src/patches/linux/linux-4.9.13-layer7.patch

   1 diff --git a/include/linux/netfilter/xt_layer7.h b/include/linux/netfilter/xt_layer7.h
   2 new file mode 100644
   3 index 0000000..147cd64
   4 --- /dev/null
   5 +++ b/include/linux/netfilter/xt_layer7.h
   6 @@ -0,0 +1,13 @@
   7 +#ifndef _XT_LAYER7_H
   8 +#define _XT_LAYER7_H
   9 +
  10 +#define MAX_PATTERN_LEN 8192
  11 +#define MAX_PROTOCOL_LEN 256
  12 +
  13 +struct xt_layer7_info {
  14 +    char protocol[MAX_PROTOCOL_LEN];
  15 +    char pattern[MAX_PATTERN_LEN];
  16 +    u_int8_t invert;
  17 +};
  18 +
  19 +#endif /* _XT_LAYER7_H */
  20 diff --git a/include/net/netfilter/nf_conntrack.h b/include/net/netfilter/nf_conntrack.h
  21 index d9d52c0..09389b6 100644
  22 --- a/include/net/netfilter/nf_conntrack.h
  23 +++ b/include/net/netfilter/nf_conntrack.h
  24 @@ -120,6 +120,22 @@ struct nf_conn {
  25         /* Extensions */
  26         struct nf_ct_ext *ext;
  27
  28 +#if defined(CONFIG_NETFILTER_XT_MATCH_LAYER7) || \
  29 +    defined(CONFIG_NETFILTER_XT_MATCH_LAYER7_MODULE)
  30 +       struct {
  31 +               /*
  32 +                * e.g. "http". NULL before decision. "unknown" after decision
  33 +                * if no match.
  34 +                */
  35 +               char *app_proto;
  36 +               /*
  37 +                * application layer data so far. NULL after match decision.
  38 +                */
  39 +               char *app_data;
  40 +               unsigned int app_data_len;
  41 +       } layer7;
  42 +#endif
  43 +
  44         /* Storage reserved for other modules, must be the last member */
  45         union nf_conntrack_proto proto;
  46  };
  47 diff --git a/net/netfilter/Kconfig b/net/netfilter/Kconfig
  48 index e8d56d9..ab4ae1d 100644
  49 --- a/net/netfilter/Kconfig
  50 +++ b/net/netfilter/Kconfig
  51 @@ -1238,6 +1238,26 @@ config NETFILTER_XT_MATCH_L2TP
  52
  53         To compile it as a module, choose M here. If unsure, say N.
  54
  55 +config NETFILTER_XT_MATCH_LAYER7
  56 +       tristate '"layer7" match support'
  57 +       depends on NETFILTER_XTABLES
  58 +       depends on NETFILTER_ADVANCED
  59 +       depends on NF_CONNTRACK
  60 +       help
  61 +         Say Y if you want to be able to classify connections (and their
  62 +         packets) based on regular expression matching of their application
  63 +         layer data.   This is one way to classify applications such as
  64 +         peer-to-peer filesharing systems that do not always use the same
  65 +         port.
  66 +
  67 +         To compile it as a module, choose M here.  If unsure, say N.
  68 +
  69 +config NETFILTER_XT_MATCH_LAYER7_DEBUG
  70 +       bool 'Layer 7 debugging output'
  71 +       depends on NETFILTER_XT_MATCH_LAYER7
  72 +       help
  73 +         Say Y to get lots of debugging output.
  74 +
  75  config NETFILTER_XT_MATCH_LENGTH
  76         tristate '"length" match support'
  77         depends on NETFILTER_ADVANCED
  78 diff --git a/net/netfilter/Makefile b/net/netfilter/Makefile
  79 index c23c3c8..916b9d5 100644
  80 --- a/net/netfilter/Makefile
  81 +++ b/net/netfilter/Makefile
  82 @@ -174,6 +174,7 @@ obj-$(CONFIG_NETFILTER_XT_MATCH_RECENT) += xt_recent.o
  83  obj-$(CONFIG_NETFILTER_XT_MATCH_SCTP) += xt_sctp.o
  84  obj-$(CONFIG_NETFILTER_XT_MATCH_SOCKET) += xt_socket.o
  85  obj-$(CONFIG_NETFILTER_XT_MATCH_STATE) += xt_state.o
  86 +obj-$(CONFIG_NETFILTER_XT_MATCH_LAYER7) += xt_layer7.o
  87  obj-$(CONFIG_NETFILTER_XT_MATCH_STATISTIC) += xt_statistic.o
  88  obj-$(CONFIG_NETFILTER_XT_MATCH_STRING) += xt_string.o
  89  obj-$(CONFIG_NETFILTER_XT_MATCH_TCPMSS) += xt_tcpmss.o
  90 diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c
  91 index 0f87e5d..1f355a0 100644
  92 --- a/net/netfilter/nf_conntrack_core.c
  93 +++ b/net/netfilter/nf_conntrack_core.c
  94 @@ -406,6 +406,11 @@ destroy_conntrack(struct nf_conntrack *nfct)
  95          */
  96         nf_ct_remove_expectations(ct);
  97
  98 +#if defined(CONFIG_NETFILTER_XT_MATCH_LAYER7) || defined(CONFIG_NETFILTER_XT_MATCH_LAYER7_MODULE)
  99 +       if(ct->layer7.app_data)
 100 +               kfree(ct->layer7.app_data);
 101 +#endif
 102 +
 103         nf_ct_del_from_dying_or_unconfirmed_list(ct);
 104
 105         local_bh_enable();
 106 diff --git a/net/netfilter/nf_conntrack_standalone.c b/net/netfilter/nf_conntrack_standalone.c
 107 index 5f446cd..92f29f9 100644
 108 --- a/net/netfilter/nf_conntrack_standalone.c
 109 +++ b/net/netfilter/nf_conntrack_standalone.c
 110 @@ -274,6 +274,11 @@ static int ct_seq_show(struct seq_file *s, void *v)
 111         ct_show_zone(s, ct, NF_CT_DEFAULT_ZONE_DIR);
 112         ct_show_delta_time(s, ct);
 113
 114 +#if defined(CONFIG_NETFILTER_XT_MATCH_LAYER7) || defined(CONFIG_NETFILTER_XT_MATCH_LAYER7_MODULE)
 115 +       if(ct->layer7.app_proto)
 116 +               seq_printf(s, "l7proto=%s ", ct->layer7.app_proto);
 117 +#endif
 118 +
 119         seq_printf(s, "use=%u\n", atomic_read(&ct->ct_general.use));
 120
 121         if (seq_has_overflowed(s))
 122 diff --git a/net/netfilter/regexp/regexp.c b/net/netfilter/regexp/regexp.c
 123 new file mode 100644
 124 index 0000000..9006988
 125 --- /dev/null
 126 +++ b/net/netfilter/regexp/regexp.c
 127 @@ -0,0 +1,1197 @@
 128 +/*
 129 + * regcomp and regexec -- regsub and regerror are elsewhere
 130 + * @(#)regexp.c        1.3 of 18 April 87
 131 + *
 132 + *     Copyright (c) 1986 by University of Toronto.
 133 + *     Written by Henry Spencer.  Not derived from licensed software.
 134 + *
 135 + *     Permission is granted to anyone to use this software for any
 136 + *     purpose on any computer system, and to redistribute it freely,
 137 + *     subject to the following restrictions:
 138 + *
 139 + *     1. The author is not responsible for the consequences of use of
 140 + *             this software, no matter how awful, even if they arise
 141 + *             from defects in it.
 142 + *
 143 + *     2. The origin of this software must not be misrepresented, either
 144 + *             by explicit claim or by omission.
 145 + *
 146 + *     3. Altered versions must be plainly marked as such, and must not
 147 + *             be misrepresented as being the original software.
 148 + *
 149 + * Beware that some of this code is subtly aware of the way operator
 150 + * precedence is structured in regular expressions.  Serious changes in
 151 + * regular-expression syntax might require a total rethink.
 152 + *
 153 + * This code was modified by Ethan Sommer to work within the kernel
 154 + * (it now uses kmalloc etc..)
 155 + *
 156 + * Modified slightly by Matthew Strait to use more modern C.
 157 + */
 158 +
 159 +#include "regexp.h"
 160 +#include "regmagic.h"
 161 +
 162 +/* added by ethan and matt.  Lets it work in both kernel and user space.
 163 +(So iptables can use it, for instance.)  Yea, it goes both ways... */
 164 +#if __KERNEL__
 165 +  #define malloc(foo) kmalloc(foo,GFP_ATOMIC)
 166 +#else
 167 +  #define printk(format,args...) printf(format,##args)
 168 +#endif
 169 +
 170 +void regerror(char * s)
 171 +{
 172 +        printk("<3>Regexp: %s\n", s);
 173 +        /* NOTREACHED */
 174 +}
 175 +
 176 +/*
 177 + * The "internal use only" fields in regexp.h are present to pass info from
 178 + * compile to execute that permits the execute phase to run lots faster on
 179 + * simple cases.  They are:
 180 + *
 181 + * regstart    char that must begin a match; '\0' if none obvious
 182 + * reganch     is the match anchored (at beginning-of-line only)?
 183 + * regmust     string (pointer into program) that match must include, or NULL
 184 + * regmlen     length of regmust string
 185 + *
 186 + * Regstart and reganch permit very fast decisions on suitable starting points
 187 + * for a match, cutting down the work a lot.  Regmust permits fast rejection
 188 + * of lines that cannot possibly match.  The regmust tests are costly enough
 189 + * that regcomp() supplies a regmust only if the r.e. contains something
 190 + * potentially expensive (at present, the only such thing detected is * or +
 191 + * at the start of the r.e., which can involve a lot of backup).  Regmlen is
 192 + * supplied because the test in regexec() needs it and regcomp() is computing
 193 + * it anyway.
 194 + */
 195 +
 196 +/*
 197 + * Structure for regexp "program".  This is essentially a linear encoding
 198 + * of a nondeterministic finite-state machine (aka syntax charts or
 199 + * "railroad normal form" in parsing technology).  Each node is an opcode
 200 + * plus a "next" pointer, possibly plus an operand.  "Next" pointers of
 201 + * all nodes except BRANCH implement concatenation; a "next" pointer with
 202 + * a BRANCH on both ends of it is connecting two alternatives.  (Here we
 203 + * have one of the subtle syntax dependencies:  an individual BRANCH (as
 204 + * opposed to a collection of them) is never concatenated with anything
 205 + * because of operator precedence.)  The operand of some types of node is
 206 + * a literal string; for others, it is a node leading into a sub-FSM.  In
 207 + * particular, the operand of a BRANCH node is the first node of the branch.
