os/kernelhwsrv/userlibandfileserver/fileserver/fs_utils/bit_vector.cpp
changeset 0 bde4ae8d615e
     1.1 --- /dev/null	Thu Jan 01 00:00:00 1970 +0000
     1.2 +++ b/os/kernelhwsrv/userlibandfileserver/fileserver/fs_utils/bit_vector.cpp	Fri Jun 15 03:10:57 2012 +0200
     1.3 @@ -0,0 +1,883 @@
     1.4 +// Copyright (c) 1995-2009 Nokia Corporation and/or its subsidiary(-ies).
     1.5 +// All rights reserved.
     1.6 +// This component and the accompanying materials are made available
     1.7 +// under the terms of the License "Eclipse Public License v1.0"
     1.8 +// which accompanies this distribution, and is available
     1.9 +// at the URL "http://www.eclipse.org/legal/epl-v10.html".
    1.10 +//
    1.11 +// Initial Contributors:
    1.12 +// Nokia Corporation - initial contribution.
    1.13 +//
    1.14 +// Contributors:
    1.15 +//
    1.16 +// Description:
    1.17 +//
    1.18 +//  Collection of common constants, utility functions, etc. for the file server and file systems.
    1.19 +//  Definitions here must be filesystem-agnostic, i.e. generic enougs to be used by every file system
    1.20 +//
    1.21 +//  This is the internal file and must not be exported.
    1.22 +
    1.23 +/**
    1.24 +    @file
    1.25 +    @internalTechnology
    1.26 +*/
    1.27 +
    1.28 +#include "bit_vector.h"
    1.29 +
    1.30 +
    1.31 +//#######################################################################################################################################
    1.32 +//#   RBitVector class implementation
    1.33 +//#######################################################################################################################################
    1.34 +
    1.35 +const TUint32 K_FFFF = 0xFFFFFFFF; //-- all one bits, beware rigth shifts of signed integers!
    1.36 +
    1.37 +RBitVector::RBitVector()
    1.38 +          :iNumBits(0), ipData(NULL), iNumWords(0)
    1.39 +    {
    1.40 +    }
    1.41 +
    1.42 +
    1.43 +RBitVector::~RBitVector()
    1.44 +    {
    1.45 +    Close();
    1.46 +    }
    1.47 +
    1.48 +/**
    1.49 +    Panics.
    1.50 +    @param aPanicCode   a panic code
    1.51 +*/
    1.52 +void RBitVector::Panic(TPanicCode aPanicCode) const
    1.53 +    {
    1.54 +    _LIT(KPanicCat,"RBitVector");
    1.55 +    User::Panic(KPanicCat, aPanicCode);
    1.56 +    }
    1.57 +
    1.58 +/** explicitly closes the object and deallocates memory */
    1.59 +void RBitVector::Close()
    1.60 +    {
    1.61 +    iNumBits = 0;
    1.62 +    iNumWords =0;
    1.63 +    User::Free(ipData);
    1.64 +    ipData = NULL;
    1.65 +    }
    1.66 +
    1.67 +//-----------------------------------------------------------------------------
    1.68 +
    1.69 +/**
    1.70 +    Comparison perator.
    1.71 +    @param  aRhs a vector to compate with.
    1.72 +    @panic ESizeMismatch in the case of different vector sizes
    1.73 +*/
    1.74 +TBool RBitVector::operator==(const RBitVector& aRhs) const
    1.75 +    {
    1.76 +    __ASSERT_ALWAYS(ipData, Panic(ENotInitialised));
    1.77 +    __ASSERT_ALWAYS(iNumBits == aRhs.iNumBits, Panic(ESizeMismatch));
    1.78 +
    1.79 +
    1.80 +    if(!iNumBits)
    1.81 +        return ETrue; //-- comparing 0-lenght arrays
    1.82 +
    1.83 +    if(this == &aRhs)
    1.84 +        {//-- comparing with itself, potential source of errors
    1.85 +        ASSERT(0);
    1.86 +        return ETrue; 
    1.87 +        }
    1.88 +    
    1.89 +    if(iNumWords >= 1)
    1.90 +        {
    1.91 +        const TUint32 cntBytes = (iNumBits >> 5) << 2; //-- bytes to compare
    1.92 +        if(memcompare((const TUint8*)ipData, cntBytes, (const TUint8*)aRhs.ipData, cntBytes))
    1.93 +            return EFalse;
    1.94 +        }
    1.95 +
    1.96 +    const TUint32 bitsRest  = iNumBits & 0x1F;
    1.97 +    if(bitsRest)
    1.98 +        {
    1.99 +        const TUint32 mask = K_FFFF >> (32-bitsRest);
   1.100 +        return ( (ipData[iNumWords-1] & mask) == (aRhs.ipData[iNumWords-1] & mask) );
   1.101 +        }
   1.102 +    
   1.103 +    return ETrue;
   1.104 +    }
   1.105 +
   1.106 +TBool RBitVector::operator!=(const RBitVector& aRhs) const  
   1.107 +    {
   1.108 +    return ! ((*this) == aRhs);
   1.109 +    } 
   1.110 +
   1.111 +//-----------------------------------------------------------------------------
   1.112 +
   1.113 +/** The same as Create(), but leaves on error */
   1.114 +void RBitVector::CreateL(TUint32 aNumBits)
   1.115 +    {
   1.116 +    User::LeaveIfError(Create(aNumBits));
   1.117 +    }
   1.118 +
   1.119 +
   1.120 +/**
   1.121 +    Create the vector with the size of aNumBits bits.
