SQLite

Check-in [af9503daf3]
Login

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

Overview
Comment:Modify the crash-recovery test code in test6.c for 3.5. Also change some other code to use the new sqlite3_io_methods interface. Lots of things are broken now. (CVS 4228)
Downloads: Tarball | ZIP archive
Timelines: family | ancestors | descendants | both | trunk
Files: files | file ages | folders
SHA1: af9503daf3f7703fcddad754bc1dc9e179830b6e
User & Date: danielk1977 2007-08-15 17:08:46.000
Context
2007-08-15
19:16
Test infrastructure for the new memory subsystem. (CVS 4229) (check-in: 9e50665672 user: drh tags: trunk)
17:08
Modify the crash-recovery test code in test6.c for 3.5. Also change some other code to use the new sqlite3_io_methods interface. Lots of things are broken now. (CVS 4228) (check-in: af9503daf3 user: danielk1977 tags: trunk)
17:07
Add a debugging memory allocator. (CVS 4227) (check-in: 8d2d1c4ff9 user: drh tags: trunk)
Changes
Unified Diff Ignore Whitespace Patch
Changes to src/os.c.
16
17
18
19
20
21
22
23
24
25
26
27

28
29
30
31
32

33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70



71
72
73
74
75
76

77
78
79

80
81
82
83
84
85
86
#define _SQLITE_OS_C_ 1
#include "sqliteInt.h"
#include "os.h"
#undef _SQLITE_OS_C_

/*
** The following routines are convenience wrappers around methods
** of the OsFile object.  This is mostly just syntactic sugar.  All
** of this would be completely automatic if SQLite were coded using
** C++ instead of plain old C.
*/
int sqlite3OsClose(OsFile **pId){

  OsFile *id;
  if( pId!=0 && (id = *pId)!=0 ){
    return id->pMethod->xClose(pId);
  }else{
    return SQLITE_OK;

  }
}
int sqlite3OsOpenDirectory(OsFile *id, const char *zName){
  return id->pMethod->xOpenDirectory(id, zName);
}
int sqlite3OsRead(OsFile *id, void *pBuf, int amt){
  return id->pMethod->xRead(id, pBuf, amt);
}
int sqlite3OsWrite(OsFile *id, const void *pBuf, int amt){
  return id->pMethod->xWrite(id, pBuf, amt);
}
int sqlite3OsSeek(OsFile *id, i64 offset){
  return id->pMethod->xSeek(id, offset);
}
int sqlite3OsTruncate(OsFile *id, i64 size){
  return id->pMethod->xTruncate(id, size);
}
int sqlite3OsSync(OsFile *id, int fullsync){
  return id->pMethod->xSync(id, fullsync);
}
void sqlite3OsSetFullSync(OsFile *id, int value){
  id->pMethod->xSetFullSync(id, value);
}
int sqlite3OsFileSize(OsFile *id, i64 *pSize){
  return id->pMethod->xFileSize(id, pSize);
}
int sqlite3OsLock(OsFile *id, int lockType){
  return id->pMethod->xLock(id, lockType);
}
int sqlite3OsUnlock(OsFile *id, int lockType){
  return id->pMethod->xUnlock(id, lockType);
}
int sqlite3OsCheckReservedLock(OsFile *id){
  return id->pMethod->xCheckReservedLock(id);
}
int sqlite3OsSectorSize(OsFile *id){
  int (*xSectorSize)(OsFile*) = id->pMethod->xSectorSize;
  return xSectorSize ? xSectorSize(id) : SQLITE_DEFAULT_SECTOR_SIZE;



}

#if defined(SQLITE_TEST) || defined(SQLITE_DEBUG)
  /* These methods are currently only used for testing and debugging. */
  int sqlite3OsFileHandle(OsFile *id){
    return id->pMethod->xFileHandle(id);

  }
  int sqlite3OsLockState(OsFile *id){
    return id->pMethod->xLockState(id);

  }
#endif

#ifdef SQLITE_ENABLE_REDEF_IO
/*
** A function to return a pointer to the virtual function table.
** This routine really does not accomplish very much since the







|



|
>
|

|
<
|
>
|
|
<
|

|
|

|
|

|
|

|
|

|
|

<
<
<
|
|

|
|

|
|

|
|

|
|

>
>
>




|
|
>

|
|
>







16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

32
33
34
35

36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52



53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
#define _SQLITE_OS_C_ 1
#include "sqliteInt.h"
#include "os.h"
#undef _SQLITE_OS_C_

/*
** The following routines are convenience wrappers around methods
** of the sqlite3_file object.  This is mostly just syntactic sugar. All
** of this would be completely automatic if SQLite were coded using
** C++ instead of plain old C.
*/
int sqlite3OsClose(sqlite3_file **pId){
  int rc = SQLITE_OK;
  sqlite3_file *id;
  if( pId!=0 && (id = *pId)!=0 ){
    rc = id->pMethods->xClose(id);

    if( rc==SQLITE_OK ){
      *pId = 0;
    }
  }

  return rc;
}
int sqlite3OsRead(sqlite3_file *id, void *pBuf, int amt, i64 offset){
  return id->pMethods->xRead(id, pBuf, amt, offset);
}
int sqlite3OsWrite(sqlite3_file *id, const void *pBuf, int amt, i64 offset){
  return id->pMethods->xWrite(id, pBuf, amt, offset);
}
int sqlite3OsTruncate(sqlite3_file *id, i64 size){
  return id->pMethods->xTruncate(id, size);
}
int sqlite3OsSync(sqlite3_file *id, int fullsync){
  return id->pMethods->xSync(id, fullsync);
}
int sqlite3OsFileSize(sqlite3_file *id, i64 *pSize){
  return id->pMethods->xFileSize(id, pSize);
}



int sqlite3OsLock(sqlite3_file *id, int lockType){
  return id->pMethods->xLock(id, lockType);
}
int sqlite3OsUnlock(sqlite3_file *id, int lockType){
  return id->pMethods->xUnlock(id, lockType);
}
int sqlite3OsBreakLock(sqlite3_file *id){
  return id->pMethods->xBreakLock(id);
}
int sqlite3OsCheckReservedLock(sqlite3_file *id){
  return id->pMethods->xCheckReservedLock(id);
}
int sqlite3OsSectorSize(sqlite3_file *id){
  int (*xSectorSize)(sqlite3_file*) = id->pMethods->xSectorSize;
  return xSectorSize ? xSectorSize(id) : SQLITE_DEFAULT_SECTOR_SIZE;
}
int sqlite3OsDeviceCharacteristics(sqlite3_file *id){
  return id->pMethods->xDeviceCharacteristics(id);
}

#if defined(SQLITE_TEST) || defined(SQLITE_DEBUG)
  /* These methods are currently only used for testing and debugging. */
  int sqlite3OsFileHandle(sqlite3_file *id){
    /* return id->pMethods->xFileHandle(id); */
    return 0;
  }
  int sqlite3OsLockState(sqlite3_file *id){
    /* return id->pMethods->xLockState(id); */
    return 0;
  }
#endif

