SQLite

        }
      }
    }
  
    if( rc!=SQLITE_OK ){
      unlockBtreeIfUnused(pBt);
    }
  }while( (rc&0xFF)==SQLITE_BUSY && pBt->inTransaction==TRANS_NONE &&
          btreeInvokeBusyHandler(pBt) );

  if( rc==SQLITE_OK ){
    if( p->inTrans==TRANS_NONE ){
      pBt->nTransaction++;
#ifndef SQLITE_OMIT_SHARED_CACHE
      if( p->sharable ){

}
int sqlite3OsShmGet(sqlite3_file *id,int reqSize,int *pSize,void volatile **pp){
  return id->pMethods->xShmGet(id, reqSize, pSize, pp);
}
int sqlite3OsShmRelease(sqlite3_file *id){
  return id->pMethods->xShmRelease(id);
}
int sqlite3OsShmLock(sqlite3_file *id, int offset, int n, int flags){
  return id->pMethods->xShmLock(id, offset, n, flags);
}
void sqlite3OsShmBarrier(sqlite3_file *id){
  id->pMethods->xShmBarrier(id);
}
int sqlite3OsShmClose(sqlite3_file *id, int deleteFlag){
  return id->pMethods->xShmClose(id, deleteFlag);
}

#define SQLITE_FCNTL_DB_UNCHANGED 0xca093fa0
int sqlite3OsSectorSize(sqlite3_file *id);
int sqlite3OsDeviceCharacteristics(sqlite3_file *id);
int sqlite3OsShmOpen(sqlite3_file *id);
int sqlite3OsShmSize(sqlite3_file *id, int, int*);
int sqlite3OsShmGet(sqlite3_file *id, int, int*, void volatile**);
int sqlite3OsShmRelease(sqlite3_file *id);
int sqlite3OsShmLock(sqlite3_file *id, int, int, int);
void sqlite3OsShmBarrier(sqlite3_file *id);
int sqlite3OsShmClose(sqlite3_file *id, int);

/* 
** Functions for accessing sqlite3_vfs methods 
*/
int sqlite3OsOpen(sqlite3_vfs *, const char *, sqlite3_file*, int, int *);

 */
static int semClose(sqlite3_file *id) {
  if( id ){
    unixFile *pFile = (unixFile*)id;
    semUnlock(id, NO_LOCK);
    assert( pFile );
    unixEnterMutex();
    releaseInodeInfo(pFile->pInode);
    unixLeaveMutex();
    closeUnixFile(id);
  }
  return SQLITE_OK;
}

#endif /* OS_VXWORKS */

      /* If there are outstanding locks, do not actually close the file just
      ** yet because that would clear those locks.  Instead, add the file
      ** descriptor to pInode->aPending.  It will be automatically closed when
      ** the last lock is cleared.
      */
      setPendingFd(pFile);
    }
    releaseInodeInfo(pFile->pInode);
    sqlite3_free(pFile->lockingContext);
    rc = closeUnixFile(id);
    unixLeaveMutex();
  }
  return rc;
}

**
** All other fields are read/write.  The unixShm.pFile->mutex must be held
** while accessing any read/write fields.
*/
struct unixShm {
  unixShmNode *pShmNode;     /* The underlying unixShmNode object */
  unixShm *pNext;            /* Next unixShm with the same unixShmNode */

  u8 hasMutex;               /* True if holding the unixShmNode mutex */
  u8 hasMutexBuf;            /* True if holding pFile->mutexBuf */
  u16 sharedMask;            /* Mask of shared locks held */
  u16 exclMask;              /* Mask of exclusive locks held */
#ifdef SQLITE_DEBUG
  u8 id;                     /* Id of this connection within its unixShmNode */
#endif
};

/*





** Constants used for locking
*/



#define UNIX_SHM_BASE   ((18+SQLITE_SHM_NLOCK)*4)         /* first lock byte */

#define UNIX_SHM_DMS    (UNIX_SHM_BASE+SQLITE_SHM_NLOCK)  /* deadman switch */


/*






























** Apply posix advisory locks for all bytes from ofst through ofst+n-1.



**
** Locks block if the mask is exactly UNIX_SHM_C and are non-blocking
** otherwise.
*/
static int unixShmSystemLock(
  unixShmNode *pShmNode, /* Apply locks to this open shared-memory segment */
  int lockType,          /* F_UNLCK, F_RDLCK, or F_WRLCK */
  int ofst,              /* First byte of the locking range */
  int n                  /* Number of bytes to lock */
){
  struct flock f;       /* The posix advisory locking structure */


  int rc = SQLITE_OK;   /* Result code form fcntl() */


  /* Access to the unixShmNode object is serialized by the caller */
  assert( sqlite3_mutex_held(pShmNode->mutex) || pShmNode->nRef==0 );

  /* Shared locks never span more than one byte */
  assert( n==1 || lockType!=F_RDLCK );

  /* Locks are within range */
  assert( n>=1 && n<SQLITE_SHM_NLOCK );

  /* Initialize the locking parameters */
  memset(&f, 0, sizeof(f));
  f.l_type = lockType;
  f.l_whence = SEEK_SET;










  f.l_start = ofst;
  f.l_len = n;












  rc = fcntl(pShmNode->h, F_SETLK, &f);
  rc = (rc!=(-1)) ? SQLITE_OK : SQLITE_BUSY;

  /* Update the global lock state and do debug tracing */
#ifdef SQLITE_DEBUG
  { u16 mask;
  OSTRACE(("SHM-LOCK "));
  mask = (1<<(ofst+n)) - (1<<ofst);
  if( rc==SQLITE_OK ){
    if( lockType==F_UNLCK ){
      OSTRACE(("unlock %d ok", ofst));
      pShmNode->exclMask &= ~mask;
      pShmNode->sharedMask &= ~mask;
    }else if( lockType==F_RDLCK ){
      OSTRACE(("read-lock %d ok", ofst));
      pShmNode->exclMask &= ~mask;
      pShmNode->sharedMask |= mask;
    }else{
      assert( lockType==F_WRLCK );
      OSTRACE(("write-lock %d ok", ofst));
      pShmNode->exclMask |= mask;
      pShmNode->sharedMask &= ~mask;
    }
  }else{
    if( lockType==F_UNLCK ){
      OSTRACE(("unlock %d failed", ofst));
    }else if( lockType==F_RDLCK ){
      OSTRACE(("read-lock failed"));
    }else{
      assert( lockType==F_WRLCK );
      OSTRACE(("write-lock %d failed", ofst));
    }
  }
  OSTRACE((" - afterwards %03x,%03x\n",

           pShmNode->sharedMask, pShmNode->exclMask));

  }
#endif

  return rc;        
}



















































































































/*
** Purge the unixShmNodeList list of all entries with unixShmNode.nRef==0.
**
** This is not a VFS shared-memory method; it is a utility function called
** by VFS shared-memory methods.
*/

      goto shm_open_err;
    }

    /* Check to see if another process is holding the dead-man switch.
    ** If not, truncate the file to zero length. 
    */
    rc = SQLITE_OK;
    if( unixShmSystemLock(pShmNode, F_WRLCK, UNIX_SHM_DMS, 1)==SQLITE_OK ){
      if( ftruncate(pShmNode->h, 0) ){
        rc = SQLITE_IOERR;
      }
    }
    if( rc==SQLITE_OK ){
      rc = unixShmSystemLock(pShmNode, F_RDLCK, UNIX_SHM_DMS, 1);
    }
    if( rc ) goto shm_open_err;
  }

  /* Make the new connection a child of the unixShmNode */
  p->pShmNode = pShmNode;
  p->pNext = pShmNode->pFirst;

  p = pDbFd->pShm;
  if( p==0 ) return SQLITE_OK;
  pShmNode = p->pShmNode;

  assert( pShmNode==pDbFd->pInode->pShmNode );
  assert( pShmNode->pInode==pDbFd->pInode );





  /* Remove connection p from the set of connections associated
  ** with pShmNode */
  sqlite3_mutex_enter(pShmNode->mutex);
  for(pp=&pShmNode->pFirst; (*pp)!=p; pp = &(*pp)->pNext){}
  *pp = p->pNext;

  /* Free the connection p */
  assert( p->hasMutexBuf==0 );
  sqlite3_free(p);
  pDbFd->pShm = 0;
  sqlite3_mutex_leave(pShmNode->mutex);

  /* If pShmNode->nRef has reached 0, then close the underlying
  ** shared-memory file, too */
  unixEnterMutex();

    }
    rc = ftruncate(pShmNode->h, reqSize);
    reqSize = -1;
  }
  return rc;
}

/*
** Release the lock held on the shared memory segment to that other
** threads are free to resize it if necessary.
**
** If the lock is not currently held, this routine is a harmless no-op.
**
** If the shared-memory object is in lock state RECOVER, then we do not
** really want to release the lock, so in that case too, this routine
** is a no-op.
*/
static int unixShmRelease(sqlite3_file *fd){
  unixFile *pDbFd = (unixFile*)fd;
  unixShm *p = pDbFd->pShm;

  if( p->hasMutexBuf ){
    assert( sqlite3_mutex_notheld(p->pShmNode->mutex) );
    sqlite3_mutex_leave(p->pShmNode->mutexBuf);
    p->hasMutexBuf = 0;
  }
  return SQLITE_OK;
}

/*
** Map the shared storage into memory. 
**
** If reqMapSize is positive, then an attempt is made to make the
** mapping at least reqMapSize bytes in size.  However, the mapping
** will never be larger than the size of the underlying shared memory

  unixShm *p = pDbFd->pShm;
  unixShmNode *pShmNode = p->pShmNode;
  int rc = SQLITE_OK;

  assert( pShmNode==pDbFd->pInode->pShmNode );
  assert( pShmNode->pInode==pDbFd->pInode );

  if( p->hasMutexBuf==0 ){
    assert( sqlite3_mutex_notheld(pShmNode->mutex) );
    sqlite3_mutex_enter(pShmNode->mutexBuf);
    p->hasMutexBuf = 1;
  }
  sqlite3_mutex_enter(pShmNode->mutex);
  if( pShmNode->szMap==0 || reqMapSize>pShmNode->szMap ){
    int actualSize;

      pShmNode->pMMapBuf = 0;
      pShmNode->szMap = 0;
    }
  }
  *pNewMapSize = pShmNode->szMap;
  *ppBuf = pShmNode->pMMapBuf;
  sqlite3_mutex_leave(pShmNode->mutex);
  if( *ppBuf==0 ){

    /* Do not hold the mutex if a NULL pointer is being returned. */










    unixShmRelease(fd);


  }





  return rc;
}

















/*
** Change the lock state for a shared-memory segment.
**
** Note that the relationship between SHAREd and EXCLUSIVE locks is a little
** different here than in posix.  In xShmLock(), one can go from unlocked
** to shared and back or from unlocked to exclusive and back.  But one may
** not go from shared to exclusive or from exclusive to shared.
*/
static int unixShmLock(
  sqlite3_file *fd,          /* Database file holding the shared memory */
  int ofst,                  /* First lock to acquire or release */
  int n,                     /* Number of locks to acquire or release */
  int flags                  /* What to do with the lock */
){
  unixFile *pDbFd = (unixFile*)fd;      /* Connection holding shared memory */
  unixShm *p = pDbFd->pShm;             /* The shared memory being locked */
  unixShm *pX;                          /* For looping over all siblings */
  unixShmNode *pShmNode = p->pShmNode;  /* The underlying file iNode */
  int rc = SQLITE_OK;                   /* Result code */
  u16 mask;                             /* Mask of locks to take or release */

  assert( pShmNode==pDbFd->pInode->pShmNode );
  assert( pShmNode->pInode==pDbFd->pInode );
  assert( ofst>=0 && ofst+n<=SQLITE_SHM_NLOCK );




  assert( n>=1 );
  assert( flags==(SQLITE_SHM_LOCK | SQLITE_SHM_SHARED)
       || flags==(SQLITE_SHM_LOCK | SQLITE_SHM_EXCLUSIVE)
       || flags==(SQLITE_SHM_UNLOCK | SQLITE_SHM_SHARED)
       || flags==(SQLITE_SHM_UNLOCK | SQLITE_SHM_EXCLUSIVE) );
  assert( n==1 || (flags & SQLITE_SHM_EXCLUSIVE)!=0 );




  mask = (1<<(ofst+n)) - (1<<ofst);
  assert( n>1 || mask==(1<<ofst) );
  sqlite3_mutex_enter(pShmNode->mutex);
  if( flags & SQLITE_SHM_UNLOCK ){
    u16 allMask = 0; /* Mask of locks held by siblings */



    /* See if any siblings hold this same lock */
    for(pX=pShmNode->pFirst; pX; pX=pX->pNext){
      if( pX==p ) continue;
      assert( (pX->exclMask & (p->exclMask|p->sharedMask))==0 );
      allMask |= pX->sharedMask;
    }

    /* Unlock the system-level locks */
    if( (mask & allMask)==0 ){
      rc = unixShmSystemLock(pShmNode, F_UNLCK, ofst+UNIX_SHM_BASE, n);
    }else{
      rc = SQLITE_OK;
    }

    /* Undo the local locks */
    if( rc==SQLITE_OK ){
      p->exclMask &= ~mask;
      p->sharedMask &= ~mask;
    } 
  }else if( flags & SQLITE_SHM_SHARED ){
    u16 allShared = 0;  /* Union of locks held by connections other than "p" */


    /* Find out which shared locks are already held by sibling connections.
    ** If any sibling already holds an exclusive lock, go ahead and return
    ** SQLITE_BUSY.

    */
    for(pX=pShmNode->pFirst; pX; pX=pX->pNext){
      if( (pX->exclMask & mask)!=0 ){
        rc = SQLITE_BUSY;

        break;
      }






      allShared |= pX->sharedMask;
    }

    /* Get shared locks at the system level, if necessary */
    if( rc==SQLITE_OK ){
      if( (allShared & mask)==0 ){
        rc = unixShmSystemLock(pShmNode, F_RDLCK, ofst+UNIX_SHM_BASE, n);

      }else{

        rc = SQLITE_OK;
      }
    }







    /* Get the local shared locks */





    if( rc==SQLITE_OK ){
      p->sharedMask |= mask;
    }
  }else{
    /* Make sure no sibling connections hold locks that will block this
    ** lock.  If any do, return SQLITE_BUSY right away.
    */
    for(pX=pShmNode->pFirst; pX; pX=pX->pNext){
      if( (pX->exclMask & mask)!=0 || (pX->sharedMask & mask)!=0 ){
        rc = SQLITE_BUSY;
        break;
      }



    }





  


    /* Get the exclusive locks at the system level.  Then if successful
    ** also mark the local connection as being locked.
    */
    if( rc==SQLITE_OK ){











      rc = unixShmSystemLock(pShmNode, F_WRLCK, ofst+UNIX_SHM_BASE, n);
      if( rc==SQLITE_OK ){
        assert( (p->sharedMask & mask)==0 );
        p->exclMask |= mask;
      }

    }
  }
  sqlite3_mutex_leave(pShmNode->mutex);
  OSTRACE(("SHM-LOCK shmid-%d, pid-%d got %03x,%03x\n",
           p->id, getpid(), p->sharedMask, p->exclMask));

  return rc;
}

/*
** Implement a memory barrier or memory fence on shared memory.  
**
** All loads and stores begun before the barrier must complete before
** any load or store begun after the barrier.
*/
static void unixShmBarrier(
  sqlite3_file *fd           /* Database file holding the shared memory */
){



  unixEnterMutex();
  unixLeaveMutex();

}


#else
# define unixShmOpen    0
# define unixShmSize    0
# define unixShmGet     0

  DWORD lastErrno;           /* The Windows errno from the last I/O error */
  int szMap;                 /* Size of the mapping of file into memory */
  char *pMMapBuf;            /* Where currently mmapped().  NULL if unmapped */
  int nRef;                  /* Number of winShm objects pointing to this */
  winShm *pFirst;            /* All winShm objects pointing to this */
  winShmNode *pNext;         /* Next in list of all winShmNode objects */
#ifdef SQLITE_DEBUG


  u8 nextShmId;              /* Next available winShm.id value */
#endif
};

