diff options
author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-07 09:06:44 +0000 |
---|---|---|
committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-07 09:06:44 +0000 |
commit | ed5640d8b587fbcfed7dd7967f3de04b37a76f26 (patch) | |
tree | 7a5f7c6c9d02226d7471cb3cc8fbbf631b415303 /sot/source/sdstor/stgstrms.cxx | |
parent | Initial commit. (diff) | |
download | libreoffice-cb75148ebd0135178ff46f89a30139c44f8d2040.tar.xz libreoffice-cb75148ebd0135178ff46f89a30139c44f8d2040.zip |
Adding upstream version 4:7.4.7.upstream/4%7.4.7upstream
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'sot/source/sdstor/stgstrms.cxx')
-rw-r--r-- | sot/source/sdstor/stgstrms.cxx | 1356 |
1 files changed, 1356 insertions, 0 deletions
diff --git a/sot/source/sdstor/stgstrms.cxx b/sot/source/sdstor/stgstrms.cxx new file mode 100644 index 000000000..717e4e7d1 --- /dev/null +++ b/sot/source/sdstor/stgstrms.cxx @@ -0,0 +1,1356 @@ +/* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- */ +/* + * This file is part of the LibreOffice project. + * + * This Source Code Form is subject to the terms of the Mozilla Public + * License, v. 2.0. If a copy of the MPL was not distributed with this + * file, You can obtain one at http://mozilla.org/MPL/2.0/. + * + * This file incorporates work covered by the following license notice: + * + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed + * with this work for additional information regarding copyright + * ownership. The ASF licenses this file to you under the Apache + * License, Version 2.0 (the "License"); you may not use this file + * except in compliance with the License. You may obtain a copy of + * the License at http://www.apache.org/licenses/LICENSE-2.0 . + */ + +#include <algorithm> + +#include <string.h> +#include <sal/log.hxx> +#include <o3tl/safeint.hxx> +#include <osl/file.hxx> +#include <unotools/tempfile.hxx> + +#include "stgelem.hxx" +#include "stgcache.hxx" +#include "stgstrms.hxx" +#include "stgdir.hxx" +#include "stgio.hxx" +#include <memory> + +///////////////////////////// class StgFAT + +// The FAT class performs FAT operations on an underlying storage stream. +// This stream is either the master FAT stream (m == true ) or a normal +// storage stream, which then holds the FAT for small data allocations. + +StgFAT::StgFAT( StgStrm& r, bool m ) : m_rStrm( r ) +{ + m_bPhys = m; + m_nPageSize = m_rStrm.GetIo().GetPhysPageSize(); + m_nEntries = m_nPageSize >> 2; + m_nOffset = 0; + m_nMaxPage = 0; + m_nLimit = 0; +} + +// Retrieve the physical page for a given byte offset. + +rtl::Reference< StgPage > StgFAT::GetPhysPage( sal_Int32 nByteOff ) +{ + rtl::Reference< StgPage > pPg; + // Position within the underlying stream + // use the Pos2Page() method of the stream + if( m_rStrm.Pos2Page( nByteOff ) ) + { + m_nOffset = m_rStrm.GetOffset(); + sal_Int32 nPhysPage = m_rStrm.GetPage(); + // get the physical page (must be present) + pPg = m_rStrm.GetIo().Get( nPhysPage, true ); + } + return pPg; +} + +// Get the follow page for a certain FAT page. + +sal_Int32 StgFAT::GetNextPage( sal_Int32 nPg ) +{ + if (nPg >= 0) + { + if (nPg > (SAL_MAX_INT32 >> 2)) + return STG_EOF; + rtl::Reference< StgPage > pPg = GetPhysPage( nPg << 2 ); + nPg = pPg.is() ? StgCache::GetFromPage( pPg, m_nOffset >> 2 ) : STG_EOF; + } + return nPg; +} + +// Find the best fit block for the given size. Return +// the starting block and its size or STG_EOF and 0. +// nLastPage is a stopper which tells the current +// underlying stream size. It is treated as a recommendation +// to abort the search to inhibit excessive file growth. + +sal_Int32 StgFAT::FindBlock( sal_Int32& nPgs ) +{ + sal_Int32 nMinStart = STG_EOF, nMinLen = 0; + sal_Int32 nMaxStart = STG_EOF, nMaxLen = 0x7FFFFFFFL; + sal_Int32 nTmpStart = STG_EOF, nTmpLen = 0; + sal_Int32 nPages = m_rStrm.GetSize() >> 2; + bool bFound = false; + rtl::Reference< StgPage > pPg; + short nEntry = 0; + for( sal_Int32 i = 0; i < nPages; i++, nEntry++ ) + { + if( !