/ Check-in [16dcf5a6]
Login

Many hyperlinks are disabled.
Use anonymous login to enable hyperlinks.

Overview
Comment:Add state diagram to comments in experimental version of pager.c.
Downloads: Tarball | ZIP archive | SQL archive
Timelines: family | ancestors | descendants | both | experimental
Files: files | file ages | folders
SHA1:16dcf5a6d3ebff9a734705693b5ac740f4ff2cc1
User & Date: dan 2010-08-03 12:48:46
Context
2010-08-03
15:57
Add comments to clarify the purpose of the pager "error state". check-in: fde4c597 user: dan tags: experimental
12:48
Add state diagram to comments in experimental version of pager.c. check-in: 16dcf5a6 user: dan tags: experimental
06:42
Incremental checkin on pager state refactoring. check-in: 0a636798 user: dan tags: experimental
Changes
Hide Diffs Unified Diffs Ignore Whitespace Patch

Changes to src/pager.c.

126
127
128
129
130
131
132

133
134
135
136
137
138
139
...
143
144
145
146
147
148
149
150

151
152
153
154
155
156
157
...
174
175
176
177
178
179
180




















181
182
183
184
185
186
187
188
...
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
...
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
...
652
653
654
655
656
657
658


659
660
661
662
663
664
665
...
667
668
669
670
671
672
673

674
675
676
677
678
679
680
...
698
699
700
701
702
703
704
705
706
707

708
709
710
711
712
713
714
715
716
717
718
719






720
721
722
723
724
725
726
727
....
2638
2639
2640
2641
2642
2643
2644









2645
2646
2647
2648
2649
2650
2651
....
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
....
3112
3113
3114
3115
3116
3117
3118

3119
3120

3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
....
3394
3395
3396
3397
3398
3399
3400
3401
3402
3403
3404
3405
3406
3407
3408
....
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437

3438
3439
3440
3441
3442
3443
3444
....
3512
3513
3514
3515
3516
3517
3518
3519




3520

3521

3522
3523
3524
3525
3526
3527
3528
....
3588
3589
3590
3591
3592
3593
3594
3595
3596
3597
3598


3599
3600
3601
3602
3603
3604
3605
....
3793
3794
3795
3796
3797
3798
3799
3800
3801
3802
3803
3804
3805
3806
3807
3808
3809
3810
3811
3812
3813
3814
3815
....
4620
4621
4622
4623
4624
4625
4626
4627
4628
4629
4630
4631
4632
4633
4634
4635
4636
4637
4638
4639
4640
4641
4642
4643
4644
4645
4646
4647
4648
4649
4650
4651
4652
....
4749
4750
4751
4752
4753
4754
4755
4756
4757
4758
4759
4760
4761
4762
4763
4764
4765
4766
4767
4768
4769
4770
4771
4772
4773
4774
4775
4776
4777
4778
....
4797
4798
4799
4800
4801
4802
4803
4804
4805
4806
4807
4808
4809
4810
4811
4812
4813
4814
4815
4816

4817
4818
4819
4820
4821
4822
4823
....
4891
4892
4893
4894
4895
4896
4897
4898
4899
4900
4901
4902
4903
4904
4905
....
4963
4964
4965
4966
4967
4968
4969

4970
4971
4972
4973
4974
4975
4976
4977
4978
4979
4980
4981
4982
4983
4984
....
5006
5007
5008
5009
5010
5011
5012
5013
5014
5015
5016
5017
5018
5019
5020
5021
5022
....
5028
5029
5030
5031
5032
5033
5034
5035
5036
5037
5038
5039
5040
5041
5042
....
5076
5077
5078
5079
5080
5081
5082



5083
5084
5085
5086
5087
5088
5089
....
5097
5098
5099
5100
5101
5102
5103
5104
5105
5106
5107
5108
5109
5110
5111
5112
5113
5114
5115
5116
5117
5118
5119
5120
5121
5122
5123
....
5508
5509
5510
5511
5512
5513
5514
5515
5516
5517
5518
5519
5520
5521
5522
....
5768
5769
5770
5771
5772
5773
5774



5775
5776
5777
5778
5779
5780
5781
5782
5783
5784
5785
5786
5787
5788
5789
....
5792
5793
5794
5795
5796
5797
5798
5799
5800
5801
5802
5803
5804
5805
5806
5807
5808
5809
5810
5811
5812
5813
#define FILEHANDLEID(fd) ((int)fd)

/*
** The Pager.eState variable stores the current 'state' of a pager. A
** pager may be in any one of the following six states:
**
**  NONE:

**    The pager starts up in this state. Nothing is guaranteed in this
**    state - the file may or may not be locked and the database size is
**    unknown. The database may not be read or written.
**
**    * No read or write transaction is active.
**    * Any lock, or no lock at all, may be held on the database file.
**    * The dbSize, dbOrigSize and dbFileSize variables may not be trusted.
................................................................................
**    rollback (non-WAL) mode are met. Unless the pager is (or recently
**    was) in exclusive-locking mode, a user-level read transaction is 
**    open. The database size is known in this state.
** 
**    * A read transaction may be active.
**    * A SHARED or greater lock is held on the database file.
**    * The dbSize variable may be trusted (even if a user-level read 
**      transaction is not active).

**
**  WRITER_INITIAL:
**    * A write transaction is active.
**    * A RESERVED or greater lock is held on the database file.
**    * The dbSize, dbOrigSize and dbFileSize variables are all valid.
**    * The contents of the pager cache have not been modified.
**
................................................................................
**    * A write transaction is active.
**    * An EXCLUSIVE or greater lock is held on the database file.
**    * All writing and syncing of journal and database data has finished.
**      If no error occured, all that remains is to finalize the journal to
**      commit the transaction. If an error did occur, the caller will need
**      to rollback the transaction. 
**




















** Allowable transitions and the [function] that performs each:
** 
**   NONE              -> READER              [PagerSharedLock]
**   READER            -> WRITER_INITIAL      [PagerBegin]
**   WRITER_INITIAL    -> WRITER_CACHEMOD     [pager_open_journal]
**   WRITER_CACHEMOD   -> WRITER_DBMOD        [syncJournal]
**   WRITER_DBMOD      -> WRITER_FINISHED     [PagerCommitPhaseOne]
** 
................................................................................
**   or SQLITE_FULL. Once one of the first three errors occurs, it persists
**   and is returned as the result of every major pager API call.  The
**   SQLITE_FULL return code is slightly different. It persists only until the
**   next successful rollback is performed on the pager cache. Also,
**   SQLITE_FULL does not affect the sqlite3PagerGet() and sqlite3PagerLookup()
**   APIs, they may still be used successfully.
**
** dbSizeValid, dbSize, dbOrigSize, dbFileSize
**
**   Managing the size of the database file in pages is a little complicated.
**   The variable Pager.dbSize contains the number of pages that the database
**   image currently contains. As the database image grows or shrinks this
**   variable is updated. The variable Pager.dbFileSize contains the number
**   of pages in the database file. This may be different from Pager.dbSize
**   if some pages have been appended to the database image but not yet written
................................................................................
  ** operates.
  **
  ** The 'state' variable is described in more detail along with the
  ** descriptions of the values it may take - PAGER_UNLOCK etc. Many of the
  ** other variables in this block are described in the comment directly 
  ** above this class definition.
  */
#if 0
  u8 state;                   /* PAGER_UNLOCK, _SHARED, _RESERVED, etc. */
  u8 dbModified;              /* True if there are any changes to the Db */
  u8 journalStarted;          /* True if header of journal is synced */
  u8 dbSizeValid;             /* Set when dbSize is correct */
#endif
  u8 eState;                  /* Pager state (NONE, READER, WRITER_INITIAL..) */
  u8 eLock;                   /* Current lock held on database file */