/*
** A global array of all winShmNode objects.
**

**
** The winShm.pFile element is initialized when the object is created
** and is read-only thereafter.
*/
struct winShm {
  winShmNode *pShmNode;      /* The underlying winShmNode object */
  winShm *pNext;             /* Next winShm with the same winShmNode */

  u8 hasMutex;               /* True if holding the winShmNode mutex */
  u8 hasMutexBuf;            /* True if holding pFile->mutexBuf */


#ifdef SQLITE_DEBUG
  u8 id;                     /* Id of this connection with its winShmNode */
#endif
};

/*





** Constants used for locking
*/



#define WIN_SHM_BASE   ((18+SQLITE_SHM_NLOCK)*4)        /* first lock byte */

#define WIN_SHM_DMS    (WIN_SHM_BASE+SQLITE_SHM_NLOCK)  /* deadman switch */


/*






























** Apply advisory locks for all n bytes beginning at ofst.






*/
#define _SHM_UNLCK  1
#define _SHM_RDLCK  2
#define _SHM_WRLCK  3
static int winShmSystemLock(
  winShmNode *pFile,    /* Apply locks to this open shared-memory segment */
  int lockType,         /* _SHM_UNLCK, _SHM_RDLCK, or _SHM_WRLCK */
  int ofst,             /* Offset to first byte to be locked/unlocked */
  int nByte             /* Number of bytes to lock or unlock */
){
  OVERLAPPED ovlp;
  DWORD dwFlags;


  int rc = 0;           /* Result code form Lock/UnlockFileEx() */


  /* Access to the winShmNode object is serialized by the caller */
  assert( sqlite3_mutex_held(pFile->mutex) || pFile->nRef==0 );

  /* Initialize the locking parameters */






  dwFlags = LOCKFILE_FAIL_IMMEDIATELY;





  if( lockType == _SHM_WRLCK ) dwFlags |= LOCKFILE_EXCLUSIVE_LOCK;

  /* Find the first bit in lockMask that is set */


  memset(&ovlp, 0, sizeof(OVERLAPPED));
  ovlp.Offset = ofst;












  /* Release/Acquire the system-level lock */
  if( lockType==_SHM_UNLCK ){

    rc = UnlockFileEx(pFile->hFile.h, 0, nByte, 0, &ovlp);


  }else{




    rc = LockFileEx(pFile->hFile.h, dwFlags, 0, nByte, 0, &ovlp);


  }
  if( !rc ){
    OSTRACE(("SHM-LOCK %d %s ERROR 0x%08lx\n", 
             pFile->hFile.h,
             lockType==_SHM_UNLCK ? "UnlockFileEx" : "LockFileEx",
             GetLastError()));




  }

  rc = (rc!=0) ? SQLITE_OK : SQLITE_BUSY;






























































































































































  return rc;
}

/*
** Purge the winShmNodeList list of all entries with winShmNode.nRef==0.
**
** This is not a VFS shared-memory method; it is a utility function called

      rc = SQLITE_CANTOPEN_BKPT;
      goto shm_open_err;
    }

    /* Check to see if another process is holding the dead-man switch.
    ** If not, truncate the file to zero length. 
    */
    if( winShmSystemLock(pShmNode, _SHM_WRLCK, WIN_SHM_DMS, 1)==SQLITE_OK ){
      rc = winTruncate((sqlite3_file *)&pShmNode->hFile, 0);
    }
    if( rc==SQLITE_OK ){
      winShmSystemLock(pShmNode, _SHM_UNLCK, WIN_SHM_DMS, 1);
      rc = winShmSystemLock(pShmNode, _SHM_RDLCK, WIN_SHM_DMS, 1);
    }
    if( rc ) goto shm_open_err;
  }

  /* Make the new connection a child of the winShmNode */
  p->pShmNode = pShmNode;
  p->pNext = pShmNode->pFirst;
#ifdef SQLITE_DEBUG
  p->id = pShmNode->nextShmId++;
#endif
  pShmNode->pFirst = p;
  pShmNode->nRef++;
  pDbFd->pShm = p;
  winShmLeaveMutex();
  return SQLITE_OK;

  /* Jump here on any error */
shm_open_err:
  winShmSystemLock(pShmNode, _SHM_UNLCK, WIN_SHM_DMS, 1);
  winShmPurge();                 /* This call frees pShmNode if required */
  sqlite3_free(p);
  sqlite3_free(pNew);
  winShmLeaveMutex();
  return rc;
}

  winShmNode *pShmNode; /* The underlying shared-memory file */
  winShm **pp;          /* For looping over sibling connections */

  pDbFd = (winFile*)fd;
  p = pDbFd->pShm;
  pShmNode = p->pShmNode;





  /* Remove connection p from the set of connections associated
  ** with pShmNode */
  sqlite3_mutex_enter(pShmNode->mutex);
  for(pp=&pShmNode->pFirst; (*pp)!=p; pp = &(*pp)->pNext){}
  *pp = p->pNext;

  /* Free the connection p */

  winShmNode *pShmNode = p->pShmNode;
  int rc = SQLITE_OK;

  *pNewSize = 0;
  if( reqSize>=0 ){
    sqlite3_int64 sz;
    rc = winFileSize((sqlite3_file *)&pShmNode->hFile, &sz);
    if( SQLITE_OK==rc && reqSize>sz ){



      rc = winTruncate((sqlite3_file *)&pShmNode->hFile, reqSize);

    }
  }
  if( SQLITE_OK==rc ){
    sqlite3_int64 sz;
    rc = winFileSize((sqlite3_file *)&pShmNode->hFile, &sz);
    if( SQLITE_OK==rc ){
      *pNewSize = (int)sz;

** *ppBuf and *pNewMapSize might be NULL and zero if no space has
** yet been allocated to the underlying storage.
*/
static int winShmGet(
  sqlite3_file *fd,        /* The database file holding the shared memory */
  int reqMapSize,          /* Requested size of mapping. -1 means don't care */
  int *pNewMapSize,        /* Write new size of mapping here */
  void volatile **ppBuf    /* Write mapping buffer origin here */
){
  winFile *pDbFd = (winFile*)fd;
  winShm *p = pDbFd->pShm;
  winShmNode *pShmNode = p->pShmNode;
  int rc = SQLITE_OK;

  if( p->hasMutexBuf==0 ){
    assert( sqlite3_mutex_notheld(pShmNode->mutex) );
    sqlite3_mutex_enter(pShmNode->mutexBuf);
    p->hasMutexBuf = 1;
  }
  sqlite3_mutex_enter(pShmNode->mutex);
  if( pShmNode->szMap==0 || reqMapSize>pShmNode->szMap ){
    int actualSize;

** If the shared-memory object is in lock state RECOVER, then we do not
** really want to release the lock, so in that case too, this routine
** is a no-op.
*/
static int winShmRelease(sqlite3_file *fd){
  winFile *pDbFd = (winFile*)fd;
  winShm *p = pDbFd->pShm;
  if( p->hasMutexBuf ){
    winShmNode *pShmNode = p->pShmNode;
    assert( sqlite3_mutex_notheld(pShmNode->mutex) );
    sqlite3_mutex_leave(pShmNode->mutexBuf);
    p->hasMutexBuf = 0;
  }
  return SQLITE_OK;
}

















/*
** Change the lock state for a shared-memory segment.
*/
static int winShmLock(
  sqlite3_file *fd,          /* Database file holding the shared memory */
  int ofst,                  /* First lock to acquire or release */
  int n,                     /* Number of locks to acquire or release */
  int flags                  /* What to do with the lock */
){
  winFile *pDbFd = (winFile*)fd;
  winShm *p = pDbFd->pShm;
  winShmNode *pShmNode = p->pShmNode;
  int rc = SQLITE_PROTOCOL;



  assert( ofst>=0 && ofst+n<=SQLITE_SHM_NLOCK );

  assert( n>=1 );
  assert( flags==(SQLITE_SHM_LOCK | SQLITE_SHM_SHARED)
       || flags==(SQLITE_SHM_LOCK | SQLITE_SHM_EXCLUSIVE)


       || flags==(SQLITE_SHM_UNLOCK | SQLITE_SHM_SHARED)













       || flags==(SQLITE_SHM_UNLOCK | SQLITE_SHM_EXCLUSIVE) );




  assert( n==1 || (flags & SQLITE_SHM_EXCLUSIVE)!=0 );





  sqlite3_mutex_enter(pShmNode->mutex);

  if( flags & SQLITE_SHM_UNLOCK ){

    rc = winShmSystemLock(pShmNode, _SHM_UNLCK, ofst+WIN_SHM_BASE, n);




  }else if( flags & SQLITE_SHM_SHARED ){







    rc = winShmSystemLock(pShmNode, _SHM_RDLCK, ofst+WIN_SHM_BASE, n);



  }else{
    rc = winShmSystemLock(pShmNode, _SHM_WRLCK, ofst+WIN_SHM_BASE, n);

  }


















































  sqlite3_mutex_leave(pShmNode->mutex);
  OSTRACE(("SHM-LOCK shmid-%d, pid-%d %s\n",

           p->id, (int)GetCurrentProcessId(), rc ? "failed" : "ok"));

  return rc;
}

/*
** Implement a memory barrier or memory fence on shared memory.  
**
** All loads and stores begun before the barrier must complete before

  return (pPager->pWal!=0);
}
#else
# define pagerUseWal(x) 0
# define pagerRollbackWal(x) 0
# define pagerWalFrames(v,w,x,y,z) 0
# define pagerOpenWalIfPresent(z) SQLITE_OK
# define pagerBeginReadTransaction(z) SQLITE_OK
#endif

/*
** Unlock the database file. This function is a no-op if the pager
** is in exclusive mode.
**
** If the pager is currently in error state, discard the contents of 

    ** this happens.  One can argue that this doesn't need to be cleared
    ** until the change-counter check fails in PagerSharedLock().
    ** Clearing the page size cache here is being conservative.
    */
    pPager->dbSizeValid = 0;

    if( pagerUseWal(pPager) ){
      sqlite3WalEndReadTransaction(pPager->pWal);
    }else{
      rc = osUnlock(pPager->fd, NO_LOCK);
    }
    if( rc ){
      pPager->errCode = rc;
    }
    IOTRACE(("UNLOCK %p\n", pPager))

  }
  sqlite3BitvecDestroy(pPager->pInJournal);
  pPager->pInJournal = 0;
  pPager->nRec = 0;
  sqlite3PcacheCleanAll(pPager->pPCache);

  if( pagerUseWal(pPager) ){
    rc2 = sqlite3WalEndWriteTransaction(pPager->pWal);
    pPager->state = PAGER_SHARED;

    /* If the connection was in locking_mode=exclusive mode but is no longer,
    ** drop the EXCLUSIVE lock held on the database file.
    */
    if( rc2==SQLITE_OK 
     && !pPager->exclusiveMode 
     && sqlite3WalExclusiveMode(pPager->pWal, 0) 
    ){

      rc2 = osUnlock(pPager->fd, SHARED_LOCK);
    }
  }else if( !pPager->exclusiveMode ){
    rc2 = osUnlock(pPager->fd, SHARED_LOCK);
    pPager->state = PAGER_SHARED;
    pPager->changeCountDone = 0;
  }else if( pPager->state==PAGER_SYNCED ){

      sqlite3BackupUpdate(pPager->pBackup, p->pgno, (u8 *)p->pData);
    }
  }
  return rc;
}

/*
** Begin a read transaction on the WAL.
**
** This routine used to be called "pagerOpenSnapshot()" because it essentially
** makes a snapshot of the database at the current point in time and preserves
** that snapshot for use by the reader in spite of concurrently changes by
** other writers or checkpointers.
*/
static int pagerBeginReadTransaction(Pager *pPager){
  int rc;                         /* Return code */
  int changed = 0;                /* True if cache must be reset */

  assert( pagerUseWal(pPager) );

  /* sqlite3WalEndReadTransaction() was not called for the previous
  ** transaction in locking_mode=EXCLUSIVE.  So call it now.  If we
  ** are in locking_mode=NORMAL and EndRead() was previously called,
  ** the duplicate call is harmless.
  */
  sqlite3WalEndReadTransaction(pPager->pWal);

  rc = sqlite3WalBeginReadTransaction(pPager->pWal, &changed);
  if( rc==SQLITE_OK ){
    int dummy;
    if( changed ){
      pager_reset(pPager);
      assert( pPager->errCode || pPager->dbSizeValid==0 );
    }
    rc = sqlite3PagerPagecount(pPager, &dummy);

    int isWal;                    /* True if WAL file exists */
    rc = pagerHasWAL(pPager, &isWal);
    if( rc==SQLITE_OK ){
      if( isWal ){
        pager_reset(pPager);
        rc = sqlite3PagerOpenWal(pPager, 0);
        if( rc==SQLITE_OK ){
          rc = pagerBeginReadTransaction(pPager);
        }
      }else if( pPager->journalMode==PAGER_JOURNALMODE_WAL ){
        pPager->journalMode = PAGER_JOURNALMODE_DELETE;
      }
    }
  }
  return rc;

      isErrorReset = 1;
    }
    pPager->errCode = SQLITE_OK;
    pager_reset(pPager);
  }

  if( pagerUseWal(pPager) ){
    rc = pagerBeginReadTransaction(pPager);
  }else if( pPager->state==PAGER_UNLOCK || isErrorReset ){
    sqlite3_vfs * const pVfs = pPager->pVfs;
    int isHotJournal = 0;
    assert( !MEMDB );
    assert( sqlite3PcacheRefCount(pPager->pPCache)==0 );
    if( pPager->noReadlock ){
      assert( pPager->readOnly );

    assert( pPager->pInJournal==0 );
    assert( !MEMDB && !pPager->tempFile );

    if( pagerUseWal(pPager) ){
      /* If the pager is configured to use locking_mode=exclusive, and an
      ** exclusive lock on the database is not already held, obtain it now.
      */
      if( pPager->exclusiveMode && sqlite3WalExclusiveMode(pPager->pWal, -1) ){
        rc = sqlite3OsLock(pPager->fd, EXCLUSIVE_LOCK);
        pPager->state = PAGER_SHARED;
        if( rc!=SQLITE_OK ){
          return rc;
        }
        sqlite3WalExclusiveMode(pPager->pWal, 1);
      }

      /* Grab the write lock on the log file. If successful, upgrade to
      ** PAGER_RESERVED state. Otherwise, return an error code to the caller.
      ** The busy-handler is not invoked if another connection already
      ** holds the write-lock. If possible, the upper layer will call it.
      **
      ** WAL mode sets Pager.state to PAGER_RESERVED when it has an open
      ** transaction, but never to PAGER_EXCLUSIVE. This is because in 
      ** PAGER_EXCLUSIVE state the code to roll back savepoint transactions
      ** may copy data from the sub-journal into the database file as well
      ** as into the page cache. Which would be incorrect in WAL mode.
      */
      rc = sqlite3WalBeginWriteTransaction(pPager->pWal);
      if( rc==SQLITE_OK ){
        pPager->dbOrigSize = pPager->dbSize;
        pPager->state = PAGER_RESERVED;
        pPager->journalOff = 0;
      }

      assert( rc!=SQLITE_OK || pPager->state==PAGER_RESERVED );

*/
int sqlite3PagerCheckpoint(Pager *pPager){
  int rc = SQLITE_OK;
  if( pPager->pWal ){
    u8 *zBuf = (u8 *)pPager->pTmpSpace;
    rc = sqlite3WalCheckpoint(pPager->pWal,
        (pPager->noSync ? 0 : pPager->sync_flags),
        pPager->pageSize, zBuf

    );
  }
  return rc;
}

int sqlite3PagerWalCallback(Pager *pPager){
  return sqlite3WalCallback(pPager->pWal);