#ifdef SQLITE_ENABLE_REDEF_IO
/*
** A function to return a pointer to the virtual function table.
** This routine really does not accomplish very much since the
Changes to src/os.h.
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361

362



363
364
365

366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
#define RESERVED_BYTE     (PENDING_BYTE+1)
#define SHARED_FIRST      (PENDING_BYTE+2)
#define SHARED_SIZE       510

/*
** Prototypes for operating system interface routines.
*/
int sqlite3OsClose(OsFile**);
int sqlite3OsOpenDirectory(OsFile*, const char*);
int sqlite3OsRead(OsFile*, void*, int amt);
int sqlite3OsWrite(OsFile*, const void*, int amt);
int sqlite3OsSeek(OsFile*, i64 offset);
int sqlite3OsTruncate(OsFile*, i64 size);
int sqlite3OsSync(OsFile*, int);
void sqlite3OsSetFullSync(OsFile *id, int setting);
int sqlite3OsFileSize(OsFile*, i64 *pSize);
int sqlite3OsLock(OsFile*, int);
int sqlite3OsUnlock(OsFile*, int);

int sqlite3OsCheckReservedLock(OsFile *id);



int sqlite3OsOpenReadWrite(const char*, OsFile**, int*);
int sqlite3OsOpenExclusive(const char*, OsFile**, int);
int sqlite3OsOpenReadOnly(const char*, OsFile**);

int sqlite3OsDelete(const char*);
int sqlite3OsFileExists(const char*);
char *sqlite3OsFullPathname(const char*);
int sqlite3OsIsDirWritable(char*);
int sqlite3OsSyncDirectory(const char*);
int sqlite3OsSectorSize(OsFile *id);
int sqlite3OsTempFileName(char*);
int sqlite3OsRandomSeed(char*);
int sqlite3OsSleep(int ms);
int sqlite3OsCurrentTime(double*);
void sqlite3OsEnterMutex(void);
void sqlite3OsLeaveMutex(void);
int sqlite3OsInMutex(int);
ThreadData *sqlite3OsThreadSpecificData(int);
void *sqlite3OsMalloc(int);
void *sqlite3OsRealloc(void *, int);
void sqlite3OsFree(void *);
int sqlite3OsAllocationSize(void *);
void *sqlite3OsDlopen(const char*);
void *sqlite3OsDlsym(void*, const char*);
int sqlite3OsDlclose(void*);

#if defined(SQLITE_TEST) || defined(SQLITE_DEBUG)
  int sqlite3OsFileHandle(OsFile *id);
  int sqlite3OsLockState(OsFile *id);
#endif

/*
** If the SQLITE_ENABLE_REDEF_IO macro is defined, then the OS-layer
** interface routines are not called directly but are invoked using
** pointers to functions.  This allows the implementation of various
** OS-layer interface routines to be modified at run-time.  There are







|
<
|
|
<
|
|
<
|
|
|
>
|
>
>
>
|
|
|
>





<

















|
|







344
345
346
347
348
349
350
351

352
353

354
355

356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372

373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
#define RESERVED_BYTE     (PENDING_BYTE+1)
#define SHARED_FIRST      (PENDING_BYTE+2)
#define SHARED_SIZE       510

/*
** Prototypes for operating system interface routines.
*/
int sqlite3OsClose(sqlite3_file**);

int sqlite3OsRead(sqlite3_file*, void*, int amt, i64 offset);
int sqlite3OsWrite(sqlite3_file*, const void*, int amt, i64 offset);

int sqlite3OsTruncate(sqlite3_file*, i64 size);
int sqlite3OsSync(sqlite3_file*, int);

int sqlite3OsFileSize(sqlite3_file*, i64 *pSize);
int sqlite3OsLock(sqlite3_file*, int);
int sqlite3OsUnlock(sqlite3_file*, int);
int sqlite3OsBreakLock(sqlite3_file*);
int sqlite3OsCheckReservedLock(sqlite3_file *id);
int sqlite3OsSectorSize(sqlite3_file *id);
int sqlite3OsDeviceCharacteristics(sqlite3_file *id);

int sqlite3OsOpenReadWrite(const char*, sqlite3_file**, int*);
int sqlite3OsOpenExclusive(const char*, sqlite3_file**, int);
int sqlite3OsOpenReadOnly(const char*, sqlite3_file**);

int sqlite3OsDelete(const char*);
int sqlite3OsFileExists(const char*);
char *sqlite3OsFullPathname(const char*);
int sqlite3OsIsDirWritable(char*);
int sqlite3OsSyncDirectory(const char*);

int sqlite3OsTempFileName(char*);
int sqlite3OsRandomSeed(char*);
int sqlite3OsSleep(int ms);
int sqlite3OsCurrentTime(double*);
void sqlite3OsEnterMutex(void);
void sqlite3OsLeaveMutex(void);
int sqlite3OsInMutex(int);
ThreadData *sqlite3OsThreadSpecificData(int);
void *sqlite3OsMalloc(int);
void *sqlite3OsRealloc(void *, int);
void sqlite3OsFree(void *);
int sqlite3OsAllocationSize(void *);
void *sqlite3OsDlopen(const char*);
void *sqlite3OsDlsym(void*, const char*);
int sqlite3OsDlclose(void*);

#if defined(SQLITE_TEST) || defined(SQLITE_DEBUG)
  int sqlite3OsFileHandle(sqlite3_file *id);
  int sqlite3OsLockState(sqlite3_file *id);
#endif

/*
** If the SQLITE_ENABLE_REDEF_IO macro is defined, then the OS-layer
** interface routines are not called directly but are invoked using
** pointers to functions.  This allows the implementation of various
** OS-layer interface routines to be modified at run-time.  There are
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
** When redefinable I/O is enabled, a single global instance of the
** following structure holds pointers to the routines that SQLite 
** uses to talk with the underlying operating system.  Modify this
** structure (before using any SQLite API!) to accomodate perculiar
** operating system interfaces or behaviors.
*/
struct sqlite3OsVtbl {
  int (*xOpenReadWrite)(const char*, OsFile**, int*);
  int (*xOpenExclusive)(const char*, OsFile**, int);
  int (*xOpenReadOnly)(const char*, OsFile**);

  int (*xDelete)(const char*);
  int (*xFileExists)(const char*);
  char *(*xFullPathname)(const char*);
  int (*xIsDirWritable)(char*);
  int (*xSyncDirectory)(const char*);
  int (*xTempFileName)(char*);







|
|
|







406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
** When redefinable I/O is enabled, a single global instance of the
** following structure holds pointers to the routines that SQLite 
** uses to talk with the underlying operating system.  Modify this
** structure (before using any SQLite API!) to accomodate perculiar
** operating system interfaces or behaviors.
*/
struct sqlite3OsVtbl {
  int (*xOpenReadWrite)(const char*, sqlite3_file**, int*);
  int (*xOpenExclusive)(const char*, sqlite3_file**, int);
  int (*xOpenReadOnly)(const char*, sqlite3_file**);

  int (*xDelete)(const char*);
  int (*xFileExists)(const char*);
  char *(*xFullPathname)(const char*);
  int (*xIsDirWritable)(char*);
  int (*xSyncDirectory)(const char*);
  int (*xTempFileName)(char*);
Changes to src/os_unix.c.
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100

