Back to index

nagios-plugins  1.4.16
base64.c
Go to the documentation of this file.
00001 /* base64.c -- Encode binary data using printable characters.
00002    Copyright (C) 1999, 2000, 2001, 2004, 2005, 2006, 2009, 2010 Free Software
00003    Foundation, Inc.
00004 
00005    This program is free software; you can redistribute it and/or modify
00006    it under the terms of the GNU General Public License as published by
00007    the Free Software Foundation; either version 3, or (at your option)
00008    any later version.
00009 
00010    This program is distributed in the hope that it will be useful,
00011    but WITHOUT ANY WARRANTY; without even the implied warranty of
00012    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
00013    GNU General Public License for more details.
00014 
00015    You should have received a copy of the GNU General Public License
00016    along with this program; if not, write to the Free Software Foundation,
00017    Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.  */
00018 
00019 /* Written by Simon Josefsson.  Partially adapted from GNU MailUtils
00020  * (mailbox/filter_trans.c, as of 2004-11-28).  Improved by review
00021  * from Paul Eggert, Bruno Haible, and Stepan Kasal.
00022  *
00023  * See also RFC 3548 <http://www.ietf.org/rfc/rfc3548.txt>.
00024  *
00025  * Be careful with error checking.  Here is how you would typically
00026  * use these functions:
00027  *
00028  * bool ok = base64_decode_alloc (in, inlen, &out, &outlen);
00029  * if (!ok)
00030  *   FAIL: input was not valid base64
00031  * if (out == NULL)
00032  *   FAIL: memory allocation error
00033  * OK: data in OUT/OUTLEN
00034  *
00035  * size_t outlen = base64_encode_alloc (in, inlen, &out);
00036  * if (out == NULL && outlen == 0 && inlen != 0)
00037  *   FAIL: input too long
00038  * if (out == NULL)
00039  *   FAIL: memory allocation error
00040  * OK: data in OUT/OUTLEN.
00041  *
00042  */
00043 
00044 #include <config.h>
00045 
00046 /* Get prototype. */
00047 #include "base64.h"
00048 
00049 /* Get malloc. */
00050 #include <stdlib.h>
00051 
00052 /* Get UCHAR_MAX. */
00053 #include <limits.h>
00054 
00055 #include <string.h>
00056 
00057 /* C89 compliant way to cast 'char' to 'unsigned char'. */
00058 static inline unsigned char
00059 to_uchar (char ch)
00060 {
00061   return ch;
00062 }
00063 
00064 /* Base64 encode IN array of size INLEN into OUT array of size OUTLEN.
00065    If OUTLEN is less than BASE64_LENGTH(INLEN), write as many bytes as
00066    possible.  If OUTLEN is larger than BASE64_LENGTH(INLEN), also zero
00067    terminate the output buffer. */
00068 void
00069 base64_encode (const char *restrict in, size_t inlen,
00070                char *restrict out, size_t outlen)
00071 {
00072   static const char b64str[64] =
00073     "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
00074 
00075   while (inlen && outlen)
00076     {
00077       *out++ = b64str[(to_uchar (in[0]) >> 2) & 0x3f];
00078       if (!--outlen)
00079         break;
00080       *out++ = b64str[((to_uchar (in[0]) << 4)
00081                        + (--inlen ? to_uchar (in[1]) >> 4 : 0))
00082                       & 0x3f];
00083       if (!--outlen)
00084         break;
00085       *out++ =
00086         (inlen
00087          ? b64str[((to_uchar (in[1]) << 2)
00088                    + (--inlen ? to_uchar (in[2]) >> 6 : 0))
00089                   & 0x3f]
00090          : '=');
00091       if (!--outlen)
00092         break;
00093       *out++ = inlen ? b64str[to_uchar (in[2]) & 0x3f] : '=';
00094       if (!--outlen)
00095         break;
00096       if (inlen)
00097         inlen--;
00098       if (inlen)
00099         in += 3;
00100     }
00101 
00102   if (outlen)
00103     *out = '\0';
00104 }
00105 
00106 /* Allocate a buffer and store zero terminated base64 encoded data
00107    from array IN of size INLEN, returning BASE64_LENGTH(INLEN), i.e.,
00108    the length of the encoded data, excluding the terminating zero.  On
00109    return, the OUT variable will hold a pointer to newly allocated
00110    memory that must be deallocated by the caller.  If output string
00111    length would overflow, 0 is returned and OUT is set to NULL.  If
00112    memory allocation failed, OUT is set to NULL, and the return value
00113    indicates length of the requested memory block, i.e.,
00114    BASE64_LENGTH(inlen) + 1. */
00115 size_t
00116 base64_encode_alloc (const char *in, size_t inlen, char **out)
00117 {
00118   size_t outlen = 1 + BASE64_LENGTH (inlen);
00119 
00120   /* Check for overflow in outlen computation.