#define SQLITE_IOERR_BLOCKED           (SQLITE_IOERR | (11<<8))
#define SQLITE_IOERR_NOMEM             (SQLITE_IOERR | (12<<8))
#define SQLITE_IOERR_ACCESS            (SQLITE_IOERR | (13<<8))
#define SQLITE_IOERR_CHECKRESERVEDLOCK (SQLITE_IOERR | (14<<8))
#define SQLITE_IOERR_LOCK              (SQLITE_IOERR | (15<<8))
#define SQLITE_IOERR_CLOSE             (SQLITE_IOERR | (16<<8))
#define SQLITE_IOERR_DIR_CLOSE         (SQLITE_IOERR | (17<<8))
#define SQLITE_LOCKED_SHAREDCACHE      (SQLITE_LOCKED |  (1<<8))
#define SQLITE_BUSY_RECOVERY           (SQLITE_BUSY   |  (1<<8))

/*
** CAPI3REF: Flags For File Open Operations
**
** These bit values are intended for use in the
** 3rd parameter to the [sqlite3_open_v2()] interface and
** in the 4th parameter to the xOpen method of the

  int (*xSectorSize)(sqlite3_file*);
  int (*xDeviceCharacteristics)(sqlite3_file*);
  /* Methods above are valid for version 1 */
  int (*xShmOpen)(sqlite3_file*);
  int (*xShmSize)(sqlite3_file*, int reqSize, int *pNewSize);
  int (*xShmGet)(sqlite3_file*, int reqSize, int *pSize, void volatile**);
  int (*xShmRelease)(sqlite3_file*);
  int (*xShmLock)(sqlite3_file*, int offset, int n, int flags);
  void (*xShmBarrier)(sqlite3_file*);
  int (*xShmClose)(sqlite3_file*, int deleteFlag);
  /* Methods above are valid for version 2 */
  /* Additional methods may be added in future releases */
};

/*

#define SQLITE_ACCESS_EXISTS    0
#define SQLITE_ACCESS_READWRITE 1
#define SQLITE_ACCESS_READ      2

/*
** CAPI3REF: Flags for the xShmLock VFS method
**
** These integer constants define the various locking operations
** allowed by the xShmLock method of [sqlite3_io_methods].  The
** following are the only legal combinations of flags to the
** xShmLock method:
**
** <ul>
** <li>  SQLITE_SHM_LOCK | SQLITE_SHM_SHARED
** <li>  SQLITE_SHM_LOCK | SQLITE_SHM_EXCLUSIVE
** <li>  SQLITE_SHM_UNLOCK | SQLITE_SHM_SHARED
** <li>  SQLITE_SHM_UNLOCK | SQLITE_SHM_EXCLUSIVE
** </ul>
**
** When unlocking, the same SHARED or EXCLUSIVE flag must be supplied as
** was given no the corresponding lock.  
**
** The xShmLock method can transition between unlocked and SHARED or
** between unlocked and EXCLUSIVE.  It cannot transition between SHARED
** and EXCLUSIVE.
*/
#define SQLITE_SHM_UNLOCK       1
#define SQLITE_SHM_LOCK         2
#define SQLITE_SHM_SHARED       4
#define SQLITE_SHM_EXCLUSIVE    8

/*
** CAPI3REF: Maximum xShmLock index
**
** The xShmLock method on [sqlite3_io_methods] may use values
** between 0 and this upper bound as its "offset" argument.
** The SQLite core will never attempt to acquire or release a
** lock outside of this range
*/
#define SQLITE_SHM_NLOCK        8


/*
** CAPI3REF: Initialize The SQLite Library
**
** ^The sqlite3_initialize() routine initializes the
** SQLite library.  ^The sqlite3_shutdown() routine
** deallocates any resources that were allocated by sqlite3_initialize().

                     Tcl_GetStringFromObj(Tcl_NewIntObj(iArg), 0), " ", 0);
    return TCL_ERROR;
  }
  return TCL_OK;  
}

/*
** tclcmd:   file_control_lockproxy_test DB PWD
**
** This TCL command runs the sqlite3_file_control interface and
** verifies correct operation of the SQLITE_GET_LOCKPROXYFILE and
** SQLITE_SET_LOCKPROXYFILE verbs.
*/
static int file_control_lockproxy_test(
  ClientData clientData, /* Pointer to sqlite3_enable_XXX function */
  Tcl_Interp *interp,    /* The TCL interpreter that invoked this command */
  int objc,              /* Number of arguments */
  Tcl_Obj *CONST objv[]  /* Command arguments */
){
  sqlite3 *db;
  const char *zPwd;
  int nPwd;
  
  if( objc!=3 ){
    Tcl_AppendResult(interp, "wrong # args: should be \"",
                     Tcl_GetStringFromObj(objv[0], 0), " DB PWD", 0);
    return TCL_ERROR;
  }
  if( getDbPointer(interp, Tcl_GetString(objv[1]), &db) ){
   return TCL_ERROR;
  }
  zPwd = Tcl_GetStringFromObj(objv[2], &nPwd);
  
#if !defined(SQLITE_ENABLE_LOCKING_STYLE)
#  if defined(__APPLE__)
#    define SQLITE_ENABLE_LOCKING_STYLE 1
#  else
#    define SQLITE_ENABLE_LOCKING_STYLE 0
#  endif
#endif
#if SQLITE_ENABLE_LOCKING_STYLE && defined(__APPLE__)
  {

    char *testPath;
    int rc;
    char proxyPath[400];
    
    if( sizeof(proxyPath)<nPwd+20 ){
      Tcl_AppendResult(interp, "PWD too big", (void*)0);
      return TCL_ERROR;
    }
    sprintf(proxyPath, "%s/test.proxy", zPwd);
    rc = sqlite3_file_control(db, NULL, SQLITE_SET_LOCKPROXYFILE, proxyPath);
    if( rc ){
      Tcl_SetObjResult(interp, Tcl_NewIntObj(rc)); 
      return TCL_ERROR;
    }
    rc = sqlite3_file_control(db, NULL, SQLITE_GET_LOCKPROXYFILE, &testPath);
    if( strncmp(proxyPath,testPath,11) ){

#if SQLITE_OS_WIN
  extern int sqlite3_os_type;
#endif
#ifdef SQLITE_DEBUG
  extern int sqlite3WhereTrace;
  extern int sqlite3OSTrace;
  extern int sqlite3VdbeAddopTrace;
  extern int sqlite3WalTrace;
#endif
#ifdef SQLITE_TEST
  extern char sqlite3_query_plan[];
  static char *query_plan = sqlite3_query_plan;
#ifdef SQLITE_ENABLE_FTS3
  extern int sqlite3_fts3_enable_parentheses;
#endif

#ifdef SQLITE_DEBUG
  Tcl_LinkVar(interp, "sqlite_addop_trace",
      (char*)&sqlite3VdbeAddopTrace, TCL_LINK_INT);
  Tcl_LinkVar(interp, "sqlite_where_trace",
      (char*)&sqlite3WhereTrace, TCL_LINK_INT);
  Tcl_LinkVar(interp, "sqlite_os_trace",
      (char*)&sqlite3OSTrace, TCL_LINK_INT);
  Tcl_LinkVar(interp, "sqlite_wal_trace",
      (char*)&sqlite3WalTrace, TCL_LINK_INT);
#endif
#ifndef SQLITE_OMIT_DISKIO
  Tcl_LinkVar(interp, "sqlite_opentemp_count",
      (char*)&sqlite3_opentemp_count, TCL_LINK_INT);
#endif
  Tcl_LinkVar(interp, "sqlite_static_bind_value",
      (char*)&sqlite_static_bind_value, TCL_LINK_STRING);

  void volatile **pp
){
  return sqlite3OsShmGet(((CrashFile*)pFile)->pRealFile, reqSize, pSize, pp);
}
static int cfShmRelease(sqlite3_file *pFile){
  return sqlite3OsShmRelease(((CrashFile*)pFile)->pRealFile);
}
static int cfShmLock(sqlite3_file *pFile, int ofst, int n, int flags){
  return sqlite3OsShmLock(((CrashFile*)pFile)->pRealFile, ofst, n, flags);
}
static void cfShmBarrier(sqlite3_file *pFile){
  sqlite3OsShmBarrier(((CrashFile*)pFile)->pRealFile);
}
static int cfShmClose(sqlite3_file *pFile, int delFlag){
  return sqlite3OsShmClose(((CrashFile*)pFile)->pRealFile, delFlag);
}

static int devsymFileControl(sqlite3_file*, int op, void *pArg);
static int devsymSectorSize(sqlite3_file*);
static int devsymDeviceCharacteristics(sqlite3_file*);
static int devsymShmOpen(sqlite3_file*);
static int devsymShmSize(sqlite3_file*,int,int*);
static int devsymShmGet(sqlite3_file*,int,int*,volatile void**);
static int devsymShmRelease(sqlite3_file*);
static int devsymShmLock(sqlite3_file*,int,int,int);
static void devsymShmBarrier(sqlite3_file*);
static int devsymShmClose(sqlite3_file*,int);

/*
** Method declarations for devsym_vfs.
*/
static int devsymOpen(sqlite3_vfs*, const char *, sqlite3_file*, int , int *);

  devsym_file *p = (devsym_file *)pFile;
  return sqlite3OsShmGet(p->pReal, reqSz, pSize, pp);
}
static int devsymShmRelease(sqlite3_file *pFile){
  devsym_file *p = (devsym_file *)pFile;
  return sqlite3OsShmRelease(p->pReal);
}
static int devsymShmLock(sqlite3_file *pFile, int ofst, int n, int flags){
  devsym_file *p = (devsym_file *)pFile;
  return sqlite3OsShmLock(p->pReal, ofst, n, flags);
}
static void devsymShmBarrier(sqlite3_file *pFile){
  devsym_file *p = (devsym_file *)pFile;
  sqlite3OsShmBarrier(p->pReal);
}
static int devsymShmClose(sqlite3_file *pFile, int delFlag){
  devsym_file *p = (devsym_file *)pFile;

static int vfslogSectorSize(sqlite3_file*);
static int vfslogDeviceCharacteristics(sqlite3_file*);

static int vfslogShmOpen(sqlite3_file *pFile);
static int vfslogShmSize(sqlite3_file *pFile, int reqSize, int *pNewSize);
static int vfslogShmGet(sqlite3_file *pFile, int,int*,volatile void **);
static int vfslogShmRelease(sqlite3_file *pFile);
static int vfslogShmLock(sqlite3_file *pFile, int ofst, int n, int flags);
static void vfslogShmBarrier(sqlite3_file*);
static int vfslogShmClose(sqlite3_file *pFile, int deleteFlag);

/*
** Method declarations for vfslog_vfs.
*/
static int vfslogOpen(sqlite3_vfs*, const char *, sqlite3_file*, int , int *);

  VfslogFile *p = (VfslogFile *)pFile;
  t = vfslog_time();
  rc = p->pReal->pMethods->xShmRelease(p->pReal);
  t = vfslog_time() - t;
  vfslog_call(p->pVfslog, OS_SHMRELEASE, p->iFileId, t, rc, 0, 0);
  return rc;
}
static int vfslogShmLock(sqlite3_file *pFile, int ofst, int n, int flags){
  int rc;
  sqlite3_uint64 t;
  VfslogFile *p = (VfslogFile *)pFile;
  t = vfslog_time();
  rc = p->pReal->pMethods->xShmLock(p->pReal, ofst, n, flags);
  t = vfslog_time() - t;
  vfslog_call(p->pVfslog, OS_SHMLOCK, p->iFileId, t, rc, 0, 0);
  return rc;
}
static void vfslogShmBarrier(sqlite3_file *pFile){
  sqlite3_uint64 t;
  VfslogFile *p = (VfslogFile *)pFile;

static int tvfsSleep(sqlite3_vfs*, int microseconds);
static int tvfsCurrentTime(sqlite3_vfs*, double*);

static int tvfsShmOpen(sqlite3_file*);
static int tvfsShmSize(sqlite3_file*, int , int *);
static int tvfsShmGet(sqlite3_file*, int , int *, volatile void **);
static int tvfsShmRelease(sqlite3_file*);
static int tvfsShmLock(sqlite3_file*, int , int, int);
static void tvfsShmBarrier(sqlite3_file*);
static int tvfsShmClose(sqlite3_file*, int);

static sqlite3_io_methods tvfs_io_methods = {
  2,                            /* iVersion */
  tvfsClose,                      /* xClose */
  tvfsRead,                       /* xRead */

  tvfsResultCode(p, &rc);

  return rc;
}

static int tvfsShmLock(
  sqlite3_file *pFile,
  int ofst,
  int n,
  int flags
){
  int rc = SQLITE_OK;
  TestvfsFile *pFd = (TestvfsFile *)pFile;
  Testvfs *p = (Testvfs *)(pFd->pVfs->pAppData);
  int nLock;
  char zLock[80];

  sqlite3_snprintf(sizeof(zLock), zLock, "%d %d", ofst, n);
  nLock = strlen(zLock);
  if( flags & SQLITE_SHM_LOCK ){
    strcpy(&zLock[nLock], " lock");
  }else{
    strcpy(&zLock[nLock], " unlock");
  }
  nLock += strlen(&zLock[nLock]);
  if( flags & SQLITE_SHM_SHARED ){
    strcpy(&zLock[nLock], " shared");
  }else{
    strcpy(&zLock[nLock], " exclusive");
  }
  tvfsExecTcl(p, "xShmLock", 
      Tcl_NewStringObj(pFd->pShm->zFile, -1), pFd->pShmId,
      Tcl_NewStringObj(zLock, -1)
  );
  tvfsResultCode(p, &rc);




  return rc;
}

static void tvfsShmBarrier(sqlite3_file *pFile){
  int rc = SQLITE_OK;
  TestvfsFile *pFd = (TestvfsFile *)pFile;
  Testvfs *p = (Testvfs *)(pFd->pVfs->pAppData);

**   VFSNAME shm FILENAME ?NEWVALUE?
**
** When the xShmLock method is invoked by SQLite, the following script is
** run:
**
**   SCRIPT xShmLock    FILENAME ID LOCK
**
** where LOCK is of the form "OFFSET NBYTE lock/unlock shared/exclusive"


*/
static int testvfs_cmd(
  ClientData cd,
  Tcl_Interp *interp,
  int objc,
  Tcl_Obj *CONST objv[]
){

** implement a loop.  This test used to be on every single instruction,
** but that meant we more testing that we needed.  By only testing the
** flag on jump instructions, we get a (small) speed improvement.
*/
#define CHECK_FOR_INTERRUPT \
   if( db->u1.isInterrupted ) goto abort_due_to_interrupt;


















#ifndef NDEBUG
/*
** This function is only called from within an assert() expression. It
** checks that the sqlite3.nTransaction variable is correctly set to
** the number of non-transaction savepoints currently in the 
** linked list starting at sqlite3.pSavepoint.