101
102
103
104
105
106
107
108
109
110
111
112
113




114
115
116
117
118
119
120
121
122

/*
** The unixFile structure is subclass of OsFile specific for the unix
** protability layer.
*/
typedef struct unixFile unixFile;
struct unixFile {
  IoMethod const *pMethod;  /* Always the first entry */
  struct openCnt *pOpen;    /* Info about all open fd's on this inode */
  struct lockInfo *pLock;   /* Info about locks on this inode */
#ifdef SQLITE_ENABLE_LOCKING_STYLE
  void *lockingContext;     /* Locking style specific state */
#endif /* SQLITE_ENABLE_LOCKING_STYLE */
  int h;                    /* The file descriptor */
  unsigned char locktype;   /* The type of lock held on this fd */
  unsigned char isOpen;     /* True if needs to be closed */
  unsigned char fullSync;   /* Use F_FULLSYNC if available */
  int dirfd;                /* File descriptor for the directory */
  i64 offset;               /* Seek offset */
#ifdef SQLITE_UNIX_THREADS
  pthread_t tid;            /* The thread that "owns" this OsFile */
#endif
};


/*
** Provide the ability to override some OS-layer functions during
** testing.  This is used to simulate OS crashes to verify that 
** commits are atomic even in the event of an OS crash.
*/
#ifdef SQLITE_CRASH_TEST
  extern int sqlite3CrashTestEnable;
  extern int sqlite3CrashOpenReadWrite(const char*, OsFile**, int*);
  extern int sqlite3CrashOpenExclusive(const char*, OsFile**, int);
  extern int sqlite3CrashOpenReadOnly(const char*, OsFile**, int);
# define CRASH_TEST_OVERRIDE(X,A,B,C) \
    if(sqlite3CrashTestEnable){ return X(A,B,C); }




#else
# define CRASH_TEST_OVERRIDE(X,A,B,C)  /* no-op */
#endif


/*
** Include code that is common to all os_*.c files
*/
#include "os_common.h"







|















>








|
<
<
|
|
>
>
>
>

|







78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110


111
112
113
114
115
116
117
118
119
120
121
122
123
124
125

/*
** The unixFile structure is subclass of OsFile specific for the unix
** protability layer.
*/
typedef struct unixFile unixFile;
struct unixFile {
  sqlite3_io_methods const *pMethod;  /* Always the first entry */
  struct openCnt *pOpen;    /* Info about all open fd's on this inode */
  struct lockInfo *pLock;   /* Info about locks on this inode */
#ifdef SQLITE_ENABLE_LOCKING_STYLE
  void *lockingContext;     /* Locking style specific state */
#endif /* SQLITE_ENABLE_LOCKING_STYLE */
  int h;                    /* The file descriptor */
  unsigned char locktype;   /* The type of lock held on this fd */
  unsigned char isOpen;     /* True if needs to be closed */
  unsigned char fullSync;   /* Use F_FULLSYNC if available */
  int dirfd;                /* File descriptor for the directory */
  i64 offset;               /* Seek offset */
#ifdef SQLITE_UNIX_THREADS
  pthread_t tid;            /* The thread that "owns" this OsFile */
#endif
};


/*
** Provide the ability to override some OS-layer functions during
** testing.  This is used to simulate OS crashes to verify that 
** commits are atomic even in the event of an OS crash.
*/
#ifdef SQLITE_CRASH_TEST
  extern int sqlite3CrashTestEnable;
  int sqlite3CrashFileWrap(sqlite3_file *, const char *, sqlite3_file **);


  static int CRASH_TEST_OVERRIDE(const char *zFile, sqlite3_file **pId, int rc){
    if( rc==SQLITE_OK && sqlite3CrashTestEnable ){
      rc = sqlite3CrashFileWrap(*pId, zFile, pId); 
    }
    return rc;
  }
#else
# define CRASH_TEST_OVERRIDE(A,B,C) C
#endif


/*
** Include code that is common to all os_*.c files
*/
#include "os_common.h"
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846


847

848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865




866
867
868
869
870
871
872
873
874
875

876

877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894

895

896
897
898
899
900
901
902
int sqlite3UnixFileExists(const char *zFilename){
  return access(zFilename, 0)==0;
}

/* Forward declaration */
static int allocateUnixFile(
  int h,                    /* File descriptor of the open file */
  OsFile **pId,             /* Write the real file descriptor here */
  const char *zFilename,    /* Name of the file being opened */
  int delFlag               /* If true, make sure the file deletes on close */
);

/*
** Attempt to open a file for both reading and writing.  If that
** fails, try opening it read-only.  If the file does not exist,
** try to create it.
**
** On success, a handle for the open file is written to *id
** and *pReadonly is set to 0 if the file was opened for reading and
** writing or 1 if the file was opened read-only.  The function returns
** SQLITE_OK.
**
** On failure, the function returns SQLITE_CANTOPEN and leaves
** *id and *pReadonly unchanged.
*/
int sqlite3UnixOpenReadWrite(
  const char *zFilename,
  OsFile **pId,
  int *pReadonly
){
  int h;
  
  CRASH_TEST_OVERRIDE(sqlite3CrashOpenReadWrite, zFilename, pId, pReadonly);
  assert( 0==*pId );
  h = open(zFilename, O_RDWR|O_CREAT|O_LARGEFILE|O_BINARY,
                        SQLITE_DEFAULT_FILE_PERMISSIONS);
  if( h<0 ){
#ifdef EISDIR
    if( errno==EISDIR ){
      return SQLITE_CANTOPEN;
    }
#endif
    h = open(zFilename, O_RDONLY|O_LARGEFILE|O_BINARY);
    if( h<0 ){
      return SQLITE_CANTOPEN; 
    }
    *pReadonly = 1;
  }else{
    *pReadonly = 0;
  }


  return allocateUnixFile(h, pId, zFilename, 0);

}


/*
** Attempt to open a new file for exclusive access by this process.
** The file will be opened for both reading and writing.  To avoid
** a potential security problem, we do not allow the file to have
** previously existed.  Nor do we allow the file to be a symbolic
** link.
**
** If delFlag is true, then make arrangements to automatically delete
** the file when it is closed.
**
** On success, write the file handle into *id and return SQLITE_OK.
**
** On failure, return SQLITE_CANTOPEN.
*/
int sqlite3UnixOpenExclusive(const char *zFilename, OsFile **pId, int delFlag){




  int h;