   1.122 +    @return system-wide error codes:
   1.123 +        KErrNoMemory    unable to allocate sufficient amount of memory for the array
   1.124 +        KErrInUse       an attempt to call Create() for non-empty vector. Close it first.
   1.125 +        KErrArgument    invalid aNumBits value == 0
   1.126 +*/
   1.127 +TInt RBitVector::Create(TUint32 aNumBits)
   1.128 +    {
   1.129 +
   1.130 +    if(ipData)
   1.131 +        return KErrInUse; //-- array is already in use. Close it first.
   1.132 +
   1.133 +    if(!aNumBits)
   1.134 +        return KErrArgument;
   1.135 +
   1.136 +    //-- memory is allocated by word (32 bit) quiantities
   1.137 +    const TUint32 numWords = (aNumBits >> 5) + ((aNumBits & 0x1F) > 0 ? 1:0);
   1.138 +    ipData = (TUint32*)User::AllocZ(numWords << 2);
   1.139 +
   1.140 +    if(!ipData)
   1.141 +        return KErrNoMemory;
   1.142 +
   1.143 +    iNumBits  = aNumBits;
   1.144 +    iNumWords = numWords;
   1.145 +
   1.146 +    return KErrNone;
   1.147 +    }
   1.148 +
   1.149 +
   1.150 +/**
   1.151 +    Fill a bit vector with a given bit value
   1.152 +    @param aVal a bit value
   1.153 +*/
   1.154 +void RBitVector::Fill(TBool aVal)
   1.155 +    {
   1.156 +    __ASSERT_ALWAYS(ipData, Panic(ENotInitialised));
   1.157 +    memset(ipData, (aVal ? 0xFF : 0x00), iNumWords << 2);
   1.158 +    }
   1.159 +
   1.160 +/** Invert all bits in a bit vector */
   1.161 +void RBitVector::Invert()
   1.162 +{
   1.163 +    __ASSERT_ALWAYS(ipData, Panic(ENotInitialised));
   1.164 +    for(TUint32 i=0; i<iNumWords; ++i)
   1.165 +        ipData[i] ^= K_FFFF;
   1.166 +}
   1.167 +
   1.168 +
   1.169 +/**
   1.170 +    Perform "And" operation between 2 vectors. They shall be the same size.
   1.171 +    @param  aRhs a vector from the right hand side
   1.172 +    @panic ESizeMismatch in the case of different vector sizes
   1.173 +*/
   1.174 +void RBitVector::And(const RBitVector& aRhs)
   1.175 +    {
   1.176 +    __ASSERT_ALWAYS(ipData, Panic(ENotInitialised));
   1.177 +    __ASSERT_ALWAYS(iNumBits == aRhs.iNumBits, Panic(ESizeMismatch));
   1.178 +    for(TUint32 i=0; i<iNumWords; ++i)
   1.179 +        {
   1.180 +        ipData[i] &= aRhs.ipData[i];
   1.181 +        }
   1.182 +    }
   1.183 +
   1.184 +/**
   1.185 +    Perform "Or" operation between 2 vectors. They shall be the same size.    
   1.186 +    @param  aRhs a vector from the right hand side
   1.187 +    @panic ESizeMismatch in the case of different vector sizes
   1.188 +*/
   1.189 +void RBitVector::Or(const RBitVector& aRhs)
   1.190 +    {
   1.191 +    __ASSERT_ALWAYS(ipData, Panic(ENotInitialised));
   1.192 +    __ASSERT_ALWAYS(iNumBits == aRhs.iNumBits, Panic(ESizeMismatch));
   1.193 +    for(TUint32 i=0; i<iNumWords; ++i)
   1.194 +        {
   1.195 +        ipData[i] |= aRhs.ipData[i];
   1.196 +        }
   1.197 +    }
   1.198 +
   1.199 +/**
   1.200 +    Perform "Xor" operation between 2 vectors. They shall be the same size.    