  CHECK_FOR_INTERRUPT;
  sqlite3VdbeIOTraceSql(p);
#ifndef SQLITE_OMIT_PROGRESS_CALLBACK
  checkProgress = db->xProgress!=0;
#endif
#ifdef SQLITE_DEBUG
  sqlite3BeginBenignMalloc();
  if( p->pc==0  && (p->db->flags & SQLITE_VdbeListing)!=0 ){


    int i;
    printf("VDBE Program Listing:\n");
    sqlite3VdbePrintSql(p);
    for(i=0; i<p->nOp; i++){
      sqlite3VdbePrintOp(stdout, i, &aOp[i]);
    }
  }



  sqlite3EndBenignMalloc();
#endif
  for(pc=p->pc; rc==SQLITE_OK; pc++){
    assert( pc>=0 && pc<p->nOp );
    if( db->mallocFailed ) goto no_mem;
#ifdef VDBE_PROFILE
    origPc = pc;

    if( p->trace ){
      if( pc==0 ){
        printf("VDBE Execution Trace:\n");
        sqlite3VdbePrintSql(p);
      }
      sqlite3VdbePrintOp(p->trace, pc, pOp);
    }







#endif
      

    /* Check to see if we need to simulate an interrupt.  This only happens
    ** if we have a special test build.
    */
#ifdef SQLITE_TEST

** being considered valid at the same time and being checkpointing together
** following a crash.
**
** READER ALGORITHM
**
** To read a page from the database (call it page number P), a reader
** first checks the WAL to see if it contains page P.  If so, then the
** last valid instance of page P that is a followed by a commit frame
** or is a commit frame itself becomes the value read.  If the WAL
** contains no copies of page P that are valid and which are a commit
** frame or are followed by a commit frame, then page P is read from
** the database file.
**
** To start a read transaction, the reader records the index of the last
** valid frame in the WAL.  The reader uses this recorded "mxFrame" value
** for all subsequent read operations.  New transactions can be appended
** to the WAL, but as long as the reader uses its original mxFrame value
** and ignores the newly appended content, it will see a consistent snapshot
** of the database from a single point in time.  This technique allows
** multiple concurrent readers to view different versions of the database
** content simultaneously.
**
** The reader algorithm in the previous paragraphs works correctly, but 
** because frames for page P can appear anywhere within the WAL, the
** reader has to scan the entire WAL looking for page P frames.  If the
** WAL is large (multiple megabytes is typical) that scan can be slow,
** and read performance suffers.  To overcome this problem, a separate
** data structure called the wal-index is maintained to expedite the
** search for frames of a particular page.
** 

** table is never more than half full.  The expected number of collisions 
** prior to finding a match is 1.  Each entry of the hash table is an
** 1-based index of an entry in the mapping section of the same
** index block.   Let K be the 1-based index of the largest entry in
** the mapping section.  (For index blocks other than the last, K will
** always be exactly HASHTABLE_NPAGE (4096) and for the last index block
** K will be (mxFrame%HASHTABLE_NPAGE).)  Unused slots of the hash table
** contain a value of 0.

**
** To look for page P in the hash table, first compute a hash iKey on
** P as follows:
**
**      iKey = (P * 383) % HASHTABLE_NSLOT
**
** Then start scanning entries of the hash table, starting with iKey

** that correspond to frames greater than the new K value are removed
** from the hash table at this point.
*/
#ifndef SQLITE_OMIT_WAL

#include "wal.h"

/*
** Trace output macros
*/
#if defined(SQLITE_TEST) && defined(SQLITE_DEBUG)
int sqlite3WalTrace = 0;
# define WALTRACE(X)  if(sqlite3WalTrace) sqlite3DebugPrintf X
#else
# define WALTRACE(X)
#endif


/*
** Indices of various locking bytes.   WAL_NREADER is the number
** of available reader locks and should be at least 3.
*/
#define WAL_WRITE_LOCK         0
#define WAL_ALL_BUT_WRITE      1
#define WAL_CKPT_LOCK          1
#define WAL_RECOVER_LOCK       2
#define WAL_READ_LOCK(I)       (3+(I))
#define WAL_NREADER            (SQLITE_SHM_NLOCK-3)


/* Object declarations */
typedef struct WalIndexHdr WalIndexHdr;
typedef struct WalIterator WalIterator;
typedef struct WalCkptInfo WalCkptInfo;


/*
** The following object holds a copy of the wal-index header content.
**
** The actual header in the wal-index consists of two copies of this
** object.
*/
struct WalIndexHdr {
  u32 iChange;                    /* Counter incremented each transaction */
  u8 isInit;                      /* 1 when initialized */
  u8 bigEndCksum;                 /* True if checksums in WAL are big-endian */
  u16 szPage;                     /* Database page size in bytes */
  u32 mxFrame;                    /* Index of last valid frame in the WAL */
  u32 nPage;                      /* Size of database in pages */
  u32 aFrameCksum[2];             /* Checksum of last frame in log */
  u32 aSalt[2];                   /* Two salt values copied from WAL header */
  u32 aCksum[2];                  /* Checksum over all prior fields */
};

/*
** A copy of the following object occurs in the wal-index immediately
** following the second copy of the WalIndexHdr.  This object stores
** information used by checkpoint.
**
** nBackfill is the number of frames in the WAL that have been written
** back into the database. (We call the act of moving content from WAL to
** database "backfilling".)  The nBackfill number is never greater than
** WalIndexHdr.mxFrame.  nBackfill can only be increased by threads
** holding the WAL_CKPT_LOCK lock (which includes a recovery thread).
** However, a WAL_WRITE_LOCK thread can move the value of nBackfill from
** mxFrame back to zero when the WAL is reset.
**
** There is one entry in aReadMark[] for each reader lock.  If a reader
** holds read-lock K, then the value in aReadMark[K] is no greater than
** the mxFrame for that reader.  aReadMark[0] is a special case.  It
** always holds zero.  Readers holding WAL_READ_LOCK(0) always ignore 
** the entire WAL and read all content directly from the database.
**
** The value of aReadMark[K] may only be changed by a thread that
** is holding an exclusive lock on WAL_READ_LOCK(K).  Thus, the value of
** aReadMark[K] cannot changed while there is a reader is using that mark
** since the reader will be holding a shared lock on WAL_READ_LOCK(K).
**
** The checkpointer may only transfer frames from WAL to database where
** the frame numbers are less than or equal to every aReadMark[] that is
** in use (that is, every aReadMark[j] for which there is a corresponding
** WAL_READ_LOCK(j)).  New readers (usually) pick the aReadMark[] with the
** largest value and will increase an unused aReadMark[] to mxFrame if there
** is not already an aReadMark[] equal to mxFrame.  The exception to the
** previous sentence is when nBackfill equals mxFrame (meaning that everything
** in the WAL has been backfilled into the database) then new readers
** will choose aReadMark[0] which has value 0 and hence such reader will
** get all their all content directly from the database file and ignore 
** the WAL.
**
** Writers normally append new frames to the end of the WAL.  However,
** if nBackfill equals mxFrame (meaning that all WAL content has been
** written back into the database) and if no readers are using the WAL
** (in other words, if there are no WAL_READ_LOCK(i) where i>0) then
** the writer will first "reset" the WAL back to the beginning and start
** writing new content beginning at frame 1.
**
** We assume that 32-bit loads are atomic and so no locks are needed in
** order to read from any aReadMark[] entries.
*/
struct WalCkptInfo {
  u32 nBackfill;                  /* Number of WAL frames backfilled into DB */
  u32 aReadMark[WAL_NREADER];     /* Reader marks */
};


/* A block of WALINDEX_LOCK_RESERVED bytes beginning at
** WALINDEX_LOCK_OFFSET is reserved for locks. Since some systems
** only support mandatory file-locks, we do not read or write data
** from the region of the file on which locks are applied.
*/
#define WALINDEX_LOCK_OFFSET   (sizeof(WalIndexHdr)*2 + sizeof(WalCkptInfo))
#define WALINDEX_LOCK_RESERVED 16
#define WALINDEX_HDR_SIZE      (WALINDEX_LOCK_OFFSET+WALINDEX_LOCK_RESERVED)

/* Size of header before each frame in wal */
#define WAL_FRAME_HDRSIZE 24

/* Size of write ahead log header */
#define WAL_HDRSIZE 24

)

/*
** An open write-ahead log file is represented by an instance of the
** following object.
*/
struct Wal {
  sqlite3_vfs *pVfs;         /* The VFS used to create pDbFd */
  sqlite3_file *pDbFd;       /* File handle for the database file */
  sqlite3_file *pWalFd;      /* File handle for WAL file */
  u32 iCallback;             /* Value to pass to log callback (or 0) */
  int szWIndex;              /* Size of the wal-index that is mapped in mem */
  volatile u32 *pWiData;     /* Pointer to wal-index content in memory */
  u16 szPage;                /* Database page size */
  i16 readLock;              /* Which read lock is being held.  -1 for none */
  u8 exclusiveMode;          /* Non-zero if connection is in exclusive mode */
  u8 isWIndexOpen;           /* True if ShmOpen() called on pDbFd */
  u8 writeLock;              /* True if in a write transaction */
  u8 ckptLock;               /* True if holding a checkpoint lock */
  WalIndexHdr hdr;           /* Wal-index header for current transaction */
  char *zWalName;            /* Name of WAL file */

  u32 nCkpt;                 /* Checkpoint sequence counter in the wal-header */
};

/*
** Return a pointer to the WalCkptInfo structure in the wal-index.
*/
static volatile WalCkptInfo *walCkptInfo(Wal *pWal){
  assert( pWal->pWiData!=0 );
  return (volatile WalCkptInfo*)&pWal->pWiData[sizeof(WalIndexHdr)/2];
}


/*
** This structure is used to implement an iterator that loops through
** all frames in the WAL in database page order. Where two or more frames
** correspond to the same database page, the iterator visits only the 
** frame most recently written to the WAL (in other words, the frame with

    }while( aData<aEnd );
  }

  aOut[0] = s1;
  aOut[1] = s2;
}
























/*
** Write the header information in pWal->hdr into the wal-index.
**
** The checksum on pWal->hdr is updated before it is written.
*/
static void walIndexWriteHdr(Wal *pWal){
  WalIndexHdr *aHdr;

  assert( pWal->writeLock );
  pWal->hdr.isInit = 1;
  walChecksumBytes(1, (u8*)&pWal->hdr, offsetof(WalIndexHdr, aCksum),

                   0, pWal->hdr.aCksum);
  aHdr = (WalIndexHdr*)pWal->pWiData;
  memcpy(&aHdr[1], &pWal->hdr, sizeof(WalIndexHdr));
  sqlite3OsShmBarrier(pWal->pDbFd);
  memcpy(&aHdr[0], &pWal->hdr, sizeof(WalIndexHdr));
}

/*
** This function encodes a single frame header and writes it to a buffer
** supplied by the caller. A frame-header is made up of a series of 
** 4-byte big-endian integers, as follows:
**

** create incompatibilities.
*/
#define HASHTABLE_NPAGE      4096  /* Must be power of 2 and multiple of 256 */
#define HASHTABLE_DATATYPE   u16
#define HASHTABLE_HASH_1     383                  /* Should be prime */
#define HASHTABLE_NSLOT      (HASHTABLE_NPAGE*2)  /* Must be a power of 2 */
#define HASHTABLE_NBYTE      (sizeof(HASHTABLE_DATATYPE)*HASHTABLE_NSLOT)

#if defined(SQLITE_TEST) && defined(SQLITE_DEBUG)
/*
** Names of locks.  This routine is used to provide debugging output and is not
** a part of an ordinary build.
*/
static const char *walLockName(int lockIdx){
  if( lockIdx==WAL_WRITE_LOCK ){
    return "WRITE-LOCK";
  }else if( lockIdx==WAL_CKPT_LOCK ){
    return "CKPT-LOCK";
  }else if( lockIdx==WAL_RECOVER_LOCK ){
    return "RECOVER-LOCK";
  }else{
    static char zName[15];
    sqlite3_snprintf(sizeof(zName), zName, "READ-LOCK[%d]",
                     lockIdx-WAL_READ_LOCK(0));
    return zName;
  }
}
#endif /*defined(SQLITE_TEST) || defined(SQLITE_DEBUG) */
    

/*
** Set or release locks on the WAL.  Locks are either shared or exclusive.
** A lock cannot be moved directly between shared and exclusive - it must go
** through the unlocked state first.
**
** In locking_mode=EXCLUSIVE, all of these routines become no-ops.
*/
static int walLockShared(Wal *pWal, int lockIdx){
  int rc;
  if( pWal->exclusiveMode ) return SQLITE_OK;
  rc = sqlite3OsShmLock(pWal->pDbFd, lockIdx, 1,
                        SQLITE_SHM_LOCK | SQLITE_SHM_SHARED);
  WALTRACE(("WAL%p: acquire SHARED-%s %s\n", pWal,
            walLockName(lockIdx), rc ? "failed" : "ok"));
  return rc;
}
static void walUnlockShared(Wal *pWal, int lockIdx){
  if( pWal->exclusiveMode ) return;
  (void)sqlite3OsShmLock(pWal->pDbFd, lockIdx, 1,
                         SQLITE_SHM_UNLOCK | SQLITE_SHM_SHARED);
  WALTRACE(("WAL%p: release SHARED-%s\n", pWal, walLockName(lockIdx)));
}
static int walLockExclusive(Wal *pWal, int lockIdx, int n){
  int rc;
  if( pWal->exclusiveMode ) return SQLITE_OK;
  rc = sqlite3OsShmLock(pWal->pDbFd, lockIdx, n,
                        SQLITE_SHM_LOCK | SQLITE_SHM_EXCLUSIVE);
  WALTRACE(("WAL%p: acquire EXCLUSIVE-%s cnt=%d %s\n", pWal,
            walLockName(lockIdx), n, rc ? "failed" : "ok"));
  return rc;
}
static void walUnlockExclusive(Wal *pWal, int lockIdx, int n){
  if( pWal->exclusiveMode ) return;
  (void)sqlite3OsShmLock(pWal->pDbFd, lockIdx, n,
                         SQLITE_SHM_UNLOCK | SQLITE_SHM_EXCLUSIVE);
  WALTRACE(("WAL%p: release EXCLUSIVE-%s cnt=%d\n", pWal,
             walLockName(lockIdx), n));
}

/*
** Return the index in the Wal.pWiData array that corresponds to 
** frame iFrame.
**
** Wal.pWiData is an array of u32 elements that is the wal-index.
** The array begins with a header and is then followed by alternating

      (WALINDEX_LOCK_OFFSET+WALINDEX_LOCK_RESERVED)/sizeof(u32)
    + (((iFrame-1)/HASHTABLE_NPAGE) * HASHTABLE_NBYTE)/sizeof(u32)
    + (iFrame-1)
  );
}

/*
** Return the minimum size of the shared-memory, in bytes, that is needed
** to support a wal-index containing frame iFrame.  The value returned
** includes the wal-index header and the complete "block" containing iFrame,
** including the hash table segment that follows the block.
*/
static int walMappingSize(u32 iFrame){
  const int nByte = (sizeof(u32)*HASHTABLE_NPAGE + HASHTABLE_NBYTE) ;
  return ( WALINDEX_LOCK_OFFSET 
         + WALINDEX_LOCK_RESERVED 
         + nByte * ((iFrame + HASHTABLE_NPAGE - 1)/HASHTABLE_NPAGE)
  );

**
** If enlargeTo is non-negative, then increase the size of the underlying
** storage to be at least as big as enlargeTo before remapping.
*/
static int walIndexRemap(Wal *pWal, int enlargeTo){
  int rc;
  int sz;
  assert( pWal->writeLock );
  rc = sqlite3OsShmSize(pWal->pDbFd, enlargeTo, &sz);
  if( rc==SQLITE_OK && sz>pWal->szWIndex ){
    walIndexUnmap(pWal);
    rc = walIndexMap(pWal, sz);
  }
  assert( pWal->szWIndex>=enlargeTo || rc!=SQLITE_OK );
  return rc;
}

/*
** Compute a hash on a page number.  The resulting hash value must land
** between 0 and (HASHTABLE_NSLOT-1).  The walHashNext() function advances
** the hash to the next value in the event of a collision.
*/
static int walHash(u32 iPage){
  assert( iPage>0 );
  assert( (HASHTABLE_NSLOT & (HASHTABLE_NSLOT-1))==0 );
  return (iPage*HASHTABLE_HASH_1) & (HASHTABLE_NSLOT-1);
}
static int walNextHash(int iPriorHash){

/*
** Remove entries from the hash table that point to WAL slots greater
** than pWal->hdr.mxFrame.
**
** This function is called whenever pWal->hdr.mxFrame is decreased due
** to a rollback or savepoint.
**
** At most only the hash table containing pWal->hdr.mxFrame needs to be
** updated.  Any later hash tables will be automatically cleared when
** pWal->hdr.mxFrame advances to the point where those hash tables are
** actually needed.
*/
static void walCleanupHash(Wal *pWal){
  volatile HASHTABLE_DATATYPE *aHash;  /* Pointer to hash table to clear */
  volatile u32 *aPgno;                 /* Unused return from walHashFind() */
  u32 iZero;                           /* frame == (aHash[x]+iZero) */
  int iLimit = 0;                      /* Zero values greater than this */

  assert( pWal->writeLock );
  testcase( pWal->hdr.mxFrame==HASHTABLE_NPAGE-1 );
  testcase( pWal->hdr.mxFrame==HASHTABLE_NPAGE );
  testcase( pWal->hdr.mxFrame==HASHTABLE_NPAGE+1 );
  if( (pWal->hdr.mxFrame % HASHTABLE_NPAGE)>0 ){
    int nByte;                    /* Number of bytes to zero in aPgno[] */
    int i;                        /* Used to iterate through aHash[] */

    walHashFind(pWal, pWal->hdr.mxFrame+1, &aHash, &aPgno, &iZero);
    iLimit = pWal->hdr.mxFrame - iZero;
    assert( iLimit>0 );
    for(i=0; i<HASHTABLE_NSLOT; i++){
      if( aHash[i]>iLimit ){
        aHash[i] = 0;
      }
    }