00121    *
00122    * If there is no overflow, outlen >= inlen.
00123    *
00124    * If the operation (inlen + 2) overflows then it yields at most +1, so
00125    * outlen is 0.
00126    *
00127    * If the multiplication overflows, we lose at least half of the
00128    * correct value, so the result is < ((inlen + 2) / 3) * 2, which is
00129    * less than (inlen + 2) * 0.66667, which is less than inlen as soon as
00130    * (inlen > 4).
00131    */
00132   if (inlen > outlen)
00133     {
00134       *out = NULL;
00135       return 0;
00136     }
00137 
00138   *out = malloc (outlen);
00139   if (!*out)
00140     return outlen;
00141 
00142   base64_encode (in, inlen, *out, outlen);
00143 
00144   return outlen - 1;
00145 }
00146 
00147 /* With this approach this file works independent of the charset used
00148    (think EBCDIC).  However, it does assume that the characters in the
00149    Base64 alphabet (A-Za-z0-9+/) are encoded in 0..255.  POSIX
00150    1003.1-2001 require that char and unsigned char are 8-bit
00151    quantities, though, taking care of that problem.  But this may be a
00152    potential problem on non-POSIX C99 platforms.
00153 
00154    IBM C V6 for AIX mishandles "#define B64(x) ...'x'...", so use "_"
00155    as the formal parameter rather than "x".  */
00156 #define B64(_)                                  \
00157   ((_) == 'A' ? 0                               \
00158    : (_) == 'B' ? 1                             \
00159    : (_) == 'C' ? 2                             \
00160    : (_) == 'D' ? 3                             \
00161    : (_) == 'E' ? 4                             \
00162    : (_) == 'F' ? 5                             \
00163    : (_) == 'G' ? 6                             \
00164    : (_) == 'H' ? 7                             \
00165    : (_) == 'I' ? 8                             \
00166    : (_) == 'J' ? 9                             \
00167    : (_) == 'K' ? 10                            \
00168    : (_) == 'L' ? 11                            \
00169    : (_) == 'M' ? 12                            \
00170    : (_) == 'N' ? 13                            \
00171    : (_) == 'O' ? 14                            \
00172    : (_) == 'P' ? 15                            \
00173    : (_) == 'Q' ? 16                            \
00174    : (_) == 'R' ? 17                            \
00175    : (_) == 'S' ? 18                            \
00176    : (_) == 'T' ? 19                            \
00177    : (_) == 'U' ? 20                            \
00178    : (_) == 'V' ? 21                            \
00179    : (_) == 'W' ? 22                            \
00180    : (_) == 'X' ? 23                            \
00181    : (_) == 'Y' ? 24                            \
00182    : (_) == 'Z' ? 25                            \
00183    : (_) == 'a' ? 26                            \
00184    : (_) == 'b' ? 27                            \
00185    : (_) == 'c' ? 28                            \
00186    : (_) == 'd' ? 29                            \
00187    : (_) == 'e' ? 30                            \
00188    : (_) == 'f' ? 31                            \
00189    : (_) == 'g' ? 32                            \
00190    : (_) == 'h' ? 33                            \
00191    : (_) == 'i' ? 34                            \
00192    : (_) == 'j' ? 35                            \
00193    : (_) == 'k' ? 36                            \
00194    : (_) == 'l' ? 37                            \