  u8 changeCountDone;         /* Set after incrementing the change-counter */
  u8 setMaster;               /* True if a m-j name has been written to jrnl */
  u8 doNotSpill;              /* Do not spill the cache when non-zero */
  u8 doNotSyncSpill;          /* Do not do a spill that requires jrnl sync */
  u8 subjInMemory;            /* True to use in-memory sub-journals */
  Pgno dbSize;                /* Number of pages in the database */
  Pgno dbOrigSize;            /* dbSize before the current transaction */
................................................................................
      assert( p->eLock>=SHARED_LOCK || p->noReadlock );
      break;

    case PAGER_WRITER_INITIAL:
      if( !pagerUseWal(pPager) ){
        assert( p->eLock>=RESERVED_LOCK );
      }


      break;

    case PAGER_WRITER_CACHEMOD:
      if( !pagerUseWal(pPager) ){
        /* It is possible that if journal_mode=wal here that neither the
        ** journal file nor the WAL file are open. This happens during
        ** a rollback transaction that switches from journal_mode=off
................................................................................
        */
        assert( p->eLock>=RESERVED_LOCK );
        assert( isOpen(p->jfd) 
             || p->journalMode==PAGER_JOURNALMODE_OFF 
             || p->journalMode==PAGER_JOURNALMODE_WAL 
        );
      }

      break;

    case PAGER_WRITER_DBMOD:
      assert( !pagerUseWal(pPager) );
      assert( p->eLock>=EXCLUSIVE_LOCK || pagerUseWal(pPager) );
      assert( isOpen(p->jfd) 
           || p->journalMode==PAGER_JOURNALMODE_OFF 
................................................................................
/*
** (gdb) printf "%s", print_pager_state(pPager)
*/
static char *print_pager_state(Pager *p){
  static char zRet[1024];

  sqlite3_snprintf(1024, zRet,
      "State: %s\n"
      "Lock:  %s\n"
      "Locking mode:  locking_mode=%s\n"

      "Backing store: tempFile=%d memDb=%d\n"
      , p->eState==PAGER_NONE            ? "NONE" :
        p->eState==PAGER_READER          ? "READER" :
        p->eState==PAGER_WRITER_INITIAL  ? "WRITER_INITIAL" :
        p->eState==PAGER_WRITER_CACHEMOD ? "WRITER_CACHEMOD" :
        p->eState==PAGER_WRITER_DBMOD    ? "WRITER_DBMOD" :
        p->eState==PAGER_WRITER_FINISHED ? "WRITER_FINISHED" : "?error?"
      , p->eLock==NO_LOCK         ? "NONE" :
        p->eLock==RESERVED_LOCK   ? "RESERVED" :
        p->eLock==EXCLUSIVE_LOCK  ? "EXCLUSIVE" :
        p->eLock==SHARED_LOCK     ? "SHARED" : "?error?"
      , p->exclusiveMode ? "exclusive" : "normal"






      , (int)p->tempFile, (int)p->memDb
  );

  return zRet;
}
#endif

/*
................................................................................
      }
    }
    nPage = (Pgno)(n / pPager->pageSize);
    if( nPage==0 && n>0 ){
      nPage = 1;
    }
  }









  *pnPage = nPage;
  return SQLITE_OK;
}


/*
** Check if the *-wal file that corresponds to the database opened by pPager
................................................................................
** Attempt to set the maximum database page count if mxPage is positive. 
** Make no changes if mxPage is zero or negative.  And never reduce the
** maximum page count below the current size of the database.
**
** Regardless of mxPage, return the current maximum page count.
*/
int sqlite3PagerMaxPageCount(Pager *pPager, int mxPage){
  int nPage;
  if( mxPage>0 ){
    pPager->mxPgno = mxPage;
  }
  if( pPager->eState!=PAGER_NONE ){
    sqlite3PagerPagecount(pPager, &nPage);
    assert( (int)pPager->mxPgno>=nPage );
  }
  return pPager->mxPgno;
}

/*
** The following set of routines are used to disable the simulated
** I/O error mechanism.  These routines are used to avoid simulated
................................................................................
      rc = SQLITE_OK;
    }
  }
  return rc;
}

/*

** Return the total number of pages in the database file associated 
** with pPager. Normally, this is calculated as (<db file size>/<page-size>).

** However, if the file is between 1 and <page-size> bytes in size, then 
** this is considered a 1 page file.
**
** If the pager is in error state when this function is called, then the
** error state error code is returned and *pnPage left unchanged. Or,
** if the file system has to be queried for the size of the file and
** the query attempt returns an IO error, the IO error code is returned
** and *pnPage is left unchanged.
**
** Otherwise, if everything is successful, then SQLITE_OK is returned
** and *pnPage is set to the number of pages in the database.
*/
int sqlite3PagerPagecount(Pager *pPager, int *pnPage){
  Pgno nPage = 0;           /* Value to return via *pnPage */

  assert( pPager->eState>=PAGER_SHARED );
  assert( pPager->eState!=PAGER_WRITER_FINISHED );

  /* Determine the number of pages in the file. Store this in nPage. */
  if( pPager->eState>PAGER_NONE ){
    nPage = pPager->dbSize;
  }else{
    int rc;                 /* Error returned by OsFileSize() */
    i64 n = 0;              /* File size in bytes returned by OsFileSize() */

    if( pagerUseWal(pPager) && pPager->eState!=PAGER_NONE ){
      nPage = sqlite3WalDbsize(pPager->pWal);
    }

    if( nPage==0 ){
      assert( isOpen(pPager->fd) || pPager->tempFile );
      if( isOpen(pPager->fd) ){
        if( SQLITE_OK!=(rc = sqlite3OsFileSize(pPager->fd, &n)) ){
          pager_error(pPager, rc);
          return rc;
        }
      }
      if( n>0 && n<pPager->pageSize ){
        nPage = 1;
      }else{
        nPage = (Pgno)(n / pPager->pageSize);
      }
    }
    if( pPager->eState!=PAGER_NONE ){
      pPager->dbSize = nPage;
      pPager->dbFileSize = nPage;
    }
  }

  /* If the current number of pages in the file is greater than the 
  ** configured maximum pager number, increase the allowed limit so
  ** that the file can be read.
  */
  if( nPage>pPager->mxPgno ){
    pPager->mxPgno = (Pgno)nPage;
  }