    /* Zero the entries in the aPgno array that correspond to frames with
    ** frame numbers greater than pWal->hdr.mxFrame. 
    */
    nByte = sizeof(u32) * (HASHTABLE_NPAGE-iLimit);
    memset((void *)&aPgno[iZero+iLimit+1], 0, nByte);
    assert( &((u8 *)&aPgno[iZero+iLimit+1])[nByte]==(u8 *)aHash );
  }

#ifdef SQLITE_ENABLE_EXPENSIVE_ASSERT
  /* Verify that the every entry in the mapping region is still reachable
  ** via the hash table even after the cleanup.
  */

  if( iLimit ){
    int i;           /* Loop counter */
    int iKey;        /* Hash key */
    for(i=1; i<=iLimit; i++){
      for(iKey=walHash(aPgno[i+iZero]); aHash[iKey]; iKey=walNextHash(iKey)){
        if( aHash[iKey]==i ) break;
      }
      assert( aHash[iKey]==i );

  return rc;
}


/*
** Recover the wal-index by reading the write-ahead log file. 
**
** This routine first tries to establish an exclusive lock on the
** wal-index to prevent other threads/processes from doing anything
** with the WAL or wal-index while recovery is running.  The
** WAL_RECOVER_LOCK is also held so that other threads will know
** that this thread is running recovery.  If unable to establish
** the necessary locks, this routine returns SQLITE_BUSY.
*/
static int walIndexRecover(Wal *pWal){
  int rc;                         /* Return Code */
  i64 nSize;                      /* Size of log file */
  u32 aFrameCksum[2] = {0, 0};
  int iLock;                      /* Lock offset to lock for checkpoint */
  int nLock;                      /* Number of locks to hold */

  /* Obtain an exclusive lock on all byte in the locking range not already
  ** locked by the caller. The caller is guaranteed to have locked the
  ** WAL_WRITE_LOCK byte, and may have also locked the WAL_CKPT_LOCK byte.
  ** If successful, the same bytes that are locked here are unlocked before
  ** this function returns.
  */
  assert( pWal->ckptLock==1 || pWal->ckptLock==0 );
  assert( WAL_ALL_BUT_WRITE==WAL_WRITE_LOCK+1 );
  assert( WAL_CKPT_LOCK==WAL_ALL_BUT_WRITE );
  assert( pWal->writeLock );
  iLock = WAL_ALL_BUT_WRITE + pWal->ckptLock;
  nLock = SQLITE_SHM_NLOCK - iLock;
  rc = walLockExclusive(pWal, iLock, nLock);
  if( rc ){
    return rc;
  }
  WALTRACE(("WAL%p: recovery begin...\n", pWal));

  memset(&pWal->hdr, 0, sizeof(WalIndexHdr));

  rc = sqlite3OsFileSize(pWal->pWalFd, &nSize);
  if( rc!=SQLITE_OK ){
    goto recovery_error;
  }

  if( nSize>WAL_HDRSIZE ){
    u8 aBuf[WAL_HDRSIZE];         /* Buffer to load WAL header into */
    u8 *aFrame = 0;               /* Malloc'd buffer to load entire frame */
    int szFrame;                  /* Number of bytes in buffer aFrame[] */
    u8 *aData;                    /* Pointer to data part of aFrame buffer */
    int iFrame;                   /* Index of last frame read */
    i64 iOffset;                  /* Next offset to read from log file */
    int szPage;                   /* Page size according to the log */
    u32 magic;                    /* Magic value read from WAL header */

    /* Read in the WAL header. */
    rc = sqlite3OsRead(pWal->pWalFd, aBuf, WAL_HDRSIZE, 0);
    if( rc!=SQLITE_OK ){
      goto recovery_error;
    }

    /* If the database page size is not a power of two, or is greater than
    ** SQLITE_MAX_PAGE_SIZE, conclude that the WAL file contains no valid 
    ** data. Similarly, if the 'magic' value is invalid, ignore the whole
    ** WAL file.
    */

        aBuf, WAL_HDRSIZE, 0, pWal->hdr.aFrameCksum
    );

    /* Malloc a buffer to read frames into. */
    szFrame = szPage + WAL_FRAME_HDRSIZE;
    aFrame = (u8 *)sqlite3_malloc(szFrame);
    if( !aFrame ){
      rc = SQLITE_NOMEM;
      goto recovery_error;
    }
    aData = &aFrame[WAL_FRAME_HDRSIZE];

    /* Read all frames from the log file. */
    iFrame = 0;
    for(iOffset=WAL_HDRSIZE; (iOffset+szFrame)<=nSize; iOffset+=szFrame){
      u32 pgno;                   /* Database page number for frame */

  if( rc==SQLITE_OK && pWal->hdr.mxFrame==0 ){
    rc = walIndexRemap(pWal, walMappingSize(1));
  }
  if( rc==SQLITE_OK ){
    pWal->hdr.aFrameCksum[0] = aFrameCksum[0];
    pWal->hdr.aFrameCksum[1] = aFrameCksum[1];
    walIndexWriteHdr(pWal);

    /* Zero the checkpoint-header. This is safe because this thread is 
    ** currently holding locks that exclude all other readers, writers and
    ** checkpointers.
    */
    memset((void *)walCkptInfo(pWal), 0, sizeof(WalCkptInfo));
  }

recovery_error:
  WALTRACE(("WAL%p: recovery %s\n", pWal, rc ? "failed" : "ok"));
  walUnlockExclusive(pWal, iLock, nLock);
  return rc;
}

/*
** Close an open wal-index.
*/
static void walIndexClose(Wal *pWal, int isDelete){
  if( pWal->isWIndexOpen ){


    sqlite3OsShmClose(pWal->pDbFd, isDelete);
    pWal->isWIndexOpen = 0;
  }
}

/* 
** Open a connection to the WAL file associated with database zDbName.
** The database file must already be opened on connection pDbFd.


**
** A SHARED lock should be held on the database file when this function
** is called. The purpose of this SHARED lock is to prevent any other
** client from unlinking the WAL or wal-index file. If another process
** were to do this just after this client opened one of these files, the
** system would be badly broken.
**
** If the log file is successfully opened, SQLITE_OK is returned and 
** *ppWal is set to point to a new WAL handle. If an error occurs,
** an SQLite error code is returned and *ppWal is left unmodified.
*/

    return SQLITE_NOMEM;
  }

  pRet->pVfs = pVfs;
  pRet->pWalFd = (sqlite3_file *)&pRet[1];
  pRet->pDbFd = pDbFd;
  pRet->szWIndex = -1;
  pRet->readLock = -1;
  sqlite3_randomness(8, &pRet->hdr.aSalt);
  pRet->zWalName = zWal = pVfs->szOsFile + (char*)pRet->pWalFd;
  sqlite3_snprintf(nWal, zWal, "%s-wal", zDbName);
  rc = sqlite3OsShmOpen(pDbFd);

  /* Open file handle on the write-ahead log file. */
  if( rc==SQLITE_OK ){
    pRet->isWIndexOpen = 1;
    flags = (SQLITE_OPEN_READWRITE|SQLITE_OPEN_CREATE|SQLITE_OPEN_MAIN_JOURNAL);
    rc = sqlite3OsOpen(pVfs, zWal, pRet->pWalFd, flags, &flags);
  }

  if( rc!=SQLITE_OK ){
    walIndexClose(pRet, 0);
    sqlite3OsClose(pRet->pWalFd);
    sqlite3_free(pRet);
  }else{
    *ppWal = pRet;
    WALTRACE(("WAL%d: opened\n", pRet));
  }
  return rc;
}

/*
** Find the smallest page number out of all pages held in the WAL that
** has not been returned by any prior invocation of this method on the

  }

  /* This routine only runs while holding SQLITE_SHM_CHECKPOINT.  No other
  ** thread is able to write to shared memory while this routine is
  ** running (or, indeed, while the WalIterator object exists).  Hence,
  ** we can cast off the volatile qualifacation from shared memory
  */
  assert( pWal->ckptLock );
  aData = (u32*)pWal->pWiData;

  /* Allocate space for the WalIterator object */
  iLast = pWal->hdr.mxFrame;
  nSegment = (iLast >> 8) + 1;
  nFinal = (iLast & 0x000000FF);
  nByte = sizeof(WalIterator) + (nSegment+1)*(sizeof(struct WalSegment)+256);

/* 
** Free an iterator allocated by walIteratorInit().
*/
static void walIteratorFree(WalIterator *p){
  sqlite3_free(p);
}


/*
** Copy as much content as we can from the WAL back into the database file
** in response to an sqlite3_wal_checkpoint() request or the equivalent.
**
** The amount of information copies from WAL to database might be limited
** by active readers.  This routine will never overwrite a database page
** that a concurrent reader might be using.
**
** All I/O barrier operations (a.k.a fsyncs) occur in this routine when
** SQLite is in WAL-mode in synchronous=NORMAL.  That means that if 
** checkpoints are always run by a background thread or background 
** process, foreground threads will never block on a lengthy fsync call.
**
** Fsync is called on the WAL before writing content out of the WAL and
** into the database.  This ensures that if the new content is persistent
** in the WAL and can be recovered following a power-loss or hard reset.
**
** Fsync is also called on the database file if (and only if) the entire
** WAL content is copied into the database file.  This second fsync makes
** it safe to delete the WAL since the new content will persist in the
** database file.
**
** This routine uses and updates the nBackfill field of the wal-index header.
** This is the only routine tha will increase the value of nBackfill.  
** (A WAL reset or recovery will revert nBackfill to zero, but not increase
** its value.)
**
** The caller must be holding sufficient locks to ensure that no other
** checkpoint is running (in any other thread or process) at the same
** time.
*/
static int walCheckpoint(
  Wal *pWal,                      /* Wal connection */
  int sync_flags,                 /* Flags for OsSync() (or 0) */
  int nBuf,                       /* Size of zBuf in bytes */
  u8 *zBuf                        /* Temporary buffer to use */
){
  int rc;                         /* Return code */
  int szPage = pWal->hdr.szPage;  /* Database page-size */
  WalIterator *pIter = 0;         /* Wal iterator context */
  u32 iDbpage = 0;                /* Next database page to write */
  u32 iFrame = 0;                 /* Wal frame containing data for iDbpage */
  u32 mxSafeFrame;                /* Max frame that can be backfilled */
  int i;                          /* Loop counter */
  volatile WalIndexHdr *pHdr;     /* The actual wal-index header in SHM */
  volatile WalCkptInfo *pInfo;    /* The checkpoint status information */

  /* Allocate the iterator */
  rc = walIteratorInit(pWal, &pIter);
  if( rc!=SQLITE_OK || pWal->hdr.mxFrame==0 ){
    walIteratorFree(pIter);
    return rc;
  }

  /*** TODO:  Move this test out to the caller.  Make it an assert() here ***/
  if( pWal->hdr.szPage!=nBuf ){
    walIteratorFree(pIter);
    return SQLITE_CORRUPT_BKPT;
  }

  /* Compute in mxSafeFrame the index of the last frame of the WAL that is
  ** safe to write into the database.  Frames beyond mxSafeFrame might
  ** overwrite database pages that are in use by active readers and thus
  ** cannot be backfilled from the WAL.
  */
  mxSafeFrame = pWal->hdr.mxFrame;
  pHdr = (volatile WalIndexHdr*)pWal->pWiData;
  pInfo = (volatile WalCkptInfo*)&pHdr[2];
  assert( pInfo==walCkptInfo(pWal) );
  for(i=1; i<WAL_NREADER; i++){
    u32 y = pInfo->aReadMark[i];
    if( y>0 && (mxSafeFrame==0 || mxSafeFrame>=y) ){
      if( y<=pWal->hdr.mxFrame
       && walLockExclusive(pWal, WAL_READ_LOCK(i), 1)==SQLITE_OK
      ){
        pInfo->aReadMark[i] = 0;
        walUnlockExclusive(pWal, WAL_READ_LOCK(i), 1);
      }else{
        mxSafeFrame = y-1;
      }
    }
  }

  if( pInfo->nBackfill<mxSafeFrame
   && (rc = walLockExclusive(pWal, WAL_READ_LOCK(0), 1))==SQLITE_OK
  ){
    u32 nBackfill = pInfo->nBackfill;

    /* Sync the WAL to disk */
    if( sync_flags ){
      rc = sqlite3OsSync(pWal->pWalFd, sync_flags);

    }

    /* Iterate through the contents of the WAL, copying data to the db file. */
    while( rc==SQLITE_OK && 0==walIteratorNext(pIter, &iDbpage, &iFrame) ){
      if( iFrame<=nBackfill || iFrame>mxSafeFrame ) continue;
      rc = sqlite3OsRead(pWal->pWalFd, zBuf, szPage, 
          walFrameOffset(iFrame, szPage) + WAL_FRAME_HDRSIZE
      );
      if( rc!=SQLITE_OK ) break;
      rc = sqlite3OsWrite(pWal->pDbFd, zBuf, szPage, (iDbpage-1)*szPage);
      if( rc!=SQLITE_OK ) break;
    }

    /* If work was actually accomplished... */
    if( rc==SQLITE_OK && pInfo->nBackfill<mxSafeFrame ){
      pInfo->nBackfill = mxSafeFrame;
      if( mxSafeFrame==pHdr[0].mxFrame && sync_flags ){
        rc = sqlite3OsTruncate(pWal->pDbFd, ((i64)pWal->hdr.nPage*(i64)szPage));



        if( rc==SQLITE_OK && sync_flags ){
          rc = sqlite3OsSync(pWal->pDbFd, sync_flags);

        }
      }
    }


    /* Release the reader lock held while backfilling */
    walUnlockExclusive(pWal, WAL_READ_LOCK(0), 1);
  }else{
    /* Reset the return code so as not to report a checkpoint failure
    ** just because active readers prevent any backfill.
    */
    rc = SQLITE_OK;
  }

  walIteratorFree(pIter);
  return rc;
}

/*
** Close a connection to a log file.
*/

    ** the database. In this case checkpoint the database and unlink both
    ** the wal and wal-index files.
    **
    ** The EXCLUSIVE lock is not released before returning.
    */
    rc = sqlite3OsLock(pWal->pDbFd, SQLITE_LOCK_EXCLUSIVE);
    if( rc==SQLITE_OK ){
      pWal->exclusiveMode = 1;
      rc = sqlite3WalCheckpoint(pWal, sync_flags, nBuf, zBuf);
      if( rc==SQLITE_OK ){
        isDelete = 1;
      }
      walIndexUnmap(pWal);
    }

    walIndexClose(pWal, isDelete);
    sqlite3OsClose(pWal->pWalFd);
    if( isDelete ){
      sqlite3OsDelete(pWal->pVfs, pWal->zWalName, 0);
    }
    WALTRACE(("WAL%p: closed\n", pWal));
    sqlite3_free(pWal);
  }
  return rc;
}

