view toys/lsb/md5sum.c @ 912:f4f5132d5ac7

Stat cleanup. From the mailing list: Ok, first thing: clean up the help text. I realize what's there is copied verbatim from the man page, but that man page sucks. ("modification time" vs "change time"?) Took a bit of finagling to fit it in 80x24, but just made it. GLOBALS() indent was still tab, change to two spaces. And I tend to put a blank line between options lib/args.c automatically fills out and normal globals. We never do anything with date_stat_format() but immediately print it, might as well make the function do it. The types[] array in do_stat() is a rough edge. Hmmm... there's no else case that sets the type in case it was unknown (such as 0). In theory, this never happens. In practice it means I can cheat slightly, given this observation: $ find linux -name stat.h | xargs grep 'S_IF[A-Z]*[ \t]' linux/include/uapi/linux/stat.h:#define S_IFMT 00170000 linux/include/uapi/linux/stat.h:#define S_IFSOCK 0140000 linux/include/uapi/linux/stat.h:#define S_IFLNK 0120000 linux/include/uapi/linux/stat.h:#define S_IFREG 0100000 linux/include/uapi/linux/stat.h:#define S_IFBLK 0060000 linux/include/uapi/linux/stat.h:#define S_IFDIR 0040000 linux/include/uapi/linux/stat.h:#define S_IFCHR 0020000 linux/include/uapi/linux/stat.h:#define S_IFIFO 0010000 I.E. the only place the I_IFBLAH constants occur a stat.h header in current linux code is in the generic stuff, it doesn't vary per target. (The access permission bits are actually subtly standardized in posix due to the command line arguments to chmod, although I'm sure cygwin finds a way to break. But the type fields, not so much. But linux has to be binary compatible with itself foreverish, and that's all I really care about.) So, we have ALMOST have this going by twos, except there's no 8 and there is a 1. so let's make the 1 the default, feed a blank string into the 8... No, duh: octal. So it's actually 2, 4, 6, 8, 10, 12. So make the loop look like: filetype = statf->st_mode & S_IFMT; TT.ftname = types; for (i = 1; filetype != (i*8192) && i < 7; i++) TT.ftname += strlen(TT.ftname)+1; Yes that's linux-specific, and I think I'm ok with that. Printing all zeroes and pretending that's nanosecond resolution... either support it or don't. Let's see, supporting it is stat->st_atim.tv_nsec and similar... no mention of nanoseconds in strftime() (et tu, posix2008?) so pass it as a second argument and append it by hand... (Need to test that against musl...) When we hit an unknown type in print_it() we print the literal character, which is right for %% but what about an unknown option? $ stat -c %q / ? Eh, I guess that's a "don't care". It didn't die with an error, that's the important thing. I have a horrible idea for compressing the switch/case blocks, but should probably check this in and get some sleep for right now...
author Rob Landley <rob@landley.net>
date Tue, 28 May 2013 00:28:45 -0500
parents 786841fdb1e0
children 5bdd2ee6f3c4
line wrap: on
line source

/* md5sum.c - Calculate RFC 1321 md5 hash and sha1 hash.
 *
 * Copyright 2012 Rob Landley <rob@landley.net>
 *
 * See http://refspecs.linuxfoundation.org/LSB_4.1.0/LSB-Core-generic/LSB-Core-generic/md5sum.html
 * and http://www.ietf.org/rfc/rfc1321.txt
 *
 * They're combined this way to share infrastructure, and because md5sum is
 * and LSB standard command, sha1sum is just a good idea.

USE_MD5SUM(NEWTOY(md5sum, NULL, TOYFLAG_USR|TOYFLAG_BIN))
USE_MD5SUM_SHA1SUM(OLDTOY(sha1sum, md5sum, NULL, TOYFLAG_USR|TOYFLAG_BIN))

config MD5SUM
  bool "md5sum"
  default y
  help
    usage: md5sum [FILE]...

