ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1706 - (show annotations)
Wed Jul 5 15:45:39 2017 UTC (3 years, 10 months ago) by ph10
File MIME type: text/html
File size: 22103 byte(s)
Final file tidies for 8.41 release.
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, 16-BIT AND 32-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">JIT FAST PATH API</a>
27 <li><a name="TOC12" href="#SEC12">SEE ALSO</a>
28 <li><a name="TOC13" href="#SEC13">AUTHOR</a>
29 <li><a name="TOC14" href="#SEC14">REVISION</a>
30 </ul>
31 <br><a name="SEC1" href="#TOC1">PCRE JUST-IN-TIME COMPILER SUPPORT</a><br>
32 <P>
33 Just-in-time compiling is a heavyweight optimization that can greatly speed up
34 pattern matching. However, it comes at the cost of extra processing before the
35 match is performed. Therefore, it is of most benefit when the same pattern is
36 going to be matched many times. This does not necessarily mean many calls of a
37 matching function; if the pattern is not anchored, matching attempts may take
38 place many times at various positions in the subject, even for a single call.
39 Therefore, if the subject string is very long, it may still pay to use JIT for
40 one-off matches.
41 </P>
42 <P>
43 JIT support applies only to the traditional Perl-compatible matching function.
44 It does not apply when the DFA matching function is being used. The code for
45 this support was written by Zoltan Herczeg.
46 </P>
47 <br><a name="SEC2" href="#TOC1">8-BIT, 16-BIT AND 32-BIT SUPPORT</a><br>
48 <P>
49 JIT support is available for all of the 8-bit, 16-bit and 32-bit PCRE
50 libraries. To keep this documentation simple, only the 8-bit interface is
51 described in what follows. If you are using the 16-bit library, substitute the
52 16-bit functions and 16-bit structures (for example, <i>pcre16_jit_stack</i>
53 instead of <i>pcre_jit_stack</i>). If you are using the 32-bit library,
54 substitute the 32-bit functions and 32-bit structures (for example,
55 <i>pcre32_jit_stack</i> instead of <i>pcre_jit_stack</i>).
56 </P>
57 <br><a name="SEC3" href="#TOC1">AVAILABILITY OF JIT SUPPORT</a><br>
58 <P>
59 JIT support is an optional feature of PCRE. The "configure" option --enable-jit
60 (or equivalent CMake option) must be set when PCRE is built if you want to use
61 JIT. The support is limited to the following hardware platforms:
62 <pre>
63 ARM v5, v7, and Thumb2
64 Intel x86 32-bit and 64-bit
65 MIPS 32-bit
66 Power PC 32-bit and 64-bit
67 SPARC 32-bit (experimental)
68 </pre>
69 If --enable-jit is set on an unsupported platform, compilation fails.
70 </P>
71 <P>
72 A program that is linked with PCRE 8.20 or later can tell if JIT support is
73 available by calling <b>pcre_config()</b> with the PCRE_CONFIG_JIT option. The
74 result is 1 when JIT is available, and 0 otherwise. However, a simple program
75 does not need to check this in order to use JIT. The normal API is implemented
76 in a way that falls back to the interpretive code if JIT is not available. For
77 programs that need the best possible performance, there is also a "fast path"
78 API that is JIT-specific.
79 </P>
80 <P>
81 If your program may sometimes be linked with versions of PCRE that are older
82 than 8.20, but you want to use JIT when it is available, you can test the
83 values of PCRE_MAJOR and PCRE_MINOR, or the existence of a JIT macro such as
84 PCRE_CONFIG_JIT, for compile-time control of your code. Also beware that the
85 <b>pcre_jit_exec()</b> function was not available at all before 8.32,
86 and may not be available at all if PCRE isn't compiled with
87 --enable-jit. See the "JIT FAST PATH API" section below for details.
88 </P>
89 <br><a name="SEC4" href="#TOC1">SIMPLE USE OF JIT</a><br>
90 <P>
91 You have to do two things to make use of the JIT support in the simplest way:
92 <pre>
93 (1) Call <b>pcre_study()</b> with the PCRE_STUDY_JIT_COMPILE option for
94 each compiled pattern, and pass the resulting <b>pcre_extra</b> block to
95 <b>pcre_exec()</b>.
97 (2) Use <b>pcre_free_study()</b> to free the <b>pcre_extra</b> block when it is
98 no longer needed, instead of just freeing it yourself. This ensures that
99 any JIT data is also freed.