( nEntry % m_nEntries ) ) + { + // load the next page for that stream + nEntry = 0; + pPg = GetPhysPage( i << 2 ); + if( !pPg.is() ) + return STG_EOF; + } + sal_Int32 nCur = StgCache::GetFromPage( pPg, nEntry ); + if( nCur == STG_FREE ) + { + // count the size of this area + if( nTmpLen ) + nTmpLen++; + else + { + nTmpStart = i; + nTmpLen = 1; + } + if( nTmpLen == nPgs + // If we already did find a block, stop when reaching the limit + || ( bFound && ( nEntry >= m_nLimit ) ) ) + break; + } + else if( nTmpLen ) + { + if( nTmpLen > nPgs && nTmpLen < nMaxLen ) + { + // block > requested size + nMaxLen = nTmpLen; + nMaxStart = nTmpStart; + bFound = true; + } + else if( nTmpLen >= nMinLen ) + { + // block < requested size + nMinLen = nTmpLen; + nMinStart = nTmpStart; + bFound = true; + if( nTmpLen == nPgs ) + break; + } + nTmpStart = STG_EOF; + nTmpLen = 0; + } + } + // Determine which block to use. + if( nTmpLen ) + { + if( nTmpLen > nPgs && nTmpLen < nMaxLen ) + { + // block > requested size + nMaxLen = nTmpLen; + nMaxStart = nTmpStart; + } + else if( nTmpLen >= nMinLen ) + { + // block < requested size + nMinLen = nTmpLen; + nMinStart = nTmpStart; + } + } + if( nMinStart != STG_EOF && nMaxStart != STG_EOF ) + { + // two areas found; return the best fit area + sal_Int32 nMinDiff = nPgs - nMinLen; + sal_Int32 nMaxDiff = nMaxLen - nPgs; + if( nMinDiff > nMaxDiff ) + nMinStart = STG_EOF; + } + if( nMinStart != STG_EOF ) + { + nPgs = nMinLen; return nMinStart; + } + else + { + return nMaxStart; + } +} + +// Set up the consecutive chain for a given block. + +bool StgFAT::MakeChain( sal_Int32 nStart, sal_Int32 nPgs ) +{ + sal_Int32 nPos = nStart << 2; + rtl::Reference< StgPage > pPg = GetPhysPage( nPos ); + if( !pPg.is() || !nPgs ) + return false; + while( --nPgs ) + { + if( m_nOffset >= m_nPageSize ) + { + pPg = GetPhysPage( nPos ); + if( !pPg.is() ) + return false; + } + m_rStrm.GetIo().SetToPage( pPg, m_nOffset >> 2, ++nStart ); + m_nOffset += 4; + nPos += 4; + } + if( m_nOffset >= m_nPageSize ) + { + pPg = GetPhysPage( nPos ); + if( !pPg.is() ) + return false; + } + m_rStrm.GetIo().SetToPage( pPg, m_nOffset >> 2, STG_EOF ); + return true; +} + +// Allocate a block of data from the given page number on. +// It the page number is != STG_EOF, chain the block. + +sal_Int32 StgFAT::AllocPages( sal_Int32 nBgn, sal_Int32 nPgs ) +{ + sal_Int32 nOrig = nBgn; + sal_Int32 nLast = nBgn; + sal_Int32 nBegin = STG_EOF; + sal_Int32 nAlloc; + sal_Int32 nPages = m_rStrm.GetSize() >> 2; + short nPasses = 0; + // allow for two passes + while( nPasses < 2 ) + { + // try to satisfy the request from the pool of free pages + while( nPgs ) + { + nAlloc = nPgs; + nBegin = FindBlock( nAlloc ); + // no more blocks left in present alloc chain + if( nBegin == STG_EOF ) + break; + if( ( nBegin + nAlloc ) > m_nMaxPage ) + m_nMaxPage = nBegin + nAlloc; + if( !MakeChain( nBegin, nAlloc ) ) + return STG_EOF; + if( nOrig == STG_EOF ) + nOrig = nBegin; + else + { + // Patch the chain + rtl::Reference< StgPage > pPg = GetPhysPage( nLast << 2 ); + if( !pPg.is() ) + return STG_EOF; + m_rStrm.GetIo().SetToPage( pPg, m_nOffset >> 2, nBegin ); + } + nLast = nBegin + nAlloc - 1; + nPgs -= nAlloc; + } + if( nPgs && !nPasses ) + { + // we need new, fresh space, so allocate and retry + if( !m_rStrm.SetSize( ( nPages + nPgs ) << 2 ) ) + return STG_EOF; + if( !m_bPhys && !InitNew( nPages ) ) + return 0; + // FIXME: this was originally "FALSE", whether or not that + // makes sense (or should be STG_EOF instead, say?) + nPages = m_rStrm.GetSize() >> 2; + nPasses++; + } + else + break; + } + // now we should have a chain for the complete block + if( nBegin == STG_EOF || nPgs ) + { + m_rStrm.GetIo().SetError( SVSTREAM_FILEFORMAT_ERROR ); + return STG_EOF; // bad structure + } + return nOrig; +} + +// Initialize newly allocated pages for a standard FAT stream +// It can be assumed that the stream size is always on +// a page boundary + +bool StgFAT::InitNew( sal_Int32 nPage1 ) +{ + sal_Int32 n = ( ( m_rStrm.GetSize() >> 2 ) - nPage1 ) / m_nEntries; + if ( n > 0 ) + { + while( n-- ) + { + rtl::Reference< StgPage > pPg; + // Position within the underlying stream + // use the Pos2Page() method of the stream + m_rStrm.Pos2Page( nPage1 << 2 ); + // Initialize the page + pPg = m_rStrm.GetIo().Copy( m_rStrm.GetPage() ); + if ( !pPg.is() ) + return false; + for( short i = 0; i < m_nEntries; i++ ) + m_rStrm.GetIo().SetToPage( pPg, i, STG_FREE ); + nPage1++; + } + } + return true; +} + +// Release a chain + +bool StgFAT::FreePages( sal_Int32 nStart, bool bAll ) +{ + while( nStart >= 0 ) + { + rtl::Reference< StgPage > pPg = GetPhysPage( nStart << 2 ); + if( !pPg.is() ) + return