SQLite

Check-in [1431be9557]
Login

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

Overview
Comment:Replace the sqlite3_io_methods.xMremap interface with sqlite3_io_methods.xFetch and xUnfetch.
Downloads: Tarball | ZIP archive
Timelines: family | ancestors | descendants | both | experimental-mmap
Files: files | file ages | folders
SHA1: 1431be95579160fb70408d43e17fc23c7b69ab4a
User & Date: dan 2013-03-23 21:00:41.457
Context
2013-03-25
13:50
Fix a case in the pager where an xFetch() reference was being leaked following an OOM error. (check-in: 5885ba6ce7 user: dan tags: experimental-mmap)
2013-03-23
21:00
Replace the sqlite3_io_methods.xMremap interface with sqlite3_io_methods.xFetch and xUnfetch. (check-in: 1431be9557 user: dan tags: experimental-mmap)
17:29
Improve a comment in wal.c. No code changes. (check-in: 60b9f5e4dd user: dan tags: experimental-mmap)
Changes
Side-by-Side Diff Ignore Whitespace Patch
Changes to src/backup.c.
516
517
518
519
520
521
522
523

524
525
526
527
528
529
530
516
517
518
519
520
521
522

523
524
525
526
527
528
529
530







-
+







            iOff+=pgszSrc
          ){
            PgHdr *pSrcPg = 0;
            const Pgno iSrcPg = (Pgno)((iOff/pgszSrc)+1);
            rc = sqlite3PagerGet(pSrcPager, iSrcPg, &pSrcPg);
            if( rc==SQLITE_OK ){
              u8 *zData = sqlite3PagerGetData(pSrcPg);
              rc = sqlite3PagerWriteData(pDestPager, zData, pgszSrc, iOff);
              rc = sqlite3OsWrite(pFile, zData, pgszSrc, iOff);
            }
            sqlite3PagerUnref(pSrcPg);
          }
          if( rc==SQLITE_OK ){
            rc = backupTruncateFile(pFile, iSize);
          }

Changes to src/os.c.
136
137
138
139
140
141
142
143
144


145
146

147
148
149


150
151

152
153
154
155
156
157
158
136
137
138
139
140
141
142


143
144


145



146
147


148
149
150
151
152
153
154
155







-
-
+
+
-
-
+
-
-
-
+
+
-
-
+







  int pgsz,
  int bExtend,                    /* True to extend file if necessary */
  void volatile **pp              /* OUT: Pointer to mapping */
){
  DO_OS_MALLOC_TEST(id);
  return id->pMethods->xShmMap(id, iPage, pgsz, bExtend, pp);
}
int sqlite3OsMremap(
  sqlite3_file *id,               /* Database file handle */

int sqlite3OsFetch(sqlite3_file *id, i64 iOff, int iAmt, void **pp){
  int flags,                      /* SQLITE_MREMAP_XXX flags */
  i64 iOff,                       /* Offset at which mapping(s) start */
  return id->pMethods->xFetch(id, iOff, iAmt, pp);
  i64 nOld,                       /* Size of old mapping */
  i64 nNew,                       /* Size of requested mapping */
  void **pp                       /* IN/OUT: Pointer to mapped region */
}
int sqlite3OsUnfetch(sqlite3_file *id, void *p){
){
  return id->pMethods->xMremap(id, flags, iOff, nOld, nNew, pp);
  return id->pMethods->xUnfetch(id, p);
}

/*
** The next group of routines are convenience wrappers around the
** VFS methods.
*/
int sqlite3OsOpen(
Changes to src/os.h.
255
256
257
258
259
260
261
262


263
264
265
266
267
268
269
255
256
257
258
259
260
261

262
263
264
265
266
267
268
269
270







-
+
+







#define SQLITE_FCNTL_DB_UNCHANGED 0xca093fa0
int sqlite3OsSectorSize(sqlite3_file *id);
int sqlite3OsDeviceCharacteristics(sqlite3_file *id);
int sqlite3OsShmMap(sqlite3_file *,int,int,int,void volatile **);
int sqlite3OsShmLock(sqlite3_file *id, int, int, int);
void sqlite3OsShmBarrier(sqlite3_file *id);
int sqlite3OsShmUnmap(sqlite3_file *id, int);
int sqlite3OsMremap(sqlite3_file *id, int, i64, i64, i64, void **);
int sqlite3OsFetch(sqlite3_file *id, i64, int, void **);
int sqlite3OsUnfetch(sqlite3_file *, void *);


/* 
** Functions for accessing sqlite3_vfs methods 
*/
int sqlite3OsOpen(sqlite3_vfs *, const char *, sqlite3_file*, int, int *);
int sqlite3OsDelete(sqlite3_vfs *, const char *, int);
Changes to src/os_unix.c.
245
246
247
248
249
250
251


252
253
254






255
256
257
258
259
260
261
245
246
247
248
249
250
251
252
253



254
255
256
257
258
259
260
261
262
263
264
265
266







+
+
-
-
-
+
+
+
+
+
+







  ** occur if a file is updated without also updating the transaction
  ** counter.  This test is made to avoid new problems similar to the
  ** one described by ticket #3584. 
  */
  unsigned char transCntrChng;   /* True if the transaction counter changed */
  unsigned char dbUpdate;        /* True if any part of database file changed */
  unsigned char inNormalWrite;   /* True if in a normal write operation */

#endif
  sqlite3_int64 mmapSize;        /* Size of xMremap() */
  void *pMapRegion;              /* Area memory mapped */
#endif
  sqlite3_int64 mmapSize;         /* Usable size of mapping at pMapRegion */
  sqlite3_int64 mmapOrigsize;     /* Actual size of mapping at pMapRegion */
  sqlite3_int64 mmapLimit;        /* Configured FCNTL_MMAP_SIZE value */
  void *pMapRegion;               /* Memory mapped region */
  int nFetchOut;                  /* Number of outstanding xFetch refs */

#ifdef SQLITE_TEST
  /* In test mode, increase the size of this structure a bit so that 
  ** it is larger than the struct CrashFile defined in test6.c.
  */
  char aPadding[32];
#endif
};
1801
1802
1803
1804
1805
1806
1807



1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819

1820
1821
1822
1823
1824
1825
1826
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835







+
+
+












+







** If the locking level of the file descriptor is already at or below
** the requested locking level, this routine is a no-op.
*/
static int unixUnlock(sqlite3_file *id, int eFileLock){
  return posixUnlock(id, eFileLock, 0);
}

static int unixMapfile(unixFile *pFd, i64 nByte);
static void unixUnmapfile(unixFile *pFd);

/*
** This function performs the parts of the "close file" operation 
** common to all locking schemes. It closes the directory and file
** handles, if they are valid, and sets all fields of the unixFile
** structure to 0.
**
** It is *not* necessary to hold the mutex when this routine is called,
** even on VxWorks.  A mutex will be acquired on VxWorks by the
** vxworksReleaseFileId() routine.
*/
static int closeUnixFile(sqlite3_file *id){
  unixFile *pFile = (unixFile*)id;
  unixUnmapfile(pFile);
  if( pFile->h>=0 ){
    robust_close(pFile, pFile->h, __LINE__);
    pFile->h = -1;
  }
#if OS_VXWORKS
  if( pFile->pId ){
    if( pFile->ctrlFlags & UNIXFILE_DELETE ){
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086















3087
3088
3089
3090
3091
3092
3093
3079
3080
3081
3082
3083
3084
3085

3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116







-









+
+
+
+
+
+
+
+
+
+
+
+
+
+
+







  void *pBuf, 
  int amt,
  sqlite3_int64 offset
){
  unixFile *pFile = (unixFile *)id;
  int got;
  assert( id );
  assert( offset>=pFile->mmapSize );  /* Never read from the mmapped region */

  /* If this is a database file (not a journal, master-journal or temp
  ** file), the bytes in the locking range should never be read or written. */
#if 0
  assert( pFile->pUnused==0
       || offset>=PENDING_BYTE+512
       || offset+amt<=PENDING_BYTE 
  );
#endif