/*
** Try to read the wal-index header.  Return 0 on success and 1 if
** there is a problem.
**
** The wal-index is in shared memory.  Another thread or process might
** be writing the header at the same time this procedure is trying to
** read it, which might result in inconsistency.  A dirty read is detected
** by verifying that both copies of the header are the same and also by
** a checksum on the header.
**
** If and only if the read is consistent and the header is different from
** pWal->hdr, then pWal->hdr is updated to the content of the new header
** and *pChanged is set to 1.
**
** If the checksum cannot be verified return non-zero. If the header
** is read successfully and the checksum verified, return zero.
*/
int walIndexTryHdr(Wal *pWal, int *pChanged){
  u32 aCksum[2];               /* Checksum on the header content */
  WalIndexHdr h1, h2;          /* Two copies of the header content */
  WalIndexHdr *aHdr;           /* Header in shared memory */

  if( pWal->szWIndex < WALINDEX_HDR_SIZE ){
    /* The wal-index is not large enough to hold the header, then assume
    ** header is invalid. */
    return 1;
  }
  assert( pWal->pWiData );

  /* Read the header. This might happen currently with a write to the
  ** same area of shared memory on a different CPU in a SMP,
  ** meaning it is possible that an inconsistent snapshot is read
  ** from the file. If this happens, return non-zero.
  **
  ** There are two copies of the header at the beginning of the wal-index.
  ** When reading, read [0] first then [1].  Writes are in the reverse order.
  ** Memory barriers are used to prevent the compiler or the hardware from
  ** reordering the reads and writes.
  */
  aHdr = (WalIndexHdr*)pWal->pWiData;
  memcpy(&h1, &aHdr[0], sizeof(h1));
  sqlite3OsShmBarrier(pWal->pDbFd);
  memcpy(&h2, &aHdr[1], sizeof(h2));

  if( memcmp(&h1, &h2, sizeof(h1))!=0 ){
    return 1;   /* Dirty read */
  }  
  if( h1.isInit==0 ){
    return 1;   /* Malformed header - probably all zeros */
  }
  walChecksumBytes(1, (u8*)&h1, sizeof(h1)-sizeof(h1.aCksum), 0, aCksum);
  if( aCksum[0]!=h1.aCksum[0] || aCksum[1]!=h1.aCksum[1] ){
    return 1;   /* Checksum does not match */
  }

** after this routine returns.
**
** If the wal-index header is successfully read, return SQLITE_OK. 
** Otherwise an SQLite error code.
*/
static int walIndexReadHdr(Wal *pWal, int *pChanged){
  int rc;                         /* Return code */
  int badHdr;                     /* True if a header read failed */


  assert( pChanged );
  rc = walIndexMap(pWal, walMappingSize(1));
  if( rc!=SQLITE_OK ){
    return rc;
  }

  /* Try once to read the header straight out.  This works most of the




  ** time.





  */
  badHdr = walIndexTryHdr(pWal, pChanged);

  /* If the first attempt failed, it might have been due to a race
  ** with a writer.  So get a WRITE lock and try again.









  */
  assert( badHdr==0 || pWal->writeLock==0 );
  if( badHdr ){
    rc = walLockExclusive(pWal, WAL_WRITE_LOCK, 1);
    if( rc==SQLITE_OK ){
      pWal->writeLock = 1;
      badHdr = walIndexTryHdr(pWal, pChanged);
      if( badHdr ){
        /* If the wal-index header is still malformed even while holding
        ** a WRITE lock, it can only mean that the header is corrupted and
        ** needs to be reconstructed.  So run recovery to do exactly that.
        */
        rc = walIndexRecover(pWal);
        *pChanged = 1;
      }
      walUnlockExclusive(pWal, WAL_WRITE_LOCK, 1);
      pWal->writeLock = 0;
    }else if( rc!=SQLITE_BUSY ){

      return rc;
    }
  }

  /* Make sure the mapping is large enough to cover the entire wal-index */
  if( rc==SQLITE_OK ){
    int szWanted = walMappingSize(pWal->hdr.mxFrame);
    if( pWal->szWIndex<szWanted ){
      rc = walIndexMap(pWal, szWanted);
    }
  }

  return rc;
}

/*
** This is the value that walTryBeginRead returns when it needs to
** be retried.
*/
#define WAL_RETRY  (-1)

/*
** Attempt to start a read transaction.  This might fail due to a race or
** other transient condition.  When that happens, it returns WAL_RETRY to
** indicate to the caller that it is safe to retry immediately.
**
** On success return SQLITE_OK.  On a permantent failure (such an
** I/O error or an SQLITE_BUSY because another process is running
** recovery) return a positive error code.
**
** On success, this routine obtains a read lock on 
** WAL_READ_LOCK(pWal->readLock).  The pWal->readLock integer is
** in the range 0 <= pWal->readLock < WAL_NREADER.  If pWal->readLock==(-1)
** that means the Wal does not hold any read lock.  The reader must not
** access any database page that is modified by a WAL frame up to and
** including frame number aReadMark[pWal->readLock].  The reader will
** use WAL frames up to and including pWal->hdr.mxFrame if pWal->readLock>0
** Or if pWal->readLock==0, then the reader will ignore the WAL
** completely and get all content directly from the database file.
** When the read transaction is completed, the caller must release the
** lock on WAL_READ_LOCK(pWal->readLock) and set pWal->readLock to -1.
**
** This routine uses the nBackfill and aReadMark[] fields of the header
** to select a particular WAL_READ_LOCK() that strives to let the
** checkpoint process do as much work as possible.  This routine might
** update values of the aReadMark[] array in the header, but if it does
** so it takes care to hold an exclusive lock on the corresponding
** WAL_READ_LOCK() while changing values.
*/
static int walTryBeginRead(Wal *pWal, int *pChanged, int useWal){
  volatile WalIndexHdr *pHdr;     /* Header of the wal-index */
  volatile WalCkptInfo *pInfo;    /* Checkpoint information in wal-index */
  u32 mxReadMark;                 /* Largest aReadMark[] value */
  int mxI;                        /* Index of largest aReadMark[] value */
  int i;                          /* Loop counter */
  int rc;                         /* Return code  */

  assert( pWal->readLock<0 );     /* Not currently locked */

  if( !useWal ){
    rc = walIndexReadHdr(pWal, pChanged);
    if( rc==SQLITE_BUSY ){
      /* If there is not a recovery running in another thread or process
      ** then convert BUSY errors to WAL_RETRY.  If recovery is known to
      ** be running, convert BUSY to BUSY_RECOVERY.  There is a race here
      ** which might cause WAL_RETRY to be returned even if BUSY_RECOVERY
      ** would be technically correct.  But the race is benign since with
      ** WAL_RETRY this routine will be called again and will probably be
      ** right on the second iteration.
      */
      rc = walLockShared(pWal, WAL_RECOVER_LOCK);
      if( rc==SQLITE_OK ){
        walUnlockShared(pWal, WAL_RECOVER_LOCK);
        rc = WAL_RETRY;
      }else if( rc==SQLITE_BUSY ){
        rc = SQLITE_BUSY_RECOVERY;
      }
    }
  }else{
    rc = walIndexMap(pWal, walMappingSize(pWal->hdr.mxFrame));
  }
  if( rc!=SQLITE_OK ){
    return rc;
  }

  pHdr = (volatile WalIndexHdr*)pWal->pWiData;
  pInfo = (volatile WalCkptInfo*)&pHdr[2];
  assert( pInfo==walCkptInfo(pWal) );
  if( !useWal && pInfo->nBackfill==pWal->hdr.mxFrame ){
    /* The WAL has been completely backfilled (or it is empty).
    ** and can be safely ignored.
    */
    rc = walLockShared(pWal, WAL_READ_LOCK(0));
    if( rc==SQLITE_OK ){
      if( pHdr->mxFrame!=pWal->hdr.mxFrame ){
        walUnlockShared(pWal, WAL_READ_LOCK(0));
        return WAL_RETRY;
      }
      pWal->readLock = 0;
      return SQLITE_OK;
    }else if( rc!=SQLITE_BUSY ){
      return rc;
    }
  }

  /* If we get this far, it means that the reader will want to use
  ** the WAL to get at content from recent commits.  The job now is
  ** to select one of the aReadMark[] entries that is closest to
  ** but not exceeding pWal->hdr.mxFrame and lock that entry.
  */
  mxReadMark = 0;
  mxI = 0;
  for(i=1; i<WAL_NREADER; i++){
    u32 thisMark = pInfo->aReadMark[i];
    if( mxReadMark<thisMark ){
      mxReadMark = thisMark;
      mxI = i;
    }
  }
  if( mxI==0 ){
    /* If we get here, it means that all of the aReadMark[] entries between
    ** 1 and WAL_NREADER-1 are zero.  Try to initialize aReadMark[1] to
    ** be mxFrame, then retry.
    */
    rc = walLockExclusive(pWal, WAL_READ_LOCK(1), 1);
    if( rc==SQLITE_OK ){
      pInfo->aReadMark[1] = pWal->hdr.mxFrame+1;
      walUnlockExclusive(pWal, WAL_READ_LOCK(1), 1);
    }
    return WAL_RETRY;
  }else{
    if( mxReadMark < pWal->hdr.mxFrame ){
      for(i=1; i<WAL_NREADER; i++){
        rc = walLockExclusive(pWal, WAL_READ_LOCK(i), 1);
        if( rc==SQLITE_OK ){
          mxReadMark = pInfo->aReadMark[i] = pWal->hdr.mxFrame+1;
          mxI = i;
          walUnlockExclusive(pWal, WAL_READ_LOCK(i), 1);
          break;
        }
      }
    }

    rc = walLockShared(pWal, WAL_READ_LOCK(mxI));
    if( rc ){
      return rc==SQLITE_BUSY ? WAL_RETRY : rc;
    }
    if( pInfo->aReadMark[mxI]!=mxReadMark
     || pHdr[0].mxFrame!=pWal->hdr.mxFrame
     || (sqlite3OsShmBarrier(pWal->pDbFd), pHdr[1].mxFrame!=pWal->hdr.mxFrame)
    ){
      walUnlockShared(pWal, WAL_READ_LOCK(mxI));
      return WAL_RETRY;
    }else{
      pWal->readLock = mxI;
    }
  }
  return rc;
}

/*
** Begin a read transaction on the database.
**
** This routine used to be called sqlite3OpenSnapshot() and with good reason:
** it takes a snapshot of the state of the WAL and wal-index for the current
** instant in time.  The current thread will continue to use this snapshot.
** Other threads might append new content to the WAL and wal-index but
** that extra content is ignored by the current thread.
**






** If the database contents have changes since the previous read

** transaction, then *pChanged is set to 1 before returning.  The
** Pager layer will use this to know that is cache is stale and

** needs to be flushed.
*/
int sqlite3WalBeginReadTransaction(Wal *pWal, int *pChanged){
  int rc;                         /* Return code */



  do{

    rc = walTryBeginRead(pWal, pChanged, 0);





  }while( rc==WAL_RETRY );
  walIndexUnmap(pWal);
  return rc;
}

/*
** Finish with a read transaction.  All this does is release the
** read-lock.
*/
void sqlite3WalEndReadTransaction(Wal *pWal){
  if( pWal->readLock>=0 ){
    walUnlockShared(pWal, WAL_READ_LOCK(pWal->readLock));

    pWal->readLock = -1;
  }
}

/*
** Read a page from the WAL, if it is present in the WAL and if the 
** current read transaction is configured to use the WAL.  
**
** The *pInWal is set to 1 if the requested page is in the WAL and
** has been loaded.  Or *pInWal is set to 0 if the page was not in 
** the WAL and needs to be read out of the database.
*/
int sqlite3WalRead(
  Wal *pWal,                      /* WAL handle */
  Pgno pgno,                      /* Database page number to read data for */
  int *pInWal,                    /* OUT: True if data is read from WAL */
  int nOut,                       /* Size of buffer pOut in bytes */
  u8 *pOut                        /* Buffer to write page data to */
){
  int rc;                         /* Return code */
  u32 iRead = 0;                  /* If !=0, WAL frame to return data from */
  u32 iLast = pWal->hdr.mxFrame;  /* Last page in WAL for this reader */
  int iHash;                      /* Used to loop through N hash tables */

  /* This routine is only called from within a read transaction */
  assert( pWal->readLock>=0 );

  /* If the "last page" field of the wal-index header snapshot is 0, then
  ** no data will be read from the wal under any circumstances. Return early
  ** in this case to avoid the walIndexMap/Unmap overhead.  Likewise, if
  ** pWal->readLock==0, then the WAL is ignored by the reader so
  ** return early, as if the WAL were empty.
  */
  if( iLast==0 || pWal->readLock==0 ){
    *pInWal = 0;
    return SQLITE_OK;
  }

  /* Ensure the wal-index is mapped. */

  rc = walIndexMap(pWal, walMappingSize(iLast));
  if( rc!=SQLITE_OK ){
    return rc;
  }

  /* Search the hash table or tables for an entry matching page number
  ** pgno. Each iteration of the following for() loop searches one

}


/* 
** Set *pPgno to the size of the database file (or zero, if unknown).
*/
void sqlite3WalDbsize(Wal *pWal, Pgno *pPgno){
  assert( pWal->readLock>=0 );

  *pPgno = pWal->hdr.nPage;
}


/* 
** This function starts a write transaction on the WAL.

**
** A read transaction must have already been started by a prior call
** to sqlite3WalBeginReadTransaction().
**
** If another thread or process has written into the database since
** the read transaction was started, then it is not possible for this
** thread to write as doing so would cause a fork.  So this routine
** returns SQLITE_BUSY in that case and no write transaction is started.
**
** There can only be a single writer active at a time.
*/
int sqlite3WalBeginWriteTransaction(Wal *pWal){
  int rc;
  volatile WalCkptInfo *pInfo;

  /* Cannot start a write transaction without first holding a read
  ** transaction. */
  assert( pWal->readLock>=0 );

  /* Only one writer allowed at a time.  Get the write lock.  Return
  ** SQLITE_BUSY if unable.
  */
  rc = walLockExclusive(pWal, WAL_WRITE_LOCK, 1);
  if( rc ){
    return rc;
  }
  pWal->writeLock = 1;

  /* If another connection has written to the database file since the
  ** time the read transaction on this connection was started, then
  ** the write is disallowed.
  */

  rc = walIndexMap(pWal, walMappingSize(pWal->hdr.mxFrame));

  if( rc ){
    walUnlockExclusive(pWal, WAL_WRITE_LOCK, 1);
    pWal->writeLock = 0;

    return rc;
  }
  if( memcmp(&pWal->hdr, (void*)pWal->pWiData, sizeof(WalIndexHdr))!=0 ){
    walUnlockExclusive(pWal, WAL_WRITE_LOCK, 1);
    pWal->writeLock = 0;
    walIndexUnmap(pWal);
    return SQLITE_BUSY;
  }

  pInfo = walCkptInfo(pWal);
  if( pWal->readLock==0 ){
    assert( pInfo->nBackfill==pWal->hdr.mxFrame );
    if( pInfo->nBackfill>0 ){
      rc = walLockExclusive(pWal, WAL_READ_LOCK(1), WAL_NREADER-1);
      if( rc==SQLITE_OK ){
        /* If all readers are using WAL_READ_LOCK(0) (in other words if no
        ** readers are currently using the WAL) */
        pWal->nCkpt++;
        pWal->hdr.mxFrame = 0;
        sqlite3Put4byte((u8*)pWal->hdr.aSalt,
                         1 + sqlite3Get4byte((u8*)pWal->hdr.aSalt));
        sqlite3_randomness(4, &pWal->hdr.aSalt[1]);
        walIndexWriteHdr(pWal);
        pInfo->nBackfill = 0;
        memset((void*)&pInfo->aReadMark[1], 0,
               sizeof(pInfo->aReadMark)-sizeof(u32));
        rc = sqlite3OsTruncate(pWal->pDbFd, 
                               ((i64)pWal->hdr.nPage*(i64)pWal->szPage));
        walUnlockExclusive(pWal, WAL_READ_LOCK(1), WAL_NREADER-1);
      }
    }
    walUnlockShared(pWal, WAL_READ_LOCK(0));
    pWal->readLock = -1;
    do{
      int notUsed;
      rc = walTryBeginRead(pWal, &notUsed, 1);
    }while( rc==WAL_RETRY );
  }
  walIndexUnmap(pWal);
  return rc;
}

