | |
1032 | ---->| A->1 |------>| 1st | | |
1033 | +-------+ | | | |
1034 | rrrrrrrrrrrrrrrrr | | | |
1035 | +-------+ | | | |
1036 | | A->1 |------>| 2nd | | |
1037 | +-------+ | | | |
1038 | : : +-------+ | |
1039 | ||
1040 | ||
1041 | The guarantee is that the second load will always come up with A == 1 if the | |
1042 | load of B came up with B == 2. No such guarantee exists for the first load of | |
1043 | A; that may come up with either A == 0 or A == 1. | |
1044 | ||
1045 | ||
1046 | READ MEMORY BARRIERS VS LOAD SPECULATION | |
1047 | ---------------------------------------- | |
1048 | ||
1049 | Many CPUs speculate with loads: that is they see that they will need to load an | |
1050 | item from memory, and they find a time where they're not using the bus for any | |
1051 | other loads, and so do the load in advance - even though they haven't actually | |
1052 | got to that point in the instruction execution flow yet. This permits the | |
1053 | actual load instruction to potentially complete immediately because the CPU | |
1054 | already has the value to hand. | |
1055 | ||
1056 | It may turn out that the CPU didn't actually need the value - perhaps because a | |
1057 | branch circumvented the load - in which case it can discard the value or just | |
1058 | cache it for later use. | |
1059 | ||
1060 | Consider: | |
1061 | ||
1062 | CPU 1 CPU 2 | |
1063 | ======================= ======================= | |
1064 | LOAD B | |
1065 | DIVIDE } Divide instructions generally | |
1066 | DIVIDE } take a long time to perform | |
1067 | LOAD A | |
1068 | ||
1069 | Which might appear as this: | |
1070 | ||
1071 | : : +-------+ | |
1072 | +-------+ | | | |
1073 | --->| B->2 |------>| | | |
1074 | +-------+ | CPU 2 | | |
1075 | : :DIVIDE | | | |
1076 | +-------+ | | | |
1077 | The CPU being busy doing a ---> --->| A->0 |~~~~ | | | |
1078 | division speculates on the +-------+ ~ | | | |
1079 | LOAD of A : : ~ | | | |
1080 | : :DIVIDE | | | |
1081 | : : ~ | | | |
1082 | Once the divisions are complete --> : : ~-->| | | |
1083 | the CPU can then perform the : : | | | |
1084 | LOAD with immediate effect : : +-------+ | |
1085 | ||
1086 | ||
1087 | Placing a read barrier or a data dependency barrier just before the second | |
1088 | load: | |
1089 | ||
1090 | CPU 1 CPU 2 | |
1091 | ======================= ======================= | |
1092 | LOAD B | |
1093 | DIVIDE | |
1094 | DIVIDE | |
1095 | <read barrier> | |
1096 | LOAD A | |
1097 | ||
1098 | will force any value speculatively obtained to be reconsidered to an extent | |
1099 | dependent on the type of barrier used. If there was no change made to the | |
1100 | speculated memory location, then the speculated value will just be used: | |
1101 | ||
1102 | : : +-------+ | |
1103 | +-------+ | | | |
1104 | --->| B->2 |------>| | | |
1105 | +-------+ | CPU 2 | | |
1106 | : :DIVIDE | | | |
1107 | +-------+ | | | |
1108 | The CPU being busy doing a ---> --->| A->0 |~~~~ | | | |
1109 | division speculates on the +-------+ ~ | | | |
1110 | LOAD of A : : ~ | | | |
1111 | : :DIVIDE | | | |
1112 | : : ~ | | | |
1113 | : : ~ | | | |
1114 | rrrrrrrrrrrrrrrr~ | | | |
1115 | : : ~ | | | |
1116 | : : ~-->| | | |
1117 | : : | | | |
1118 | : : +-------+ | |
1119 | ||
1120 | ||
1121 | but if there was an update or an invalidation from another CPU pending, then | |
1122 | the speculation will be cancelled and the value reloaded: | |
1123 | ||
1124 | : : +-------+ | |
1125 | +-------+ | | | |
1126 | --->| B->2 |------>| | | |
1127 | +-------+ | CPU 2 | | |
1128 | : :DIVIDE | | | |
1129 | +-------+ | | | |
1130 | The CPU being busy doing a ---> --->| A->0 |~~~~ | | | |
1131 | division speculates on the +-------+ ~ | | | |
1132 | LOAD of A : : ~ | | | |
1133 | : :DIVIDE | | | |
1134 | : : ~ | | | |
1135 | : : ~ | | | |
1136 | rrrrrrrrrrrrrrrrr | | | |
1137 | +-------+ | | | |
1138 | The speculation is discarded ---> --->| A->1 |------>| | | |
1139 | and an updated value is +-------+ | | | |
1140 | retrieved : : +-------+ | |
1141 | ||
1142 | ||
1143 | TRANSITIVITY | |
1144 | ------------ | |
1145 | ||
1146 | Transitivity is a deeply intuitive notion about ordering that is not | |
1147 | always provided by real computer systems. The following example | |
1148 | demonstrates transitivity (also called "cumulativity"): | |
1149 | ||
1150 | CPU 1 CPU 2 CPU 3 | |
1151 | ======================= ======================= ======================= | |
1152 | { X = 0, Y = 0 } | |
1153 | STORE X=1 LOAD X STORE Y=1 | |
1154 | <general barrier> <general barrier> | |
1155 | LOAD Y LOAD X | |
1156 | ||
1157 | Suppose that CPU 2's load from X returns 1 and its load from Y returns 0. | |
1158 | This indicates that CPU 2's load from X in some sense follows CPU 1's | |
1159 | store to X and that CPU 2's load from Y in some sense preceded CPU 3's | |
1160 | store to Y. The question is then "Can CPU 3's load from X return 0?" | |
1161 | ||
1162 | Because CPU 2's load from X in some sense came after CPU 1's store, it | |
1163 | is natural to expect that CPU 3's load from X must therefore return 1. | |
1164 | This expectation is an example of transitivity: if a load executing on | |
1165 | CPU A follows a load from the same variable executing on CPU B, then | |
1166 | CPU A's load must either return the same value that CPU B's load did, | |
1167 | or must return some later value. | |
1168 | ||
1169 | In the Linux kernel, use of general memory barriers guarantees | |
1170 | transitivity. Therefore, in the above example, if CPU 2's load from X | |
1171 | returns 1 and its load from Y returns 0, then CPU 3's load from X must | |
1172 | also return 1. | |
1173 | ||
1174 | However, transitivity is -not- guaranteed for read or write barriers. | |
1175 | For example, suppose that CPU 2's general barrier in the above example | |
1176 | is changed to a read barrier as shown below: | |
1177 | ||
1178 | CPU 1 CPU 2 CPU 3 | |
1179 | ======================= ======================= ======================= | |
1180 | { X = 0, Y = 0 } | |
1181 | STORE X=1 LOAD X STORE Y=1 | |
1182 | <read barrier> <general barrier> | |
1183 | LOAD Y LOAD X | |
1184 | ||
1185 | This substitution destroys transitivity: in this example, it is perfectly | |
1186 | legal for CPU 2's load from X to return 1, its load from Y to return 0, | |
1187 | and CPU 3's load from X to return 0. | |
1188 | ||
1189 | The key point is that although CPU 2's read barrier orders its pair | |
1190 | of loads, it does not guarantee to order CPU 1's store. Therefore, if | |
1191 | this example runs on a system where CPUs 1 and 2 share a store buffer | |
1192 | or a level of cache, CPU 2 might have early access to CPU 1's writes. | |
1193 | General barriers are therefore required to ensure that all CPUs agree | |
1194 | on the combined order of CPU 1's and CPU 2's accesses. | |
1195 | ||
1196 | To reiterate, if your code requires transitivity, use general barriers | |
1197 | throughout. | |
1198 | ||
1199 | ||
1200 | ======================== | |
1201 | EXPLICIT KERNEL BARRIERS | |
1202 | ======================== | |
1203 | ||
1204 | The Linux kernel has a variety of different barriers that act at different | |
1205 | levels: | |
1206 | ||
1207 | (*) Compiler barrier. | |
1208 | ||
1209 | (*) CPU memory barriers. | |
1210 | ||
1211 | (*) MMIO write barrier. | |
1212 | ||
1213 | ||
1214 | COMPILER BARRIER | |
1215 | ---------------- | |
1216 | ||
1217 | The Linux kernel has an explicit compiler barrier function that prevents the | |
1218 | compiler from moving the memory accesses either side of it to the other side: | |
1219 | ||
1220 | barrier(); | |
1221 | ||
1222 | This is a general barrier -- there are no read-read or write-write variants | |
1223 | of barrier(). However, ACCESS_ONCE() can be thought of as a weak form | |
1224 | for barrier() that affects only the specific accesses flagged by the | |
1225 | ACCESS_ONCE(). | |
1226 | ||
1227 | The barrier() function has the following effects: | |
1228 | ||
1229 | (*) Prevents the compiler from reordering accesses following the | |
1230 | barrier() to precede any accesses preceding the barrier(). | |
1231 | One example use for this property is to ease communication between | |
1232 | interrupt-handler code and the code that was interrupted. | |
1233 | ||
1234 | (*) Within a loop, forces the compiler to load the variables used | |
1235 | in that loop's conditional on each pass through that loop. | |
1236 | ||
1237 | The ACCESS_ONCE() function can prevent any number of optimizations that, | |
1238 | while perfectly safe in single-threaded code, can be fatal in concurrent | |
1239 | code. Here are some examples of these sorts of optimizations: | |
1240 | ||
1241 | (*) The compiler is within its rights to merge successive loads from | |
1242 | the same variable. Such merging can cause the compiler to "optimize" | |
1243 | the following code: | |
1244 | ||
1245 | while (tmp = a) | |
1246 | do_something_with(tmp); | |
1247 | ||
1248 | into the following code, which, although in some sense legitimate | |
1249 | for single-threaded code, is almost certainly not what the developer | |
1250 | intended: | |
1251 | ||
1252 | if (tmp = a) | |
1253 | for (;;) | |
1254 | do_something_with(tmp); | |
1255 | ||
1256 | Use ACCESS_ONCE() to prevent the compiler from doing this to you: | |
1257 | ||
1258 | while (tmp = ACCESS_ONCE(a)) | |
1259 | do_something_with(tmp); | |
1260 | ||
1261 | (*) The compiler is within its rights to reload a variable, for example, | |
1262 | in cases where high register pressure prevents the compiler from | |
1263 | keeping all data of interest in registers. The compiler might | |
1264 | therefore optimize the variable 'tmp' out of our previous example: | |