  CRASH_TEST_OVERRIDE(sqlite3CrashOpenExclusive, zFilename, pId, delFlag);
  assert( 0==*pId );
  h = open(zFilename,
                O_RDWR|O_CREAT|O_EXCL|O_NOFOLLOW|O_LARGEFILE|O_BINARY,
                delFlag ? 0600 : SQLITE_DEFAULT_FILE_PERMISSIONS);
  if( h<0 ){
    return SQLITE_CANTOPEN;
  }

  return allocateUnixFile(h, pId, zFilename, delFlag);

}

/*
** Attempt to open a new file for read-only access.
**
** On success, write the file handle into *id and return SQLITE_OK.
**
** On failure, return SQLITE_CANTOPEN.
*/
int sqlite3UnixOpenReadOnly(const char *zFilename, OsFile **pId){
  int h;
  
  CRASH_TEST_OVERRIDE(sqlite3CrashOpenReadOnly, zFilename, pId, 0);
  assert( 0==*pId );
  h = open(zFilename, O_RDONLY|O_LARGEFILE|O_BINARY);
  if( h<0 ){
    return SQLITE_CANTOPEN;
  }

  return allocateUnixFile(h, pId, zFilename, 0);

}

/*
** Attempt to open a file descriptor for the directory that contains a
** file.  This file descriptor can be used to fsync() the directory
** in order to make sure the creation of a new file is actually written
** to disk.







|



















|




<

















>
>
|
>

















|
>
>
>
>


<







>
|
>









|


<





>
|
>







800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831

832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876

877
878
879
880
881
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
int sqlite3UnixFileExists(const char *zFilename){
  return access(zFilename, 0)==0;
}

/* Forward declaration */
static int allocateUnixFile(
  int h,                    /* File descriptor of the open file */
  sqlite3_file **pId,       /* Write the real file descriptor here */
  const char *zFilename,    /* Name of the file being opened */
  int delFlag               /* If true, make sure the file deletes on close */
);

/*
** Attempt to open a file for both reading and writing.  If that
** fails, try opening it read-only.  If the file does not exist,
** try to create it.
**
** On success, a handle for the open file is written to *id
** and *pReadonly is set to 0 if the file was opened for reading and
** writing or 1 if the file was opened read-only.  The function returns
** SQLITE_OK.
**
** On failure, the function returns SQLITE_CANTOPEN and leaves
** *id and *pReadonly unchanged.
*/
int sqlite3UnixOpenReadWrite(
  const char *zFilename,
  sqlite3_file **pId,
  int *pReadonly
){
  int h;
  

  assert( 0==*pId );
  h = open(zFilename, O_RDWR|O_CREAT|O_LARGEFILE|O_BINARY,
                        SQLITE_DEFAULT_FILE_PERMISSIONS);
  if( h<0 ){
#ifdef EISDIR
    if( errno==EISDIR ){
      return SQLITE_CANTOPEN;
    }
#endif
    h = open(zFilename, O_RDONLY|O_LARGEFILE|O_BINARY);
    if( h<0 ){
      return SQLITE_CANTOPEN; 
    }
    *pReadonly = 1;
  }else{
    *pReadonly = 0;
  }

  return CRASH_TEST_OVERRIDE(
    zFilename, pId, allocateUnixFile(h, pId, zFilename, 0)
  );
}


/*
** Attempt to open a new file for exclusive access by this process.
** The file will be opened for both reading and writing.  To avoid
** a potential security problem, we do not allow the file to have
** previously existed.  Nor do we allow the file to be a symbolic
** link.
**
** If delFlag is true, then make arrangements to automatically delete
** the file when it is closed.
**
** On success, write the file handle into *id and return SQLITE_OK.
**
** On failure, return SQLITE_CANTOPEN.
*/
int sqlite3UnixOpenExclusive(
  const char *zFilename, 
  sqlite3_file **pId, 
  int delFlag
){
  int h;


  assert( 0==*pId );
  h = open(zFilename,
                O_RDWR|O_CREAT|O_EXCL|O_NOFOLLOW|O_LARGEFILE|O_BINARY,
                delFlag ? 0600 : SQLITE_DEFAULT_FILE_PERMISSIONS);
  if( h<0 ){
    return SQLITE_CANTOPEN;
  }
  return CRASH_TEST_OVERRIDE(
    zFilename, pId, allocateUnixFile(h, pId, zFilename, delFlag)
  );
}

/*
** Attempt to open a new file for read-only access.
**
** On success, write the file handle into *id and return SQLITE_OK.
**
** On failure, return SQLITE_CANTOPEN.
*/
int sqlite3UnixOpenReadOnly(const char *zFilename, sqlite3_file **pId){
  int h;
  

  assert( 0==*pId );
  h = open(zFilename, O_RDONLY|O_LARGEFILE|O_BINARY);
  if( h<0 ){
    return SQLITE_CANTOPEN;
  }
  return CRASH_TEST_OVERRIDE(
    zFilename, pId, allocateUnixFile(h, pId, zFilename, 0)
  );
}

/*
** Attempt to open a file descriptor for the directory that contains a
** file.  This file descriptor can be used to fsync() the directory
** in order to make sure the creation of a new file is actually written
** to disk.
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028





1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074





1075
1076
1077
1078
1079

1080
1081
1082
1083
1084
1085
1086
  return 1;
}

/*
** Seek to the offset in id->offset then read cnt bytes into pBuf.
** Return the number of bytes actually read.  Update the offset.
*/
static int seekAndRead(unixFile *id, void *pBuf, int cnt){
  int got;
  i64 newOffset;
  TIMER_START;
#if defined(USE_PREAD)
  got = pread(id->h, pBuf, cnt, id->offset);
  SimulateIOError( got = -1 );
#elif defined(USE_PREAD64)
  got = pread64(id->h, pBuf, cnt, id->offset);
  SimulateIOError( got = -1 );
#else
  newOffset = lseek(id->h, id->offset, SEEK_SET);
  SimulateIOError( newOffset-- );
  if( newOffset!=id->offset ){
    return -1;
  }
  got = read(id->h, pBuf, cnt);
#endif
  TIMER_END;
  OSTRACE5("READ    %-3d %5d %7lld %d\n", id->h, got, id->offset, TIMER_ELAPSED);
  if( got>0 ){
    id->offset += got;
  }
  return got;
}

/*
** Read data from a file into a buffer.  Return SQLITE_OK if all
** bytes were read successfully and SQLITE_IOERR if anything goes
** wrong.
*/
static int unixRead(OsFile *id, void *pBuf, int amt){





  int got;
  assert( id );
  got = seekAndRead((unixFile*)id, pBuf, amt);
  if( got==amt ){
    return SQLITE_OK;
  }else if( got<0 ){
    return SQLITE_IOERR_READ;
  }else{
    memset(&((char*)pBuf)[got], 0, amt-got);
    return SQLITE_IOERR_SHORT_READ;
  }
}