 208 + * (NB this is *not* a tree structure:  the tail of the branch connects
 209 + * to the thing following the set of BRANCHes.)  The opcodes are:
 210 + */
 211 +
 212 +/* definition  number  opnd?   meaning */
 213 +#define        END     0       /* no   End of program. */
 214 +#define        BOL     1       /* no   Match "" at beginning of line. */
 215 +#define        EOL     2       /* no   Match "" at end of line. */
 216 +#define        ANY     3       /* no   Match any one character. */
 217 +#define        ANYOF   4       /* str  Match any character in this string. */
 218 +#define        ANYBUT  5       /* str  Match any character not in this string. */
 219 +#define        BRANCH  6       /* node Match this alternative, or the next... */
 220 +#define        BACK    7       /* no   Match "", "next" ptr points backward. */
 221 +#define        EXACTLY 8       /* str  Match this string. */
 222 +#define        NOTHING 9       /* no   Match empty string. */
 223 +#define        STAR    10      /* node Match this (simple) thing 0 or more times. */
 224 +#define        PLUS    11      /* node Match this (simple) thing 1 or more times. */
 225 +#define        OPEN    20      /* no   Mark this point in input as start of #n. */
 226 +                       /*      OPEN+1 is number 1, etc. */
 227 +#define        CLOSE   30      /* no   Analogous to OPEN. */
 228 +
 229 +/*
 230 + * Opcode notes:
 231 + *
 232 + * BRANCH      The set of branches constituting a single choice are hooked
 233 + *             together with their "next" pointers, since precedence prevents
 234 + *             anything being concatenated to any individual branch.  The
 235 + *             "next" pointer of the last BRANCH in a choice points to the
 236 + *             thing following the whole choice.  This is also where the
 237 + *             final "next" pointer of each individual branch points; each
 238 + *             branch starts with the operand node of a BRANCH node.
 239 + *
 240 + * BACK                Normal "next" pointers all implicitly point forward; BACK
 241 + *             exists to make loop structures possible.
 242 + *
 243 + * STAR,PLUS   '?', and complex '*' and '+', are implemented as circular
 244 + *             BRANCH structures using BACK.  Simple cases (one character
 245 + *             per match) are implemented with STAR and PLUS for speed
 246 + *             and to minimize recursive plunges.
 247 + *
 248 + * OPEN,CLOSE  ...are numbered at compile time.
 249 + */
 250 +
 251 +/*
 252 + * A node is one char of opcode followed by two chars of "next" pointer.
 253 + * "Next" pointers are stored as two 8-bit pieces, high order first.  The
 254 + * value is a positive offset from the opcode of the node containing it.
 255 + * An operand, if any, simply follows the node.  (Note that much of the
 256 + * code generation knows about this implicit relationship.)
 257 + *
 258 + * Using two bytes for the "next" pointer is vast overkill for most things,
 259 + * but allows patterns to get big without disasters.
 260 + */
 261 +#define        OP(p)   (*(p))
 262 +#define        NEXT(p) (((*((p)+1)&0377)<<8) + (*((p)+2)&0377))
 263 +#define        OPERAND(p)      ((p) + 3)
 264 +
 265 +/*
 266 + * See regmagic.h for one further detail of program structure.
 267 + */
 268 +
 269 +
 270 +/*
 271 + * Utility definitions.
 272 + */
 273 +#ifndef CHARBITS
 274 +#define        UCHARAT(p)      ((int)*(unsigned char *)(p))
 275 +#else
 276 +#define        UCHARAT(p)      ((int)*(p)&CHARBITS)
 277 +#endif
 278 +
 279 +#define        FAIL(m) { regerror(m); return(NULL); }
 280 +#define        ISMULT(c)       ((c) == '*' || (c) == '+' || (c) == '?')
 281 +#define        META    "^$.[()|?+*\\"
 282 +
 283 +/*
 284 + * Flags to be passed up and down.
 285 + */
 286 +#define        HASWIDTH        01      /* Known never to match null string. */
 287 +#define        SIMPLE          02      /* Simple enough to be STAR/PLUS operand. */
 288 +#define        SPSTART         04      /* Starts with * or +. */
 289 +#define        WORST           0       /* Worst case. */
 290 +
 291 +/*
 292 + * Global work variables for regcomp().
 293 + */
 294 +struct match_globals {
 295 +char *reginput;                /* String-input pointer. */
 296 +char *regbol;          /* Beginning of input, for ^ check. */
 297 +char **regstartp;      /* Pointer to startp array. */
 298 +char **regendp;                /* Ditto for endp. */
 299 +char *regparse;                /* Input-scan pointer. */
 300 +int regnpar;           /* () count. */
 301 +char regdummy;
 302 +char *regcode;         /* Code-emit pointer; &regdummy = don't. */
 303 +long regsize;          /* Code size. */
 304 +};
 305 +
 306 +/*
 307 + * Forward declarations for regcomp()'s friends.
 308 + */
 309 +#ifndef STATIC
 310 +#define        STATIC  static
 311 +#endif
 312 +STATIC char *reg(struct match_globals *g, int paren,int *flagp);
 313 +STATIC char *regbranch(struct match_globals *g, int *flagp);
 314 +STATIC char *regpiece(struct match_globals *g, int *flagp);
 315 +STATIC char *regatom(struct match_globals *g, int *flagp);
 316 +STATIC char *regnode(struct match_globals *g, char op);
 317 +STATIC char *regnext(struct match_globals *g, char *p);
 318 +STATIC void regc(struct match_globals *g, char b);
 319 +STATIC void reginsert(struct match_globals *g, char op, char *opnd);
 320 +STATIC void regtail(struct match_globals *g, char *p, char *val);
 321 +STATIC void regoptail(struct match_globals *g, char *p, char *val);
 322 +
 323 +
 324 +__kernel_size_t my_strcspn(const char *s1,const char *s2)
 325 +{
 326 +        char *scan1;
 327 +        char *scan2;
 328 +        int count;
 329 +
 330 +        count = 0;
 331 +        for (scan1 = (char *)s1; *scan1 != '\0'; scan1++) {
 332 +                for (scan2 = (char *)s2; *scan2 != '\0';)       /* ++ moved down. */
 333 +                        if (*scan1 == *scan2++)
 334 +                                return(count);
 335 +                count++;
 336 +        }
 337 +        return(count);
 338 +}
 339 +
 340 +/*
 341 + - regcomp - compile a regular expression into internal code
 342 + *
 343 + * We can't allocate space until we know how big the compiled form will be,
 344 + * but we can't compile it (and thus know how big it is) until we've got a
 345 + * place to put the code.  So we cheat:  we compile it twice, once with code
 346 + * generation turned off and size counting turned on, and once "for real".
 347 + * This also means that we don't allocate space until we are sure that the
 348 + * thing really will compile successfully, and we never have to move the
 349 + * code and thus invalidate pointers into it.  (Note that it has to be in
 350 + * one piece because free() must be able to free it all.)
 351 + *
 352 + * Beware that the optimization-preparation code in here knows about some
 353 + * of the structure of the compiled regexp.
 354 + */
 355 +regexp *
 356 +regcomp(char *exp,int *patternsize)
 357 +{
 358 +       register regexp *r;
 359 +       register char *scan;
 360 +       register char *longest;
 361 +       register int len;
 362 +       int flags;
 363 +       struct match_globals g;
 364 +
 365 +       /* commented out by ethan
 366 +          extern char *malloc();
 367 +       */
 368 +
 369 +       if (exp == NULL)
 370 +               FAIL("NULL argument");
 371 +
 372 +       /* First pass: determine size, legality. */
 373 +       g.regparse = exp;
 374 +       g.regnpar = 1;
 375 +       g.regsize = 0L;
 376 +       g.regcode = &g.regdummy;
 377 +       regc(&g, MAGIC);
 378 +       if (reg(&g, 0, &flags) == NULL)
 379 +               return(NULL);
 380 +
 381 +       /* Small enough for pointer-storage convention? */
 382 +       if (g.regsize >= 32767L)                /* Probably could be 65535L. */
 383 +               FAIL("regexp too big");
 384 +
 385 +       /* Allocate space. */
 386 +       *patternsize=sizeof(regexp) + (unsigned)g.regsize;
 387 +       r = (regexp *)malloc(sizeof(regexp) + (unsigned)g.regsize);
 388 +       if (r == NULL)
 389 +               FAIL("out of space");
 390 +
 391 +       /* Second pass: emit code. */
 392 +       g.regparse = exp;
 393 +       g.regnpar = 1;
 394 +       g.regcode = r->program;
 395 +       regc(&g, MAGIC);
 396 +       if (reg(&g, 0, &flags) == NULL)
 397 +               return(NULL);
 398 +
 399 +       /* Dig out information for optimizations. */
 400 +       r->regstart = '\0';     /* Worst-case defaults. */
 401 +       r->reganch = 0;
 402 +       r->regmust = NULL;
 403 +       r->regmlen = 0;
 404 +       scan = r->program+1;                    /* First BRANCH. */
 405 +       if (OP(regnext(&g, scan)) == END) {             /* Only one top-level choice. */
 406 +               scan = OPERAND(scan);
 407 +
 408 +               /* Starting-point info. */
 409 +               if (OP(scan) == EXACTLY)
 410 +                       r->regstart = *OPERAND(scan);
 411 +               else if (OP(scan) == BOL)
 412 +                       r->reganch++;
 413 +
 414 +               /*
 415 +                * If there's something expensive in the r.e., find the
 416 +                * longest literal string that must appear and make it the
 417 +                * regmust.  Resolve ties in favor of later strings, since
 418 +                * the regstart check works with the beginning of the r.e.
 419 +                * and avoiding duplication strengthens checking.  Not a
 420 +                * strong reason, but sufficient in the absence of others.
 421 +                */
 422 +               if (flags&SPSTART) {
 423 +                       longest = NULL;
 424 +                       len = 0;
 425 +                       for (; scan != NULL; scan = regnext(&g, scan))
 426 +                               if (OP(scan) == EXACTLY && strlen(OPERAND(scan)) >= len) {
 427 +                                       longest = OPERAND(scan);
 428 +                                       len = strlen(OPERAND(scan));
 429 +                               }
 430 +                       r->regmust = longest;
 431 +                       r->regmlen = len;
 432 +               }
 433 +       }
 434 +
 435 +       return(r);
 436 +}
 437 +
 438 +/*
 439 + - reg - regular expression, i.e. main body or parenthesized thing
 440 + *
 441 + * Caller must absorb opening parenthesis.
 442 + *
 443 + * Combining parenthesis handling with the base level of regular expression
 444 + * is a trifle forced, but the need to tie the tails of the branches to what
 445 + * follows makes it hard to avoid.
