Gitiles
Code Review Sign In
gerrit-public.fairphone.software / fp2-dev / platform / external / llvm / 53cae1362dca8aa312c3e36c10b106ea7d349f93 / . / docs / Atomics.html
blob: 967ebdddb1a3e7e37d540d1c5880495103c04cb6 [file] [log] [blame]
Eli Friedman138515d2011-08-09 21:07:10 +0000[diff] [blame]1<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
3<html>
4<head>
5 <title>LLVM Atomic Instructions and Concurrency Guide</title>
Eli Friedman21006d42011-08-09 23:02:53 +0000[diff] [blame]6 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
Eli Friedman138515d2011-08-09 21:07:10 +0000[diff] [blame]7 <link rel="stylesheet" href="llvm.css" type="text/css">
8</head>
9<body>
10
11<h1>
12 LLVM Atomic Instructions and Concurrency Guide
13</h1>
14
15<ol>
16 <li><a href="#introduction">Introduction</a></li>
17 <li><a href="#loadstore">Load and store</a></li>
Eli Friedman138515d2011-08-09 21:07:10 +0000[diff] [blame]18 <li><a href="#otherinst">Other atomic instructions</a></li>
Eli Friedman1bf4ad42011-08-11 23:44:25 +0000[diff] [blame]19 <li><a href="#ordering">Atomic orderings</a></li>
Eli Friedman138515d2011-08-09 21:07:10 +0000[diff] [blame]20 <li><a href="#iropt">Atomics and IR optimization</a></li>
21 <li><a href="#codegen">Atomics and Codegen</a></li>
22</ol>
23
24<div class="doc_author">
25 <p>Written by Eli Friedman</p>
26</div>
27
28<!-- *********************************************************************** -->
29<h2>
30 <a name="introduction">Introduction</a>
31</h2>
32<!-- *********************************************************************** -->
33
34<div>
35
36<p>Historically, LLVM has not had very strong support for concurrency; some
37minimal intrinsics were provided, and <code>volatile</code> was used in some
38cases to achieve rough semantics in the presence of concurrency. However, this
39is changing; there are now new instructions which are well-defined in the
40presence of threads and asynchronous signals, and the model for existing
41instructions has been clarified in the IR.</p>
42
43<p>The atomic instructions are designed specifically to provide readable IR and
44 optimized code generation for the following:</p>
45<ul>
Eli Friedman1bf4ad42011-08-11 23:44:25 +0000[diff] [blame]46 <li>The new C++0x <code>&lt;atomic&gt;</code> header.
47 (<a href="http://www.open-std.org/jtc1/sc22/wg21/">C++0x draft available here</a>.)
48 (<a href="http://www.open-std.org/jtc1/sc22/wg14/">C1x draft available here</a>)</li>
Eli Friedman138515d2011-08-09 21:07:10 +0000[diff] [blame]49 <li>Proper semantics for Java-style memory, for both <code>volatile</code> and
Eli Friedman1bf4ad42011-08-11 23:44:25 +0000[diff] [blame]50 regular shared variables.
51 (<a href="http://java.sun.com/docs/books/jls/third_edition/html/memory.html">Java Specification</a>)</li>
52 <li>gcc-compatible <code>__sync_*</code> builtins.
53 (<a href="http://gcc.gnu.org/onlinedocs/gcc/Atomic-Builtins.html">Description</a>)</li>
Eli Friedman138515d2011-08-09 21:07:10 +0000[diff] [blame]54 <li>Other scenarios with atomic semantics, including <code>static</code>
55 variables with non-trivial constructors in C++.</li>
56</ul>
57
Eli Friedman1bf4ad42011-08-11 23:44:25 +0000[diff] [blame]58<p>Atomic and volatile in the IR are orthogonal; "volatile" is the C/C++
59 volatile, which ensures that every volatile load and store happens and is
60 performed in the stated order. A couple examples: if a
61 SequentiallyConsistent store is immediately followed by another
62 SequentiallyConsistent store to the same address, the first store can
63 be erased. This transformation is not allowed for a pair of volatile
64 stores. On the other hand, a non-volatile non-atomic load can be moved
65 across a volatile load freely, but not an Acquire load.</p>
66
Eli Friedman138515d2011-08-09 21:07:10 +0000[diff] [blame]67<p>This document is intended to provide a guide to anyone either writing a
68 frontend for LLVM or working on optimization passes for LLVM with a guide
69 for how to deal with instructions with special semantics in the presence of
70 concurrency. This is not intended to be a precise guide to the semantics;
71 the details can get extremely complicated and unreadable, and are not
72 usually necessary.</p>
73
74</div>
75
76<!-- *********************************************************************** -->
77<h2>
78 <a name="loadstore">Load and store</a>
79</h2>
80<!-- *********************************************************************** -->
81
82<div>
83
84<p>The basic <code>'load'</code> and <code>'store'</code> allow a variety of
85 optimizations, but can have unintuitive results in a concurrent environment.