  /* Deal with as much of this write request as possible by transfering
  ** data to the memory mapping using memcpy().  */
  if( offset<pFile->mmapSize ){
    if( offset+amt <= pFile->mmapSize ){
      memcpy(pBuf, &((u8 *)(pFile->pMapRegion))[offset], amt);
      return SQLITE_OK;
    }else{
      int nCopy = pFile->mmapSize - offset;
      memcpy(pBuf, &((u8 *)(pFile->pMapRegion))[offset], nCopy);
      pBuf = &((u8 *)pBuf)[nCopy];
      amt -= nCopy;
      offset += nCopy;
    }
  }

  got = seekAndRead(pFile, offset, pBuf, amt);
  if( got==amt ){
    return SQLITE_OK;
  }else if( got<0 ){
    /* lastErrno set by seekAndRead */
    return SQLITE_IOERR_READ;
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3176
3177
3178
3179
3180
3181
3182

3183
3184
3185
3186
3187
3188
3189







-







  int amt,
  sqlite3_int64 offset 
){
  unixFile *pFile = (unixFile*)id;
  int wrote = 0;
  assert( id );
  assert( amt>0 );
  assert( offset>=pFile->mmapSize );   /* Never write into the mmapped region */

  /* If this is a database file (not a journal, master-journal or temp
  ** file), the bytes in the locking range should never be read or written. */
#if 0
  assert( pFile->pUnused==0
       || offset>=PENDING_BYTE+512
       || offset+amt<=PENDING_BYTE 
3185
3186
3187
3188
3189
3190
3191















3192
3193
3194
3195
3196
3197
3198
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235







+
+
+
+
+
+
+
+
+
+
+
+
+
+
+







      SimulateIOErrorBenign(0);
      if( rc!=4 || memcmp(oldCntr, &((char*)pBuf)[24-offset], 4)!=0 ){
        pFile->transCntrChng = 1;  /* The transaction counter has changed */
      }
    }
  }
#endif

  /* Deal with as much of this write request as possible by transfering
  ** data from the memory mapping using memcpy().  */
  if( offset<pFile->mmapSize ){
    if( offset+amt <= pFile->mmapSize ){
      memcpy(&((u8 *)(pFile->pMapRegion))[offset], pBuf, amt);
      return SQLITE_OK;
    }else{
      int nCopy = pFile->mmapSize - offset;
      memcpy(&((u8 *)(pFile->pMapRegion))[offset], pBuf, nCopy);
      pBuf = &((u8 *)pBuf)[nCopy];
      amt -= nCopy;
      offset += nCopy;
    }
  }

  while( amt>0 && (wrote = seekAndWrite(pFile, offset, pBuf, amt))>0 ){
    amt -= wrote;
    offset += wrote;
    pBuf = &((char*)pBuf)[wrote];
  }
  SimulateIOError(( wrote=(-1), amt=1 ));
3466
3467
3468
3469
3470
3471
3472

3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518

3519
3520
3521
3522
3523
3524
3525







+








-







    ** that effectively updates the change counter.  This might happen
    ** when restoring a database using the backup API from a zero-length
    ** source.
    */
    if( pFile->inNormalWrite && nByte==0 ){
      pFile->transCntrChng = 1;
    }
#endif

    /* If the file was just truncated to a size smaller than the currently
    ** mapped region, reduce the effective mapping size as well. SQLite will
    ** use read() and write() to access data beyond this point from now on.  
    */
    if( nByte<pFile->mmapSize ){
      pFile->mmapSize = nByte;
    }
#endif

    return SQLITE_OK;
  }
}

/*
** Determine the current size of a file in bytes
3563
3564
3565
3566
3567
3568
3569













3570
3571
3572
3573
3574
3575
3576
3600
3601
3602
3603
3604
3605
3606
3607
3608
3609
3610
3611
3612
3613
3614
3615
3616
3617
3618
3619
3620
3621
3622
3623
3624
3625
3626







+
+
+
+
+
+
+
+
+
+
+
+
+







        int nWrite = seekAndWrite(pFile, iWrite, "", 1);
        if( nWrite!=1 ) return SQLITE_IOERR_WRITE;
        iWrite += nBlk;
      }
#endif
    }
  }

  if( pFile->mmapLimit>0 ){
    int rc;
    if( pFile->szChunk<=0 ){
      if( robust_ftruncate(pFile->h, nByte) ){
        pFile->lastErrno = errno;
        return unixLogError(SQLITE_IOERR_TRUNCATE, "ftruncate", pFile->zPath);
      }
    }

    rc = unixMapfile(pFile, nByte);
    return rc;
  }

  return SQLITE_OK;
}

/*
** If *pArg is inititially negative then this is a query.  Set *pArg to
** 1 or 0 depending on whether or not bit mask of pFile->ctrlFlags is set.
3631
3632
3633
3634
3635
3636
3637
3638
3639


3640
3641
3642
3643
3644
3645
3646
3681
3682
3683
3684
3685
3686
3687


3688
3689
3690
3691
3692
3693
3694
3695
3696







-
-
+
+







      char *zTFile = sqlite3_malloc( pFile->pVfs->mxPathname );
      if( zTFile ){
        unixGetTempname(pFile->pVfs->mxPathname, zTFile);
        *(char**)pArg = zTFile;
      }
      return SQLITE_OK;
    }
    case SQLITE_FCNTL_GETFD: {
      *(int*)pArg = pFile->h;
    case SQLITE_FCNTL_MMAP_SIZE: {
      pFile->mmapLimit = *(i64*)pArg;
      return SQLITE_OK;
    }
#ifdef SQLITE_DEBUG
    /* The pager calls this method to signal that it has done
    ** a rollback and that the database is therefore unchanged and
    ** it hence it is OK for the transaction change counter to be
    ** unchanged.
4447
4448
4449
4450
4451
4452
4453
4454
4455
4456
4457











4458
4459
4460
4461

4462
4463
4464
4465
4466

4467
4468
4469
4470
4471

4472
4473

4474
4475
4476
4477
4478
4479

4480
4481
4482
4483



4484
4485

4486


4487
4488


4489
4490
4491
4492
4493
4494
4495





4496
4497
4498
4499

4500
4501
4502
4503
4504
4505
4506














4507
4508
4509
4510
4511
4512
4513
4514
4515













4516
4517
4518
4519
4520
4521

4522
4523


4524
4525
4526
4527
4528






4529
4530
4531
4532
4533
4534
4535
4536






4537
4538

4539
4540
4541
4542
4543
4544
4545
4497
4498
4499
4500
4501
4502
4503




4504
4505
4506
4507
4508
4509
4510
4511
4512
4513
4514




4515





4516



4517

4518


4519
4520





4521
4522



4523
4524
4525
4526

4527
4528
4529
4530


4531
4532







4533
4534
4535
4536
4537




4538







4539
4540
4541
4542
4543
4544
4545
4546
4547
4548
4549
4550
4551
4552
4553








4554
4555
4556
4557
4558
4559
4560
4561
4562
4563
4564
4565
4566





4567
4568


4569
4570





4571
4572
4573
4574
4575
4576








4577
4578
4579
4580
4581
4582


4583
4584
4585
4586
4587
4588
4589
4590







-
-
-
-
+
+
+
+
+
+
+
+
+
+
+
-
-
-
-
+
-
-
-
-
-
+
-
-
-

-
+
-
-
+

-
-
-
-
-
+

-
-
-
+
+
+

-
+

+
+
-
-
+
+
-
-
-
-
-
-
-
+
+
+
+
+
-
-
-
-
+
-
-
-
-
-
-
-
+
+
+
+
+
+
+
+
+
+
+
+
+
+

-
-
-
-
-
-
-
-
+
+
+
+
+
+
+
+
+
+
+
+
+
-
-
-
-
-

+
-
-
+
+
-
-
-
-
-
+
+
+
+
+
+
-
-
-
-
-
-
-
-
+
+
+
+
+
+
-
-
+







**
**     ROUNDUP(0,  8) ->  0
**     ROUNDUP(13, 8) -> 16
**     ROUNDUP(32, 8) -> 32
*/
#define ROUNDUP(x,y)     (((x)+y-1)&~(y-1))

