CRC32Stream.st
author Claus Gittinger <cg@exept.de>
Sat, 16 Mar 2019 21:42:14 +0100
changeset 4845 36770d6f14bf
parent 4238 ab8390ef187a
child 4847 c76689ab85bf
permissions -rw-r--r--
#REFACTORING by cg class: CRC32Stream class definition added: #generatorPolynom:initValue: removed: #generatorPolynom #hashValue #initialize #reset changed: #generatorPolynom: #nextPutBytes:from:startingAt: category of: #generatorPolynom: class: CRC32Stream class added: #generatorPolynomMSB: #generatorPolynomMSB:initValue: #generatorPolynomMSB:initValue:xorOut: #new removed: #crcTableFor: #generatorPolynom: #initialize comment/format in: #documentation #examples #newCrc32c

"{ Encoding: utf8 }"

"
 COPYRIGHT (c) 2003 by eXept Software AG
	      All Rights Reserved

 This software is furnished under a license and may be used
 only in accordance with the terms of that license and with the
 inclusion of the above copyright notice.   This software may not
 be provided or otherwise made available to, or used by, any
 other person.  No title to or ownership of the software is
 hereby transferred.
"
"{ Package: 'stx:libbasic2' }"

"{ NameSpace: Smalltalk }"

CRCStream subclass:#CRC32Stream
	instanceVariableNames:''
	classVariableNames:''
	poolDictionaries:''
	category:'System-Crypt-Hashing'
!

!CRC32Stream class methodsFor:'documentation'!

copyright
"
 COPYRIGHT (c) 2003 by eXept Software AG
	      All Rights Reserved

 This software is furnished under a license and may be used
 only in accordance with the terms of that license and with the
 inclusion of the above copyright notice.   This software may not
 be provided or otherwise made available to, or used by, any
 other person.  No title to or ownership of the software is
 hereby transferred.
"
!

documentation
"
    Standard CRC method as defined by ISO 3309 [ISO-3309] or ITU-T V.42 [ITU-T-V42].
    The default CRC polynomial employed is

        x^32+x^26+x^23+x^22+x^16+x^12+x^11+x^10+x^8+x^7+x^5+x^4+x^2+x+1
        (or 16r04C11DB7)

    You can also create an instace performing the Castagnioli CRC-32C 
    (used in iSCSI & SCTP [RFC3720], G.hn payload, SSE4.2):

        self newCrc32c

        poly: 16r1edc6f41
        = x32 + x28 + x27 + x26 + x25 + x23 + x22 + x20 + x19 + x18 + x14 + x13 + x11 + x10 + x9 + x8 + x6 + 1

    Only use CRC to protect against communication errors;
    do NOT use CRC for cryptography - use SHA1Stream or MD5Stream instead.

    Notice that this CRC is also used with PNG images - therefore, its performance
    directly affects png image processing.

    throughput:
        220 Mb/s on MacBook Pro (2.6Ghz I7)
        157000 Kb/s on a 2.5Ghz 64X2 Athlon 4800+ (64bit)
        150000 Kb/s on 2Ghz Duo

    [author:]
        Stefan Vogel (stefan@zwerg)

    [instance variables:]

    [class variables:]

    [see also:]
        SHA1Stream
        MD5Stream

"
!

examples
"

  expect 60C1D0A0
                                                                [exBegin]
    self information:(CRC32Stream hashValueOf:'resume') hexPrintString
                                                                [exEnd]

  expect 16r60C1D0A0
                                                                [exBegin]
    self information:(CRC32Stream new
                            nextPut:$r;
                            nextPut:$e;
                            nextPut:$s;
                            nextPut:$u;
                            nextPut:$m;
                            nextPut:$e;
                            hashValue) hexPrintString
                                                                [exEnd]

  expect 16r70E46888:
                                                                [exBegin]
    self information:(CRC32Stream hashValueOf:#[1 2 3 4 5 6 7]) hexPrintString
                                                                [exEnd]