  /* Set the output variable and return SQLITE_OK */
  *pnPage = nPage;
  return SQLITE_OK;
}


static int pagerLock(Pager *pPager, int eLock){
  int rc;
  assert( eLock==SHARED_LOCK || eLock==RESERVED_LOCK || eLock==EXCLUSIVE_LOCK );
................................................................................
}

/*
** Sync the journal. In other words, make sure all the pages that have
** been written to the journal have actually reached the surface of the
** disk and can be restored in the event of a hot-journal rollback.
**
** If the Pager.noSync flag is set, then this function is a no-op. 
** Otherwise, the actions required depend on the journal-mode and the 
** device characteristics of the the file-system, as follows:
**
**   * If the journal file is an in-memory journal file, no action need
**     be taken.
**
**   * Otherwise, if the device does not support the SAFE_APPEND property,
................................................................................
**     if( NOT SEQUENTIAL ) xSync(<journal file>);
**   }
**
** If successful, this routine clears the PGHDR_NEED_SYNC flag of every 
** page currently held in memory before returning SQLITE_OK. If an IO
** error is encountered, then the IO error code is returned to the caller.
*/
static int syncJournal(Pager *pPager){
  int rc;                         /* Return code */

  assert( pPager->eState==PAGER_WRITER_CACHEMOD
       || pPager->eState==PAGER_WRITER_DBMOD
  );
  assert( assert_pager_state(pPager) );


  rc = sqlite3PagerExclusiveLock(pPager);
  if( rc!=SQLITE_OK ) return rc;

  if( !pPager->noSync ){
    assert( !pPager->tempFile );
    if( isOpen(pPager->jfd) && pPager->journalMode!=PAGER_JOURNALMODE_MEMORY ){
................................................................................
        PAGERTRACE(("SYNC journal of %d\n", PAGERID(pPager)));
        IOTRACE(("JSYNC %p\n", pPager))
        rc = sqlite3OsSync(pPager->jfd, pPager->sync_flags| 
          (pPager->sync_flags==SQLITE_SYNC_FULL?SQLITE_SYNC_DATAONLY:0)
        );
        if( rc!=SQLITE_OK ) return rc;
      }
    }






    pPager->journalHdr = pPager->journalOff;

  }

  /* Unless the pager is in noSync mode, the journal file was just 
  ** successfully synced. Either way, clear the PGHDR_NEED_SYNC flag on 
  ** all pages.
  */
  sqlite3PcacheClearSyncFlags(pPager->pPCache);
................................................................................
  ** However it may be called in WRITER_CACHEMOD state if the page being
  ** written (and all other pages that reside on the same disk sector) was
  ** a free-list leaf page at the start of the transaction. In that case
  ** the database file is not really being modified, so it is Ok to write
  ** to it in CACHEMOD state.
  */
  assert( !pagerUseWal(pPager) );
  assert( pPager->eState==PAGER_WRITER_CACHEMOD
       || pPager->eState==PAGER_WRITER_DBMOD 
  );
  assert( pPager->eState==PAGER_WRITER_DBMOD || pList->pDirty==0 );



  rc = pager_wait_on_lock(pPager, EXCLUSIVE_LOCK);

  /* If the file is a temp-file has not yet been opened, open it now. It
  ** is not possible for rc to be other than SQLITE_OK if this branch
  ** is taken, as pager_wait_on_lock() is a no-op for temp-files.
  */
................................................................................
    if( rc==SQLITE_OK ){
      rc = pagerWalFrames(pPager, pPg, 0, 0, 0);
    }
  }else{
  
    /* Sync the journal file if required. */
    if( pPg->flags&PGHDR_NEED_SYNC ){
      rc = syncJournal(pPager);
      if( rc==SQLITE_OK && 
        !pPager->noSync &&
        !(pPager->journalMode==PAGER_JOURNALMODE_MEMORY) &&
        !(sqlite3OsDeviceCharacteristics(pPager->fd)&SQLITE_IOCAP_SAFE_APPEND)
      ){
        pPager->nRec = 0;
        rc = writeJournalHdr(pPager);
      }
    }
  
    /* If the page number of this page is larger than the current size of
    ** the database image, it may need to be written to the sub-journal.
    ** This is because the call to pager_write_pagelist() below will not
    ** actually write data to the file in this case.
    **
................................................................................
    assert( pgno<=PAGER_MAX_PGNO && pgno!=PAGER_MJ_PGNO(pPager) );
    PAGER_INCR(pPager->nHit);
    return SQLITE_OK;

  }else{
    /* The pager cache has created a new page. Its content needs to 
    ** be initialized.  */
    int nMax;

    PAGER_INCR(pPager->nMiss);
    pPg = *ppPage;
    pPg->pPager = pPager;

    /* The maximum page number is 2^31. Return SQLITE_CORRUPT if a page
    ** number greater than this, or the unused locking-page, is requested. */
    if( pgno>PAGER_MAX_PGNO || pgno==PAGER_MJ_PGNO(pPager) ){
      rc = SQLITE_CORRUPT_BKPT;
      goto pager_acquire_err;
    }

    rc = sqlite3PagerPagecount(pPager, &nMax);
    if( rc!=SQLITE_OK ){
      goto pager_acquire_err;
    }

    if( MEMDB || nMax<(int)pgno || noContent || !isOpen(pPager->fd) ){
      if( pgno>pPager->mxPgno ){
        rc = SQLITE_FULL;
        goto pager_acquire_err;
      }
      if( noContent ){
        /* Failure to set the bits in the InJournal bit-vectors is benign.
        ** It merely means that we might do some extra work to journal a 
................................................................................
**
** Return SQLITE_OK if everything is successful. Otherwise, return 
** SQLITE_NOMEM if the attempt to allocate Pager.pInJournal fails, or 
** an IO error code if opening or writing the journal file fails.
*/
static int pager_open_journal(Pager *pPager){
  int rc = SQLITE_OK;                        /* Return code */
  int nPage;                                 /* Size of database file */
  sqlite3_vfs * const pVfs = pPager->pVfs;   /* Local cache of vfs pointer */

  assert( pPager->eState==PAGER_WRITER_INITIAL );
  assert( assert_pager_state(pPager) );
  assert( pPager->pInJournal==0 );
  
  /* If already in the error state, this function is a no-op.  But on
  ** the other hand, this routine is never called if we are already in
  ** an error state. */
  if( NEVER(pPager->errCode) ) return pPager->errCode;

  if( !pagerUseWal(pPager) && pPager->journalMode!=PAGER_JOURNALMODE_OFF ){
    rc = sqlite3PagerPagecount(pPager, &nPage);
    if( rc ) return rc;
    pPager->pInJournal = sqlite3BitvecCreate(nPage);
    if( pPager->pInJournal==0 ){
      return SQLITE_NOMEM;
    }
  
    /* Open the journal file if it is not already open. */
    if( !isOpen(pPager->jfd) ){
      if( pPager->journalMode==PAGER_JOURNALMODE_MEMORY ){
................................................................................
  
