SQLite

/*
** 2004 April 6
**
** The author disclaims copyright to this source code.  In place of
** a legal notice, here is a blessing:
**
**    May you do good and not evil.
**    May you find forgiveness for yourself and forgive others.
**    May you share freely, never taking more than you give.
**
*************************************************************************
** $Id: btree.c,v 1.658 2009/07/08 01:49:12 drh Exp $
**
** This file implements a external (disk-based) database using BTrees.
** See the header comment on "btreeInt.h" for additional information.
** Including a description of file format and an overview of operation.
*/
#include "btreeInt.h"

** This routine works only for pages that do not contain overflow cells.
*/
#define findCell(P,I) \
  ((P)->aData + ((P)->maskPage & get2byte(&(P)->aData[(P)->cellOffset+2*(I)])))

/*
** This a more complex version of findCell() that works for
** pages that do contain overflow cells.
*/
static u8 *findOverflowCell(MemPage *pPage, int iCell){
  int i;
  assert( sqlite3_mutex_held(pPage->pBt->mutex) );
  for(i=pPage->nOverflow-1; i>=0; i--){
    int k;
    struct _OvflCell *pOvfl;

  }else{
    pInfo->nData = 0;
    n += getVarint32(&pCell[n], nPayload);
    pInfo->nKey = nPayload;
  }
  pInfo->nPayload = nPayload;
  pInfo->nHeader = n;
  testcase( nPayload==pPage->maxLocal );
  testcase( nPayload==pPage->maxLocal+1 );
  if( likely(nPayload<=pPage->maxLocal) ){
    /* This is the (easy) common case where the entire payload fits
    ** on the local page.  No overflow is required.
    */
    int nSize;          /* Total size of cell content in bytes */
    nSize = nPayload + n;
    pInfo->nLocal = (u16)nPayload;

    int minLocal;  /* Minimum amount of payload held locally */
    int maxLocal;  /* Maximum amount of payload held locally */
    int surplus;   /* Overflow payload available for local storage */

    minLocal = pPage->minLocal;
    maxLocal = pPage->maxLocal;
    surplus = minLocal + (nPayload - minLocal)%(pPage->pBt->usableSize - 4);
    testcase( surplus==maxLocal );
    testcase( surplus==maxLocal+1 );
    if( surplus <= maxLocal ){
      pInfo->nLocal = (u16)surplus;
    }else{
      pInfo->nLocal = (u16)minLocal;
    }
    pInfo->iOverflow = (u16)(pInfo->nLocal + n);
    pInfo->nSize = pInfo->iOverflow + 4;

    ** past the end of the key value. */
    pEnd = &pIter[9];
    while( (*pIter++)&0x80 && pIter<pEnd );
  }else{
    pIter += getVarint32(pIter, nSize);
  }

  testcase( nSize==pPage->maxLocal );
  testcase( nSize==pPage->maxLocal+1 );
  if( nSize>pPage->maxLocal ){
    int minLocal = pPage->minLocal;
    nSize = minLocal + (nSize - minLocal) % (pPage->pBt->usableSize - 4);
    testcase( nSize==pPage->maxLocal );
    testcase( nSize==pPage->maxLocal+1 );
    if( nSize>pPage->maxLocal ){
      nSize = minLocal;
    }
    nSize += 4;
  }
  nSize += (u32)(pIter - pCell);

  cbrk = usableSize;
  iCellFirst = cellOffset + 2*nCell;
  iCellLast = usableSize - 4;
  for(i=0; i<nCell; i++){
    u8 *pAddr;     /* The i-th cell pointer */
    pAddr = &data[cellOffset + i*2];
    pc = get2byte(pAddr);
    testcase( pc==iCellFirst );
    testcase( pc==iCellLast );
#if !defined(SQLITE_ENABLE_OVERSIZE_CELL_CHECK)
    /* These conditions have already been verified in sqlite3BtreeInitPage()
    ** if SQLITE_ENABLE_OVERSIZE_CELL_CHECK is defined 
    */
    if( pc<iCellFirst || pc>iCellLast ){
      return SQLITE_CORRUPT_BKPT;
    }
#endif
    assert( pc>=iCellFirst && pc<=iCellLast );
    size = cellSizePtr(pPage, &temp[pc]);
    cbrk -= size;
#if defined(SQLITE_ENABLE_OVERSIZE_CELL_CHECK)
    if( cbrk<iCellFirst ){
      return SQLITE_CORRUPT_BKPT;
    }
#else
    if( cbrk<iCellFirst || pc+size>usableSize ){
      return SQLITE_CORRUPT_BKPT;
    }
#endif
    assert( cbrk+size<=usableSize && cbrk>=iCellFirst );
    testcase( cbrk+size==usableSize );
    testcase( cbrk==iCellFirst );
    testcase( pc+size==usableSize );
    memcpy(&data[cbrk], &temp[pc], size);
    put2byte(pAddr, cbrk);
  }
  assert( cbrk>=iCellFirst );
  put2byte(&data[hdr+5], cbrk);
  data[hdr+1] = 0;
  data[hdr+2] = 0;
  data[hdr+7] = 0;
  memset(&data[iCellFirst], 0, cbrk-iCellFirst);
  assert( sqlite3PagerIswriteable(pPage->pDbPage) );
  if( cbrk-iCellFirst!=pPage->nFree ){
    return SQLITE_CORRUPT_BKPT;
  }
  return SQLITE_OK;
}