/*
** End a write transaction.  The commit has already been done.  This
** routine merely releases the lock.
*/
int sqlite3WalEndWriteTransaction(Wal *pWal){
  walUnlockExclusive(pWal, WAL_WRITE_LOCK, 1);
  pWal->writeLock = 0;
  return SQLITE_OK;
}

/*
** If any data has been written (but not committed) to the log file, this
** function moves the write-pointer back to the start of the transaction.
**
** Additionally, the callback function is invoked for each frame written
** to the WAL since the start of the transaction. If the callback returns
** other than SQLITE_OK, it is not invoked again and the error code is
** returned to the caller.
**
** Otherwise, if the callback function does not return an error, this
** function returns SQLITE_OK.
*/
int sqlite3WalUndo(Wal *pWal, int (*xUndo)(void *, Pgno), void *pUndoCtx){
  int rc = SQLITE_OK;
  if( pWal->writeLock ){
    int unused;
    Pgno iMax = pWal->hdr.mxFrame;
    Pgno iFrame;
  
    assert( pWal->pWiData==0 );
    rc = walIndexReadHdr(pWal, &unused);
    if( rc==SQLITE_OK ){
      rc = walIndexMap(pWal, walMappingSize(iMax));
    }
    if( rc==SQLITE_OK ){
      for(iFrame=pWal->hdr.mxFrame+1; rc==SQLITE_OK && iFrame<=iMax; iFrame++){
        assert( pWal->writeLock );
        rc = xUndo(pUndoCtx, pWal->pWiData[walIndexEntry(iFrame)]);
      }
      walCleanupHash(pWal);
    }
    walIndexUnmap(pWal);
  }
  return rc;
}

/* 
** Argument aWalData must point to an array of WAL_SAVEPOINT_NDATA u32 
** values. This function populates the array with values required to 
** "rollback" the write position of the WAL handle back to the current 
** point in the event of a savepoint rollback (via WalSavepointUndo()).
*/
void sqlite3WalSavepoint(Wal *pWal, u32 *aWalData){
  assert( pWal->writeLock );
  aWalData[0] = pWal->hdr.mxFrame;
  aWalData[1] = pWal->hdr.aFrameCksum[0];
  aWalData[2] = pWal->hdr.aFrameCksum[1];
}

/* 
** Move the write position of the WAL back to the point identified by
** the values in the aWalData[] array. aWalData must point to an array
** of WAL_SAVEPOINT_NDATA u32 values that has been previously populated
** by a call to WalSavepoint().
*/
int sqlite3WalSavepointUndo(Wal *pWal, u32 *aWalData){
  int rc = SQLITE_OK;
  assert( pWal->writeLock );

  assert( aWalData[0]<=pWal->hdr.mxFrame );
  if( aWalData[0]<pWal->hdr.mxFrame ){
    rc = walIndexMap(pWal, walMappingSize(pWal->hdr.mxFrame));
    pWal->hdr.mxFrame = aWalData[0];
    pWal->hdr.aFrameCksum[0] = aWalData[1];
    pWal->hdr.aFrameCksum[1] = aWalData[2];

  u32 iFrame;                     /* Next frame address */
  u8 aFrame[WAL_FRAME_HDRSIZE];   /* Buffer to assemble frame-header in */
  PgHdr *p;                       /* Iterator to run through pList with. */
  PgHdr *pLast = 0;               /* Last frame in list */
  int nLast = 0;                  /* Number of extra copies of last page */

  assert( pList );
  assert( pWal->writeLock );
  assert( pWal->pWiData==0 );

#if defined(SQLITE_TEST) && defined(SQLITE_DEBUG)
  { int cnt; for(cnt=0, p=pList; p; p=p->pDirty, cnt++){}
    WALTRACE(("WAL%p: frame write begin. %d frames. mxFrame=%d. %s\n",
              pWal, cnt, pWal->hdr.mxFrame, isCommit ? "Commit" : "Spill"));
  }
#endif

  /* If this is the first frame written into the log, write the WAL
  ** header to the start of the WAL file. See comments at the top of
  ** this source file for a description of the WAL header format.
  */
  iFrame = pWal->hdr.mxFrame;
  if( iFrame==0 ){
    u8 aWalHdr[WAL_HDRSIZE];        /* Buffer to assembly wal-header in */
    sqlite3Put4byte(&aWalHdr[0], (WAL_MAGIC | SQLITE_BIGENDIAN));
    sqlite3Put4byte(&aWalHdr[4], 3007000);
    sqlite3Put4byte(&aWalHdr[8], szPage);
    pWal->szPage = szPage;
    pWal->hdr.bigEndCksum = SQLITE_BIGENDIAN;
    sqlite3Put4byte(&aWalHdr[12], pWal->nCkpt);
    memcpy(&aWalHdr[16], pWal->hdr.aSalt, 8);
    rc = sqlite3OsWrite(pWal->pWalFd, aWalHdr, sizeof(aWalHdr), 0);
    WALTRACE(("WAL%p: wal-header write %s\n", pWal, rc ? "failed" : "ok"));
    if( rc!=SQLITE_OK ){
      return rc;
    }
    walChecksumBytes(1, aWalHdr, sizeof(aWalHdr), 0, pWal->hdr.aFrameCksum);
  }
  assert( pWal->szPage==szPage );

    if( isCommit ){
      walIndexWriteHdr(pWal);
      pWal->iCallback = iFrame;
    }
  }

  walIndexUnmap(pWal);
  WALTRACE(("WAL%p: frame write %s\n", pWal, rc ? "failed" : "ok"));
  return rc;
}

/* 
** This routine is called to implement sqlite3_wal_checkpoint() and
** related interfaces.
**
** Obtain a CHECKPOINT lock and then backfill as much information as
** we can from WAL into the database.


*/
int sqlite3WalCheckpoint(
  Wal *pWal,                      /* Wal connection */
  int sync_flags,                 /* Flags to sync db file with (or 0) */
  int nBuf,                       /* Size of temporary buffer */
  u8 *zBuf                        /* Temporary buffer to use */


){
  int rc;                         /* Return code */
  int isChanged = 0;              /* True if a new wal-index header is loaded */

  assert( pWal->pWiData==0 );
  assert( pWal->ckptLock==0 );











  WALTRACE(("WAL%p: checkpoint begins\n", pWal));
  rc = walLockExclusive(pWal, WAL_CKPT_LOCK, 1);



  if( rc ){
    /* Usually this is SQLITE_BUSY meaning that another thread or process
    ** is already running a checkpoint, or maybe a recovery.  But it might
    ** also be SQLITE_IOERR. */
    return rc;
  }
  pWal->ckptLock = 1;

  /* Copy data from the log to the database file. */
  rc = walIndexReadHdr(pWal, &isChanged);
  if( rc==SQLITE_OK ){
    rc = walCheckpoint(pWal, sync_flags, nBuf, zBuf);
  }
  if( isChanged ){
    /* If a new wal-index header was loaded before the checkpoint was 
    ** performed, then the pager-cache associated with pWal is now
    ** out of date. So zero the cached wal-index header to ensure that
    ** next time the pager opens a snapshot on this database it knows that
    ** the cache needs to be reset.
    */
    memset(&pWal->hdr, 0, sizeof(WalIndexHdr));
  }

  /* Release the locks. */
  walIndexUnmap(pWal);
  walUnlockExclusive(pWal, WAL_CKPT_LOCK, 1);
  pWal->ckptLock = 0;
  WALTRACE(("WAL%p: checkpoint %s\n", pWal, rc ? "failed" : "ok"));
  return rc;
}

/* Return the value to pass to a sqlite3_wal_hook callback, the
** number of frames in the WAL at the point of the last commit since
** sqlite3WalCallback() was called.  If no commits have occurred since
** the last call, then return 0.
*/
int sqlite3WalCallback(Wal *pWal){
  u32 ret = 0;
  if( pWal ){
    ret = pWal->iCallback;
    pWal->iCallback = 0;
  }
  return (int)ret;
}

/*
** This function is called to change the WAL subsystem into or out



** of locking_mode=EXCLUSIVE.

**
** If op is zero, then attempt to change from locking_mode=EXCLUSIVE
** into locking_mode=NORMAL.  This means that we must acquire a lock
** on the pWal->readLock byte.  If the WAL is already in locking_mode=NORMAL
** or if the acquisition of the lock fails, then return 0.  If the
** transition out of exclusive-mode is successful, return 1.  This
** operation must occur while the pager is still holding the exclusive
** lock on the main database file.
**

** If op is one, then change from locking_mode=NORMAL into 
** locking_mode=EXCLUSIVE.  This means that the pWal->readLock must
** be released.  Return 1 if the transition is made and 0 if the
** WAL is already in exclusive-locking mode - meaning that this
** routine is a no-op.  The pager must already hold the exclusive lock

** on the main database file before invoking this operation.
**


** If op is negative, then do a dry-run of the op==1 case but do
** not actually change anything.  The pager uses this to see if it
** should acquire the database exclusive lock prior to invoking
** the op==1 case.
*/
int sqlite3WalExclusiveMode(Wal *pWal, int op){
  int rc;
  assert( pWal->writeLock==0 && pWal->readLock>=0 );
  if( op==0 ){
    if( pWal->exclusiveMode ){
      pWal->exclusiveMode = 0;
      if( walLockShared(pWal, WAL_READ_LOCK(pWal->readLock))!=SQLITE_OK ){
        pWal->exclusiveMode = 1;
      }
      rc = pWal->exclusiveMode==0;
    }else{
      /* No changes.  Either already in locking_mode=NORMAL or else the 
      ** acquisition of the read-lock failed.  The pager must continue to
      ** hold the database exclusive lock. */
      rc = 0;
    }
  }else if( op>0 ){
    assert( pWal->exclusiveMode==0 );
    walUnlockShared(pWal, WAL_READ_LOCK(pWal->readLock));
    pWal->exclusiveMode = 1;
    rc = 1;
  }else{
    rc = pWal->exclusiveMode==0;
  }
  return rc;
}

#endif /* #ifndef SQLITE_OMIT_WAL */

#ifndef _WAL_H_
#define _WAL_H_

#include "sqliteInt.h"

#ifdef SQLITE_OMIT_WAL
# define sqlite3WalOpen(x,y,z)                 0
# define sqlite3WalClose(w,x,y,z)              0
# define sqlite3WalBeginReadTransaction(y,z)   0
# define sqlite3WalEndReadTransaction(z)
# define sqlite3WalRead(v,w,x,y,z)             0
# define sqlite3WalDbsize(y,z)
# define sqlite3WalBeginWriteTransaction(y)    0
# define sqlite3WalEndWRiteTransaction(x)      0
# define sqlite3WalUndo(x,y,z)                 0
# define sqlite3WalSavepoint(y,z)
# define sqlite3WalSavepointUndo(y,z)          0
# define sqlite3WalFrames(u,v,w,x,y,z)         0
# define sqlite3WalCheckpoint(u,v,w,x)         0
# define sqlite3WalCallback(z)                 0
#else

#define WAL_SAVEPOINT_NDATA 3

/* Connection to a write-ahead log (WAL) file. 
** There is one object of this type for each pager. 
*/
typedef struct Wal Wal;

/* Open and close a connection to a write-ahead log. */
int sqlite3WalOpen(sqlite3_vfs*, sqlite3_file*, const char *zName, Wal**);
int sqlite3WalClose(Wal *pWal, int sync_flags, int, u8 *);

/* Used by readers to open (lock) and close (unlock) a snapshot.  A 
** snapshot is like a read-transaction.  It is the state of the database
** at an instant in time.  sqlite3WalOpenSnapshot gets a read lock and
** preserves the current state even if the other threads or processes
** write to or checkpoint the WAL.  sqlite3WalCloseSnapshot() closes the
** transaction and releases the lock.
*/
int sqlite3WalBeginReadTransaction(Wal *pWal, int *);
void sqlite3WalEndReadTransaction(Wal *pWal);

/* Read a page from the write-ahead log, if it is present. */
int sqlite3WalRead(Wal *pWal, Pgno pgno, int *pInWal, int nOut, u8 *pOut);

/* Return the size of the database as it existed at the beginning
** of the snapshot */
void sqlite3WalDbsize(Wal *pWal, Pgno *pPgno);

/* Obtain or release the WRITER lock. */
int sqlite3WalBeginWriteTransaction(Wal *pWal);
int sqlite3WalEndWriteTransaction(Wal *pWal);

/* Undo any frames written (but not committed) to the log */
int sqlite3WalUndo(Wal *pWal, int (*xUndo)(void *, Pgno), void *pUndoCtx);

/* Return an integer that records the current (uncommitted) write
** position in the WAL */
void sqlite3WalSavepoint(Wal *pWal, u32 *aWalData);

/* Move the write position of the WAL back to iFrame.  Called in
** response to a ROLLBACK TO command. */
int sqlite3WalSavepointUndo(Wal *pWal, u32 *aWalData);

/* Write a frame or frames to the log. */
int sqlite3WalFrames(Wal *pWal, int, PgHdr *, Pgno, int, int);

/* Copy pages from the log to the database file */ 
int sqlite3WalCheckpoint(
  Wal *pWal,                      /* Write-ahead log connection */
  int sync_flags,                 /* Flags to sync db file with (or 0) */
  int nBuf,                       /* Size of buffer nBuf */
  u8 *zBuf                        /* Temporary buffer to use */


);

/* Return the value to pass to a sqlite3_wal_hook callback, the
** number of frames in the WAL at the point of the last commit since
** sqlite3WalCallback() was called.  If no commits have occurred since
** the last call, then return 0.
*/

do_test filectrl-1.4 {
  sqlite3 db test.db
  file_control_lasterrno_test db
} {}
do_test filectrl-1.5 {
  db close
  sqlite3 db test_control_lockproxy.db
  file_control_lockproxy_test db [pwd]
} {}
db close
file delete -force .test_control_lockproxy.db-conch test.proxy
finish_test

      return "ERROR: Child process hung up"
    }
    append r $line
  }
}

proc testfixture_nb_cb {varname chan} {
  if {[eof $chan]} {
    append ::tfnb($chan) "ERROR: Child process hung up"
    set line "OVER"
  } else {
    set line [gets $chan]
  }

  if { $line == "OVER" } {
    set $varname $::tfnb($chan)
    unset ::tfnb($chan)
    close $chan
  } else {
    append ::tfnb($chan) $line
  }

  # checkpointing the database. But not from writing to it.
  #
  do_test wal-10.$tn.11 {
    sql2 { BEGIN; SELECT * FROM t1 }
  } {1 2 3 4 5 6 7 8 9 10}
  do_test wal-10.$tn.12 {
    catchsql { PRAGMA wal_checkpoint } 
  } {0 {}}   ;# Reader no longer block checkpoints
  do_test wal-10.$tn.13 {
    execsql { INSERT INTO t1 VALUES(11, 12) }
    sql2 {SELECT * FROM t1}
  } {1 2 3 4 5 6 7 8 9 10}




  # Writers do not block checkpoints any more either.
  #






  do_test wal-10.$tn.14 {
    catchsql { PRAGMA wal_checkpoint } 
  } {0 {}}






  # The following series of test cases used to verify another blocking
  # case in WAL - a case which no longer blocks.
  #

  do_test wal-10.$tn.15 {
    sql2 { COMMIT; BEGIN; SELECT * FROM t1; }
  } {1 2 3 4 5 6 7 8 9 10 11 12}
  do_test wal-10.$tn.16 {
    catchsql { PRAGMA wal_checkpoint } 