 446 + */
 447 +static char *
 448 +reg(struct match_globals *g, int paren, int *flagp /* Parenthesized? */ )
 449 +{
 450 +       register char *ret;
 451 +       register char *br;
 452 +       register char *ender;
 453 +       register int parno = 0; /* 0 makes gcc happy */
 454 +       int flags;
 455 +
 456 +       *flagp = HASWIDTH;      /* Tentatively. */
 457 +
 458 +       /* Make an OPEN node, if parenthesized. */
 459 +       if (paren) {
 460 +               if (g->regnpar >= NSUBEXP)
 461 +                       FAIL("too many ()");
 462 +               parno = g->regnpar;
 463 +               g->regnpar++;
 464 +               ret = regnode(g, OPEN+parno);
 465 +       } else
 466 +               ret = NULL;
 467 +
 468 +       /* Pick up the branches, linking them together. */
 469 +       br = regbranch(g, &flags);
 470 +       if (br == NULL)
 471 +               return(NULL);
 472 +       if (ret != NULL)
 473 +               regtail(g, ret, br);    /* OPEN -> first. */
 474 +       else
 475 +               ret = br;
 476 +       if (!(flags&HASWIDTH))
 477 +               *flagp &= ~HASWIDTH;
 478 +       *flagp |= flags&SPSTART;
 479 +       while (*g->regparse == '|') {
 480 +               g->regparse++;
 481 +               br = regbranch(g, &flags);
 482 +               if (br == NULL)
 483 +                       return(NULL);
 484 +               regtail(g, ret, br);    /* BRANCH -> BRANCH. */
 485 +               if (!(flags&HASWIDTH))
 486 +                       *flagp &= ~HASWIDTH;
 487 +               *flagp |= flags&SPSTART;
 488 +       }
 489 +
 490 +       /* Make a closing node, and hook it on the end. */
 491 +       ender = regnode(g, (paren) ? CLOSE+parno : END);
 492 +       regtail(g, ret, ender);
 493 +
 494 +       /* Hook the tails of the branches to the closing node. */
 495 +       for (br = ret; br != NULL; br = regnext(g, br))
 496 +               regoptail(g, br, ender);
 497 +
 498 +       /* Check for proper termination. */
 499 +       if (paren && *g->regparse++ != ')') {
 500 +               FAIL("unmatched ()");
 501 +       } else if (!paren && *g->regparse != '\0') {
 502 +               if (*g->regparse == ')') {
 503 +                       FAIL("unmatched ()");
 504 +               } else
 505 +                       FAIL("junk on end");    /* "Can't happen". */
 506 +               /* NOTREACHED */
 507 +       }
 508 +
 509 +       return(ret);
 510 +}
 511 +
 512 +/*
 513 + - regbranch - one alternative of an | operator
 514 + *
 515 + * Implements the concatenation operator.
 516 + */
 517 +static char *
 518 +regbranch(struct match_globals *g, int *flagp)
 519 +{
 520 +       register char *ret;
 521 +       register char *chain;
 522 +       register char *latest;
 523 +       int flags;
 524 +
 525 +       *flagp = WORST;         /* Tentatively. */
 526 +
 527 +       ret = regnode(g, BRANCH);
 528 +       chain = NULL;
 529 +       while (*g->regparse != '\0' && *g->regparse != '|' && *g->regparse != ')') {
 530 +               latest = regpiece(g, &flags);
 531 +               if (latest == NULL)
 532 +                       return(NULL);
 533 +               *flagp |= flags&HASWIDTH;
 534 +               if (chain == NULL)      /* First piece. */
 535 +                       *flagp |= flags&SPSTART;
 536 +               else
 537 +                       regtail(g, chain, latest);
 538 +               chain = latest;
 539 +       }
 540 +       if (chain == NULL)      /* Loop ran zero times. */
 541 +               (void) regnode(g, NOTHING);
 542 +
 543 +       return(ret);
 544 +}
 545 +
 546 +/*
 547 + - regpiece - something followed by possible [*+?]
 548 + *
 549 + * Note that the branching code sequences used for ? and the general cases
 550 + * of * and + are somewhat optimized:  they use the same NOTHING node as
 551 + * both the endmarker for their branch list and the body of the last branch.
 552 + * It might seem that this node could be dispensed with entirely, but the
 553 + * endmarker role is not redundant.
 554 + */
 555 +static char *
 556 +regpiece(struct match_globals *g, int *flagp)
 557 +{
 558 +       register char *ret;
 559 +       register char op;
 560 +       register char *next;
 561 +       int flags;
 562 +
 563 +       ret = regatom(g, &flags);
 564 +       if (ret == NULL)
 565 +               return(NULL);
 566 +
 567 +       op = *g->regparse;
 568 +       if (!ISMULT(op)) {
 569 +               *flagp = flags;
 570 +               return(ret);
 571 +       }
 572 +
 573 +       if (!(flags&HASWIDTH) && op != '?')
 574 +               FAIL("*+ operand could be empty");
 575 +       *flagp = (op != '+') ? (WORST|SPSTART) : (WORST|HASWIDTH);
 576 +
 577 +       if (op == '*' && (flags&SIMPLE))
 578 +               reginsert(g, STAR, ret);
 579 +       else if (op == '*') {
 580 +               /* Emit x* as (x&|), where & means "self". */
 581 +               reginsert(g, BRANCH, ret);                      /* Either x */
 582 +               regoptail(g, ret, regnode(g, BACK));            /* and loop */
 583 +               regoptail(g, ret, ret);                 /* back */
 584 +               regtail(g, ret, regnode(g, BRANCH));            /* or */
 585 +               regtail(g, ret, regnode(g, NOTHING));           /* null. */
 586 +       } else if (op == '+' && (flags&SIMPLE))
 587 +               reginsert(g, PLUS, ret);
 588 +       else if (op == '+') {
 589 +               /* Emit x+ as x(&|), where & means "self". */
 590 +               next = regnode(g, BRANCH);                      /* Either */
 591 +               regtail(g, ret, next);
 592 +               regtail(g, regnode(g, BACK), ret);              /* loop back */
 593 +               regtail(g, next, regnode(g, BRANCH));           /* or */
 594 +               regtail(g, ret, regnode(g, NOTHING));           /* null. */
 595 +       } else if (op == '?') {
 596 +               /* Emit x? as (x|) */
 597 +               reginsert(g, BRANCH, ret);                      /* Either x */
 598 +               regtail(g, ret, regnode(g, BRANCH));            /* or */
 599 +               next = regnode(g, NOTHING);             /* null. */
 600 +               regtail(g, ret, next);
 601 +               regoptail(g, ret, next);
 602 +       }
 603 +       g->regparse++;
 604 +       if (ISMULT(*g->regparse))
 605 +               FAIL("nested *?+");
 606 +
 607 +       return(ret);
 608 +}
 609 +
 610 +/*
 611 + - regatom - the lowest level
 612 + *
 613 + * Optimization:  gobbles an entire sequence of ordinary characters so that
 614 + * it can turn them into a single node, which is smaller to store and
 615 + * faster to run.  Backslashed characters are exceptions, each becoming a
 616 + * separate node; the code is simpler that way and it's not worth fixing.
 617 + */
 618 +static char *
 619 +regatom(struct match_globals *g, int *flagp)
 620 +{
 621 +       register char *ret;
 622 +       int flags;
 623 +
 624 +       *flagp = WORST;         /* Tentatively. */
 625 +
 626 +       switch (*g->regparse++) {
 627 +       case '^':
 628 +               ret = regnode(g, BOL);
 629 +               break;
 630 +       case '$':
 631 +               ret = regnode(g, EOL);
 632 +               break;
 633 +       case '.':
 634 +               ret = regnode(g, ANY);
 635 +               *flagp |= HASWIDTH|SIMPLE;
 636 +               break;
 637 +       case '[': {
 638 +                       register int class;
 639 +                       register int classend;
 640 +
 641 +                       if (*g->regparse == '^') {      /* Complement of range. */
 642 +                               ret = regnode(g, ANYBUT);
 643 +                               g->regparse++;
 644 +                       } else
 645 +                               ret = regnode(g, ANYOF);
 646 +                       if (*g->regparse == ']' || *g->regparse == '-')
 647 +                               regc(g, *g->regparse++);
 648 +                       while (*g->regparse != '\0' && *g->regparse != ']') {
 649 +                               if (*g->regparse == '-') {
 650 +                                       g->regparse++;
 651 +                                       if (*g->regparse == ']' || *g->regparse == '\0')
 652 +                                               regc(g, '-');
 653 +                                       else {
 654 +                                               class = UCHARAT(g->regparse-2)+1;
 655 +                                               classend = UCHARAT(g->regparse);
 656 +                                               if (class > classend+1)
 657 +                                                       FAIL("invalid [] range");
 658 +                                               for (; class <= classend; class++)
 659 +                                                       regc(g, class);
 660 +                                               g->regparse++;
 661 +                                       }
 662 +                               } else
 663 +                                       regc(g, *g->regparse++);
 664 +                       }
 665 +                       regc(g, '\0');
 666 +                       if (*g->regparse != ']')
 667 +                               FAIL("unmatched []");
 668 +                       g->regparse++;
 669 +                       *flagp |= HASWIDTH|SIMPLE;
 670 +               }
 671 +               break;
 672 +       case '(':
 673 +               ret = reg(g, 1, &flags);
 674 +               if (ret == NULL)
 675 +                       return(NULL);
 676 +               *flagp |= flags&(HASWIDTH|SPSTART);
 677 +               break;
 678 +       case '\0':
 679 +       case '|':
 680 +       case ')':
 681 +               FAIL("internal urp");   /* Supposed to be caught earlier. */
 682 +               break;
 683 +       case '?':
 684 +       case '+':
 685 +       case '*':
 686 +               FAIL("?+* follows nothing");
 687 +               break;
 688 +       case '\\':
 689 +               if (*g->regparse == '\0')
 690 +                       FAIL("trailing \\");
 691 +               ret = regnode(g, EXACTLY);
 692 +               regc(g, *g->regparse++);
 693 +               regc(g, '\0');
 694 +               *flagp |= HASWIDTH|SIMPLE;
 695 +               break;
 696 +       default: {
 697 +                       register int len;
 698 +                       register char ender;
 699 +
 700 +                       g->regparse--;
 701 +                       len = my_strcspn((const char *)g->regparse, (const char *)META);
 702 +                       if (len <= 0)
 703 +                               FAIL("internal disaster");
 704 +                       ender = *(g->regparse+len);
 705 +                       if (len > 1 && ISMULT(ender))
 706 +                               len--;          /* Back off clear of ?+* operand. */
 707 +                       *flagp |= HASWIDTH;
 708 +                       if (len == 1)
 709 +                               *flagp |= SIMPLE;
 710 +                       ret = regnode(g, EXACTLY);
 711 +                       while (len > 0) {
 712 +                               regc(g, *g->regparse++);
 713 +                               len--;
 714 +                       }
 715 +                       regc(g, '\0');
 716 +               }
 717 +               break;
 718 +       }
 719 +
 720 +       return(ret);
 721 +}
 722 +
 723 +/*
 724 + - regnode - emit a node
 725 + */
 726 +static char *                  /* Location. */
 727 +regnode(struct match_globals *g, char op)
 728 +{
 729 +       register char *ret;
 730 +       register char *ptr;
 731 +
 732 +       ret = g->regcode;
 733 +       if (ret == &g->regdummy) {
 734 +               g->regsize += 3;
 735 +               return(ret);
 736 +       }
 737 +
 738 +       ptr = ret;
 739 +       *ptr++ = op;
 740 +       *ptr++ = '\0';          /* Null "next" pointer. */
 741 +       *ptr++ = '\0';
 742 +       g->regcode = ptr;
 743 +
 744 +       return(ret);
 745 +}
 746 +
 747 +/*
 748 + - regc - emit (if appropriate) a byte of code
 749 + */
 750 +static void
 751 +regc(struct match_globals *g, char b)
 752 +{
 753 +       if (g->regcode != &g->regdummy)
 754 +               *g->regcode++ = b;
 755 +       else
 756 +               g->regsize++;
 757 +}
 758 +
 759 +/*
 760 + - reginsert - insert an operator in front of already-emitted operand
 761 + *
 762 + * Means relocating the operand.