/*
** Allocate nByte bytes of space from within the B-Tree page passed
** as the first argument. Write into *pIdx the index into pPage->aData[]
** of the first byte of allocated space. Return either SQLITE_OK or
** an error code (usually SQLITE_CORRUPT).
**
** The caller guarantees that there is sufficient space to make the

** allocation.  This routine might need to defragment in order to bring
** all the space together, however.  This routine will avoid using
** the first two bytes past the cell pointer area since presumably this
** allocation is being made in order to insert a new cell, so we will


** also end up needing a new cell pointer.
*/
static int allocateSpace(MemPage *pPage, int nByte, int *pIdx){
  const int hdr = pPage->hdrOffset;    /* Local cache of pPage->hdrOffset */
  u8 * const data = pPage->aData;      /* Local cache of pPage->aData */
  int nFrag;                           /* Number of fragmented bytes on pPage */
  int top;                             /* First byte of cell content area */
  int gap;        /* First byte of gap between cell pointers and cell content */
  int rc;         /* Integer return code */
  
  assert( sqlite3PagerIswriteable(pPage->pDbPage) );
  assert( pPage->pBt );
  assert( sqlite3_mutex_held(pPage->pBt->mutex) );
  assert( nByte>=0 );  /* Minimum cell size is 4 */
  assert( pPage->nFree>=nByte );
  assert( pPage->nOverflow==0 );

  nFrag = data[hdr+7];


  assert( pPage->cellOffset == hdr + 12 - 4*pPage->leaf );
  gap = pPage->cellOffset + 2*pPage->nCell;
  top = get2byte(&data[hdr+5]);
  assert( gap<=top );
  testcase( gap+2==top );
  testcase( gap+1==top );
  testcase( gap==top );


  if( nFrag>=60 ){
    /* Always defragment highly fragmented pages */
    rc = defragmentPage(pPage);
    if( rc ) return rc;
    top = get2byte(&data[hdr+5]);
  }else if( gap+2<=top ){
    /* Search the freelist looking for a free slot big enough to satisfy 
    ** the request. The allocation is made from the first free slot in 
    ** the list that is large enough to accomadate it.
    */
    int pc, addr;
    for(addr=hdr+1; (pc = get2byte(&data[addr]))>0; addr=pc){
      int size = get2byte(&data[pc+2]);     /* Size of free slot */
      if( size>=nByte ){
        int x = size - nByte;
        testcase( x==4 );
        testcase( x==3 );
        if( x<4 ){
          /* Remove the slot from the free-list. Update the number of
          ** fragmented bytes within the page. */
          memcpy(&data[addr], &data[pc], 2);
          data[hdr+7] = (u8)(nFrag + x);
        }else{
          /* The slot remains on the free-list. Reduce its size to account
          ** for the portion used by the new allocation. */
          put2byte(&data[pc+2], x);
        }
        *pIdx = pc + x;
        return SQLITE_OK;
      }
    }
  }

  /* Check to make sure there is enough space in the gap to satisfy
  ** the allocation.  If not, defragment.
  */
  testcase( gap+2+nByte==top );
  if( gap+2+nByte>top ){
    rc = defragmentPage(pPage);
    if( rc ) return rc;
    top = get2byte(&data[hdr+5]);
    assert( gap+nByte<=top );
  }


  /* Allocate memory from the gap in between the cell pointer array
  ** and the cell content area.
  */
  top -= nByte;
  put2byte(&data[hdr+5], top);
  *pIdx = top;
  return SQLITE_OK;
}