    Calculate md5 hash for each input file, reading from stdin if none.
    Output one hash (16 hex digits) for each input file, followed by
    filename.

config MD5SUM_SHA1SUM
  bool "sha1sum"
  default y
  depends on MD5SUM
  help
    usage: sha1sum [FILE]...

    calculate sha1 hash for each input file, reading from stdin if one.
    Output one hash (20 hex digits) for each input file, followed by
    filename.
*/

#define FOR_md5sum
#include "toys.h"

GLOBALS(
  unsigned state[5];
  unsigned oldstate[5];
  uint64_t count;
  union {
    char c[64];
    unsigned i[16];
  } buffer;
)

// for(i=0; i<64; i++) md5table[i] = abs(sin(i+1))*(1<<32);  But calculating
// that involves not just floating point but pulling in -lm (and arguing with
// C about whether 1<<32 is a valid thing to do on 32 bit platforms) so:

static uint32_t md5table[64] = {
  0xd76aa478, 0xe8c7b756, 0x242070db, 0xc1bdceee, 0xf57c0faf, 0x4787c62a,
  0xa8304613, 0xfd469501, 0x698098d8, 0x8b44f7af, 0xffff5bb1, 0x895cd7be,
  0x6b901122, 0xfd987193, 0xa679438e, 0x49b40821, 0xf61e2562, 0xc040b340,
  0x265e5a51, 0xe9b6c7aa, 0xd62f105d, 0x02441453, 0xd8a1e681, 0xe7d3fbc8,
  0x21e1cde6, 0xc33707d6, 0xf4d50d87, 0x455a14ed, 0xa9e3e905, 0xfcefa3f8,
  0x676f02d9, 0x8d2a4c8a, 0xfffa3942, 0x8771f681, 0x6d9d6122, 0xfde5380c,
  0xa4beea44, 0x4bdecfa9, 0xf6bb4b60, 0xbebfbc70, 0x289b7ec6, 0xeaa127fa,
  0xd4ef3085, 0x04881d05, 0xd9d4d039, 0xe6db99e5, 0x1fa27cf8, 0xc4ac5665,
  0xf4292244, 0x432aff97, 0xab9423a7, 0xfc93a039, 0x655b59c3, 0x8f0ccc92,
  0xffeff47d, 0x85845dd1, 0x6fa87e4f, 0xfe2ce6e0, 0xa3014314, 0x4e0811a1,
  0xf7537e82, 0xbd3af235, 0x2ad7d2bb, 0xeb86d391
};

// Mix next 64 bytes of data into md5 hash

static void md5_transform(void)
{
  unsigned x[4], *b = (unsigned *)TT.buffer.c;
  int i;

  memcpy(x, TT.state, sizeof(x));

  for (i=0; i<64; i++) {
    unsigned int in, a, rot, temp;

    a = (-i)&3;
    if (i<16) {
      in = i;
      rot = 7+(5*(i&3));
      temp = x[(a+1)&3];
      temp = (temp & x[(a+2)&3]) | ((~temp) & x[(a+3)&3]);
    } else if (i<32) {
      in = (1+(5*i))&15;
      temp = (i&3)+1;
      rot = temp*5;
      if (temp&2) rot--;
      temp = x[(a+3)&3];
      temp = (x[(a+1)&3] & temp) | (x[(a+2)&3] & ~temp);
    } else if (i<48) {
      in = (5+(3*(i&15)))&15;
      rot = i&3;
      rot = 4+(5*rot)+((rot+1)&6);
      temp = x[(a+1)&3] ^ x[(a+2)&3] ^ x[(a+3)&3];
    } else {
      in = (7*(i&15))&15;
      rot = (i&3)+1;
      rot = (5*rot)+(((rot+2)&2)>>1);
      temp = x[(a+2)&3] ^ (x[(a+1)&3] | ~x[(a+3)&3]);
    }
    temp += x[a] + b[in] + md5table[i];
    x[a] = x[(a+1)&3] + ((temp<<rot) | (temp>>(32-rot)));
  }
  for (i=0; i<4; i++) TT.state[i] += x[i];
}

// Mix next 64 bytes of data into sha1 hash.