 763 + */
 764 +static void
 765 +reginsert(struct match_globals *g, char op, char* opnd)
 766 +{
 767 +       register char *src;
 768 +       register char *dst;
 769 +       register char *place;
 770 +
 771 +       if (g->regcode == &g->regdummy) {
 772 +               g->regsize += 3;
 773 +               return;
 774 +       }
 775 +
 776 +       src = g->regcode;
 777 +       g->regcode += 3;
 778 +       dst = g->regcode;
 779 +       while (src > opnd)
 780 +               *--dst = *--src;
 781 +
 782 +       place = opnd;           /* Op node, where operand used to be. */
 783 +       *place++ = op;
 784 +       *place++ = '\0';
 785 +       *place++ = '\0';
 786 +}
 787 +
 788 +/*
 789 + - regtail - set the next-pointer at the end of a node chain
 790 + */
 791 +static void
 792 +regtail(struct match_globals *g, char *p, char *val)
 793 +{
 794 +       register char *scan;
 795 +       register char *temp;
 796 +       register int offset;
 797 +
 798 +       if (p == &g->regdummy)
 799 +               return;
 800 +
 801 +       /* Find last node. */
 802 +       scan = p;
 803 +       for (;;) {
 804 +               temp = regnext(g, scan);
 805 +               if (temp == NULL)
 806 +                       break;
 807 +               scan = temp;
 808 +       }
 809 +
 810 +       if (OP(scan) == BACK)
 811 +               offset = scan - val;
 812 +       else
 813 +               offset = val - scan;
 814 +       *(scan+1) = (offset>>8)&0377;
 815 +       *(scan+2) = offset&0377;
 816 +}
 817 +
 818 +/*
 819 + - regoptail - regtail on operand of first argument; nop if operandless
 820 + */
 821 +static void
 822 +regoptail(struct match_globals *g, char *p, char *val)
 823 +{
 824 +       /* "Operandless" and "op != BRANCH" are synonymous in practice. */
 825 +       if (p == NULL || p == &g->regdummy || OP(p) != BRANCH)
 826 +               return;
 827 +       regtail(g, OPERAND(p), val);
 828 +}
 829 +
 830 +/*
 831 + * regexec and friends
 832 + */
 833 +
 834 +
 835 +/*
 836 + * Forwards.
 837 + */
 838 +STATIC int regtry(struct match_globals *g, regexp *prog, char *string);
 839 +STATIC int regmatch(struct match_globals *g, char *prog);
 840 +STATIC int regrepeat(struct match_globals *g, char *p);
 841 +
 842 +#ifdef DEBUG
 843 +int regnarrate = 0;
 844 +void regdump();
 845 +STATIC char *regprop(char *op);
 846 +#endif
 847 +
 848 +/*
 849 + - regexec - match a regexp against a string
 850 + */
 851 +int
 852 +regexec(regexp *prog, char *string)
 853 +{
 854 +       register char *s;
 855 +       struct match_globals g;
 856 +
 857 +       /* Be paranoid... */
 858 +       if (prog == NULL || string == NULL) {
 859 +               printk("<3>Regexp: NULL parameter\n");
 860 +               return(0);
 861 +       }
 862 +
 863 +       /* Check validity of program. */
 864 +       if (UCHARAT(prog->program) != MAGIC) {
 865 +               printk("<3>Regexp: corrupted program\n");
 866 +               return(0);
 867 +       }
 868 +
 869 +       /* If there is a "must appear" string, look for it. */
 870 +       if (prog->regmust != NULL) {
 871 +               s = string;
 872 +               while ((s = strchr(s, prog->regmust[0])) != NULL) {
 873 +                       if (strncmp(s, prog->regmust, prog->regmlen) == 0)
 874 +                               break;  /* Found it. */
 875 +                       s++;
 876 +               }
 877 +               if (s == NULL)  /* Not present. */
 878 +                       return(0);
 879 +       }
 880 +
 881 +       /* Mark beginning of line for ^ . */
 882 +       g.regbol = string;
 883 +
 884 +       /* Simplest case:  anchored match need be tried only once. */
 885 +       if (prog->reganch)
 886 +               return(regtry(&g, prog, string));
 887 +
 888 +       /* Messy cases:  unanchored match. */
 889 +       s = string;
 890 +       if (prog->regstart != '\0')
 891 +               /* We know what char it must start with. */
 892 +               while ((s = strchr(s, prog->regstart)) != NULL) {
 893 +                       if (regtry(&g, prog, s))
 894 +                               return(1);
 895 +                       s++;
 896 +               }
 897 +       else
 898 +               /* We don't -- general case. */
 899 +               do {
 900 +                       if (regtry(&g, prog, s))
 901 +                               return(1);
 902 +               } while (*s++ != '\0');
 903 +
 904 +       /* Failure. */
 905 +       return(0);
 906 +}
 907 +
 908 +/*
 909 + - regtry - try match at specific point
 910 + */
 911 +static int                     /* 0 failure, 1 success */
 912 +regtry(struct match_globals *g, regexp *prog, char *string)
 913 +{
 914 +       register int i;
 915 +       register char **sp;
 916 +       register char **ep;
 917 +
 918 +       g->reginput = string;
 919 +       g->regstartp = prog->startp;
 920 +       g->regendp = prog->endp;
 921 +
 922 +       sp = prog->startp;
 923 +       ep = prog->endp;
 924 +       for (i = NSUBEXP; i > 0; i--) {
 925 +               *sp++ = NULL;
 926 +               *ep++ = NULL;
 927 +       }
 928 +       if (regmatch(g, prog->program + 1)) {
 929 +               prog->startp[0] = string;
 930 +               prog->endp[0] = g->reginput;
 931 +               return(1);
 932 +       } else
 933 +               return(0);
 934 +}
 935 +
 936 +/*
 937 + - regmatch - main matching routine
 938 + *
 939 + * Conceptually the strategy is simple:  check to see whether the current
 940 + * node matches, call self recursively to see whether the rest matches,
 941 + * and then act accordingly.  In practice we make some effort to avoid
 942 + * recursion, in particular by going through "ordinary" nodes (that don't
 943 + * need to know whether the rest of the match failed) by a loop instead of
 944 + * by recursion.
 945 + */
 946 +static int                     /* 0 failure, 1 success */
 947 +regmatch(struct match_globals *g, char *prog)
 948 +{
 949 +       register char *scan = prog; /* Current node. */
 950 +       char *next;                 /* Next node. */
 951 +
 952 +#ifdef DEBUG
 953 +       if (scan != NULL && regnarrate)
 954 +               fprintf(stderr, "%s(\n", regprop(scan));
 955 +#endif
 956 +       while (scan != NULL) {
 957 +#ifdef DEBUG
 958 +               if (regnarrate)
 959 +                       fprintf(stderr, "%s...\n", regprop(scan));
 960 +#endif
 961 +               next = regnext(g, scan);
 962 +
 963 +               switch (OP(scan)) {
 964 +               case BOL:
 965 +                       if (g->reginput != g->regbol)
 966 +                               return(0);
 967 +                       break;
 968 +               case EOL:
 969 +                       if (*g->reginput != '\0')
 970 +                               return(0);
 971 +                       break;
 972 +               case ANY:
 973 +                       if (*g->reginput == '\0')
 974 +                               return(0);
 975 +                       g->reginput++;
 976 +                       break;
 977 +               case EXACTLY: {
 978 +                               register int len;
 979 +                               register char *opnd;
 980 +
 981 +                               opnd = OPERAND(scan);
 982 +                               /* Inline the first character, for speed. */
 983 +                               if (*opnd != *g->reginput)
 984 +                                       return(0);
 985 +                               len = strlen(opnd);
 986 +                               if (len > 1 && strncmp(opnd, g->reginput, len) != 0)
 987 +                                       return(0);
 988 +                               g->reginput += len;
 989 +                       }
 990 +                       break;
 991 +               case ANYOF:
 992 +                       if (*g->reginput == '\0' || strchr(OPERAND(scan), *g->reginput) == NULL)
 993 +                               return(0);
 994 +                       g->reginput++;
 995 +                       break;
 996 +               case ANYBUT:
 997 +                       if (*g->reginput == '\0' || strchr(OPERAND(scan), *g->reginput) != NULL)
 998 +                               return(0);
 999 +                       g->reginput++;
1000 +                       break;
1001 +               case NOTHING:
1002 +               case BACK:
1003 +                       break;
1004 +               case OPEN+1:
1005 +               case OPEN+2:
1006 +               case OPEN+3:
1007 +               case OPEN+4:
1008 +               case OPEN+5:
1009 +               case OPEN+6:
1010 +               case OPEN+7:
1011 +               case OPEN+8:
1012 +               case OPEN+9: {
1013 +                               register int no;
1014 +                               register char *save;
1015 +
1016 +                               no = OP(scan) - OPEN;
1017 +                               save = g->reginput;
1018 +
1019 +                               if (regmatch(g, next)) {
1020 +                                       /*
1021 +                                        * Don't set startp if some later
1022 +                                        * invocation of the same parentheses
1023 +                                        * already has.
1024 +                                        */
1025 +                                       if (g->regstartp[no] == NULL)
1026 +                                               g->regstartp[no] = save;
1027 +                                       return(1);
1028 +                               } else
1029 +                                       return(0);
1030 +                       }
1031 +                       break;
1032 +               case CLOSE+1:
1033 +               case CLOSE+2:
1034 +               case CLOSE+3:
1035 +               case CLOSE+4:
1036 +               case CLOSE+5:
1037 +               case CLOSE+6:
1038 +               case CLOSE+7:
1039 +               case CLOSE+8:
1040 +               case CLOSE+9:
1041 +                       {
1042 +                               register int no;
1043 +                               register char *save;
1044 +
1045 +                               no = OP(scan) - CLOSE;
1046 +                               save = g->reginput;
1047 +
1048 +                               if (regmatch(g, next)) {
1049 +                                       /*
1050 +                                        * Don't set endp if some later
1051 +                                        * invocation of the same parentheses
1052 +                                        * already has.
1053 +                                        */
1054 +                                       if (g->regendp[no] == NULL)
1055 +                                               g->regendp[no] = save;
1056 +                                       return(1);
1057 +                               } else
1058 +                                       return(0);
1059 +                       }
1060 +                       break;
1061 +               case BRANCH: {
1062 +                               register char *save;
1063 +
1064 +                               if (OP(next) != BRANCH)         /* No choice. */
1065 +                                       next = OPERAND(scan);   /* Avoid recursion. */
1066 +                               else {
1067 +                                       do {
1068 +                                               save = g->reginput;
1069 +                                               if (regmatch(g, OPERAND(scan)))
1070 +                                                       return(1);
1071 +                                               g->reginput = save;
1072 +                                               scan = regnext(g, scan);
1073 +                                       } while (scan != NULL && OP(scan) == BRANCH);
1074 +                                       return(0);
1075 +                                       /* NOTREACHED */
1076 +                               }
1077 +                       }
1078 +                       break;
1079 +               case STAR:
1080 +               case PLUS: {
1081 +                               register char nextch;
1082 +                               register int no;
1083 +                               register char *save;
1084 +                               register int min;
1085 +
1086 +                               /*
1087 +                                * Lookahead to avoid useless match attempts
1088 +                                * when we know what character comes next.
