ViewVC logotype

Contents of /code/trunk/doc/pcrejit.3

Parent Directory Parent Directory | Revision Log Revision Log

Revision 686 - (show annotations)
Tue Sep 6 16:06:36 2011 UTC (10 years, 1 month ago) by ph10
File size: 9827 byte(s)
More documentation updates.
3 PCRE - Perl-compatible regular expressions
5 .rs
6 .sp
7 Just-in-time compiling is a heavyweight optimization that can greatly speed up
8 pattern matching. However, it comes at the cost of extra processing before the
9 match is performed. Therefore, it is of most benefit when the same pattern is
10 going to be matched many times. This does not necessarily mean many calls of
11 \fPpcre_exec()\fP; if the pattern is not anchored, matching attempts may take
12 place many times at various positions in the subject, even for a single call to
13 \fBpcre_exec()\fP. If the subject string is very long, it may still pay to use
14 JIT for one-off matches.
15 .P
16 JIT support applies only to the traditional matching function,
17 \fBpcre_exec()\fP. It does not apply when \fBpcre_dfa_exec()\fP is being used.
18 The code for this support was written by Zoltan Herczeg.
19 .
20 .
22 .rs
23 .sp
24 JIT support is an optional feature of PCRE. The "configure" option --enable-jit
25 (or equivalent CMake option) must be set when PCRE is built if you want to use
26 JIT. The support is limited to the following hardware platforms:
27 .sp
28 ARM v5, v7, and Thumb2
29 Intel x86 32-bit and 64-bit
30 MIPS 32-bit
31 Power PC 32-bit and 64-bit
32 .sp
33 If --enable-jit is set on an unsupported platform, compilation fails.
34 .P
35 A program can tell if JIT support is available by calling \fBpcre_config()\fP
36 with the PCRE_CONFIG_JIT option. The result is 1 when JIT is available, and 0
37 otherwise. However, a simple program does not need to check this in order to
38 use JIT. The API is implemented in a way that falls back to the ordinary PCRE
39 code if JIT is not available.
40 .
41 .
43 .rs
44 .sp
45 You have to do two things to make use of the JIT support in the simplest way:
46 .sp
47 (1) Call \fBpcre_study()\fP with the PCRE_STUDY_JIT_COMPILE option for
48 each compiled pattern, and pass the resulting \fBpcre_extra\fP block to
49 \fBpcre_exec()\fP.
51 (2) Use \fBpcre_free_study()\fP to free the \fBpcre_extra\fP block when it is
52 no longer needed instead of just freeing it yourself. This
53 ensures that any JIT data is also freed.
54 .sp
55 In some circumstances you may need to call additional functions. These are
56 described in the section entitled
57 .\" HTML <a href="#stackcontrol">
58 .\" </a>
59 "Controlling the JIT stack"
60 .\"
61 below.
62 .P
63 If JIT support is not available, PCRE_STUDY_JIT_COMPILE is ignored, and no JIT
64 data is set up. Otherwise, the compiled pattern is passed to the JIT compiler,
65 which turns it into machine code that executes much faster than the normal
66 interpretive code. When \fBpcre_exec()\fP is passed a \fBpcre_extra\fP block
67 containing a pointer to JIT code, it obeys that instead of the normal code. The
68 result is identical, but the code runs much faster.
69 .P
70 There are some \fBpcre_exec()\fP options that are not supported for JIT
71 execution. There are also some pattern items that JIT cannot handle. Details
72 are given below. In both cases, execution automatically falls back to the
73 interpretive code.
74 .P
75 If the JIT compiler finds an unsupported item, no JIT data is generated. You
76 can find out if JIT execution is available after studying a pattern by calling
77 \fBpcre_fullinfo()\fP with the PCRE_INFO_JIT option. A result of 1 means that
78 JIT compilationw was successful. A result of 0 means that JIT support is not
79 available, or the pattern was not studied with PCRE_STUDY_JIT_COMPILE, or the
80 JIT compiler was not able to handle the pattern.
81 .
82 .
84 .rs
85 .sp
86 The only \fBpcre_exec()\fP options that are supported for JIT execution are
88 PCRE_NOTEMPTY_ATSTART. Note in particular that partial matching is not
89 supported.
90 .P
91 The unsupported pattern items are:
92 .sp
93 \eC match a single byte, even in UTF-8 mode
94 (?Cn) callouts
95 (?(<name>)... conditional test on setting of a named subpattern
96 (?(R)... conditional test on whole pattern recursion
97 (?(Rn)... conditional test on recursion, by number
98 (?(R&name)... conditional test on recursion, by name
99 (*COMMIT) )
100 (*MARK) )
101 (*PRUNE) ) the backtracking control verbs
102 (*SKIP) )
103 (*THEN) )
104 .sp
105 Support for some of these may be added in future.
106 .
107 .
109 .rs
110 .sp
111 When a pattern is matched using JIT execution, the return values are the same
112 as those given by the interpretive \fBpcre_exec()\fP code, with the addition of
113 one new error code: PCRE_ERROR_JIT_STACKLIMIT. This means that the memory used
114 for the JIT stack was insufficient. See
115 .\" HTML <a href="#stackcontrol">
116 .\" </a>
117 "Controlling the JIT stack"
118 .\"
119 below for a discussion of JIT stack usage. For compatibility with the
120 interpretive \fBpcre_exec()\fP code, no more than two-thirds of the
121 \fIovector\fP argument is used for passing back captured substrings.
122 .P
