|
|
<html><head><title>pcre2jit specification</title></head><body bgcolor="#FFFFFF" text="#00005A" link="#0066FF" alink="#3399FF" vlink="#2222BB"><h1>pcre2jit man page</h1><p>Return to the <a href="index.html">PCRE2 index page</a>.</p><p>This page is part of the PCRE2 HTML documentation. It was generatedautomatically from the original man page. If there is any nonsense in it,please consult the man page, in case the conversion went wrong.<br><ul><li><a name="TOC1" href="#SEC1">PCRE2 JUST-IN-TIME COMPILER SUPPORT</a><li><a name="TOC2" href="#SEC2">AVAILABILITY OF JIT SUPPORT</a><li><a name="TOC3" href="#SEC3">SIMPLE USE OF JIT</a><li><a name="TOC4" href="#SEC4">MATCHING SUBJECTS CONTAINING INVALID UTF</a><li><a name="TOC5" href="#SEC5">UNSUPPORTED OPTIONS AND PATTERN ITEMS</a><li><a name="TOC6" href="#SEC6">RETURN VALUES FROM JIT MATCHING</a><li><a name="TOC7" href="#SEC7">CONTROLLING THE JIT STACK</a><li><a name="TOC8" href="#SEC8">JIT STACK FAQ</a><li><a name="TOC9" href="#SEC9">FREEING JIT SPECULATIVE MEMORY</a><li><a name="TOC10" href="#SEC10">EXAMPLE CODE</a><li><a name="TOC11" href="#SEC11">JIT FAST PATH API</a><li><a name="TOC12" href="#SEC12">SEE ALSO</a><li><a name="TOC13" href="#SEC13">AUTHOR</a><li><a name="TOC14" href="#SEC14">REVISION</a></ul><br><a name="SEC1" href="#TOC1">PCRE2 JUST-IN-TIME COMPILER SUPPORT</a><br><P>Just-in-time compiling is a heavyweight optimization that can greatly speed uppattern matching. However, it comes at the cost of extra processing before thematch is performed, so it is of most benefit when the same pattern is going tobe matched many times. This does not necessarily mean many calls of a matchingfunction; if the pattern is not anchored, matching attempts may take place manytimes at various positions in the subject, even for a single call. Therefore,if the subject string is very long, it may still pay to use JIT even forone-off matches. JIT support is available for all of the 8-bit, 16-bit and32-bit PCRE2 libraries.</P><P>JIT support applies only to the traditional Perl-compatible matching function.It does not apply when the DFA matching function is being used. The code forJIT support was written by Zoltan Herczeg.</P><br><a name="SEC2" href="#TOC1">AVAILABILITY OF JIT SUPPORT</a><br><P>JIT support is an optional feature of PCRE2. The "configure" option--enable-jit (or equivalent CMake option) must be set when PCRE2 is built ifyou want to use JIT. The support is limited to the following hardwareplatforms:<pre> ARM 32-bit (v7, and Thumb2) ARM 64-bit IBM s390x 64 bit Intel x86 32-bit and 64-bit LoongArch 64 bit MIPS 32-bit and 64-bit Power PC 32-bit and 64-bit RISC-V 32-bit and 64-bit</pre>If --enable-jit is set on an unsupported platform, compilation fails.</P><P>A client program can tell if JIT support is available by calling<b>pcre2_config()</b> with the PCRE2_CONFIG_JIT option. The result is one ifPCRE2 was built with JIT support, and zero otherwise. However, having the JITcode available does not guarantee that it will be used for any particularmatch. One reason for this is that there are a number of options and patternitems that are<a href="#unsupported">not supported by JIT</a>(see below). Another reason is that in some environments JIT is unable to getmemory in which to build its compiled code. The only guarantee from<b>pcre2_config()</b> is that if it returns zero, JIT will definitely <i>not</i>be used.</P><P>A simple program does not need to check availability in order to use JIT whenpossible. The API is implemented in a way that falls back to the interpretivecode if JIT is not available or cannot be used for a given match. For programsthat need the best possible performance, there is a<a href="#fastpath">"fast path"</a>API that is JIT-specific.</P><br><a name="SEC3" href="#TOC1">SIMPLE USE OF JIT</a><br><P>To make use of the JIT support in the simplest way, all you have to do is tocall <b>pcre2_jit_compile()</b> after successfully compiling a pattern with<b>pcre2_compile()</b>. This function has two arguments: the first is thecompiled pattern pointer that was returned by <b>pcre2_compile()</b>, and thesecond is zero or more of the following option bits: PCRE2_JIT_COMPLETE,PCRE2_JIT_PARTIAL_HARD, or PCRE2_JIT_PARTIAL_SOFT.