  
    /* Write the first journal header to the journal file and open 
    ** the sub-journal if necessary.
    */
    if( rc==SQLITE_OK ){
      /* TODO: Check if all of these are really required. */
      pPager->dbOrigSize = pPager->dbSize;
      pPager->nRec = 0;
      pPager->journalOff = 0;
      pPager->setMaster = 0;
      pPager->journalHdr = 0;
      rc = writeJournalHdr(pPager);
    }
  }

  if( rc!=SQLITE_OK ){
    sqlite3BitvecDestroy(pPager->pInJournal);
    pPager->pInJournal = 0;
  }else{

    pPager->eState = PAGER_WRITER_CACHEMOD;
  }

  return rc;
}

/*
................................................................................
      ** when it has an open transaction, but never to DBMOD or FINISHED.
      ** This is because in those states 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.
      */
      pPager->eState = PAGER_WRITER_INITIAL;
      pPager->dbOrigSize = pPager->dbSize;
      pPager->journalOff = 0;
    }else{
      /* Ignore any IO error that occurs within pager_end_transaction(). The
      ** purpose of this call is to reset the internal state of the pager
      ** sub-system. It doesn't matter if the journal-file is not properly
      ** finalized at this point (since it is not a valid journal file anyway).
      */
................................................................................
    ** yet be open. Open it now if this is the case.
    */
    if( pPager->eState==PAGER_WRITER_INITIAL ){
      rc = pager_open_journal(pPager);
      if( rc!=SQLITE_OK ) return rc;
    }
    assert( pPager->eState>=PAGER_WRITER_CACHEMOD );

  
    /* The transaction journal now exists and we have a RESERVED or an
    ** EXCLUSIVE lock on the main database file.  Write the current page to
    ** the transaction journal if it is not there already.
    */
    if( !pageInJournal(pPg) && isOpen(pPager->jfd) ){
      assert( !pagerUseWal(pPager) );
      if( pPg->pgno<=pPager->dbOrigSize ){
        u32 cksum;
        char *pData2;

        /* We should never write to the journal file the page that
        ** contains the database locks.  The following assert verifies
        ** that we do not. */
        assert( pPg->pgno!=PAGER_MJ_PGNO(pPager) );
................................................................................
        /* Even if an IO or diskfull error occurred while journalling the
        ** page in the block above, set the need-sync flag for the page.
        ** Otherwise, when the transaction is rolled back, the logic in
        ** playback_one_page() will think that the page needs to be restored
        ** in the database file. And if an IO error occurs while doing so,
        ** then corruption may follow.
        */
        if( 1 || !pPager->noSync ){
          pPg->flags |= PGHDR_NEED_SYNC;
        }

        /* An error has occurred writing to the journal file. The 
        ** transaction will be rolled back by the layer above.
        */
        if( rc!=SQLITE_OK ){
          return rc;
        }
................................................................................
        assert( rc==SQLITE_OK || rc==SQLITE_NOMEM );
        rc |= addToSavepointBitvecs(pPager, pPg->pgno);
        if( rc!=SQLITE_OK ){
          assert( rc==SQLITE_NOMEM );
          return rc;
        }
      }else{
        if( pPager->eState!=PAGER_WRITER_DBMOD && (!pPager->noSync || 1)){
          pPg->flags |= PGHDR_NEED_SYNC;
        }
        PAGERTRACE(("APPEND %d page %d needSync=%d\n",
                PAGERID(pPager), pPg->pgno,
               ((pPg->flags&PGHDR_NEED_SYNC)?1:0)));
      }
    }
................................................................................
int sqlite3PagerWrite(DbPage *pDbPage){
  int rc = SQLITE_OK;

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




  if( nPagePerSector>1 ){
    Pgno nPageCount;          /* Total number of pages in database file */
    Pgno pg1;                 /* First page of the sector pPg is located on. */
    int nPage = 0;            /* Number of pages starting at pg1 to journal */
    int ii;                   /* Loop counter */
    int needSync = 0;         /* True if any page has PGHDR_NEED_SYNC */

................................................................................

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

    rc = sqlite3PagerPagecount(pPager, (int *)&nPageCount);
    if( rc==SQLITE_OK ){
      if( pPg->pgno>nPageCount ){
        nPage = (pPg->pgno - pg1)+1;
      }else if( (pg1+nPagePerSector-1)>nPageCount ){
        nPage = nPageCount+1-pg1;
      }else{
        nPage = nPagePerSector;
      }
      assert(nPage>0);
      assert(pg1<=pPg->pgno);
      assert((pg1+nPage)>pPg->pgno);
    }

    for(ii=0; ii<nPage && rc==SQLITE_OK; ii++){
      Pgno pg = pg1+ii;
      PgHdr *pPage;
      if( pg==pPg->pgno || !sqlite3BitvecTest(pPager->pInJournal, pg) ){
        if( pg!=PAGER_MJ_PGNO(pPager) ){
          rc = sqlite3PagerGet(pPager, pg, &pPage);
................................................................................
      ** Because the change-counter page was just modified, unless the
      ** atomic-update optimization is used it is almost certain that the
      ** journal requires a sync here. However, in locking_mode=exclusive
      ** on a system under memory pressure it is just possible that this is 
      ** not the case. In this case it is likely enough that the redundant
      ** xSync() call will be changed to a no-op by the OS anyhow. 
      */
      rc = syncJournal(pPager);
      if( rc!=SQLITE_OK ) goto commit_phase_one_exit;

      rc = pager_write_pagelist(pPager,sqlite3PcacheDirtyList(pPager->pPCache));
      if( rc!=SQLITE_OK ){
        assert( rc!=SQLITE_IOERR_BLOCKED );
        goto commit_phase_one_exit;
      }
................................................................................
** If a memory allocation fails, SQLITE_NOMEM is returned. If an error 
** occurs while opening the sub-journal file, then an IO error code is
** returned. Otherwise, SQLITE_OK.
*/
int sqlite3PagerOpenSavepoint(Pager *pPager, int nSavepoint){
  int rc = SQLITE_OK;                       /* Return code */
  int nCurrent = pPager->nSavepoint;        /* Current number of savepoints */




  if( nSavepoint>nCurrent && pPager->useJournal ){
    int ii;                                 /* Iterator variable */
    PagerSavepoint *aNew;                   /* New Pager.aSavepoint array */
    int nPage;                              /* Size of database file */

    rc = sqlite3PagerPagecount(pPager, &nPage);
    if( rc ) return rc;

    /* Grow the Pager.aSavepoint array using realloc(). Return SQLITE_NOMEM
    ** if the allocation fails. Otherwise, zero the new portion in case a 
    ** malloc failure occurs while populating it in the for(...) loop below.
    */
    aNew = (PagerSavepoint *)sqlite3Realloc(
        pPager->aSavepoint, sizeof(PagerSavepoint)*nSavepoint
................................................................................
      return SQLITE_NOMEM;
    }
    memset(&aNew[nCurrent], 0, (nSavepoint-nCurrent) * sizeof(PagerSavepoint));
    pPager->aSavepoint = aNew;