/*
** Seek to the offset in id->offset then read cnt bytes into pBuf.
** Return the number of bytes actually read.  Update the offset.
*/
static int seekAndWrite(unixFile *id, const void *pBuf, int cnt){
  int got;
  i64 newOffset;
  TIMER_START;
#if defined(USE_PREAD)
  got = pwrite(id->h, pBuf, cnt, id->offset);
#elif defined(USE_PREAD64)
  got = pwrite64(id->h, pBuf, cnt, id->offset);
#else
  newOffset = lseek(id->h, id->offset, SEEK_SET);
  if( newOffset!=id->offset ){
    return -1;
  }
  got = write(id->h, pBuf, cnt);
#endif
  TIMER_END;
  OSTRACE5("WRITE   %-3d %5d %7lld %d\n", id->h, got, id->offset, TIMER_ELAPSED);
  if( got>0 ){
    id->offset += got;
  }
  return got;
}


/*
** Write data from a buffer into a file.  Return SQLITE_OK on success
** or some other error code on failure.
*/
static int unixWrite(OsFile *id, const void *pBuf, int amt){





  int wrote = 0;
  assert( id );
  assert( amt>0 );
  while( amt>0 && (wrote = seekAndWrite((unixFile*)id, pBuf, amt))>0 ){
    amt -= wrote;

    pBuf = &((char*)pBuf)[wrote];
  }
  SimulateIOError(( wrote=(-1), amt=1 ));
  SimulateDiskfullError(( wrote=0, amt=1 ));
  if( amt>0 ){
    if( wrote<0 ){
      return SQLITE_IOERR_WRITE;







|




|


|


|

|






<
<
<








|
>
>
>
>
>


|














|




|

|

|
|





|
<
<
<








|
>
>
>
>
>



|

>







1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027



1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075



1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
  return 1;
}

/*
** Seek to the offset in id->offset then read cnt bytes into pBuf.
** Return the number of bytes actually read.  Update the offset.
*/
static int seekAndRead(unixFile *id, sqlite3_int64 offset, void *pBuf, int cnt){
  int got;
  i64 newOffset;
  TIMER_START;
#if defined(USE_PREAD)
  got = pread(id->h, pBuf, cnt, offset);
  SimulateIOError( got = -1 );
#elif defined(USE_PREAD64)
  got = pread64(id->h, pBuf, cnt, offset);
  SimulateIOError( got = -1 );
#else
  newOffset = lseek(id->h, offset, SEEK_SET);
  SimulateIOError( newOffset-- );
  if( newOffset!=offset ){
    return -1;
  }
  got = read(id->h, pBuf, cnt);
#endif
  TIMER_END;
  OSTRACE5("READ    %-3d %5d %7lld %d\n", id->h, got, id->offset, TIMER_ELAPSED);



  return got;
}

/*
** Read data from a file into a buffer.  Return SQLITE_OK if all
** bytes were read successfully and SQLITE_IOERR if anything goes
** wrong.
*/
static int unixRead(
  sqlite3_file *id, 
  void *pBuf, 
  int amt,
  sqlite3_int64 offset
){
  int got;
  assert( id );
  got = seekAndRead((unixFile*)id, offset, pBuf, amt);
  if( got==amt ){
    return SQLITE_OK;
  }else if( got<0 ){
    return SQLITE_IOERR_READ;
  }else{
    memset(&((char*)pBuf)[got], 0, amt-got);
    return SQLITE_IOERR_SHORT_READ;
  }
}

/*
** Seek to the offset in id->offset then read cnt bytes into pBuf.
** Return the number of bytes actually read.  Update the offset.
*/
static int seekAndWrite(unixFile *id, i64 offset, const void *pBuf, int cnt){
  int got;
  i64 newOffset;
  TIMER_START;
#if defined(USE_PREAD)
  got = pwrite(id->h, pBuf, cnt, offset);
#elif defined(USE_PREAD64)
  got = pwrite64(id->h, pBuf, cnt, offset);
#else
  newOffset = lseek(id->h, offset, SEEK_SET);
  if( newOffset!=offset ){
    return -1;
  }
  got = write(id->h, pBuf, cnt);
#endif
  TIMER_END;
  OSTRACE5("WRITE   %-3d %5d %7lld %d\n", id->h, got, offset, TIMER_ELAPSED);



  return got;
}


/*
** Write data from a buffer into a file.  Return SQLITE_OK on success
** or some other error code on failure.
*/
static int unixWrite(
  sqlite3_file *id, 
  const void *pBuf, 
  int amt,
  sqlite3_int64 offset 
){
  int wrote = 0;
  assert( id );
  assert( amt>0 );
  while( amt>0 && (wrote = seekAndWrite((unixFile*)id, offset, pBuf, amt))>0 ){
    amt -= wrote;
    offset += wrote;
    pBuf = &((char*)pBuf)[wrote];
  }
  SimulateIOError(( wrote=(-1), amt=1 ));
  SimulateDiskfullError(( wrote=0, amt=1 ));
  if( amt>0 ){
    if( wrote<0 ){
      return SQLITE_IOERR_WRITE;
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
** has been created by fsync-ing the directory that contains the file.
** If we do not do this and we encounter a power failure, the directory
** entry for the journal might not exist after we reboot.  The next
** SQLite to access the file will not know that the journal exists (because
** the directory entry for the journal was never created) and the transaction
** will not roll back - possibly leading to database corruption.
*/
static int unixSync(OsFile *id, int dataOnly){
  int rc;
  unixFile *pFile = (unixFile*)id;
  assert( pFile );
  OSTRACE2("SYNC    %-3d\n", pFile->h);
  rc = full_fsync(pFile->h, pFile->fullSync, dataOnly);
  SimulateIOError( rc=1 );
  if( rc ){







|







1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
** has been created by fsync-ing the directory that contains the file.
** If we do not do this and we encounter a power failure, the directory
** entry for the journal might not exist after we reboot.  The next
** SQLite to access the file will not know that the journal exists (because
** the directory entry for the journal was never created) and the transaction
** will not roll back - possibly leading to database corruption.
*/
static int unixSync(sqlite3_file *id, int dataOnly){
  int rc;
  unixFile *pFile = (unixFile*)id;
  assert( pFile );
  OSTRACE2("SYNC    %-3d\n", pFile->h);
  rc = full_fsync(pFile->h, pFile->fullSync, dataOnly);
  SimulateIOError( rc=1 );
  if( rc ){
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
  }
#endif
}

/*
** Truncate an open file to a specified size
*/
static int unixTruncate(OsFile *id, i64 nByte){
  int rc;
  assert( id );
  rc = ftruncate(((unixFile*)id)->h, (off_t)nByte);
  SimulateIOError( rc=1 );
  if( rc ){
    return SQLITE_IOERR_TRUNCATE;
  }else{
    return SQLITE_OK;
  }
}