100 </pre>
101 For a program that may be linked with pre-8.20 versions of PCRE, you can insert
102 <pre>
105 #endif
106 </pre>
107 so that no option is passed to <b>pcre_study()</b>, and then use something like
108 this to free the study data:
109 <pre>
110 #ifdef PCRE_CONFIG_JIT
111 pcre_free_study(study_ptr);
112 #else
113 pcre_free(study_ptr);
114 #endif
115 </pre>
116 PCRE_STUDY_JIT_COMPILE requests the JIT compiler to generate code for complete
117 matches. If you want to run partial matches using the PCRE_PARTIAL_HARD or
118 PCRE_PARTIAL_SOFT options of <b>pcre_exec()</b>, you should set one or both of
119 the following options in addition to, or instead of, PCRE_STUDY_JIT_COMPILE
120 when you call <b>pcre_study()</b>:
121 <pre>
124 </pre>
125 If using <b>pcre_jit_exec()</b> and supporting a pre-8.32 version of
126 PCRE, you can insert:
127 <pre>
128 #if PCRE_MAJOR &#62;= 8 && PCRE_MINOR &#62;= 32
129 pcre_jit_exec(...);
130 #else
131 pcre_exec(...)
132 #endif
133 </pre>
134 but as described in the "JIT FAST PATH API" section below this assumes
135 version 8.32 and later are compiled with --enable-jit, which may
136 break.
137 <br>
138 <br>
139 The JIT compiler generates different optimized code for each of the three
140 modes (normal, soft partial, hard partial). When <b>pcre_exec()</b> is called,
141 the appropriate code is run if it is available. Otherwise, the pattern is
142 matched using interpretive code.
143 </P>
144 <P>
145 In some circumstances you may need to call additional functions. These are
146 described in the section entitled
147 <a href="#stackcontrol">"Controlling the JIT stack"</a>
148 below.
149 </P>
150 <P>
151 If JIT support is not available, PCRE_STUDY_JIT_COMPILE etc. are ignored, and
152 no JIT data is created. Otherwise, the compiled pattern is passed to the JIT
153 compiler, which turns it into machine code that executes much faster than the
154 normal interpretive code. When <b>pcre_exec()</b> is passed a <b>pcre_extra</b>
155 block containing a pointer to JIT code of the appropriate mode (normal or
156 hard/soft partial), it obeys that code instead of running the interpreter. The
157 result is identical, but the compiled JIT code runs much faster.
158 </P>
159 <P>
160 There are some <b>pcre_exec()</b> options that are not supported for JIT
161 execution. There are also some pattern items that JIT cannot handle. Details
162 are given below. In both cases, execution automatically falls back to the
163 interpretive code. If you want to know whether JIT was actually used for a
164 particular match, you should arrange for a JIT callback function to be set up
165 as described in the section entitled
166 <a href="#stackcontrol">"Controlling the JIT stack"</a>
167 below, even if you do not need to supply a non-default JIT stack. Such a
168 callback function is called whenever JIT code is about to be obeyed. If the
169 execution options are not right for JIT execution, the callback function is not
170 obeyed.
171 </P>
172 <P>
173 If the JIT compiler finds an unsupported item, no JIT data is generated. You
174 can find out if JIT execution is available after studying a pattern by calling
175 <b>pcre_fullinfo()</b> with the PCRE_INFO_JIT option. A result of 1 means that
176 JIT compilation was successful. A result of 0 means that JIT support is not
177 available, or the pattern was not studied with PCRE_STUDY_JIT_COMPILE etc., or
178 the JIT compiler was not able to handle the pattern.
179 </P>
180 <P>
181 Once a pattern has been studied, with or without JIT, it can be used as many
182 times as you like for matching different subject strings.
183 </P>
184 <br><a name="SEC5" href="#TOC1">UNSUPPORTED OPTIONS AND PATTERN ITEMS</a><br>
185 <P>
186 The only <b>pcre_exec()</b> options that are supported for JIT execution are
190 </P>
191 <P>
192 The only unsupported pattern items are \C (match a single data unit) when
193 running in a UTF mode, and a callout immediately before an assertion condition
194 in a conditional group.
195 </P>
196 <br><a name="SEC6" href="#TOC1">RETURN VALUES FROM JIT EXECUTION</a><br>
197 <P>
198 When a pattern is matched using JIT execution, the return values are the same
199 as those given by the interpretive <b>pcre_exec()</b> code, with the addition of
200 one new error code: PCRE_ERROR_JIT_STACKLIMIT. This means that the memory used
201 for the JIT stack was insufficient. See
202 <a href="#stackcontrol">"Controlling the JIT stack"</a>
203 below for a discussion of JIT stack usage. For compatibility with the
204 interpretive <b>pcre_exec()</b> code, no more than two-thirds of the
205 <i>ovector</i> argument is used for passing back captured substrings.