86 For a frontend writer, the rule is essentially that all memory accessed
87 with basic loads and stores by multiple threads should be protected by a
88 lock or other synchronization; otherwise, you are likely to run into
89 undefined behavior. (Do not use volatile as a substitute for atomics; it
90 might work on some platforms, but does not provide the necessary guarantees
91 in general.)</p>
92
93<p>From the optimizer's point of view, the rule is that if there
Eli Friedman1bf4ad42011-08-11 23:44:25 +0000[diff] [blame]94 are not any instructions with atomic ordering involved, concurrency does
95 not matter, with one exception: if a variable might be visible to another
Eli Friedman138515d2011-08-09 21:07:10 +0000[diff] [blame]96 thread or signal handler, a store cannot be inserted along a path where it
Eli Friedman1bf4ad42011-08-11 23:44:25 +0000[diff] [blame]97 might not execute otherwise. For example, suppose LICM wants to take all the
98 loads and stores in a loop to and from a particular address and promote them
99 to registers. LICM is not allowed to insert an unconditional store after
100 the loop with the computed value unless a store unconditionally executes
101 within the loop. Note that speculative loads are allowed; a load which
102 is part of a race returns <code>undef</code>, but does not have undefined
103 behavior.</p>
Eli Friedman138515d2011-08-09 21:07:10 +0000[diff] [blame]104
105<p>For cases where simple loads and stores are not sufficient, LLVM provides
106 atomic loads and stores with varying levels of guarantees.</p>
107
108</div>
109
110<!-- *********************************************************************** -->
111<h2>
Eli Friedman138515d2011-08-09 21:07:10 +0000[diff] [blame]112 <a name="otherinst">Other atomic instructions</a>
113</h2>
114<!-- *********************************************************************** -->
115
116<div>
117
118<p><code>cmpxchg</code> and <code>atomicrmw</code> are essentially like an
119 atomic load followed by an atomic store (where the store is conditional for
Eli Friedmane2d8cf72011-08-10 20:17:43 +0000[diff] [blame]120 <code>cmpxchg</code>), but no other memory operation can happen between
121 the load and store. Note that our cmpxchg does not have quite as many
122 options for making cmpxchg weaker as the C++0x version.</p>
Eli Friedman138515d2011-08-09 21:07:10 +0000[diff] [blame]123
124<p>A <code>fence</code> provides Acquire and/or Release ordering which is not
125 part of another operation; it is normally used along with Monotonic memory
126 operations. A Monotonic load followed by an Acquire fence is roughly
127 equivalent to an Acquire load.</p>
128
129<p>Frontends generating atomic instructions generally need to be aware of the
130 target to some degree; atomic instructions are guaranteed to be lock-free,
131 and therefore an instruction which is wider than the target natively supports
132 can be impossible to generate.</p>
133
134</div>
135
136<!-- *********************************************************************** -->
137<h2>
Eli Friedman1bf4ad42011-08-11 23:44:25 +0000[diff] [blame]138 <a name="ordering">Atomic orderings</a>
139</h2>
140<!-- *********************************************************************** -->
141
142<div>
143
144<p>In order to achieve a balance between performance and necessary guarantees,
145 there are six levels of atomicity. They are listed in order of strength;
146 each level includes all the guarantees of the previous level except for
147 Acquire/Release.</p>
148
149<!-- ======================================================================= -->
150<h3>
151 <a name="o_unordered">Unordered</a>
152</h3>
153
154<div>
155
156<p>Unordered is the lowest level of atomicity. It essentially guarantees that
157 races produce somewhat sane results instead of having undefined behavior.