  expect 16r8CD04C73:
                                                                [exBegin]
    self information:((CRC32Stream hashValueOf:#[16rFF 16rFF 16rFF 16rFF 16rFF 16rFF 16rFF 16rFF 16rFF 16rFF
             16rFF 16rFF 16rFF 16rFF 16rFF 16rFF 16rFF 16rFF 16rFF 16rFF
             16rFF 16rFF 16rFF 16rFF 16rFF 16rFF 16rFF 16rFF 16rFF 16rFF
             16rFF 16rFF 16rFF 16rFF 16rFF 16rFF 16rFF 16rFF 16rFF 16rFF]) hexPrintString)
                                                                [exEnd]

  timing throughput:
                                                                [exBegin]
    |hashStream n t|

    hashStream := CRC32Stream new.
    n := 1000000.
    t := Time millisecondsToRun:[
            n timesRepeat:[
                hashStream nextPutAll:'12345678901234567890123456789012345678901234567890'.
            ].
         ].
    t := (t / 1000) asFloat.
    Transcript show:'crc32:'; showCR: hashStream hashValue hexPrintString.
    Transcript show:t; show:' seconds for '; show:(50*n/1024) asFloat; showCR:' Kb'.
    Transcript show:(n*50/1024 / t); showCR:' Kb/s'
                                                                [exEnd]

  test vectors from https://tools.ietf.org/html/rfc3720#page-217:
  
  expect 0
                                                                [exBegin]
    self information:(CRC32Stream newCrc32c hashValueOf:'') hexPrintString
                                                                [exEnd]
  expect C1D04330
                                                                [exBegin]
    self information:(CRC32Stream newCrc32c hashValueOf:'a') hexPrintString
                                                                [exEnd]
  expect E3069283
                                                                [exBegin]
    self information:(CRC32Stream newCrc32c hashValueOf:'123456789') hexPrintString
                                                                [exEnd]
  expect 8A9136AA
                                                                [exBegin]
    self information:(CRC32Stream newCrc32c hashValueOf:(ByteArray new:32 withAll:0)) hexPrintString
                                                                [exEnd]
  expect 62a8ab43
                                                                [exBegin]
    self information:(CRC32Stream newCrc32c hashValueOf:(ByteArray new:32 withAll:16rFF)) hexPrintString
                                                                [exEnd]
  expect 46dd794e
                                                                [exBegin]
    self information:(CRC32Stream newCrc32c hashValueOf:(0 to:31) asByteArray) hexPrintString
                                                                [exEnd]

"
! !

!CRC32Stream class methodsFor:'instance creation'!

generatorPolynomMSB:anInteger
    "notice, in literature, the generator polynom is usually specified as an MSB number"
    
    ^ self generatorPolynom:(anInteger bitReversed32)

    "
       self assert:((self generatorPolynom:16r82F63B78)
                                nextPut:'123456789';
                                hashValue)    = 16rE3069283
    "

    "Created: / 16-03-2019 / 20:55:46 / Claus Gittinger"
!

generatorPolynomMSB:anInteger initValue:initValueArg
    "notice, in literature, the generator polynom is usually specified as an MSB number"
    
    ^ self generatorPolynom:(anInteger bitReversed32) initValue:initValueArg

    "Created: / 16-03-2019 / 21:11:31 / Claus Gittinger"
!

generatorPolynomMSB:anInteger initValue:initValueArg xorOut:xorOut
    "notice, in literature, the generator polynom is usually specified as an MSB number"
    
    ^ self generatorPolynom:(anInteger bitReversed32) initValue:initValueArg xorOut:xorOut

    "Created: / 16-03-2019 / 21:21:27 / Claus Gittinger"
!

new
    "return an instance of the ITU-T CRC-32
        x^32+x^26+x^23+x^22+x^16+x^12+x^11+x^10+x^8+x^7+x^5+x^4+x^2+x+1"

    "/ 16r4C11DB7 bitReversed32 -> 16rEDB88320
    ^ self generatorPolynom:16rEDB88320 initValue:16rFFFFFFFF xorOut:16rFFFFFFFF

    "Created: / 16-03-2019 / 21:09:19 / Claus Gittinger"
!