</P><P>If JIT support is not available, a call to <b>pcre2_jit_compile()</b> doesnothing and returns PCRE2_ERROR_JIT_BADOPTION. Otherwise, the compiled patternis passed to the JIT compiler, which turns it into machine code that executesmuch faster than the normal interpretive code, but yields exactly the sameresults. The returned value from <b>pcre2_jit_compile()</b> is zero on success,or a negative error code.</P><P>There is a limit to the size of pattern that JIT supports, imposed by the sizeof machine stack that it uses. The exact rules are not documented because theymay change at any time, in particular, when new optimizations are introduced.If a pattern is too big, a call to <b>pcre2_jit_compile()</b> returnsPCRE2_ERROR_NOMEMORY.</P><P>PCRE2_JIT_COMPLETE requests the JIT compiler to generate code for completematches. If you want to run partial matches using the PCRE2_PARTIAL_HARD orPCRE2_PARTIAL_SOFT options of <b>pcre2_match()</b>, you should set one or bothof the other options as well as, or instead of PCRE2_JIT_COMPLETE. The JITcompiler generates different optimized code for each of the three modes(normal, soft partial, hard partial). When <b>pcre2_match()</b> is called, theappropriate code is run if it is available. Otherwise, the pattern is matchedusing interpretive code.</P><P>You can call <b>pcre2_jit_compile()</b> multiple times for the same compiledpattern. It does nothing if it has previously compiled code for any of theoption bits. For example, you can call it once with PCRE2_JIT_COMPLETE and(perhaps later, when you find you need partial matching) again withPCRE2_JIT_COMPLETE and PCRE2_JIT_PARTIAL_HARD. This time it will ignorePCRE2_JIT_COMPLETE and just compile code for partial matching. If<b>pcre2_jit_compile()</b> is called with no option bits set, it immediatelyreturns zero. This is an alternative way of testing whether JIT is available.</P><P>At present, it is not possible to free JIT compiled code except when the entirecompiled pattern is freed by calling <b>pcre2_code_free()</b>.</P><P>In some circumstances you may need to call additional functions. These aredescribed in the section entitled<a href="#stackcontrol">"Controlling the JIT stack"</a>below.</P><P>There are some <b>pcre2_match()</b> options that are not supported by JIT, andthere are also some pattern items that JIT cannot handle. Details are given<a href="#unsupported">below.</a>In both cases, matching automatically falls back to the interpretive code. Ifyou want to know whether JIT was actually used for a particular match, youshould arrange for a JIT callback function to be set up as described in thesection entitled<a href="#stackcontrol">"Controlling the JIT stack"</a>below, even if you do not need to supply a non-default JIT stack. Such acallback function is called whenever JIT code is about to be obeyed. If thematch-time options are not right for JIT execution, the callback function isnot obeyed.</P><P>If the JIT compiler finds an unsupported item, no JIT data is generated. Youcan find out if JIT compilation was successful for a compiled pattern bycalling <b>pcre2_pattern_info()</b> with the PCRE2_INFO_JITSIZE option. Anon-zero result means that JIT compilation was successful. A result of 0 meansthat JIT support is not available, or the pattern was not processed by<b>pcre2_jit_compile()</b>, or the JIT compiler was not able to handle thepattern. Successful JIT compilation does not, however, guarantee the use of JITat match time because there are some match time options that are not supportedby JIT.</P><br><a name="SEC4" href="#TOC1">MATCHING SUBJECTS CONTAINING INVALID UTF</a><br><P>When a pattern is compiled with the PCRE2_UTF option, subject strings arenormally expected to be a valid sequence of UTF code units. By default, this ischecked at the start of matching and an error is generated if invalid UTF isdetected. The PCRE2_NO_UTF_CHECK option can be passed to <b>pcre2_match()</b> toskip the check (for improved performance) if you are sure that a subject stringis valid. If this option is used with an invalid string, the result isundefined. The calling program may crash or loop or otherwise misbehave.