false; + nStart = StgCache::GetFromPage( pPg, m_nOffset >> 2 ); + // The first released page is either set to EOF or FREE + m_rStrm.GetIo().SetToPage( pPg, m_nOffset >> 2, bAll ? STG_FREE : STG_EOF ); + bAll = true; + } + return true; +} + +///////////////////////////// class StgStrm + +// The base stream class provides basic functionality for seeking +// and accessing the data on a physical basis. It uses the built-in +// FAT class for the page allocations. + +StgStrm::StgStrm( StgIo& r ) + : m_nPos(0), + m_bBytePosValid(true), + m_rIo(r), + m_pEntry(nullptr), + m_nStart(STG_EOF), + m_nSize(0), + m_nPage(STG_EOF), + m_nOffset(0), + m_nPageSize(m_rIo.GetPhysPageSize()) +{ +} + +StgStrm::~StgStrm() +{ +} + +// Attach the stream to the given entry. + +void StgStrm::SetEntry( StgDirEntry& r ) +{ + r.m_aEntry.SetLeaf( STG_DATA, m_nStart ); + r.m_aEntry.SetSize( m_nSize ); + m_pEntry = &r; + r.SetDirty(); +} + +/* + * The page chain, is basically a singly linked list of slots each + * point to the next page. Instead of traversing the file structure + * for this each time build a simple flat in-memory vector list + * of pages. + */ +sal_Int32 StgStrm::scanBuildPageChainCache() +{ + if (m_nSize > 0) + { + m_aPagesCache.reserve(m_nSize/m_nPageSize); + m_aUsedPageNumbers.reserve(m_nSize/m_nPageSize); + } + + bool bError = false; + sal_Int32 nBgn = m_nStart; + sal_Int32 nOptSize = 0; + + // Track already scanned PageNumbers here and use them to + // see if an already counted page is re-visited + while( nBgn >= 0 && !bError ) + { + m_aPagesCache.push_back(nBgn); + nBgn = m_pFat->GetNextPage( nBgn ); + + //returned second is false if it already exists + if (!m_aUsedPageNumbers.insert(nBgn).second) + { + SAL_WARN ("sot", "Error: page number " << nBgn << " already in chain for stream"); + bError = true; + } + + nOptSize += m_nPageSize; + } + if (bError) + { + SAL_WARN("sot", "returning wrong format error"); + m_rIo.SetError( ERRCODE_IO_WRONGFORMAT ); + m_aPagesCache.clear(); + m_aUsedPageNumbers.clear(); + } + return nOptSize; +} + +// Compute page number and offset for the given byte position. +// If the position is behind the size, set the stream right +// behind the EOF. +bool StgStrm::Pos2Page( sal_Int32 nBytePos ) +{ + if ( !m_pFat ) + return false; + + // Values < 0 seek to the end + if( nBytePos < 0 || nBytePos >= m_nSize ) + nBytePos = m_nSize; + // Adjust the position back to offset 0 + m_nPos -= m_nOffset; + sal_Int32 nMask = ~( m_nPageSize - 1 ); + sal_Int32 nOld = m_nPos & nMask; + sal_Int32 nNew = nBytePos & nMask; + m_nOffset = static_cast<short>( nBytePos & ~nMask ); + m_nPos = nBytePos; + if (nOld == nNew) + return m_bBytePosValid; + + // See fdo#47644 for a .doc with a vast amount of pages where seeking around the + // document takes a colossal amount of time + + // Please Note: we build the pagescache incrementally as we go if necessary, + // so that a corrupted FAT doesn't poison the stream state for earlier reads + size_t nIdx = nNew / m_nPageSize; + if( nIdx >= m_aPagesCache.size() ) + { + // Extend the FAT cache ! ... + size_t nToAdd = nIdx + 1; + + if (m_aPagesCache.empty()) + { + m_aPagesCache.push_back( m_nStart ); + assert(m_aUsedPageNumbers.empty()); + m_aUsedPageNumbers.insert(m_nStart); + } + + nToAdd -= m_aPagesCache.size(); + + sal_Int32 nBgn = m_aPagesCache.back(); + + // Start adding pages while we can + while (nToAdd > 0 && nBgn >= 0) + { + sal_Int32 nOldBgn = nBgn; + nBgn = m_pFat->GetNextPage(nOldBgn); + if( nBgn >= 0 ) + { + //returned second is false if it already exists + if (!m_aUsedPageNumbers.insert(nBgn).second) + { + SAL_WARN ("sot", "Error: page number " << nBgn << " already in chain for stream"); + break; + } + + //very much the normal case + m_aPagesCache.push_back(nBgn); + --nToAdd; + } + } + } + + if ( nIdx > m_aPagesCache.size() ) + { + SAL_WARN("sot", "seek to index " << nIdx << + " beyond page cache size " << m_aPagesCache.size()); + // fdo#84229 - handle seek to end and back as eg. XclImpStream expects + m_nPage = STG_EOF; + m_nOffset = 0; + // Intriguingly in the past we didn't reset nPos to match the real + // length of the stream thus: + // nIdx = m_aPagesCache.size(); + // nPos = nPageSize * nIdx; + // so retain this behavior for now. + m_bBytePosValid = false; + return false; + } + + // special case: seek to 1st byte of new, unallocated