newCrc32c
    "return an instance of the Castagnoli CRC-32
        x32 + x28 + x27 + x26 + x25 + x23 + x22 + x20 + x19 + x18 + x14 + x13 + x11 + x10 + x9 + x8 + x6 + 1
     (used in iSCSI & SCTP, G.hn payload, SSE4.2)"

    "/ 16r1edc6f41 bitReversed32 -> 16r82F63B78
    ^ self generatorPolynom:16r82F63B78 initValue:16rFFFFFFFF xorOut:16rFFFFFFFF

    "
     Castagnoli crc:
       self assert:((self newCrc32c)
                                nextPut:'123456789';
                                hashValue) = 3808858755. '16rE3069283'

     default crc:
       self assert:((self new)
                                nextPut:'123456789';
                                hashValue) = 3421780262. '16rCBF43926'
    "

    "Modified (comment): / 17-05-2012 / 12:48:53 / cg"
    "Modified: / 16-03-2019 / 21:21:07 / Claus Gittinger"
! !

!CRC32Stream methodsFor:'initialization'!

generatorPolynom:anInteger
    "set the generator polynom for this instance.
     Note: you have to set the bit-reversed value, so the LSB must be first"

    self generatorPolynom:anInteger initValue:16rFFFFFFFF xorOut:16rFFFFFFFF.

    "Modified: / 16-03-2019 / 21:23:25 / Claus Gittinger"
!

generatorPolynom:anInteger initValue:initValueArg
    "set the generator polynom for this instance.
     Note: you have to set the bit-reversed value, so the LSB must be first"

    self generatorPolynom:anInteger initValue:initValueArg xorOut:16rFFFFFFFF

    "Created: / 16-03-2019 / 21:06:16 / Claus Gittinger"
! !

!CRC32Stream methodsFor:'writing'!

nextPutBytes:count from:anObject startingAt:start
    "add the hash of anObject to the computed hash so far."

%{
    if (__bothSmallInteger(count, start)) {
        int len, offs;
        int objSize;
        unsigned char *extPtr;

        len = __intVal(count);
        offs = __intVal(start) - 1;

        if (__isExternalBytesLike(anObject)) {
            OBJ sz = __externalBytesSize(anObject);

            extPtr = (unsigned char *)__externalBytesAddress(anObject);
            if (__isSmallInteger(sz)) {
                objSize = __intVal(sz);
            } else {
                objSize = 0; /* unknown */
            }
        } else {
            int nInstVars, nInstBytes;
            OBJ oClass = __Class(anObject);

            switch (__intVal(__ClassInstPtr(oClass)->c_flags) & ARRAYMASK) {
                case BYTEARRAY:
                case WORDARRAY:
                case LONGARRAY:
                case SWORDARRAY:
                case SLONGARRAY:
                case FLOATARRAY:
                case DOUBLEARRAY:
                    break;
                default:
                    goto bad;
            }
            nInstVars = __intVal(__ClassInstPtr(oClass)->c_ninstvars);
            nInstBytes = __OBJS2BYTES__(nInstVars);
            // nInstBytes is the number of bytes occupied by pointer instance variables
            // subtract from size and add to byte-pointer
            objSize = __qSize(anObject) - OHDR_SIZE - nInstBytes;
            extPtr = (unsigned char *)__byteArrayVal(anObject)+nInstBytes;
        }

        if ((offs >= 0) && (len >= 0) && (objSize >= (len + offs))) {
            unsigned int _crc;
            unsigned int *_crcTable = __integerArrayVal( __INST(crcTable) );
            unsigned char *cp = extPtr+offs;
            unsigned int n = len;

            if (sizeof(INT) == 8 || __isSmallInteger(__INST(crc)) ) {
                _crc = (unsigned int) (__intVal( __INST(crc) ));
            } else {
                _crc = __unsignedLongIntVal( __INST(crc) );
            }