206 </P>
207 <P>
208 The error code PCRE_ERROR_MATCHLIMIT is returned by the JIT code if searching a
209 very large pattern tree goes on for too long, as it is in the same circumstance
210 when JIT is not used, but the details of exactly what is counted are not the
211 same. The PCRE_ERROR_RECURSIONLIMIT error code is never returned by JIT
212 execution.
213 </P>
214 <br><a name="SEC7" href="#TOC1">SAVING AND RESTORING COMPILED PATTERNS</a><br>
215 <P>
216 The code that is generated by the JIT compiler is architecture-specific, and is
217 also position dependent. For those reasons it cannot be saved (in a file or
218 database) and restored later like the bytecode and other data of a compiled
219 pattern. Saving and restoring compiled patterns is not something many people
220 do. More detail about this facility is given in the
221 <a href="pcreprecompile.html"><b>pcreprecompile</b></a>
222 documentation. It should be possible to run <b>pcre_study()</b> on a saved and
223 restored pattern, and thereby recreate the JIT data, but because JIT
224 compilation uses significant resources, it is probably not worth doing this;
225 you might as well recompile the original pattern.
226 <a name="stackcontrol"></a></P>
227 <br><a name="SEC8" href="#TOC1">CONTROLLING THE JIT STACK</a><br>
228 <P>
229 When the compiled JIT code runs, it needs a block of memory to use as a stack.
230 By default, it uses 32K on the machine stack. However, some large or
231 complicated patterns need more than this. The error PCRE_ERROR_JIT_STACKLIMIT
232 is given when there is not enough stack. Three functions are provided for
233 managing blocks of memory for use as JIT stacks. There is further discussion
234 about the use of JIT stacks in the section entitled
235 <a href="#stackcontrol">"JIT stack FAQ"</a>
236 below.
237 </P>
238 <P>
239 The <b>pcre_jit_stack_alloc()</b> function creates a JIT stack. Its arguments
240 are a starting size and a maximum size, and it returns a pointer to an opaque
241 structure of type <b>pcre_jit_stack</b>, or NULL if there is an error. The
242 <b>pcre_jit_stack_free()</b> function can be used to free a stack that is no
243 longer needed. (For the technically minded: the address space is allocated by
244 mmap or VirtualAlloc.)
245 </P>
246 <P>
247 JIT uses far less memory for recursion than the interpretive code,
248 and a maximum stack size of 512K to 1M should be more than enough for any
249 pattern.
250 </P>
251 <P>
252 The <b>pcre_assign_jit_stack()</b> function specifies which stack JIT code
253 should use. Its arguments are as follows:
254 <pre>
255 pcre_extra *extra
256 pcre_jit_callback callback
257 void *data
258 </pre>
259 The <i>extra</i> argument must be the result of studying a pattern with
260 PCRE_STUDY_JIT_COMPILE etc. There are three cases for the values of the other
261 two options:
262 <pre>
263 (1) If <i>callback</i> is NULL and <i>data</i> is NULL, an internal 32K block
264 on the machine stack is used.
266 (2) If <i>callback</i> is NULL and <i>data</i> is not NULL, <i>data</i> must be
267 a valid JIT stack, the result of calling <b>pcre_jit_stack_alloc()</b>.
269 (3) If <i>callback</i> is not NULL, it must point to a function that is
270 called with <i>data</i> as an argument at the start of matching, in
271 order to set up a JIT stack. If the return from the callback
272 function is NULL, the internal 32K stack is used; otherwise the
273 return value must be a valid JIT stack, the result of calling
274 <b>pcre_jit_stack_alloc()</b>.
275 </pre>
276 A callback function is obeyed whenever JIT code is about to be run; it is not
277 obeyed when <b>pcre_exec()</b> is called with options that are incompatible for
278 JIT execution. A callback function can therefore be used to determine whether a
279 match operation was executed by JIT or by the interpreter.
280 </P>
281 <P>
282 You may safely use the same JIT stack for more than one pattern (either by
283 assigning directly or by callback), as long as the patterns are all matched
284 sequentially in the same thread. In a multithread application, if you do not
285 specify a JIT stack, or if you assign or pass back NULL from a callback, that
286 is thread-safe, because each thread has its own machine stack. However, if you
287 assign or pass back a non-NULL JIT stack, this must be a different stack for
288 each thread so that the application is thread-safe.
289 </P>
290 <P>
291 Strictly speaking, even more is allowed. You can assign the same non-NULL stack
292 to any number of patterns as long as they are not used for matching by multiple
293 threads at the same time. For example, you can assign the same stack to all
294 compiled patterns, and use a global mutex in the callback to wait until the
295 stack is available for use. However, this is an inefficient solution, and not
296 recommended.
