ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 691 - (show annotations)
Sun Sep 11 14:31:21 2011 UTC (5 years, 4 months ago) by ph10
File MIME type: text/html
File size: 11386 byte(s)
Final source and document tidies for 8.20-RC1.
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">AVAILABILITY OF JIT SUPPORT</a>
18 <li><a name="TOC3" href="#SEC3">SIMPLE USE OF JIT</a>
20 <li><a name="TOC5" href="#SEC5">RETURN VALUES FROM JIT EXECUTION</a>
22 <li><a name="TOC7" href="#SEC7">CONTROLLING THE JIT STACK</a>
23 <li><a name="TOC8" href="#SEC8">EXAMPLE CODE</a>
24 <li><a name="TOC9" href="#SEC9">SEE ALSO</a>
25 <li><a name="TOC10" href="#SEC10">AUTHOR</a>
26 <li><a name="TOC11" href="#SEC11">REVISION</a>
27 </ul>
28 <br><a name="SEC1" href="#TOC1">PCRE JUST-IN-TIME COMPILER SUPPORT</a><br>
29 <P>
30 Just-in-time compiling is a heavyweight optimization that can greatly speed up
31 pattern matching. However, it comes at the cost of extra processing before the
32 match is performed. Therefore, it is of most benefit when the same pattern is
33 going to be matched many times. This does not necessarily mean many calls of
34 \fPpcre_exec()\fP; if the pattern is not anchored, matching attempts may take
35 place many times at various positions in the subject, even for a single call to
36 <b>pcre_exec()</b>. If the subject string is very long, it may still pay to use
37 JIT for one-off matches.
38 </P>
39 <P>
40 JIT support applies only to the traditional matching function,
41 <b>pcre_exec()</b>. It does not apply when <b>pcre_dfa_exec()</b> is being used.
42 The code for this support was written by Zoltan Herczeg.
43 </P>
44 <br><a name="SEC2" href="#TOC1">AVAILABILITY OF JIT SUPPORT</a><br>
45 <P>
46 JIT support is an optional feature of PCRE. The "configure" option --enable-jit
47 (or equivalent CMake option) must be set when PCRE is built if you want to use
48 JIT. The support is limited to the following hardware platforms:
49 <pre>
50 ARM v5, v7, and Thumb2
51 Intel x86 32-bit and 64-bit
52 MIPS 32-bit
53 Power PC 32-bit and 64-bit
54 </pre>
55 If --enable-jit is set on an unsupported platform, compilation fails.
56 </P>
57 <P>
58 A program can tell if JIT support is available by calling <b>pcre_config()</b>
59 with the PCRE_CONFIG_JIT option. The result is 1 when JIT is available, and 0
60 otherwise. However, a simple program does not need to check this in order to
61 use JIT. The API is implemented in a way that falls back to the ordinary PCRE
62 code if JIT is not available.
63 </P>
64 <br><a name="SEC3" href="#TOC1">SIMPLE USE OF JIT</a><br>
65 <P>
66 You have to do two things to make use of the JIT support in the simplest way:
67 <pre>
68 (1) Call <b>pcre_study()</b> with the PCRE_STUDY_JIT_COMPILE option for
69 each compiled pattern, and pass the resulting <b>pcre_extra</b> block to
70 <b>pcre_exec()</b>.
72 (2) Use <b>pcre_free_study()</b> to free the <b>pcre_extra</b> block when it is
73 no longer needed instead of just freeing it yourself. This
74 ensures that any JIT data is also freed.
75 </pre>
76 In some circumstances you may need to call additional functions. These are
77 described in the section entitled
78 <a href="#stackcontrol">"Controlling the JIT stack"</a>
79 below.
80 </P>
81 <P>
82 If JIT support is not available, PCRE_STUDY_JIT_COMPILE is ignored, and no JIT
83 data is set up. Otherwise, the compiled pattern is passed to the JIT compiler,
84 which turns it into machine code that executes much faster than the normal
85 interpretive code. When <b>pcre_exec()</b> is passed a <b>pcre_extra</b> block
86 containing a pointer to JIT code, it obeys that instead of the normal code. The
87 result is identical, but the code runs much faster.