1265 | ||
1266 | while (tmp = a) | |
1267 | do_something_with(tmp); | |
1268 | ||
1269 | This could result in the following code, which is perfectly safe in | |
1270 | single-threaded code, but can be fatal in concurrent code: | |
1271 | ||
1272 | while (a) | |
1273 | do_something_with(a); | |
1274 | ||
1275 | For example, the optimized version of this code could result in | |
1276 | passing a zero to do_something_with() in the case where the variable | |
1277 | a was modified by some other CPU between the "while" statement and | |
1278 | the call to do_something_with(). | |
1279 | ||
1280 | Again, use ACCESS_ONCE() to prevent the compiler from doing this: | |
1281 | ||
1282 | while (tmp = ACCESS_ONCE(a)) | |
1283 | do_something_with(tmp); | |
1284 | ||
1285 | Note that if the compiler runs short of registers, it might save | |
1286 | tmp onto the stack. The overhead of this saving and later restoring | |
1287 | is why compilers reload variables. Doing so is perfectly safe for | |
1288 | single-threaded code, so you need to tell the compiler about cases | |
1289 | where it is not safe. | |
1290 | ||
1291 | (*) The compiler is within its rights to omit a load entirely if it knows | |
1292 | what the value will be. For example, if the compiler can prove that | |
1293 | the value of variable 'a' is always zero, it can optimize this code: | |
1294 | ||
1295 | while (tmp = a) | |
1296 | do_something_with(tmp); | |
1297 | ||
1298 | Into this: | |
1299 | ||
1300 | do { } while (0); | |
1301 | ||
1302 | This transformation is a win for single-threaded code because it gets | |
1303 | rid of a load and a branch. The problem is that the compiler will | |
1304 | carry out its proof assuming that the current CPU is the only one | |
1305 | updating variable 'a'. If variable 'a' is shared, then the compiler's | |
1306 | proof will be erroneous. Use ACCESS_ONCE() to tell the compiler | |
1307 | that it doesn't know as much as it thinks it does: | |
1308 | ||
1309 | while (tmp = ACCESS_ONCE(a)) | |
1310 | do_something_with(tmp); | |
1311 | ||
1312 | But please note that the compiler is also closely watching what you | |
1313 | do with the value after the ACCESS_ONCE(). For example, suppose you | |
1314 | do the following and MAX is a preprocessor macro with the value 1: | |
1315 | ||
1316 | while ((tmp = ACCESS_ONCE(a)) % MAX) | |
1317 | do_something_with(tmp); | |
1318 | ||
1319 | Then the compiler knows that the result of the "%" operator applied | |
1320 | to MAX will always be zero, again allowing the compiler to optimize | |
1321 | the code into near-nonexistence. (It will still load from the | |
1322 | variable 'a'.) | |
1323 | ||
1324 | (*) Similarly, the compiler is within its rights to omit a store entirely | |
1325 | if it knows that the variable already has the value being stored. | |
1326 | Again, the compiler assumes that the current CPU is the only one | |
1327 | storing into the variable, which can cause the compiler to do the | |
1328 | wrong thing for shared variables. For example, suppose you have | |
1329 | the following: | |
1330 | ||
1331 | a = 0; | |
1332 | /* Code that does not store to variable a. */ | |
1333 | a = 0; | |
1334 | ||
1335 | The compiler sees that the value of variable 'a' is already zero, so | |
1336 | it might well omit the second store. This would come as a fatal | |
1337 | surprise if some other CPU might have stored to variable 'a' in the | |
1338 | meantime. | |
1339 | ||
1340 | Use ACCESS_ONCE() to prevent the compiler from making this sort of | |
1341 | wrong guess: | |
1342 | ||
1343 | ACCESS_ONCE(a) = 0; | |
1344 | /* Code that does not store to variable a. */ | |
1345 | ACCESS_ONCE(a) = 0; | |
1346 | ||
1347 | (*) The compiler is within its rights to reorder memory accesses unless | |
1348 | you tell it not to. For example, consider the following interaction | |
1349 | between process-level code and an interrupt handler: | |
1350 | ||
1351 | void process_level(void) | |
1352 | { | |
1353 | msg = get_message(); | |
1354 | flag = true; | |
1355 | } | |
1356 | ||
1357 | void interrupt_handler(void) | |
1358 | { | |
1359 | if (flag) | |
1360 | process_message(msg); | |
1361 | } | |
1362 | ||
1363 | There is nothing to prevent the the compiler from transforming | |
1364 | process_level() to the following, in fact, this might well be a | |
1365 | win for single-threaded code: | |
1366 | ||
1367 | void process_level(void) | |
1368 | { | |
1369 | flag = true; | |
1370 | msg = get_message(); | |
1371 | } | |
1372 | ||
1373 | If the interrupt occurs between these two statement, then | |
1374 | interrupt_handler() might be passed a garbled msg. Use ACCESS_ONCE() | |
1375 | to prevent this as follows: | |
1376 | ||
1377 | void process_level(void) | |
1378 | { | |
1379 | ACCESS_ONCE(msg) = get_message(); | |
1380 | ACCESS_ONCE(flag) = true; | |
1381 | } | |
1382 | ||
1383 | void interrupt_handler(void) | |
1384 | { | |
1385 | if (ACCESS_ONCE(flag)) | |
1386 | process_message(ACCESS_ONCE(msg)); | |
1387 | } | |
1388 | ||
1389 | Note that the ACCESS_ONCE() wrappers in interrupt_handler() | |
1390 | are needed if this interrupt handler can itself be interrupted | |
1391 | by something that also accesses 'flag' and 'msg', for example, | |
1392 | a nested interrupt or an NMI. Otherwise, ACCESS_ONCE() is not | |
1393 | needed in interrupt_handler() other than for documentation purposes. | |
1394 | (Note also that nested interrupts do not typically occur in modern | |
1395 | Linux kernels, in fact, if an interrupt handler returns with | |
1396 | interrupts enabled, you will get a WARN_ONCE() splat.) | |
1397 | ||
1398 | You should assume that the compiler can move ACCESS_ONCE() past | |
1399 | code not containing ACCESS_ONCE(), barrier(), or similar primitives. | |
1400 | ||
1401 | This effect could also be achieved using barrier(), but ACCESS_ONCE() | |
1402 | is more selective: With ACCESS_ONCE(), the compiler need only forget | |
1403 | the contents of the indicated memory locations, while with barrier() | |
1404 | the compiler must discard the value of all memory locations that | |
1405 | it has currented cached in any machine registers. Of course, | |
1406 | the compiler must also respect the order in which the ACCESS_ONCE()s | |
1407 | occur, though the CPU of course need not do so. | |
1408 | ||
1409 | (*) The compiler is within its rights to invent stores to a variable, | |
1410 | as in the following example: | |
1411 | ||
1412 | if (a) | |
1413 | b = a; | |
1414 | else | |
1415 | b = 42; | |
1416 | ||
1417 | The compiler might save a branch by optimizing this as follows: | |
1418 | ||
1419 | b = 42; | |
1420 | if (a) | |
1421 | b = a; | |
1422 | ||
1423 | In single-threaded code, this is not only safe, but also saves | |
1424 | a branch. Unfortunately, in concurrent code, this optimization | |
1425 | could cause some other CPU to see a spurious value of 42 -- even | |
1426 | if variable 'a' was never zero -- when loading variable 'b'. | |
1427 | Use ACCESS_ONCE() to prevent this as follows: | |
1428 | ||
1429 | if (a) | |
1430 | ACCESS_ONCE(b) = a; | |
1431 | else | |
1432 | ACCESS_ONCE(b) = 42; | |
1433 | ||
1434 | The compiler can also invent loads. These are usually less | |
1435 | damaging, but they can result in cache-line bouncing and thus in | |
1436 | poor performance and scalability. Use ACCESS_ONCE() to prevent | |
1437 | invented loads. | |
1438 | ||
1439 | (*) For aligned memory locations whose size allows them to be accessed | |
1440 | with a single memory-reference instruction, prevents "load tearing" | |
1441 | and "store tearing," in which a single large access is replaced by | |
1442 | multiple smaller accesses. For example, given an architecture having | |
1443 | 16-bit store instructions with 7-bit immediate fields, the compiler | |
1444 | might be tempted to use two 16-bit store-immediate instructions to | |
1445 | implement the following 32-bit store: | |
1446 | ||
1447 | p = 0x00010002; | |
1448 | ||
1449 | Please note that GCC really does use this sort of optimization, | |
1450 | which is not surprising given that it would likely take more | |
1451 | than two instructions to build the constant and then store it. | |
1452 | This optimization can therefore be a win in single-threaded code. | |
1453 | In fact, a recent bug (since fixed) caused GCC to incorrectly use | |
1454 | this optimization in a volatile store. In the absence of such bugs, | |
1455 | use of ACCESS_ONCE() prevents store tearing in the following example: | |
1456 | ||
1457 | ACCESS_ONCE(p) = 0x00010002; | |
1458 | ||
1459 | Use of packed structures can also result in load and store tearing, | |
1460 | as in this example: | |
1461 | ||
1462 | struct __attribute__((__packed__)) foo { | |
1463 | short a; | |
1464 | int b; | |
1465 | short c; | |
1466 | }; | |
1467 | struct foo foo1, foo2; | |
1468 | ... | |
1469 | ||
1470 | foo2.a = foo1.a; | |
1471 | foo2.b = foo1.b; | |
1472 | foo2.c = foo1.c; | |
1473 | ||
1474 | Because there are no ACCESS_ONCE() wrappers and no volatile markings, | |
1475 | the compiler would be well within its rights to implement these three | |
1476 | assignment statements as a pair of 32-bit loads followed by a pair | |
1477 | of 32-bit stores. This would result in load tearing on 'foo1.b' | |
1478 | and store tearing on 'foo2.b'. ACCESS_ONCE() again prevents tearing | |
1479 | in this example: | |
1480 | ||
1481 | foo2.a = foo1.a; | |
1482 | ACCESS_ONCE(foo2.b) = ACCESS_ONCE(foo1.b); | |
1483 | foo2.c = foo1.c; | |
1484 | ||
1485 | All that aside, it is never necessary to use ACCESS_ONCE() on a variable | |
1486 | that has been marked volatile. For example, because 'jiffies' is marked | |