    /* Populate the PagerSavepoint structures just allocated. */
    for(ii=nCurrent; ii<nSavepoint; ii++){
      aNew[ii].nOrig = nPage;
      if( isOpen(pPager->jfd) && pPager->journalOff>0 ){
        aNew[ii].iOffset = pPager->journalOff;
      }else{
        aNew[ii].iOffset = JOURNAL_HDR_SZ(pPager);
      }
      aNew[ii].iSubRec = pPager->nSubRec;
      aNew[ii].pInSavepoint = sqlite3BitvecCreate(nPage);
      if( !aNew[ii].pInSavepoint ){
        return SQLITE_NOMEM;
      }
      if( pagerUseWal(pPager) ){
        sqlite3WalSavepoint(pPager->pWal, aNew[ii].aWalData);
      }
      pPager->nSavepoint = ii+1;







>







 







|
>







 







>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
|







 







|







 







<
<
<
<
<
<


<







 







>
>







 







>







 







|
|

>
|











>
>
>
>
>
>
|







 







>
>
>
>
>
>
>
>
>







 







<



|
<
|







 







>
|
<
>


<
<
<
<
<
<
<
<
<


<
<


<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
|
<
<
<
<
<
<
<
<
<
<
<
<
<
<







 







|







 







|






>







 







|
>
>
>
>
|
>
|
>







 







|
|

|
>
>







 







|
<
<
<
<
<
<
<
<







 







<












<
<
<
<
<
|







 







<












<
<
|







 







<












>







 







|







 







>





|
|
|







 







<
|
<







 







|







 







>
>
>







 







|
<
|
|
|
|
|
|
|
|
|
|
<







 







|







 







>
>
>




<
<
<
<







 







|






|







126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
...
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
...
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
...
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
...
462
463
464
465
466
467
468






469
470

471
472
473
474
475
476
477
...
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
...
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
...
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
....
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
....
3075
3076
3077
3078
3079
3080
3081

3082
3083
3084
3085

3086
3087
3088
3089
3090
3091
3092
3093
....
3144
3145
3146
3147
3148
3149
3150
3151
3152

3153
3154
3155









3156
3157


3158
3159



























3160














3161
3162
3163
3164
3165
3166
3167
....
3375
3376
3377
3378
3379
3380
3381
3382
3383
3384
3385
3386
3387
3388
3389
....
3405
3406
3407
3408
3409
3410
3411
3412
3413
3414
3415
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
....
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
....
3576
3577
3578
3579
3580
3581
3582
3583
3584
3585
3586
3587
3588
3589
3590
3591
3592
3593
3594
3595
....
3783
3784
3785
3786
3787
3788
3789
3790








3791
3792
3793
3794
3795
3796
3797
....
4602
4603
4604
4605
4606
4607
4608

4609
4610
4611
4612
4613
4614
4615
4616
4617
4618
4619
4620





4621
4622
4623
4624
4625
4626
4627
4628
....
4725
4726
4727
4728
4729
4730
4731

4732
4733
4734
4735
4736
4737
4738
4739
4740
4741
4742
4743


4744
4745
4746
4747
4748
4749
4750
4751
....
4770
4771
4772
4773
4774
4775
4776

4777
4778
4779
4780
4781
4782
4783
4784
4785
4786
4787
4788
4789
4790
4791
4792
4793
4794
4795
4796
....
4864
4865
4866
4867
4868
4869
4870
4871
4872
4873
4874
4875
4876
4877
4878
....
4936
4937
4938
4939
4940
4941
4942
4943
4944
4945
4946
4947
4948
4949
4950
4951
4952
4953
4954
4955
4956
4957
4958
....
4980
4981
4982
4983
4984
4985
4986

4987

4988
4989
4990
4991
4992
4993
4994
....
5000
5001
5002
5003
5004
5005
5006
5007
5008
5009
5010
5011
5012
5013
5014
....
5048
5049
5050
5051
5052
5053
5054
5055
5056
5057
5058
5059
5060
5061
5062
5063
5064
....
5072
5073
5074
5075
5076
5077
5078
5079

5080
5081
5082
5083
5084
5085
5086
5087
5088
5089

5090
5091
5092
5093
5094
5095
5096
....
5481
5482
5483
5484
5485
5486
5487
5488
5489
5490
5491
5492
5493
5494
5495
....
5741
5742
5743
5744
5745
5746
5747
5748
5749
5750
5751
5752
5753
5754




5755
5756
5757
5758
5759
5760
5761
....
5764
5765
5766
5767
5768
5769
5770
5771
5772
5773
5774
5775
5776
5777
5778
5779
5780
5781
5782
5783
5784
5785
#define FILEHANDLEID(fd) ((int)fd)

/*
** The Pager.eState variable stores the current 'state' of a pager. A
** pager may be in any one of the following six states:
**
**  NONE:
**
**    The pager starts up in this state. Nothing is guaranteed in this
**    state - the file may or may not be locked and the database size is
**    unknown. The database may not be read or written.
**
**    * No read or write transaction is active.
**    * Any lock, or no lock at all, may be held on the database file.
**    * The dbSize, dbOrigSize and dbFileSize variables may not be trusted.
................................................................................
**    rollback (non-WAL) mode are met. Unless the pager is (or recently
**    was) in exclusive-locking mode, a user-level read transaction is 
**    open. The database size is known in this state.
** 
**    * A read transaction may be active.
**    * A SHARED or greater lock is held on the database file.
**    * The dbSize variable may be trusted (even if a user-level read 
**      transaction is not active). The dbOrigSize and dbFileSize variables
**      may not be trusted at this point.
**
**  WRITER_INITIAL:
**    * A write transaction is active.
**    * A RESERVED or greater lock is held on the database file.
**    * The dbSize, dbOrigSize and dbFileSize variables are all valid.
**    * The contents of the pager cache have not been modified.
**
................................................................................
**    * A write transaction is active.
**    * An EXCLUSIVE or greater lock is held on the database file.
**    * All writing and syncing of journal and database data has finished.
**      If no error occured, all that remains is to finalize the journal to
**      commit the transaction. If an error did occur, the caller will need
**      to rollback the transaction. 
**
** State diagram:
**
**                            NONE <-------------+
**                              |                |
**                              V                |
**               +---------> READER------------->|
**               |              |                |
**               |              V                |
**               |<-------WRITER_INITIAL-----> ERROR
**               |              |                ^  
**               |              V                |
**               |<------WRITER_CACHEMOD-------->|
**               |              |                |
**               |              V                |
**               |<-------WRITER_DBMOD---------->|
**               |              |                |
**               |              V                |
**               +<------WRITER_FINISHED-------->+
**
**
** State transitions and the [function] that performs each:
** 
**   NONE              -> READER              [PagerSharedLock]
**   READER            -> WRITER_INITIAL      [PagerBegin]
**   WRITER_INITIAL    -> WRITER_CACHEMOD     [pager_open_journal]
**   WRITER_CACHEMOD   -> WRITER_DBMOD        [syncJournal]
**   WRITER_DBMOD      -> WRITER_FINISHED     [PagerCommitPhaseOne]
** 
................................................................................
**   or SQLITE_FULL. Once one of the first three errors occurs, it persists
**   and is returned as the result of every major pager API call.  The
**   SQLITE_FULL return code is slightly different. It persists only until the
**   next successful rollback is performed on the pager cache. Also,
**   SQLITE_FULL does not affect the sqlite3PagerGet() and sqlite3PagerLookup()
**   APIs, they may still be used successfully.
**
** dbSize, dbOrigSize, dbFileSize
**
**   Managing the size of the database file in pages is a little complicated.
**   The variable Pager.dbSize contains the number of pages that the database
**   image currently contains. As the database image grows or shrinks this
**   variable is updated. The variable Pager.dbFileSize contains the number
**   of pages in the database file. This may be different from Pager.dbSize
**   if some pages have been appended to the database image but not yet written
................................................................................
  ** operates.
  **
  ** The 'state' variable is described in more detail along with the
  ** descriptions of the values it may take - PAGER_UNLOCK etc. Many of the
  ** other variables in this block are described in the comment directly 
  ** above this class definition.
  */