88 </P>
89 <P>
90 There are some <b>pcre_exec()</b> options that are not supported for JIT
91 execution. There are also some pattern items that JIT cannot handle. Details
92 are given below. In both cases, execution automatically falls back to the
93 interpretive code.
94 </P>
95 <P>
96 If the JIT compiler finds an unsupported item, no JIT data is generated. You
97 can find out if JIT execution is available after studying a pattern by calling
98 <b>pcre_fullinfo()</b> with the PCRE_INFO_JIT option. A result of 1 means that
99 JIT compilationw was successful. A result of 0 means that JIT support is not
100 available, or the pattern was not studied with PCRE_STUDY_JIT_COMPILE, or the
101 JIT compiler was not able to handle the pattern.
102 </P>
103 <br><a name="SEC4" href="#TOC1">UNSUPPORTED OPTIONS AND PATTERN ITEMS</a><br>
104 <P>
105 The only <b>pcre_exec()</b> options that are supported for JIT execution are
107 PCRE_NOTEMPTY_ATSTART. Note in particular that partial matching is not
108 supported.
109 </P>
110 <P>
111 The unsupported pattern items are:
112 <pre>
113 \C match a single byte, even in UTF-8 mode
114 (?Cn) callouts
115 (?(&#60;name&#62;)... conditional test on setting of a named subpattern
116 (?(R)... conditional test on whole pattern recursion
117 (?(Rn)... conditional test on recursion, by number
118 (?(R&name)... conditional test on recursion, by name
119 (*COMMIT) )
120 (*MARK) )
121 (*PRUNE) ) the backtracking control verbs
122 (*SKIP) )
123 (*THEN) )
124 </pre>
125 Support for some of these may be added in future.
126 </P>
127 <br><a name="SEC5" href="#TOC1">RETURN VALUES FROM JIT EXECUTION</a><br>
128 <P>
129 When a pattern is matched using JIT execution, the return values are the same
130 as those given by the interpretive <b>pcre_exec()</b> code, with the addition of
131 one new error code: PCRE_ERROR_JIT_STACKLIMIT. This means that the memory used
132 for the JIT stack was insufficient. See
133 <a href="#stackcontrol">"Controlling the JIT stack"</a>
134 below for a discussion of JIT stack usage. For compatibility with the
135 interpretive <b>pcre_exec()</b> code, no more than two-thirds of the
136 <i>ovector</i> argument is used for passing back captured substrings.
137 </P>
138 <P>
139 The error code PCRE_ERROR_MATCHLIMIT is returned by the JIT code if searching a
140 very large pattern tree goes on for too long, as it is in the same circumstance
141 when JIT is not used, but the details of exactly what is counted are not the
142 same. The PCRE_ERROR_RECURSIONLIMIT error code is never returned by JIT
143 execution.
144 </P>
145 <br><a name="SEC6" href="#TOC1">SAVING AND RESTORING COMPILED PATTERNS</a><br>
146 <P>
147 The code that is generated by the JIT compiler is architecture-specific, and is
148 also position dependent. For those reasons it cannot be saved and restored like
149 the bytecode and other data of a compiled pattern. You should be able run
150 <b>pcre_study()</b> on a saved and restored pattern, and thereby recreate the
151 JIT data, but because JIT compilation uses significant resources, it is
152 probably not worth doing this.
153 <a name="stackcontrol"></a></P>
154 <br><a name="SEC7" href="#TOC1">CONTROLLING THE JIT STACK</a><br>
155 <P>
156 When the compiled JIT code runs, it needs a block of memory to use as a stack.
157 By default, it uses 32K on the machine stack. However, some large or
158 complicated patterns need more than this. The error PCRE_ERROR_JIT_STACKLIMIT
159 is given when there is not enough stack. Three functions are provided for
160 managing blocks of memory for use as JIT stacks.
161 </P>
162 <P>
163 The <b>pcre_jit_stack_alloc()</b> function creates a JIT stack. Its arguments
164 are a starting size and a maximum size, and it returns a pointer to an opaque
165 structure of type <b>pcre_jit_stack</b>, or NULL if there is an error. The
166 <b>pcre_jit_stack_free()</b> function can be used to free a stack that is no
167 longer needed. (For the technically minded: the address space is allocated by
168 mmap or VirtualAlloc.)