1487 | volatile, it is never necessary to say ACCESS_ONCE(jiffies). The reason | |
1488 | for this is that ACCESS_ONCE() is implemented as a volatile cast, which | |
1489 | has no effect when its argument is already marked volatile. | |
1490 | ||
1491 | Please note that these compiler barriers have no direct effect on the CPU, | |
1492 | which may then reorder things however it wishes. | |
1493 | ||
1494 | ||
1495 | CPU MEMORY BARRIERS | |
1496 | ------------------- | |
1497 | ||
1498 | The Linux kernel has eight basic CPU memory barriers: | |
1499 | ||
1500 | TYPE MANDATORY SMP CONDITIONAL | |
1501 | =============== ======================= =========================== | |
1502 | GENERAL mb() smp_mb() | |
1503 | WRITE wmb() smp_wmb() | |
1504 | READ rmb() smp_rmb() | |
1505 | DATA DEPENDENCY read_barrier_depends() smp_read_barrier_depends() | |
1506 | ||
1507 | ||
1508 | All memory barriers except the data dependency barriers imply a compiler | |
1509 | barrier. Data dependencies do not impose any additional compiler ordering. | |
1510 | ||
1511 | Aside: In the case of data dependencies, the compiler would be expected to | |
1512 | issue the loads in the correct order (eg. `a[b]` would have to load the value | |
1513 | of b before loading a[b]), however there is no guarantee in the C specification | |
1514 | that the compiler may not speculate the value of b (eg. is equal to 1) and load | |
1515 | a before b (eg. tmp = a[1]; if (b != 1) tmp = a[b]; ). There is also the | |
1516 | problem of a compiler reloading b after having loaded a[b], thus having a newer | |
1517 | copy of b than a[b]. A consensus has not yet been reached about these problems, | |
1518 | however the ACCESS_ONCE macro is a good place to start looking. | |
1519 | ||
1520 | SMP memory barriers are reduced to compiler barriers on uniprocessor compiled | |
1521 | systems because it is assumed that a CPU will appear to be self-consistent, | |
1522 | and will order overlapping accesses correctly with respect to itself. | |
1523 | ||
1524 | [!] Note that SMP memory barriers _must_ be used to control the ordering of | |
1525 | references to shared memory on SMP systems, though the use of locking instead | |
1526 | is sufficient. | |
1527 | ||
1528 | Mandatory barriers should not be used to control SMP effects, since mandatory | |
1529 | barriers unnecessarily impose overhead on UP systems. They may, however, be | |
1530 | used to control MMIO effects on accesses through relaxed memory I/O windows. | |
1531 | These are required even on non-SMP systems as they affect the order in which | |
1532 | memory operations appear to a device by prohibiting both the compiler and the | |
1533 | CPU from reordering them. | |
1534 | ||
1535 | ||
1536 | There are some more advanced barrier functions: | |
1537 | ||
1538 | (*) set_mb(var, value) | |
1539 | ||
1540 | This assigns the value to the variable and then inserts a full memory | |
1541 | barrier after it, depending on the function. It isn't guaranteed to | |
1542 | insert anything more than a compiler barrier in a UP compilation. | |
1543 | ||
1544 | ||
1545 | (*) smp_mb__before_atomic_dec(); | |
1546 | (*) smp_mb__after_atomic_dec(); | |
1547 | (*) smp_mb__before_atomic_inc(); | |
1548 | (*) smp_mb__after_atomic_inc(); | |
1549 | ||
1550 | These are for use with atomic add, subtract, increment and decrement | |
1551 | functions that don't return a value, especially when used for reference | |
1552 | counting. These functions do not imply memory barriers. | |
1553 | ||
1554 | As an example, consider a piece of code that marks an object as being dead | |
1555 | and then decrements the object's reference count: | |
1556 | ||
1557 | obj->dead = 1; | |
1558 | smp_mb__before_atomic_dec(); | |
1559 | atomic_dec(&obj->ref_count); | |
1560 | ||
1561 | This makes sure that the death mark on the object is perceived to be set | |
1562 | *before* the reference counter is decremented. | |
1563 | ||
1564 | See Documentation/atomic_ops.txt for more information. See the "Atomic | |
1565 | operations" subsection for information on where to use these. | |
1566 | ||
1567 | ||
1568 | (*) smp_mb__before_clear_bit(void); | |
1569 | (*) smp_mb__after_clear_bit(void); | |
1570 | ||
1571 | These are for use similar to the atomic inc/dec barriers. These are | |
1572 | typically used for bitwise unlocking operations, so care must be taken as | |
1573 | there are no implicit memory barriers here either. | |
1574 | ||
1575 | Consider implementing an unlock operation of some nature by clearing a | |
1576 | locking bit. The clear_bit() would then need to be barriered like this: | |
1577 | ||
1578 | smp_mb__before_clear_bit(); | |
1579 | clear_bit( ... ); | |
1580 | ||
1581 | This prevents memory operations before the clear leaking to after it. See | |
1582 | the subsection on "Locking Functions" with reference to UNLOCK operation | |
1583 | implications. | |
1584 | ||
1585 | See Documentation/atomic_ops.txt for more information. See the "Atomic | |
1586 | operations" subsection for information on where to use these. | |
1587 | ||
1588 | ||
1589 | MMIO WRITE BARRIER | |
1590 | ------------------ | |
1591 | ||
1592 | The Linux kernel also has a special barrier for use with memory-mapped I/O | |
1593 | writes: | |
1594 | ||
1595 | mmiowb(); | |
1596 | ||
1597 | This is a variation on the mandatory write barrier that causes writes to weakly | |
1598 | ordered I/O regions to be partially ordered. Its effects may go beyond the | |
1599 | CPU->Hardware interface and actually affect the hardware at some level. | |
1600 | ||
1601 | See the subsection "Locks vs I/O accesses" for more information. | |
1602 | ||
1603 | ||
1604 | =============================== | |
1605 | IMPLICIT KERNEL MEMORY BARRIERS | |
1606 | =============================== | |
1607 | ||
1608 | Some of the other functions in the linux kernel imply memory barriers, amongst | |
1609 | which are locking and scheduling functions. | |
1610 | ||
1611 | This specification is a _minimum_ guarantee; any particular architecture may | |
1612 | provide more substantial guarantees, but these may not be relied upon outside | |
1613 | of arch specific code. | |
1614 | ||
1615 | ||
1616 | LOCKING FUNCTIONS | |
1617 | ----------------- | |
1618 | ||
1619 | The Linux kernel has a number of locking constructs: | |
1620 | ||
1621 | (*) spin locks | |
1622 | (*) R/W spin locks | |
1623 | (*) mutexes | |
1624 | (*) semaphores | |
1625 | (*) R/W semaphores | |
1626 | (*) RCU | |
1627 | ||
1628 | In all cases there are variants on "LOCK" operations and "UNLOCK" operations | |
1629 | for each construct. These operations all imply certain barriers: | |
1630 | ||
1631 | (1) LOCK operation implication: | |
1632 | ||
1633 | Memory operations issued after the LOCK will be completed after the LOCK | |
1634 | operation has completed. | |
1635 | ||
1636 | Memory operations issued before the LOCK may be completed after the LOCK | |
1637 | operation has completed. | |
1638 | ||
1639 | (2) UNLOCK operation implication: | |
1640 | ||
1641 | Memory operations issued before the UNLOCK will be completed before the | |
1642 | UNLOCK operation has completed. | |
1643 | ||
1644 | Memory operations issued after the UNLOCK may be completed before the | |
1645 | UNLOCK operation has completed. | |
1646 | ||
1647 | (3) LOCK vs LOCK implication: | |
1648 | ||
1649 | All LOCK operations issued before another LOCK operation will be completed | |
1650 | before that LOCK operation. | |
1651 | ||
1652 | (4) LOCK vs UNLOCK implication: | |
1653 | ||
1654 | All LOCK operations issued before an UNLOCK operation will be completed | |
1655 | before the UNLOCK operation. | |
1656 | ||
1657 | All UNLOCK operations issued before a LOCK operation will be completed | |
1658 | before the LOCK operation. | |
1659 | ||
1660 | (5) Failed conditional LOCK implication: | |
1661 | ||
1662 | Certain variants of the LOCK operation may fail, either due to being | |
1663 | unable to get the lock immediately, or due to receiving an unblocked | |
1664 | signal whilst asleep waiting for the lock to become available. Failed | |
1665 | locks do not imply any sort of barrier. | |
1666 | ||
1667 | Therefore, from (1), (2) and (4) an UNLOCK followed by an unconditional LOCK is | |
1668 | equivalent to a full barrier, but a LOCK followed by an UNLOCK is not. | |
1669 | ||
1670 | [!] Note: one of the consequences of LOCKs and UNLOCKs being only one-way | |
1671 | barriers is that the effects of instructions outside of a critical section | |
1672 | may seep into the inside of the critical section. | |
1673 | ||
1674 | A LOCK followed by an UNLOCK may not be assumed to be full memory barrier | |
1675 | because it is possible for an access preceding the LOCK to happen after the | |
1676 | LOCK, and an access following the UNLOCK to happen before the UNLOCK, and the | |
1677 | two accesses can themselves then cross: | |
1678 | ||
1679 | *A = a; | |
1680 | LOCK | |
1681 | UNLOCK | |
1682 | *B = b; | |
1683 | ||
1684 | may occur as: | |
1685 | ||
1686 | LOCK, STORE *B, STORE *A, UNLOCK | |
1687 | ||
1688 | Locks and semaphores may not provide any guarantee of ordering on UP compiled | |
1689 | systems, and so cannot be counted on in such a situation to actually achieve | |
1690 | anything at all - especially with respect to I/O accesses - unless combined | |
1691 | with interrupt disabling operations. | |
1692 | ||
1693 | See also the section on "Inter-CPU locking barrier effects". | |
1694 | ||
1695 | ||
1696 | As an example, consider the following: | |
1697 | ||
1698 | *A = a; | |
1699 | *B = b; | |
1700 | LOCK | |
1701 | *C = c; | |
1702 | *D = d; | |
1703 | UNLOCK | |
1704 | *E = e; | |
