/[pcre]/code/trunk/pcre_compile.c
ViewVC logotype

Diff of /code/trunk/pcre_compile.c

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

revision 772 by ph10, Tue Nov 22 13:36:51 2011 UTC revision 773 by ph10, Wed Nov 30 18:10:27 2011 UTC
# Line 88  so this number is very generous. Line 88  so this number is very generous.
88  The same workspace is used during the second, actual compile phase for  The same workspace is used during the second, actual compile phase for
89  remembering forward references to groups so that they can be filled in at the  remembering forward references to groups so that they can be filled in at the
90  end. Each entry in this list occupies LINK_SIZE bytes, so even when LINK_SIZE  end. Each entry in this list occupies LINK_SIZE bytes, so even when LINK_SIZE
91  is 4 there is plenty of room. */  is 4 there is plenty of room for most patterns. However, the memory can get
92    filled up by repetitions of forward references, for example patterns like
93    /(?1){0,1999}(b)/, and one user did hit the limit. The code has been changed so
94    that the workspace is expanded using malloc() in this situation. The value
95    below is therefore a minimum, and we put a maximum on it for safety. The
96    minimum is now also defined in terms of LINK_SIZE so that the use of malloc()
97    kicks in at the same number of forward references in all cases. */
98    
99  #define COMPILE_WORK_SIZE (4096)  #define COMPILE_WORK_SIZE (2048*LINK_SIZE)
100    #define COMPILE_WORK_SIZE_MAX (100*COMPILE_WORK_SIZE)
101    
102  /* The overrun tests check for a slightly smaller size so that they detect the  /* The overrun tests check for a slightly smaller size so that they detect the
103  overrun before it actually does run off the end of the data block. */  overrun before it actually does run off the end of the data block. */
104    
105  #define WORK_SIZE_CHECK (COMPILE_WORK_SIZE - 100)  #define WORK_SIZE_SAFETY_MARGIN (100)
106    
107    
108  /* Table for handling escaped characters in the range '0'-'z'. Positive returns  /* Table for handling escaped characters in the range '0'-'z'. Positive returns
# Line 582  return s; Line 589  return s;
589    
590    
591  /*************************************************  /*************************************************
592    *           Expand the workspace                 *
593    *************************************************/
594    
595    /* This function is called during the second compiling phase, if the number of
596    forward references fills the existing workspace, which is originally a block on
597    the stack. A larger block is obtained from malloc() unless the ultimate limit
598    has been reached or the increase will be rather small.
599    
600    Argument: pointer to the compile data block
601    Returns:  0 if all went well, else an error number
602    */
603    
604    static int
605    expand_workspace(compile_data *cd)
606    {
607    uschar *newspace;
608    int newsize = cd->workspace_size * 2;
609    
610    if (newsize > COMPILE_WORK_SIZE_MAX) newsize = COMPILE_WORK_SIZE_MAX;
611    if (cd->workspace_size >= COMPILE_WORK_SIZE_MAX ||
612        newsize - cd->workspace_size < WORK_SIZE_SAFETY_MARGIN)
613     return ERR72;
614    
615    newspace = (pcre_malloc)(newsize);
616    if (newspace == NULL) return ERR21;
617    
618    memcpy(newspace, cd->start_workspace, cd->workspace_size);
619    cd->hwm = (uschar *)newspace + (cd->hwm - cd->start_workspace);
620    if (cd->workspace_size > COMPILE_WORK_SIZE)
621      (pcre_free)((void *)cd->start_workspace);
622    cd->start_workspace = newspace;
623    cd->workspace_size = newsize;
624    return 0;
625    }
626    
627    
628    
629    /*************************************************
630  *            Check for counted repeat            *  *            Check for counted repeat            *
631  *************************************************/  *************************************************/
632    
# Line 3332  for (;; ptr++) Line 3377  for (;; ptr++)
3377  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
3378      if (code > cd->hwm) cd->hwm = code;                 /* High water info */      if (code > cd->hwm) cd->hwm = code;                 /* High water info */
3379  #endif  #endif
3380      if (code > cd->start_workspace + WORK_SIZE_CHECK)   /* Check for overrun */      if (code > cd->start_workspace + cd->workspace_size -
3381            WORK_SIZE_SAFETY_MARGIN)                       /* Check for overrun */
3382        {        {
3383        *errorcodeptr = ERR52;        *errorcodeptr = ERR52;
3384        goto FAILED;        goto FAILED;
# Line 3382  for (;; ptr++) Line 3428  for (;; ptr++)
3428    /* In the real compile phase, just check the workspace used by the forward    /* In the real compile phase, just check the workspace used by the forward