  u8 eState;                  /* Pager state (NONE, READER, WRITER_INITIAL..) */
  u8 eLock;                   /* Current lock held on database file */

  u8 changeCountDone;         /* Set after incrementing the change-counter */
  u8 setMaster;               /* True if a m-j name has been written to jrnl */
  u8 doNotSpill;              /* Do not spill the cache when non-zero */
  u8 doNotSyncSpill;          /* Do not do a spill that requires jrnl sync */
  u8 subjInMemory;            /* True to use in-memory sub-journals */
  Pgno dbSize;                /* Number of pages in the database */
  Pgno dbOrigSize;            /* dbSize before the current transaction */
................................................................................
      assert( p->eLock>=SHARED_LOCK || p->noReadlock );
      break;

    case PAGER_WRITER_INITIAL:
      if( !pagerUseWal(pPager) ){
        assert( p->eLock>=RESERVED_LOCK );
      }
      assert( pPager->dbSize==pPager->dbOrigSize );
      assert( pPager->dbOrigSize==pPager->dbFileSize );
      break;

    case PAGER_WRITER_CACHEMOD:
      if( !pagerUseWal(pPager) ){
        /* It is possible that if journal_mode=wal here that neither the
        ** journal file nor the WAL file are open. This happens during
        ** a rollback transaction that switches from journal_mode=off
................................................................................
        */
        assert( p->eLock>=RESERVED_LOCK );
        assert( isOpen(p->jfd) 
             || p->journalMode==PAGER_JOURNALMODE_OFF 
             || p->journalMode==PAGER_JOURNALMODE_WAL 
        );
      }
      assert( pPager->dbOrigSize==pPager->dbFileSize );
      break;

    case PAGER_WRITER_DBMOD:
      assert( !pagerUseWal(pPager) );
      assert( p->eLock>=EXCLUSIVE_LOCK || pagerUseWal(pPager) );
      assert( isOpen(p->jfd) 
           || p->journalMode==PAGER_JOURNALMODE_OFF 
................................................................................
/*
** (gdb) printf "%s", print_pager_state(pPager)
*/
static char *print_pager_state(Pager *p){
  static char zRet[1024];

  sqlite3_snprintf(1024, zRet,
      "State:         %s\n"
      "Lock:          %s\n"
      "Locking mode:  locking_mode=%s\n"
      "Journal mode:  journal_mode=%s\n"
      "Backing store: tempFile=%d memDb=%d useJournal=%d\n"
      , p->eState==PAGER_NONE            ? "NONE" :
        p->eState==PAGER_READER          ? "READER" :
        p->eState==PAGER_WRITER_INITIAL  ? "WRITER_INITIAL" :
        p->eState==PAGER_WRITER_CACHEMOD ? "WRITER_CACHEMOD" :
        p->eState==PAGER_WRITER_DBMOD    ? "WRITER_DBMOD" :
        p->eState==PAGER_WRITER_FINISHED ? "WRITER_FINISHED" : "?error?"
      , p->eLock==NO_LOCK         ? "NONE" :
        p->eLock==RESERVED_LOCK   ? "RESERVED" :
        p->eLock==EXCLUSIVE_LOCK  ? "EXCLUSIVE" :
        p->eLock==SHARED_LOCK     ? "SHARED" : "?error?"
      , p->exclusiveMode ? "exclusive" : "normal"
      , p->journalMode==PAGER_JOURNALMODE_MEMORY   ? "memory" :
        p->journalMode==PAGER_JOURNALMODE_OFF      ? "off" :
        p->journalMode==PAGER_JOURNALMODE_DELETE   ? "delete" :
        p->journalMode==PAGER_JOURNALMODE_PERSIST  ? "persist" :
        p->journalMode==PAGER_JOURNALMODE_TRUNCATE ? "truncate" :
        p->journalMode==PAGER_JOURNALMODE_WAL      ? "wal" : "?error?"
      , (int)p->tempFile, (int)p->memDb, (int)p->useJournal
  );

  return zRet;
}
#endif

/*
................................................................................
      }
    }
    nPage = (Pgno)(n / pPager->pageSize);
    if( nPage==0 && n>0 ){
      nPage = 1;
    }
  }

  /* If the current number of pages in the file is greater than the 
  ** configured maximum pager number, increase the allowed limit so
  ** that the file can be read.
  */
  if( nPage>pPager->mxPgno ){
    pPager->mxPgno = (Pgno)nPage;
  }

  *pnPage = nPage;
  return SQLITE_OK;
}


/*
** Check if the *-wal file that corresponds to the database opened by pPager
................................................................................
** Attempt to set the maximum database page count if mxPage is positive. 
** Make no changes if mxPage is zero or negative.  And never reduce the
** maximum page count below the current size of the database.
**
** Regardless of mxPage, return the current maximum page count.
*/
int sqlite3PagerMaxPageCount(Pager *pPager, int mxPage){

  if( mxPage>0 ){
    pPager->mxPgno = mxPage;
  }
  if( pPager->eState!=PAGER_NONE && pPager->mxPgno<pPager->dbSize ){

    pPager->mxPgno = pPager->dbSize;
  }
  return pPager->mxPgno;
}

/*
** The following set of routines are used to disable the simulated
** I/O error mechanism.  These routines are used to avoid simulated
................................................................................
      rc = SQLITE_OK;
    }
  }
  return rc;
}

/*
** This function may only be called when a read-transaction is open on
** the pager. It returns the total number of pages in the database.