  } {0 {}}

  do_test wal-10.$tn.17 {
    execsql { PRAGMA wal_checkpoint } 
  } {}
  do_test wal-10.$tn.18 {
    sql3 { BEGIN; SELECT * FROM t1 }
  } {1 2 3 4 5 6 7 8 9 10 11 12}
  do_test wal-10.$tn.19 {
    catchsql { INSERT INTO t1 VALUES(13, 14) }
  } {0 {}}
  do_test wal-10.$tn.20 {
    execsql { SELECT * FROM t1 }
  } {1 2 3 4 5 6 7 8 9 10 11 12 13 14}
  do_test wal-10.$tn.21 {
    sql3 COMMIT
    sql2 COMMIT
  } {}
  do_test wal-10.$tn.22 {

    execsql { SELECT * FROM t1 }
  } {1 2 3 4 5 6 7 8 9 10 11 12 13 14}

  # Another series of tests that used to demonstrate blocking behavior
  # but which now work.

  #
  do_test wal-10.$tn.23 {
    execsql { PRAGMA wal_checkpoint }
  } {}
  do_test wal-10.$tn.24 {
    sql2 { BEGIN; SELECT * FROM t1; }
  } {1 2 3 4 5 6 7 8 9 10 11 12 13 14}
  do_test wal-10.$tn.25 {
    execsql { PRAGMA wal_checkpoint }
  } {}
  do_test wal-10.$tn.26 {
    catchsql { INSERT INTO t1 VALUES(15, 16) }
  } {0 {}}
  do_test wal-10.$tn.27 {
    sql3 { INSERT INTO t1 VALUES(17, 18) }
  } {}
  do_test wal-10.$tn.28 {
    code3 {
      set ::STMT [sqlite3_prepare db3 "SELECT * FROM t1" -1 TAIL]
      sqlite3_step $::STMT
    }

    execsql { SELECT * FROM t1 }
  } {1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18}

  do_test wal-10.$tn.29 {
    execsql { INSERT INTO t1 VALUES(19, 20) }
    catchsql { PRAGMA wal_checkpoint }
  } {0 {}}
  do_test wal-10.$tn.30 {
    code3 { sqlite3_finalize $::STMT }
    execsql { PRAGMA wal_checkpoint }
  } {}

  # At one point, if a reader failed to upgrade to a writer because it
  # was reading an old snapshot, the write-locks were not being released.
  # Test that this bug has been fixed.
  #
  do_test wal-10.$tn.31 {
    sql2 COMMIT
    execsql { BEGIN ; SELECT * FROM t1 }
    sql2 { INSERT INTO t1 VALUES(21, 22) }
    catchsql { INSERT INTO t1 VALUES(23, 24) }
  } {1 {database is locked}}
  do_test wal-10.$tn.32 {
    # This statement would fail when the bug was present.
    sql2 { INSERT INTO t1 VALUES(23, 24) }
  } {}
  do_test wal-10.$tn.33 {
    execsql { SELECT * FROM t1 ; COMMIT }
  } {1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20}
  do_test wal-10.$tn.34 {
    execsql { SELECT * FROM t1 }
  } {1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24}

  # Test that if a checkpointer cannot obtain the required locks, it
  # releases all locks before returning a busy error.
  #
  do_test wal-10.$tn.35 {
    execsql { 
      DELETE FROM t1;
      INSERT INTO t1 VALUES('a', 'b');
      INSERT INTO t1 VALUES('c', 'd');
    }
    sql2 {
      BEGIN;
        SELECT * FROM t1;
    }
  } {a b c d}


  do_test wal-10.$tn.36 {
    catchsql { PRAGMA wal_checkpoint }
  } {0 {}}
  do_test wal-10.$tn.36 {
    sql3 { INSERT INTO t1 VALUES('e', 'f') }
    sql2 { SELECT * FROM t1 }
  } {a b c d}
  do_test wal-10.$tn.37 {
    sql2 COMMIT
    execsql { PRAGMA wal_checkpoint }

do_test wal-15.3.2 {
  sqlite3_errcode db
} {SQLITE_LOCKED}
do_test wal-15.3.3 {
  sqlite3_errmsg db
} {database table is locked}

# Earlier versions returned an error is returned if the db cannot be 
# checkpointed because of locks held by another connection. Check that
# this is no longer the case.
#
sqlite3 db2 test.db
do_test wal-15.4.1 {
  execsql {
    BEGIN;
    SELECT * FROM t1;
  } db2
} {1 2}
do_test wal-15.4.2 {
  execsql { COMMIT }
  sqlite3_wal_checkpoint db
} {SQLITE_OK}
do_test wal-15.4.3 {
  sqlite3_errmsg db
} {not an error}

# After [db2] drops its lock, [db] may checkpoint the db.
#
do_test wal-15.4.4 {
  execsql { COMMIT } db2
  sqlite3_wal_checkpoint db
} {SQLITE_OK}

source $testdir/tester.tcl
source $testdir/lock_common.tcl
ifcapable !wal {finish_test ; return }

proc set_tvfs_hdr {file args} {

  # Set $nHdr to the number of bytes in the wal-index header:
  set nHdr 40
  set nInt [expr {$nHdr/4}]

  if {[llength $args]>1} {
    return -code error {wrong # args: should be "set_tvfs_hdr fileName ?val?"}
  }

  set blob [tvfs shm $file]

  if {[llength $args]} {
    set ia [lindex $args 0]
    binary scan $blob a[expr $nHdr*2]a* dummy tail
    set blob [binary format i${nInt}i${nInt}a* $ia $ia $tail]
    tvfs shm $file $blob
  }

  binary scan $blob i${nInt} ints
  return $ints
}

    SELECT count(a), sum(a) FROM t1;
  }
} {4 10}
do_test wal2-1.1 {
  execsql { SELECT count(a), sum(a) FROM t1 } db2
} {4 10}

set RECOVER [list                                      \
  {0 1 lock exclusive}   {1 7 lock exclusive}          \
  {1 7 unlock exclusive} {0 1 unlock exclusive}        \
]
set READ [list                                         \
  {4 1 lock exclusive} {4 1 unlock exclusive}          \
  {4 1 lock shared}    {4 1 unlock shared}             \
]

foreach {tn iInsert res wal_index_hdr_mod wal_locks} "
         2    5   {5 15}    0             {$RECOVER $READ}
         3    6   {6 21}    1             {$RECOVER $READ}
         4    7   {7 28}    2             {$RECOVER $READ}
         5    8   {8 36}    3             {$RECOVER $READ}
         6    9   {9 45}    4             {$RECOVER $READ}
         7   10   {10 55}   5             {$RECOVER $READ}
         8   11   {11 66}   6             {$RECOVER $READ}
         9   12   {12 78}   7             {$RECOVER $READ}
        10   13   {13 91}   8             {$RECOVER $READ}
        11   14   {14 105}  9             {$RECOVER $READ}
        12   15   {15 120}  -1            {$READ}
" {

  do_test wal2-1.$tn.1 {
    execsql { INSERT INTO t1 VALUES($iInsert) }

    set ::locks [list]
    set ::cb_done 0

    proc tvfs_cb {method args} {
      if {$::cb_done == 0 && $method == "xShmGet"} {
        set ::cb_done 1
        if {$::wal_index_hdr_mod >= 0} {
          incr_tvfs_hdr [lindex $args 0] $::wal_index_hdr_mod 1
        }
      }

      if {$method == "xShmLock"} { lappend ::locks [lindex $args 2] }
      return SQLITE_OK
    }

    execsql { SELECT count(a), sum(a) FROM t1 } db2
  } $res

# it simply drops back to a READ lock and proceeds. But because the
# header is out-of-date, the reader reads the out-of-date snapshot.
#
# After this, the header is corrupted again and the reader is allowed
# to run recovery. This time, it sees an up-to-date snapshot of the
# database file.
#
set WRITER [list 0 1 lock exclusive]
set LOCKS  [list \
  {0 1 lock exclusive} {0 1 unlock exclusive} \
  {4 1 lock shared}    {4 1 unlock shared}    \
]
do_test wal2-2.0 {

  testvfs tvfs tvfs_cb
  proc tvfs_cb {method args} {
    if {$method == "xShmOpen"} { set ::shm_file [lindex $args 0] }
    return SQLITE_OK
  }

        if {$::wal_index_hdr_mod >= 0} {
          incr_tvfs_hdr $::shm_file $::wal_index_hdr_mod 1
        }
      }
      if {$method == "xShmLock"} {
        set lock [lindex $args 2]
        lappend ::locks $lock
        if {$lock == $::WRITER} {
          set_tvfs_hdr $::shm_file $::oldhdr
        }
      }
      return SQLITE_OK
    }

    execsql { SELECT count(a), sum(a) FROM t1 } db2
  } $res0

  do_test wal2-2.$tn.3 {
    set ::locks
  } $LOCKS

  do_test wal2-2.$tn.4 {
    set ::locks [list]
    set ::cb_done 0
    proc tvfs_cb {method args} {
      if {$::cb_done == 0 && $method == "xShmGet"} {
        set ::cb_done 1

  } $res1
}
db close
db2 close
tvfs delete
file delete -force test.db test.db-wal test.db-journal


if 0 {
#-------------------------------------------------------------------------
# This test case - wal2-3.* - tests the response of the library to an
# SQLITE_BUSY when attempting to obtain a READ or RECOVER lock.
#
#   wal2-3.0 - 2: SQLITE_BUSY when obtaining a READ lock
#   wal2-3.3 - 6: SQLITE_BUSY when obtaining a RECOVER lock
#

} {4 10}
do_test wal2-3.5 {
  list [info exists ::sabotage] [info exists ::locked]
} {0 0}
db close
tvfs delete
file delete -force test.db test.db-wal test.db-journal

}

#-------------------------------------------------------------------------
# Test that a database connection using a VFS that does not support the
# xShmXXX interfaces cannot open a WAL database.
#
do_test wal2-4.1 {
  sqlite3 db test.db

db close
tvfs delete

#-------------------------------------------------------------------------
# Test that if a database connection is forced to run recovery before it
# can perform a checkpoint, it does not transition into RECOVER state.
#
# UPDATE: This has now changed. When running a checkpoint, if recovery is
# required the client grabs all exclusive locks (just as it would for a
# recovery performed as a pre-cursor to a normal database transaction).
#
set expected_locks [list]
lappend expected_locks {1 1 lock exclusive}   ;# Lock checkpoint
lappend expected_locks {0 1 lock exclusive}   ;# Lock writer
lappend expected_locks {2 6 lock exclusive}   ;# Lock recovery & all aReadMark[]
lappend expected_locks {2 6 unlock exclusive} ;# Unlock recovery & aReadMark[]
lappend expected_locks {0 1 unlock exclusive} ;# Unlock writer
lappend expected_locks {3 1 lock exclusive}   ;# Lock aReadMark[0]
lappend expected_locks {3 1 unlock exclusive} ;# Unlock aReadMark[0]
lappend expected_locks {1 1 unlock exclusive} ;# Unlock checkpoint
do_test wal2-5.1 {
  proc tvfs_cb {method args} {
    set ::shm_file [lindex $args 0]
    if {$method == "xShmLock"} { lappend ::locks [lindex $args 2] }
    return $::tvfs_cb_return
  }
  set tvfs_cb_return SQLITE_OK

    INSERT INTO x VALUES(1);
  }

  incr_tvfs_hdr $::shm_file 1 1
  set ::locks [list]
  execsql { PRAGMA wal_checkpoint }
  set ::locks
} $expected_locks
db close
tvfs delete

#-------------------------------------------------------------------------
# This block, test cases wal2-6.*, tests the operation of WAL with
# "PRAGMA locking_mode=EXCLUSIVE" set.
#

  list [file exists test.db-wal] [file exists test.db-journal]
} {0 1}
do_test wal2-6.3.7 {
  execsql { PRAGMA lock_status }
} {main exclusive temp closed}
db close


# This test - wal2-6.4.* - uses a single database connection and the
# [testvfs] instrumentation to test that xShmLock() is being called
# as expected when a WAL database is used with locking_mode=exclusive.
#
do_test wal2-6.4.1 {
  file delete -force test.db test.db-wal test.db-journal
  proc tvfs_cb {method args} {
    set ::shm_file [lindex $args 0]
    if {$method == "xShmLock"} { lappend ::locks [lindex $args 2] }
    return "SQLITE_OK"
  }
  testvfs tvfs tvfs_cb
  sqlite3 db test.db -vfs tvfs
} {}

set RECOVERY {
  {0 1 lock exclusive} {1 7 lock exclusive} 
  {1 7 unlock exclusive} {0 1 unlock exclusive}
}
set READMARK0_READ {
  {3 1 lock shared} {3 1 unlock shared}
}
set READMARK0_WRITE {
  {3 1 lock shared} 
  {0 1 lock exclusive} {3 1 unlock shared} 
  {4 1 lock exclusive} {4 1 unlock exclusive} {4 1 lock shared} 
  {0 1 unlock exclusive} {4 1 unlock shared}
}
set READMARK1_SET {
  {4 1 lock exclusive} {4 1 unlock exclusive}
}
set READMARK1_READ {
  {4 1 lock shared} {4 1 unlock shared}
}

foreach {tn sql res expected_locks} {
  2 {
    PRAGMA journal_mode = WAL;
    BEGIN;
      CREATE TABLE t1(x);
      INSERT INTO t1 VALUES('Leonard');
      INSERT INTO t1 VALUES('Arthur');
    COMMIT;
  } {wal} {
    $RECOVERY 
    $READMARK0_READ 
    $READMARK0_WRITE
  }




  3 {
    # This test should do the READMARK1_SET locking to populate the 
    # aReadMark[1] slot with the current mxFrame value. Followed by
    # READMARK1_READ to read the database.
    #
    SELECT * FROM t1
  } {Leonard Arthur} {


    $READMARK1_SET
    $READMARK1_READ
  }

  4 {

    # aReadMark[1] is already set to mxFrame. So just READMARK1_READ
    # this time, not READMARK1_SET.
    #
    SELECT * FROM t1 ORDER BY x
  } {Arthur Leonard} { 
    $READMARK1_READ 
  }

  5 {
    PRAGMA locking_mode = exclusive
  } {exclusive} { } 

  6 {
    INSERT INTO t1 VALUES('Julius Henry');
    SELECT * FROM t1;
  } {Leonard Arthur {Julius Henry}} {
    $READMARK1_READ
  }


  7 {
    INSERT INTO t1 VALUES('Karl');
    SELECT * FROM t1;
  } {Leonard Arthur {Julius Henry} Karl} { }

  8 {
    PRAGMA locking_mode = normal
  } {normal} { }

  9 {
    SELECT * FROM t1 ORDER BY x
  } {Arthur {Julius Henry} Karl Leonard} { }

  10 {
    DELETE FROM t1
  } {} {
    $READMARK1_READ
  }

  11 {
    SELECT * FROM t1
  } {} {
    $READMARK1_SET
    $READMARK1_READ
  }
} {

  set L [list]
  foreach el [subst $expected_locks] { lappend L $el }

  set S ""
  foreach sq [split $sql "\n"] { 
    set sq [string trim $sq]
    if {[string match {#*} $sq]==0} {append S "$sq\n"}
  }

  set ::locks [list]
  do_test wal2-6.4.$tn.1 { execsql $S } $res
  do_test wal2-6.4.$tn.2 { set ::locks  } $L
}

db close
tvfs delete

do_test wal2-6.5.1 {
  sqlite3 db test.db
  execsql {
    PRAGMA journal_mode = wal;

        INSERT INTO t1 VALUES(randomblob(100));
        INSERT INTO t1 VALUES(randomblob(100));
        INSERT INTO t1 SELECT md5sum(x) FROM t1;
      COMMIT;
    }
  }

  # Turn off auto-checkpoint. Otherwise, an auto-checkpoint run by a
  # writer may cause the dedicated checkpoint thread to return an
  # SQLITE_BUSY error.
  #
  db eval { PRAGMA wal_autocheckpoint = 0 }

  set nRun 0
  while {[tt_continue]} {
    read_transaction
    write_transaction 
    incr nRun
  }
  set nRun

    INSERT INTO t1 VALUES(0, 0, 0);
  }
} -thread t 10 {

  set nextwrite $E(pid)

  proc wal_hook {zDb nEntry} {
    if {$nEntry>10} { 
      set rc [catch { db eval {PRAGMA wal_checkpoint} } msg]
      if {$rc && $msg != "database is locked"} { error $msg }
    }
    return 0
  }
  db wal_hook wal_hook

  while {[tt_continue]} {
    set max 0
    while { $max != ($nextwrite-1) && [tt_continue] } {