ViewVC logotype

Contents of /code/trunk/doc/html/pcrejit.html

Parent Directory Parent Directory | Revision Log Revision Log

Revision 869 - (show annotations)
Sat Jan 14 11:16:23 2012 UTC (3 years, 8 months ago) by ph10
File MIME type: text/html
File size: 16407 byte(s)
Error occurred while calculating annotation data.
Bring HTML docs up to date.
1 <html>
2 <head>
3 <title>pcrejit specification</title>
4 </head>
5 <body bgcolor="#FFFFFF" text="#00005A" link="#0066FF" alink="#3399FF" vlink="#2222BB">
6 <h1>pcrejit man page</h1>
7 <p>
8 Return to the <a href="index.html">PCRE index page</a>.
9 </p>
10 <p>
11 This page is part of the PCRE HTML documentation. It was generated automatically
12 from the original man page. If there is any nonsense in it, please consult the
13 man page, in case the conversion went wrong.
14 <br>
15 <ul>
16 <li><a name="TOC1" href="#SEC1">PCRE JUST-IN-TIME COMPILER SUPPORT</a>
17 <li><a name="TOC2" href="#SEC2">8-BIT and 16-BIT SUPPORT</a>
18 <li><a name="TOC3" href="#SEC3">AVAILABILITY OF JIT SUPPORT</a>
19 <li><a name="TOC4" href="#SEC4">SIMPLE USE OF JIT</a>
21 <li><a name="TOC6" href="#SEC6">RETURN VALUES FROM JIT EXECUTION</a>
23 <li><a name="TOC8" href="#SEC8">CONTROLLING THE JIT STACK</a>
24 <li><a name="TOC9" href="#SEC9">JIT STACK FAQ</a>
25 <li><a name="TOC10" href="#SEC10">EXAMPLE CODE</a>
26 <li><a name="TOC11" href="#SEC11">SEE ALSO</a>
27 <li><a name="TOC12" href="#SEC12">AUTHOR</a>
28 <li><a name="TOC13" href="#SEC13">REVISION</a>
29 </ul>
30 <br><a name="SEC1" href="#TOC1">PCRE JUST-IN-TIME COMPILER SUPPORT</a><br>
31 <P>
32 Just-in-time compiling is a heavyweight optimization that can greatly speed up
33 pattern matching. However, it comes at the cost of extra processing before the
34 match is performed. Therefore, it is of most benefit when the same pattern is
35 going to be matched many times. This does not necessarily mean many calls of a
36 matching function; if the pattern is not anchored, matching attempts may take
37 place many times at various positions in the subject, even for a single call.
38 Therefore, if the subject string is very long, it may still pay to use JIT for
39 one-off matches.
40 </P>
41 <P>
42 JIT support applies only to the traditional Perl-compatible matching function.
43 It does not apply when the DFA matching function is being used. The code for
44 this support was written by Zoltan Herczeg.
45 </P>
46 <br><a name="SEC2" href="#TOC1">8-BIT and 16-BIT SUPPORT</a><br>
47 <P>
48 JIT support is available for both the 8-bit and 16-bit PCRE libraries. To keep
49 this documentation simple, only the 8-bit interface is described in what
50 follows. If you are using the 16-bit library, substitute the 16-bit functions
51 and 16-bit structures (for example, <i>pcre16_jit_stack</i> instead of
52 <i>pcre_jit_stack</i>).
53 </P>
54 <br><a name="SEC3" href="#TOC1">AVAILABILITY OF JIT SUPPORT</a><br>
55 <P>
56 JIT support is an optional feature of PCRE. The "configure" option --enable-jit
57 (or equivalent CMake option) must be set when PCRE is built if you want to use
58 JIT. The support is limited to the following hardware platforms:
59 <pre>
60 ARM v5, v7, and Thumb2
61 Intel x86 32-bit and 64-bit
62 MIPS 32-bit
63 Power PC 32-bit and 64-bit
64 </pre>
65 The Power PC support is designated as experimental because it has not been
66 fully tested. If --enable-jit is set on an unsupported platform, compilation
67 fails.
68 </P>
69 <P>
70 A program that is linked with PCRE 8.20 or later can tell if JIT support is
71 available by calling <b>pcre_config()</b> with the PCRE_CONFIG_JIT option. The
72 result is 1 when JIT is available, and 0 otherwise. However, a simple program
73 does not need to check this in order to use JIT. The API is implemented in a
74 way that falls back to the ordinary PCRE code if JIT is not available.
75 </P>
76 <P>
77 If your program may sometimes be linked with versions of PCRE that are older
78 than 8.20, but you want to use JIT when it is available, you can test
79 the values of PCRE_MAJOR and PCRE_MINOR, or the existence of a JIT macro such
80 as PCRE_CONFIG_JIT, for compile-time control of your code.