/*
** Map, remap or unmap part of the database file.
*/
static int unixMremap(
static void unixUnmapfile(unixFile *pFd){
  assert( pFd->nFetchOut==0 );
  if( pFd->pMapRegion ){
    munmap(pFd->pMapRegion, pFd->mmapOrigsize);
    pFd->pMapRegion = 0;
    pFd->mmapSize = 0;
    pFd->mmapOrigsize = 0;
  }
}

static int unixMapfile(unixFile *pFd, i64 nByte){
  sqlite3_file *fd,               /* Main database file */
  int flags,                      /* Mask of SQLITE_MREMAP_XXX flags */
  sqlite3_int64 iOff,             /* Offset to start mapping at */
  sqlite3_int64 nOld,             /* Size of old mapping, or zero */
  i64 nMap = nByte;
  sqlite3_int64 nNew,             /* Size of new mapping, or zero */
  void **ppMap                    /* IN/OUT: Old/new mappings */
){
  unixFile *p = (unixFile *)fd;   /* The underlying database file */
  int rc = SQLITE_OK;             /* Return code */
  int rc;
  void *pNew = 0;                 /* New mapping */
  i64 nNewRnd;                    /* nNew rounded up */
  i64 nOldRnd;                    /* nOld rounded up */

  assert( iOff==0 );
  assert( nMap>=0 || pFd->nFetchOut==0 );
  /* assert( p->mmapSize==nOld ); */
  assert( p->pMapRegion==0 || p->pMapRegion==(*ppMap) );
  if( pFd->nFetchOut>0 ) return SQLITE_OK;

  /* If the SQLITE_MREMAP_EXTEND flag is set, then the size of the requested 
  ** mapping (nNew bytes) may be greater than the size of the database file.
  ** If this is the case, extend the file on disk using ftruncate().  */
  assert( nNew>0 || (flags & SQLITE_MREMAP_EXTEND)==0 );
  if( flags & SQLITE_MREMAP_EXTEND ){
  if( nMap<0 ){
    struct stat statbuf;          /* Low-level file information */
    rc = osFstat(p->h, &statbuf);
    if( rc==SQLITE_OK && nNew>statbuf.st_size ){
      rc = robust_ftruncate(p->h, nNew);
    rc = osFstat(pFd->h, &statbuf);
    if( rc!=SQLITE_OK ){
      return SQLITE_IOERR_FSTAT;
    }
    if( rc!=SQLITE_OK ) return rc;
    nMap = statbuf.st_size;
  }
  if( nMap>pFd->mmapLimit ){
    nMap = pFd->mmapLimit;

  /* According to some sources, the effect of changing the size of the
  }

  ** underlying file on mapped regions that correspond to the added or
  ** removed pages is undefined. However, there is reason to believe that
  ** on modern platforms like Linux or OSX, things just work. For example,
  ** it is possible to create a mapping larger than the file on disk and
  ** extend the file on disk later on.
  **
  ** Exploit this on Linux and OSX to reduce the number of munmap()/mmap() 
  if( nMap!=pFd->mmapSize ){
    void *pNew;
    unixUnmapfile(pFd);

    if( nMap>0 ){
  ** calls required if the file size is changing. In this case all mappings 
  ** are rounded up to the nearest 4MB. And if a new mapping is requested 
  ** that has the same rounded size as an old mapping, the old mapping can 
  ** be reused as is. */
      void *pNew;
#if defined(__APPLE__) || defined(__linux__)
  nNewRnd = ROUNDUP(nNew, 4096*1024);
  nOldRnd = ROUNDUP(nOld, 4096*1024);
#else
  nNewRnd = ROUNDUP(nNew, 4096*1);
  nOldRnd = ROUNDUP(nOld, 4096*1);
#endif
      int flags = PROT_READ;
      if( (pFd->ctrlFlags & UNIXFILE_RDONLY)==0 ) flags |= PROT_WRITE;
      pNew = mmap(0, ROUNDUP(nMap, 4096), flags, MAP_SHARED, pFd->h, 0);
      if( pNew==MAP_FAILED ){
        return SQLITE_IOERR_MREMAP;
      }

      pFd->pMapRegion = pNew;
      pFd->mmapOrigsize = pFd->mmapSize = nMap;
    }
  }

  return SQLITE_OK;
}

  /* On OSX or Linux, reuse the old mapping if it is the right size. */
#if defined(__APPLE__) || defined(__linux__)
  if( nNewRnd==nOldRnd ){
    VVA_ONLY( p->mmapSize = nNew; )
    return SQLITE_OK;
  }
#endif

static int unixFetch(sqlite3_file *fd, i64 iOff, int nAmt, void **pp){
  unixFile *pFd = (unixFile *)fd;   /* The underlying database file */
  *pp = 0;

  if( pFd->mmapLimit>0 ){
    if( pFd->pMapRegion==0 ){
      int rc = unixMapfile(pFd, -1);
      if( rc!=SQLITE_OK ) return rc;
    }
    if( pFd->mmapSize >= iOff+nAmt ){
      *pp = &((u8 *)pFd->pMapRegion)[iOff];
      pFd->nFetchOut++;
    }
  /* If we get this far, unmap any old mapping. */
  if( nOldRnd!=0 ){
    void *pOld = *ppMap;
    munmap(pOld, nOldRnd);
    VVA_ONLY( p->mmapSize = 0; p->pMapRegion = 0; );
  }
  return SQLITE_OK;

  /* And, if required, use mmap() to create a new mapping. */
}

  if( nNewRnd>0 ){
    int flags = PROT_READ;
    if( (p->ctrlFlags & UNIXFILE_RDONLY)==0 ) flags |= PROT_WRITE;
    pNew = mmap(0, nNewRnd, flags, MAP_SHARED, p->h, iOff);
    if( pNew==MAP_FAILED ){
static int unixUnfetch(sqlite3_file *fd, void *p){
  unixFile *pFd = (unixFile *)fd;   /* The underlying database file */

  assert( (p==0)==(pFd->nFetchOut==0) );

  if( p ){
      pNew = 0;
      VVA_ONLY( p->mmapSize = 0; p->pMapRegion = 0; )
      rc = SQLITE_IOERR_MREMAP;
    }else{
      VVA_ONLY( p->mmapSize = nNew; p->pMapRegion = pNew; )
    }
  }

    pFd->nFetchOut--;
  }else{
    unixUnmapfile(pFd);
  }

  assert( pFd->nFetchOut>=0 );
  *ppMap = pNew;
  return rc;
  return SQLITE_OK;
}

/*
** Here ends the implementation of all sqlite3_file methods.
**
********************** End sqlite3_file Methods *******************************
******************************************************************************/
4593
4594
4595
4596
4597
4598
4599

4600

4601
4602
4603
4604
4605
4606
4607
4638
4639
4640
4641
4642
4643
4644
4645

4646
4647
4648
4649
4650
4651
4652
4653







+
-
+







   unixFileControl,            /* xFileControl */                            \
   unixSectorSize,             /* xSectorSize */                             \
   unixDeviceCharacteristics,  /* xDeviceCapabilities */                     \
   unixShmMap,                 /* xShmMap */                                 \
   unixShmLock,                /* xShmLock */                                \
   unixShmBarrier,             /* xShmBarrier */                             \
   unixShmUnmap,               /* xShmUnmap */                               \
   unixFetch,                  /* xFetch */                                  \
   unixMremap,                 /* xMremap */                                 \
   unixUnfetch,                /* xUnfetch */                                \
};                                                                           \
static const sqlite3_io_methods *FINDER##Impl(const char *z, unixFile *p){   \
  UNUSED_PARAMETER(z); UNUSED_PARAMETER(p);                                  \
  return &METHOD;                                                            \
}                                                                            \
static const sqlite3_io_methods *(*const FINDER)(const char*,unixFile *p)    \
    = FINDER##Impl;
4861
4862
4863
4864
4865
4866
4867
4868
4869
4870
4871
4872
4873
4874
4875
4907
4908
4909
4910
4911
4912
4913

4914
4915
4916
4917
4918
4919
4920







-







  assert( zFilename!=0 || (ctrlFlags & UNIXFILE_NOLOCK)!=0 );