**
** However, if the file is between 1 and <page-size> bytes in size, then 
** this is considered a 1 page file.









*/
int sqlite3PagerPagecount(Pager *pPager, int *pnPage){


  assert( pPager->eState>=PAGER_SHARED );
  assert( pPager->eState!=PAGER_WRITER_FINISHED );



























  *pnPage = (int)pPager->dbSize;














  return SQLITE_OK;
}


static int pagerLock(Pager *pPager, int eLock){
  int rc;
  assert( eLock==SHARED_LOCK || eLock==RESERVED_LOCK || eLock==EXCLUSIVE_LOCK );
................................................................................
}

/*
** Sync the journal. In other words, make sure all the pages that have
** been written to the journal have actually reached the surface of the
** disk and can be restored in the event of a hot-journal rollback.
**
** If the Pager.noSync flag is set, then this function is a no-op.
** Otherwise, the actions required depend on the journal-mode and the 
** device characteristics of the the file-system, as follows:
**
**   * If the journal file is an in-memory journal file, no action need
**     be taken.
**
**   * Otherwise, if the device does not support the SAFE_APPEND property,
................................................................................
**     if( NOT SEQUENTIAL ) xSync(<journal file>);
**   }
**
** If successful, this routine clears the PGHDR_NEED_SYNC flag of every 
** page currently held in memory before returning SQLITE_OK. If an IO
** error is encountered, then the IO error code is returned to the caller.
*/
static int syncJournal(Pager *pPager, int newHdr){
  int rc;                         /* Return code */

  assert( pPager->eState==PAGER_WRITER_CACHEMOD
       || pPager->eState==PAGER_WRITER_DBMOD
  );
  assert( assert_pager_state(pPager) );
  assert( !pagerUseWal(pPager) );

  rc = sqlite3PagerExclusiveLock(pPager);
  if( rc!=SQLITE_OK ) return rc;

  if( !pPager->noSync ){
    assert( !pPager->tempFile );
    if( isOpen(pPager->jfd) && pPager->journalMode!=PAGER_JOURNALMODE_MEMORY ){
................................................................................
        PAGERTRACE(("SYNC journal of %d\n", PAGERID(pPager)));
        IOTRACE(("JSYNC %p\n", pPager))
        rc = sqlite3OsSync(pPager->jfd, pPager->sync_flags| 
          (pPager->sync_flags==SQLITE_SYNC_FULL?SQLITE_SYNC_DATAONLY:0)
        );
        if( rc!=SQLITE_OK ) return rc;
      }

      pPager->journalHdr = pPager->journalOff;
      if( newHdr && 0==(iDc&SQLITE_IOCAP_SAFE_APPEND) ){
        pPager->nRec = 0;
        rc = writeJournalHdr(pPager);
      }
    }else{
      pPager->journalHdr = pPager->journalOff;
    }
  }

  /* Unless the pager is in noSync mode, the journal file was just 
  ** successfully synced. Either way, clear the PGHDR_NEED_SYNC flag on 
  ** all pages.
  */
  sqlite3PcacheClearSyncFlags(pPager->pPCache);
................................................................................
  ** However it may be called in WRITER_CACHEMOD state if the page being
  ** written (and all other pages that reside on the same disk sector) was
  ** a free-list leaf page at the start of the transaction. In that case
  ** the database file is not really being modified, so it is Ok to write
  ** to it in CACHEMOD state.
  */
  assert( !pagerUseWal(pPager) );
  assert( pPager->eState==PAGER_WRITER_DBMOD
       || pPager->eState==PAGER_WRITER_CACHEMOD 
  );
  assert( pPager->eState==PAGER_WRITER_DBMOD 
       || (pList->pDirty==0 && pList->pgno<=pPager->dbFileSize)
  );

  rc = pager_wait_on_lock(pPager, EXCLUSIVE_LOCK);

  /* If the file is a temp-file has not yet been opened, open it now. It
  ** is not possible for rc to be other than SQLITE_OK if this branch
  ** is taken, as pager_wait_on_lock() is a no-op for temp-files.
  */
................................................................................
    if( rc==SQLITE_OK ){
      rc = pagerWalFrames(pPager, pPg, 0, 0, 0);
    }
  }else{
  
    /* Sync the journal file if required. */
    if( pPg->flags&PGHDR_NEED_SYNC ){
      rc = syncJournal(pPager, 1);








    }
  
    /* If the page number of this page is larger than the current size of
    ** the database image, it may need to be written to the sub-journal.
    ** This is because the call to pager_write_pagelist() below will not
    ** actually write data to the file in this case.
    **
................................................................................
    assert( pgno<=PAGER_MAX_PGNO && pgno!=PAGER_MJ_PGNO(pPager) );
    PAGER_INCR(pPager->nHit);
    return SQLITE_OK;

  }else{
    /* The pager cache has created a new page. Its content needs to 
    ** be initialized.  */


    PAGER_INCR(pPager->nMiss);
    pPg = *ppPage;
    pPg->pPager = pPager;

    /* The maximum page number is 2^31. Return SQLITE_CORRUPT if a page
    ** number greater than this, or the unused locking-page, is requested. */
    if( pgno>PAGER_MAX_PGNO || pgno==PAGER_MJ_PGNO(pPager) ){
      rc = SQLITE_CORRUPT_BKPT;
      goto pager_acquire_err;
    }






    if( MEMDB || pPager->dbSize<(int)pgno || noContent || !isOpen(pPager->fd) ){
      if( pgno>pPager->mxPgno ){
        rc = SQLITE_FULL;
        goto pager_acquire_err;
      }
      if( noContent ){
        /* Failure to set the bits in the InJournal bit-vectors is benign.
        ** It merely means that we might do some extra work to journal a 
................................................................................
**
** Return SQLITE_OK if everything is successful. Otherwise, return 
** SQLITE_NOMEM if the attempt to allocate Pager.pInJournal fails, or 
** an IO error code if opening or writing the journal file fails.
*/
static int pager_open_journal(Pager *pPager){
  int rc = SQLITE_OK;                        /* Return code */

  sqlite3_vfs * const pVfs = pPager->pVfs;   /* Local cache of vfs pointer */

  assert( pPager->eState==PAGER_WRITER_INITIAL );
  assert( assert_pager_state(pPager) );
  assert( pPager->pInJournal==0 );
  
  /* If already in the error state, this function is a no-op.  But on
  ** the other hand, this routine is never called if we are already in
  ** an error state. */
  if( NEVER(pPager->errCode) ) return pPager->errCode;

  if( !pagerUseWal(pPager) && pPager->journalMode!=PAGER_JOURNALMODE_OFF ){


    pPager->pInJournal = sqlite3BitvecCreate(pPager->dbSize);
    if( pPager->pInJournal==0 ){
      return SQLITE_NOMEM;
    }
  
    /* Open the journal file if it is not already open. */
    if( !isOpen(pPager->jfd) ){
      if( pPager->journalMode==PAGER_JOURNALMODE_MEMORY ){
................................................................................
  