00195    : (_) == 'm' ? 38                            \
00196    : (_) == 'n' ? 39                            \
00197    : (_) == 'o' ? 40                            \
00198    : (_) == 'p' ? 41                            \
00199    : (_) == 'q' ? 42                            \
00200    : (_) == 'r' ? 43                            \
00201    : (_) == 's' ? 44                            \
00202    : (_) == 't' ? 45                            \
00203    : (_) == 'u' ? 46                            \
00204    : (_) == 'v' ? 47                            \
00205    : (_) == 'w' ? 48                            \
00206    : (_) == 'x' ? 49                            \
00207    : (_) == 'y' ? 50                            \
00208    : (_) == 'z' ? 51                            \
00209    : (_) == '0' ? 52                            \
00210    : (_) == '1' ? 53                            \
00211    : (_) == '2' ? 54                            \
00212    : (_) == '3' ? 55                            \
00213    : (_) == '4' ? 56                            \
00214    : (_) == '5' ? 57                            \
00215    : (_) == '6' ? 58                            \
00216    : (_) == '7' ? 59                            \
00217    : (_) == '8' ? 60                            \
00218    : (_) == '9' ? 61                            \
00219    : (_) == '+' ? 62                            \
00220    : (_) == '/' ? 63                            \
00221    : -1)
00222 
00223 static const signed char b64[0x100] = {
00224   B64 (0), B64 (1), B64 (2), B64 (3),
00225   B64 (4), B64 (5), B64 (6), B64 (7),
00226   B64 (8), B64 (9), B64 (10), B64 (11),
00227   B64 (12), B64 (13), B64 (14), B64 (15),
00228   B64 (16), B64 (17), B64 (18), B64 (19),
00229   B64 (20), B64 (21), B64 (22), B64 (23),
00230   B64 (24), B64 (25), B64 (26), B64 (27),
00231   B64 (28), B64 (29), B64 (30), B64 (31),
00232   B64 (32), B64 (33), B64 (34), B64 (35),
00233   B64 (36), B64 (37), B64 (38), B64 (39),
00234   B64 (40), B64 (41), B64 (42), B64 (43),
00235   B64 (44), B64 (45), B64 (46), B64 (47),
00236   B64 (48), B64 (49), B64 (50), B64 (51),
00237   B64 (52), B64 (53), B64 (54), B64 (55),
00238   B64 (56), B64 (57), B64 (58), B64 (59),
00239   B64 (60), B64 (61), B64 (62), B64 (63),
00240   B64 (64), B64 (65), B64 (66), B64 (67),
00241   B64 (68), B64 (69), B64 (70), B64 (71),
00242   B64 (72), B64 (73), B64 (74), B64 (75),
00243   B64 (76), B64 (77), B64 (78), B64 (79),
00244   B64 (80), B64 (81), B64 (82), B64 (83),
00245   B64 (84), B64 (85), B64 (86), B64 (87),
00246   B64 (88), B64 (89), B64 (90), B64 (91),
00247   B64 (92), B64 (93), B64 (94), B64 (95),
00248   B64 (96), B64 (97), B64 (98), B64 (99),
00249   B64 (100), B64 (101), B64 (102), B64 (103),
00250   B64 (104), B64 (105), B64 (106), B64 (107),
00251   B64 (108), B64 (109), B64 (110), B64 (111),
00252   B64 (112), B64 (113), B64 (114), B64 (115),
00253   B64 (116), B64 (117), B64 (118), B64 (119),
00254   B64 (120), B64 (121), B64 (122), B64 (123),
00255   B64 (124), B64 (125), B64 (126), B64 (127),
00256   B64 (128), B64 (129), B64 (130), B64 (131),
00257   B64 (132), B64 (133), B64 (134), B64 (135),
00258   B64 (136), B64 (137), B64 (138), B64 (139),
00259   B64 (140), B64 (141), B64 (142), B64 (143),
00260   B64 (144), B64 (145), B64 (146), B64 (147),
00261   B64 (148), B64 (149), B64 (150), B64 (151),
00262   B64 (152), B64 (153), B64 (154), B64 (155),