1705 | *F = f; | |
1706 | ||
1707 | The following sequence of events is acceptable: | |
1708 | ||
1709 | LOCK, {*F,*A}, *E, {*C,*D}, *B, UNLOCK | |
1710 | ||
1711 | [+] Note that {*F,*A} indicates a combined access. | |
1712 | ||
1713 | But none of the following are: | |
1714 | ||
1715 | {*F,*A}, *B, LOCK, *C, *D, UNLOCK, *E | |
1716 | *A, *B, *C, LOCK, *D, UNLOCK, *E, *F | |
1717 | *A, *B, LOCK, *C, UNLOCK, *D, *E, *F | |
1718 | *B, LOCK, *C, *D, UNLOCK, {*F,*A}, *E | |
1719 | ||
1720 | ||
1721 | ||
1722 | INTERRUPT DISABLING FUNCTIONS | |
1723 | ----------------------------- | |
1724 | ||
1725 | Functions that disable interrupts (LOCK equivalent) and enable interrupts | |
1726 | (UNLOCK equivalent) will act as compiler barriers only. So if memory or I/O | |
1727 | barriers are required in such a situation, they must be provided from some | |
1728 | other means. | |
1729 | ||
1730 | ||
1731 | SLEEP AND WAKE-UP FUNCTIONS | |
1732 | --------------------------- | |
1733 | ||
1734 | Sleeping and waking on an event flagged in global data can be viewed as an | |
1735 | interaction between two pieces of data: the task state of the task waiting for | |
1736 | the event and the global data used to indicate the event. To make sure that | |
1737 | these appear to happen in the right order, the primitives to begin the process | |
1738 | of going to sleep, and the primitives to initiate a wake up imply certain | |
1739 | barriers. | |
1740 | ||
1741 | Firstly, the sleeper normally follows something like this sequence of events: | |
1742 | ||
1743 | for (;;) { | |
1744 | set_current_state(TASK_UNINTERRUPTIBLE); | |
1745 | if (event_indicated) | |
1746 | break; | |
1747 | schedule(); | |
1748 | } | |
1749 | ||
1750 | A general memory barrier is interpolated automatically by set_current_state() | |
1751 | after it has altered the task state: | |
1752 | ||
1753 | CPU 1 | |
1754 | =============================== | |
1755 | set_current_state(); | |
1756 | set_mb(); | |
1757 | STORE current->state | |
1758 | <general barrier> | |
1759 | LOAD event_indicated | |
1760 | ||
1761 | set_current_state() may be wrapped by: | |
1762 | ||
1763 | prepare_to_wait(); | |
1764 | prepare_to_wait_exclusive(); | |
1765 | ||
1766 | which therefore also imply a general memory barrier after setting the state. | |
1767 | The whole sequence above is available in various canned forms, all of which | |
1768 | interpolate the memory barrier in the right place: | |
1769 | ||
1770 | wait_event(); | |
1771 | wait_event_interruptible(); | |
1772 | wait_event_interruptible_exclusive(); | |
1773 | wait_event_interruptible_timeout(); | |
1774 | wait_event_killable(); | |
1775 | wait_event_timeout(); | |
1776 | wait_on_bit(); | |
1777 | wait_on_bit_lock(); | |
1778 | ||
1779 | ||
1780 | Secondly, code that performs a wake up normally follows something like this: | |
1781 | ||
1782 | event_indicated = 1; | |
1783 | wake_up(&event_wait_queue); | |
1784 | ||
1785 | or: | |
1786 | ||
1787 | event_indicated = 1; | |
1788 | wake_up_process(event_daemon); | |
1789 | ||
1790 | A write memory barrier is implied by wake_up() and co. if and only if they wake | |
1791 | something up. The barrier occurs before the task state is cleared, and so sits | |
1792 | between the STORE to indicate the event and the STORE to set TASK_RUNNING: | |
1793 | ||
1794 | CPU 1 CPU 2 | |
1795 | =============================== =============================== | |
1796 | set_current_state(); STORE event_indicated | |
1797 | set_mb(); wake_up(); | |
1798 | STORE current->state <write barrier> | |
1799 | <general barrier> STORE current->state | |
1800 | LOAD event_indicated | |
1801 | ||
1802 | The available waker functions include: | |
1803 | ||
1804 | complete(); | |
1805 | wake_up(); | |
1806 | wake_up_all(); | |
1807 | wake_up_bit(); | |
1808 | wake_up_interruptible(); | |
1809 | wake_up_interruptible_all(); | |
1810 | wake_up_interruptible_nr(); | |
1811 | wake_up_interruptible_poll(); | |
1812 | wake_up_interruptible_sync(); | |
1813 | wake_up_interruptible_sync_poll(); | |
1814 | wake_up_locked(); | |
1815 | wake_up_locked_poll(); | |
1816 | wake_up_nr(); | |
1817 | wake_up_poll(); | |
1818 | wake_up_process(); | |
1819 | ||
1820 | ||
1821 | [!] Note that the memory barriers implied by the sleeper and the waker do _not_ | |
1822 | order multiple stores before the wake-up with respect to loads of those stored | |
1823 | values after the sleeper has called set_current_state(). For instance, if the | |
1824 | sleeper does: | |
1825 | ||
1826 | set_current_state(TASK_INTERRUPTIBLE); | |
1827 | if (event_indicated) | |
1828 | break; | |
1829 | __set_current_state(TASK_RUNNING); | |
1830 | do_something(my_data); | |
1831 | ||
1832 | and the waker does: | |
1833 | ||
1834 | my_data = value; | |
1835 | event_indicated = 1; | |
1836 | wake_up(&event_wait_queue); | |
1837 | ||
1838 | there's no guarantee that the change to event_indicated will be perceived by | |
1839 | the sleeper as coming after the change to my_data. In such a circumstance, the | |
1840 | code on both sides must interpolate its own memory barriers between the | |
1841 | separate data accesses. Thus the above sleeper ought to do: | |
1842 | ||
1843 | set_current_state(TASK_INTERRUPTIBLE); | |
1844 | if (event_indicated) { | |
1845 | smp_rmb(); | |
1846 | do_something(my_data); | |
1847 | } | |
1848 | ||
1849 | and the waker should do: | |
1850 | ||
1851 | my_data = value; | |
1852 | smp_wmb(); | |
1853 | event_indicated = 1; | |
1854 | wake_up(&event_wait_queue); | |
1855 | ||
1856 | ||
1857 | MISCELLANEOUS FUNCTIONS | |
1858 | ----------------------- | |
1859 | ||
1860 | Other functions that imply barriers: | |
1861 | ||
1862 | (*) schedule() and similar imply full memory barriers. | |
1863 | ||
1864 | ||
1865 | ================================= | |
1866 | INTER-CPU LOCKING BARRIER EFFECTS | |
1867 | ================================= | |
1868 | ||
1869 | On SMP systems locking primitives give a more substantial form of barrier: one | |
1870 | that does affect memory access ordering on other CPUs, within the context of | |
1871 | conflict on any particular lock. | |
1872 | ||
1873 | ||
1874 | LOCKS VS MEMORY ACCESSES | |
1875 | ------------------------ | |
1876 | ||
1877 | Consider the following: the system has a pair of spinlocks (M) and (Q), and | |
1878 | three CPUs; then should the following sequence of events occur: | |
1879 | ||
1880 | CPU 1 CPU 2 | |
1881 | =============================== =============================== | |
1882 | ACCESS_ONCE(*A) = a; ACCESS_ONCE(*E) = e; | |
1883 | LOCK M LOCK Q | |
1884 | ACCESS_ONCE(*B) = b; ACCESS_ONCE(*F) = f; | |
1885 | ACCESS_ONCE(*C) = c; ACCESS_ONCE(*G) = g; | |
1886 | UNLOCK M UNLOCK Q | |
1887 | ACCESS_ONCE(*D) = d; ACCESS_ONCE(*H) = h; | |
1888 | ||
1889 | Then there is no guarantee as to what order CPU 3 will see the accesses to *A | |
1890 | through *H occur in, other than the constraints imposed by the separate locks | |
1891 | on the separate CPUs. It might, for example, see: | |
1892 | ||
1893 | *E, LOCK M, LOCK Q, *G, *C, *F, *A, *B, UNLOCK Q, *D, *H, UNLOCK M | |
1894 | ||
1895 | But it won't see any of: | |
1896 | ||
1897 | *B, *C or *D preceding LOCK M | |
1898 | *A, *B or *C following UNLOCK M | |
1899 | *F, *G or *H preceding LOCK Q | |
1900 | *E, *F or *G following UNLOCK Q | |
1901 | ||
1902 | ||
1903 | However, if the following occurs: | |
1904 | ||
1905 | CPU 1 CPU 2 | |
1906 | =============================== =============================== | |
1907 | ACCESS_ONCE(*A) = a; | |
1908 | LOCK M [1] | |
1909 | ACCESS_ONCE(*B) = b; | |
1910 | ACCESS_ONCE(*C) = c; | |
1911 | UNLOCK M [1] | |
1912 | ACCESS_ONCE(*D) = d; ACCESS_ONCE(*E) = e; | |
1913 | LOCK M [2] | |
1914 | ACCESS_ONCE(*F) = f; | |
1915 | ACCESS_ONCE(*G) = g; | |
1916 | UNLOCK M [2] | |
1917 | ACCESS_ONCE(*H) = h; | |
1918 | ||
1919 | CPU 3 might see: | |
1920 | ||
1921 | *E, LOCK M [1], *C, *B, *A, UNLOCK M [1], | |
1922 | LOCK M [2], *H, *F, *G, UNLOCK M [2], *D | |
1923 | ||
1924 | But assuming CPU 1 gets the lock first, CPU 3 won't see any of: | |
1925 | ||
1926 | *B, *C, *D, *F, *G or *H preceding LOCK M [1] | |
1927 | *A, *B or *C following UNLOCK M [1] | |
1928 | *F, *G or *H preceding LOCK M [2] | |
1929 | *A, *B, *C, *E, *F or *G following UNLOCK M [2] | |
1930 | ||
1931 | ||
1932 | LOCKS VS I/O ACCESSES | |
1933 | --------------------- | |
1934 | ||
1935 | Under certain circumstances (especially involving NUMA), I/O accesses within | |
1936 | two spinlocked sections on two different CPUs may be seen as interleaved by the | |
1937 | PCI bridge, because the PCI bridge does not necessarily participate in the | |
1938 | cache-coherence protocol, and is therefore incapable of issuing the required | |
1939 | read memory barriers. | |
1940 | ||
1941 | For example: | |
1942 | ||
1943 | CPU 1 CPU 2 | |
1944 | =============================== =============================== | |
1945 | spin_lock(Q) | |
1946 | writel(0, ADDR) | |
1947 | writel(1, DATA); | |
1948 | spin_unlock(Q); | |
1949 | spin_lock(Q); | |
1950 | writel(4, ADDR); | |
1951 | writel(5, DATA); | |
1952 | spin_unlock(Q); | |
1953 | ||
1954 | may be seen by the PCI bridge as follows: | |
1955 | ||
1956 | STORE *ADDR = 0, STORE *ADDR = 4, STORE *DATA = 1, STORE *DATA = 5 | |
1957 | ||
1958 | which would probably cause the hardware to malfunction. | |
1959 | ||
1960 | ||
1961 | What is necessary here is to intervene with an mmiowb() before dropping the | |
1962 | spinlock, for example: | |
1963 | ||
1964 | CPU 1 CPU 2 | |
1965 | =============================== =============================== | |
1966 | spin_lock(Q) | |
1967 | writel(0, ADDR) | |
1968 | writel(1, DATA); | |