   1.201 +    @param  aRhs a vector from the right hand side
   1.202 +    @panic ESizeMismatch in the case of different vector sizes
   1.203 +*/
   1.204 +void RBitVector::Xor(const RBitVector& aRhs)
   1.205 +    {
   1.206 +    __ASSERT_ALWAYS(ipData, Panic(ENotInitialised));
   1.207 +    __ASSERT_ALWAYS(iNumBits == aRhs.iNumBits, Panic(ESizeMismatch));
   1.208 +    for(TUint32 i=0; i<iNumWords; ++i)
   1.209 +        {
   1.210 +        ipData[i] ^= aRhs.ipData[i];
   1.211 +        }
   1.212 +    }
   1.213 +
   1.214 +//-----------------------------------------------------------------------------
   1.215 +/**
   1.216 +    Fill a range from bit number "aIndexFrom" to "aIndexTo" inclusively with the value of aVal
   1.217 +    
   1.218 +    @param  aIndexFrom  start bit number (inclusive)
   1.219 +    @param  aIndexTo    end bit number (inclusive)
   1.220 +    @param  aVal        the value to be used to fill the range (0s or 1s)
   1.221 +*/
   1.222 +void RBitVector::Fill(TUint32 aIndexFrom, TUint32 aIndexTo, TBool aVal)
   1.223 +    {
   1.224 +    __ASSERT_ALWAYS(ipData, Panic(ENotInitialised));
   1.225 +
   1.226 +    //-- swap indexes if they are not in order
   1.227 +    if(aIndexFrom > aIndexTo)
   1.228 +        {
   1.229 +        const TUint32 tmp = aIndexFrom;
   1.230 +        aIndexFrom = aIndexTo;
   1.231 +        aIndexTo = tmp;
   1.232 +        }
   1.233 +
   1.234 +    __ASSERT_ALWAYS((aIndexFrom < iNumBits) && (aIndexTo < iNumBits), Panic(EIndexOutOfRange));
   1.235 +
   1.236 +    const TUint32 wordStart = WordNum(aIndexFrom);
   1.237 +    const TUint32 wordTo    = WordNum(aIndexTo);
   1.238 +
   1.239 +    if(aVal)
   1.240 +        {//-- filling a range with '1'
   1.241 +        
   1.242 +        TUint32 shift = BitInWord(aIndexFrom);
   1.243 +        const TUint32 mask1 = (K_FFFF >> shift) << shift;
   1.244 +
   1.245 +        TUint32 mask2 = K_FFFF;
   1.246 +        shift = 1+BitInWord(aIndexTo);
   1.247 +        if(shift < 32)
   1.248 +            {
   1.249 +            mask2 = ~((mask2 >> shift) << shift);
   1.250 +            }
   1.251 +
   1.252 +        if(wordTo == wordStart)
   1.253 +            {//-- a special case, filling is in the same word
   1.254 +            ipData[wordStart] |= (mask1 & mask2);
   1.255 +            }
   1.256 +        else
   1.257 +            {
   1.258 +            ipData[wordStart] |= mask1; 
   1.259 +            ipData[wordTo]    |= mask2;
   1.260 +            
   1.261 +            const TUint32 wholeWordsBetween = wordTo - wordStart - 1; //-- whole words that can be bulk filled
   1.262 +
   1.263 +            if(wholeWordsBetween)
   1.264 +                memset(ipData+wordStart+1, 0xFF, wholeWordsBetween << 2);
   1.265 +                            
   1.266 +            }
   1.267 +        }
   1.268 +    else
   1.269 +        {//-- filling a range with '0'
   1.270 +        
   1.271 +        TUint32 shift = BitInWord(aIndexFrom);
   1.272 +        const TUint32 mask1 = ~((K_FFFF >> shift) << shift);
   1.273 +
   1.274 +        TUint32 mask2 = 0;
   1.275 +        shift = 1+BitInWord(aIndexTo);
   1.276 +        if(shift < 32)
   1.277 +            {
   1.278 +            mask2 = ((K_FFFF >> shift) << shift);
   1.279 +            }
   1.280 +
   1.281 +        if(wordTo == wordStart)
   1.282 +            {//-- a special case, filling is in the same word
   1.283 +            ipData[wordStart] &= (mask1 | mask2);
   1.284 +            }
   1.285 +        else
   1.286 +            {
   1.287 +            ipData[wordStart] &= mask1; 
   1.288 +            ipData[wordTo]    &= mask2;
   1.289 +            
   1.290 +            const TUint32 wholeWordsBetween = wordTo - wordStart - 1; //-- whole words that can be bulk filled
   1.291 +
   1.292 +            if(wholeWordsBetween)
   1.293 +                memset(ipData+wordStart+1, 0x00, wholeWordsBetween << 2);
   1.294 +                            
   1.295 +            }
   1.296 +        }
   1.297 +
   1.298 +    }
   1.299 +
   1.300 +//-----------------------------------------------------------------------------
   1.301 +
   1.302 +/**
   1.303 +    Search for a specified bit value ('0' or '1') in the vector from the given position.
   1.304 +    @param  aStartPos   zero-based index; from this position the search will start. This position isn't included to the search.
   1.305 +                        On return may contain a new position if the specified bit is found in specified direction.
   1.306 +    @param  aBitVal     zero or non-zero bit to search.
   1.307 +    @param  aDir        Specifies the search direction
   1.308 +
   1.309 +    @return ETrue if the specified bit value is found; aStartPos gets updated.
   1.310 +            EFalse otherwise.