00263   B64 (156), B64 (157), B64 (158), B64 (159),
00264   B64 (160), B64 (161), B64 (162), B64 (163),
00265   B64 (164), B64 (165), B64 (166), B64 (167),
00266   B64 (168), B64 (169), B64 (170), B64 (171),
00267   B64 (172), B64 (173), B64 (174), B64 (175),
00268   B64 (176), B64 (177), B64 (178), B64 (179),
00269   B64 (180), B64 (181), B64 (182), B64 (183),
00270   B64 (184), B64 (185), B64 (186), B64 (187),
00271   B64 (188), B64 (189), B64 (190), B64 (191),
00272   B64 (192), B64 (193), B64 (194), B64 (195),
00273   B64 (196), B64 (197), B64 (198), B64 (199),
00274   B64 (200), B64 (201), B64 (202), B64 (203),
00275   B64 (204), B64 (205), B64 (206), B64 (207),
00276   B64 (208), B64 (209), B64 (210), B64 (211),
00277   B64 (212), B64 (213), B64 (214), B64 (215),
00278   B64 (216), B64 (217), B64 (218), B64 (219),
00279   B64 (220), B64 (221), B64 (222), B64 (223),
00280   B64 (224), B64 (225), B64 (226), B64 (227),
00281   B64 (228), B64 (229), B64 (230), B64 (231),
00282   B64 (232), B64 (233), B64 (234), B64 (235),
00283   B64 (236), B64 (237), B64 (238), B64 (239),
00284   B64 (240), B64 (241), B64 (242), B64 (243),
00285   B64 (244), B64 (245), B64 (246), B64 (247),
00286   B64 (248), B64 (249), B64 (250), B64 (251),
00287   B64 (252), B64 (253), B64 (254), B64 (255)
00288 };
00289 
00290 #if UCHAR_MAX == 255
00291 # define uchar_in_range(c) true
00292 #else
00293 # define uchar_in_range(c) ((c) <= 255)
00294 #endif
00295 
00296 /* Return true if CH is a character from the Base64 alphabet, and
00297    false otherwise.  Note that '=' is padding and not considered to be
00298    part of the alphabet.  */
00299 bool
00300 isbase64 (char ch)
00301 {
00302   return uchar_in_range (to_uchar (ch)) && 0 <= b64[to_uchar (ch)];
00303 }
00304 
00305 /* Initialize decode-context buffer, CTX.  */
00306 void
00307 base64_decode_ctx_init (struct base64_decode_context *ctx)
00308 {
00309   ctx->i = 0;
00310 }
00311 
00312 /* If CTX->i is 0 or 4, there are four or more bytes in [*IN..IN_END), and
00313    none of those four is a newline, then return *IN.  Otherwise, copy up to
00314    4 - CTX->i non-newline bytes from that range into CTX->buf, starting at
00315    index CTX->i and setting CTX->i to reflect the number of bytes copied,
00316    and return CTX->buf.  In either case, advance *IN to point to the byte
00317    after the last one processed, and set *N_NON_NEWLINE to the number of
00318    verified non-newline bytes accessible through the returned pointer.  */
00319 static inline char *
00320 get_4 (struct base64_decode_context *ctx,
00321        char const *restrict *in, char const *restrict in_end,
00322        size_t *n_non_newline)
00323 {
00324   if (ctx->i == 4)
00325     ctx->i = 0;
00326 
00327   if (ctx->i == 0)
00328     {
00329       char const *t = *in;
00330       if (4 <= in_end - *in && memchr (t, '\n', 4) == NULL)
00331         {
00332           /* This is the common case: no newline.  */
00333           *in += 4;
00334           *n_non_newline = 4;
00335           return (char *) t;
00336         }
00337     }
00338 
00339   {
00340     /* Copy non-newline bytes into BUF.  */
00341     char const *p = *in;
00342     while (p < in_end)
00343       {
00344         char c = *p++;
00345         if (c != '\n')
00346           {
00347             ctx->buf[ctx->i++] = c;
00348             if (ctx->i == 4)
00349               break;
00350           }
00351       }
00352 
00353     *in = p;
00354     *n_non_newline = ctx->i;