  OSTRACE(("OPEN    %-3d %s\n", h, zFilename));
  pNew->h = h;
  pNew->pVfs = pVfs;
  pNew->zPath = zFilename;
  pNew->ctrlFlags = (u8)ctrlFlags;
  VVA_ONLY( pNew->mmapSize = 0; )
  if( sqlite3_uri_boolean(((ctrlFlags & UNIXFILE_URI) ? zFilename : 0),
                           "psow", SQLITE_POWERSAFE_OVERWRITE) ){
    pNew->ctrlFlags |= UNIXFILE_PSOW;
  }
  if( strcmp(pVfs->zName,"unix-excl")==0 ){
    pNew->ctrlFlags |= UNIXFILE_EXCL;
  }
Changes to src/pager.c.
652
653
654
655
656
657
658
659
660

661
662
663
664
665
666
667
668
669
670
671
672
673
652
653
654
655
656
657
658


659


660
661
662

663
664
665
666
667
668
669







-
-
+
-
-



-







  i64 journalOff;             /* Current write offset in the journal file */
  i64 journalHdr;             /* Byte offset to previous journal header */
  sqlite3_backup *pBackup;    /* Pointer to list of ongoing backup processes */
  PagerSavepoint *aSavepoint; /* Array of active savepoints */
  int nSavepoint;             /* Number of elements in aSavepoint[] */
  char dbFileVers[16];        /* Changes whenever database file changes */

  void *pMap;                 /* Memory mapped prefix of database file */
  i64 nMap;                   /* Size of mapping at pMap in bytes */ 
  u8 bUseFetch;               /* True to use xFetch() */
  i64 nMapValid;              /* Bytes at pMap known to be valid */
  i64 nMapLimit;              /* Maximum permitted mapping size */
  int nMapCfgLimit;           /* Configured limit value */
  int nMmapOut;               /* Number of mmap pages currently outstanding */
  PgHdr *pFree;               /* List of free mmap page headers (pDirty) */
  int bMapResize;             /* Check if the mapping should be resized */
  /*
  ** End of the routinely-changing class members
  ***************************************************************************/

  u16 nExtra;                 /* Add this many bytes to each in-memory page */
  i16 nReserve;               /* Number of unused bytes at end of each page */
  u32 vfsFlags;               /* Flags for sqlite3_vfs.xOpen() */
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2079
2080
2081
2082
2083
2084
2085


















2086
2087
2088
2089
2090
2091
2092







-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-







                           (int)pPager->nReserve);
  }
}
#else
# define pagerReportSize(X)     /* No-op if we do not support a codec */
#endif

/*
** Write nBuf bytes of data from buffer pBuf to offset iOff of the 
** database file. If this part of the database file is memory mapped,
** use memcpy() to do so. Otherwise, call sqlite3OsWrite().
**
** Return SQLITE_OK if successful, or an SQLite error code if an error 
** occurs.
*/
int sqlite3PagerWriteData(Pager *pPager, const void *pBuf, int nBuf, i64 iOff){
  int rc = SQLITE_OK;
  if( pPager->nMapValid>=(iOff+nBuf) ){
    memcpy(&((u8 *)(pPager->pMap))[iOff], pBuf, nBuf);
  }else{
    rc = sqlite3OsWrite(pPager->fd, pBuf, nBuf, iOff);
  }
  return rc;
}

/*
** Read a single page from either the journal file (if isMainJrnl==1) or
** from the sub-journal (if isMainJrnl==0) and playback that page.
** The page begins at offset *pOffset into the file. The *pOffset
** value is increased to the start of the next page in the journal.
**
** The main rollback journal uses checksums - the statement journal does 
2275
2276
2277
2278
2279
2280
2281
2282

2283
2284
2285
2286
2287
2288
2289
2253
2254
2255
2256
2257
2258
2259

2260
2261
2262
2263
2264
2265
2266
2267







-
+







  if( isOpen(pPager->fd)
   && (pPager->eState>=PAGER_WRITER_DBMOD || pPager->eState==PAGER_OPEN)
   && isSynced
  ){
    i64 ofst = (pgno-1)*(i64)pPager->pageSize;
    testcase( !isSavepnt && pPg!=0 && (pPg->flags&PGHDR_NEED_SYNC)!=0 );
    assert( !pagerUseWal(pPager) );
    rc = sqlite3PagerWriteData(pPager, aData, pPager->pageSize, ofst);
    rc = sqlite3OsWrite(pPager->fd, (u8 *)aData, pPager->pageSize, ofst);
    if( pgno>pPager->dbFileSize ){
      pPager->dbFileSize = pgno;
    }
    if( pPager->pBackup ){
      CODEC1(pPager, aData, pgno, 3, rc=SQLITE_NOMEM);
      sqlite3BackupUpdate(pPager->pBackup, pgno, (u8*)aData);
      CODEC2(pPager, aData, pgno, 7, rc=SQLITE_NOMEM, aData);
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2508
2509
2510
2511
2512
2513
2514



2515
2516
2517
2518
2519
2520
2521







-
-
-







    assert( pPager->eLock==EXCLUSIVE_LOCK );
    /* TODO: Is it safe to use Pager.dbFileSize here? */
    rc = sqlite3OsFileSize(pPager->fd, &currentSize);
    newSize = szPage*(i64)nPage;
    if( rc==SQLITE_OK && currentSize!=newSize ){
      if( currentSize>newSize ){
        rc = sqlite3OsTruncate(pPager->fd, newSize);
        if( newSize<pPager->nMapValid ){
          pPager->nMapValid = newSize;
        }
      }else if( (currentSize+szPage)<=newSize ){
        char *pTmp = pPager->pTmpSpace;
        memset(pTmp, 0, szPage);
        testcase( (newSize-szPage) == currentSize );
        testcase( (newSize-szPage) >  currentSize );
        rc = sqlite3OsWrite(pPager->fd, pTmp, szPage, newSize-szPage);
      }
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892



2893
2894
2895
2896
2897
2898
2899
2900
2855
2856
2857
2858
2859
2860
2861






2862
2863
2864

2865
2866
2867
2868
2869
2870
2871







-
-
-
-
-
-
+
+
+
-







  }

  if( iFrame ){
    /* Try to pull the page from the write-ahead log. */
    rc = sqlite3WalReadFrame(pPager->pWal, iFrame, pgsz, pPg->pData);
  }else{
    i64 iOffset = (pgno-1)*(i64)pPager->pageSize;
    if( pPager->pMap && pPager->nMapValid>=iOffset+pPager->pageSize ){
      memcpy(pPg->pData, &((u8 *)(pPager->pMap))[iOffset], pPager->pageSize);
    }else{
      rc = sqlite3OsRead(pPager->fd, pPg->pData, pgsz, iOffset);
      if( rc==SQLITE_IOERR_SHORT_READ ){
        rc = SQLITE_OK;
    rc = sqlite3OsRead(pPager->fd, pPg->pData, pgsz, iOffset);
    if( rc==SQLITE_IOERR_SHORT_READ ){
      rc = SQLITE_OK;
      }
    }
  }

  if( pgno==1 ){
    if( rc ){
      /* If the read is unsuccessful, set the dbFileVers[] to something
      ** that will never be a valid file version.  dbFileVers[] is a copy
3116
3117
3118
3119
3120
3121
3122

3123
3124
3125
3126
3127
3128
3129
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101







+







  ** the duplicate call is harmless.
  */
  sqlite3WalEndReadTransaction(pPager->pWal);

  rc = sqlite3WalBeginReadTransaction(pPager->pWal, &changed);
  if( rc!=SQLITE_OK || changed ){
    pager_reset(pPager);
    if( pPager->bUseFetch ) sqlite3OsUnfetch(pPager->fd, 0);
  }

  return rc;
}
#endif

/*
3378
3379
3380
3381
3382
3383
3384
3385
3386

3387
3388
3389
3390

3391
3392


3393
3394

3395
3396
3397
3398
3399
3400











3401
3402
3403
3404
3405
3406
3407
3350
3351
3352
3353
3354
3355
3356


3357


3358
3359
3360


3361
3362


3363






3364
3365
3366
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
3378
3379
3380
3381