  
    /* Write the first journal header to the journal file and open 
    ** the sub-journal if necessary.
    */
    if( rc==SQLITE_OK ){
      /* TODO: Check if all of these are really required. */

      pPager->nRec = 0;
      pPager->journalOff = 0;
      pPager->setMaster = 0;
      pPager->journalHdr = 0;
      rc = writeJournalHdr(pPager);
    }
  }

  if( rc!=SQLITE_OK ){
    sqlite3BitvecDestroy(pPager->pInJournal);
    pPager->pInJournal = 0;
  }else{
    assert( pPager->eState==PAGER_WRITER_INITIAL );
    pPager->eState = PAGER_WRITER_CACHEMOD;
  }

  return rc;
}

/*
................................................................................
      ** when it has an open transaction, but never to DBMOD or FINISHED.
      ** This is because in those states 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.
      */
      pPager->eState = PAGER_WRITER_INITIAL;
      pPager->dbFileSize = pPager->dbOrigSize = pPager->dbSize;
      pPager->journalOff = 0;
    }else{
      /* Ignore any IO error that occurs within pager_end_transaction(). The
      ** purpose of this call is to reset the internal state of the pager
      ** sub-system. It doesn't matter if the journal-file is not properly
      ** finalized at this point (since it is not a valid journal file anyway).
      */
................................................................................
    ** yet be open. Open it now if this is the case.
    */
    if( pPager->eState==PAGER_WRITER_INITIAL ){
      rc = pager_open_journal(pPager);
      if( rc!=SQLITE_OK ) return rc;
    }
    assert( pPager->eState>=PAGER_WRITER_CACHEMOD );
    assert( assert_pager_state(pPager) );
  
    /* The transaction journal now exists and we have a RESERVED or an
    ** EXCLUSIVE lock on the main database file.  Write the current page to
    ** the transaction journal if it is not there already.
    */
    if( !pageInJournal(pPg) && !pagerUseWal(pPager) ){
      assert( pagerUseWal(pPager)==0 );
      if( pPg->pgno<=pPager->dbOrigSize && isOpen(pPager->jfd) ){
        u32 cksum;
        char *pData2;

        /* We should never write to the journal file the page that
        ** contains the database locks.  The following assert verifies
        ** that we do not. */
        assert( pPg->pgno!=PAGER_MJ_PGNO(pPager) );
................................................................................
        /* Even if an IO or diskfull error occurred while journalling the
        ** page in the block above, set the need-sync flag for the page.
        ** Otherwise, when the transaction is rolled back, the logic in
        ** playback_one_page() will think that the page needs to be restored
        ** in the database file. And if an IO error occurs while doing so,
        ** then corruption may follow.
        */

        pPg->flags |= PGHDR_NEED_SYNC;


        /* An error has occurred writing to the journal file. The 
        ** transaction will be rolled back by the layer above.
        */
        if( rc!=SQLITE_OK ){
          return rc;
        }
................................................................................
        assert( rc==SQLITE_OK || rc==SQLITE_NOMEM );
        rc |= addToSavepointBitvecs(pPager, pPg->pgno);
        if( rc!=SQLITE_OK ){
          assert( rc==SQLITE_NOMEM );
          return rc;
        }
      }else{
        if( pPager->eState!=PAGER_WRITER_DBMOD ){
          pPg->flags |= PGHDR_NEED_SYNC;
        }
        PAGERTRACE(("APPEND %d page %d needSync=%d\n",
                PAGERID(pPager), pPg->pgno,
               ((pPg->flags&PGHDR_NEED_SYNC)?1:0)));
      }
    }
................................................................................
int sqlite3PagerWrite(DbPage *pDbPage){
  int rc = SQLITE_OK;

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

  assert( pPager->eState>=PAGER_WRITER_INITIAL );
  assert( assert_pager_state(pPager) );

  if( nPagePerSector>1 ){
    Pgno nPageCount;          /* Total number of pages in database file */
    Pgno pg1;                 /* First page of the sector pPg is located on. */
    int nPage = 0;            /* Number of pages starting at pg1 to journal */
    int ii;                   /* Loop counter */
    int needSync = 0;         /* True if any page has PGHDR_NEED_SYNC */

................................................................................

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

    nPageCount = pPager->dbSize;

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


    for(ii=0; ii<nPage && rc==SQLITE_OK; ii++){
      Pgno pg = pg1+ii;
      PgHdr *pPage;
      if( pg==pPg->pgno || !sqlite3BitvecTest(pPager->pInJournal, pg) ){
        if( pg!=PAGER_MJ_PGNO(pPager) ){
          rc = sqlite3PagerGet(pPager, pg, &pPage);
................................................................................
      ** Because the change-counter page was just modified, unless the
      ** atomic-update optimization is used it is almost certain that the
      ** journal requires a sync here. However, in locking_mode=exclusive
      ** on a system under memory pressure it is just possible that this is 
      ** not the case. In this case it is likely enough that the redundant
      ** xSync() call will be changed to a no-op by the OS anyhow. 
      */
      rc = syncJournal(pPager, 0);
      if( rc!=SQLITE_OK ) goto commit_phase_one_exit;

      rc = pager_write_pagelist(pPager,sqlite3PcacheDirtyList(pPager->pPCache));
      if( rc!=SQLITE_OK ){
        assert( rc!=SQLITE_IOERR_BLOCKED );
        goto commit_phase_one_exit;
      }
................................................................................
** If a memory allocation fails, SQLITE_NOMEM is returned. If an error 
** occurs while opening the sub-journal file, then an IO error code is
** returned. Otherwise, SQLITE_OK.
*/
int sqlite3PagerOpenSavepoint(Pager *pPager, int nSavepoint){
  int rc = SQLITE_OK;                       /* Return code */
  int nCurrent = pPager->nSavepoint;        /* Current number of savepoints */

  assert( pPager->eState>=PAGER_WRITER_INITIAL );
  assert( assert_pager_state(pPager) );

  if( nSavepoint>nCurrent && pPager->useJournal ){
    int ii;                                 /* Iterator variable */
    PagerSavepoint *aNew;                   /* New Pager.aSavepoint array */





    /* Grow the Pager.aSavepoint array using realloc(). Return SQLITE_NOMEM
    ** if the allocation fails. Otherwise, zero the new portion in case a 
    ** malloc failure occurs while populating it in the for(...) loop below.
    */
    aNew = (PagerSavepoint *)sqlite3Realloc(
        pPager->aSavepoint, sizeof(PagerSavepoint)*nSavepoint
................................................................................
      return SQLITE_NOMEM;
    }
    memset(&aNew[nCurrent], 0, (nSavepoint-nCurrent) * sizeof(PagerSavepoint));
    pPager->aSavepoint = aNew;

    /* Populate the PagerSavepoint structures just allocated. */
    for(ii=nCurrent; ii<nSavepoint; ii++){
      aNew[ii].nOrig = pPager->dbSize;
      if( isOpen(pPager->jfd) && pPager->journalOff>0 ){
        aNew[ii].iOffset = pPager->journalOff;
      }else{
        aNew[ii].iOffset = JOURNAL_HDR_SZ(pPager);
      }
      aNew[ii].iSubRec = pPager->nSubRec;
      aNew[ii].pInSavepoint = sqlite3BitvecCreate(pPager->dbSize);
      if( !aNew[ii].pInSavepoint ){
        return SQLITE_NOMEM;
      }
      if( pagerUseWal(pPager) ){
        sqlite3WalSavepoint(pPager->pWal, aNew[ii].aWalData);
      }
      pPager->nSavepoint = ii+1;