3429    reference list. */    reference list. */
3430    
3431    else if (cd->hwm > cd->start_workspace + WORK_SIZE_CHECK)    else if (cd->hwm > cd->start_workspace + cd->workspace_size -
3432               WORK_SIZE_SAFETY_MARGIN)
3433      {      {
3434      *errorcodeptr = ERR52;      *errorcodeptr = ERR52;
3435      goto FAILED;      goto FAILED;
# Line 4885  for (;; ptr++) Line 4932  for (;; ptr++)
4932              }              }
4933    
4934            /* This is compiling for real. If there is a set first byte for            /* This is compiling for real. If there is a set first byte for
4935            the group, and we have not yet set a "required byte", set it. */            the group, and we have not yet set a "required byte", set it. Make
4936              sure there is enough workspace for copying forward references before
4937              doing the copy. */
4938    
4939            else            else
4940              {              {
4941              if (groupsetfirstbyte && reqbyte < 0) reqbyte = firstbyte;              if (groupsetfirstbyte && reqbyte < 0) reqbyte = firstbyte;
4942    
4943              for (i = 1; i < repeat_min; i++)              for (i = 1; i < repeat_min; i++)
4944                {                {
4945                uschar *hc;                uschar *hc;
4946                uschar *this_hwm = cd->hwm;                uschar *this_hwm = cd->hwm;
4947                memcpy(code, previous, len);                memcpy(code, previous, len);
4948    
4949                  while (cd->hwm > cd->start_workspace + cd->workspace_size -
4950                         WORK_SIZE_SAFETY_MARGIN - (this_hwm - save_hwm))
4951                    {
4952                    int save_offset = save_hwm - cd->start_workspace;
4953                    int this_offset = this_hwm - cd->start_workspace;
4954                    *errorcodeptr = expand_workspace(cd);
4955                    if (*errorcodeptr != 0) goto FAILED;
4956                    save_hwm = (uschar *)cd->start_workspace + save_offset;
4957                    this_hwm = (uschar *)cd->start_workspace + this_offset;
4958                    }
4959    
4960                for (hc = save_hwm; hc < this_hwm; hc += LINK_SIZE)                for (hc = save_hwm; hc < this_hwm; hc += LINK_SIZE)
4961                  {                  {
                 if (cd->hwm >= cd->start_workspace + WORK_SIZE_CHECK)  
                   {  
                   *errorcodeptr = ERR72;  
                   goto FAILED;  
                   }  
4962                  PUT(cd->hwm, 0, GET(hc, 0) + len);                  PUT(cd->hwm, 0, GET(hc, 0) + len);
4963                  cd->hwm += LINK_SIZE;                  cd->hwm += LINK_SIZE;
4964                  }                  }
# Line 4967  for (;; ptr++) Line 5024  for (;; ptr++)
5024              }              }
5025    
5026            memcpy(code, previous, len);            memcpy(code, previous, len);
5027    
5028              /* Ensure there is enough workspace for forward references before
5029              copying them. */
5030    
5031              while (cd->hwm > cd->start_workspace + cd->workspace_size -
5032                     WORK_SIZE_SAFETY_MARGIN - (this_hwm - save_hwm))
5033                {
5034                int save_offset = save_hwm - cd->start_workspace;
5035                int this_offset = this_hwm - cd->start_workspace;
5036                *errorcodeptr = expand_workspace(cd);
5037                if (*errorcodeptr != 0) goto FAILED;
5038                save_hwm = (uschar *)cd->start_workspace + save_offset;
5039                this_hwm = (uschar *)cd->start_workspace + this_offset;
5040                }
5041    
5042            for (hc = save_hwm; hc < this_hwm; hc += LINK_SIZE)            for (hc = save_hwm; hc < this_hwm; hc += LINK_SIZE)
5043              {              {
             if (cd->hwm >= cd->start_workspace + WORK_SIZE_CHECK)  
               {  
               *errorcodeptr = ERR72;  
               goto FAILED;  
               }  
5044              PUT(cd->hwm, 0, GET(hc, 0) + len + ((i != 0)? 2+LINK_SIZE : 1));              PUT(cd->hwm, 0, GET(hc, 0) + len + ((i != 0)? 2+LINK_SIZE : 1));
5045              cd->hwm += LINK_SIZE;              cd->hwm += LINK_SIZE;
5046              }              }
# Line 5991  for (;; ptr++) Line 6058  for (;; ptr++)
6058                of the group. Then remember the forward reference. */                of the group. Then remember the forward reference. */
6059    
6060                called = cd->start_code + recno;                called = cd->start_code + recno;
6061                if (cd->hwm >= cd->start_workspace + WORK_SIZE_CHECK)                if (cd->hwm >= cd->start_workspace + cd->workspace_size -