-
-
+
-
-


+
-
-
+
+
-
-
+
-
-
-
-
-
-
+
+
+
+
+
+
+
+
+
+
+







** Change the maximum number of in-memory pages that are allowed.
*/
void sqlite3PagerSetCachesize(Pager *pPager, int mxPage){
  sqlite3PcacheSetCachesize(pPager->pPCache, mxPage);
}

/*
** Set Pager.nMapLimit, the maximum permitted mapping size, based on the
** current values of Pager.nMapCfgLimit and Pager.pageSize.
** Invoke SQLITE_FCNTL_MMAP_SIZE based on the current value of nMapCfgLimit.
**
** If this connection should not use mmap at all, set nMapLimit to zero.
*/
static void pagerFixMaplimit(Pager *pPager){
  sqlite3_file *fd = pPager->fd;
  if( isOpen(pPager->fd)==0 
   || pPager->fd->pMethods->iVersion<3 
  if( isOpen(fd) ){
    pPager->bUseFetch = (fd->pMethods->iVersion>=3) && pPager->nMapCfgLimit!=0;
   || pPager->fd->pMethods->xMremap==0 
   || pPager->tempFile 
    if( pPager->bUseFetch ){
  ){
    pPager->nMapLimit = 0;
  }else if( pPager->nMapCfgLimit<0 ){
    pPager->nMapLimit = (i64)pPager->nMapCfgLimit * -1024;
  }else{
    pPager->nMapLimit = (i64)pPager->nMapCfgLimit * pPager->pageSize;
      void *p;
      i64 nMapLimit;
      if( pPager->nMapCfgLimit<0 ){
        nMapLimit = (i64)pPager->nMapCfgLimit * -1024;
      }else{
        nMapLimit = (i64)pPager->nMapCfgLimit * pPager->pageSize;
      }

      p = (void *)&nMapLimit;
      sqlite3OsFileControlHint(pPager->fd, SQLITE_FCNTL_MMAP_SIZE, p);
    }
  }
}

/*
** Change the maximum size of any memory mapping made of the database file.
*/
void sqlite3PagerSetMmapsize(Pager *pPager, int nMap){
3866
3867
3868
3869
3870
3871
3872
3873
3874
3875
3876
3877
3878
3879
3880
3881
3882
3883
3884
3885
3886
3887
3888
3889
3890
3891
3892
3893
3894
3895
3896
3897
3898
3899
3900
3901
3902
3903
3904
3905
3906
3907
3908
3909
3910
3911
3912
3913
3914
3915
3916
3917
3918
3919
3920
3921
3922
3923
3924
3925
3926


3927
3928
3929
3930
3931
3932
3933






3934
3935
3936
3937
3938
3939
3940
3840
3841
3842
3843
3844
3845
3846
3847


















































3848


3849
3850
3851
3852
3853
3854
3855
3856

3857
3858
3859
3860
3861
3862
3863
3864
3865
3866
3867
3868
3869








-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-

-
-
+
+






-
+
+
+
+
+
+







  }
  if( rc==SQLITE_OK ){
    rc = sqlite3OsFileSize(pPager->jfd, &pPager->journalHdr);
  }
  return rc;
}

/*
** Unmap any memory mapping of the database file.
*/
static int pagerUnmap(Pager *pPager){
  assert( pPager->nMmapOut==0 );
  if( pPager->pMap ){
    sqlite3OsMremap(pPager->fd, 0, 0, pPager->nMap, 0, &pPager->pMap);
    pPager->nMap = 0;
    pPager->nMapValid = 0;
  }
  return SQLITE_OK;
}

/*
** Create, or recreate, the memory mapping of the database file.
*/
static int pagerMap(Pager *pPager, int bExtend){
  int rc = SQLITE_OK;             /* Return code */
  Pgno nPg;                       /* Size of mapping to request in pages */
  i64 sz;                         /* Size of mapping to request in bytes */

  assert( isOpen(pPager->fd) && pPager->tempFile==0 );
  assert( pPager->pMap==0 || pPager->nMap>0 );
  /* assert( pPager->eState>=1 ); */
  assert( pPager->nMmapOut==0 );
  assert( pPager->nMapLimit>0 );

  /* Figure out how large a mapping to request. Set variable sz to this 
  ** value in bytes. */
  nPg = (pPager->eState==1) ? pPager->dbSize : pPager->dbFileSize;
  sz = (i64)nPg * pPager->pageSize;
  if( sz>pPager->nMapLimit ) sz = pPager->nMapLimit;

  if( sz!=pPager->nMapValid ){
    int flags = (bExtend ? SQLITE_MREMAP_EXTEND : 0);
    rc = sqlite3OsMremap(pPager->fd, flags, 0, pPager->nMap, sz, &pPager->pMap);
    if( rc==SQLITE_OK ){
      assert( pPager->pMap!=0 );
      pPager->nMap = sz;
    }else{
      assert( pPager->pMap==0 );
      pPager->nMap = 0;
    }
    pPager->nMapValid = pPager->nMap;
  }
  pPager->bMapResize = 0;

  return rc;
}

/*
** Obtain a reference to a memory mapped page object for page number pgno. 
** The caller must ensure that page pgno lies within the currently mapped 
** region. If successful, set *ppPage to point to the new page reference
** The new object will use the pointer pData, obtained from xFetch().
** If successful, set *ppPage to point to the new page reference
** and return SQLITE_OK. Otherwise, return an SQLite error code and set
** *ppPage to zero.
**
** Page references obtained by calling this function should be released
** by calling pagerReleaseMapPage().
*/
static int pagerAcquireMapPage(Pager *pPager, Pgno pgno, PgHdr **ppPage){
static int pagerAcquireMapPage(
  Pager *pPager,                  /* Pager object */
  Pgno pgno,                      /* Page number */
  void *pData,                    /* xFetch()'d data for this page */
  PgHdr **ppPage                  /* OUT: Acquired page object */
){
  PgHdr *p;                       /* Memory mapped page to return */

  if( pPager->pFree ){
    *ppPage = p = pPager->pFree;
    pPager->pFree = p->pDirty;
    p->pDirty = 0;
    memset(p->pExtra, 0, pPager->nExtra);
3951
3952
3953
3954
3955
3956
3957
3958
3959

3960
3961
3962
3963
3964
3965
3966
3967
3968
3969
3970
3971
3972
3973



3974
3975
3976
3977
3978
3979
3980
3880
3881
3882
3883
3884
3885
3886

3887
3888
3889
3890
3891
3892
3893
3894
3895
3896
3897
3898
3899
3900
3901
3902
3903
3904
3905
3906
3907
3908
3909
3910
3911
3912







-

+














+
+
+








  assert( p->pExtra==(void *)&p[1] );
  assert( p->pPage==0 );
  assert( p->flags==PGHDR_MMAP );
  assert( p->pPager==pPager );
  assert( p->nRef==1 );

  p->pData = &((u8 *)pPager->pMap)[(i64)(pgno-1) * pPager->pageSize];
  p->pgno = pgno;
  p->pData = pData;
  pPager->nMmapOut++;

  return SQLITE_OK;
}

/*
** Release a reference to page pPg. pPg must have been returned by an 
** earlier call to pagerAcquireMapPage().
*/
static void pagerReleaseMapPage(PgHdr *pPg){
  Pager *pPager = pPg->pPager;
  pPager->nMmapOut--;
  pPg->pDirty = pPager->pFree;
  pPager->pFree = pPg;

  assert( pPager->fd->pMethods->iVersion>=3 );
  sqlite3OsUnfetch(pPager->fd, pPg->pData);
}

/*
** Free all PgHdr objects stored in the Pager.pFree list.
*/
static void pagerFreeMapHdrs(Pager *pPager){
  PgHdr *p;
4002
4003
4004
4005
4006
4007
4008
4009
4010
4011
4012
4013
4014
4015
4016
3934
3935
3936
3937
3938
3939
3940

