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 1.339 2007/03/19 11:25:20 danielk1977 Exp $
**
** This file implements a external (disk-based) database using BTrees.
** For a detailed discussion of BTrees, refer to
**
**     Donald E. Knuth, THE ART OF COMPUTER PROGRAMMING, Volume 3:
**     "Sorting And Searching", pages 473-480. Addison-Wesley
**     Publishing Company, Reading, Massachusetts.

  assert( pNewParent!=0 );
  if( pgno==0 ) return SQLITE_OK;
  assert( pBt->pPager!=0 );
  aData = sqlite3pager_lookup(pBt->pPager, pgno);
  if( aData ){
    pThis = (MemPage*)&aData[pBt->pageSize];

    if( pThis->isInit ){
      assert( pThis->aData==aData );
      if( pThis->pParent!=pNewParent ){
        if( pThis->pParent ) sqlite3pager_unref(pThis->pParent->aData);
        pThis->pParent = pNewParent;
        sqlite3pager_ref(pNewParent->aData);
      }
      pThis->idxParent = idx;
    }

** The pager is used to access a database disk file.  It implements
** atomic commit and rollback through the use of a journal file that
** is separate from the database file.  The pager also implements file
** locking to prevent two processes from writing the same database
** file simultaneously, or one process from reading the database while
** another is writing.
**
** @(#) $Id: pager.c,v 1.290 2007/03/19 11:25:20 danielk1977 Exp $
*/
#ifndef SQLITE_OMIT_DISKIO
#include "sqliteInt.h"
#include "os.h"
#include "pager.h"
#include <assert.h>
#include <string.h>

  void *pCodecArg;            /* First argument to xCodec() */
  int nHash;                  /* Size of the pager hash table */
  PgHdr **aHash;              /* Hash table to map page number to PgHdr */
#ifdef SQLITE_ENABLE_MEMORY_MANAGEMENT
  Pager *pNext;               /* Linked list of pagers in this thread */
#endif
  char *pTmpSpace;            /* Pager.pageSize bytes of space for tmp use */
  int doNotSync;
};

/*
** If SQLITE_TEST is defined then increment the variable given in
** the argument
*/
#ifdef SQLITE_TEST

** - 4 bytes: Initial database page count.
** - 4 bytes: Sector size used by the process that wrote this journal.
** 
** Followed by (JOURNAL_HDR_SZ - 24) bytes of unused space.
*/
static int writeJournalHdr(Pager *pPager){
  char zHeader[sizeof(aJournalMagic)+16];
  int rc;

  if( pPager->stmtHdrOff==0 ){
    pPager->stmtHdrOff = pPager->journalOff;
  }

  rc = seekJournalHdr(pPager);
  if( rc ) return rc;

  pPager->journalHdr = pPager->journalOff;



  pPager->journalOff += JOURNAL_HDR_SZ(pPager);

  /* FIX ME: 
  **
  ** Possibly for a pager not in no-sync mode, the journal magic should not
  ** be written until nRec is filled in as part of next syncJournal(). 
  **

    i64 os_szJ;
    rc = sqlite3OsFileSize(pPager->jfd, &os_szJ);
    if( rc!=SQLITE_OK ) return rc;
    assert( szJ==os_szJ );
  }
#endif

  /* Set hdrOff to be the offset just after the end of the last journal
  ** page written before the first journal-header for this statement
  ** transaction was written, or the end of the file if no journal
  ** header was written.
  */
  hdrOff = pPager->stmtHdrOff;
  assert( pPager->fullSync || !hdrOff );
  if( !hdrOff ){
    hdrOff = szJ;
  }

  */
  rc = sqlite3OsSeek(pPager->jfd, pPager->stmtJSize);
  if( rc!=SQLITE_OK ){
    goto end_stmt_playback;
  }
  pPager->journalOff = pPager->stmtJSize;
  pPager->cksumInit = pPager->stmtCksum;

  while( pPager->journalOff < hdrOff ){
    rc = pager_playback_one_page(pPager, pPager->jfd, 1);
    assert( rc!=SQLITE_DONE );
    if( rc!=SQLITE_OK ) goto end_stmt_playback;
  }

  while( pPager->journalOff < szJ ){
    u32 nJRec;         /* Number of Journal Records */

