| 1 | commit 7af8e8717def179fd7b69e173abd347c1a3547cb |
| 2 | Author: ph10 <ph10@2f5784b3-3f2a-0410-8824-cb99058d5e15> |
| 3 | Date: Wed Aug 5 15:38:32 2015 +0000 |
| 4 | |
| 5 | Fix buffer overflow for named references in (?| situations. |
| 6 | |
| 7 | |
| 8 | git-svn-id: svn://vcs.exim.org/pcre/code/trunk@1585 2f5784b3-3f2a-0410-8824-cb99058d5e15 |
| 9 | |
| 10 | diff --git a/pcre_compile.c b/pcre_compile.c |
| 11 | index 7d9f276..89ca8f1 100644 |
| 12 | --- a/pcre_compile.c |
| 13 | +++ b/pcre_compile.c |
| 14 | @@ -6668,6 +6668,7 @@ for (;; ptr++) |
| 15 | /* ------------------------------------------------------------ */ |
| 16 | case CHAR_VERTICAL_LINE: /* Reset capture count for each branch */ |
| 17 | reset_bracount = TRUE; |
| 18 | + cd->dupgroups = TRUE; /* Record (?| encountered */ |
| 19 | /* Fall through */ |
| 20 | |
| 21 | /* ------------------------------------------------------------ */ |
| 22 | @@ -7178,7 +7179,8 @@ for (;; ptr++) |
| 23 | if (lengthptr != NULL) |
| 24 | { |
| 25 | named_group *ng; |
| 26 | - |
| 27 | + recno = 0; |
| 28 | + |
| 29 | if (namelen == 0) |
| 30 | { |
| 31 | *errorcodeptr = ERR62; |
| 32 | @@ -7195,32 +7197,6 @@ for (;; ptr++) |
| 33 | goto FAILED; |
| 34 | } |
| 35 | |
| 36 | - /* The name table does not exist in the first pass; instead we must |
| 37 | - scan the list of names encountered so far in order to get the |
| 38 | - number. If the name is not found, set the value to 0 for a forward |
| 39 | - reference. */ |
| 40 | - |
| 41 | - recno = 0; |
| 42 | - ng = cd->named_groups; |
| 43 | - for (i = 0; i < cd->names_found; i++, ng++) |
| 44 | - { |
| 45 | - if (namelen == ng->length && |
| 46 | - STRNCMP_UC_UC(name, ng->name, namelen) == 0) |
| 47 | - { |
| 48 | - open_capitem *oc; |
| 49 | - recno = ng->number; |
| 50 | - if (is_recurse) break; |
| 51 | - for (oc = cd->open_caps; oc != NULL; oc = oc->next) |
| 52 | - { |
| 53 | - if (oc->number == recno) |
| 54 | - { |
| 55 | - oc->flag = TRUE; |
| 56 | - break; |
| 57 | - } |
| 58 | - } |
| 59 | - } |
| 60 | - } |
| 61 | - |
| 62 | /* Count named back references. */ |
| 63 | |
| 64 | if (!is_recurse) cd->namedrefcount++; |
| 65 | @@ -7242,7 +7218,44 @@ for (;; ptr++) |
| 66 | issue is fixed "properly" in PCRE2. As PCRE1 is now in maintenance |
| 67 | only mode, we finesse the bug by allowing more memory always. */ |
| 68 | |
| 69 | - /* if (recno == 0) */ *lengthptr += 2 + 2*LINK_SIZE; |
| 70 | + *lengthptr += 2 + 2*LINK_SIZE; |
| 71 | + |
| 72 | + /* It is even worse than that. The current reference may be to an |
| 73 | + existing named group with a different number (so apparently not |
| 74 | + recursive) but which later on is also attached to a group with the |
| 75 | + current number. This can only happen if $(| has been previous |
| 76 | + encountered. In that case, we allow yet more memory, just in case. |
| 77 | + (Again, this is fixed "properly" in PCRE2. */ |
| 78 | + |
| 79 | + if (cd->dupgroups) *lengthptr += 2 + 2*LINK_SIZE; |
| 80 | + |
| 81 | + /* Otherwise, check for recursion here. The name table does not exist |
| 82 | + in the first pass; instead we must scan the list of names encountered |