</P><P>However, a way of running matches on strings that may contain invalid UTFsequences is available. Calling <b>pcre2_compile()</b> with thePCRE2_MATCH_INVALID_UTF option has two effects: it tells the interpreter in<b>pcre2_match()</b> to support invalid UTF, and, if <b>pcre2_jit_compile()</b>is subsequently called, the compiled JIT code also supports invalid UTF.Details of how this support works, in both the JIT and the interpretive cases,is given in the<a href="pcre2unicode.html"><b>pcre2unicode</b></a>documentation.</P><P>There is also an obsolete option for <b>pcre2_jit_compile()</b> calledPCRE2_JIT_INVALID_UTF, which currently exists only for backward compatibility.It is superseded by the <b>pcre2_compile()</b> option PCRE2_MATCH_INVALID_UTFand should no longer be used. It may be removed in future.<a name="unsupported"></a></P><br><a name="SEC5" href="#TOC1">UNSUPPORTED OPTIONS AND PATTERN ITEMS</a><br><P>The <b>pcre2_match()</b> options that are supported for JIT matching arePCRE2_COPY_MATCHED_SUBJECT, PCRE2_NOTBOL, PCRE2_NOTEOL, PCRE2_NOTEMPTY,PCRE2_NOTEMPTY_ATSTART, PCRE2_NO_UTF_CHECK, PCRE2_PARTIAL_HARD, andPCRE2_PARTIAL_SOFT. The PCRE2_ANCHORED and PCRE2_ENDANCHORED options are notsupported at match time.</P><P>If the PCRE2_NO_JIT option is passed to <b>pcre2_match()</b> it disables theuse of JIT, forcing matching by the interpreter code.</P><P>The only unsupported pattern items are \C (match a single data unit) whenrunning in a UTF mode, and a callout immediately before an assertion conditionin a conditional group.</P><br><a name="SEC6" href="#TOC1">RETURN VALUES FROM JIT MATCHING</a><br><P>When a pattern is matched using JIT, the return values are the same as thosegiven by the interpretive <b>pcre2_match()</b> code, with the addition of onenew error code: PCRE2_ERROR_JIT_STACKLIMIT. This means that the memory used forthe JIT stack was insufficient. See<a href="#stackcontrol">"Controlling the JIT stack"</a>below for a discussion of JIT stack usage.</P><P>The error code PCRE2_ERROR_MATCHLIMIT is returned by the JIT code if searchinga very large pattern tree goes on for too long, as it is in the samecircumstance when JIT is not used, but the details of exactly what is countedare not the same. The PCRE2_ERROR_DEPTHLIMIT error code is never returnedwhen JIT matching is used.<a name="stackcontrol"></a></P><br><a name="SEC7" href="#TOC1">CONTROLLING THE JIT STACK</a><br><P>When the compiled JIT code runs, it needs a block of memory to use as a stack.By default, it uses 32KiB on the machine stack. However, some large orcomplicated patterns need more than this. The error PCRE2_ERROR_JIT_STACKLIMITis given when there is not enough stack. Three functions are provided formanaging blocks of memory for use as JIT stacks. There is further discussionabout the use of JIT stacks in the section entitled<a href="#stackfaq">"JIT stack FAQ"</a>below.</P><P>The <b>pcre2_jit_stack_create()</b> function creates a JIT stack. Its argumentsare a starting size, a maximum size, and a general context (for memoryallocation functions, or NULL for standard memory allocation). It returns apointer to an opaque structure of type <b>pcre2_jit_stack</b>, or NULL if thereis an error. The <b>pcre2_jit_stack_free()</b> function is used to free a stackthat is no longer needed. If its argument is NULL, this function returnsimmediately, without doing anything. (For the technically minded: the addressspace is allocated by mmap or VirtualAlloc.) A maximum stack size of 512KiB to1MiB should be more than enough for any pattern.</P><P>The <b>pcre2_jit_stack_assign()</b> function specifies which stack JIT codeshould use. Its arguments are as follows:<pre> pcre2_match_context *mcontext pcre2_jit_callback callback void *data</pre>The first argument is a pointer to a match context. When this is subsequentlypassed to a matching function, its information determines which JIT stack isused. If this argument is NULL, the function returns immediately, without doinganything. There are three cases for the values of the other two options:<pre> (1) If <i>callback</i> is NULL and <i>data</i> is NULL, an internal 32KiB block on the machine stack is used. This is the default when a match context is created.
(2) If <i>callback</i> is NULL and <i>data</i> is not NULL, <i>data</i> must be a pointer to a valid JIT stack, the result of calling <b>pcre2_jit_stack_create()</b>.