3941
3942
3943
3944
3945
3946
3947







-







*/
int sqlite3PagerClose(Pager *pPager){
  u8 *pTmp = (u8 *)pPager->pTmpSpace;

  assert( assert_pager_state(pPager) );
  disable_simulated_io_errors();
  sqlite3BeginBenignMalloc();
  pagerUnmap(pPager);
  pagerFreeMapHdrs(pPager);
  /* pPager->errCode = 0; */
  pPager->exclusiveMode = 0;
#ifndef SQLITE_OMIT_WAL
  sqlite3WalClose(pPager->pWal, pPager->ckptSyncFlags, pPager->pageSize, pTmp);
  pPager->pWal = 0;
#endif
4212
4213
4214
4215
4216
4217
4218
4219
4220
4221
4222
4223
4224
4225
4226
4227
4228
4229
4230
4231
4232
4233
4234
4235
4236
4237
4238
4239
4240
4241
4242
4243
4244
4245
4246
4247
4248
4249
4250
4251
4252
4253
4254
4255
4256
4257
4258
4259
4260
4261
4262
4263
4264
4265
4143
4144
4145
4146
4147
4148
4149








































4150
4151
4152
4153
4154
4155
4156







-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-







  */
  sqlite3PcacheClearSyncFlags(pPager->pPCache);
  pPager->eState = PAGER_WRITER_DBMOD;
  assert( assert_pager_state(pPager) );
  return SQLITE_OK;
}

/*
** This is called by the wal.c module at the start of a checkpoint. If the
** checkpoint runs to completion, it will set the database file size to
** szReq bytes. This function performs two tasks:
**
**   * If the file is currently less than szReq bytes in size, an
**     xFileControl(SQLITE_FNCTL_SIZE_HINT) is issued to inform the OS
**     layer of the expected file size, and
**
**   * If mmap is being used, then the mapping is extended to szReq
**     bytes in size.
**
** SQLITE_OK is returned if successful, or an error code if an error occurs.
*/
int sqlite3PagerSetFilesize(Pager *pPager, i64 szReq){
  int rc;
  i64 sz;                         /* Size of file on disk in bytes */

  assert( pPager->eState==PAGER_OPEN );
  assert( pPager->nMmapOut==0 );

  rc = sqlite3OsFileSize(pPager->fd, &sz);
  if( rc==SQLITE_OK ){
    if( sz>szReq ){
      sqlite3OsFileControlHint(pPager->fd, SQLITE_FCNTL_SIZE_HINT, &sz);
    }
  }


  if( rc==SQLITE_OK ){
    i64 szMap = (szReq > pPager->nMapLimit) ? pPager->nMapLimit : szReq;
    if( pPager->nMapValid!=pPager->nMap || szMap!=pPager->nMap ){
      pPager->dbFileSize = (szReq / pPager->pageSize);
      rc = pagerMap(pPager, 1);
    }
  }

  return rc;
}

/*
** The argument is the first in a linked list of dirty pages connected
** by the PgHdr.pDirty pointer. This function writes each one of the
** in-memory pages in the list to the database file. The argument may
** be NULL, representing an empty list. In this case this function is
** a no-op.
**
4311
4312
4313
4314
4315
4316
4317
4318
4319
4320
4321
4322
4323
4324
4325
4326
4327
4328
4329
4202
4203
4204
4205
4206
4207
4208





4209
4210
4211
4212
4213
4214
4215







-
-
-
-
-







  assert( rc!=SQLITE_OK || isOpen(pPager->fd) );
  if( rc==SQLITE_OK 
   && (pList->pDirty ? pPager->dbSize : pList->pgno+1)>pPager->dbHintSize 
  ){
    sqlite3_int64 szFile = pPager->pageSize * (sqlite3_int64)pPager->dbSize;
    sqlite3OsFileControlHint(pPager->fd, SQLITE_FCNTL_SIZE_HINT, &szFile);
    pPager->dbHintSize = pPager->dbSize;

    if( pPager->nMmapOut==0 && pPager->nMapLimit>0 ){
      pPager->dbFileSize = pPager->dbSize;
      rc = pagerMap(pPager, 1);
    }
  }

  while( rc==SQLITE_OK && pList ){
    Pgno pgno = pList->pgno;

    /* If there are dirty pages in the page cache with page numbers greater
    ** than Pager.dbSize, this means sqlite3PagerTruncateImage() was called to
4340
4341
4342
4343
4344
4345
4346
4347

4348
4349
4350
4351
4352
4353
4354
4226
4227
4228
4229
4230
4231
4232

4233
4234
4235
4236
4237
4238
4239
4240







-
+







      assert( (pList->flags&PGHDR_NEED_SYNC)==0 );
      if( pList->pgno==1 ) pager_write_changecounter(pList);

      /* Encode the database */
      CODEC2(pPager, pList->pData, pgno, 6, return SQLITE_NOMEM, pData);

      /* Write out the page data. */
      rc = sqlite3PagerWriteData(pPager, pData, pPager->pageSize, offset);
      rc = sqlite3OsWrite(pPager->fd, pData, pPager->pageSize, offset);

      /* If page 1 was just written, update Pager.dbFileVers to match
      ** the value now stored in the database file. If writing this 
      ** page caused the database file to grow, update dbFileSize. 
      */
      if( pgno==1 ){
        memcpy(&pPager->dbFileVers, &pData[24], sizeof(pPager->dbFileVers));
5160
5161
5162
5163
5164
5165
5166
5167

5168
5169
5170
5171
5172
5173
5174
5046
5047
5048
5049
5050
5051
5052

5053
5054
5055
5056
5057
5058
5059
5060







-
+







           || (pPager->exclusiveMode && pPager->eLock>SHARED_LOCK)
      );
    }

    if( !pPager->tempFile && (
        pPager->pBackup 
     || sqlite3PcachePagecount(pPager->pPCache)>0 
     || pPager->pMap
     || pPager->bUseFetch /* TODO: Currently required for xUnfetch(0) only. */
    )){
      /* The shared-lock has just been acquired on the database file
      ** and there are already pages in the cache (from a previous
      ** read or write transaction).  Check to see if the database
      ** has been modified.  If the database has changed, flush the
      ** cache.
      **
5184
5185
5186
5187
5188
5189
5190
5191

5192
5193
5194
5195
5196

5197
5198
5199
5200
5201
5202
5203
5204
5205
5206
5207
5208
5209
5210
5211
5212
5213
5214
5215
5216



5217
5218
5219
5220
5221
5222
5223
5070
5071
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
5097
5098
5099
5100
5101
5102
5103
5104
5105







-
+

-
-
-
-
+
-
















-
-
-
+
+
+







      */
      Pgno nPage = 0;
      char dbFileVers[sizeof(pPager->dbFileVers)];

      rc = pagerPagecount(pPager, &nPage);
      if( rc ) goto failed;

      if( nPage>0 || pPager->pMap ){
      if( nPage>0 ){
        IOTRACE(("CKVERS %p %d\n", pPager, sizeof(dbFileVers)));
        if( pPager->pMap ){
          memcpy(&dbFileVers, &((u8 *)(pPager->pMap))[24], sizeof(dbFileVers));
        }else{
          rc = sqlite3OsRead(pPager->fd, &dbFileVers, sizeof(dbFileVers), 24);
        rc = sqlite3OsRead(pPager->fd, &dbFileVers, sizeof(dbFileVers), 24);
        }
        if( rc!=SQLITE_OK ){
          goto failed;
        }
      }else{
        memset(dbFileVers, 0, sizeof(dbFileVers));
      }

      if( memcmp(pPager->dbFileVers, dbFileVers, sizeof(dbFileVers))!=0 ){
        pager_reset(pPager);

        /* Unmap the database file. It is possible that external processes
        ** may have truncated the database file and then extended it back
        ** to its original size while this process was not holding a lock.
        ** In this case there may exist a Pager.pMap mapping that appears
        ** to be the right size but is not actually valid. Avoid this
        ** possibility by unmapping the db here. */
        pagerUnmap(pPager);
      }else if( pPager->pMap ){
        pPager->bMapResize = 1;
        if( pPager->bUseFetch ){
          sqlite3OsUnfetch(pPager->fd, 0);
        }
      }
    }

    /* If there is a WAL file in the file-system, open this database in WAL
    ** mode. Otherwise, the following function call is a no-op.
    */
    rc = pagerOpenWalIfPresent(pPager);
5321
5322
5323
5324
5325
5326
5327
5328