158 It also guarantees the operation to be lock-free, so it do not depend on
159 the data being part of a special atomic structure or depend on a separate
160 per-process global lock. Note that code generation will fail for
161 unsupported atomic operations; if you need such an operation, use explicit
162 locking.</p>
163
164<dl>
165 <dt>Relevant standard</dt>
166 <dd>This is intended to match the Java memory model for shared
167 variables.</dd>
168 <dt>Notes for frontends</dt>
169 <dd>This cannot be used for synchronization, but is useful for Java and
170 other "safe" languages which need to guarantee that the generated
171 code never exhibits undefined behavior. Note that this guarantee
172 is cheap on common platforms for loads of a native width, but can
173 be expensive or unavailable for wider loads, like a 64-bit store
174 on ARM. (A frontend for Java or other "safe" languages would normally
175 split a 64-bit store on ARM into two 32-bit unordered stores.)
176 <dt>Notes for optimizers</dt>
177 <dd>In terms of the optimizer, this prohibits any transformation that
178 transforms a single load into multiple loads, transforms a store
179 into multiple stores, narrows a store, or stores a value which
180 would not be stored otherwise. Some examples of unsafe optimizations
181 are narrowing an assignment into a bitfield, rematerializing
182 a load, and turning loads and stores into a memcpy call. Reordering
183 unordered operations is safe, though, and optimizers should take
184 advantage of that because unordered operations are common in
185 languages that need them.</dd>
186 <dt>Notes for code generation</dt>
187 <dd>These operations are required to be atomic in the sense that if you
188 use unordered loads and unordered stores, a load cannot see a value
189 which was never stored. A normal load or store instruction is usually
190 sufficient, but note that an unordered load or store cannot
191 be split into multiple instructions (or an instruction which
192 does multiple memory operations, like <code>LDRD</code> on ARM).</dd>
193</dl>
194
195</div>
196
197<!-- ======================================================================= -->
198<h3>
199 <a name="o_monotonic">Monotonic</a>
200</h3>
201
202<div>
203
204<p>Monotonic is the weakest level of atomicity that can be used in
205 synchronization primitives, although it does not provide any general
206 synchronization. It essentially guarantees that if you take all the
207 operations affecting a specific address, a consistent ordering exists.
208
209<dl>
210 <dt>Relevant standard</dt>
211 <dd>This corresponds to the C++0x/C1x <code>memory_order_relaxed</code>;
212 see those standards for the exact definition.
213 <dt>Notes for frontends</dt>
214 <dd>If you are writing a frontend which uses this directly, use with caution.
215 The guarantees in terms of synchronization are very weak, so make
216 sure these are only used in a pattern which you know is correct.
217 Generally, these would either be used for atomic operations which
218 do not protect other memory (like an atomic counter), or along with
219 a <code>fence</code>.</dd>
220 <dt>Notes for optimizers</dt>
221 <dd>In terms of the optimizer, this can be treated as a read+write on the
222 relevant memory location (and alias analysis will take advantage of
223 that). In addition, it is legal to reorder non-atomic and Unordered
224 loads around Monotonic loads. CSE/DSE and a few other optimizations
225 are allowed, but Monotonic operations are unlikely to be used in ways
226 which would make those optimizations useful.</dd>
227 <dt>Notes for code generation</dt>
228 <dd>Code generation is essentially the same as that for unordered for loads
229 and stores. No fences is required. <code>cmpxchg</code> and
230 <code>atomicrmw</code> are required to appear as a single operation.</dd>
231</dl>
232
233</div>
234
235<!-- ======================================================================= -->
236<h3>
237 <a name="o_acquire">Acquire</a>
238</h3>
239
240<div>
241
242<p>Acquire provides a barrier of the sort necessary to acquire a lock to access
243 other memory with normal loads and stores.
244
245<dl>
246 <dt>Relevant standard</dt>
247 <dd>This corresponds to the C++0x/C1x <code>memory_order_acquire</code>. It
248 should also be used for C++0x/C1x <code>memory_order_consume</code>.
249 <dt>Notes for frontends</dt>
250 <dd>If you are writing a frontend which uses this directly, use with caution.