   1.311 +
   1.312 +*/
   1.313 +TBool RBitVector::Find(TUint32& aStartPos, TBool aBitVal, TFindDirection aDir) const
   1.314 +    {
   1.315 +    __ASSERT_ALWAYS(aStartPos < iNumBits, Panic(EIndexOutOfRange));
   1.316 +    ASSERT(iNumWords && ipData);
   1.317 +
   1.318 +    switch(aDir)
   1.319 +        {
   1.320 +        case ERight:    //-- Search from the given position to the right
   1.321 +            return FindToRight(aStartPos, aBitVal);
   1.322 +
   1.323 +        case ELeft:     //-- Search from the given position to the left (towards lower index)
   1.324 +            return FindToLeft(aStartPos, aBitVal);
   1.325 +
   1.326 +        case ENearestL: //-- Search for the nearest value in both directions starting from left
   1.327 +            return FindNearest(aStartPos, aBitVal, ETrue);
   1.328 +
   1.329 +        case ENearestR: //-- Search for the nearest value in both directions starting from right
   1.330 +            return FindNearest(aStartPos, aBitVal, EFalse);
   1.331 +
   1.332 +        default:
   1.333 +            Panic(EWrondFindDirection);
   1.334 +            return EFalse;
   1.335 +
   1.336 +        };
   1.337 +    
   1.338 +    }
   1.339 +
   1.340 +//-----------------------------------------------------------------------------
   1.341 +/**
   1.342 +    Internal method to look for a given bit value in the right direction.
   1.343 +    see TBool RBitVector::Find(...)
   1.344 +*/
   1.345 +TBool RBitVector::FindToRight(TUint32& aStartPos, TBool aBitVal) const
   1.346 +    {
   1.347 +    if(aStartPos >= iNumBits-1)
   1.348 +        return EFalse; //-- no way to the right
   1.349 +
   1.350 +    const TUint32 startPos = aStartPos+1;
   1.351 +    const TUint32 fInvert = aBitVal ? 0 : K_FFFF; //-- invert everything if we are looking for '0' bit
   1.352 +
   1.353 +    TUint32 wordNum = WordNum(startPos);
   1.354 +    TUint32 val = ipData[wordNum] ^ fInvert;
   1.355 +
   1.356 +    if(wordNum == iNumWords-1)
   1.357 +        {//-- process the last word in the array, some higher bits might not belong to the bit vector
   1.358 +        val = MaskLastWord(val);
   1.359 +        }
   1.360 +
   1.361 +    const TUint32 shift = BitInWord(startPos);
   1.362 +    val = (val >> shift) << shift; //-- mask unused low bits
   1.363 +
   1.364 +    if(val)
   1.365 +        {//-- there are '1' bits in the current word
   1.366 +        goto found;
   1.367 +        }
   1.368 +    else
   1.369 +        {//-- search in higher words
   1.370 +        wordNum++;
   1.371 +
   1.372 +        while(iNumWords-wordNum > 1)
   1.373 +            {
   1.374 +            val = ipData[wordNum] ^ fInvert;
   1.375 +            if(val)
   1.376 +                goto found;
   1.377 +
   1.378 +            wordNum++;
   1.379 +            }
   1.380 +
   1.381 +        if(wordNum == iNumWords-1)
   1.382 +            {//-- process the last word in the array, some higher bith might not belong to the bit vector
   1.383 +            val = ipData[wordNum] ^ fInvert;
   1.384 +            val = MaskLastWord(val);
   1.385 +
   1.386 +            if(val)
   1.387 +                goto found;
   1.388 +            }
   1.389 +        }
   1.390 +
   1.391 +    return EFalse; //-- haven't found anything
   1.392 +
   1.393 +  found:
   1.394 +
   1.395 +    val &= (~val+1); //-- select rightmost bit
   1.396 +    aStartPos = (wordNum << 5)+Log2(val);
   1.397 +    return ETrue;
   1.398 +    }
   1.399 +
   1.400 +
   1.401 +//-----------------------------------------------------------------------------
   1.402 +
   1.403 +/**
   1.404 +    Internal method to look for a given bit value in the left direction.
   1.405 +    see TBool RBitVector::Find(...)
   1.406 +*/
   1.407 +TBool RBitVector::FindToLeft(TUint32& aStartPos, TBool aBitVal) const
   1.408 +{
   1.409 +    if(!aStartPos)
   1.410 +        return EFalse; //-- no way to the left
   1.411 +    
   1.412 +    const TUint32 startPos=aStartPos-1;
   1.413 +    const TUint32 fInvert = aBitVal ? 0 : K_FFFF; //-- invert everything if we are looking for '0' bit
   1.414 +
   1.415 +    TUint32 wordNum = WordNum(startPos);
   1.416 +    TUint32 val = ipData[wordNum] ^ fInvert;
   1.417 +
   1.418 +    const TUint32 shift = 31-(BitInWord(startPos));
   1.419 +    val = (val << shift) >> shift; //-- mask unused high bits
   1.420 +
   1.421 +    if(val)
   1.422 +    {//-- there are '1' bits in the current word
   1.423 +        goto found;
   1.424 +    }
   1.425 +    else
   1.426 +    {//-- search in the lower words
   1.427 +        while(wordNum)
   1.428 +        {
   1.429 +            wordNum--;
   1.430 +            val=ipData[wordNum] ^ fInvert;
   1.431 +            if(val)
   1.432 +                goto found;
   1.433 +        }
   1.434 +    }
   1.435 +
   1.436 +    return EFalse; //-- nothing found
   1.437 +
   1.438 + found:
   1.439 +    aStartPos = (wordNum << 5)+Log2(val);
   1.440 +    return ETrue;
   1.441 +}
   1.442 +
   1.443 +//-----------------------------------------------------------------------------
   1.444 +
   1.445 +/**
   1.446 +    Internal method to look for a given bit value in the both directions.