5329
5330
5331
5332
5333
5334
5335
5203
5204
5205
5206
5207
5208
5209

5210
5211
5212
5213
5214
5215
5216
5217







-
+







  u32 iFrame = 0;                 /* Frame to read from WAL file */
  const int noContent = (flags & PAGER_ACQUIRE_NOCONTENT);

  /* It is acceptable to use a read-only (mmap) page for any page except
  ** page 1 if there is no write-transaction open or the ACQUIRE_READONLY
  ** flag was specified by the caller. And so long as the db is not a 
  ** temporary or in-memory database.  */
  const int bMmapOk = (pPager->nMapLimit>0 && pgno!=1
  const int bMmapOk = (pgno!=1 && pPager->bUseFetch
   && (pPager->eState==PAGER_READER || (flags & PAGER_ACQUIRE_READONLY))
  );

  assert( pPager->eState>=PAGER_READER );
  assert( assert_pager_state(pPager) );
  assert( noContent==0 || bMmapOk==0 );

5345
5346
5347
5348
5349
5350
5351

5352
5353
5354
5355






5356
5357
5358
5359
5360



5361
5362
5363
5364
5365
5366
5367
5227
5228
5229
5230
5231
5232
5233
5234




5235
5236
5237
5238
5239
5240
5241
5242
5243
5244

5245
5246
5247
5248
5249
5250
5251
5252
5253
5254







+
-
-
-
-
+
+
+
+
+
+




-
+
+
+








    if( bMmapOk && pagerUseWal(pPager) ){
      rc = sqlite3WalFindFrame(pPager->pWal, pgno, &iFrame);
      if( rc!=SQLITE_OK ) goto pager_acquire_err;
    }

    if( iFrame==0 && bMmapOk ){
      void *pData = 0;
      if( pPager->pMap==0 || (pPager->bMapResize && pPager->nMmapOut==0) ){
        rc = pagerMap(pPager, 0);
      }
      if( rc==SQLITE_OK && pPager->nMap>=((i64)pgno * pPager->pageSize) ){

      rc = sqlite3OsFetch(pPager->fd, 
          (i64)(pgno-1) * pPager->pageSize, pPager->pageSize, &pData
      );

      if( rc==SQLITE_OK && pData ){
        if( pPager->eState>PAGER_READER ){
          (void)sqlite3PcacheFetch(pPager->pPCache, pgno, 0, &pPg);
        }
        if( pPg==0 ){
          rc = pagerAcquireMapPage(pPager, pgno, &pPg);
          rc = pagerAcquireMapPage(pPager, pgno, pData, &pPg);
        }else{
          sqlite3OsUnfetch(pPager->fd, pData);
        }
        if( pPg ){
          assert( rc==SQLITE_OK );
          *ppPage = pPg;
          return SQLITE_OK;
        }
      }
7113
7114
7115
7116
7117
7118
7119
7120

7121
7122
7123
7124
7125
7126
7127
7000
7001
7002
7003
7004
7005
7006

7007
7008
7009
7010
7011
7012
7013
7014







-
+







    rc = pagerExclusiveLock(pPager);
  }

  /* Open the connection to the log file. If this operation fails, 
  ** (e.g. due to malloc() failure), return an error code.
  */
  if( rc==SQLITE_OK ){
    rc = sqlite3WalOpen(pPager->pVfs, pPager,
    rc = sqlite3WalOpen(pPager->pVfs,
        pPager->fd, pPager->zWal, pPager->exclusiveMode,
        pPager->journalSizeLimit, &pPager->pWal
    );
  }
  pagerFixMaplimit(pPager);

  return rc;
7151
7152
7153
7154
7155
7156
7157
7158
7159
7160
7161
7162
7163
7164
7165
7166
7038
7039
7040
7041
7042
7043
7044


7045
7046
7047
7048
7049
7050
7051







-
-








  assert( assert_pager_state(pPager) );
  assert( pPager->eState==PAGER_OPEN   || pbOpen );
  assert( pPager->eState==PAGER_READER || !pbOpen );
  assert( pbOpen==0 || *pbOpen==0 );
  assert( pbOpen!=0 || (!pPager->tempFile && !pPager->pWal) );

  pagerUnmap(pPager);

  if( !pPager->tempFile && !pPager->pWal ){
    if( !sqlite3PagerWalSupported(pPager) ) return SQLITE_CANTOPEN;

    /* Close any rollback journal previously open */
    sqlite3OsClose(pPager->jfd);

    rc = pagerOpenWal(pPager);
Changes to src/pager.h.
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
172
173
174
175
176
177
178



179
180
181
182
183
184
185







-
-
-







void sqlite3PagerClearCache(Pager *);
int sqlite3SectorSize(sqlite3_file *);

/* Functions used to truncate the database file. */
void sqlite3PagerTruncateImage(Pager*,Pgno);
int sqlite3PagerSetFilesize(Pager *, i64);

/* Write data to the database file */
int sqlite3PagerWriteData(Pager *pPager, const void *pBuf, int nBuf, i64 iOff);

#if defined(SQLITE_HAS_CODEC) && !defined(SQLITE_OMIT_WAL)
void *sqlite3PagerCodec(DbPage *);
#endif

/* Functions to support testing and debugging. */
#if !defined(NDEBUG) || defined(SQLITE_TEST)
  Pgno sqlite3PagerPagenumber(DbPage*);
Changes to src/sqlite.h.in.
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
728
729
730
731
732
733
734
703
704
705
706
707
708
709


















710
711
712
713
714
715
716







-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-







** to xWrite().
**
** If xRead() returns SQLITE_IOERR_SHORT_READ it must also fill
** in the unread portions of the buffer with zeros.  A VFS that
** fails to zero-fill short reads might seem to work.  However,
** failure to zero-fill short reads will eventually lead to
** database corruption.
**
** Assuming parameter nNew is non-zero, the xMremap method should attempt
** to memory map a region nNew bytes in size starting at offset iOffset
** of the file.  If successful, it should set *ppMap to point to the
** mapping and return SQLITE_OK. If the file is opened for read-write
** access, then the mapping should also be read-write.
**
** If nOld is non-zero, then the initial value of *ppMap points to a
** mapping returned by a previous call to xMremap. The existing mapping
** is nOld bytes in size and starts at offset iOffset of the file. In
** this case the xMremap method is expected to unmap the existing mapping
** and overwrite *ppMap with the pointer to the new mapping. If nOld is
** zero, then the initial value of *ppMap is undefined.
**
** If nNew is zero, then no new mapping should be created. Any old
** mapping must still be unmapped if nOld is non-zero. If the nOld
** parameter is non-zero, then the existing mapping is always unmapped -
** even if an error occurs.
*/
typedef struct sqlite3_io_methods sqlite3_io_methods;
struct sqlite3_io_methods {
  int iVersion;
  int (*xClose)(sqlite3_file*);
  int (*xRead)(sqlite3_file*, void*, int iAmt, sqlite3_int64 iOfst);
  int (*xWrite)(sqlite3_file*, const void*, int iAmt, sqlite3_int64 iOfst);
743
744
745
746
747
748
749
750
751


752
753
754
755
756
757
758
759
760
761
762
763
764
725
726
727
728
729
730
731


732
733
734
735
736
737


738
739
740
741
742
743
744







-
-
+
+




-
-







  int (*xDeviceCharacteristics)(sqlite3_file*);
  /* Methods above are valid for version 1 */
  int (*xShmMap)(sqlite3_file*, int iPg, int pgsz, int, void volatile**);
  int (*xShmLock)(sqlite3_file*, int offset, int n, int flags);
  void (*xShmBarrier)(sqlite3_file*);
  int (*xShmUnmap)(sqlite3_file*, int deleteFlag);
  /* Methods above are valid for version 2 */
  int (*xMremap)(sqlite3_file *fd, int flags,
      sqlite3_int64 iOff, sqlite3_int64 nOld, sqlite3_int64 nNew, void **ppMap);
  int (*xFetch)(sqlite3_file*, sqlite3_int64 iOfst, int iAmt, void **pp);
  int (*xUnfetch)(sqlite3_file*, void *p);
  /* Methods above are valid for version 3 */
  /* Additional methods may be added in future releases */
};