81 </P>
82 <br><a name="SEC4" href="#TOC1">SIMPLE USE OF JIT</a><br>
83 <P>
84 You have to do two things to make use of the JIT support in the simplest way:
85 <pre>
86 (1) Call <b>pcre_study()</b> with the PCRE_STUDY_JIT_COMPILE option for
87 each compiled pattern, and pass the resulting <b>pcre_extra</b> block to
88 <b>pcre_exec()</b>.
90 (2) Use <b>pcre_free_study()</b> to free the <b>pcre_extra</b> block when it is
91 no longer needed instead of just freeing it yourself. This
92 ensures that any JIT data is also freed.
93 </pre>
94 For a program that may be linked with pre-8.20 versions of PCRE, you can insert
95 <pre>
98 #endif
99 </pre>
100 so that no option is passed to <b>pcre_study()</b>, and then use something like
101 this to free the study data:
102 <pre>
103 #ifdef PCRE_CONFIG_JIT
104 pcre_free_study(study_ptr);
105 #else
106 pcre_free(study_ptr);
107 #endif
108 </pre>
109 In some circumstances you may need to call additional functions. These are
110 described in the section entitled
111 <a href="#stackcontrol">"Controlling the JIT stack"</a>
112 below.
113 </P>
114 <P>
115 If JIT support is not available, PCRE_STUDY_JIT_COMPILE is ignored, and no JIT
116 data is set up. Otherwise, the compiled pattern is passed to the JIT compiler,
117 which turns it into machine code that executes much faster than the normal
118 interpretive code. When <b>pcre_exec()</b> is passed a <b>pcre_extra</b> block
119 containing a pointer to JIT code, it obeys that instead of the normal code. The
120 result is identical, but the code runs much faster.
121 </P>
122 <P>
123 There are some <b>pcre_exec()</b> options that are not supported for JIT
124 execution. There are also some pattern items that JIT cannot handle. Details
125 are given below. In both cases, execution automatically falls back to the
126 interpretive code.
127 </P>
128 <P>
129 If the JIT compiler finds an unsupported item, no JIT data is generated. You
130 can find out if JIT execution is available after studying a pattern by calling
131 <b>pcre_fullinfo()</b> with the PCRE_INFO_JIT option. A result of 1 means that
132 JIT compilation was successful. A result of 0 means that JIT support is not
133 available, or the pattern was not studied with PCRE_STUDY_JIT_COMPILE, or the
134 JIT compiler was not able to handle the pattern.
135 </P>
136 <P>
137 Once a pattern has been studied, with or without JIT, it can be used as many
138 times as you like for matching different subject strings.
139 </P>
140 <br><a name="SEC5" href="#TOC1">UNSUPPORTED OPTIONS AND PATTERN ITEMS</a><br>
141 <P>
142 The only <b>pcre_exec()</b> options that are supported for JIT execution are
144 PCRE_NOTEMPTY_ATSTART. Note in particular that partial matching is not
145 supported.
146 </P>
147 <P>
148 The unsupported pattern items are:
149 <pre>
150 \C match a single byte; not supported in UTF-8 mode
151 (?Cn) callouts
152 (*COMMIT) )
153 (*MARK) )
154 (*PRUNE) ) the backtracking control verbs
155 (*SKIP) )
156 (*THEN) )
157 </pre>
158 Support for some of these may be added in future.
159 </P>
160 <br><a name="SEC6" href="#TOC1">RETURN VALUES FROM JIT EXECUTION</a><br>
161 <P>
162 When a pattern is matched using JIT execution, the return values are the same
163 as those given by the interpretive <b>pcre_exec()</b> code, with the addition of
164 one new error code: PCRE_ERROR_JIT_STACKLIMIT. This means that the memory used
165 for the JIT stack was insufficient. See
166 <a href="#stackcontrol">"Controlling the JIT stack"</a>
167 below for a discussion of JIT stack usage. For compatibility with the
168 interpretive <b>pcre_exec()</b> code, no more than two-thirds of the
169 <i>ovector</i> argument is used for passing back captured substrings.
170 </P>
171 <P>
172 The error code PCRE_ERROR_MATCHLIMIT is returned by the JIT code if searching a
173 very large pattern tree goes on for too long, as it is in the same circumstance
174 when JIT is not used, but the details of exactly what is counted are not the
175 same. The PCRE_ERROR_RECURSIONLIMIT error code is never returned by JIT
176 execution.