   1.447 +    see TBool RBitVector::Find(...)
   1.448 +*/
   1.449 +TBool RBitVector::FindNearest(TUint32& aStartPos, TBool aBitVal, TBool aToLeft) const
   1.450 +{
   1.451 +    if(iNumBits < 2)
   1.452 +        return EFalse;
   1.453 +
   1.454 +    if(aStartPos == 0)
   1.455 +        return FindToRight(aStartPos, aBitVal);
   1.456 +
   1.457 +    if(aStartPos == iNumBits-1)
   1.458 +        return FindToLeft(aStartPos, aBitVal);
   1.459 +
   1.460 +    
   1.461 +    const TUint32 fInvert = aBitVal ? 0 : K_FFFF; //-- invert everything if we are looking for '0' bit
   1.462 +    
   1.463 +    TUint32 wordNum = WordNum(aStartPos);
   1.464 +    TUint32 l_Idx; //-- index of the word to the left
   1.465 +    TUint32 r_Idx; //-- index of the word to the right
   1.466 +    
   1.467 +    l_Idx = r_Idx = wordNum;
   1.468 +
   1.469 +    TBool   noWayLeft  = (wordNum == 0);            //-- if we are in the first word
   1.470 +    TBool   noWayRight = (wordNum == iNumWords-1);  //-- if we are in the last word
   1.471 +
   1.472 +    //-- look in the current word first
   1.473 +    TUint32 val = ipData[wordNum] ^ fInvert;
   1.474 +    
   1.475 +    if(noWayRight)
   1.476 +    {   //-- this is the last word in the array, mask unused high bits in the last word
   1.477 +        val = MaskLastWord(val);
   1.478 +    }
   1.479 +
   1.480 +    const TUint32 bitPos = aStartPos & 0x1F;
   1.481 +    val &= ~(1<<bitPos); //-- mask the bit at current position
   1.482 +    
   1.483 +    if(val == 0)
   1.484 +    {//-- no '1' bits in the current word
   1.485 +        noWayLeft  = ItrLeft(l_Idx);
   1.486 +        noWayRight = ItrRight(r_Idx);
   1.487 +    }
   1.488 +    else if(bitPos == 0)
   1.489 +    {
   1.490 +        noWayLeft = ItrLeft(l_Idx); //-- move to the previous word
   1.491 +    }
   1.492 +    else if(bitPos == 31)
   1.493 +    {
   1.494 +        noWayRight = ItrRight(r_Idx); //-- move to the next word
   1.495 +    }
   1.496 +    else
   1.497 +    {//-- look in the current word, in both halves to the left and right from the start position
   1.498 +        
   1.499 +        const TUint32 shift1 = 32-bitPos;
   1.500 +        const TUint32 partLo = (val << shift1) >> shift1; //-- towards lower bits
   1.501 +
   1.502 +        const TUint32 shift2 = bitPos+1;
   1.503 +        const TUint32 partHi = (val >> shift2) << shift2; //-- towards higher bits 
   1.504 +        
   1.505 +
   1.506 +        if(partLo && !partHi) //-- only lower part has '1' bits   
   1.507 +        {
   1.508 +            aStartPos = (wordNum << 5)+Log2(partLo);
   1.509 +            return ETrue;
   1.510 +        }
   1.511 +        else if(!partLo && partHi) //-- only higher part has '1' bits
   1.512 +        {
   1.513 +            aStartPos = (wordNum << 5)+Log2( (partHi & (~partHi+1)) );
   1.514 +            return ETrue;
   1.515 +        }
   1.516 +        else if(partLo && partHi) //-- both parts contain '1' bits, select the nearest one
   1.517 +        {
   1.518 +            const TUint32 posL = (wordNum << 5)+Log2(partLo);
   1.519 +            const TUint32 posR = (wordNum << 5)+Log2( (partHi & (~partHi+1)) );
   1.520 +        
   1.521 +            ASSERT(aStartPos > posL);
   1.522 +            ASSERT(posR > aStartPos);
   1.523 +            const TUint32 distL = aStartPos-posL;
   1.524 +            const TUint32 distR = posR-aStartPos;
   1.525 +
   1.526 +            if(distL < distR)
   1.527 +            {
   1.528 +                aStartPos = posL;
   1.529 +                return ETrue;
   1.530 +            }
   1.531 +            else if(distL > distR)
   1.532 +            {
   1.533 +                aStartPos = posR;
   1.534 +                return ETrue;
   1.535 +            }
   1.536 +            else
   1.537 +            {//-- distL == distR, take into account search priority
   1.538 +                aStartPos = aToLeft ? posL : posR;
   1.539 +                return ETrue;
   1.540 +            }
   1.541 +        }
   1.542 +        else //-- (!partLo && !partHi), nothing in the current word
   1.543 +        {
   1.544 +            ASSERT(0);
   1.545 +        }
   1.546 +
   1.547 +    }// if(bitPos > 0 && bitPos < 31)
   1.548 +
   1.549 +    //-- now we are processing separate words from both sides of the search position
   1.550 +    for(;;)
   1.551 +    { 
   1.552 +        TUint32 wL = ipData[l_Idx] ^ fInvert;
   1.553 +        TUint32 wR = ipData[r_Idx] ^ fInvert;