#define SQLITE_MREMAP_EXTEND  0x0001        /* xMremap call may extend file */

/*
** CAPI3REF: Standard File Control Opcodes
**
** These integer constants are opcodes for the xFileControl method
** of the [sqlite3_io_methods] object and for the [sqlite3_file_control()]
** interface.
**
902
903
904
905
906
907
908




909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927

928
929
930
931
932
933
934
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910

911
912
913
914
915
916
917
918







+
+
+
+


















-
+







** ^Application can invoke this file-control to have SQLite generate a
** temporary filename using the same algorithm that is followed to generate
** temporary filenames for TEMP tables and other internal uses.  The
** argument should be a char** which will be filled with the filename
** written into memory obtained from [sqlite3_malloc()].  The caller should
** invoke [sqlite3_free()] on the result to avoid a memory leak.
**
** <li>[[SQLITE_FCNTL_MMAP_SIZE]]
** The argument is assumed to point to a value of type sqlite3_int64. An
** advisory maximum amount of this file to memory map in bytes.
**
** </ul>
*/
#define SQLITE_FCNTL_LOCKSTATE               1
#define SQLITE_GET_LOCKPROXYFILE             2
#define SQLITE_SET_LOCKPROXYFILE             3
#define SQLITE_LAST_ERRNO                    4
#define SQLITE_FCNTL_SIZE_HINT               5
#define SQLITE_FCNTL_CHUNK_SIZE              6
#define SQLITE_FCNTL_FILE_POINTER            7
#define SQLITE_FCNTL_SYNC_OMITTED            8
#define SQLITE_FCNTL_WIN32_AV_RETRY          9
#define SQLITE_FCNTL_PERSIST_WAL            10
#define SQLITE_FCNTL_OVERWRITE              11
#define SQLITE_FCNTL_VFSNAME                12
#define SQLITE_FCNTL_POWERSAFE_OVERWRITE    13
#define SQLITE_FCNTL_PRAGMA                 14
#define SQLITE_FCNTL_BUSYHANDLER            15
#define SQLITE_FCNTL_TEMPFILENAME           16
#define SQLITE_FCNTL_GETFD                  17
#define SQLITE_FCNTL_MMAP_SIZE              18

/*
** CAPI3REF: Mutex Handle
**
** The mutex module within SQLite defines [sqlite3_mutex] to be an
** abstract type for a mutex object.  The SQLite core never looks
** at the internal representation of an [sqlite3_mutex].  It only
Changes to src/wal.c.
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
408
409
410
411
412
413
414

415
416
417
418
419
420
421







-







** 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 */
  Pager *pPager;             /* Pager object */
  u32 iCallback;             /* Value to pass to log callback (or 0) */
  i64 mxWalSize;             /* Truncate WAL to this size upon reset */
  int nWiData;               /* Size of array apWiData */
  int szFirstBlock;          /* Size of first block written to WAL file */
  volatile u32 **apWiData;   /* Pointer to wal-index content in memory */
  u32 szPage;                /* Database page size */
  i16 readLock;              /* Which read lock is being held.  -1 for none */
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1247
1248
1249
1250
1251
1252
1253

1254
1255
1256
1257
1258
1259
1260







-







**
** 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.
*/
int sqlite3WalOpen(
  sqlite3_vfs *pVfs,              /* vfs module to open wal and wal-index */
  Pager *pPager,                  /* Pager object handle */
  sqlite3_file *pDbFd,            /* The open database file */
  const char *zWalName,           /* Name of the WAL file */
  int bNoShm,                     /* True to run in heap-memory mode */
  i64 mxWalSize,                  /* Truncate WAL to this size on reset */
  Wal **ppWal                     /* OUT: Allocated Wal handle */
){
  int rc;                         /* Return Code */
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1287
1288
1289
1290
1291
1292
1293

1294
1295
1296
1297
1298
1299
1300







-







  pRet->pWalFd = (sqlite3_file *)&pRet[1];
  pRet->pDbFd = pDbFd;
  pRet->readLock = -1;
  pRet->mxWalSize = mxWalSize;
  pRet->zWalName = zWalName;
  pRet->syncHeader = 1;
  pRet->padToSectorBoundary = 1;
  pRet->pPager = pPager;
  pRet->exclusiveMode = (bNoShm ? WAL_HEAPMEMORY_MODE: WAL_NORMAL_MODE);

  /* Open file handle on the write-ahead log file. */
  flags = (SQLITE_OPEN_READWRITE|SQLITE_OPEN_CREATE|SQLITE_OPEN_WAL);
  rc = sqlite3OsOpen(pVfs, zWalName, pRet->pWalFd, flags, &flags);
  if( rc==SQLITE_OK && flags&SQLITE_OPEN_READONLY ){
    pRet->readOnly = WAL_RDONLY;
1721
1722
1723
1724
1725
1726
1727
1728
1729


1730
1731

1732
1733
1734
1735
1736
1737
1738
1739






1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752

1753
1754
1755
1756
1757
1758
1759
1718
1719
1720
1721
1722
1723
1724


1725
1726


1727




1728
1729


1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747

1748
1749
1750
1751
1752
1753
1754
1755







-
-
+
+
-
-
+
-
-
-
-


-
-
+
+
+
+
+
+












-
+







    u32 nBackfill = pInfo->nBackfill;

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

    /* If the database file is currently smaller than mxPage pages in size,
    ** the call below issues an SQLITE_FCNTL_SIZE_HINT to the OS layer to
    /* If the database may grow as a result of this checkpoint, hint
    ** about the eventual size of the db file to the VFS layer.
    ** inform it that it is likely to grow to that size.
    **
    */
    ** Additionally, if the pager is using mmap(), then the call to 
    ** SetFilesize() guarantees that the mapping is not larger than mxPage
    ** pages. This makes the sqlite3OsTruncate() call below safe - no pages
    ** that are part of the mapped region will be truncated away.  */
    if( rc==SQLITE_OK ){
      i64 nReq = ((i64)mxPage * szPage);
      rc = sqlite3PagerSetFilesize(pWal->pPager, nReq);
    }
      rc = sqlite3OsFileSize(pWal->pDbFd, &nSize);
      if( rc==SQLITE_OK && nSize<nReq ){
        sqlite3OsFileControlHint(pWal->pDbFd, SQLITE_FCNTL_SIZE_HINT, &nReq);
      }
    }


    /* Iterate through the contents of the WAL, copying data to the db file. */
    while( rc==SQLITE_OK && 0==walIteratorNext(pIter, &iDbpage, &iFrame) ){
      i64 iOffset;
      assert( walFramePgno(pWal, iFrame)==iDbpage );
      if( iFrame<=nBackfill || iFrame>mxSafeFrame || iDbpage>mxPage ) continue;
      iOffset = walFrameOffset(iFrame, szPage) + WAL_FRAME_HDRSIZE;
      /* testcase( IS_BIG_INT(iOffset) ); // requires a 4GiB WAL file */
      rc = sqlite3OsRead(pWal->pWalFd, zBuf, szPage, iOffset);
      if( rc!=SQLITE_OK ) break;
      iOffset = (iDbpage-1)*(i64)szPage;
      testcase( IS_BIG_INT(iOffset) );
      rc = sqlite3PagerWriteData(pWal->pPager, zBuf, szPage, iOffset);
      rc = sqlite3OsWrite(pWal->pDbFd, zBuf, szPage, iOffset);
      if( rc!=SQLITE_OK ) break;
    }

    /* If work was actually accomplished... */
    if( rc==SQLITE_OK ){
      if( mxSafeFrame==walIndexHdr(pWal)->mxFrame ){
        i64 szDb = pWal->hdr.nPage*(i64)szPage;
Changes to src/wal.h.
49
50
51
52
53
54
55
56
57

58
59
60
61
62
63
64
49
50
51
52
53
54
55


56
57
58
59
60
61
62
63







-
-
+








/* 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*, Pager *, sqlite3_file*, const char *, int, i64, Wal**);
int sqlite3WalOpen(sqlite3_vfs*, sqlite3_file*, const char *, int, i64, Wal**);
int sqlite3WalClose(Wal *pWal, int sync_flags, int, u8 *);

/* Set the limiting size of a WAL file. */
void sqlite3WalLimit(Wal*, i64);

/* 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