177 </P>
178 <br><a name="SEC7" href="#TOC1">SAVING AND RESTORING COMPILED PATTERNS</a><br>
179 <P>
180 The code that is generated by the JIT compiler is architecture-specific, and is
181 also position dependent. For those reasons it cannot be saved (in a file or
182 database) and restored later like the bytecode and other data of a compiled
183 pattern. Saving and restoring compiled patterns is not something many people
184 do. More detail about this facility is given in the
185 <a href="pcreprecompile.html"><b>pcreprecompile</b></a>
186 documentation. It should be possible to run <b>pcre_study()</b> on a saved and
187 restored pattern, and thereby recreate the JIT data, but because JIT
188 compilation uses significant resources, it is probably not worth doing this;
189 you might as well recompile the original pattern.
190 <a name="stackcontrol"></a></P>
191 <br><a name="SEC8" href="#TOC1">CONTROLLING THE JIT STACK</a><br>
192 <P>
193 When the compiled JIT code runs, it needs a block of memory to use as a stack.
194 By default, it uses 32K on the machine stack. However, some large or
195 complicated patterns need more than this. The error PCRE_ERROR_JIT_STACKLIMIT
196 is given when there is not enough stack. Three functions are provided for
197 managing blocks of memory for use as JIT stacks. There is further discussion
198 about the use of JIT stacks in the section entitled
199 <a href="#stackcontrol">"JIT stack FAQ"</a>
200 below.
201 </P>
202 <P>
203 The <b>pcre_jit_stack_alloc()</b> function creates a JIT stack. Its arguments
204 are a starting size and a maximum size, and it returns a pointer to an opaque
205 structure of type <b>pcre_jit_stack</b>, or NULL if there is an error. The
206 <b>pcre_jit_stack_free()</b> function can be used to free a stack that is no
207 longer needed. (For the technically minded: the address space is allocated by
208 mmap or VirtualAlloc.)
209 </P>
210 <P>
211 JIT uses far less memory for recursion than the interpretive code,
212 and a maximum stack size of 512K to 1M should be more than enough for any
213 pattern.
214 </P>
215 <P>
216 The <b>pcre_assign_jit_stack()</b> function specifies which stack JIT code
217 should use. Its arguments are as follows:
218 <pre>
219 pcre_extra *extra
220 pcre_jit_callback callback
221 void *data
222 </pre>
223 The <i>extra</i> argument must be the result of studying a pattern with
224 PCRE_STUDY_JIT_COMPILE. There are three cases for the values of the other two
225 options:
226 <pre>
227 (1) If <i>callback</i> is NULL and <i>data</i> is NULL, an internal 32K block
228 on the machine stack is used.
230 (2) If <i>callback</i> is NULL and <i>data</i> is not NULL, <i>data</i> must be
231 a valid JIT stack, the result of calling <b>pcre_jit_stack_alloc()</b>.
233 (3) If <i>callback</i> not NULL, it must point to a function that is called
234 with <i>data</i> as an argument at the start of matching, in order to
235 set up a JIT stack. If the result is NULL, the internal 32K stack
236 is used; otherwise the return value must be a valid JIT stack,
237 the result of calling <b>pcre_jit_stack_alloc()</b>.
238 </pre>
239 You may safely assign the same JIT stack to more than one pattern, as long as
240 they are all matched sequentially in the same thread. In a multithread
241 application, each thread must use its own JIT stack.
242 </P>
243 <P>
244 Strictly speaking, even more is allowed. You can assign the same stack to any
245 number of patterns as long as they are not used for matching by multiple
246 threads at the same time. For example, you can assign the same stack to all
247 compiled patterns, and use a global mutex in the callback to wait until the
248 stack is available for use. However, this is an inefficient solution, and
249 not recommended.
250 </P>
251 <P>
252 This is a suggestion for how a typical multithreaded program might operate:
253 <pre>
254 During thread initalization
255 thread_local_var = pcre_jit_stack_alloc(...)
257 During thread exit
258 pcre_jit_stack_free(thread_local_var)
260 Use a one-line callback function
261 return thread_local_var
262 </pre>
263 All the functions described in this section do nothing if JIT is not available,
264 and <b>pcre_assign_jit_stack()</b> does nothing unless the <b>extra</b> argument
265 is non-NULL and points to a <b>pcre_extra</b> block that is the result of a
266 successful study with PCRE_STUDY_JIT_COMPILE.
267 <a name="stackfaq"></a></P>
268 <br><a name="SEC9" href="#TOC1">JIT STACK FAQ</a><br>
269 <P>
270 (1) Why do we need JIT stacks?
271 <br>
272 <br>
273 PCRE (and JIT) is a recursive, depth-first engine, so it needs a stack where
274 the local data of the current node is pushed before checking its child nodes.
275 Allocating real machine stack on some platforms is difficult. For example, the
276 stack chain needs to be updated every time if we extend the stack on PowerPC.
277 Although it is possible, its updating time overhead decreases performance. So
278 we do the recursion in memory.