169 </P>
170 <P>
171 JIT uses far less memory for recursion than the interpretive code,
172 and a maximum stack size of 512K to 1M should be more than enough for any
173 pattern.
174 </P>
175 <P>
176 The <b>pcre_assign_jit_stack()</b> function specifies which stack JIT code
177 should use. Its arguments are as follows:
178 <pre>
179 pcre_extra *extra
180 pcre_jit_callback callback
181 void *data
182 </pre>
183 The <i>extra</i> argument must be the result of studying a pattern with
184 PCRE_STUDY_JIT_COMPILE. There are three cases for the values of the other two
185 options:
186 <pre>
187 (1) If <i>callback</i> is NULL and <i>data</i> is NULL, an internal 32K block
188 on the machine stack is used.
190 (2) If <i>callback</i> is NULL and <i>data</i> is not NULL, <i>data</i> must be
191 a valid JIT stack, the result of calling <b>pcre_jit_stack_alloc()</b>.
193 (3) If <i>callback</i> not NULL, it must point to a function that is called
194 with <i>data</i> as an argument at the start of matching, in order to
195 set up a JIT stack. If the result is NULL, the internal 32K stack
196 is used; otherwise the return value must be a valid JIT stack,
197 the result of calling <b>pcre_jit_stack_alloc()</b>.
198 </pre>
199 You may safely assign the same JIT stack to more than one pattern, as long as
200 they are all matched sequentially in the same thread. In a multithread
201 application, each thread must use its own JIT stack.
202 </P>
203 <P>
204 Strictly speaking, even more is allowed. You can assign the same stack to any
205 number of patterns as long as they are not used for matching by multiple
206 threads at the same time. For example, you can assign the same stack to all
207 compiled patterns, and use a global mutex in the callback to wait until the
208 stack is available for use. However, this is an inefficient solution, and
209 not recommended.
210 </P>
211 <P>
212 This is a suggestion for how a typical multithreaded program might operate:
213 <pre>
214 During thread initalization
215 thread_local_var = pcre_jit_stack_alloc(...)
217 During thread exit
218 pcre_jit_stack_free(thread_local_var)
220 Use a one-line callback function
221 return thread_local_var
222 </pre>
223 All the functions described in this section do nothing if JIT is not available,
224 and <b>pcre_assign_jit_stack()</b> does nothing unless the <b>extra</b> argument
225 is non-NULL and points to a <b>pcre_extra</b> block that is the result of a
226 successful study with PCRE_STUDY_JIT_COMPILE.
227 </P>
228 <br><a name="SEC8" href="#TOC1">EXAMPLE CODE</a><br>
229 <P>
230 This is a single-threaded example that specifies a JIT stack without using a
231 callback.
232 <pre>
233 int rc;
234 int ovector[30];
235 pcre *re;
236 pcre_extra *extra;
237 pcre_jit_stack *jit_stack;
239 re = pcre_compile(pattern, 0, &error, &erroffset, NULL);
240 /* Check for errors */
241 extra = pcre_study(re, PCRE_STUDY_JIT_COMPILE, &error);
242 jit_stack = pcre_jit_stack_alloc(32*1024, 512*1024);
243 /* Check for error (NULL) */
244 pcre_assign_jit_stack(extra, NULL, jit_stack);
245 rc = pcre_exec(re, extra, subject, length, 0, 0, ovector, 30);
246 /* Check results */
247 pcre_free(re);
248 pcre_free_study(extra);
249 pcre_jit_stack_free(jit_stack);
251 </PRE>
252 </P>
253 <br><a name="SEC9" href="#TOC1">SEE ALSO</a><br>
254 <P>
255 <b>pcreapi</b>(3)
256 </P>
257 <br><a name="SEC10" href="#TOC1">AUTHOR</a><br>
258 <P>
259 Philip Hazel
260 <br>
261 University Computing Service
262 <br>
263 Cambridge CB2 3QH, England.
264 <br>
265 </P>
266 <br><a name="SEC11" href="#TOC1">REVISION</a><br>
267 <P>
268 Last updated: 06 September 2011
269 <br>
270 Copyright &copy; 1997-2011 University of Cambridge.
271 <br>
272 <p>
273 Return to the <a href="index.html">PCRE index page</a>.
274 </p>

  ViewVC Help
Powered by ViewVC 1.1.5