00355     return ctx->buf;
00356   }
00357 }
00358 
00359 #define return_false                            \
00360   do                                            \
00361     {                                           \
00362       *outp = out;                              \
00363       return false;                             \
00364     }                                           \
00365   while (false)
00366 
00367 /* Decode up to four bytes of base64-encoded data, IN, of length INLEN
00368    into the output buffer, *OUT, of size *OUTLEN bytes.  Return true if
00369    decoding is successful, false otherwise.  If *OUTLEN is too small,
00370    as many bytes as possible are written to *OUT.  On return, advance
00371    *OUT to point to the byte after the last one written, and decrement
00372    *OUTLEN to reflect the number of bytes remaining in *OUT.  */
00373 static inline bool
00374 decode_4 (char const *restrict in, size_t inlen,
00375           char *restrict *outp, size_t *outleft)
00376 {
00377   char *out = *outp;
00378   if (inlen < 2)
00379     return false;
00380 
00381   if (!isbase64 (in[0]) || !isbase64 (in[1]))
00382     return false;
00383 
00384   if (*outleft)
00385     {
00386       *out++ = ((b64[to_uchar (in[0])] << 2)
00387                 | (b64[to_uchar (in[1])] >> 4));
00388       --*outleft;
00389     }
00390 
00391   if (inlen == 2)
00392     return_false;
00393 
00394   if (in[2] == '=')
00395     {
00396       if (inlen != 4)
00397         return_false;
00398 
00399       if (in[3] != '=')
00400         return_false;
00401     }
00402   else
00403     {
00404       if (!isbase64 (in[2]))
00405         return_false;
00406 
00407       if (*outleft)
00408         {
00409           *out++ = (((b64[to_uchar (in[1])] << 4) & 0xf0)
00410                     | (b64[to_uchar (in[2])] >> 2));
00411           --*outleft;
00412         }
00413 
00414       if (inlen == 3)
00415         return_false;
00416 
00417       if (in[3] == '=')
00418         {
00419           if (inlen != 4)
00420             return_false;
00421         }
00422       else
00423         {
00424           if (!isbase64 (in[3]))
00425             return_false;
00426 
00427           if (*outleft)
00428             {
00429               *out++ = (((b64[to_uchar (in[2])] << 6) & 0xc0)
00430                         | b64[to_uchar (in[3])]);
00431               --*outleft;
00432             }
00433         }
00434     }
00435 
00436   *outp = out;
00437   return true;
00438 }
00439 
00440 /* Decode base64-encoded input array IN of length INLEN to output array
00441    OUT that can hold *OUTLEN bytes.  The input data may be interspersed
00442    with newlines.  Return true if decoding was successful, i.e. if the
00443    input was valid base64 data, false otherwise.  If *OUTLEN is too
00444    small, as many bytes as possible will be written to OUT.  On return,
00445    *OUTLEN holds the length of decoded bytes in OUT.  Note that as soon
00446    as any non-alphabet, non-newline character is encountered, decoding
00447    is stopped and false is returned.  If INLEN is zero, then process
00448    only whatever data is stored in CTX.
00449 
00450    Initially, CTX must have been initialized via base64_decode_ctx_init.
00451    Subsequent calls to this function must reuse whatever state is recorded
00452    in that buffer.  It is necessary for when a quadruple of base64 input
00453    bytes spans two input buffers.