/*
** Determine the current size of a file in bytes
*/
static int unixFileSize(OsFile *id, i64 *pSize){
  int rc;
  struct stat buf;
  assert( id );
  rc = fstat(((unixFile*)id)->h, &buf);
  SimulateIOError( rc=1 );
  if( rc!=0 ){
    return SQLITE_IOERR_FSTAT;
  }
  *pSize = buf.st_size;
  return SQLITE_OK;
}

/*
** This routine checks if there is a RESERVED lock held on the specified
** file by this or any other process. If such a lock is held, return
** non-zero.  If the file is unlocked or holds only SHARED locks, then
** return zero.
*/
static int unixCheckReservedLock(OsFile *id){
  int r = 0;
  unixFile *pFile = (unixFile*)id;

  assert( pFile );
  sqlite3OsEnterMutex(); /* Because pFile->pLock is shared across threads */

  /* Check if a thread in this process holds such a lock */







|














|


















|







1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
  }
#endif
}

/*
** Truncate an open file to a specified size
*/
static int unixTruncate(sqlite3_file *id, i64 nByte){
  int rc;
  assert( id );
  rc = ftruncate(((unixFile*)id)->h, (off_t)nByte);
  SimulateIOError( rc=1 );
  if( rc ){
    return SQLITE_IOERR_TRUNCATE;
  }else{
    return SQLITE_OK;
  }
}

/*
** Determine the current size of a file in bytes
*/
static int unixFileSize(sqlite3_file *id, i64 *pSize){
  int rc;
  struct stat buf;
  assert( id );
  rc = fstat(((unixFile*)id)->h, &buf);
  SimulateIOError( rc=1 );
  if( rc!=0 ){
    return SQLITE_IOERR_FSTAT;
  }
  *pSize = buf.st_size;
  return SQLITE_OK;
}

/*
** This routine checks if there is a RESERVED lock held on the specified
** file by this or any other process. If such a lock is held, return
** non-zero.  If the file is unlocked or holds only SHARED locks, then
** return zero.
*/
static int unixCheckReservedLock(sqlite3_file *id){
  int r = 0;
  unixFile *pFile = (unixFile*)id;

  assert( pFile );
  sqlite3OsEnterMutex(); /* Because pFile->pLock is shared across threads */

  /* Check if a thread in this process holds such a lock */
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
**    SHARED -> (PENDING) -> EXCLUSIVE
**    RESERVED -> (PENDING) -> EXCLUSIVE
**    PENDING -> EXCLUSIVE
**
** This routine will only increase a lock.  Use the sqlite3OsUnlock()
** routine to lower a locking level.
*/
static int unixLock(OsFile *id, int locktype){
  /* The following describes the implementation of the various locks and
  ** lock transitions in terms of the POSIX advisory shared and exclusive
  ** lock primitives (called read-locks and write-locks below, to avoid
  ** confusion with SQLite lock names). The algorithms are complicated
  ** slightly in order to be compatible with windows systems simultaneously
  ** accessing the same database file, in case that is ever required.
  **







|







1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
**    SHARED -> (PENDING) -> EXCLUSIVE
**    RESERVED -> (PENDING) -> EXCLUSIVE
**    PENDING -> EXCLUSIVE
**
** This routine will only increase a lock.  Use the sqlite3OsUnlock()
** routine to lower a locking level.
*/
static int unixLock(sqlite3_file *id, int locktype){
  /* The following describes the implementation of the various locks and
  ** lock transitions in terms of the POSIX advisory shared and exclusive
  ** lock primitives (called read-locks and write-locks below, to avoid
  ** confusion with SQLite lock names). The algorithms are complicated
  ** slightly in order to be compatible with windows systems simultaneously
  ** accessing the same database file, in case that is ever required.
  **
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
/*
** Lower the locking level on file descriptor pFile to locktype.  locktype
** must be either NO_LOCK or SHARED_LOCK.
**
** 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(OsFile *id, int locktype){
  struct lockInfo *pLock;
  struct flock lock;
  int rc = SQLITE_OK;
  unixFile *pFile = (unixFile*)id;

  assert( pFile );
  OSTRACE7("UNLOCK  %d %d was %d(%d,%d) pid=%d\n", pFile->h, locktype,







|







1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
/*
** Lower the locking level on file descriptor pFile to locktype.  locktype
** must be either NO_LOCK or SHARED_LOCK.
**
** 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 locktype){
  struct lockInfo *pLock;
  struct flock lock;
  int rc = SQLITE_OK;
  unixFile *pFile = (unixFile*)id;

  assert( pFile );
  OSTRACE7("UNLOCK  %d %d was %d(%d,%d) pid=%d\n", pFile->h, locktype,
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
  pFile->locktype = locktype;
  return rc;
}

/*
** Close a file.
*/
static int unixClose(OsFile **pId){
  unixFile *id = (unixFile*)*pId;

  if( !id ) return SQLITE_OK;
  unixUnlock(*pId, NO_LOCK);
  if( id->dirfd>=0 ) close(id->dirfd);
  id->dirfd = -1;
  sqlite3OsEnterMutex();

  if( id->pOpen->nLock ){
    /* If there are outstanding locks, do not actually close the file just
    ** yet because that would clear those locks.  Instead, add the file
    ** descriptor to pOpen->aPending.  It will be automatically closed when
    ** the last lock is cleared.
    */
    int *aNew;
    struct openCnt *pOpen = id->pOpen;
    aNew = realloc( pOpen->aPending, (pOpen->nPending+1)*sizeof(int) );
    if( aNew==0 ){
      /* If a malloc fails, just leak the file descriptor */
    }else{
      pOpen->aPending = aNew;
      pOpen->aPending[pOpen->nPending] = id->h;
      pOpen->nPending++;
    }
  }else{
    /* There are no outstanding locks so we can close the file immediately */
    close(id->h);
  }
  releaseLockInfo(id->pLock);
  releaseOpenCnt(id->pOpen);

  sqlite3OsLeaveMutex();
  id->isOpen = 0;
  OSTRACE2("CLOSE   %-3d\n", id->h);
  OpenCounter(-1);
  sqlite3ThreadSafeFree(id);
  *pId = 0;
  return SQLITE_OK;
}