static const unsigned rconsts[]={0x5A827999,0x6ED9EBA1,0x8F1BBCDC,0xCA62C1D6};
#define rol(value, bits) (((value) << (bits)) | ((value) >> (32 - (bits))))

static void sha1_transform(void)
{
  int i, j, k, count;
  unsigned *block = TT.buffer.i;
  unsigned *rot[5], *temp;

  // Copy context->state[] to working vars
  for (i=0; i<5; i++) {
    TT.oldstate[i] = TT.state[i];
    rot[i] = TT.state + i;
  }
  // 4 rounds of 20 operations each.
  for (i=count=0; i<4; i++) {
    for (j=0; j<20; j++) {
      unsigned work;

      work = *rot[2] ^ *rot[3];
      if (!i) work = (work & *rot[1]) ^ *rot[3];
      else {
        if (i==2) work = ((*rot[1]|*rot[2])&*rot[3])|(*rot[1]&*rot[2]);
        else work ^= *rot[1];
      }

      if (!i && j<16)
        work += block[count] = (rol(block[count],24)&0xFF00FF00)
                             | (rol(block[count],8)&0x00FF00FF);
      else
        work += block[count&15] = rol(block[(count+13)&15]
              ^ block[(count+8)&15] ^ block[(count+2)&15] ^ block[count&15], 1);
      *rot[4] += work + rol(*rot[0],5) + rconsts[i];
      *rot[1] = rol(*rot[1],30);

      // Rotate by one for next time.
      temp = rot[4];
      for (k=4; k; k--) rot[k] = rot[k-1];
      *rot = temp;
      count++;
    }
  }
  // Add the previous values of state[]
  for (i=0; i<5; i++) TT.state[i] += TT.oldstate[i];
}

// Fill the 64-byte working buffer and call transform() when full.

static void hash_update(char *data, unsigned int len, void (*transform)(void))
{
  unsigned int i, j;

  j = TT.count & 63;
  TT.count += len;

  // Enough data to process a frame?
  if ((j + len) > 63) {
    i = 64-j;
    memcpy(TT.buffer.c + j, data, i);
    transform();
    for ( ; i + 63 < len; i += 64) {
      memcpy(TT.buffer.c, data + i, 64);
      transform();
    }
    j = 0;
  } else i = 0;
  // Grab remaining chunk
  memcpy(TT.buffer.c + j, data + i, len - i);
}

// Callback for loopfiles()

static void do_hash(int fd, char *name)
{
  uint64_t count;
  int i, sha1=toys.which->name[0]=='s';;
  char buf;
  void (*transform)(void);

  /* SHA1 initialization constants  (md5sum uses first 4) */
  TT.state[0] = 0x67452301;
  TT.state[1] = 0xEFCDAB89;
  TT.state[2] = 0x98BADCFE;
  TT.state[3] = 0x10325476;
  TT.state[4] = 0xC3D2E1F0;
  TT.count = 0;

  transform = sha1 ? sha1_transform : md5_transform;
  for (;;) {
    i = read(fd, toybuf, sizeof(toybuf));
    if (i<1) break;
    hash_update(toybuf, i, transform);
  }

  count = TT.count << 3;

  // End the message by appending a "1" bit to the data, ending with the
  // message size (in bits, big endian), and adding enough zero bits in
  // between to pad to the end of the next 64-byte frame.
  //
  // Since our input up to now has been in whole bytes, we can deal with
  // bytes here too.

  buf = 0x80;
  do {
    hash_update(&buf, 1, transform);
    buf = 0;
  } while ((TT.count & 63) != 56);
  if (sha1) count=bswap_64(count);
  for (i = 0; i < 8; i++) TT.buffer.c[56+i] = count >> (8*i);
  transform();

  if (sha1)
    for (i = 0; i < 20; i++)
      printf("%02x", 255&(TT.state[i>>2] >> ((3-(i & 3)) * 8)));
  else for (i=0; i<4; i++) printf("%08x", SWAP_BE32(TT.state[i]));

  // Wipe variables.  Cryptographer paranoia.
  memset(&TT, 0, sizeof(TT));

  printf("  %s\n", name);
}

void md5sum_main(void)
{
  loopfiles(toys.optargs, do_hash);
}