page + // (in case the file size is a multiple of the page size) + if( nBytePos == m_nSize && !m_nOffset && nIdx > 0 && nIdx == m_aPagesCache.size() ) + { + nIdx--; + m_nOffset = m_nPageSize; + } + else if ( nIdx == m_aPagesCache.size() ) + { + m_nPage = STG_EOF; + m_bBytePosValid = false; + return false; + } + + m_nPage = m_aPagesCache[ nIdx ]; + + m_bBytePosValid = m_nPage >= 0; + return m_bBytePosValid; +} + +// Copy an entire stream. Both streams are allocated in the FAT. +// The target stream is this stream. + +bool StgStrm::Copy( sal_Int32 nFrom, sal_Int32 nBytes ) +{ + if ( !m_pFat ) + return false; + + m_aPagesCache.clear(); + m_aUsedPageNumbers.clear(); + + sal_Int32 nTo = m_nStart; + sal_Int32 nPgs = ( nBytes + m_nPageSize - 1 ) / m_nPageSize; + while( nPgs-- ) + { + if( nTo < 0 ) + { + m_rIo.SetError( SVSTREAM_FILEFORMAT_ERROR ); + return false; + } + m_rIo.Copy( nTo, nFrom ); + if( nFrom >= 0 ) + { + nFrom = m_pFat->GetNextPage( nFrom ); + if( nFrom < 0 ) + { + m_rIo.SetError( SVSTREAM_FILEFORMAT_ERROR ); + return false; + } + } + nTo = m_pFat->GetNextPage( nTo ); + } + return true; +} + +bool StgStrm::SetSize( sal_Int32 nBytes ) +{ + if ( nBytes < 0 || !m_pFat ) + return false; + + m_aPagesCache.clear(); + m_aUsedPageNumbers.clear(); + + // round up to page size + sal_Int32 nOld = ( ( m_nSize + m_nPageSize - 1 ) / m_nPageSize ) * m_nPageSize; + sal_Int32 nNew = ( ( nBytes + m_nPageSize - 1 ) / m_nPageSize ) * m_nPageSize; + if( nNew > nOld ) + { + if( !Pos2Page( m_nSize ) ) + return false; + sal_Int32 nBgn = m_pFat->AllocPages( m_nPage, ( nNew - nOld ) / m_nPageSize ); + if( nBgn == STG_EOF ) + return false; + if( m_nStart == STG_EOF ) + m_nStart = m_nPage = nBgn; + } + else if( nNew < nOld ) + { + bool bAll = ( nBytes == 0 ); + if( !Pos2Page( nBytes ) || !m_pFat->FreePages( m_nPage, bAll ) ) + return false; + if( bAll ) + m_nStart = m_nPage = STG_EOF; + } + if( m_pEntry ) + { + // change the dir entry? + if( !m_nSize || !nBytes ) + m_pEntry->m_aEntry.SetLeaf( STG_DATA, m_nStart ); + m_pEntry->m_aEntry.SetSize( nBytes ); + m_pEntry->SetDirty(); + } + m_nSize = nBytes; + m_pFat->SetLimit( GetPages() ); + return true; +} + +// Return the # of allocated pages + + +//////////////////////////// class StgFATStrm + +// The FAT stream class provides physical access to the master FAT. +// Since this access is implemented as a StgStrm, we can use the +// FAT allocator. + +StgFATStrm::StgFATStrm(StgIo& r, sal_Int32 nFatStrmSize) : StgStrm( r ) +{ + m_pFat.reset( new StgFAT( *this, true ) ); + m_nSize = nFatStrmSize; +} + +bool StgFATStrm::Pos2Page( sal_Int32 nBytePos ) +{ + // Values < 0 seek to the end + if( nBytePos < 0 || nBytePos >= m_nSize ) + nBytePos = m_nSize ? m_nSize - 1 : 0; + m_nPage = nBytePos / m_nPageSize; + m_nOffset = static_cast<short>( nBytePos % m_nPageSize ); + m_nPage = GetPage(m_nPage, false); + bool bValid = m_nPage >= 0; + SetPos(nBytePos, bValid); + return bValid; +} + +// Get the page number entry for the given page offset. + +sal_Int32 StgFATStrm::GetPage(sal_Int32 nOff, bool bMake, sal_uInt16 *pnMasterAlloc) +{ + OSL_ENSURE( nOff >= 0, "The offset may not be negative!" ); + if( pnMasterAlloc ) *pnMasterAlloc = 0; + if( nOff < StgHeader::GetFAT1Size() ) + return m_rIo.m_aHdr.GetFATPage( nOff ); + sal_Int32 nMaxPage = m_nSize >> 2; + nOff = nOff - StgHeader::GetFAT1Size(); + // number of master pages that we need to iterate through + sal_uInt16 nMasterCount = ( m_nPageSize >> 2 ) - 1; + sal_uInt16 nBlocks = nOff / nMasterCount; + // offset in the last master page + nOff = nOff % nMasterCount; + + rtl::Reference< StgPage > pOldPage; + rtl::Reference< StgPage > pMaster; + sal_Int32 nFAT = m_rIo.m_aHdr.GetFATChain(); + for( sal_uInt16 nCount = 0; nCount <= nBlocks; nCount++ ) + { + if( nFAT == STG_EOF || nFAT == STG_FREE ) + { + if( bMake ) + { + m_aPagesCache.clear(); + m_aUsedPageNumbers.clear(); + + // create a new master page + nFAT = nMaxPage++; + pMaster = m_rIo.Copy( nFAT ); + if ( pMaster.is() ) + { + for( short k = 0; k < static_cast<short>( m_nPageSize >> 2 ); k++ ) + m_rIo.SetToPage( pMaster, k, STG_FREE ); + // chaining + if( !pOldPage.is() ) + m_rIo.m_aHdr.SetFATChain( nFAT ); + else + m_rIo.SetToPage( pOldPage, nMasterCount, nFAT ); + if( nMaxPage >= m_rIo.GetPhysPages() ) + if( !m_rIo.SetSize( nMaxPage ) ) + return