1089 +                                */
1090 +                               nextch = '\0';
1091 +                               if (OP(next) == EXACTLY)
1092 +                                       nextch = *OPERAND(next);
1093 +                               min = (OP(scan) == STAR) ? 0 : 1;
1094 +                               save = g->reginput;
1095 +                               no = regrepeat(g, OPERAND(scan));
1096 +                               while (no >= min) {
1097 +                                       /* If it could work, try it. */
1098 +                                       if (nextch == '\0' || *g->reginput == nextch)
1099 +                                               if (regmatch(g, next))
1100 +                                                       return(1);
1101 +                                       /* Couldn't or didn't -- back up. */
1102 +                                       no--;
1103 +                                       g->reginput = save + no;
1104 +                               }
1105 +                               return(0);
1106 +                       }
1107 +                       break;
1108 +               case END:
1109 +                       return(1);      /* Success! */
1110 +                       break;
1111 +               default:
1112 +                       printk("<3>Regexp: memory corruption\n");
1113 +                       return(0);
1114 +                       break;
1115 +               }
1116 +
1117 +               scan = next;
1118 +       }
1119 +
1120 +       /*
1121 +        * We get here only if there's trouble -- normally "case END" is
1122 +        * the terminating point.
1123 +        */
1124 +       printk("<3>Regexp: corrupted pointers\n");
1125 +       return(0);
1126 +}
1127 +
1128 +/*
1129 + - regrepeat - repeatedly match something simple, report how many
1130 + */
1131 +static int
1132 +regrepeat(struct match_globals *g, char *p)
1133 +{
1134 +       register int count = 0;
1135 +       register char *scan;
1136 +       register char *opnd;
1137 +
1138 +       scan = g->reginput;
1139 +       opnd = OPERAND(p);
1140 +       switch (OP(p)) {
1141 +       case ANY:
1142 +               count = strlen(scan);
1143 +               scan += count;
1144 +               break;
1145 +       case EXACTLY:
1146 +               while (*opnd == *scan) {
1147 +                       count++;
1148 +                       scan++;
1149 +               }
1150 +               break;
1151 +       case ANYOF:
1152 +               while (*scan != '\0' && strchr(opnd, *scan) != NULL) {
1153 +                       count++;
1154 +                       scan++;
1155 +               }
1156 +               break;
1157 +       case ANYBUT:
1158 +               while (*scan != '\0' && strchr(opnd, *scan) == NULL) {
1159 +                       count++;
1160 +                       scan++;
1161 +               }
1162 +               break;
1163 +       default:                /* Oh dear.  Called inappropriately. */
1164 +               printk("<3>Regexp: internal foulup\n");
1165 +               count = 0;      /* Best compromise. */
1166 +               break;
1167 +       }
1168 +       g->reginput = scan;
1169 +
1170 +       return(count);
1171 +}
1172 +
1173 +/*
1174 + - regnext - dig the "next" pointer out of a node
1175 + */
1176 +static char*
1177 +regnext(struct match_globals *g, char *p)
1178 +{
1179 +       register int offset;
1180 +
1181 +       if (p == &g->regdummy)
1182 +               return(NULL);
1183 +
1184 +       offset = NEXT(p);
1185 +       if (offset == 0)
1186 +               return(NULL);
1187 +
1188 +       if (OP(p) == BACK)
1189 +               return(p-offset);
1190 +       else
1191 +               return(p+offset);
1192 +}
1193 +
1194 +#ifdef DEBUG
1195 +
1196 +STATIC char *regprop();
1197 +
1198 +/*
1199 + - regdump - dump a regexp onto stdout in vaguely comprehensible form
1200 + */
1201 +void
1202 +regdump(regexp *r)
1203 +{
1204 +       register char *s;
1205 +       register char op = EXACTLY;     /* Arbitrary non-END op. */
1206 +       register char *next;
1207 +       /* extern char *strchr(); */
1208 +
1209 +
1210 +       s = r->program + 1;
1211 +       while (op != END) {     /* While that wasn't END last time... */
1212 +               op = OP(s);
1213 +               printf("%2d%s", s-r->program, regprop(s));      /* Where, what. */
1214 +               next = regnext(s);
1215 +               if (next == NULL)               /* Next ptr. */
1216 +                       printf("(0)");
1217 +               else
1218 +                       printf("(%d)", (s-r->program)+(next-s));
1219 +               s += 3;
1220 +               if (op == ANYOF || op == ANYBUT || op == EXACTLY) {
1221 +                       /* Literal string, where present. */
1222 +                       while (*s != '\0') {
1223 +                               putchar(*s);
1224 +                               s++;
1225 +                       }
1226 +                       s++;
1227 +               }
1228 +               putchar('\n');
1229 +       }
1230 +
1231 +       /* Header fields of interest. */
1232 +       if (r->regstart != '\0')
1233 +               printf("start `%c' ", r->regstart);
1234 +       if (r->reganch)
1235 +               printf("anchored ");
1236 +       if (r->regmust != NULL)
1237 +               printf("must have \"%s\"", r->regmust);
1238 +       printf("\n");
1239 +}
1240 +
1241 +/*
1242 + - regprop - printable representation of opcode
1243 + */
1244 +static char *
1245 +regprop(char *op)
1246 +{
1247 +#define BUFLEN 50
1248 +       register char *p;
1249 +       static char buf[BUFLEN];
1250 +
1251 +       strcpy(buf, ":");
1252 +
1253 +       switch (OP(op)) {
1254 +       case BOL:
1255 +               p = "BOL";
1256 +               break;
1257 +       case EOL:
1258 +               p = "EOL";
1259 +               break;
1260 +       case ANY:
1261 +               p = "ANY";
1262 +               break;
1263 +       case ANYOF:
1264 +               p = "ANYOF";
1265 +               break;
1266 +       case ANYBUT:
1267 +               p = "ANYBUT";
1268 +               break;
1269 +       case BRANCH:
1270 +               p = "BRANCH";
1271 +               break;
1272 +       case EXACTLY:
1273 +               p = "EXACTLY";
1274 +               break;
1275 +       case NOTHING:
1276 +               p = "NOTHING";
1277 +               break;
1278 +       case BACK:
1279 +               p = "BACK";
1280 +               break;
1281 +       case END:
1282 +               p = "END";
1283 +               break;
1284 +       case OPEN+1:
1285 +       case OPEN+2:
1286 +       case OPEN+3:
1287 +       case OPEN+4:
1288 +       case OPEN+5:
1289 +       case OPEN+6:
1290 +       case OPEN+7:
1291 +       case OPEN+8:
1292 +       case OPEN+9:
1293 +               snprintf(buf+strlen(buf),BUFLEN-strlen(buf), "OPEN%d", OP(op)-OPEN);
1294 +               p = NULL;
1295 +               break;
1296 +       case CLOSE+1:
1297 +       case CLOSE+2:
1298 +       case CLOSE+3:
1299 +       case CLOSE+4:
1300 +       case CLOSE+5:
1301 +       case CLOSE+6:
1302 +       case CLOSE+7:
1303 +       case CLOSE+8:
1304 +       case CLOSE+9:
1305 +               snprintf(buf+strlen(buf),BUFLEN-strlen(buf), "CLOSE%d", OP(op)-CLOSE);
1306 +               p = NULL;
1307 +               break;
1308 +       case STAR:
1309 +               p = "STAR";
1310 +               break;
1311 +       case PLUS:
1312 +               p = "PLUS";
1313 +               break;
1314 +       default:
1315 +               printk("<3>Regexp: corrupted opcode\n");
1316 +               break;
1317 +       }
1318 +       if (p != NULL)
1319 +               strncat(buf, p, BUFLEN-strlen(buf));
1320 +       return(buf);
1321 +}
1322 +#endif
1323 +
1324 +
1325 diff --git a/net/netfilter/regexp/regexp.h b/net/netfilter/regexp/regexp.h
1326 new file mode 100644
1327 index 0000000..a72eba7
1328 --- /dev/null
1329 +++ b/net/netfilter/regexp/regexp.h
1330 @@ -0,0 +1,41 @@
1331 +/*
1332 + * Definitions etc. for regexp(3) routines.
1333 + *
1334 + * Caveat:  this is V8 regexp(3) [actually, a reimplementation thereof],
1335 + * not the System V one.
1336 + */
1337 +
1338 +#ifndef REGEXP_H
1339 +#define REGEXP_H
1340 +
1341 +
1342 +/*
1343 +http://www.opensource.apple.com/darwinsource/10.3/expect-1/expect/expect.h ,
1344 +which contains a version of this library, says:
1345 +
1346 + *
1347 + * NSUBEXP must be at least 10, and no greater than 117 or the parser
1348 + * will not work properly.
1349 + *
1350 +
1351 +However, it looks rather like this library is limited to 10.  If you think
1352 +otherwise, let us know.
1353 +*/
1354 +
1355 +#define NSUBEXP  10
1356 +typedef struct regexp {
1357 +       char *startp[NSUBEXP];
1358 +       char *endp[NSUBEXP];
1359 +       char regstart;          /* Internal use only. */
1360 +       char reganch;           /* Internal use only. */
1361 +       char *regmust;          /* Internal use only. */
1362 +       int regmlen;            /* Internal use only. */
1363 +       char program[1];        /* Unwarranted chumminess with compiler. */
1364 +} regexp;
1365 +
1366 +regexp * regcomp(char *exp, int *patternsize);
1367 +int regexec(regexp *prog, char *string);
1368 +void regsub(regexp *prog, char *source, char *dest);
1369 +void regerror(char *s);
1370 +
1371 +#endif
1372 diff --git a/net/netfilter/regexp/regmagic.h b/net/netfilter/regexp/regmagic.h
1373 new file mode 100644
1374 index 0000000..5acf447
1375 --- /dev/null
1376 +++ b/net/netfilter/regexp/regmagic.h
1377 @@ -0,0 +1,5 @@
1378 +/*
1379 + * The first byte of the regexp internal "program" is actually this magic
1380 + * number; the start node begins in the second byte.
1381 + */
1382 +#define        MAGIC   0234
1383 diff --git a/net/netfilter/regexp/regsub.c b/net/netfilter/regexp/regsub.c
1384 new file mode 100644
1385 index 0000000..339631f
1386 --- /dev/null
1387 +++ b/net/netfilter/regexp/regsub.c
1388 @@ -0,0 +1,95 @@
1389 +/*
1390 + * regsub
1391 + * @(#)regsub.c        1.3 of 2 April 86
1392 + *
1393 + *     Copyright (c) 1986 by University of Toronto.
1394 + *     Written by Henry Spencer.  Not derived from licensed software.