| 83 | + so far in order to get the number. If the name is not found, leave |
| 84 | + the value of recno as 0 for a forward reference. */ |
| 85 | + |
| 86 | + else |
| 87 | + { |
| 88 | + ng = cd->named_groups; |
| 89 | + for (i = 0; i < cd->names_found; i++, ng++) |
| 90 | + { |
| 91 | + if (namelen == ng->length && |
| 92 | + STRNCMP_UC_UC(name, ng->name, namelen) == 0) |
| 93 | + { |
| 94 | + open_capitem *oc; |
| 95 | + recno = ng->number; |
| 96 | + if (is_recurse) break; |
| 97 | + for (oc = cd->open_caps; oc != NULL; oc = oc->next) |
| 98 | + { |
| 99 | + if (oc->number == recno) |
| 100 | + { |
| 101 | + oc->flag = TRUE; |
| 102 | + break; |
| 103 | + } |
| 104 | + } |
| 105 | + } |
| 106 | + } |
| 107 | + } |
| 108 | } |
| 109 | |
| 110 | /* In the real compile, search the name table. We check the name |
| 111 | @@ -7289,8 +7302,6 @@ for (;; ptr++) |
| 112 | for (i++; i < cd->names_found; i++) |
| 113 | { |
| 114 | if (STRCMP_UC_UC(slot + IMM2_SIZE, cslot + IMM2_SIZE) != 0) break; |
| 115 | - |
| 116 | - |
| 117 | count++; |
| 118 | cslot += cd->name_entry_size; |
| 119 | } |
| 120 | @@ -9239,6 +9250,7 @@ cd->names_found = 0; |
| 121 | cd->name_entry_size = 0; |
| 122 | cd->name_table = NULL; |
| 123 | cd->dupnames = FALSE; |
| 124 | +cd->dupgroups = FALSE; |
| 125 | cd->namedrefcount = 0; |
| 126 | cd->start_code = cworkspace; |
| 127 | cd->hwm = cworkspace; |
| 128 | @@ -9273,7 +9285,7 @@ if (errorcode != 0) goto PCRE_EARLY_ERROR_RETURN; |
| 129 | |
| 130 | DPRINTF(("end pre-compile: length=%d workspace=%d\n", length, |
| 131 | (int)(cd->hwm - cworkspace))); |
| 132 | - |
| 133 | + |
| 134 | if (length > MAX_PATTERN_SIZE) |
| 135 | { |
| 136 | errorcode = ERR20; |
| 137 | diff --git a/pcre_internal.h b/pcre_internal.h |
| 138 | index 80e2420..544d9c0 100644 |
| 139 | --- a/pcre_internal.h |
| 140 | +++ b/pcre_internal.h |
| 141 | @@ -2454,6 +2454,7 @@ typedef struct compile_data { |
| 142 | BOOL had_pruneorskip; /* (*PRUNE) or (*SKIP) encountered */ |
| 143 | BOOL check_lookbehind; /* Lookbehinds need later checking */ |
| 144 | BOOL dupnames; /* Duplicate names exist */ |
| 145 | + BOOL dupgroups; /* Duplicate groups exist: (?| found */ |
| 146 | BOOL iscondassert; /* Next assert is a condition */ |
| 147 | int nltype; /* Newline type */ |
| 148 | int nllen; /* Newline string length */ |
| 149 | diff --git a/testdata/testinput2 b/testdata/testinput2 |
| 150 | index df2c1cc..e119bd9 100644 |
| 151 | --- a/testdata/testinput2 |
| 152 | +++ b/testdata/testinput2 |
| 153 | @@ -4194,4 +4194,6 @@ backtracking verbs. --/ |
| 154 | |
| 155 | /(?1){3918}(((((0(\k'R'))))(?J)(?'R'(?'R'\3){99})))/I |
| 156 | |
| 157 | +/(?J:(?|(:(?|(?'R')(\k'R')|((?'R')))H'Rk'Rf)|s(?'R')))/ |
| 158 | + |
| 159 | /-- End of testinput2 --/ |
| 160 | diff --git a/testdata/testoutput2 b/testdata/testoutput2 |
| 161 | index d3fc254..54db2cc 100644 |
| 162 | --- a/testdata/testoutput2 |
| 163 | +++ b/testdata/testoutput2 |
| 164 | @@ -14537,4 +14537,6 @@ Duplicate name status changes |
| 165 | No first char |
| 166 | Need char = '0' |
| 167 | |
| 168 | +/(?J:(?|(:(?|(?'R')(\k'R')|((?'R')))H'Rk'Rf)|s(?'R')))/ |
| 169 | + |
| 170 | /-- End of testinput2 --/ |