123 The error code PCRE_ERROR_MATCHLIMIT is returned by the JIT code if searching a
124 very large pattern tree goes on for too long, as it is in the same circumstance
125 when JIT is not used, but the details of exactly what is counted are not the
126 same. The PCRE_ERROR_RECURSIONLIMIT error code is never returned by JIT
127 execution.
128 .
129 .
131 .rs
132 .sp
133 The code that is generated by the JIT compiler is architecture-specific, and is
134 also position dependent. For those reasons it cannot be saved and restored like
135 the bytecode and other data of a compiled pattern. You should be able run
136 \fBpcre_study()\fP on a saved and restored pattern, and thereby recreate the
137 JIT data, but because JIT compilation uses significant resources, it is
138 probably not worth doing this.
139 .
140 .
141 .\" HTML <a name="stackcontrol"></a>
143 .rs
144 .sp
145 When the compiled JIT code runs, it needs a block of memory to use as a stack.
146 By default, it uses 32K on the machine stack. However, some large or
147 complicated patterns need more than this. The error PCRE_ERROR_JIT_STACKLIMIT
148 is given when there is not enough stack. Three functions are provided for
149 managing blocks of memory for use as JIT stacks.
150 .P
151 The \fBpcre_jit_stack_alloc()\fP function creates a JIT stack. Its arguments
152 are a starting size and a maximum size, and it returns a pointer to an opaque
153 structure of type \fBpcre_jit_stack\fP, or NULL if there is an error. The
154 \fBpcre_jit_stack_free()\fP function can be used to free a stack that is no
155 longer needed. (For the technically minded: the address space is allocated by
156 mmap or VirtualAlloc.)
157 .P
158 JIT uses far less memory for recursion than the interpretive code,
159 and a maximum stack size of 512K to 1M should be more than enough for any
160 pattern.
161 .P
162 The \fBpcre_assign_jit_stack()\fP function specifies which stack JIT code
163 should use. Its arguments are as follows:
164 .sp
165 pcre_extra *extra
166 pcre_jit_callback callback
167 void *data
168 .sp
169 The \fIextra\fP argument must be the result of studying a pattern with
170 PCRE_STUDY_JIT_COMPILE. There are three cases for the values of the other two
171 options:
172 .sp
173 (1) If \fIcallback\fP is NULL and \fIdata\fP is NULL, an internal 32K block
174 on the machine stack is used.
175 .sp
176 (2) If \fIcallback\fP is NULL and \fIdata\fP is not NULL, \fIdata\fP must be
177 a valid JIT stack, the result of calling \fBpcre_jit_stack_alloc()\fP.
178 .sp
179 (3) If \fIcallback\fP not NULL, it must point to a function that is called
180 with \fIdata\fP as an argument at the start of matching, in order to
181 set up a JIT stack. If the result is NULL, the internal 32K stack
182 is used; otherwise the return value must be a valid JIT stack,
183 the result of calling \fBpcre_jit_stack_alloc()\fP.
184 .sp
185 You may safely assign the same JIT stack to more than one pattern, as long as
186 they are all matched sequentially in the same thread. In a multithread
187 application, each thread must use its own JIT stack.
188 .P
189 Strictly speaking, even more is allowed. You can assign the same stack to any
190 number of patterns as long as they are not used for matching by multiple
191 threads at the same time. For example, you can assign the same stack to all
192 compiled patterns, and use a global mutex in the callback to wait until the
193 stack is available for use. However, this is an inefficient solution, and
194 not recommended.
195 .P
196 This is a suggestion for how a typical multithreaded program might operate:
197 .sp
198 During thread initalization
199 thread_local_var = pcre_jit_stack_alloc(...)
201 During thread exit
202 pcre_jit_stack_free(thread_local_var)
204 Use a one-line callback function
205 return thread_local_var
206 .sp
207 All the functions described in this section do nothing if JIT is not available,
208 and \fBpcre_assign_jit_stack()\fP does nothing unless the \fBextra\fP argument
209 is non-NULL and points to a \fBpcre_extra\fP block that is the result of a
210 successful study with PCRE_STUDY_JIT_COMPILE.
211 .
212 .
214 .rs
215 .sp
216 This is a single-threaded example that specifies a JIT stack without using a
217 callback.
218 .sp
219 int rc;
220 int ovector[30];
221 pcre *re;
222 pcre_extra *extra;
223 pcre_jit_stack *jit_stack;
224 .sp
225 re = pcre_compile(pattern, 0, &error, &erroffset, NULL);
226 /* Check for errors */
227 extra = pcre_study(re, PCRE_STUDY_JIT_COMPILE, &error);
228 jit_stack = pcre_jit_stack_alloc(32*1024, 512*1024);
229 /* Check for error (NULL) */
230 pcre_assign_jit_stack(extra, NULL, jit_stack);
231 rc = pcre_exec(re, extra, subject, length, 0, 0, ovector, 30);
232 /* Check results */
233 pcre_free(re);
234 pcre_free_study(extra);
235 pcre_jit_stack_free(jit_stack);
236 .sp
237 .
238 .
239 .SH "SEE ALSO"
240 .rs
241 .sp
242 \fBpcreapi\fP(3)
243 .
244 .
246 .rs
247 .sp
248 .nf
249 Philip Hazel
250 University Computing Service
251 Cambridge CB2 3QH, England.
252 .fi
253 .
254 .
256 .rs
257 .sp
258 .nf
259 Last updated: 06 September 2011
260 Copyright (c) 1997-2011 University of Cambridge.
261 .fi

  ViewVC Help
Powered by ViewVC 1.1.5