1395 + *
1396 + *     Permission is granted to anyone to use this software for any
1397 + *     purpose on any computer system, and to redistribute it freely,
1398 + *     subject to the following restrictions:
1399 + *
1400 + *     1. The author is not responsible for the consequences of use of
1401 + *             this software, no matter how awful, even if they arise
1402 + *             from defects in it.
1403 + *
1404 + *     2. The origin of this software must not be misrepresented, either
1405 + *             by explicit claim or by omission.
1406 + *
1407 + *     3. Altered versions must be plainly marked as such, and must not
1408 + *             be misrepresented as being the original software.
1409 + *
1410 + *
1411 + * This code was modified by Ethan Sommer to work within the kernel
1412 + * (it now uses kmalloc etc..)
1413 + *
1414 + */
1415 +#include "regexp.h"
1416 +#include "regmagic.h"
1417 +#include <linux/string.h>
1418 +
1419 +
1420 +#ifndef CHARBITS
1421 +#define        UCHARAT(p)      ((int)*(unsigned char *)(p))
1422 +#else
1423 +#define        UCHARAT(p)      ((int)*(p)&CHARBITS)
1424 +#endif
1425 +
1426 +#if 0
1427 +//void regerror(char * s)
1428 +//{
1429 +//        printk("regexp(3): %s", s);
1430 +//        /* NOTREACHED */
1431 +//}
1432 +#endif
1433 +
1434 +/*
1435 + - regsub - perform substitutions after a regexp match
1436 + */
1437 +void
1438 +regsub(regexp * prog, char * source, char * dest)
1439 +{
1440 +       register char *src;
1441 +       register char *dst;
1442 +       register char c;
1443 +       register int no;
1444 +       register int len;
1445 +
1446 +       /* Not necessary and gcc doesn't like it -MLS */
1447 +       /*extern char *strncpy();*/
1448 +
1449 +       if (prog == NULL || source == NULL || dest == NULL) {
1450 +               regerror("NULL parm to regsub");
1451 +               return;
1452 +       }
1453 +       if (UCHARAT(prog->program) != MAGIC) {
1454 +               regerror("damaged regexp fed to regsub");
1455 +               return;
1456 +       }
1457 +
1458 +       src = source;
1459 +       dst = dest;
1460 +       while ((c = *src++) != '\0') {
1461 +               if (c == '&')
1462 +                       no = 0;
1463 +               else if (c == '\\' && '0' <= *src && *src <= '9')
1464 +                       no = *src++ - '0';
1465 +               else
1466 +                       no = -1;
1467 +
1468 +               if (no < 0) {   /* Ordinary character. */
1469 +                       if (c == '\\' && (*src == '\\' || *src == '&'))
1470 +                               c = *src++;
1471 +                       *dst++ = c;
1472 +               } else if (prog->startp[no] != NULL && prog->endp[no] != NULL) {
1473 +                       len = prog->endp[no] - prog->startp[no];
1474 +                       (void) strncpy(dst, prog->startp[no], len);
1475 +                       dst += len;
1476 +                       if (len != 0 && *(dst-1) == '\0') {     /* strncpy hit NUL. */
1477 +                               regerror("damaged match string");
1478 +                               return;
1479 +                       }
1480 +               }
1481 +       }
1482 +       *dst++ = '\0';
1483 +}
1484 diff --git a/net/netfilter/xt_layer7.c b/net/netfilter/xt_layer7.c
1485 new file mode 100644
1486 index 0000000..ddf7fec
1487 --- /dev/null
1488 +++ b/net/netfilter/xt_layer7.c
1489 @@ -0,0 +1,683 @@
1490 +/*
1491 +  Kernel module to match application layer (OSI layer 7) data in connections.
1492 +
1493 +  http://l7-filter.sf.net
1494 +
1495 +  (C) 2003-2009 Matthew Strait and Ethan Sommer.
1496 +
1497 +  This program is free software; you can redistribute it and/or
1498 +  modify it under the terms of the GNU General Public License
1499 +  as published by the Free Software Foundation; either version
1500 +  2 of the License, or (at your option) any later version.
1501 +  http://www.gnu.org/licenses/gpl.txt
1502 +
1503 +  Based on ipt_string.c (C) 2000 Emmanuel Roger <winfield@freegates.be>,
1504 +  xt_helper.c (C) 2002 Harald Welte and cls_layer7.c (C) 2003 Matthew Strait,
1505 +  Ethan Sommer, Justin Levandoski.
1506 +*/
1507 +
1508 +#include <linux/spinlock.h>
1509 +#include <linux/version.h>
1510 +#include <net/ip.h>
1511 +#include <net/tcp.h>
1512 +#include <linux/module.h>
1513 +#include <linux/skbuff.h>
1514 +#include <linux/netfilter.h>
1515 +#include <net/netfilter/nf_conntrack.h>
1516 +#include <net/netfilter/nf_conntrack_core.h>
1517 +#if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 27)
1518 +#include <net/netfilter/nf_conntrack_extend.h>
1519 +#include <net/netfilter/nf_conntrack_acct.h>
1520 +#endif
1521 +#include <linux/netfilter/x_tables.h>
1522 +#include <linux/netfilter/xt_layer7.h>
1523 +#include <linux/ctype.h>
1524 +#include <linux/proc_fs.h>
1525 +
1526 +#include "regexp/regexp.c"
1527 +
1528 +MODULE_LICENSE("GPL");
1529 +MODULE_AUTHOR("Matthew Strait <quadong@users.sf.net>, Ethan Sommer <sommere@users.sf.net>, Arne Fitzenreiter <arne_f@ipfire.org>");
1530 +MODULE_DESCRIPTION("iptables application layer match module");
1531 +MODULE_ALIAS("ipt_layer7");
1532 +MODULE_VERSION("2.30");
1533 +
1534 +static int maxdatalen = 2048; // this is the default
1535 +module_param(maxdatalen, int, 0444);
1536 +MODULE_PARM_DESC(maxdatalen, "maximum bytes of data looked at by l7-filter");
1537 +#ifdef CONFIG_NETFILTER_XT_MATCH_LAYER7_DEBUG
1538 +       #define DPRINTK(format,args...) printk(format,##args)
1539 +#else
1540 +       #define DPRINTK(format,args...)
1541 +#endif
1542 +
1543 +/* Number of packets whose data we look at.
1544 +This can be modified through /proc/net/layer7_numpackets */
1545 +static int num_packets = 10;
1546 +
1547 +static struct pattern_cache {
1548 +       char * regex_string;
1549 +       regexp * pattern;
1550 +       struct pattern_cache * next;
1551 +} * first_pattern_cache = NULL;
1552 +
1553 +static struct proto_cache {
1554 +       char * proto_string;
1555 +       struct proto_cache * next;
1556 +} * first_proto_cache = NULL;
1557 +
1558 +DEFINE_SPINLOCK(l7_lock);
1559 +
1560 +static int total_acct_packets(struct nf_conn *ct)
1561 +{
1562 +#if LINUX_VERSION_CODE <= KERNEL_VERSION(2, 6, 26)
1563 +       BUG_ON(ct == NULL);
1564 +       return (ct->counters[IP_CT_DIR_ORIGINAL].packets + ct->counters[IP_CT_DIR_REPLY].packets);
1565 +#else
1566 +       struct nf_conn_counter *acct;
1567 +
1568 +       BUG_ON(ct == NULL);
1569 +       acct = nf_conn_acct_find(ct);
1570 +       if (!acct)
1571 +               return 0;
1572 +        return (atomic64_read(&acct[IP_CT_DIR_ORIGINAL].packets) + atomic64_read(&acct[IP_CT_DIR_REPLY].packets));
1573 +#endif
1574 +}
1575 +
1576 +#ifdef CONFIG_IP_NF_MATCH_LAYER7_DEBUG
1577 +/* Converts an unfriendly string into a friendly one by
1578 +replacing unprintables with periods and all whitespace with " ". */
1579 +static char * friendly_print(unsigned char * s)
1580 +{
1581 +       char * f = kmalloc(strlen(s) + 1, GFP_ATOMIC);
1582 +       int i;
1583 +
1584 +       if(!f) {
1585 +               if (net_ratelimit())
1586 +                       printk(KERN_ERR "layer7: out of memory in "
1587 +                                       "friendly_print, bailing.\n");
1588 +               return NULL;
1589 +       }
1590 +
1591 +       for(i = 0; i < strlen(s); i++){
1592 +               if(isprint(s[i]) && s[i] < 128) f[i] = s[i];
1593 +               else if(isspace(s[i]))          f[i] = ' ';
1594 +               else                            f[i] = '.';
1595 +       }
1596 +       f[i] = '\0';
1597 +       return f;
1598 +}
1599 +
1600 +static char dec2hex(int i)
1601 +{
1602 +       switch (i) {
1603 +               case 0 ... 9:
1604 +                       return (i + '0');
1605 +                       break;
1606 +               case 10 ... 15:
1607 +                       return (i - 10 + 'a');
1608 +                       break;
1609 +               default:
1610 +                       if (net_ratelimit())
1611 +                               printk("layer7: Problem in dec2hex\n");
1612 +                       return '\0';
1613 +       }
1614 +}
1615 +
1616 +static char * hex_print(unsigned char * s)
1617 +{
1618 +       char * g = kmalloc(strlen(s)*3 + 1, GFP_ATOMIC);
1619 +       int i;
1620 +
1621 +       if(!g) {
1622 +              if (net_ratelimit())
1623 +                       printk(KERN_ERR "layer7: out of memory in hex_print, "
1624 +                                       "bailing.\n");
1625 +              return NULL;
1626 +       }
1627 +
1628 +       for(i = 0; i < strlen(s); i++) {
1629 +               g[i*3    ] = dec2hex(s[i]/16);
1630 +               g[i*3 + 1] = dec2hex(s[i]%16);
1631 +               g[i*3 + 2] = ' ';
1632 +       }
1633 +       g[i*3] = '\0';
1634 +
1635 +       return g;
1636 +}
1637 +#endif // DEBUG
1638 +
1639 +/* Use instead of regcomp.  As we expect to be seeing the same regexps over and
1640 +over again, it make sense to cache the results. */
1641 +static regexp * compile_and_cache(const char * regex_string,
1642 +                                  const char * protocol)
1643 +{
1644 +       struct pattern_cache * node               = first_pattern_cache;
1645 +       struct pattern_cache * last_pattern_cache = first_pattern_cache;
1646 +       struct pattern_cache * tmp;
1647 +       unsigned int len;
1648 +
1649 +       while (node != NULL) {
1650 +               if (!strcmp(node->regex_string, regex_string))
1651 +               return node->pattern;
1652 +
1653 +               last_pattern_cache = node;/* points at the last non-NULL node */
1654 +               node = node->next;
1655 +       }
1656 +
1657 +       /* If we reach the end of the list, then we have not yet cached
1658 +          the pattern for this regex. Let's do that now.