#ifdef SQLITE_ENABLE_LOCKING_STYLE
#pragma mark AFP Support








|
|
<
|
|
|
|


|






|





|




|

|
|


|
|


<







1662
1663
1664
1665
1666
1667
1668
1669
1670

1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704

1705
1706
1707
1708
1709
1710
1711
  pFile->locktype = locktype;
  return rc;
}

/*
** Close a file.
*/
static int unixClose(sqlite3_file *id){
  unixFile *pFile = (unixFile *)id;

  if( !pFile ) return SQLITE_OK;
  unixUnlock(id, NO_LOCK);
  if( pFile->dirfd>=0 ) close(pFile->dirfd);
  pFile->dirfd = -1;
  sqlite3OsEnterMutex();

  if( pFile->pOpen->nLock ){
    /* If there are outstanding locks, do not actually close the file just
    ** yet because that would clear those locks.  Instead, add the file
    ** descriptor to pOpen->aPending.  It will be automatically closed when
    ** the last lock is cleared.
    */
    int *aNew;
    struct openCnt *pOpen = pFile->pOpen;
    aNew = realloc( pOpen->aPending, (pOpen->nPending+1)*sizeof(int) );
    if( aNew==0 ){
      /* If a malloc fails, just leak the file descriptor */
    }else{
      pOpen->aPending = aNew;
      pOpen->aPending[pOpen->nPending] = pFile->h;
      pOpen->nPending++;
    }
  }else{
    /* There are no outstanding locks so we can close the file immediately */
    close(pFile->h);
  }
  releaseLockInfo(pFile->pLock);
  releaseOpenCnt(pFile->pOpen);

  sqlite3OsLeaveMutex();
  pFile->isOpen = 0;
  OSTRACE2("CLOSE   %-3d\n", pFile->h);
  OpenCounter(-1);
  sqlite3ThreadSafeFree(id);

  return SQLITE_OK;
}


#ifdef SQLITE_ENABLE_LOCKING_STYLE
#pragma mark AFP Support

2344
2345
2346
2347
2348
2349
2350
2351
2352
2353









2354
2355
2356
2357
2358
2359

2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373

2374

2375
2376
2377
2378
2379
2380
2381
** larger for some devices.
**
** SQLite code assumes this function cannot fail. It also assumes that
** if two files are created in the same file-system directory (i.e.
** a database and it's journal file) that the sector size will be the
** same for both.
*/
static int unixSectorSize(OsFile *id){
  return SQLITE_DEFAULT_SECTOR_SIZE;
}










/*
** This vector defines all the methods that can operate on an OsFile
** for unix.
*/
static const IoMethod sqlite3UnixIoMethod = {

  unixClose,
  unixOpenDirectory,
  unixRead,
  unixWrite,
  unixSeek,
  unixTruncate,
  unixSync,
  unixSetFullSync,
  unixFileHandle,
  unixFileSize,
  unixLock,
  unixUnlock,
  unixLockState,
  unixCheckReservedLock,

  unixSectorSize,

};

#ifdef SQLITE_ENABLE_LOCKING_STYLE
/*
 ** This vector defines all the methods that can operate on an OsFile
 ** for unix with AFP style file locking.
 */







|


>
>
>
>
>
>
>
>
>





|
>

<


<


<
<



<

>

>







2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384

2385
2386

2387
2388


2389
2390
2391

2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
** larger for some devices.
**
** SQLite code assumes this function cannot fail. It also assumes that
** if two files are created in the same file-system directory (i.e.
** a database and it's journal file) that the sector size will be the
** same for both.
*/
static int unixSectorSize(sqlite3_file *id){
  return SQLITE_DEFAULT_SECTOR_SIZE;
}

static int unixDeviceCharacteristics(sqlite3_file *id){
  return 0;
}

static int unixBreakLock(sqlite3_file *id){
  assert(!"TODO: unixBreakLock()");
  return 0;
}

/*
** This vector defines all the methods that can operate on an OsFile
** for unix.
*/
static const sqlite3_io_methods sqlite3UnixIoMethod = {
  1,                        /* iVersion */
  unixClose,

  unixRead,
  unixWrite,

  unixTruncate,
  unixSync,


  unixFileSize,
  unixLock,
  unixUnlock,

  unixCheckReservedLock,
  unixBreakLock,
  unixSectorSize,
  unixDeviceCharacteristics
};

#ifdef SQLITE_ENABLE_LOCKING_STYLE
/*
 ** This vector defines all the methods that can operate on an OsFile
 ** for unix with AFP style file locking.
 */
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
    OpenCounter(+1);
    return SQLITE_OK;
  }
}
#else /* SQLITE_ENABLE_LOCKING_STYLE */
static int allocateUnixFile(
  int h,                 /* Open file descriptor on file being opened */
  OsFile **pId,          /* Write the resul unixFile structure here */
  const char *zFilename, /* Name of the file being opened */
  int delFlag            /* If true, delete the file on or before closing */
){
  unixFile *pNew;
  unixFile f;
  int rc;








|







2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
    OpenCounter(+1);
    return SQLITE_OK;
  }
}
#else /* SQLITE_ENABLE_LOCKING_STYLE */
static int allocateUnixFile(
  int h,                 /* Open file descriptor on file being opened */
  sqlite3_file **pId,    /* Write the resul unixFile structure here */
  const char *zFilename, /* Name of the file being opened */
  int delFlag            /* If true, delete the file on or before closing */
){
  unixFile *pNew;
  unixFile f;
  int rc;

2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
    releaseOpenCnt(f.pOpen);
    sqlite3OsLeaveMutex();
    *pId = 0;
    return SQLITE_NOMEM;
  }else{
    *pNew = f;
    pNew->pMethod = &sqlite3UnixIoMethod;
    *pId = (OsFile*)pNew;
    OpenCounter(+1);
    return SQLITE_OK;
  }
}
#endif /* SQLITE_ENABLE_LOCKING_STYLE */

#endif /* SQLITE_OMIT_DISKIO */







|







2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
    releaseOpenCnt(f.pOpen);
    sqlite3OsLeaveMutex();
    *pId = 0;
    return SQLITE_NOMEM;
  }else{
    *pNew = f;
    pNew->pMethod = &sqlite3UnixIoMethod;
    *pId = (sqlite3_file*)pNew;
    OpenCounter(+1);
    return SQLITE_OK;
  }
}
#endif /* SQLITE_ENABLE_LOCKING_STYLE */

#endif /* SQLITE_OMIT_DISKIO */
Changes to src/pager.c.
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
** The pager is used to access a database disk file.  It implements
** atomic commit and rollback through the use of a journal file that
** is separate from the database file.  The pager also implements file
** locking to prevent two processes from writing the same database
** file simultaneously, or one process from reading the database while
** another is writing.
**
** @(#) $Id: pager.c,v 1.355 2007/08/11 00:26:21 drh Exp $
*/
#ifndef SQLITE_OMIT_DISKIO
#include "sqliteInt.h"
#include "os.h"
#include "pager.h"
#include <assert.h>
#include <string.h>







|







14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
** The pager is used to access a database disk file.  It implements
** atomic commit and rollback through the use of a journal file that
** is separate from the database file.  The pager also implements file
** locking to prevent two processes from writing the same database
** file simultaneously, or one process from reading the database while
** another is writing.
**
** @(#) $Id: pager.c,v 1.356 2007/08/15 17:08:46 danielk1977 Exp $
*/
#ifndef SQLITE_OMIT_DISKIO
#include "sqliteInt.h"
#include "os.h"
#include "pager.h"
#include <assert.h>
#include <string.h>
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
#endif