1969 | mmiowb(); | |
1970 | spin_unlock(Q); | |
1971 | spin_lock(Q); | |
1972 | writel(4, ADDR); | |
1973 | writel(5, DATA); | |
1974 | mmiowb(); | |
1975 | spin_unlock(Q); | |
1976 | ||
1977 | this will ensure that the two stores issued on CPU 1 appear at the PCI bridge | |
1978 | before either of the stores issued on CPU 2. | |
1979 | ||
1980 | ||
1981 | Furthermore, following a store by a load from the same device obviates the need | |
1982 | for the mmiowb(), because the load forces the store to complete before the load | |
1983 | is performed: | |
1984 | ||
1985 | CPU 1 CPU 2 | |
1986 | =============================== =============================== | |
1987 | spin_lock(Q) | |
1988 | writel(0, ADDR) | |
1989 | a = readl(DATA); | |
1990 | spin_unlock(Q); | |
1991 | spin_lock(Q); | |
1992 | writel(4, ADDR); | |
1993 | b = readl(DATA); | |
1994 | spin_unlock(Q); | |
1995 | ||
1996 | ||
1997 | See Documentation/DocBook/deviceiobook.tmpl for more information. | |
1998 | ||
1999 | ||
2000 | ================================= | |
2001 | WHERE ARE MEMORY BARRIERS NEEDED? | |
2002 | ================================= | |
2003 | ||
2004 | Under normal operation, memory operation reordering is generally not going to | |
2005 | be a problem as a single-threaded linear piece of code will still appear to | |
2006 | work correctly, even if it's in an SMP kernel. There are, however, four | |
2007 | circumstances in which reordering definitely _could_ be a problem: | |
2008 | ||
2009 | (*) Interprocessor interaction. | |
2010 | ||
2011 | (*) Atomic operations. | |
2012 | ||
2013 | (*) Accessing devices. | |
2014 | ||
2015 | (*) Interrupts. | |
2016 | ||
2017 | ||
2018 | INTERPROCESSOR INTERACTION | |
2019 | -------------------------- | |
2020 | ||
2021 | When there's a system with more than one processor, more than one CPU in the | |
2022 | system may be working on the same data set at the same time. This can cause | |
2023 | synchronisation problems, and the usual way of dealing with them is to use | |
2024 | locks. Locks, however, are quite expensive, and so it may be preferable to | |
2025 | operate without the use of a lock if at all possible. In such a case | |
2026 | operations that affect both CPUs may have to be carefully ordered to prevent | |
2027 | a malfunction. | |
2028 | ||
2029 | Consider, for example, the R/W semaphore slow path. Here a waiting process is | |
2030 | queued on the semaphore, by virtue of it having a piece of its stack linked to | |
2031 | the semaphore's list of waiting processes: | |
2032 | ||
2033 | struct rw_semaphore { | |
2034 | ... | |
2035 | spinlock_t lock; | |
2036 | struct list_head waiters; | |
2037 | }; | |
2038 | ||
2039 | struct rwsem_waiter { | |
2040 | struct list_head list; | |
2041 | struct task_struct *task; | |
2042 | }; | |
2043 | ||
2044 | To wake up a particular waiter, the up_read() or up_write() functions have to: | |
2045 | ||
2046 | (1) read the next pointer from this waiter's record to know as to where the | |
2047 | next waiter record is; | |
2048 | ||
2049 | (2) read the pointer to the waiter's task structure; | |
2050 | ||
2051 | (3) clear the task pointer to tell the waiter it has been given the semaphore; | |
2052 | ||
2053 | (4) call wake_up_process() on the task; and | |
2054 | ||
2055 | (5) release the reference held on the waiter's task struct. | |
2056 | ||
2057 | In other words, it has to perform this sequence of events: | |
2058 | ||
2059 | LOAD waiter->list.next; | |
2060 | LOAD waiter->task; | |
2061 | STORE waiter->task; | |
2062 | CALL wakeup | |
2063 | RELEASE task | |
2064 | ||
2065 | and if any of these steps occur out of order, then the whole thing may | |
2066 | malfunction. | |
2067 | ||
2068 | Once it has queued itself and dropped the semaphore lock, the waiter does not | |
2069 | get the lock again; it instead just waits for its task pointer to be cleared | |
2070 | before proceeding. Since the record is on the waiter's stack, this means that | |
2071 | if the task pointer is cleared _before_ the next pointer in the list is read, | |
2072 | another CPU might start processing the waiter and might clobber the waiter's | |
2073 | stack before the up*() function has a chance to read the next pointer. | |
2074 | ||
2075 | Consider then what might happen to the above sequence of events: | |
2076 | ||
2077 | CPU 1 CPU 2 | |
2078 | =============================== =============================== | |
2079 | down_xxx() | |
2080 | Queue waiter | |
2081 | Sleep | |
2082 | up_yyy() | |
2083 | LOAD waiter->task; | |
2084 | STORE waiter->task; | |
2085 | Woken up by other event | |
2086 | <preempt> | |
2087 | Resume processing | |
2088 | down_xxx() returns | |
2089 | call foo() | |
2090 | foo() clobbers *waiter | |
2091 | </preempt> | |
2092 | LOAD waiter->list.next; | |
2093 | --- OOPS --- | |
2094 | ||
2095 | This could be dealt with using the semaphore lock, but then the down_xxx() | |
2096 | function has to needlessly get the spinlock again after being woken up. | |
2097 | ||
2098 | The way to deal with this is to insert a general SMP memory barrier: | |
2099 | ||
2100 | LOAD waiter->list.next; | |
2101 | LOAD waiter->task; | |
2102 | smp_mb(); | |
2103 | STORE waiter->task; | |
2104 | CALL wakeup | |
2105 | RELEASE task | |
2106 | ||
2107 | In this case, the barrier makes a guarantee that all memory accesses before the | |
2108 | barrier will appear to happen before all the memory accesses after the barrier | |
2109 | with respect to the other CPUs on the system. It does _not_ guarantee that all | |
2110 | the memory accesses before the barrier will be complete by the time the barrier | |
2111 | instruction itself is complete. | |
2112 | ||
2113 | On a UP system - where this wouldn't be a problem - the smp_mb() is just a | |
2114 | compiler barrier, thus making sure the compiler emits the instructions in the | |
2115 | right order without actually intervening in the CPU. Since there's only one | |
2116 | CPU, that CPU's dependency ordering logic will take care of everything else. | |
2117 | ||
2118 | ||
2119 | ATOMIC OPERATIONS | |
2120 | ----------------- | |
2121 | ||
2122 | Whilst they are technically interprocessor interaction considerations, atomic | |
2123 | operations are noted specially as some of them imply full memory barriers and | |
2124 | some don't, but they're very heavily relied on as a group throughout the | |
2125 | kernel. | |
2126 | ||
2127 | Any atomic operation that modifies some state in memory and returns information | |
2128 | about the state (old or new) implies an SMP-conditional general memory barrier | |
2129 | (smp_mb()) on each side of the actual operation (with the exception of | |
2130 | explicit lock operations, described later). These include: | |
2131 | ||
2132 | xchg(); | |
2133 | cmpxchg(); | |
2134 | atomic_xchg(); atomic_long_xchg(); | |
2135 | atomic_cmpxchg(); atomic_long_cmpxchg(); | |
2136 | atomic_inc_return(); atomic_long_inc_return(); | |
2137 | atomic_dec_return(); atomic_long_dec_return(); | |
2138 | atomic_add_return(); atomic_long_add_return(); | |
2139 | atomic_sub_return(); atomic_long_sub_return(); | |
2140 | atomic_inc_and_test(); atomic_long_inc_and_test(); | |
2141 | atomic_dec_and_test(); atomic_long_dec_and_test(); | |
2142 | atomic_sub_and_test(); atomic_long_sub_and_test(); | |
2143 | atomic_add_negative(); atomic_long_add_negative(); | |
2144 | test_and_set_bit(); | |
2145 | test_and_clear_bit(); | |
2146 | test_and_change_bit(); | |
2147 | ||
2148 | /* when succeeds (returns 1) */ | |
2149 | atomic_add_unless(); atomic_long_add_unless(); | |
2150 | ||
2151 | These are used for such things as implementing LOCK-class and UNLOCK-class | |
2152 | operations and adjusting reference counters towards object destruction, and as | |
2153 | such the implicit memory barrier effects are necessary. | |
2154 | ||
2155 | ||
2156 | The following operations are potential problems as they do _not_ imply memory | |
2157 | barriers, but might be used for implementing such things as UNLOCK-class | |
2158 | operations: | |
2159 | ||
2160 | atomic_set(); | |
2161 | set_bit(); | |
2162 | clear_bit(); | |
2163 | change_bit(); | |
2164 | ||
2165 | With these the appropriate explicit memory barrier should be used if necessary | |
2166 | (smp_mb__before_clear_bit() for instance). | |
2167 | ||
2168 | ||
2169 | The following also do _not_ imply memory barriers, and so may require explicit | |
2170 | memory barriers under some circumstances (smp_mb__before_atomic_dec() for | |
2171 | instance): | |
2172 | ||
2173 | atomic_add(); | |
2174 | atomic_sub(); | |
2175 | atomic_inc(); | |
2176 | atomic_dec(); | |
2177 | ||
2178 | If they're used for statistics generation, then they probably don't need memory | |
2179 | barriers, unless there's a coupling between statistical data. | |
2180 | ||
2181 | If they're used for reference counting on an object to control its lifetime, | |
2182 | they probably don't need memory barriers because either the reference count | |
2183 | will be adjusted inside a locked section, or the caller will already hold | |
2184 | sufficient references to make the lock, and thus a memory barrier unnecessary. | |
2185 | ||
2186 | If they're used for constructing a lock of some description, then they probably | |
2187 | do need memory barriers as a lock primitive generally has to do things in a | |
2188 | specific order. | |
2189 | ||
2190 | Basically, each usage case has to be carefully considered as to whether memory | |
2191 | barriers are needed or not. | |
2192 | ||
2193 | The following operations are special locking primitives: | |
2194 | ||
2195 | test_and_set_bit_lock(); | |
2196 | clear_bit_unlock(); | |
2197 | __clear_bit_unlock(); | |