297 </P>
298 <P>
299 This is a suggestion for how a multithreaded program that needs to set up
300 non-default JIT stacks might operate:
301 <pre>
302 During thread initalization
303 thread_local_var = pcre_jit_stack_alloc(...)
305 During thread exit
306 pcre_jit_stack_free(thread_local_var)
308 Use a one-line callback function
309 return thread_local_var
310 </pre>
311 All the functions described in this section do nothing if JIT is not available,
312 and <b>pcre_assign_jit_stack()</b> does nothing unless the <b>extra</b> argument
313 is non-NULL and points to a <b>pcre_extra</b> block that is the result of a
314 successful study with PCRE_STUDY_JIT_COMPILE etc.
315 <a name="stackfaq"></a></P>
316 <br><a name="SEC9" href="#TOC1">JIT STACK FAQ</a><br>
317 <P>
318 (1) Why do we need JIT stacks?
319 <br>
320 <br>
321 PCRE (and JIT) is a recursive, depth-first engine, so it needs a stack where
322 the local data of the current node is pushed before checking its child nodes.
323 Allocating real machine stack on some platforms is difficult. For example, the
324 stack chain needs to be updated every time if we extend the stack on PowerPC.
325 Although it is possible, its updating time overhead decreases performance. So
326 we do the recursion in memory.
327 </P>
328 <P>
329 (2) Why don't we simply allocate blocks of memory with <b>malloc()</b>?
330 <br>
331 <br>
332 Modern operating systems have a nice feature: they can reserve an address space
333 instead of allocating memory. We can safely allocate memory pages inside this
334 address space, so the stack could grow without moving memory data (this is
335 important because of pointers). Thus we can allocate 1M address space, and use
336 only a single memory page (usually 4K) if that is enough. However, we can still
337 grow up to 1M anytime if needed.
338 </P>
339 <P>
340 (3) Who "owns" a JIT stack?
341 <br>
342 <br>
343 The owner of the stack is the user program, not the JIT studied pattern or
344 anything else. The user program must ensure that if a stack is used by
345 <b>pcre_exec()</b>, (that is, it is assigned to the pattern currently running),
346 that stack must not be used by any other threads (to avoid overwriting the same
347 memory area). The best practice for multithreaded programs is to allocate a
348 stack for each thread, and return this stack through the JIT callback function.
349 </P>
350 <P>
351 (4) When should a JIT stack be freed?
352 <br>
353 <br>
354 You can free a JIT stack at any time, as long as it will not be used by
355 <b>pcre_exec()</b> again. When you assign the stack to a pattern, only a pointer
356 is set. There is no reference counting or any other magic. You can free the
357 patterns and stacks in any order, anytime. Just <i>do not</i> call
358 <b>pcre_exec()</b> with a pattern pointing to an already freed stack, as that
359 will cause SEGFAULT. (Also, do not free a stack currently used by
360 <b>pcre_exec()</b> in another thread). You can also replace the stack for a
361 pattern at any time. You can even free the previous stack before assigning a
362 replacement.
363 </P>
364 <P>
365 (5) Should I allocate/free a stack every time before/after calling
366 <b>pcre_exec()</b>?
367 <br>
368 <br>
369 No, because this is too costly in terms of resources. However, you could
370 implement some clever idea which release the stack if it is not used in let's
371 say two minutes. The JIT callback can help to achieve this without keeping a
372 list of the currently JIT studied patterns.
373 </P>
374 <P>
375 (6) OK, the stack is for long term memory allocation. But what happens if a
376 pattern causes stack overflow with a stack of 1M? Is that 1M kept until the
377 stack is freed?
378 <br>
379 <br>
380 Especially on embedded sytems, it might be a good idea to release memory
381 sometimes without freeing the stack. There is no API for this at the moment.
382 Probably a function call which returns with the currently allocated memory for
383 any stack and another which allows releasing memory (shrinking the stack) would
384 be a good idea if someone needs this.
385 </P>
386 <P>
387 (7) This is too much of a headache. Isn't there any better solution for JIT
388 stack handling?
389 <br>
390 <br>
391 No, thanks to Windows. If POSIX threads were used everywhere, we could throw
392 out this complicated API.
393 </P>
394 <br><a name="SEC10" href="#TOC1">EXAMPLE CODE</a><br>
395 <P>
396 This is a single-threaded example that specifies a JIT stack without using a
397 callback.