   1.554 +        if(r_Idx == iNumWords-1)
   1.555 +        {   //-- this is the last word in the array, mask unused high bits in the last word
   1.556 +            wR = MaskLastWord(wR);
   1.557 +        }
   1.558 +
   1.559 +        if(wL && !wR)
   1.560 +        {
   1.561 +            aStartPos = (l_Idx << 5)+Log2(wL);
   1.562 +            return ETrue;
   1.563 +        }
   1.564 +        else if(!wL && wR)
   1.565 +        {
   1.566 +            aStartPos = (r_Idx << 5)+Log2( (wR & (~wR+1)) );
   1.567 +            return ETrue;
   1.568 +        }
   1.569 +        else if(wL && wR)
   1.570 +        {
   1.571 +            const TUint32 posL = (l_Idx << 5)+Log2(wL);
   1.572 +            const TUint32 posR = (r_Idx << 5)+Log2( (wR & (~wR+1)) );
   1.573 +        
   1.574 +            ASSERT(aStartPos > posL);
   1.575 +            ASSERT(posR > aStartPos);
   1.576 +            const TUint32 distL = aStartPos-posL;
   1.577 +            const TUint32 distR = posR-aStartPos;
   1.578 +
   1.579 +            if(distL < distR)
   1.580 +            {
   1.581 +                aStartPos = posL;
   1.582 +                return ETrue;
   1.583 +            }
   1.584 +            else if(distL > distR)
   1.585 +            {
   1.586 +                aStartPos = posR;
   1.587 +                return ETrue;
   1.588 +            }
   1.589 +            else
   1.590 +            {//-- distL == distR, take into account search priority
   1.591 +                aStartPos = aToLeft ? posL : posR;
   1.592 +                return ETrue;
   1.593 +            }
   1.594 +
   1.595 +        }//else if(wL && wR)
   1.596 +
   1.597 +
   1.598 +        if(noWayLeft)
   1.599 +        {
   1.600 +            aStartPos = r_Idx << 5;
   1.601 +            return FindToRight(aStartPos, aBitVal);
   1.602 +        }
   1.603 +        else
   1.604 +        {
   1.605 +            noWayLeft  = ItrLeft(l_Idx);
   1.606 +        }
   1.607 +
   1.608 +        if(noWayRight)
   1.609 +        {
   1.610 +            aStartPos = l_Idx << 5;
   1.611 +            return FindToLeft(aStartPos, aBitVal);
   1.612 +        }
   1.613 +        else
   1.614 +        {    
   1.615 +            noWayRight = ItrRight(r_Idx);
   1.616 +        }
   1.617 +
   1.618 +   }//for(;;)
   1.619 +
   1.620 +    //return EFalse;
   1.621 +}
   1.622 +
   1.623 +//-----------------------------------------------------------------------------
   1.624 +/**
   1.625 +    Find out if two vectors are different.
   1.626 +
   1.627 +    @param  aRhs        vector to compare with
   1.628 +    @param  aDiffIndex  if there is a differene, here will be the number of the first different bit
   1.629 +    @return ETrue if vectors differ, EFalse, if they are identical.
   1.630 +*/
   1.631 +TBool RBitVector::Diff(const RBitVector& aRhs, TUint32& aDiffIndex) const
   1.632 +{
   1.633 +    __ASSERT_ALWAYS(ipData, Panic(ENotInitialised));
   1.634 +    __ASSERT_ALWAYS(iNumBits == aRhs.iNumBits, Panic(ESizeMismatch));
   1.635 +    ASSERT(iNumWords > 0);
   1.636 +
   1.637 +    TUint32 diffWord=0;
   1.638 +    TUint32 wordNum=0;
   1.639 +
   1.640 +    //-- compare all but the last word in the array
   1.641 +    for(wordNum=0; wordNum < iNumWords-1; ++wordNum)
   1.642 +    {
   1.643 +        diffWord = ipData[wordNum] ^ aRhs.ipData[wordNum];
   1.644 +        if(diffWord)
   1.645 +            break;  //-- found difference
   1.646 +    }
   1.647 +
   1.648 +    //-- process the last word in the array
   1.649 +    if(!diffWord)
   1.650 +    {
   1.651 +        diffWord = MaskLastWord(ipData[wordNum]) ^ MaskLastWord(aRhs.ipData[wordNum]);
   1.652 +    }
   1.653 +
   1.654 +    if(!diffWord)
   1.655 +        return EFalse; //-- vectors are the same
   1.656 +
   1.657 +    //-- calculate the position of the bit that different.
   1.658 +    diffWord &= (~diffWord+1); //-- select rightmost bit
   1.659 +    aDiffIndex = (wordNum << 5)+Log2(diffWord);
   1.660 +    
   1.661 +    return ETrue;
   1.662 +}
   1.663 +
   1.664 +//-----------------------------------------------------------------------------
   1.665 +
   1.666 +/**
   1.667 +    Iterate to the left (towards lower index) in the array of words ipData
   1.668 +
   1.669 +    @param  aIdx index within ipData array to be decremented; if it's possible to move left, it will be decreased
   1.670 +    @return ETrue if there is no way left i.e. aIdx is 0. EFalse otherwise and aIdx decreased.