6062                      WORK_SIZE_SAFETY_MARGIN)
6063                  {                  {
6064                  *errorcodeptr = ERR72;                  *errorcodeptr = expand_workspace(cd);
6065                  goto FAILED;                  if (*errorcodeptr != 0) goto FAILED;
6066                  }                  }
6067                PUTINC(cd->hwm, 0, (int)(code + 1 - cd->start_code));                PUTINC(cd->hwm, 0, (int)(code + 1 - cd->start_code));
6068                }                }
# Line 7246  compile_data *cd = &compile_block; Line 7314  compile_data *cd = &compile_block;
7314  computing the amount of memory that is needed. Compiled items are thrown away  computing the amount of memory that is needed. Compiled items are thrown away
7315  as soon as possible, so that a fairly large buffer should be sufficient for  as soon as possible, so that a fairly large buffer should be sufficient for
7316  this purpose. The same space is used in the second phase for remembering where  this purpose. The same space is used in the second phase for remembering where
7317  to fill in forward references to subpatterns. */  to fill in forward references to subpatterns. That may overflow, in which case
7318    new memory is obtained from malloc(). */
7319    
7320  uschar cworkspace[COMPILE_WORK_SIZE];  uschar cworkspace[COMPILE_WORK_SIZE];
7321    
# Line 7436  cd->bracount = cd->final_bracount = 0; Line 7505  cd->bracount = cd->final_bracount = 0;
7505  cd->names_found = 0;  cd->names_found = 0;
7506  cd->name_entry_size = 0;  cd->name_entry_size = 0;
7507  cd->name_table = NULL;  cd->name_table = NULL;
 cd->start_workspace = cworkspace;  
7508  cd->start_code = cworkspace;  cd->start_code = cworkspace;
7509  cd->hwm = cworkspace;  cd->hwm = cworkspace;
7510    cd->start_workspace = cworkspace;
7511    cd->workspace_size = COMPILE_WORK_SIZE;
7512  cd->start_pattern = (const uschar *)pattern;  cd->start_pattern = (const uschar *)pattern;
7513  cd->end_pattern = (const uschar *)(pattern + strlen(pattern));  cd->end_pattern = (const uschar *)(pattern + strlen(pattern));
7514  cd->req_varyopt = 0;  cd->req_varyopt = 0;
# Line 7516  cd->names_found = 0; Line 7586  cd->names_found = 0;
7586  cd->name_table = (uschar *)re + re->name_table_offset;  cd->name_table = (uschar *)re + re->name_table_offset;
7587  codestart = cd->name_table + re->name_entry_size * re->name_count;  codestart = cd->name_table + re->name_entry_size * re->name_count;
7588  cd->start_code = codestart;  cd->start_code = codestart;
7589  cd->hwm = cworkspace;  cd->hwm = (uschar *)(cd->start_workspace);
7590  cd->req_varyopt = 0;  cd->req_varyopt = 0;
7591  cd->had_accept = FALSE;  cd->had_accept = FALSE;
7592  cd->check_lookbehind = FALSE;  cd->check_lookbehind = FALSE;
# Line 7550  if debugging, leave the test till after Line 7620  if debugging, leave the test till after
7620  if (code - codestart > length) errorcode = ERR23;  if (code - codestart > length) errorcode = ERR23;
7621  #endif  #endif
7622    
7623  /* Fill in any forward references that are required. */  /* Fill in any forward references that are required. There may be repeated
7624    references; optimize for them, as searching a large regex takes time. */
7625    
7626  while (errorcode == 0 && cd->hwm > cworkspace)  if (cd->hwm > cd->start_workspace)
7627    {    {
7628    int offset, recno;    int prev_recno = -1;
7629    const uschar *groupptr;    const uschar *groupptr = NULL;
7630    cd->hwm -= LINK_SIZE;    while (errorcode == 0 && cd->hwm > cd->start_workspace)
7631    offset = GET(cd->hwm, 0);      {
7632    recno = GET(codestart, offset);      int offset, recno;
7633    groupptr = _pcre_find_bracket(codestart, utf8, recno);      cd->hwm -= LINK_SIZE;
7634    if (groupptr == NULL) errorcode = ERR53;      offset = GET(cd->hwm, 0);
7635      else PUT(((uschar *)codestart), offset, (int)(groupptr - codestart));      recno = GET(codestart, offset);
7636    }      if (recno != prev_recno)
7637          {
7638          groupptr = _pcre_find_bracket(codestart, utf8, recno);
7639          prev_recno = recno;
7640          }
7641        if (groupptr == NULL) errorcode = ERR53;
7642          else PUT(((uschar *)codestart), offset, (int)(groupptr - codestart));
7643        }
7644      }
7645    
7646    /* If the workspace had to be expanded, free the new memory. */
7647    
7648    if (cd->workspace_size > COMPILE_WORK_SIZE)
7649      (pcre_free)((void *)cd->start_workspace);
7650    
7651  /* Give an error if there's back reference to a non-existent capturing  /* Give an error if there's back reference to a non-existent capturing
7652  subpattern. */  subpattern. */
Legend:
Removed from v.772  
changed lines
  Added in v.773
  ViewVC Help
Powered by ViewVC 1.1.5