/*
** Return a section of the pPage->aData to the freelist.
** The first byte of the new free block is pPage->aDisk[start]
** and the size of the block is "size" bytes.
**
** Most of the effort here is involved in coalesing adjacent
** free blocks into a single big free block.
*/
static int freeSpace(MemPage *pPage, int start, int size){
  int addr, pbegin, hdr;
  int iLast;                        /* Largest possible freeblock offset */
  unsigned char *data = pPage->aData;

  assert( pPage->pBt!=0 );
  assert( sqlite3PagerIswriteable(pPage->pDbPage) );
  assert( start>=pPage->hdrOffset+6+(pPage->leaf?0:4) );
  assert( (start + size)<=pPage->pBt->usableSize );
  assert( sqlite3_mutex_held(pPage->pBt->mutex) );
  assert( size>=0 );   /* Minimum cell size is 4 */

#ifdef SQLITE_SECURE_DELETE
  /* Overwrite deleted information with zeros when the SECURE_DELETE 
  ** option is enabled at compile-time */
  memset(&data[start], 0, size);
#endif

  /* Add the space back into the linked list of freeblocks.  Note that
  ** even though the freeblock list was checked by sqlite3BtreeInitPage(),
  ** sqlite3BtreeInitPage() did not detect overlapping freeblocks or
  ** freeblocks that overlapped cells.  If there was overlap then
  ** subsequent insert operations might have corrupted the freelist.
  ** So we do need to check for corruption while scanning the freelist.
  */
  hdr = pPage->hdrOffset;
  addr = hdr + 1;
  iLast = pPage->pBt->usableSize - 4;
  while( (pbegin = get2byte(&data[addr]))<start && pbegin>0 ){

    if( pbegin>iLast || pbegin<addr+4 ){
      return SQLITE_CORRUPT_BKPT;
    }
    addr = pbegin;
  }
  if( pbegin>iLast ){
    return SQLITE_CORRUPT_BKPT;
  }
  assert( pbegin>addr || pbegin==0 );
  put2byte(&data[addr], start);
  put2byte(&data[start], pbegin);
  put2byte(&data[start+2], size);
  pPage->nFree = pPage->nFree + (u16)size;

  /* Coalesce adjacent free blocks */
  addr = hdr + 1;
  while( (pbegin = get2byte(&data[addr]))>0 ){
    int pnext, psize, x;
    assert( pbegin>addr );
    assert( pbegin<=pPage->pBt->usableSize-4 );
    pnext = get2byte(&data[pbegin]);
    psize = get2byte(&data[pbegin+2]);
    if( pbegin + psize + 3 >= pnext && pnext>0 ){
      int frag = pnext - (pbegin+psize);
      if( (frag<0) || (frag>(int)data[hdr+7]) ){
        return SQLITE_CORRUPT_BKPT;
      }
      data[hdr+7] -= (u8)frag;
      x = get2byte(&data[pnext]);
      put2byte(&data[pbegin], x);
      x = pnext + get2byte(&data[pnext+2]) - pbegin;
      put2byte(&data[pbegin+2], x);
    }else{
      addr = pbegin;
    }

    u8 hdr;            /* Offset to beginning of page header */
    u8 *data;          /* Equal to pPage->aData */
    BtShared *pBt;        /* The main btree structure */
    u16 usableSize;    /* Amount of usable space on each page */
    u16 cellOffset;    /* Offset from start of page to first cell pointer */
    u16 nFree;         /* Number of unused bytes on the page */
    u16 top;           /* First byte of the cell content area */
    int iCellFirst;    /* First allowable cell or freeblock offset */
    int iCellLast;     /* Last possible cell or freeblock offset */

    pBt = pPage->pBt;

    hdr = pPage->hdrOffset;
    data = pPage->aData;
    if( decodeFlags(pPage, data[hdr]) ) return SQLITE_CORRUPT_BKPT;
    assert( pBt->pageSize>=512 && pBt->pageSize<=32768 );

    /* A malformed database page might cause use to read past the end
    ** of page when parsing a cell.  
    **
    ** The following block of code checks early to see if a cell extends
    ** past the end of a page boundary and causes SQLITE_CORRUPT to be 
    ** returned if it does.
    */
    iCellFirst = cellOffset + 2*pPage->nCell;
    iCellLast = usableSize - 4;
#if defined(SQLITE_ENABLE_OVERSIZE_CELL_CHECK)
    {


      int i;            /* Index into the cell pointer array */
      int sz;           /* Size of a cell */



      if( !pPage->leaf ) iCellLast--;
      for(i=0; i<pPage->nCell; i++){
        pc = get2byte(&data[cellOffset+i*2]);
        testcase( pc==iCellFirst );
        testcase( pc==iCellLast );
        if( pc<iCellFirst || pc>iCellLast ){
          return SQLITE_CORRUPT_BKPT;
        }
        sz = cellSizePtr(pPage, &data[pc]);
        testcase( pc+sz==usableSize );
        if( pc+sz>usableSize ){
          return SQLITE_CORRUPT_BKPT;
        }
      }
      if( !pPage->leaf ) iCellLast++;
    }  
#endif