STG_EOF; + // mark the page as used + // make space for Masterpage + if( !pnMasterAlloc ) // create space oneself + { + if( !Pos2Page( nFAT << 2 ) ) + return STG_EOF; + rtl::Reference< StgPage > pPg = m_rIo.Get( m_nPage, true ); + if( !pPg.is() ) + return STG_EOF; + m_rIo.SetToPage( pPg, m_nOffset >> 2, STG_MASTER ); + } + else + (*pnMasterAlloc)++; + m_rIo.m_aHdr.SetMasters( nCount + 1 ); + pOldPage = pMaster; + } + } + } + else + { + pMaster = m_rIo.Get( nFAT, true ); + if ( pMaster.is() ) + { + nFAT = StgCache::GetFromPage( pMaster, nMasterCount ); + pOldPage = pMaster; + } + } + } + if( pMaster.is() ) + return StgCache::GetFromPage( pMaster, nOff ); + m_rIo.SetError( SVSTREAM_GENERALERROR ); + return STG_EOF; +} + + +// Set the page number entry for the given page offset. + +bool StgFATStrm::SetPage( short nOff, sal_Int32 nNewPage ) +{ + OSL_ENSURE( nOff >= 0, "The offset may not be negative!" ); + m_aPagesCache.clear(); + m_aUsedPageNumbers.clear(); + + bool bRes = true; + if( nOff < StgHeader::GetFAT1Size() ) + m_rIo.m_aHdr.SetFATPage( nOff, nNewPage ); + else + { + nOff = nOff - StgHeader::GetFAT1Size(); + // number of master pages that we need to iterate through + sal_uInt16 nMasterCount = ( m_nPageSize >> 2 ) - 1; + sal_uInt16 nBlocks = nOff / nMasterCount; + // offset in the last master page + nOff = nOff % nMasterCount; + + rtl::Reference< StgPage > pMaster; + sal_Int32 nFAT = m_rIo.m_aHdr.GetFATChain(); + for( sal_uInt16 nCount = 0; nCount <= nBlocks; nCount++ ) + { + if( nFAT == STG_EOF || nFAT == STG_FREE ) + { + pMaster = nullptr; + break; + } + pMaster = m_rIo.Get( nFAT, true ); + if ( pMaster.is() ) + nFAT = StgCache::GetFromPage( pMaster, nMasterCount ); + } + if( pMaster.is() ) + m_rIo.SetToPage( pMaster, nOff, nNewPage ); + else + { + m_rIo.SetError( SVSTREAM_GENERALERROR ); + bRes = false; + } + } + + // lock the page against access + if( bRes ) + { + Pos2Page( nNewPage << 2 ); + rtl::Reference< StgPage > pPg = m_rIo.Get( m_nPage, true ); + if( pPg.is() ) + m_rIo.SetToPage( pPg, m_nOffset >> 2, STG_FAT ); + else + bRes = false; + } + return bRes; +} + +bool StgFATStrm::SetSize( sal_Int32 nBytes ) +{ + if ( nBytes < 0 ) + return false; + + m_aPagesCache.clear(); + m_aUsedPageNumbers.clear(); + + // Set the number of entries to a multiple of the page size + short nOld = static_cast<short>( ( m_nSize + ( m_nPageSize - 1 ) ) / m_nPageSize ); + short nNew = static_cast<short>( + ( nBytes + ( m_nPageSize - 1 ) ) / m_nPageSize ) ; + if( nNew < nOld ) + { + // release master pages + for( short i = nNew; i < nOld; i++ ) + SetPage( i, STG_FREE ); + } + else + { + while( nOld < nNew ) + { + // allocate master pages + // find a free master page slot + sal_Int32 nPg = 0; + sal_uInt16 nMasterAlloc = 0; + nPg = GetPage( nOld, true, &nMasterAlloc ); + if( nPg == STG_EOF ) + return false; + // 4 Bytes have been used for Allocation of each MegaMasterPage + nBytes += nMasterAlloc << 2; + + // find a free page using the FAT allocator + sal_Int32 n = 1; + OSL_ENSURE( m_pFat, "The pointer is always initializer here!" ); + sal_Int32 nNewPage = m_pFat->FindBlock( n ); + if( nNewPage == STG_EOF ) + { + // no free pages found; create a new page + // Since all pages are allocated, extend + // the file size for the next page! + nNewPage = m_nSize >> 2; + // if a MegaMasterPage was created avoid taking + // the same Page + nNewPage += nMasterAlloc; + // adjust the file size if necessary + if( nNewPage >= m_rIo.GetPhysPages() ) + if( !m_rIo.SetSize( nNewPage + 1 ) ) + return false; + } + // Set up the page with empty entries + rtl::Reference< StgPage > pPg = m_rIo.Copy( nNewPage ); + if ( !pPg.is() ) + return false; + for( short j = 0; j < static_cast<short>( m_nPageSize >> 2 ); j++ ) + m_rIo.SetToPage( pPg, j, STG_FREE ); + + // store the page number into the master FAT + // Set the size before so the correct FAT can be found + m_nSize = ( nOld + 1 ) * m_nPageSize; + SetPage( nOld, nNewPage ); + + // MegaMasterPages were created, mark it them as used + + sal_uInt32 nMax = m_rIo.m_aHdr.GetMasters( ); + sal_uInt32 nFAT = m_rIo.m_aHdr.GetFATChain(); + if( nMasterAlloc ) + for( sal_uInt32 nCount = 0; nCount < nMax; nCount++ ) + { + if( !Pos2Page( nFAT << 2 ) ) + return false; + if( nMax - nCount <= nMasterAlloc ) + { + rtl::Reference< StgPage > piPg = m_rIo.Get( m_nPage, true ); + if( !piPg.is() ) + return false; + m_rIo.SetToPage( piPg, m_nOffset >> 2, STG_MASTER ); + } + rtl::Reference< StgPage > pPage = m_rIo.Get( nFAT, true ); + if( !pPage.is() ) return false; + nFAT = StgCache::GetFromPage( pPage, (m_nPageSize >> 2 ) - 1 ); + } + + nOld++; + // We have used up 4 bytes for the STG_FAT entry + nBytes += 4; + nNew = static_cast<short>( + ( nBytes + ( m_nPageSize - 1 ) ) / m_nPageSize ); + } + } + m_nSize = nNew * m_nPageSize; + m_rIo.m_aHdr.SetFATSize( nNew ); + return true; +} + +/////////////////////////// class StgDataStrm + +// This class is a normal physical stream which can be initialized +// either with an existing dir entry or an existing FAT chain. +// The stream has a size increment which normally is 1, but which can be +// set to any value is you want the size to be incremented by certain values. + +StgDataStrm::StgDataStrm( StgIo& r, sal_Int32 nBgn, sal_Int32 nLen ) : StgStrm( r ) +{ + Init( nBgn, nLen ); +} + +StgDataStrm::StgDataStrm( StgIo& r, StgDirEntry& p ) : StgStrm( r ) +{ + m_pEntry = &p; + Init( p.m_aEntry.GetLeaf( STG_DATA ), + p.m_aEntry.GetSize() ); +} + +void StgDataStrm::Init( sal_Int32 nBgn, sal_Int32 nLen ) +{ + if ( m_rIo.m_pFAT ) + m_pFat.reset( new StgFAT( *m_rIo.m_pFAT, true ) ); + + OSL_ENSURE( m_pFat, "The pointer should not be empty!" ); + + m_nStart = m_nPage = nBgn; + m_nSize = nLen; + m_nIncr = 1; + m_nOffset = 0; + if( nLen < 0 && m_pFat ) + { + // determine the actual size of the stream by scanning + // the FAT chain and counting the # of pages allocated + m_nSize = scanBuildPageChainCache(); + } +} + +// Set the size of a physical stream. + +bool StgDataStrm::SetSize( sal_Int32 nBytes ) +{ + if ( !m_pFat ) + return false; + + nBytes = ( ( nBytes + m_nIncr - 1 ) / m_nIncr ) * m_nIncr; + sal_Int32 nOldSz = m_nSize; + if( nOldSz != nBytes ) + { + if( !StgStrm::SetSize( nBytes ) ) + return false; + sal_Int32 nMaxPage = m_pFat->GetMaxPage(); + if( nMaxPage > m_rIo.GetPhysPages() ) + if( !m_rIo.SetSize( nMaxPage ) ) + return false; + // If we only allocated one page or less, create this + // page in the cache for faster throughput. The current + // position is the former EOF point. + if( ( m_nSize - 1 ) / m_nPageSize - ( nOldSz - 1 ) / m_nPageSize == 1 ) + { + Pos2Page( nBytes ); + if( m_nPage >= 0 ) + m_rIo.Copy( m_nPage ); + } + } + return true; +} + +// Get the address of the data byte at a specified offset. +// If bForce = true, a read of non-existent data causes +// a read fault. + +void* StgDataStrm::GetPtr( sal_Int32 Pos, bool bDirty ) +{ + if( Pos2Page( Pos ) ) + { + rtl::Reference< StgPage > pPg = m_rIo.Get( m_nPage, true/*bForce*/ ); + if (pPg.is() && m_nOffset < pPg->GetSize()) + { + if( bDirty ) + m_rIo.SetDirty( pPg ); + return static_cast<sal_uInt8 *>(pPg->GetData()) + m_nOffset; + } + } + return nullptr; +} + +// This could easily be adapted to a better algorithm by determining +// the amount of consecutable blocks before doing a read. The result +// is the number of bytes read. No error is generated on EOF. + +sal_Int32 StgDataStrm::Read( void* pBuf, sal_Int32 n ) +{ + if ( n < 0 ) + return 0; + + const auto nAvailable = m_nSize - GetPos(); + if (n > nAvailable) + n = nAvailable; + sal_Int32 nDone = 0; + while( n ) + { + short nBytes = m_nPageSize - m_nOffset; + rtl::Reference< StgPage > pPg; + if( static_cast<sal_Int32>(nBytes) > n ) + nBytes = static_cast<short>(n); + if( nBytes ) + { + short nRes; + void *p = static_cast<sal_uInt8 *>(pBuf) + nDone; + if( nBytes == m_nPageSize ) + { + pPg = m_rIo.Find( m_nPage ); + if( pPg.is() ) + { + // data is present, so use the cached data + memcpy( p, pPg->GetData(), nBytes ); + nRes = nBytes; + } + else + // do a direct (unbuffered) read + nRes = static_cast<short>(m_rIo.Read( m_nPage, p )) * m_nPageSize; + } + else + { + // partial block read through the cache. + pPg = m_rIo.Get( m_nPage, false ); + if( !pPg.is() ) + break; + memcpy( p, static_cast<sal_uInt8*>(pPg->GetData()) + m_nOffset, nBytes ); + nRes = nBytes; + } + nDone += nRes; + SetPos(GetPos() + nRes, true); + n -= nRes; + m_nOffset = m_nOffset + nRes; + if( nRes != nBytes ) + break; // read error or EOF + } + // Switch to next page if necessary + if (m_nOffset >= m_nPageSize && !Pos2Page(GetPos())) + break; + } + return nDone; +} + +sal_Int32 StgDataStrm::Write( const void* pBuf, sal_Int32 n ) +{ + if ( n < 0 ) + return 0; + + sal_Int32 nDone = 0; + if( ( GetPos() + n ) > m_nSize ) + { + sal_Int32 nOld = GetPos(); + if( !SetSize( nOld + n ) ) + return 0; + Pos2Page( nOld ); + } + while( n ) + { + short nBytes = m_nPageSize - m_nOffset; + rtl::Reference< StgPage > pPg; + if( static_cast<sal_Int32>(nBytes) > n ) + nBytes = static_cast<short>(n); + if( nBytes ) + { + short nRes; + const void *p = static_cast<const sal_uInt8 *>(pBuf) + nDone; + if( nBytes == m_nPageSize ) + { + pPg = m_rIo.Find( m_nPage ); + if( pPg.is() ) + { + // data is present, so use the cached data + memcpy( pPg->GetData(), p, nBytes ); + m_rIo.SetDirty( pPg ); + nRes = nBytes; + } + else + // do a direct (unbuffered) write + nRes = static_cast<short>(m_rIo.Write( m_nPage, p )) * m_nPageSize; + } + else + { + // partial block read through the cache. + pPg = m_rIo.Get( m_nPage, false ); + if( !pPg.is() ) + break; + memcpy( static_cast<sal_uInt8*>(pPg->GetData()) + m_nOffset, p, nBytes ); + m_rIo.SetDirty( pPg ); + nRes = nBytes; + } + nDone += nRes; + SetPos(GetPos() + nRes, true); + n -= nRes; + m_nOffset = m_nOffset + nRes; + if( nRes != nBytes ) + break; // read error + } + // Switch to next page if necessary + if( m_nOffset >= m_nPageSize && !Pos2Page(GetPos()) ) + break; + } + return nDone; +} + +//////////////////////////// class StgSmallStream + +// The small stream class provides access to streams with a size < 4096 bytes. +// This stream is a StgStream containing small pages. The FAT for this stream +// is also a StgStream. The start of the FAT is in the header at DataRootPage, +// the stream itself is pointed to by the root entry (it holds start & size). + +StgSmallStrm::StgSmallStrm( StgIo& r, sal_Int32 nBgn ) : StgStrm( r ) +{ + Init( nBgn, 0 ); +} + +StgSmallStrm::StgSmallStrm( StgIo& r, StgDirEntry& p ) : StgStrm( r ) +{ + m_pEntry = &p; + Init( p.m_aEntry.GetLeaf( STG_DATA ), + p.m_aEntry.GetSize() ); +} + +void StgSmallStrm::Init( sal_Int32 nBgn, sal_Int32 nLen ) +{ + if ( m_rIo.m_pDataFAT ) + m_pFat.reset( new StgFAT( *m_rIo.m_pDataFAT, false ) ); + m_pData = m_rIo.m_pDataStrm; + OSL_ENSURE( m_pFat && m_pData, "The pointers should not be empty!" ); + + m_nPageSize = m_rIo.GetDataPageSize(); + m_nStart = + m_nPage = nBgn; + m_nSize = nLen; +} + +// This could easily be adapted to a better algorithm by determining +// the amount of consecutable blocks before doing a read. The result +// is the number of bytes read. No error is generated on EOF. + +sal_Int32 StgSmallStrm::Read( void* pBuf, sal_Int32 n ) +{ + // We can safely assume that reads are not huge, since the + // small stream is likely to be < 64 KBytes. + sal_Int32 nBytePos = GetPos(); + if( ( nBytePos + n ) > m_nSize ) + n = m_nSize - nBytePos; + sal_Int32 nDone = 0; + while( n ) + { + short nBytes = m_nPageSize - m_nOffset; + if( static_cast<sal_Int32>(nBytes) > n ) + nBytes = static_cast<short>(n); + if( nBytes ) + { + if (!m_pData) + break; + sal_Int32 nPos; + if (o3tl::checked_multiply<sal_Int32>(m_nPage, m_nPageSize, nPos)) + break; + if (!m_pData->Pos2Page(nPos + m_nOffset)) + break; + // all reading through the stream + short nRes = static_cast<short>(m_pData->Read( static_cast<sal_uInt8*>(pBuf) + nDone, nBytes )); + nDone += nRes; + SetPos(GetPos() + nRes, true); + n -= nRes; + m_nOffset = m_nOffset + nRes; + // read problem? + if( nRes != nBytes ) + break; + } + // Switch to next page if necessary + if (m_nOffset >= m_nPageSize && !Pos2Page(GetPos())) + break; + } + return nDone; +} + +sal_Int32 StgSmallStrm::Write( const void* pBuf, sal_Int32 n ) +{ + // you can safely assume that reads are not huge, since the + // small stream is likely to be < 64 KBytes. + sal_Int32 nDone = 0; + sal_Int32 nOldPos = GetPos(); + if( ( nOldPos + n ) > m_nSize ) + { + if (!SetSize(nOldPos + n)) + return 0; + Pos2Page(nOldPos); + } + while( n ) + { + short nBytes = m_nPageSize - m_nOffset; + if( static_cast<sal_Int32>(nBytes) > n ) + nBytes = static_cast<short>(n); + if( nBytes ) + { + // all writing goes through the stream + sal_Int32 nDataPos = m_nPage * m_nPageSize + m_nOffset; + if ( !m_pData + || ( m_pData->GetSize() < ( nDataPos + nBytes ) + && !m_pData->SetSize( nDataPos + nBytes ) ) ) + break; + if( !m_pData->Pos2Page( nDataPos ) ) + break; + short nRes = static_cast<short>(m_pData->Write( static_cast<sal_uInt8 const *>(pBuf) + nDone, nBytes )); + nDone += nRes; + SetPos(GetPos() + nRes, true); + n -= nRes; + m_nOffset = m_nOffset + nRes; + // write problem? + if( nRes != nBytes ) + break; + } + // Switch to next page if necessary + if( m_nOffset >= m_nPageSize && !Pos2Page(GetPos()) ) + break; + } + return nDone; +} + +/////////////////////////// class StgTmpStrm + +// The temporary stream uses a memory stream if < 32K, otherwise a +// temporary file. + +#define THRESHOLD 32768L + +StgTmpStrm::StgTmpStrm( sal_uInt64 nInitSize ) + : SvMemoryStream( nInitSize > THRESHOLD + ? 