#ifdef __LSBFIRST__
# if __POINTER_SIZE__ == 8
            if (((unsigned INT)cp & 7) == 0) {
                // longword aligned
                for ( ; n >= 8 ; n -= 8, cp += 8) {
                    unsigned INT lWord;
                    unsigned char _idx;

                    lWord = ((unsigned INT *)cp)[0];
                    _idx = (_crc ^ lWord) & 0xFF;
                    _crc = _crcTable[_idx] ^ (_crc >> 8);
                    _idx = (_crc ^ (lWord>>8)) & 0xFF;
                    _crc = _crcTable[_idx] ^ (_crc >> 8);
                    _idx = (_crc ^ (lWord>>16)) & 0xFF;
                    _crc = _crcTable[_idx] ^ (_crc >> 8);
                    _idx = (_crc ^ (lWord>>24)) & 0xFF;
                    _crc = _crcTable[_idx] ^ (_crc >> 8);
                    
                    _idx = (_crc ^ (lWord>>32)) & 0xFF;
                    _crc = _crcTable[_idx] ^ (_crc >> 8);
                    _idx = (_crc ^ (lWord>>40)) & 0xFF;
                    _crc = _crcTable[_idx] ^ (_crc >> 8);
                    _idx = (_crc ^ (lWord>>48)) & 0xFF;
                    _crc = _crcTable[_idx] ^ (_crc >> 8);
                    _idx = (_crc ^ (lWord>>56)) & 0xFF;
                    _crc = _crcTable[_idx] ^ (_crc >> 8);
                }
            }
# endif            
            if (((unsigned INT)cp & 3) == 0) {
                // word aligned
                for ( ; n >= 4 ; n -= 4, cp += 4) {
                    unsigned int word;
                    unsigned char _idx;

                    word = ((unsigned int *)cp)[0];
                    _idx = (_crc ^ word) & 0xFF;
                    _crc = _crcTable[_idx] ^ (_crc >> 8);
                    _idx = (_crc ^ (word>>8)) & 0xFF;
                    _crc = _crcTable[_idx] ^ (_crc >> 8);
                    _idx = (_crc ^ (word>>16)) & 0xFF;
                    _crc = _crcTable[_idx] ^ (_crc >> 8);
                    _idx = (_crc ^ (word>>24)) & 0xFF;
                    _crc = _crcTable[_idx] ^ (_crc >> 8);
                }
            }
#endif
            for ( ; n >= 4 ; n -= 4, cp += 4) {
                unsigned char _idx;

                _idx = (_crc ^ cp[0]) & 0xFF;
                _crc = _crcTable[_idx] ^ (_crc >> 8);
                _idx = (_crc ^ cp[1]) & 0xFF;
                _crc = _crcTable[_idx] ^ (_crc >> 8);
                _idx = (_crc ^ cp[2]) & 0xFF;
                _crc = _crcTable[_idx] ^ (_crc >> 8);
                _idx = (_crc ^ cp[3]) & 0xFF;
                _crc = _crcTable[_idx] ^ (_crc >> 8);
            }
            while (n-- > 0) {
                unsigned char _idx = (_crc ^ *cp++) & 0xFF;
                _crc = _crcTable[_idx] ^ (_crc >> 8);
            }

            if (sizeof(INT) == 8 || _crc <= _MAX_INT)
                __INST(crc) = __MKSMALLINT(_crc);
            else {
                // this code fails with gcc 4.7.2:
                // __INST(crc) = __MKUINT(_crc); __STORESELF(crc);
                OBJ temp = __MKUINT(_crc); 
                __INST(crc) = temp; __STORESELF(crc);
            }

            // if (__unsignedLongIntVal(__INST(crc)) == 4294967295) printf("inst: %d crc: %d new: %d\n", __unsignedLongIntVal(__INST(crc)), _crc, __unsignedLongIntVal(temp));
            RETURN (count);
        }
    }
bad: ;
%}.
    self error:'invalid argument'

    "Created: / 09-01-2012 / 16:48:35 / cg"
    "Modified: / 16-03-2019 / 21:40:02 / Claus Gittinger"
! !

!CRC32Stream class methodsFor:'documentation'!

version
    ^ '$Header$'
!

version_CVS
    ^ '$Header$'
! !