/[pcre]/code/trunk/doc/pcrejit.3
ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 683 by ph10, Tue Sep 6 10:37:15 2011 UTC revision 836 by ph10, Wed Dec 28 17:16:11 2011 UTC
# Line 28  JIT. The support is limited to the follo Line 28  JIT. The support is limited to the follo
28    ARM v5, v7, and Thumb2    ARM v5, v7, and Thumb2
29    Intel x86 32-bit and 64-bit    Intel x86 32-bit and 64-bit
30    MIPS 32-bit    MIPS 32-bit
31    Power PC 32-bit and 64-bit    Power PC 32-bit and 64-bit (experimental)
32  .sp  .sp
33  If --enable-jit is set on an unsupported platform, compilation fails.  The Power PC support is designated as experimental because it has not been
34  .P  fully tested. If --enable-jit is set on an unsupported platform, compilation
35  A program can tell if JIT support is available by calling \fBpcre_config()\fP  fails.
36  with the PCRE_CONFIG_JIT option. The result is 1 when JIT is available, and 0  .P
37  otherwise. However, a simple program does not need to check this in order to  A program that is linked with PCRE 8.20 or later can tell if JIT support is
38  use JIT. The API is implemented in a way that falls back to the ordinary PCRE  available by calling \fBpcre_config()\fP with the PCRE_CONFIG_JIT option. The
39  code if JIT is not available.  result is 1 when JIT is available, and 0 otherwise. However, a simple program
40    does not need to check this in order to use JIT. The API is implemented in a
41    way that falls back to the ordinary PCRE code if JIT is not available.
42    .P
43    If your program may sometimes be linked with versions of PCRE that are older
44    than 8.20, but you want to use JIT when it is available, you can test
45    the values of PCRE_MAJOR and PCRE_MINOR, or the existence of a JIT macro such
46    as PCRE_CONFIG_JIT, for compile-time control of your code.
47  .  .
48  .  .
49  .SH "SIMPLE USE OF JIT"  .SH "SIMPLE USE OF JIT"
# Line 47  You have to do two things to make use of Line 54  You have to do two things to make use of
54    (1) Call \fBpcre_study()\fP with the PCRE_STUDY_JIT_COMPILE option for    (1) Call \fBpcre_study()\fP with the PCRE_STUDY_JIT_COMPILE option for
55        each compiled pattern, and pass the resulting \fBpcre_extra\fP block to        each compiled pattern, and pass the resulting \fBpcre_extra\fP block to
56        \fBpcre_exec()\fP.        \fBpcre_exec()\fP.
57    .sp
58    (2) Use \fBpcre_free_study()\fP to free the \fBpcre_extra\fP block when it is    (2) Use \fBpcre_free_study()\fP to free the \fBpcre_extra\fP block when it is
59        no longer needed instead of just freeing it yourself. This        no longer needed instead of just freeing it yourself. This
60        ensures that any JIT data is also freed.        ensures that any JIT data is also freed.
61  .sp  .sp
62    For a program that may be linked with pre-8.20 versions of PCRE, you can insert
63    .sp
64      #ifndef PCRE_STUDY_JIT_COMPILE
65      #define PCRE_STUDY_JIT_COMPILE 0
66      #endif
67    .sp
68    so that no option is passed to \fBpcre_study()\fP, and then use something like
69    this to free the study data:
70    .sp
71      #ifdef PCRE_CONFIG_JIT
72          pcre_free_study(study_ptr);
73      #else
74          pcre_free(study_ptr);
75      #endif
76    .sp
77  In some circumstances you may need to call additional functions. These are  In some circumstances you may need to call additional functions. These are
78  described in the section entitled  described in the section entitled
79  .\" HTML <a href="#stackcontrol">  .\" HTML <a href="#stackcontrol">
# Line 75  interpretive code. Line 97  interpretive code.
97  If the JIT compiler finds an unsupported item, no JIT data is generated. You  If the JIT compiler finds an unsupported item, no JIT data is generated. You
98  can find out if JIT execution is available after studying a pattern by calling  can find out if JIT execution is available after studying a pattern by calling
99  \fBpcre_fullinfo()\fP with the PCRE_INFO_JIT option. A result of 1 means that  \fBpcre_fullinfo()\fP with the PCRE_INFO_JIT option. A result of 1 means that
100  JIT compilationw was successful. A result of 0 means that JIT support is not  JIT compilation was successful. A result of 0 means that JIT support is not
101  available, or the pattern was not studied with PCRE_STUDY_JIT_COMPILE, or the  available, or the pattern was not studied with PCRE_STUDY_JIT_COMPILE, or the
102  JIT compiler was not able to handle the pattern.  JIT compiler was not able to handle the pattern.
103    .P
104    Once a pattern has been studied, with or without JIT, it can be used as many
105    times as you like for matching different subject strings.
106  .  .
107  .  .