2198 | ||
2199 | These implement LOCK-class and UNLOCK-class operations. These should be used in | |
2200 | preference to other operations when implementing locking primitives, because | |
2201 | their implementations can be optimised on many architectures. | |
2202 | ||
2203 | [!] Note that special memory barrier primitives are available for these | |
2204 | situations because on some CPUs the atomic instructions used imply full memory | |
2205 | barriers, and so barrier instructions are superfluous in conjunction with them, | |
2206 | and in such cases the special barrier primitives will be no-ops. | |
2207 | ||
2208 | See Documentation/atomic_ops.txt for more information. | |
2209 | ||
2210 | ||
2211 | ACCESSING DEVICES | |
2212 | ----------------- | |
2213 | ||
2214 | Many devices can be memory mapped, and so appear to the CPU as if they're just | |
2215 | a set of memory locations. To control such a device, the driver usually has to | |
2216 | make the right memory accesses in exactly the right order. | |
2217 | ||
2218 | However, having a clever CPU or a clever compiler creates a potential problem | |
2219 | in that the carefully sequenced accesses in the driver code won't reach the | |
2220 | device in the requisite order if the CPU or the compiler thinks it is more | |
2221 | efficient to reorder, combine or merge accesses - something that would cause | |
2222 | the device to malfunction. | |
2223 | ||
2224 | Inside of the Linux kernel, I/O should be done through the appropriate accessor | |
2225 | routines - such as inb() or writel() - which know how to make such accesses | |
2226 | appropriately sequential. Whilst this, for the most part, renders the explicit | |
2227 | use of memory barriers unnecessary, there are a couple of situations where they | |
2228 | might be needed: | |
2229 | ||
2230 | (1) On some systems, I/O stores are not strongly ordered across all CPUs, and | |
2231 | so for _all_ general drivers locks should be used and mmiowb() must be | |
2232 | issued prior to unlocking the critical section. | |
2233 | ||
2234 | (2) If the accessor functions are used to refer to an I/O memory window with | |
2235 | relaxed memory access properties, then _mandatory_ memory barriers are | |
2236 | required to enforce ordering. | |
2237 | ||
2238 | See Documentation/DocBook/deviceiobook.tmpl for more information. | |
2239 | ||
2240 | ||
2241 | INTERRUPTS | |
2242 | ---------- | |
2243 | ||
2244 | A driver may be interrupted by its own interrupt service routine, and thus the | |
2245 | two parts of the driver may interfere with each other's attempts to control or | |
2246 | access the device. | |
2247 | ||
2248 | This may be alleviated - at least in part - by disabling local interrupts (a | |
2249 | form of locking), such that the critical operations are all contained within | |
2250 | the interrupt-disabled section in the driver. Whilst the driver's interrupt | |
2251 | routine is executing, the driver's core may not run on the same CPU, and its | |
2252 | interrupt is not permitted to happen again until the current interrupt has been | |
2253 | handled, thus the interrupt handler does not need to lock against that. | |
2254 | ||
2255 | However, consider a driver that was talking to an ethernet card that sports an | |
2256 | address register and a data register. If that driver's core talks to the card | |
2257 | under interrupt-disablement and then the driver's interrupt handler is invoked: | |
2258 | ||
2259 | LOCAL IRQ DISABLE | |
2260 | writew(ADDR, 3); | |
2261 | writew(DATA, y); | |
2262 | LOCAL IRQ ENABLE | |
2263 | <interrupt> | |
2264 | writew(ADDR, 4); | |
2265 | q = readw(DATA); | |
2266 | </interrupt> | |
2267 | ||
2268 | The store to the data register might happen after the second store to the | |
2269 | address register if ordering rules are sufficiently relaxed: | |
2270 | ||
2271 | STORE *ADDR = 3, STORE *ADDR = 4, STORE *DATA = y, q = LOAD *DATA | |
2272 | ||
2273 | ||
2274 | If ordering rules are relaxed, it must be assumed that accesses done inside an | |
2275 | interrupt disabled section may leak outside of it and may interleave with | |
2276 | accesses performed in an interrupt - and vice versa - unless implicit or | |
2277 | explicit barriers are used. | |
2278 | ||
2279 | Normally this won't be a problem because the I/O accesses done inside such | |
2280 | sections will include synchronous load operations on strictly ordered I/O | |
2281 | registers that form implicit I/O barriers. If this isn't sufficient then an | |
2282 | mmiowb() may need to be used explicitly. | |
2283 | ||
2284 | ||
2285 | A similar situation may occur between an interrupt routine and two routines | |
2286 | running on separate CPUs that communicate with each other. If such a case is | |
2287 | likely, then interrupt-disabling locks should be used to guarantee ordering. | |
2288 | ||
2289 | ||
2290 | ========================== | |
2291 | KERNEL I/O BARRIER EFFECTS | |
2292 | ========================== | |
2293 | ||
2294 | When accessing I/O memory, drivers should use the appropriate accessor | |
2295 | functions: | |
2296 | ||
2297 | (*) inX(), outX(): | |
2298 | ||
2299 | These are intended to talk to I/O space rather than memory space, but | |
2300 | that's primarily a CPU-specific concept. The i386 and x86_64 processors do | |
2301 | indeed have special I/O space access cycles and instructions, but many | |
2302 | CPUs don't have such a concept. | |
2303 | ||
2304 | The PCI bus, amongst others, defines an I/O space concept which - on such | |
2305 | CPUs as i386 and x86_64 - readily maps to the CPU's concept of I/O | |
2306 | space. However, it may also be mapped as a virtual I/O space in the CPU's | |
2307 | memory map, particularly on those CPUs that don't support alternate I/O | |
2308 | spaces. | |
2309 | ||
2310 | Accesses to this space may be fully synchronous (as on i386), but | |
2311 | intermediary bridges (such as the PCI host bridge) may not fully honour | |
2312 | that. | |
2313 | ||
2314 | They are guaranteed to be fully ordered with respect to each other. | |
2315 | ||
2316 | They are not guaranteed to be fully ordered with respect to other types of | |
2317 | memory and I/O operation. | |
2318 | ||
2319 | (*) readX(), writeX(): | |
2320 | ||
2321 | Whether these are guaranteed to be fully ordered and uncombined with | |
2322 | respect to each other on the issuing CPU depends on the characteristics | |
2323 | defined for the memory window through which they're accessing. On later | |
2324 | i386 architecture machines, for example, this is controlled by way of the | |
2325 | MTRR registers. | |
2326 | ||
2327 | Ordinarily, these will be guaranteed to be fully ordered and uncombined, | |
2328 | provided they're not accessing a prefetchable device. | |
2329 | ||
2330 | However, intermediary hardware (such as a PCI bridge) may indulge in | |
2331 | deferral if it so wishes; to flush a store, a load from the same location | |
2332 | is preferred[*], but a load from the same device or from configuration | |
2333 | space should suffice for PCI. | |
2334 | ||
2335 | [*] NOTE! attempting to load from the same location as was written to may | |
2336 | cause a malfunction - consider the 16550 Rx/Tx serial registers for | |
2337 | example. | |
2338 | ||
2339 | Used with prefetchable I/O memory, an mmiowb() barrier may be required to | |
2340 | force stores to be ordered. | |
2341 | ||
2342 | Please refer to the PCI specification for more information on interactions | |
2343 | between PCI transactions. | |
2344 | ||
2345 | (*) readX_relaxed() | |
2346 | ||
2347 | These are similar to readX(), but are not guaranteed to be ordered in any | |
2348 | way. Be aware that there is no I/O read barrier available. | |
2349 | ||
2350 | (*) ioreadX(), iowriteX() | |
2351 | ||
2352 | These will perform appropriately for the type of access they're actually | |
2353 | doing, be it inX()/outX() or readX()/writeX(). | |
2354 | ||
2355 | ||
2356 | ======================================== | |
2357 | ASSUMED MINIMUM EXECUTION ORDERING MODEL | |
2358 | ======================================== | |
2359 | ||
2360 | It has to be assumed that the conceptual CPU is weakly-ordered but that it will | |
2361 | maintain the appearance of program causality with respect to itself. Some CPUs | |
2362 | (such as i386 or x86_64) are more constrained than others (such as powerpc or | |
2363 | frv), and so the most relaxed case (namely DEC Alpha) must be assumed outside | |
2364 | of arch-specific code. | |
2365 | ||
2366 | This means that it must be considered that the CPU will execute its instruction | |
2367 | stream in any order it feels like - or even in parallel - provided that if an | |
2368 | instruction in the stream depends on an earlier instruction, then that | |
2369 | earlier instruction must be sufficiently complete[*] before the later | |
2370 | instruction may proceed; in other words: provided that the appearance of | |
2371 | causality is maintained. | |
2372 | ||
2373 | [*] Some instructions have more than one effect - such as changing the | |
2374 | condition codes, changing registers or changing memory - and different | |
2375 | instructions may depend on different effects. | |
2376 | ||
2377 | A CPU may also discard any instruction sequence that winds up having no | |
2378 | ultimate effect. For example, if two adjacent instructions both load an | |
2379 | immediate value into the same register, the first may be discarded. | |
2380 | ||
2381 | ||
2382 | Similarly, it has to be assumed that compiler might reorder the instruction | |
2383 | stream in any way it sees fit, again provided the appearance of causality is | |
2384 | maintained. | |