1659 +          Be paranoid about running out of memory to avoid list corruption. */
1660 +       tmp = kmalloc(sizeof(struct pattern_cache), GFP_ATOMIC);
1661 +
1662 +       if(!tmp) {
1663 +               if (net_ratelimit())
1664 +                       printk(KERN_ERR "layer7: out of memory in "
1665 +                                       "compile_and_cache, bailing.\n");
1666 +               return NULL;
1667 +       }
1668 +
1669 +       tmp->regex_string  = kmalloc(strlen(regex_string) + 1, GFP_ATOMIC);
1670 +       tmp->pattern       = kmalloc(sizeof(struct regexp),    GFP_ATOMIC);
1671 +       tmp->next = NULL;
1672 +
1673 +       if(!tmp->regex_string || !tmp->pattern) {
1674 +               if (net_ratelimit())
1675 +                       printk(KERN_ERR "layer7: out of memory in "
1676 +                                       "compile_and_cache, bailing.\n");
1677 +               kfree(tmp->regex_string);
1678 +               kfree(tmp->pattern);
1679 +               kfree(tmp);
1680 +               return NULL;
1681 +       }
1682 +
1683 +       /* Ok.  The new node is all ready now. */
1684 +       node = tmp;
1685 +
1686 +       if(first_pattern_cache == NULL) /* list is empty */
1687 +               first_pattern_cache = node; /* make node the beginning */
1688 +       else
1689 +               last_pattern_cache->next = node; /* attach node to the end */
1690 +
1691 +       /* copy the string and compile the regex */
1692 +       len = strlen(regex_string);
1693 +       DPRINTK("About to compile this: \"%s\"\n", regex_string);
1694 +       node->pattern = regcomp((char *)regex_string, &len);
1695 +       if ( !node->pattern ) {
1696 +               if (net_ratelimit())
1697 +                       printk(KERN_ERR "layer7: Error compiling regexp "
1698 +                                       "\"%s\" (%s)\n",
1699 +                                       regex_string, protocol);
1700 +               /* pattern is now cached as NULL, so we won't try again. */
1701 +       }
1702 +
1703 +       strcpy(node->regex_string, regex_string);
1704 +       return node->pattern;
1705 +}
1706 +
1707 +static char * get_protostr_ptr(const char * protocol)
1708 +{
1709 +       struct proto_cache * node             = first_proto_cache;
1710 +       struct proto_cache * last_proto_cache = first_proto_cache;
1711 +       struct proto_cache * tmp;
1712 +
1713 +       while (node != NULL) {
1714 +               if (!strcmp(node->proto_string, protocol))
1715 +               return node->proto_string;
1716 +
1717 +               last_proto_cache = node;/* points at the last non-NULL node */
1718 +               node = node->next;
1719 +       }
1720 +
1721 +       /* If we reach the end of the list, then we have not yet cached protocol
1722 +          Be paranoid about running out of memory to avoid list corruption. */
1723 +       tmp = kmalloc(sizeof(struct proto_cache), GFP_ATOMIC);
1724 +
1725 +       if(!tmp) {
1726 +               if (net_ratelimit())
1727 +                       printk(KERN_ERR "layer7: out of memory in "
1728 +                                       "proto_cache add, bailing.\n");
1729 +               return NULL;
1730 +       }
1731 +
1732 +       tmp->proto_string = kmalloc(strlen(protocol) + 1   , GFP_ATOMIC);
1733 +       tmp->next = NULL;
1734 +
1735 +       if(!tmp->proto_string) {
1736 +               if (net_ratelimit())
1737 +                       printk(KERN_ERR "layer7: out of memory in "
1738 +                                       "proto_cache add, bailing.\n");
1739 +               kfree(tmp->proto_string);
1740 +               kfree(tmp);
1741 +               return NULL;
1742 +       }
1743 +
1744 +       /* Ok.  The new node is all ready now. */
1745 +       node = tmp;
1746 +
1747 +       if(first_proto_cache == NULL) /* list is empty */
1748 +               first_proto_cache = node; /* make node the beginning */
1749 +       else
1750 +               last_proto_cache->next = node; /* attach node to the end */
1751 +
1752 +       strcpy(node->proto_string, protocol);
1753 +       return node->proto_string;
1754 +}
1755 +
1756 +static int can_handle(const struct sk_buff *skb)
1757 +{
1758 +       if(!ip_hdr(skb)) /* not IP */
1759 +               return 0;
1760 +       if(ip_hdr(skb)->protocol != IPPROTO_TCP &&
1761 +          ip_hdr(skb)->protocol != IPPROTO_UDP &&
1762 +          ip_hdr(skb)->protocol != IPPROTO_ICMP)
1763 +               return 0;
1764 +       return 1;
1765 +}
1766 +
1767 +/* Returns offset the into the skb->data that the application data starts */
1768 +static int app_data_offset(const struct sk_buff *skb)
1769 +{
1770 +       /* In case we are ported somewhere (ebtables?) where ip_hdr(skb)
1771 +       isn't set, this can be gotten from 4*(skb->data[0] & 0x0f) as well. */
1772 +       int ip_hl = 4*ip_hdr(skb)->ihl;
1773 +
1774 +       if( ip_hdr(skb)->protocol == IPPROTO_TCP ) {
1775 +               /* 12 == offset into TCP header for the header length field.
1776 +               Can't get this with skb->h.th->doff because the tcphdr
1777 +               struct doesn't get set when routing (this is confirmed to be
1778 +               true in Netfilter as well as QoS.) */
1779 +               int tcp_hl = 4*(skb->data[ip_hl + 12] >> 4);
1780 +
1781 +               return ip_hl + tcp_hl;
1782 +       } else if( ip_hdr(skb)->protocol == IPPROTO_UDP  ) {
1783 +               return ip_hl + 8; /* UDP header is always 8 bytes */
1784 +       } else if( ip_hdr(skb)->protocol == IPPROTO_ICMP ) {
1785 +               return ip_hl + 8; /* ICMP header is 8 bytes */
1786 +       } else {
1787 +               if (net_ratelimit())
1788 +                       printk(KERN_ERR "layer7: tried to handle unknown "
1789 +                                       "protocol!\n");
1790 +               return ip_hl + 8; /* something reasonable */
1791 +       }
1792 +}
1793 +
1794 +/* handles whether there's a match when we aren't appending data anymore */
1795 +static int match_no_append(struct nf_conn * conntrack,
1796 +                           struct nf_conn * master_conntrack,
1797 +                           enum ip_conntrack_info ctinfo,
1798 +                           enum ip_conntrack_info master_ctinfo,
1799 +                           const struct xt_layer7_info * info)
1800 +{
1801 +       /* If we're in here, throw the app data away */
1802 +       if(master_conntrack->layer7.app_data != NULL) {
1803 +
1804 +       #ifdef CONFIG_IP_NF_MATCH_LAYER7_DEBUG
1805 +               if(!master_conntrack->layer7.app_proto) {
1806 +                       char * f =
1807 +                         friendly_print(master_conntrack->layer7.app_data);
1808 +                       char * g =
1809 +                         hex_print(master_conntrack->layer7.app_data);
1810 +                       DPRINTK("\nl7-filter gave up after %d bytes "
1811 +                               "(%d packets):\n%s\n",
1812 +                               strlen(f), total_acct_packets(master_conntrack), f);
1813 +                       kfree(f);
1814 +                       DPRINTK("In hex: %s\n", g);
1815 +                       kfree(g);
1816 +               }
1817 +       #endif
1818 +
1819 +               kfree(master_conntrack->layer7.app_data);
1820 +               master_conntrack->layer7.app_data = NULL; /* don't free again */
1821 +       }
1822 +
1823 +       if(master_conntrack->layer7.app_proto){
1824 +               /* Here child connections set their .app_proto (for /proc) */
1825 +               if(!conntrack->layer7.app_proto) {
1826 +                       conntrack->layer7.app_proto = master_conntrack->layer7.app_proto;
1827 +               }
1828 +
1829 +               return (!strcmp(master_conntrack->layer7.app_proto,
1830 +                               info->protocol));
1831 +       }
1832 +       else {
1833 +               /* If not classified, set to "unknown" to distinguish from
1834 +               connections that are still being tested. */
1835 +               master_conntrack->layer7.app_proto = get_protostr_ptr("unknown");
1836 +               return 0;
1837 +       }
1838 +}
1839 +
1840 +/* add the new app data to the conntrack.  Return number of bytes added. */
1841 +static int add_data(struct nf_conn * master_conntrack,
1842 +                    char * app_data, int appdatalen)
1843 +{
1844 +       int length = 0, i;
1845 +       int oldlength = master_conntrack->layer7.app_data_len;
1846 +
1847 +       /* This is a fix for a race condition by Deti Fliegl. However, I'm not
1848 +          clear on whether the race condition exists or whether this really
1849 +          fixes it.  I might just be being dense... Anyway, if it's not really
1850 +          a fix, all it does is waste a very small amount of time. */
1851 +       if(!master_conntrack->layer7.app_data) return 0;
1852 +
1853 +       /* Strip nulls. Make everything lower case (our regex lib doesn't
1854 +       do case insensitivity).  Add it to the end of the current data. */
1855 +       for(i = 0; i < maxdatalen-oldlength-1 &&
1856 +                  i < appdatalen; i++) {
1857 +               if(app_data[i] != '\0') {
1858 +                       /* the kernel version of tolower mungs 'upper ascii' */
1859 +                       master_conntrack->layer7.app_data[length+oldlength] =
1860 +                               isascii(app_data[i])?