108  .SH "UNSUPPORTED OPTIONS AND PATTERN ITEMS"  .SH "UNSUPPORTED OPTIONS AND PATTERN ITEMS"
# Line 90  supported. Line 115  supported.
115  .P  .P
116  The unsupported pattern items are:  The unsupported pattern items are:
117  .sp  .sp
118    \eC            match a single byte, even in UTF-8 mode    \eC             match a single byte; not supported in UTF-8 mode
119    (?Cn)          callouts    (?Cn)          callouts
   (?(<name>)...  conditional test on setting of a named subpattern  
   (?(R)...       conditional test on whole pattern recursion  
   (?(Rn)...      conditional test on recursion, by number  
   (?(R&name)...  conditional test on recursion, by name  
120    (*COMMIT)      )    (*COMMIT)      )
121    (*MARK)        )    (*MARK)        )
122    (*PRUNE)       ) the backtracking control verbs    (*PRUNE)       ) the backtracking control verbs
# Line 131  execution. Line 152  execution.
152  .rs  .rs
153  .sp  .sp
154  The code that is generated by the JIT compiler is architecture-specific, and is  The code that is generated by the JIT compiler is architecture-specific, and is
155  also position dependent. For those reasons it cannot be saved and restored like  also position dependent. For those reasons it cannot be saved (in a file or
156  the bytecode and other data of a compiled pattern. You should be able run  database) and restored later like the bytecode and other data of a compiled
157  \fBpcre_study()\fP on a saved and restored pattern, and thereby recreate the  pattern. Saving and restoring compiled patterns is not something many people
158  JIT data, but because JIT compilation uses significant resources, it is  do. More detail about this facility is given in the
159  probably not worth doing this.  .\" HREF
160    \fBpcreprecompile\fP
161    .\"
162    documentation. It should be possible to run \fBpcre_study()\fP on a saved and
163    restored pattern, and thereby recreate the JIT data, but because JIT
164    compilation uses significant resources, it is probably not worth doing this;
165    you might as well recompile the original pattern.
166  .  .
167  .  .
168  .\" HTML <a name="stackcontrol"></a>  .\" HTML <a name="stackcontrol"></a>
# Line 146  When the compiled JIT code runs, it need Line 173  When the compiled JIT code runs, it need
173  By default, it uses 32K on the machine stack. However, some large or  By default, it uses 32K on the machine stack. However, some large or
174  complicated patterns need more than this. The error PCRE_ERROR_JIT_STACKLIMIT  complicated patterns need more than this. The error PCRE_ERROR_JIT_STACKLIMIT
175  is given when there is not enough stack. Three functions are provided for  is given when there is not enough stack. Three functions are provided for
176  managing blocks of memory for use as JIT stacks.  managing blocks of memory for use as JIT stacks. There is further discussion
177    about the use of JIT stacks in the section entitled
178    .\" HTML <a href="#stackcontrol">
179    .\" </a>
180    "JIT stack FAQ"
181    .\"
182    below.
183  .P  .P
184  The \fBpcre_jit_stack_alloc()\fP function creates a JIT stack. Its arguments  The \fBpcre_jit_stack_alloc()\fP function creates a JIT stack. Its arguments
185  are a starting size and a maximum size, and it returns an opaque value  are a starting size and a maximum size, and it returns a pointer to an opaque
186  of type \fBpcre_jit_stack\fP that represents a JIT stack, or NULL if there is  structure of type \fBpcre_jit_stack\fP, or NULL if there is an error. The
187  an error. The \fBpcre_jit_stack_free()\fP function can be used to free a stack  \fBpcre_jit_stack_free()\fP function can be used to free a stack that is no
188  that is no longer needed. (For the technically minded: the address space is  longer needed. (For the technically minded: the address space is allocated by
189  allocated by mmap or VirtualAlloc.)  mmap or VirtualAlloc.)
190  .P  .P
191  JIT uses far less memory for recursion than the interpretive code,  JIT uses far less memory for recursion than the interpretive code,
192  and a maximum stack size of 512K to 1M should be more than enough for any  and a maximum stack size of 512K to 1M should be more than enough for any
193  pattern.  pattern.
194  .P  .P
# Line 197  This is a suggestion for how a typical m Line 230  This is a suggestion for how a typical m
230  .sp  .sp
231    During thread initalization    During thread initalization
232      thread_local_var = pcre_jit_stack_alloc(...)      thread_local_var = pcre_jit_stack_alloc(...)
233    .sp
234    During thread exit    During thread exit
235      pcre_jit_stack_free(thread_local_var)      pcre_jit_stack_free(thread_local_var)
236    .sp
237    Use a one-line callback function    Use a one-line callback function
238      return thread_local_var      return thread_local_var
239  .sp  .sp
# Line 210  is non-NULL and points to a \fBpcre_extr Line 243  is non-NULL and points to a \fBpcre_extr
243  successful study with PCRE_STUDY_JIT_COMPILE.  successful study with PCRE_STUDY_JIT_COMPILE.