(3) If <i>callback</i> is not NULL, it must point to a function that is called with <i>data</i> as an argument at the start of matching, in order to set up a JIT stack. If the return from the callback function is NULL, the internal 32KiB stack is used; otherwise the return value must be a valid JIT stack, the result of calling <b>pcre2_jit_stack_create()</b>.</pre>A callback function is obeyed whenever JIT code is about to be run; it is notobeyed when <b>pcre2_match()</b> is called with options that are incompatiblefor JIT matching. A callback function can therefore be used to determinewhether a match operation was executed by JIT or by the interpreter.</P><P>You may safely use the same JIT stack for more than one pattern (either byassigning directly or by callback), as long as the patterns are matchedsequentially in the same thread. Currently, the only way to set upnon-sequential matches in one thread is to use callouts: if a callout functionstarts another match, that match must use a different JIT stack to the one usedfor currently suspended match(es).</P><P>In a multithread application, if you do not specify a JIT stack, or if youassign or pass back NULL from a callback, that is thread-safe, because eachthread has its own machine stack. However, if you assign or pass back anon-NULL JIT stack, this must be a different stack for each thread so that theapplication is thread-safe.</P><P>Strictly speaking, even more is allowed. You can assign the same non-NULL stackto a match context that is used by any number of patterns, as long as they arenot used for matching by multiple threads at the same time. For example, youcould use the same stack in all compiled patterns, with a global mutex in thecallback to wait until the stack is available for use. However, this is aninefficient solution, and not recommended.</P><P>This is a suggestion for how a multithreaded program that needs to set upnon-default JIT stacks might operate:<pre> During thread initialization thread_local_var = pcre2_jit_stack_create(...)
During thread exit pcre2_jit_stack_free(thread_local_var)
Use a one-line callback function return thread_local_var</pre>All the functions described in this section do nothing if JIT is not available.<a name="stackfaq"></a></P><br><a name="SEC8" href="#TOC1">JIT STACK FAQ</a><br><P>(1) Why do we need JIT stacks?<br><br>PCRE2 (and JIT) is a recursive, depth-first engine, so it needs a stack wherethe local data of the current node is pushed before checking its child nodes.Allocating real machine stack on some platforms is difficult. For example, thestack chain needs to be updated every time if we extend the stack on PowerPC.Although it is possible, its updating time overhead decreases performance. Sowe do the recursion in memory.</P><P>(2) Why don't we simply allocate blocks of memory with <b>malloc()</b>?<br><br>Modern operating systems have a nice feature: they can reserve an address spaceinstead of allocating memory. We can safely allocate memory pages inside thisaddress space, so the stack could grow without moving memory data (this isimportant because of pointers). Thus we can allocate 1MiB address space, anduse only a single memory page (usually 4KiB) if that is enough. However, we canstill grow up to 1MiB anytime if needed.</P><P>(3) Who "owns" a JIT stack?<br><br>The owner of the stack is the user program, not the JIT studied pattern oranything else. The user program must ensure that if a stack is being used by<b>pcre2_match()</b>, (that is, it is assigned to a match context that is passedto the pattern currently running), that stack must not be used by any otherthreads (to avoid overwriting the same memory area). The best practice formultithreaded programs is to allocate a stack for each thread, and return thisstack through the JIT callback function.</P><P>(4) When should a JIT stack be freed?<br><br>You can free a JIT stack at any time, as long as it will not be used by<b>pcre2_match()</b> again. When you assign the stack to a match context, only apointer is set. There is no reference counting or any other magic. You can freecompiled patterns, contexts, and stacks in any order, anytime.Just <i>do not</i> call <b>pcre2_match()</b> with a match context pointing to analready freed stack, as that will cause SEGFAULT. (Also, do not free a stackcurrently used by <b>pcre2_match()</b> in another thread). You can also replacethe stack in a context at any time when it is not in use. You should free theprevious stack before assigning a replacement.</P><P>(5) Should I allocate/free a stack every time before/after calling<b>pcre2_match()</b>?<br><br>No, because this is too costly in terms of resources. However, you couldimplement some clever idea which release the stack if it is not used in let'ssay two minutes. The JIT callback can help to achieve this without keeping alist of patterns.