251 Acquire only provides a semantic guarantee when paired with a Release
252 operation.</dd>
253 <dt>Notes for optimizers</dt>
Eli Friedman79d7de72011-08-12 03:38:32 +0000[diff] [blame]254 <dd>Optimizers not aware of atomics can treat this like a nothrow call.
Chris Lattner9a5ffbf2011-08-12 19:48:19 +0000[diff] [blame]255 It is also possible to move stores from before an Acquire load
Eli Friedman79d7de72011-08-12 03:38:32 +0000[diff] [blame]256 or read-modify-write operation to after it, and move non-Acquire
257 loads from before an Acquire operation to after it.</dd>
Eli Friedman1bf4ad42011-08-11 23:44:25 +0000[diff] [blame]258 <dt>Notes for code generation</dt>
259 <dd>Architectures with weak memory ordering (essentially everything relevant
260 today except x86 and SPARC) require some sort of fence to maintain
261 the Acquire semantics. The precise fences required varies widely by
262 architecture, but for a simple implementation, most architectures provide
263 a barrier which is strong enough for everything (<code>dmb</code> on ARM,
264 <code>sync</code> on PowerPC, etc.). Putting such a fence after the
265 equivalent Monotonic operation is sufficient to maintain Acquire
266 semantics for a memory operation.</dd>
267</dl>
268
269</div>
270
271<!-- ======================================================================= -->
272<h3>
273 <a name="o_acquire">Release</a>
274</h3>
275
276<div>
277
278<p>Release is similar to Acquire, but with a barrier of the sort necessary to
279 release a lock.
280
281<dl>
282 <dt>Relevant standard</dt>
283 <dd>This corresponds to the C++0x/C1x <code>memory_order_release</code>.</dd>
284 <dt>Notes for frontends</dt>
285 <dd>If you are writing a frontend which uses this directly, use with caution.
286 Release only provides a semantic guarantee when paired with a Acquire
287 operation.</dd>
288 <dt>Notes for optimizers</dt>
Eli Friedman79d7de72011-08-12 03:38:32 +0000[diff] [blame]289 <dd>Optimizers not aware of atomics can treat this like a nothrow call.
290 It is also possible to move loads from after a Release store
291 or read-modify-write operation to before it, and move non-Release
292 stores from after an Release operation to before it.</dd>
Eli Friedman1bf4ad42011-08-11 23:44:25 +0000[diff] [blame]293 <dt>Notes for code generation</dt>
Eli Friedmand577a062011-08-12 01:26:06 +0000[diff] [blame]294 <dd>See the section on Acquire; a fence before the relevant operation is
Eli Friedman79d7de72011-08-12 03:38:32 +0000[diff] [blame]295 usually sufficient for Release. Note that a store-store fence is not
Eli Friedmand577a062011-08-12 01:26:06 +0000[diff] [blame]296 sufficient to implement Release semantics; store-store fences are
297 generally not exposed to IR because they are extremely difficult to
Eli Friedman1bf4ad42011-08-11 23:44:25 +0000[diff] [blame]298 use correctly.</dd>
299</dl>
300
301</div>
302
303<!-- ======================================================================= -->
304<h3>
305 <a name="o_acqrel">AcquireRelease</a>
306</h3>
307
308<div>
309
310<p>AcquireRelease (<code>acq_rel</code> in IR) provides both an Acquire and a
311 Release barrier (for fences and operations which both read and write memory).
312
313<dl>
314 <dt>Relevant standard</dt>
315 <dd>This corresponds to the C++0x/C1x <code>memory_order_acq_rel</code>.
316 <dt>Notes for frontends</dt>
317 <dd>If you are writing a frontend which uses this directly, use with caution.
318 Acquire only provides a semantic guarantee when paired with a Release
319 operation, and vice versa.</dd>
320 <dt>Notes for optimizers</dt>
321 <dd>In general, optimizers should treat this like a nothrow call; the
322 the possible optimizations are usually not interesting.</dd>
323 <dt>Notes for code generation</dt>
324 <dd>This operation has Acquire and Release semantics; see the sections on
Eli Friedman5093fe62011-08-11 23:48:52 +0000[diff] [blame]325 Acquire and Release.</dd>
Eli Friedman1bf4ad42011-08-11 23:44:25 +0000[diff] [blame]326</dl>
327
328</div>
329
330<!-- ======================================================================= -->
331<h3>
332 <a name="o_seqcst">SequentiallyConsistent</a>
333</h3>
334
335<div>
336
Andrew Tricka1b953b2011-08-12 00:36:38 +0000[diff] [blame]337<p>SequentiallyConsistent (<code>seq_cst</code> in IR) provides
338 Acquire semantics for loads and Release semantics for
339 stores. Additionally, it guarantees that a total ordering exists
340 between all SequentiallyConsistent operations.