    /* Compute the total free space on the page */
    pc = get2byte(&data[hdr+1]);
    nFree = data[hdr+7] + top;
    while( pc>0 ){
      u16 next, size;
      if( pc<iCellFirst || pc>iCellLast ){
        /* Free block is off the page */
        return SQLITE_CORRUPT_BKPT; 
      }
      next = get2byte(&data[pc]);
      size = get2byte(&data[pc+2]);
      if( next>0 && next<=pc+size+3 ){
        /* Free blocks must be in ascending order */
        return SQLITE_CORRUPT_BKPT; 
      }
      nFree = nFree + size;
      pc = next;
    }

    /* At this point, nFree contains the sum of the offset to the start
    ** of the cell-content area plus the number of free bytes within
    ** the cell-content area. If this is greater than the usable-size
    ** of the page, then the page must be corrupted. This check also
    ** serves to verify that the offset to the start of the cell-content
    ** area, according to the page header, lies within the page.
    */
    if( nFree>usableSize ){
      return SQLITE_CORRUPT_BKPT; 
    }
    pPage->nFree = nFree - iCellFirst;





















    pPage->isInit = 1;
  }
  return SQLITE_OK;
}

/*
** Set up a raw page so that it looks like a database page holding

static int btreeCursor(
  Btree *p,                              /* The btree */
  int iTable,                            /* Root page of table to open */
  int wrFlag,                            /* 1 to write. 0 read-only */
  struct KeyInfo *pKeyInfo,              /* First arg to comparison function */
  BtCursor *pCur                         /* Space for new cursor */
){

  BtShared *pBt = p->pBt;                /* Shared b-tree handle */

  assert( sqlite3BtreeHoldsMutex(p) );
  assert( wrFlag==0 || wrFlag==1 );

  /* The following assert statements verify that if this is a sharable 
  ** b-tree database, the connection is holding the required table locks, 

  u8 *pCell,        /* Content of the new cell */
  int sz,           /* Bytes of content in pCell */
  u8 *pTemp,        /* Temp storage space for pCell, if needed */
  Pgno iChild       /* If non-zero, replace first 4 bytes with this value */
){
  int idx;          /* Where to write new cell content in data[] */
  int j;            /* Loop counter */

  int end;          /* First byte past the last cell pointer in data[] */
  int ins;          /* Index in data[] where new cell pointer is inserted */

  int cellOffset;   /* Address of first cell pointer in data[] */
  u8 *data;         /* The content of the whole page */
  u8 *ptr;          /* Used for moving information around in data[] */

  int nSkip = (iChild ? 4 : 0);

  assert( i>=0 && i<=pPage->nCell+pPage->nOverflow );

  }else{
    int rc = sqlite3PagerWrite(pPage->pDbPage);
    if( rc!=SQLITE_OK ){
      return rc;
    }
    assert( sqlite3PagerIswriteable(pPage->pDbPage) );
    data = pPage->aData;


    cellOffset = pPage->cellOffset;
    end = cellOffset + 2*pPage->nCell;
    ins = cellOffset + 2*i;

    rc = allocateSpace(pPage, sz, &idx);

    if( rc ) return rc;





    assert( idx>=end+2 );

    if( idx+sz > pPage->pBt->usableSize ){
      return SQLITE_CORRUPT_BKPT;
    }
    pPage->nCell++;
    pPage->nFree -= (u16)(2 + sz);
    memcpy(&data[idx+nSkip], pCell+nSkip, sz-nSkip);
    if( iChild ){
      put4byte(&data[idx], iChild);
    }
    for(j=end, ptr=&data[j]; j>ins; j-=2, ptr-=2){
      ptr[0] = ptr[-2];
      ptr[1] = ptr[-1];
    }
    put2byte(&data[ins], idx);
    put2byte(&data[pPage->hdrOffset+3], pPage->nCell);
#ifndef SQLITE_OMIT_AUTOVACUUM
    if( pPage->pBt->autoVacuum ){
      /* The cell may contain a pointer to an overflow page. If so, write
      ** the entry for the overflow page into the pointer map.
      */
      return ptrmapPutOvflPtr(pPage, pCell);
    }