244  .  .
245  .  .
246    .\" HTML <a name="stackfaq"></a>
247    .SH "JIT STACK FAQ"
248    .rs
249    .sp
250    (1) Why do we need JIT stacks?
251    .sp
252    PCRE (and JIT) is a recursive, depth-first engine, so it needs a stack where
253    the local data of the current node is pushed before checking its child nodes.
254    Allocating real machine stack on some platforms is difficult. For example, the
255    stack chain needs to be updated every time if we extend the stack on PowerPC.
256    Although it is possible, its updating time overhead decreases performance. So
257    we do the recursion in memory.
258    .P
259    (2) Why don't we simply allocate blocks of memory with \fBmalloc()\fP?
260    .sp
261    Modern operating systems have a nice feature: they can reserve an address space
262    instead of allocating memory. We can safely allocate memory pages inside this
263    address space, so the stack could grow without moving memory data (this is
264    important because of pointers). Thus we can allocate 1M address space, and use
265    only a single memory page (usually 4K) if that is enough. However, we can still
266    grow up to 1M anytime if needed.
267    .P
268    (3) Who "owns" a JIT stack?
269    .sp
270    The owner of the stack is the user program, not the JIT studied pattern or
271    anything else. The user program must ensure that if a stack is used by
272    \fBpcre_exec()\fP, (that is, it is assigned to the pattern currently running),
273    that stack must not be used by any other threads (to avoid overwriting the same
274    memory area). The best practice for multithreaded programs is to allocate a
275    stack for each thread, and return this stack through the JIT callback function.
276    .P
277    (4) When should a JIT stack be freed?
278    .sp
279    You can free a JIT stack at any time, as long as it will not be used by
280    \fBpcre_exec()\fP again. When you assign the stack to a pattern, only a pointer
281    is set. There is no reference counting or any other magic. You can free the
282    patterns and stacks in any order, anytime. Just \fIdo not\fP call
283    \fBpcre_exec()\fP with a pattern pointing to an already freed stack, as that
284    will cause SEGFAULT. (Also, do not free a stack currently used by
285    \fBpcre_exec()\fP in another thread). You can also replace the stack for a
286    pattern at any time. You can even free the previous stack before assigning a
287    replacement.
288    .P
289    (5) Should I allocate/free a stack every time before/after calling
290    \fBpcre_exec()\fP?
291    .sp
292    No, because this is too costly in terms of resources. However, you could
293    implement some clever idea which release the stack if it is not used in let's
294    say two minutes. The JIT callback can help to achive this without keeping a
295    list of the currently JIT studied patterns.
296    .P
297    (6) OK, the stack is for long term memory allocation. But what happens if a
298    pattern causes stack overflow with a stack of 1M? Is that 1M kept until the
299    stack is freed?
300    .sp
301    Especially on embedded sytems, it might be a good idea to release
302    memory sometimes without freeing the stack. There is no API for this at the
303    moment. Probably a function call which returns with the currently allocated
304    memory for any stack and another which allows releasing memory (shrinking the
305    stack) would be a good idea if someone needs this.
306    .P
307    (7) This is too much of a headache. Isn't there any better solution for JIT
308    stack handling?
309    .sp
310    No, thanks to Windows. If POSIX threads were used everywhere, we could throw
311    out this complicated API.
312    .
313    .
314  .SH "EXAMPLE CODE"  .SH "EXAMPLE CODE"
315  .rs  .rs
316  .sp  .sp
317  This is a single-threaded example that specifies a JIT stack without using a  This is a single-threaded example that specifies a JIT stack without using a
318  callback.  callback.
319  .sp  .sp
320    int rc;    int rc;
321    int ovector[30];    int ovector[30];
# Line 232  callback. Line 333  callback.
333    /* Check results */    /* Check results */
334    pcre_free(re);    pcre_free(re);
335    pcre_free_study(extra);    pcre_free_study(extra);
336      pcre_jit_stack_free(jit_stack);
337  .sp  .sp
338  .  .
339  .  .
# Line 245  callback. Line 347  callback.
347  .rs  .rs
348  .sp  .sp
349  .nf  .nf
350  Philip Hazel  Philip Hazel (FAQ by Zoltan Herczeg)
351  University Computing Service  University Computing Service
352  Cambridge CB2 3QH, England.  Cambridge CB2 3QH, England.
353  .fi  .fi
# Line 255  Cambridge CB2 3QH, England. Line 357  Cambridge CB2 3QH, England.
357  .rs  .rs
358  .sp  .sp
359  .nf  .nf
360  Last updated: 06 September 2011  Last updated: 26 November 2011
361  Copyright (c) 1997-2011 University of Cambridge.  Copyright (c) 1997-2011 University of Cambridge.
362  .fi  .fi

Legend:
Removed from v.683  
changed lines
  Added in v.836

  ViewVC Help
Powered by ViewVC 1.1.5