      ** journal file. This is done to avoid ever modifying a journal
      ** header that is involved in the rollback of pages that have
      ** already been written to the database (in case the header is
      ** trashed when the nRec field is updated).
      */
      pPager->nRec = 0;
      assert( pPager->journalOff > 0 );
      assert( pPager->doNotSync==0 );
      rc = writeJournalHdr(pPager);
      if( rc!=0 ){
        return rc;
      }
    }
    pPg = pPager->pFirst;
  }

      pPager->state = PAGER_SHARED;
    }
  }
  if( pPg==0 ){
    /* The requested page is not in the page cache. */
    int h;
    TEST_INCR(pPager->nMiss);
    if( pPager->nPage<pPager->mxPage || pPager->pFirst==0 || MEMDB ||
        (pPager->pFirstSynced==0 && pPager->doNotSync)
    ){
      /* Create a new page */
      if( pPager->nPage>=pPager->nHash ){
        pager_resize_hash_table(pPager,
           pPager->nHash<256 ? 256 : pPager->nHash*2);
        if( pPager->nHash==0 ){
          return SQLITE_NOMEM;
        }

**
** If the journal file could not be written because the disk is full,
** then this routine returns SQLITE_FULL and does an immediate rollback.
** All subsequent write attempts also return SQLITE_FULL until there
** is a call to sqlite3pager_commit() or sqlite3pager_rollback() to
** reset.
*/
static int pager_write(void *pData){
  PgHdr *pPg = DATA_TO_PGHDR(pData);
  Pager *pPager = pPg->pPager;
  int rc = SQLITE_OK;

  /* Check for errors
  */
  if( pPager->errCode ){ 

    pPager->dbSize = pPg->pgno;
    if( !MEMDB && pPager->dbSize==PENDING_BYTE/pPager->pageSize ){
      pPager->dbSize++;
    }
  }
  return rc;
}

/*
** This function is used to mark a data-page as writable. It uses 
** pager_write() to open a journal file (if it is not already open)
** and write the page *pData to the journal.
**
** The difference between this function and pager_write() is that this
** function also deals with the special case where 2 or more pages
** fit on a single disk sector. In this case all co-resident pages
** must have been written to the journal file before returning.
*/
int sqlite3pager_write(void *pData){
  int rc = SQLITE_OK;

  PgHdr *pPg = DATA_TO_PGHDR(pData);
  Pager *pPager = pPg->pPager;
  Pgno nPagePerSector = (pPager->sectorSize/pPager->pageSize);

  if( !MEMDB && nPagePerSector>1 ){
    Pgno nPageCount;          /* Total number of pages in database file */
    Pgno pg1;                 /* First page of the sector pPg is located on. */
    int nPage;                /* Number of pages starting at pg1 to journal */
    int ii;

    /* Set the doNotSync flag to 1. This is because we cannot allow a journal
    ** header to be written between the pages journaled by this function.
    */
    assert( pPager->doNotSync==0 );
    pPager->doNotSync = 1;

    /* This trick assumes that both the page-size and sector-size are
    ** an integer power of 2. It sets variable pg1 to the identifier
    ** of the first page of the sector pPg is located on.
    */
    pg1 = ((pPg->pgno-1) & ~(nPagePerSector-1)) + 1;

    nPageCount = sqlite3pager_pagecount(pPager);
    if( pPg->pgno>nPageCount ){
      nPage = (pPg->pgno - pg1)+1;
    }else if( (pg1+nPagePerSector-1)>nPageCount ){
      nPage = nPageCount+1-pg1;
    }else{
      nPage = nPagePerSector;
    }
    assert(nPage>0);
    assert(pg1<=pPg->pgno);
    assert((pg1+nPage)>pPg->pgno);

    for(ii=0; ii<nPage && rc==SQLITE_OK; ii++){
      Pgno pg = pg1+ii;
      if( !pPager->aInJournal || pg==pPg->pgno || 
          pg>pPager->origDbSize || !(pPager->aInJournal[pg/8]&(1<<(pg&7)))
      ) {
        if( pg!=PAGER_MJ_PGNO(pPager) ){
          void *pPage;
          rc = sqlite3pager_get(pPager, pg, &pPage);
          if( rc==SQLITE_OK ){
            rc = pager_write(pPage);
            sqlite3pager_unref(pPage);
          }
        }
      }
    }

    assert( pPager->doNotSync==1 );
    pPager->doNotSync = 0;
  }else{
    rc = pager_write(pData);
  }
  return rc;
}

/*
** Return TRUE if the page given in the argument was previously passed
** to sqlite3pager_write().  In other words, return TRUE if it is ok
** to change the content of the page.
*/
#ifndef NDEBUG