279 </P>
280 <P>
281 (2) Why don't we simply allocate blocks of memory with <b>malloc()</b>?
282 <br>
283 <br>
284 Modern operating systems have a nice feature: they can reserve an address space
285 instead of allocating memory. We can safely allocate memory pages inside this
286 address space, so the stack could grow without moving memory data (this is
287 important because of pointers). Thus we can allocate 1M address space, and use
288 only a single memory page (usually 4K) if that is enough. However, we can still
289 grow up to 1M anytime if needed.
290 </P>
291 <P>
292 (3) Who "owns" a JIT stack?
293 <br>
294 <br>
295 The owner of the stack is the user program, not the JIT studied pattern or
296 anything else. The user program must ensure that if a stack is used by
297 <b>pcre_exec()</b>, (that is, it is assigned to the pattern currently running),
298 that stack must not be used by any other threads (to avoid overwriting the same
299 memory area). The best practice for multithreaded programs is to allocate a
300 stack for each thread, and return this stack through the JIT callback function.
301 </P>
302 <P>
303 (4) When should a JIT stack be freed?
304 <br>
305 <br>
306 You can free a JIT stack at any time, as long as it will not be used by
307 <b>pcre_exec()</b> again. When you assign the stack to a pattern, only a pointer
308 is set. There is no reference counting or any other magic. You can free the
309 patterns and stacks in any order, anytime. Just <i>do not</i> call
310 <b>pcre_exec()</b> with a pattern pointing to an already freed stack, as that
311 will cause SEGFAULT. (Also, do not free a stack currently used by
312 <b>pcre_exec()</b> in another thread). You can also replace the stack for a
313 pattern at any time. You can even free the previous stack before assigning a
314 replacement.
315 </P>
316 <P>
317 (5) Should I allocate/free a stack every time before/after calling
318 <b>pcre_exec()</b>?
319 <br>
320 <br>
321 No, because this is too costly in terms of resources. However, you could
322 implement some clever idea which release the stack if it is not used in let's
323 say two minutes. The JIT callback can help to achive this without keeping a
324 list of the currently JIT studied patterns.
325 </P>
326 <P>
327 (6) OK, the stack is for long term memory allocation. But what happens if a
328 pattern causes stack overflow with a stack of 1M? Is that 1M kept until the
329 stack is freed?
330 <br>
331 <br>
332 Especially on embedded sytems, it might be a good idea to release
333 memory sometimes without freeing the stack. There is no API for this at the
334 moment. Probably a function call which returns with the currently allocated
335 memory for any stack and another which allows releasing memory (shrinking the
336 stack) would be a good idea if someone needs this.
337 </P>
338 <P>
339 (7) This is too much of a headache. Isn't there any better solution for JIT
340 stack handling?
341 <br>
342 <br>
343 No, thanks to Windows. If POSIX threads were used everywhere, we could throw
344 out this complicated API.
345 </P>
346 <br><a name="SEC10" href="#TOC1">EXAMPLE CODE</a><br>
347 <P>
348 This is a single-threaded example that specifies a JIT stack without using a
349 callback.
350 <pre>
351 int rc;
352 int ovector[30];
353 pcre *re;
354 pcre_extra *extra;
355 pcre_jit_stack *jit_stack;
357 re = pcre_compile(pattern, 0, &error, &erroffset, NULL);
358 /* Check for errors */
359 extra = pcre_study(re, PCRE_STUDY_JIT_COMPILE, &error);
360 jit_stack = pcre_jit_stack_alloc(32*1024, 512*1024);
361 /* Check for error (NULL) */
362 pcre_assign_jit_stack(extra, NULL, jit_stack);
363 rc = pcre_exec(re, extra, subject, length, 0, 0, ovector, 30);
364 /* Check results */
365 pcre_free(re);
366 pcre_free_study(extra);
367 pcre_jit_stack_free(jit_stack);
369 </PRE>
370 </P>
371 <br><a name="SEC11" href="#TOC1">SEE ALSO</a><br>
372 <P>
373 <b>pcreapi</b>(3)
374 </P>
375 <br><a name="SEC12" href="#TOC1">AUTHOR</a><br>
376 <P>
377 Philip Hazel (FAQ by Zoltan Herczeg)
378 <br>
379 University Computing Service
380 <br>
381 Cambridge CB2 3QH, England.
382 <br>
383 </P>
384 <br><a name="SEC13" href="#TOC1">REVISION</a><br>
385 <P>
386 Last updated: 08 January 2012
387 <br>
388 Copyright &copy; 1997-2012 University of Cambridge.
389 <br>
390 <p>
391 Return to the <a href="index.html">PCRE index page</a>.
392 </p>

  ViewVC Help
Powered by ViewVC 1.1.5