   1.671 +*/
   1.672 +TBool RBitVector::ItrLeft(TUint32& aIdx) const
   1.673 +{
   1.674 +    if(aIdx == 0)
   1.675 +        return ETrue;
   1.676 +    else
   1.677 +    {
   1.678 +        aIdx--;
   1.679 +        return EFalse;
   1.680 +    }
   1.681 +}
   1.682 +
   1.683 +//-----------------------------------------------------------------------------
   1.684 +
   1.685 +/**
   1.686 +    Iterate to the right (towards higher index) in the array of words ipData
   1.687 +
   1.688 +    @param  aIdx index within ipData array to be incremented; if it's possible to move right, it will be increased
   1.689 +    @return ETrue if there is no way right i.e. aIdx corresponds to the last word. EFalse otherwise and aIdx increased.
   1.690 +*/
   1.691 +TBool RBitVector::ItrRight(TUint32& aIdx) const
   1.692 +{
   1.693 +    if(aIdx < iNumWords-1)
   1.694 +    {
   1.695 +        aIdx++;
   1.696 +        return EFalse;
   1.697 +    }
   1.698 +    else
   1.699 +        return ETrue;
   1.700 +}
   1.701 +
   1.702 +//-----------------------------------------------------------------------------
   1.703 +
   1.704 +/**
   1.705 +    Import data to the internal bit vector representation.
   1.706 +    Just replaces number of bytes from apData to the ipData.
   1.707 +
   1.708 +    @param aStartBit starting bit number. Must have 8-bit alignment.
   1.709 +    @param aNumBits  number of bits to import; granularity: 1 bit, i.e. it can be 177, for example.
   1.710 +    @param apData    pointer to the data (bitstream) to import.
   1.711 +
   1.712 +*/
   1.713 +void RBitVector::DoImportData(TUint32 aStartBit, TUint32 aNumBits, const TAny* apData)
   1.714 +{
   1.715 +    ASSERT(aNumBits);
   1.716 +    __ASSERT_ALWAYS(ipData, Panic(ENotInitialised));
   1.717 +    
   1.718 +    //-- check parameters granularity. aStartBit must have 8-bit alignment
   1.719 +    __ASSERT_ALWAYS(!(aStartBit & 0x07), Panic(EDataAlignment));
   1.720 +
   1.721 +
   1.722 +    __ASSERT_ALWAYS(iNumWords && (aStartBit+aNumBits <= iNumBits), Panic(EIndexOutOfRange));
   1.723 +
   1.724 +    const TUint     bitsTail = aNumBits & 0x07;
   1.725 +    const TUint32   nBytes = aNumBits >> 3;
   1.726 + 
   1.727 +    if(nBytes)
   1.728 +    {//-- copy full array of bytes
   1.729 +        const TUint32   startByte = aStartBit >> 3;
   1.730 +        Mem::Copy(((TUint8*)ipData) + startByte, apData, nBytes);
   1.731 +    }
   1.732 +
   1.733 +    if(bitsTail)
   1.734 +    {//-- we need to copy trailing bits from the input data to the corresponding byte of the internal array
   1.735 +        const TUint8 mask   = (TUint8)(0xFF >> (8-bitsTail));
   1.736 +        const TUint8 orMask = (TUint8)( *((const TUint8*)apData + nBytes) & mask);
   1.737 +        const TUint8 andMask= (TUint8)~mask;
   1.738 +
   1.739 +        TUint8* pbData = (TUint8*)ipData + nBytes;
   1.740 +        *pbData &= andMask;
   1.741 +        *pbData |= orMask;
   1.742 +    }
   1.743 +
   1.744 +}
   1.745 +
   1.746 +//-----------------------------------------------------------------------------
   1.747 +
   1.748 +/**
   1.749 +    Export data from the internal bit vector buffer to the external one.
   1.750 +
   1.751 +    @param aStartBit starting bit number. Must have 8-bit alignment.
   1.752 +    @param aNumBits  number of bits to export, must comprise the whole byte, i.e. be multiple of 8.
   1.753 +                     The client is responsible for masking extra bits it doesn't need. 
   1.754 +                     Another implication: e.g. if the bitvector consists of 3 bits, this value must be 8.
   1.755 +                     The value of bits 3-7 in the aData[0] will be undefined.
   1.756 +
   1.757 +    @param aData     destination data descriptor   
   1.758 +*/
   1.759 +void  RBitVector::DoExportData(TUint32 aStartBit, TUint32 aNumBits, TDes8& aData) const
   1.760 +{
   1.761 +    ASSERT(aNumBits);
   1.762 +    __ASSERT_ALWAYS(ipData, Panic(ENotInitialised));
   1.763 +
   1.764 +    //-- check parameters granularity.