Eli Friedman1bf4ad42011-08-11 23:44:25 +0000[diff] [blame]341
342<dl>
343 <dt>Relevant standard</dt>
344 <dd>This corresponds to the C++0x/C1x <code>memory_order_seq_cst</code>,
345 Java volatile, and the gcc-compatible <code>__sync_*</code> builtins
346 which do not specify otherwise.
347 <dt>Notes for frontends</dt>
348 <dd>If a frontend is exposing atomic operations, these are much easier to
349 reason about for the programmer than other kinds of operations, and using
350 them is generally a practical performance tradeoff.</dd>
351 <dt>Notes for optimizers</dt>
Eli Friedman79d7de72011-08-12 03:38:32 +0000[diff] [blame]352 <dd>Optimizers not aware of atomics can treat this like a nothrow call.
353 For SequentiallyConsistent loads and stores, the same reorderings are
354 allowed as for Acquire loads and Release stores, except that
355 SequentiallyConsistent operations may not be reordered.</dd>
Eli Friedman1bf4ad42011-08-11 23:44:25 +0000[diff] [blame]356 <dt>Notes for code generation</dt>
Andrew Tricka1b953b2011-08-12 00:36:38 +0000[diff] [blame]357 <dd>SequentiallyConsistent loads minimally require the same barriers
Eli Friedman79d7de72011-08-12 03:38:32 +0000[diff] [blame]358 as Acquire operations and SequeuentiallyConsistent stores require
359 Release barriers. Additionally, the code generator must enforce
360 ordering between SequeuentiallyConsistent stores followed by
361 SequeuentiallyConsistent loads. This is usually done by emitting
362 either a full fence before the loads or a full fence after the
363 stores; which is preferred varies by architecture.</dd>
Eli Friedman1bf4ad42011-08-11 23:44:25 +0000[diff] [blame]364</dl>
365
366</div>
367
368</div>
369
370<!-- *********************************************************************** -->
371<h2>
Eli Friedman138515d2011-08-09 21:07:10 +0000[diff] [blame]372 <a name="iropt">Atomics and IR optimization</a>
373</h2>
374<!-- *********************************************************************** -->
375
376<div>
377
378<p>Predicates for optimizer writers to query:
379<ul>
380 <li>isSimple(): A load or store which is not volatile or atomic. This is
381 what, for example, memcpyopt would check for operations it might
382 transform.
383 <li>isUnordered(): A load or store which is not volatile and at most
384 Unordered. This would be checked, for example, by LICM before hoisting
385 an operation.
386 <li>mayReadFromMemory()/mayWriteToMemory(): Existing predicate, but note
Eli Friedmane2d8cf72011-08-10 20:17:43 +0000[diff] [blame]387 that they return true for any operation which is volatile or at least
Eli Friedman138515d2011-08-09 21:07:10 +0000[diff] [blame]388 Monotonic.
389 <li>Alias analysis: Note that AA will return ModRef for anything Acquire or
390 Release, and for the address accessed by any Monotonic operation.
391</ul>
392
393<p>There are essentially two components to supporting atomic operations. The
394 first is making sure to query isSimple() or isUnordered() instead
395 of isVolatile() before transforming an operation. The other piece is
396 making sure that a transform does not end up replacing, for example, an
397 Unordered operation with a non-atomic operation. Most of the other
398 necessary checks automatically fall out from existing predicates and
399 alias analysis queries.</p>
400
401<p>Some examples of how optimizations interact with various kinds of atomic
402 operations:
403<ul>
404 <li>memcpyopt: An atomic operation cannot be optimized into part of a
405 memcpy/memset, including unordered loads/stores. It can pull operations
406 across some atomic operations.
407 <li>LICM: Unordered loads/stores can be moved out of a loop. It just treats
408 monotonic operations like a read+write to a memory location, and anything
409 stricter than that like a nothrow call.