2385 | ||
2386 | ||
2387 | ============================ | |
2388 | THE EFFECTS OF THE CPU CACHE | |
2389 | ============================ | |
2390 | ||
2391 | The way cached memory operations are perceived across the system is affected to | |
2392 | a certain extent by the caches that lie between CPUs and memory, and by the | |
2393 | memory coherence system that maintains the consistency of state in the system. | |
2394 | ||
2395 | As far as the way a CPU interacts with another part of the system through the | |
2396 | caches goes, the memory system has to include the CPU's caches, and memory | |
2397 | barriers for the most part act at the interface between the CPU and its cache | |
2398 | (memory barriers logically act on the dotted line in the following diagram): | |
2399 | ||
2400 | <--- CPU ---> : <----------- Memory -----------> | |
2401 | : | |
2402 | +--------+ +--------+ : +--------+ +-----------+ | |
2403 | | | | | : | | | | +--------+ | |
2404 | | CPU | | Memory | : | CPU | | | | | | |
2405 | | Core |--->| Access |----->| Cache |<-->| | | | | |
2406 | | | | Queue | : | | | |--->| Memory | | |
2407 | | | | | : | | | | | | | |
2408 | +--------+ +--------+ : +--------+ | | | | | |
2409 | : | Cache | +--------+ | |
2410 | : | Coherency | | |
2411 | : | Mechanism | +--------+ | |
2412 | +--------+ +--------+ : +--------+ | | | | | |
2413 | | | | | : | | | | | | | |
2414 | | CPU | | Memory | : | CPU | | |--->| Device | | |
2415 | | Core |--->| Access |----->| Cache |<-->| | | | | |
2416 | | | | Queue | : | | | | | | | |
2417 | | | | | : | | | | +--------+ | |
2418 | +--------+ +--------+ : +--------+ +-----------+ | |
2419 | : | |
2420 | : | |
2421 | ||
2422 | Although any particular load or store may not actually appear outside of the | |
2423 | CPU that issued it since it may have been satisfied within the CPU's own cache, | |
2424 | it will still appear as if the full memory access had taken place as far as the | |
2425 | other CPUs are concerned since the cache coherency mechanisms will migrate the | |
2426 | cacheline over to the accessing CPU and propagate the effects upon conflict. | |
2427 | ||
2428 | The CPU core may execute instructions in any order it deems fit, provided the | |
2429 | expected program causality appears to be maintained. Some of the instructions | |
2430 | generate load and store operations which then go into the queue of memory | |
2431 | accesses to be performed. The core may place these in the queue in any order | |
2432 | it wishes, and continue execution until it is forced to wait for an instruction | |
2433 | to complete. | |
2434 | ||
2435 | What memory barriers are concerned with is controlling the order in which | |
2436 | accesses cross from the CPU side of things to the memory side of things, and | |
2437 | the order in which the effects are perceived to happen by the other observers | |
2438 | in the system. | |
2439 | ||
2440 | [!] Memory barriers are _not_ needed within a given CPU, as CPUs always see | |
2441 | their own loads and stores as if they had happened in program order. | |
2442 | ||
2443 | [!] MMIO or other device accesses may bypass the cache system. This depends on | |
2444 | the properties of the memory window through which devices are accessed and/or | |
2445 | the use of any special device communication instructions the CPU may have. | |
2446 | ||
2447 | ||
2448 | CACHE COHERENCY | |
2449 | --------------- | |
2450 | ||
2451 | Life isn't quite as simple as it may appear above, however: for while the | |
2452 | caches are expected to be coherent, there's no guarantee that that coherency | |
2453 | will be ordered. This means that whilst changes made on one CPU will | |
2454 | eventually become visible on all CPUs, there's no guarantee that they will | |
2455 | become apparent in the same order on those other CPUs. | |
2456 | ||
2457 | ||
2458 | Consider dealing with a system that has a pair of CPUs (1 & 2), each of which | |
2459 | has a pair of parallel data caches (CPU 1 has A/B, and CPU 2 has C/D): | |
2460 | ||
2461 | : | |
2462 | : +--------+ | |
2463 | : +---------+ | | | |
2464 | +--------+ : +--->| Cache A |<------->| | | |
2465 | | | : | +---------+ | | | |
2466 | | CPU 1 |<---+ | | | |
2467 | | | : | +---------+ | | | |
2468 | +--------+ : +--->| Cache B |<------->| | | |
2469 | : +---------+ | | | |
2470 | : | Memory | | |
2471 | : +---------+ | System | | |
2472 | +--------+ : +--->| Cache C |<------->| | | |
2473 | | | : | +---------+ | | | |
2474 | | CPU 2 |<---+ | | | |
2475 | | | : | +---------+ | | | |
2476 | +--------+ : +--->| Cache D |<------->| | | |
2477 | : +---------+ | | | |
2478 | : +--------+ | |
2479 | : | |
2480 | ||
2481 | Imagine the system has the following properties: | |
2482 | ||
2483 | (*) an odd-numbered cache line may be in cache A, cache C or it may still be | |
2484 | resident in memory; | |
2485 | ||
2486 | (*) an even-numbered cache line may be in cache B, cache D or it may still be | |
2487 | resident in memory; | |
2488 | ||
2489 | (*) whilst the CPU core is interrogating one cache, the other cache may be | |
2490 | making use of the bus to access the rest of the system - perhaps to | |
2491 | displace a dirty cacheline or to do a speculative load; | |
2492 | ||
2493 | (*) each cache has a queue of operations that need to be applied to that cache | |
2494 | to maintain coherency with the rest of the system; | |
2495 | ||
2496 | (*) the coherency queue is not flushed by normal loads to lines already | |
2497 | present in the cache, even though the contents of the queue may | |
2498 | potentially affect those loads. | |
2499 | ||
2500 | Imagine, then, that two writes are made on the first CPU, with a write barrier | |
2501 | between them to guarantee that they will appear to reach that CPU's caches in | |
2502 | the requisite order: | |
2503 | ||
2504 | CPU 1 CPU 2 COMMENT | |
2505 | =============== =============== ======================================= | |
2506 | u == 0, v == 1 and p == &u, q == &u | |
2507 | v = 2; | |
2508 | smp_wmb(); Make sure change to v is visible before | |
2509 | change to p | |
2510 | <A:modify v=2> v is now in cache A exclusively | |
2511 | p = &v; | |
2512 | <B:modify p=&v> p is now in cache B exclusively | |
2513 | ||
2514 | The write memory barrier forces the other CPUs in the system to perceive that | |
2515 | the local CPU's caches have apparently been updated in the correct order. But | |
2516 | now imagine that the second CPU wants to read those values: | |
2517 | ||
2518 | CPU 1 CPU 2 COMMENT | |
2519 | =============== =============== ======================================= | |
2520 | ... | |
2521 | q = p; | |
2522 | x = *q; | |
2523 | ||
2524 | The above pair of reads may then fail to happen in the expected order, as the | |
2525 | cacheline holding p may get updated in one of the second CPU's caches whilst | |
2526 | the update to the cacheline holding v is delayed in the other of the second | |
2527 | CPU's caches by some other cache event: | |
2528 | ||
2529 | CPU 1 CPU 2 COMMENT | |
2530 | =============== =============== ======================================= | |
2531 | u == 0, v == 1 and p == &u, q == &u | |
2532 | v = 2; | |
2533 | smp_wmb(); | |
2534 | <A:modify v=2> <C:busy> | |
2535 | <C:queue v=2> | |
2536 | p = &v; q = p; | |
2537 | <D:request p> | |
2538 | <B:modify p=&v> <D:commit p=&v> | |
2539 | <D:read p> | |
2540 | x = *q; | |
2541 | <C:read *q> Reads from v before v updated in cache | |
2542 | <C:unbusy> | |
2543 | <C:commit v=2> | |
2544 | ||
2545 | Basically, whilst both cachelines will be updated on CPU 2 eventually, there's | |
2546 | no guarantee that, without intervention, the order of update will be the same | |
2547 | as that committed on CPU 1. | |
2548 | ||
2549 | ||
2550 | To intervene, we need to interpolate a data dependency barrier or a read | |
2551 | barrier between the loads. This will force the cache to commit its coherency | |
2552 | queue before processing any further requests: | |
2553 | ||
2554 | CPU 1 CPU 2 COMMENT | |
2555 | =============== =============== ======================================= | |
2556 | u == 0, v == 1 and p == &u, q == &u | |
2557 | v = 2; | |
2558 | smp_wmb(); | |
2559 | <A:modify v=2> <C:busy> | |
2560 | <C:queue v=2> | |
2561 | p = &v; q = p; | |
2562 | <D:request p> | |
2563 | <B:modify p=&v> <D:commit p=&v> | |
2564 | <D:read p> | |
2565 | smp_read_barrier_depends() | |
2566 | <C:unbusy> | |
2567 | <C:commit v=2> | |
2568 | x = *q; | |
2569 | <C:read *q> Reads from v after v updated in cache | |
2570 | ||
2571 | ||
2572 | This sort of problem can be encountered on DEC Alpha processors as they have a | |
2573 | split cache that improves performance by making better use of the data bus. | |
2574 | Whilst most CPUs do imply a data dependency barrier on the read when a memory | |
2575 | access depends on a read, not all do, so it may not be relied on. | |
2576 | ||
2577 | Other CPUs may also have split caches, but must coordinate between the various | |
2578 | cachelets for normal memory accesses. The semantics of the Alpha removes the | |
2579 | need for coordination in the absence of memory barriers. | |
2580 | ||
2581 | ||
2582 | CACHE COHERENCY VS DMA | |
2583 | ---------------------- | |
2584 | ||
2585 | Not all systems maintain cache coherency with respect to devices doing DMA. In | |
2586 | such cases, a device attempting DMA may obtain stale data from RAM because | |
2587 | dirty cache lines may be resident in the caches of various CPUs, and may not | |
2588 | have been written back to RAM yet. To deal with this, the appropriate part of | |
2589 | the kernel must flush the overlapping bits of cache on each CPU (and maybe | |