   1.765 +    __ASSERT_ALWAYS(!(aStartBit & 0x07), Panic(EDataAlignment)); //-- aStartBit must have 8-bit alignment
   1.766 +    __ASSERT_ALWAYS(!(aNumBits & 0x07), Panic(EDataAlignment));  //-- number of bits shall comprise a byte
   1.767 +
   1.768 +    __ASSERT_ALWAYS(iNumWords && (aStartBit+aNumBits <= (iNumWords << (KBitsInByteLog2+sizeof(TUint32))) ), Panic(EIndexOutOfRange));
   1.769 +
   1.770 +    const TUint32 nBytes = aNumBits >> 3;
   1.771 +    const TUint32 startByte = aStartBit >> 3;
   1.772 +    
   1.773 +    aData.SetLength(nBytes);
   1.774 +    aData.Copy(((const TUint8*)ipData) + startByte, nBytes);
   1.775 +}
   1.776 +
   1.777 +//-----------------------------------------------------------------------------
   1.778 +
   1.779 +/**
   1.780 +    @return number of bits set to '1' in the vector
   1.781 +*/
   1.782 +TUint32 RBitVector::Num1Bits() const
   1.783 +{
   1.784 +    __ASSERT_ALWAYS(ipData, Panic(ENotInitialised));
   1.785 +    if(!iNumBits)
   1.786 +        return 0;
   1.787 +
   1.788 +    TUint32 cntBits = 0;
   1.789 +
   1.790 +    TUint32 wordNum;
   1.791 +    for(wordNum=0; wordNum < iNumWords-1; ++wordNum)
   1.792 +    {
   1.793 +        cntBits += Count1Bits(ipData[wordNum]);
   1.794 +    }
   1.795 +
   1.796 +    //-- process the last word, it shall be masked
   1.797 +    cntBits += Count1Bits(MaskLastWord(ipData[wordNum]));
   1.798 +    
   1.799 +    return cntBits;
   1.800 +}
   1.801 +
   1.802 +//-----------------------------------------------------------------------------
   1.803 +/**
   1.804 +    @return number of bits set to '0' in the vector
   1.805 +*/
   1.806 +TUint32 RBitVector::Num0Bits() const
   1.807 +{
   1.808 +    return iNumBits - Num1Bits();
   1.809 +}
   1.810 +
   1.811 +
   1.812 +//-----------------------------------------------------------------------------
   1.813 +/**
   1.814 +    Calculate number of '1' bits in the range from aIndexFrom to aIndexTo inclusively
   1.815 +
   1.816 +    @param  aIndexFrom  starting index; bit[aIndexFrom] is included to the search
   1.817 +    @param  aIndexTo    ending index;   bit[aIndexTo] is included to the search
   1.818 +    @return number of bits set to '1' in the slice
   1.819 +*/
   1.820 +TUint32 RBitVector::Num1Bits(TUint32 aIndexFrom, TUint32 aIndexTo) const
   1.821 +{
   1.822 +    __ASSERT_ALWAYS(ipData, Panic(ENotInitialised));
   1.823 +    __ASSERT_ALWAYS(aIndexTo < iNumBits && aIndexFrom <= aIndexTo, Panic(EIndexOutOfRange));
   1.824 +
   1.825 +    const TUint32 wordFrom = WordNum(aIndexFrom); //?const
   1.826 +    const TUint32 wordTo   = WordNum(aIndexTo);
   1.827 +
   1.828 +    if(wordFrom == wordTo)
   1.829 +    {//-- the same word
   1.830 +        TUint32 word = ipData[wordFrom];
   1.831 +
   1.832 +        const TUint32 shMaskR = BitInWord(aIndexFrom);
   1.833 +        word >>= shMaskR; word <<= shMaskR; //-- zero low bits
   1.834 +
   1.835 +        const TUint32 shMaskL = 31-BitInWord(aIndexTo);
   1.836 +        word <<= shMaskL;  //-- zero high bits
   1.837 +
   1.838 +        return Count1Bits(word);
   1.839 +    } 
   1.840 +
   1.841 +    TUint32 bitsCnt = 0;
   1.842 +    TUint32 wordsBetween = wordTo - wordFrom - 1;
   1.843 +
   1.844 +    //-- count '1' bits in the termial words
   1.845 +    TUint32 word = ipData[wordFrom];
   1.846 +    const TUint32 shMaskR = BitInWord(aIndexFrom);
   1.847 +    word >>= shMaskR; //-- zero low bits
   1.848 +    bitsCnt += Count1Bits(word);
   1.849 +
   1.850 +    word = ipData[wordTo];
   1.851 +    const TUint32 shMaskL = 31-BitInWord(aIndexTo);
   1.852 +    word <<= shMaskL;  //-- zero high bits
   1.853 +    bitsCnt += Count1Bits(word);
   1.854 +
   1.855 +    //-- count '1' bits in the words between terminal ones
   1.856 +    TUint32 wordIdx = wordFrom+1;
   1.857 +    while(wordsBetween--)
   1.858 +    {
   1.859 +        bitsCnt += Count1Bits(ipData[wordIdx]);
   1.860 +    }
   1.861 +    
   1.862 +
   1.863 +    return bitsCnt;
   1.864 +}
   1.865 +
   1.866 +
   1.867 +
   1.868 +
   1.869 +
   1.870 +
   1.871 +
   1.872 +
   1.873 +
   1.874 +
   1.875 +
   1.876 +
   1.877 +
   1.878 +
   1.879 +
   1.880 +
   1.881 +
   1.882 +
   1.883 +
   1.884 +
   1.885 +
   1.886 +