410 <li>DSE: Unordered stores can be DSE'ed like normal stores. Monotonic stores
411 can be DSE'ed in some cases, but it's tricky to reason about, and not
412 especially important.
413 <li>Folding a load: Any atomic load from a constant global can be
414 constant-folded, because it cannot be observed. Similar reasoning allows
415 scalarrepl with atomic loads and stores.
416</ul>
417
418</div>
419
420<!-- *********************************************************************** -->
421<h2>
422 <a name="codegen">Atomics and Codegen</a>
423</h2>
424<!-- *********************************************************************** -->
425
426<div>
427
428<p>Atomic operations are represented in the SelectionDAG with
429 <code>ATOMIC_*</code> opcodes. On architectures which use barrier
430 instructions for all atomic ordering (like ARM), appropriate fences are
431 split out as the DAG is built.</p>
432
433<p>The MachineMemOperand for all atomic operations is currently marked as
434 volatile; this is not correct in the IR sense of volatile, but CodeGen
435 handles anything marked volatile very conservatively. This should get
436 fixed at some point.</p>
437
Eli Friedman1bf4ad42011-08-11 23:44:25 +0000[diff] [blame]438<p>Common architectures have some way of representing at least a pointer-sized
439 lock-free <code>cmpxchg</code>; such an operation can be used to implement
440 all the other atomic operations which can be represented in IR up to that
441 size. Backends are expected to implement all those operations, but not
442 operations which cannot be implemented in a lock-free manner. It is
443 expected that backends will give an error when given an operation which
444 cannot be implemented. (The LLVM code generator is not very helpful here
445 at the moment, but hopefully that will change.)</p>
446
Eli Friedman138515d2011-08-09 21:07:10 +0000[diff] [blame]447<p>The implementation of atomics on LL/SC architectures (like ARM) is currently
448 a bit of a mess; there is a lot of copy-pasted code across targets, and
449 the representation is relatively unsuited to optimization (it would be nice
450 to be able to optimize loops involving cmpxchg etc.).</p>
451
452<p>On x86, all atomic loads generate a <code>MOV</code>.
453 SequentiallyConsistent stores generate an <code>XCHG</code>, other stores
454 generate a <code>MOV</code>. SequentiallyConsistent fences generate an
455 <code>MFENCE</code>, other fences do not cause any code to be generated.
456 cmpxchg uses the <code>LOCK CMPXCHG</code> instruction.
457 <code>atomicrmw xchg</code> uses <code>XCHG</code>,
458 <code>atomicrmw add</code> and <code>atomicrmw sub</code> use
459 <code>XADD</code>, and all other <code>atomicrmw</code> operations generate
460 a loop with <code>LOCK CMPXCHG</code>. Depending on the users of the
461 result, some <code>atomicrmw</code> operations can be translated into
462 operations like <code>LOCK AND</code>, but that does not work in
463 general.</p>
464
465<p>On ARM, MIPS, and many other RISC architectures, Acquire, Release, and
466 SequentiallyConsistent semantics require barrier instructions
467 for every such operation. Loads and stores generate normal instructions.
Eli Friedman1bf4ad42011-08-11 23:44:25 +0000[diff] [blame]468 <code>cmpxchg</code> and <code>atomicrmw</code> can be represented using
469 a loop with LL/SC-style instructions which take some sort of exclusive
470 lock on a cache line (<code>LDREX</code> and <code>STREX</code> on
471 ARM, etc.). At the moment, the IR does not provide any way to represent a
472 weak <code>cmpxchg</code> which would not require a loop.</p>
Eli Friedman138515d2011-08-09 21:07:10 +0000[diff] [blame]473</div>
474
475<!-- *********************************************************************** -->
476
477<hr>
478<address>
479 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
480 src="http://jigsaw.w3.org/css-validator/images/vcss-blue" alt="Valid CSS"></a>
481 <a href="http://validator.w3.org/check/referer"><img
482 src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01"></a>
483
484 <a href="http://llvm.org/">LLVM Compiler Infrastructure</a><br>
485 Last modified: $Date: 2011-08-09 02:07:00 -0700 (Tue, 09 Aug 2011) $
486</address>
487
488</body>
489</html>