00454 
00455    If CTX is NULL then newlines are treated as garbage and the input
00456    buffer is processed as a unit.  */
00457 
00458 bool
00459 base64_decode_ctx (struct base64_decode_context *ctx,
00460                    const char *restrict in, size_t inlen,
00461                    char *restrict out, size_t *outlen)
00462 {
00463   size_t outleft = *outlen;
00464   bool ignore_newlines = ctx != NULL;
00465   bool flush_ctx = false;
00466   unsigned int ctx_i = 0;
00467 
00468   if (ignore_newlines)
00469     {
00470       ctx_i = ctx->i;
00471       flush_ctx = inlen == 0;
00472     }
00473 
00474 
00475   while (true)
00476     {
00477       size_t outleft_save = outleft;
00478       if (ctx_i == 0 && !flush_ctx)
00479         {
00480           while (true)
00481             {
00482               /* Save a copy of outleft, in case we need to re-parse this
00483                  block of four bytes.  */
00484               outleft_save = outleft;
00485               if (!decode_4 (in, inlen, &out, &outleft))
00486                 break;
00487 
00488               in += 4;
00489               inlen -= 4;
00490             }
00491         }
00492 
00493       if (inlen == 0 && !flush_ctx)
00494         break;
00495 
00496       /* Handle the common case of 72-byte wrapped lines.
00497          This also handles any other multiple-of-4-byte wrapping.  */
00498       if (inlen && *in == '\n' && ignore_newlines)
00499         {
00500           ++in;
00501           --inlen;
00502           continue;
00503         }
00504 
00505       /* Restore OUT and OUTLEFT.  */
00506       out -= outleft_save - outleft;
00507       outleft = outleft_save;
00508 
00509       {
00510         char const *in_end = in + inlen;
00511         char const *non_nl;
00512 
00513         if (ignore_newlines)
00514           non_nl = get_4 (ctx, &in, in_end, &inlen);
00515         else
00516           non_nl = in;  /* Might have nl in this case. */
00517 
00518         /* If the input is empty or consists solely of newlines (0 non-newlines),
00519            then we're done.  Likewise if there are fewer than 4 bytes when not
00520            flushing context and not treating newlines as garbage.  */
00521         if (inlen == 0 || (inlen < 4 && !flush_ctx && ignore_newlines))
00522           {
00523             inlen = 0;
00524             break;
00525           }
00526         if (!decode_4 (non_nl, inlen, &out, &outleft))
00527           break;
00528 
00529         inlen = in_end - in;
00530       }
00531     }
00532 
00533   *outlen -= outleft;
00534 
00535   return inlen == 0;
00536 }
00537 
00538 /* Allocate an output buffer in *OUT, and decode the base64 encoded
00539    data stored in IN of size INLEN to the *OUT buffer.  On return, the
00540    size of the decoded data is stored in *OUTLEN.  OUTLEN may be NULL,
00541    if the caller is not interested in the decoded length.  *OUT may be
00542    NULL to indicate an out of memory error, in which case *OUTLEN
00543    contains the size of the memory block needed.  The function returns
00544    true on successful decoding and memory allocation errors.  (Use the
00545    *OUT and *OUTLEN parameters to differentiate between successful
00546    decoding and memory error.)  The function returns false if the
00547    input was invalid, in which case *OUT is NULL and *OUTLEN is
00548    undefined. */
00549 bool
00550 base64_decode_alloc_ctx (struct base64_decode_context *ctx,
00551                          const char *in, size_t inlen, char **out,
00552                          size_t *outlen)
00553 {
00554   /* This may allocate a few bytes too many, depending on input,
00555      but it's not worth the extra CPU time to compute the exact size.
00556      The exact size is 3 * inlen / 4, minus 1 if the input ends
00557      with "=" and minus another 1 if the input ends with "==".
00558      Dividing before multiplying avoids the possibility of overflow.  */
00559   size_t needlen = 3 * (inlen / 4) + 2;
00560 
00561   *out = malloc (needlen);
00562   if (!*out)
00563     return true;
00564 
00565   if (!base64_decode_ctx (ctx, in, inlen, *out, &needlen))
00566     {
00567       free (*out);
00568       *out = NULL;
00569       return false;
00570     }
00571 
00572   if (outlen)
00573     *outlen = needlen;
00574 
00575   return true;
00576 }