16 + : ( nInitSize ? nInitSize : 16 ), 4096 ) +{ + m_pStrm = nullptr; + // this calls FlushData, so all members should be set by this time + SetBufferSize( 0 ); + if( nInitSize > THRESHOLD ) + SetSize( nInitSize ); +} + +bool StgTmpStrm::Copy( StgTmpStrm& rSrc ) +{ + sal_uInt64 n = rSrc.GetSize(); + const sal_uInt64 nCur = rSrc.Tell(); + SetSize( n ); + if( GetError() == ERRCODE_NONE ) + { + std::unique_ptr<sal_uInt8[]> p(new sal_uInt8[ 4096 ]); + rSrc.Seek( 0 ); + Seek( 0 ); + while( n ) + { + const sal_uInt64 nn = std::min<sal_uInt64>(n, 4096); + if (rSrc.ReadBytes( p.get(), nn ) != nn) + break; + if (WriteBytes( p.get(), nn ) != nn) + break; + n -= nn; + } + p.reset(); + rSrc.Seek( nCur ); + Seek( nCur ); + return n == 0; + } + else + return false; +} + +StgTmpStrm::~StgTmpStrm() +{ + if( m_pStrm ) + { + m_pStrm->Close(); + osl::File::remove( m_aName ); + delete m_pStrm; + } +} + +sal_uInt64 StgTmpStrm::GetSize() const +{ + sal_uInt64 n; + if( m_pStrm ) + { + n = m_pStrm->TellEnd(); + } + else + n = nEndOfData; + return n; +} + +void StgTmpStrm::SetSize(sal_uInt64 n) +{ + if( m_pStrm ) + m_pStrm->SetStreamSize( n ); + else + { + if( n > THRESHOLD ) + { + m_aName = utl::TempFile(nullptr, false).GetURL(); + std::unique_ptr<SvFileStream> s(new SvFileStream( m_aName, StreamMode::READWRITE )); + const sal_uInt64 nCur = Tell(); + sal_uInt64 i = nEndOfData; + std::unique_ptr<sal_uInt8[]> p(new sal_uInt8[ 4096 ]); + if( i ) + { + Seek( 0 ); + while( i ) + { + const sal_uInt64 nb = std::min<sal_uInt64>(i, 4096); + if (ReadBytes(p.get(), nb) == nb + && s->WriteBytes(p.get(), nb) == nb) + i -= nb; + else + break; + } + } + if( !i && n > nEndOfData ) + { + // We have to write one byte at the end of the file + // if the file is bigger than the memstream to see + // if it fits on disk + s->Seek(nEndOfData); + memset(p.get(), 0x00, 4096); + i = n - nEndOfData; + while (i) + { + const sal_uInt64 nb = std::min<sal_uInt64>(i, 4096); + if (s->WriteBytes(p.get(), nb) == nb) + i -= nb; + else + break; // error + } + s->Flush(); + if( s->GetError() != ERRCODE_NONE ) + i = 1; + } + Seek( nCur ); + s->Seek( nCur ); + if( i ) + { + SetError( s->GetError() ); + return; + } + m_pStrm = s.release(); + // Shrink the memory to 16 bytes, which seems to be the minimum + ReAllocateMemory( - ( static_cast<tools::Long>(nEndOfData) - 16 ) ); + } + else + { + if( n > nEndOfData ) + { + SvMemoryStream::SetSize(n); + } + else + nEndOfData = n; + } + } +} + +std::size_t StgTmpStrm::GetData( void* pData, std::size_t n ) +{ + if( m_pStrm ) + { + n = m_pStrm->ReadBytes( pData, n ); + SetError( m_pStrm->GetError() ); + return n; + } + else + return SvMemoryStream::GetData( pData, n ); +} + +std::size_t StgTmpStrm::PutData( const void* pData, std::size_t n ) +{ + sal_uInt32 nCur = Tell(); + sal_uInt32 nNew = nCur + n; + if( nNew > THRESHOLD && !m_pStrm ) + { + SetSize( nNew ); + if( GetError() != ERRCODE_NONE ) + return 0; + } + if( m_pStrm ) + { + nNew = m_pStrm->WriteBytes( pData, n ); + SetError( m_pStrm->GetError() ); + } + else + nNew = SvMemoryStream::PutData( pData, n ); + return nNew; +} + +sal_uInt64 StgTmpStrm::SeekPos(sal_uInt64 n) +{ + // check if a truncated STREAM_SEEK_TO_END was passed + assert(n != SAL_MAX_UINT32); + if( n == STREAM_SEEK_TO_END ) + n = GetSize(); + if( n > THRESHOLD && !m_pStrm ) + { + SetSize( n ); + if( GetError() != ERRCODE_NONE ) + return Tell(); + else + return n; + } + else if( m_pStrm ) + { + n = m_pStrm->Seek( n ); + SetError( m_pStrm->GetError() ); + return n; + } + else + return SvMemoryStream::SeekPos( n ); +} + +void StgTmpStrm::FlushData() +{ + if( m_pStrm ) + { + m_pStrm->Flush(); + SetError( m_pStrm->GetError() ); + } + else + SvMemoryStream::FlushData(); +} + +/* vim:set shiftwidth=4 softtabstop=4 expandtab: */ |