2590 | invalidate them as well). | |
2591 | ||
2592 | In addition, the data DMA'd to RAM by a device may be overwritten by dirty | |
2593 | cache lines being written back to RAM from a CPU's cache after the device has | |
2594 | installed its own data, or cache lines present in the CPU's cache may simply | |
2595 | obscure the fact that RAM has been updated, until at such time as the cacheline | |
2596 | is discarded from the CPU's cache and reloaded. To deal with this, the | |
2597 | appropriate part of the kernel must invalidate the overlapping bits of the | |
2598 | cache on each CPU. | |
2599 | ||
2600 | See Documentation/cachetlb.txt for more information on cache management. | |
2601 | ||
2602 | ||
2603 | CACHE COHERENCY VS MMIO | |
2604 | ----------------------- | |
2605 | ||
2606 | Memory mapped I/O usually takes place through memory locations that are part of | |
2607 | a window in the CPU's memory space that has different properties assigned than | |
2608 | the usual RAM directed window. | |
2609 | ||
2610 | Amongst these properties is usually the fact that such accesses bypass the | |
2611 | caching entirely and go directly to the device buses. This means MMIO accesses | |
2612 | may, in effect, overtake accesses to cached memory that were emitted earlier. | |
2613 | A memory barrier isn't sufficient in such a case, but rather the cache must be | |
2614 | flushed between the cached memory write and the MMIO access if the two are in | |
2615 | any way dependent. | |
2616 | ||
2617 | ||
2618 | ========================= | |
2619 | THE THINGS CPUS GET UP TO | |
2620 | ========================= | |
2621 | ||
2622 | A programmer might take it for granted that the CPU will perform memory | |
2623 | operations in exactly the order specified, so that if the CPU is, for example, | |
2624 | given the following piece of code to execute: | |
2625 | ||
2626 | a = ACCESS_ONCE(*A); | |
2627 | ACCESS_ONCE(*B) = b; | |
2628 | c = ACCESS_ONCE(*C); | |
2629 | d = ACCESS_ONCE(*D); | |
2630 | ACCESS_ONCE(*E) = e; | |
2631 | ||
2632 | they would then expect that the CPU will complete the memory operation for each | |
2633 | instruction before moving on to the next one, leading to a definite sequence of | |
2634 | operations as seen by external observers in the system: | |
2635 | ||
2636 | LOAD *A, STORE *B, LOAD *C, LOAD *D, STORE *E. | |
2637 | ||
2638 | ||
2639 | Reality is, of course, much messier. With many CPUs and compilers, the above | |
2640 | assumption doesn't hold because: | |
2641 | ||
2642 | (*) loads are more likely to need to be completed immediately to permit | |
2643 | execution progress, whereas stores can often be deferred without a | |
2644 | problem; | |
2645 | ||
2646 | (*) loads may be done speculatively, and the result discarded should it prove | |
2647 | to have been unnecessary; | |
2648 | ||
2649 | (*) loads may be done speculatively, leading to the result having been fetched | |
2650 | at the wrong time in the expected sequence of events; | |
2651 | ||
2652 | (*) the order of the memory accesses may be rearranged to promote better use | |
2653 | of the CPU buses and caches; | |
2654 | ||
2655 | (*) loads and stores may be combined to improve performance when talking to | |
2656 | memory or I/O hardware that can do batched accesses of adjacent locations, | |
2657 | thus cutting down on transaction setup costs (memory and PCI devices may | |
2658 | both be able to do this); and | |
2659 | ||
2660 | (*) the CPU's data cache may affect the ordering, and whilst cache-coherency | |
2661 | mechanisms may alleviate this - once the store has actually hit the cache | |
2662 | - there's no guarantee that the coherency management will be propagated in | |
2663 | order to other CPUs. | |
2664 | ||
2665 | So what another CPU, say, might actually observe from the above piece of code | |
2666 | is: | |
2667 | ||
2668 | LOAD *A, ..., LOAD {*C,*D}, STORE *E, STORE *B | |
2669 | ||
2670 | (Where "LOAD {*C,*D}" is a combined load) | |
2671 | ||
2672 | ||
2673 | However, it is guaranteed that a CPU will be self-consistent: it will see its | |
2674 | _own_ accesses appear to be correctly ordered, without the need for a memory | |
2675 | barrier. For instance with the following code: | |
2676 | ||
2677 | U = ACCESS_ONCE(*A); | |
2678 | ACCESS_ONCE(*A) = V; | |
2679 | ACCESS_ONCE(*A) = W; | |
2680 | X = ACCESS_ONCE(*A); | |
2681 | ACCESS_ONCE(*A) = Y; | |
2682 | Z = ACCESS_ONCE(*A); | |
2683 | ||
2684 | and assuming no intervention by an external influence, it can be assumed that | |
2685 | the final result will appear to be: | |
2686 | ||
2687 | U == the original value of *A | |
2688 | X == W | |
2689 | Z == Y | |
2690 | *A == Y | |
2691 | ||
2692 | The code above may cause the CPU to generate the full sequence of memory | |
2693 | accesses: | |
2694 | ||
2695 | U=LOAD *A, STORE *A=V, STORE *A=W, X=LOAD *A, STORE *A=Y, Z=LOAD *A | |
2696 | ||
2697 | in that order, but, without intervention, the sequence may have almost any | |
2698 | combination of elements combined or discarded, provided the program's view of | |
2699 | the world remains consistent. Note that ACCESS_ONCE() is -not- optional | |
2700 | in the above example, as there are architectures where a given CPU might | |
2701 | interchange successive loads to the same location. On such architectures, | |
2702 | ACCESS_ONCE() does whatever is necessary to prevent this, for example, on | |
2703 | Itanium the volatile casts used by ACCESS_ONCE() cause GCC to emit the | |
2704 | special ld.acq and st.rel instructions that prevent such reordering. | |
2705 | ||
2706 | The compiler may also combine, discard or defer elements of the sequence before | |
2707 | the CPU even sees them. | |
2708 | ||
2709 | For instance: | |
2710 | ||
2711 | *A = V; | |
2712 | *A = W; | |
2713 | ||
2714 | may be reduced to: | |
2715 | ||
2716 | *A = W; | |
2717 | ||
2718 | since, without either a write barrier or an ACCESS_ONCE(), it can be | |
2719 | assumed that the effect of the storage of V to *A is lost. Similarly: | |
2720 | ||
2721 | *A = Y; | |
2722 | Z = *A; | |
2723 | ||
2724 | may, without a memory barrier or an ACCESS_ONCE(), be reduced to: | |
2725 | ||
2726 | *A = Y; | |
2727 | Z = Y; | |
2728 | ||
2729 | and the LOAD operation never appear outside of the CPU. | |
2730 | ||
2731 | ||
2732 | AND THEN THERE'S THE ALPHA | |
2733 | -------------------------- | |
2734 | ||
2735 | The DEC Alpha CPU is one of the most relaxed CPUs there is. Not only that, | |
2736 | some versions of the Alpha CPU have a split data cache, permitting them to have | |
2737 | two semantically-related cache lines updated at separate times. This is where | |
2738 | the data dependency barrier really becomes necessary as this synchronises both | |
2739 | caches with the memory coherence system, thus making it seem like pointer | |
2740 | changes vs new data occur in the right order. | |
2741 | ||
2742 | The Alpha defines the Linux kernel's memory barrier model. | |
2743 | ||
2744 | See the subsection on "Cache Coherency" above. | |
2745 | ||
2746 | ||
2747 | ============ | |
2748 | EXAMPLE USES | |
2749 | ============ | |
2750 | ||
2751 | CIRCULAR BUFFERS | |
2752 | ---------------- | |
2753 | ||
2754 | Memory barriers can be used to implement circular buffering without the need | |
2755 | of a lock to serialise the producer with the consumer. See: | |
2756 | ||
2757 | Documentation/circular-buffers.txt | |
2758 | ||
2759 | for details. | |
2760 | ||
2761 | ||
2762 | ========== | |
2763 | REFERENCES | |
2764 | ========== | |
2765 | ||
2766 | Alpha AXP Architecture Reference Manual, Second Edition (Sites & Witek, | |
2767 | Digital Press) | |
2768 | Chapter 5.2: Physical Address Space Characteristics | |
2769 | Chapter 5.4: Caches and Write Buffers | |
2770 | Chapter 5.5: Data Sharing | |
2771 | Chapter 5.6: Read/Write Ordering | |
2772 | ||
2773 | AMD64 Architecture Programmer's Manual Volume 2: System Programming | |
2774 | Chapter 7.1: Memory-Access Ordering | |
2775 | Chapter 7.4: Buffering and Combining Memory Writes | |
2776 | ||
2777 | IA-32 Intel Architecture Software Developer's Manual, Volume 3: | |
2778 | System Programming Guide | |
2779 | Chapter 7.1: Locked Atomic Operations | |
2780 | Chapter 7.2: Memory Ordering | |
2781 | Chapter 7.4: Serializing Instructions | |
2782 | ||
2783 | The SPARC Architecture Manual, Version 9 | |
2784 | Chapter 8: Memory Models | |
2785 | Appendix D: Formal Specification of the Memory Models | |
2786 | Appendix J: Programming with the Memory Models | |
2787 | ||
2788 | UltraSPARC Programmer Reference Manual | |
2789 | Chapter 5: Memory Accesses and Cacheability | |
2790 | Chapter 15: Sparc-V9 Memory Models | |
2791 | ||
2792 | UltraSPARC III Cu User's Manual | |
2793 | Chapter 9: Memory Models | |
2794 | ||
2795 | UltraSPARC IIIi Processor User's Manual | |
2796 | Chapter 8: Memory Models | |
2797 | ||
2798 | UltraSPARC Architecture 2005 | |
2799 | Chapter 9: Memory | |
2800 | Appendix D: Formal Specifications of the Memory Models | |
2801 | ||
2802 | UltraSPARC T1 Supplement to the UltraSPARC Architecture 2005 | |
2803 | Chapter 8: Memory Models | |
2804 | Appendix F: Caches and Cache Coherency | |
2805 | ||
2806 | Solaris Internals, Core Kernel Architecture, p63-68: | |
2807 | Chapter 3.3: Hardware Considerations for Locks and | |
2808 | Synchronization | |
2809 | ||
2810 | Unix Systems for Modern Architectures, Symmetric Multiprocessing and Caching | |
2811 | for Kernel Programmers: | |
2812 | Chapter 13: Other Memory Models | |
2813 | ||
2814 | Intel Itanium Architecture Software Developer's Manual: Volume 1: | |
2815 | Section 2.6: Speculation | |
2816 | Section 4.4: Memory Access |