398 <pre>
399 int rc;
400 int ovector[30];
401 pcre *re;
402 pcre_extra *extra;
403 pcre_jit_stack *jit_stack;
405 re = pcre_compile(pattern, 0, &error, &erroffset, NULL);
406 /* Check for errors */
407 extra = pcre_study(re, PCRE_STUDY_JIT_COMPILE, &error);
408 jit_stack = pcre_jit_stack_alloc(32*1024, 512*1024);
409 /* Check for error (NULL) */
410 pcre_assign_jit_stack(extra, NULL, jit_stack);
411 rc = pcre_exec(re, extra, subject, length, 0, 0, ovector, 30);
412 /* Check results */
413 pcre_free(re);
414 pcre_free_study(extra);
415 pcre_jit_stack_free(jit_stack);
417 </PRE>
418 </P>
419 <br><a name="SEC11" href="#TOC1">JIT FAST PATH API</a><br>
420 <P>
421 Because the API described above falls back to interpreted execution when JIT is
422 not available, it is convenient for programs that are written for general use
423 in many environments. However, calling JIT via <b>pcre_exec()</b> does have a
424 performance impact. Programs that are written for use where JIT is known to be
425 available, and which need the best possible performance, can instead use a
426 "fast path" API to call JIT execution directly instead of calling
427 <b>pcre_exec()</b> (obviously only for patterns that have been successfully
428 studied by JIT).
429 </P>
430 <P>
431 The fast path function is called <b>pcre_jit_exec()</b>, and it takes exactly
432 the same arguments as <b>pcre_exec()</b>, plus one additional argument that
433 must point to a JIT stack. The JIT stack arrangements described above do not
434 apply. The return values are the same as for <b>pcre_exec()</b>.
435 </P>
436 <P>
437 When you call <b>pcre_exec()</b>, as well as testing for invalid options, a
438 number of other sanity checks are performed on the arguments. For example, if
439 the subject pointer is NULL, or its length is negative, an immediate error is
440 given. Also, unless PCRE_NO_UTF[8|16|32] is set, a UTF subject string is tested
441 for validity. In the interests of speed, these checks do not happen on the JIT
442 fast path, and if invalid data is passed, the result is undefined.
443 </P>
444 <P>
445 Bypassing the sanity checks and the <b>pcre_exec()</b> wrapping can give
446 speedups of more than 10%.
447 </P>
448 <P>
449 Note that the <b>pcre_jit_exec()</b> function is not available in versions of
450 PCRE before 8.32 (released in November 2012). If you need to support versions
451 that old you must either use the slower <b>pcre_exec()</b>, or switch between
452 the two codepaths by checking the values of PCRE_MAJOR and PCRE_MINOR.
453 </P>
454 <P>
455 Due to an unfortunate implementation oversight, even in versions 8.32
456 and later there will be no <b>pcre_jit_exec()</b> stub function defined
457 when PCRE is compiled with --disable-jit, which is the default, and
458 there's no way to detect whether PCRE was compiled with --enable-jit
459 via a macro.
460 </P>
461 <P>
462 If you need to support versions older than 8.32, or versions that may
463 not build with --enable-jit, you must either use the slower
464 <b>pcre_exec()</b>, or switch between the two codepaths by checking the
465 values of PCRE_MAJOR and PCRE_MINOR.
466 </P>
467 <P>
468 Switching between the two by checking the version assumes that all the
469 versions being targeted are built with --enable-jit. To also support
470 builds that may use --disable-jit either <b>pcre_exec()</b> must be
471 used, or a compile-time check for JIT via <b>pcre_config()</b> (which
472 assumes the runtime environment will be the same), or as the Git
473 project decided to do, simply assume that <b>pcre_jit_exec()</b> is
474 present in 8.32 or later unless a compile-time flag is provided, see
475 the "grep: un-break building with PCRE &#62;= 8.32 without --enable-jit"
476 commit in git.git for an example of that.
477 </P>
478 <br><a name="SEC12" href="#TOC1">SEE ALSO</a><br>
479 <P>
480 <b>pcreapi</b>(3)
481 </P>
482 <br><a name="SEC13" href="#TOC1">AUTHOR</a><br>
483 <P>
484 Philip Hazel (FAQ by Zoltan Herczeg)
485 <br>
486 University Computing Service
487 <br>
488 Cambridge CB2 3QH, England.
489 <br>
490 </P>
491 <br><a name="SEC14" href="#TOC1">REVISION</a><br>
492 <P>
493 Last updated: 05 July 2017
494 <br>
495 Copyright &copy; 1997-2017 University of Cambridge.
496 <br>
497 <p>
498 Return to the <a href="index.html">PCRE index page</a>.
499 </p>

  ViewVC Help
Powered by ViewVC 1.1.5