1861 +                                       tolower(app_data[i]) : app_data[i];
1862 +                       length++;
1863 +               }
1864 +       }
1865 +
1866 +       master_conntrack->layer7.app_data[length+oldlength] = '\0';
1867 +       master_conntrack->layer7.app_data_len = length + oldlength;
1868 +
1869 +       return length;
1870 +}
1871 +
1872 +/* taken from drivers/video/modedb.c */
1873 +static int my_atoi(const char *s)
1874 +{
1875 +       int val = 0;
1876 +
1877 +       for (;; s++) {
1878 +               switch (*s) {
1879 +                       case '0'...'9':
1880 +                       val = 10*val+(*s-'0');
1881 +                       break;
1882 +               default:
1883 +                       return val;
1884 +               }
1885 +       }
1886 +}
1887 +
1888 +static int layer7_numpackets_proc_show(struct seq_file *s, void *p) {
1889 +       seq_printf(s, "%d\n", num_packets);
1890 +
1891 +       return 0;
1892 +}
1893 +
1894 +static int layer7_numpackets_proc_open(struct inode *inode, struct file *file) {
1895 +       return single_open(file, layer7_numpackets_proc_show, NULL);
1896 +}
1897 +
1898 +/* Read in num_packets from userland */
1899 +static ssize_t layer7_numpackets_write_proc(struct file* file, const char __user *buffer,
1900 +               size_t count, loff_t *data) {
1901 +       char value[1024];
1902 +       int new_num_packets;
1903 +
1904 +       if (copy_from_user(&value, buffer, sizeof(value)))
1905 +               return -EFAULT;
1906 +
1907 +       new_num_packets = my_atoi(value);
1908 +
1909 +       if ((new_num_packets < 1) || (new_num_packets > 99)) {
1910 +               printk(KERN_WARNING "layer7: numpackets must be between 1 and 99\n");
1911 +               return -EFAULT;
1912 +       }
1913 +
1914 +       num_packets = new_num_packets;
1915 +
1916 +       return count;
1917 +}
1918 +
1919 +static bool
1920 +#if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 35)
1921 +match(const struct sk_buff *skbin, struct xt_action_param *par)
1922 +#elif LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 28)
1923 +match(const struct sk_buff *skbin, const struct xt_match_param *par)
1924 +#else
1925 +match(const struct sk_buff *skbin,
1926 +      const struct net_device *in,
1927 +      const struct net_device *out,
1928 +      const struct xt_match *match,
1929 +      const void *matchinfo,
1930 +      int offset,
1931 +      unsigned int protoff,
1932 +      bool *hotdrop)
1933 +#endif
1934 +{
1935 +       /* sidestep const without getting a compiler warning... */
1936 +       struct sk_buff * skb = (struct sk_buff *)skbin;
1937 +
1938 +       const struct xt_layer7_info * info =
1939 +       #if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 28)
1940 +               par->matchinfo;
1941 +       #else
1942 +               matchinfo;
1943 +       #endif
1944 +
1945 +       enum ip_conntrack_info master_ctinfo, ctinfo;
1946 +       struct nf_conn *master_conntrack, *conntrack;
1947 +       unsigned char * app_data;
1948 +       unsigned int pattern_result, appdatalen;
1949 +       regexp * comppattern;
1950 +
1951 +       /* Be paranoid/incompetent - lock the entire match function. */
1952 +       spin_lock_bh(&l7_lock);
1953 +
1954 +       if(!can_handle(skb)){
1955 +               DPRINTK("layer7: This is some protocol I can't handle.\n");
1956 +               spin_unlock_bh(&l7_lock);
1957 +               return info->invert;
1958 +       }
1959 +
1960 +       /* Treat parent & all its children together as one connection, except
1961 +       for the purpose of setting conntrack->layer7.app_proto in the actual
1962 +       connection. This makes /proc/net/ip_conntrack more satisfying. */
1963 +       if(!(conntrack = nf_ct_get(skb, &ctinfo)) ||
1964 +          !(master_conntrack=nf_ct_get(skb,&master_ctinfo))){
1965 +               DPRINTK("layer7: couldn't get conntrack.\n");
1966 +               spin_unlock_bh(&l7_lock);
1967 +               return info->invert;
1968 +       }
1969 +
1970 +       /* Try to get a master conntrack (and its master etc) for FTP, etc. */
1971 +       while (master_ct(master_conntrack) != NULL)
1972 +               master_conntrack = master_ct(master_conntrack);
1973 +
1974 +       /* if we've classified it or seen too many packets */
1975 +       if(total_acct_packets(master_conntrack) > num_packets ||
1976 +          master_conntrack->layer7.app_proto) {
1977 +
1978 +               pattern_result = match_no_append(conntrack, master_conntrack,
1979 +                                                ctinfo, master_ctinfo, info);
1980 +
1981 +               /* skb->cb[0] == seen. Don't do things twice if there are
1982 +               multiple l7 rules. I'm not sure that using cb for this purpose
1983 +               is correct, even though it says "put your private variables
1984 +               there". But it doesn't look like it is being used for anything
1985 +               else in the skbs that make it here. */
1986 +               skb->cb[0] = 1; /* marking it seen here's probably irrelevant */
1987 +
1988 +               spin_unlock_bh(&l7_lock);
1989 +               return (pattern_result ^ info->invert);
1990 +       }
1991 +
1992 +       if(skb_is_nonlinear(skb)){
1993 +               if(skb_linearize(skb) != 0){
1994 +                       if (net_ratelimit())
1995 +                               printk(KERN_ERR "layer7: failed to linearize "
1996 +                                               "packet, bailing.\n");
1997 +                       spin_unlock_bh(&l7_lock);
1998 +                       return info->invert;
1999 +               }
2000 +       }
2001 +
2002 +       /* now that the skb is linearized, it's safe to set these. */
2003 +       app_data = skb->data + app_data_offset(skb);
2004 +       appdatalen = skb_tail_pointer(skb) - app_data;
2005 +
2006 +       /* the return value gets checked later, when we're ready to use it */
2007 +       comppattern = compile_and_cache(info->pattern, info->protocol);
2008 +
2009 +       /* On the first packet of a connection, allocate space for app data */
2010 +       if(total_acct_packets(master_conntrack) == 1 && !skb->cb[0] &&
2011 +          !master_conntrack->layer7.app_data){
2012 +               master_conntrack->layer7.app_data =
2013 +                       kmalloc(maxdatalen, GFP_ATOMIC);
2014 +               if(!master_conntrack->layer7.app_data){
2015 +                       if (net_ratelimit())
2016 +                               printk(KERN_ERR "layer7: out of memory in "
2017 +                                               "match, bailing.\n");
2018 +                       spin_unlock_bh(&l7_lock);
2019 +                       return info->invert;
2020 +               }
2021 +
2022 +               master_conntrack->layer7.app_data[0] = '\0';
2023 +       }
2024 +
2025 +       /* Can be here, but unallocated, if numpackets is increased near
2026 +       the beginning of a connection */
2027 +       if(master_conntrack->layer7.app_data == NULL){
2028 +               spin_unlock_bh(&l7_lock);
2029 +               return info->invert; /* unmatched */
2030 +       }
2031 +
2032 +       if(!skb->cb[0]){
2033 +               int newbytes;
2034 +               newbytes = add_data(master_conntrack, app_data, appdatalen);
2035 +               if(newbytes == 0) { /* didn't add any data */
2036 +                       skb->cb[0] = 1;
2037 +                       /* Didn't match before, not going to match now */
2038 +                       spin_unlock_bh(&l7_lock);
2039 +                       return info->invert;
2040 +               }
2041 +       }
2042 +
2043 +       /* If looking for "unknown", then never match.  "Unknown" means that
2044 +       we've given up; we're still trying with these packets. */
2045 +       if(!strcmp(info->protocol, "unknown")) {
2046 +               pattern_result = 0;
2047 +       /* If looking for "unset", then always match. "Unset" means that we
2048 +       haven't yet classified the connection. */
2049 +       } else if(!strcmp(info->protocol, "unset")) {
2050 +               pattern_result = 2;
2051 +               DPRINTK("layer7: matched unset: not yet classified "
2052 +                       "(%d/%d packets)\n",
2053 +                        total_acct_packets(master_conntrack), num_packets);
2054 +       /* If the regexp failed to compile, don't bother running it */
2055 +       } else if(comppattern &&
2056 +                 regexec(comppattern, master_conntrack->layer7.app_data)){
2057 +               DPRINTK("layer7: matched %s\n", info->protocol);
2058 +               pattern_result = 1;
2059 +       } else pattern_result = 0;
2060 +
2061 +       if(pattern_result == 1) {
2062 +               master_conntrack->layer7.app_proto=get_protostr_ptr(info->protocol);
2063 +       } else if(pattern_result > 1) { /* cleanup from "unset" */
2064 +               pattern_result = 1;
2065 +       }
2066 +
2067 +       /* mark the packet seen */
2068 +       skb->cb[0] = 1;
2069 +
2070 +       spin_unlock_bh(&l7_lock);
2071 +       return (pattern_result ^ info->invert);
2072 +}
2073 +
2074 +// load nf_conntrack_ipv4
2075 +#if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 35)
2076 +static int
2077 +#else
2078 +static bool
2079 +#endif
2080 +#if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 28)
2081 +check(const struct xt_mtchk_param *par)
2082 +{
2083 +        if (nf_ct_l3proto_try_module_get(par->match->family) < 0) {
2084 +                printk(KERN_WARNING "can't load conntrack support for "
2085 +                                    "proto=%d\n", par->match->family);
2086 +#else
2087 +check(const char *tablename, const void *inf,
2088 +                const struct xt_match *match, void *matchinfo,
2089 +                unsigned int hook_mask)
2090 +{
2091 +        if (nf_ct_l3proto_try_module_get(match->family) < 0) {
2092 +                printk(KERN_WARNING "can't load conntrack support for "
2093 +                                    "proto=%d\n", match->family);
2094 +#endif
2095 +#if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 35)
2096 +               return -EINVAL;
2097 +       }
2098 +       return 0;
2099 +#else
2100 +                return 0;
2101 +        }
2102 +       return 1;
2103 +#endif
2104 +}
2105 +
2106 +
2107 +#if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 28)
2108 +       static void destroy(const struct xt_mtdtor_param *par)
2109 +       {
2110 +               nf_ct_l3proto_module_put(par->match->family);
2111 +       }
2112 +#else
2113 +       static void destroy(const struct xt_match *match, void *matchinfo)
2114 +       {
2115 +               nf_ct_l3proto_module_put(match->family);
2116 +       }
2117 +#endif
2118 +
2119 +static struct xt_match xt_layer7_match[] __read_mostly = {
2120 +{
2121 +       .name           = "layer7",
2122 +       .family         = NFPROTO_IPV4,
2123 +       .checkentry     = check,
2124 +       .match          = match,
2125 +       .destroy        = destroy,
2126 +       .matchsize      = sizeof(struct xt_layer7_info),
2127 +       .me             = THIS_MODULE
2128 +}
2129 +};
2130 +
2131 +static const struct file_operations layer7_numpackets_proc_fops = {
2132 +       .owner   = THIS_MODULE,
2133 +       .open    = layer7_numpackets_proc_open,
2134 +       .read    = seq_read,
2135 +       .llseek  = seq_lseek,
2136 +       .release = single_release,
2137 +       .write   = layer7_numpackets_write_proc,
2138 +};
2139 +
2140 +static int __init xt_layer7_init(void)
2141 +{
2142 +       need_conntrack();
2143 +
2144 +       // Register proc interface
2145 +       proc_create_data("layer7_numpackets", 0644,
2146 +               init_net.proc_net, &layer7_numpackets_proc_fops, NULL);
2147 +
2148 +       if(maxdatalen < 1) {
2149 +               printk(KERN_WARNING "layer7: maxdatalen can't be < 1, "
2150 +                       "using 1\n");
2151 +               maxdatalen = 1;
2152 +       }
2153 +       /* This is not a hard limit.  It's just here to prevent people from
2154 +       bringing their slow machines to a grinding halt. */
2155 +       else if(maxdatalen > 65536) {
2156 +               printk(KERN_WARNING "layer7: maxdatalen can't be > 65536, "
2157 +                       "using 65536\n");
2158 +               maxdatalen = 65536;
2159 +       }
2160 +       return xt_register_matches(xt_layer7_match,
2161 +                                  ARRAY_SIZE(xt_layer7_match));
2162 +}
2163 +
2164 +static void __exit xt_layer7_fini(void)
2165 +{
2166 +       remove_proc_entry("layer7_numpackets", init_net.proc_net);
2167 +       xt_unregister_matches(xt_layer7_match, ARRAY_SIZE(xt_layer7_match));
2168 +}
2169 +
2170 +module_init(xt_layer7_init);
2171 +module_exit(xt_layer7_fini);
2172 +