</P><P>(6) OK, the stack is for long term memory allocation. But what happens if apattern causes stack overflow with a stack of 1MiB? Is that 1MiB kept until thestack is freed?<br><br>Especially on embedded systems, it might be a good idea to release memorysometimes without freeing the stack. There is no API for this at the moment.Probably a function call which returns with the currently allocated memory forany stack and another which allows releasing memory (shrinking the stack) wouldbe a good idea if someone needs this.</P><P>(7) This is too much of a headache. Isn't there any better solution for JITstack handling?<br><br>No, thanks to Windows. If POSIX threads were used everywhere, we could throwout this complicated API.</P><br><a name="SEC9" href="#TOC1">FREEING JIT SPECULATIVE MEMORY</a><br><P><b>void pcre2_jit_free_unused_memory(pcre2_general_context *<i>gcontext</i>);</b></P><P>The JIT executable allocator does not free all memory when it is possible. Itexpects new allocations, and keeps some free memory around to improveallocation speed. However, in low memory conditions, it might be better to freeall possible memory. You can cause this to happen by callingpcre2_jit_free_unused_memory(). Its argument is a general context, for custommemory management, or NULL for standard memory management.</P><br><a name="SEC10" href="#TOC1">EXAMPLE CODE</a><br><P>This is a single-threaded example that specifies a JIT stack without using acallback. A real program should include error checking after all the functioncalls.<pre> int rc; pcre2_code *re; pcre2_match_data *match_data; pcre2_match_context *mcontext; pcre2_jit_stack *jit_stack;
re = pcre2_compile(pattern, PCRE2_ZERO_TERMINATED, 0, &errornumber, &erroffset, NULL); rc = pcre2_jit_compile(re, PCRE2_JIT_COMPLETE); mcontext = pcre2_match_context_create(NULL); jit_stack = pcre2_jit_stack_create(32*1024, 512*1024, NULL); pcre2_jit_stack_assign(mcontext, NULL, jit_stack); match_data = pcre2_match_data_create(re, 10); rc = pcre2_match(re, subject, length, 0, 0, match_data, mcontext); /* Process result */
pcre2_code_free(re); pcre2_match_data_free(match_data); pcre2_match_context_free(mcontext); pcre2_jit_stack_free(jit_stack);
<a name="fastpath"></a></PRE></P><br><a name="SEC11" href="#TOC1">JIT FAST PATH API</a><br><P>Because the API described above falls back to interpreted matching when JIT isnot available, it is convenient for programs that are written for general usein many environments. However, calling JIT via <b>pcre2_match()</b> does have aperformance impact. Programs that are written for use where JIT is known to beavailable, and which need the best possible performance, can instead use a"fast path" API to call JIT matching directly instead of calling<b>pcre2_match()</b> (obviously only for patterns that have been successfullyprocessed by <b>pcre2_jit_compile()</b>).</P><P>The fast path function is called <b>pcre2_jit_match()</b>, and it takes exactlythe same arguments as <b>pcre2_match()</b>. However, the subject string must bespecified with a length; PCRE2_ZERO_TERMINATED is not supported. Unsupportedoption bits (for example, PCRE2_ANCHORED and PCRE2_ENDANCHORED) are ignored, asis the PCRE2_NO_JIT option. The return values are also the same as for<b>pcre2_match()</b>, plus PCRE2_ERROR_JIT_BADOPTION if a matching mode (partialor complete) is requested that was not compiled.</P><P>When you call <b>pcre2_match()</b>, as well as testing for invalid options, anumber of other sanity checks are performed on the arguments. For example, ifthe subject pointer is NULL but the length is non-zero, an immediate error isgiven. Also, unless PCRE2_NO_UTF_CHECK is set, a UTF subject string is testedfor validity. In the interests of speed, these checks do not happen on the JITfast path. If invalid UTF data is passed when PCRE2_MATCH_INVALID_UTF was notset for <b>pcre2_compile()</b>, the result is undefined. The program may crashor loop or give wrong results. In the absence of PCRE2_MATCH_INVALID_UTF youshould call <b>pcre2_jit_match()</b> in UTF mode only if you are sure thesubject is valid.</P><P>Bypassing the sanity checks and the <b>pcre2_match()</b> wrapping can givespeedups of more than 10%.</P><br><a name="SEC12" href="#TOC1">SEE ALSO</a><br><P><b>pcre2api</b>(3), <b>pcre2unicode</b>(3)</P><br><a name="SEC13" href="#TOC1">AUTHOR</a><br><P>Philip Hazel (FAQ by Zoltan Herczeg)<br>Retired from University Computing Service<br>Cambridge, England.<br></P><br><a name="SEC14" href="#TOC1">REVISION</a><br><P>Last updated: 21 February 2024<br>Copyright © 1997-2024 University of Cambridge.<br><p>Return to the <a href="index.html">PCRE2 index page</a>.</p>
|