/*
** The following two macros are used within the PAGERTRACEX() macros above
** to print out file-descriptors. 
**
** PAGERID() takes a pointer to a Pager struct as it's argument. The
** associated file-descriptor is returned. FILEHANDLEID() takes an OsFile
** struct as it's argument.
*/
#define PAGERID(p) ((int)(p->fd))
#define FILEHANDLEID(fd) ((int)fd)

/*
** The page cache as a whole is always in one of the following







|







46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
#endif

/*
** The following two macros are used within the PAGERTRACEX() macros above
** to print out file-descriptors. 
**
** PAGERID() takes a pointer to a Pager struct as it's argument. The
** associated file-descriptor is returned. FILEHANDLEID() takes an sqlite3_file
** struct as it's argument.
*/
#define PAGERID(p) ((int)(p->fd))
#define FILEHANDLEID(fd) ((int)fd)

/*
** The page cache as a whole is always in one of the following
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
  int mxPage;                 /* Maximum number of pages to hold in cache */
  Pgno mxPgno;                /* Maximum allowed size of the database */
  u8 *aInJournal;             /* One bit for each page in the database file */
  u8 *aInStmt;                /* One bit for each page in the database */
  char *zFilename;            /* Name of the database file */
  char *zJournal;             /* Name of the journal file */
  char *zDirectory;           /* Directory hold database and journal files */
  OsFile *fd, *jfd;           /* File descriptors for database and journal */
  OsFile *stfd;               /* File descriptor for the statement subjournal*/
  BusyHandler *pBusyHandler;  /* Pointer to sqlite.busyHandler */
  PgHdr *pFirst, *pLast;      /* List of free pages */
  PgHdr *pFirstSynced;        /* First free page with PgHdr.needSync==0 */
  PgHdr *pAll;                /* List of all pages */
  PgHdr *pStmt;               /* List of pages in the statement subjournal */
  PgHdr *pDirty;              /* List of all dirty pages */
  i64 journalOff;             /* Current byte offset in the journal file */







|
|







332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
  int mxPage;                 /* Maximum number of pages to hold in cache */
  Pgno mxPgno;                /* Maximum allowed size of the database */
  u8 *aInJournal;             /* One bit for each page in the database file */
  u8 *aInStmt;                /* One bit for each page in the database */
  char *zFilename;            /* Name of the database file */
  char *zJournal;             /* Name of the journal file */
  char *zDirectory;           /* Directory hold database and journal files */
  sqlite3_file *fd, *jfd;     /* File descriptors for database and journal */
  sqlite3_file *stfd;         /* File descriptor for the statement subjournal*/
  BusyHandler *pBusyHandler;  /* Pointer to sqlite.busyHandler */
  PgHdr *pFirst, *pLast;      /* List of free pages */
  PgHdr *pFirstSynced;        /* First free page with PgHdr.needSync==0 */
  PgHdr *pAll;                /* List of all pages */
  PgHdr *pStmt;               /* List of pages in the statement subjournal */
  PgHdr *pDirty;              /* List of all dirty pages */
  i64 journalOff;             /* Current byte offset in the journal file */
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
/*
** Read a 32-bit integer from the given file descriptor.  Store the integer
** that is read in *pRes.  Return SQLITE_OK if everything worked, or an
** error code is something goes wrong.
**
** All values are stored on disk as big-endian.
*/
static int read32bits(OsFile *fd, u32 *pRes){
  unsigned char ac[4];
  int rc = sqlite3OsRead(fd, ac, sizeof(ac));
  if( rc==SQLITE_OK ){
    *pRes = sqlite3Get4byte(ac);
  }
  return rc;
}

/*
** Write a 32-bit integer into a string buffer in big-endian byte order.
*/
#define put32bits(A,B)  sqlite3Put4byte((u8*)A,B)

/*
** Write a 32-bit integer into the given file descriptor.  Return SQLITE_OK
** on success or an error code is something goes wrong.
*/
static int write32bits(OsFile *fd, u32 val){
  char ac[4];
  put32bits(ac, val);
  return sqlite3OsWrite(fd, ac, 4);
}

/*
** This function should be called when an error occurs within the pager
** code. The first argument is a pointer to the pager structure, the
** second the error-code about to be returned by a pager API function. 
** The value returned is a copy of the second argument to this function. 







|

|















|


|







522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
/*
** Read a 32-bit integer from the given file descriptor.  Store the integer
** that is read in *pRes.  Return SQLITE_OK if everything worked, or an
** error code is something goes wrong.
**
** All values are stored on disk as big-endian.
*/
static int read32bits(sqlite3_file *fd, i64 offset, u32 *pRes){
  unsigned char ac[4];
  int rc = sqlite3OsRead(fd, ac, sizeof(ac), offset);
  if( rc==SQLITE_OK ){
    *pRes = sqlite3Get4byte(ac);
  }
  return rc;
}

/*
** Write a 32-bit integer into a string buffer in big-endian byte order.
*/
#define put32bits(A,B)  sqlite3Put4byte((u8*)A,B)

/*
** Write a 32-bit integer into the given file descriptor.  Return SQLITE_OK
** on success or an error code is something goes wrong.
*/
static int write32bits(sqlite3_file *fd, i64 offset, u32 val){
  char ac[4];
  put32bits(ac, val);
  return sqlite3OsWrite(fd, ac, 4, offset);
}

/*
** This function should be called when an error occurs within the pager
** code. The first argument is a pointer to the pager structure, the
** second the error-code about to be returned by a pager API function. 
** The value returned is a copy of the second argument to this function. 
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
** written into memory obtained from sqliteMalloc(). *pzMaster is
** set to point at the memory and SQLITE_OK returned. The caller must
** sqliteFree() *pzMaster.
**
** If no master journal file name is present *pzMaster is set to 0 and
** SQLITE_OK returned.
*/
static int readMasterJournal(OsFile *pJrnl, char **pzMaster){
  int rc;
  u32 len;
  i64 szJ;
  u32 cksum;
  int i;
  unsigned char aMagic[8]; /* A buffer to hold the magic header */

  *pzMaster = 0;

  rc = sqlite3OsFileSize(pJrnl, &szJ);
  if( rc!=SQLITE_OK || szJ<16 ) return rc;

  rc = sqlite3OsSeek(pJrnl, szJ-16);
  if( rc!=SQLITE_OK ) return rc;
 
  rc = read32bits(pJrnl, &len);
  if( rc!=SQLITE_OK ) return rc;

  rc = read32bits(pJrnl, &cksum);
  if( rc!=SQLITE_OK ) return rc;

  rc = sqlite3OsRead(pJrnl, aMagic, 8);
  if( rc!=SQLITE_OK || memcmp(aMagic, aJournalMagic, 8) ) return rc;

  rc = sqlite3OsSeek(pJrnl, szJ-16-len);
  if( rc!=SQLITE_OK ) return rc;

  *pzMaster = (char *)sqliteMalloc(len+1);
  if( !*pzMaster ){
    return SQLITE_NOMEM;
  }
  rc = sqlite3OsRead(pJrnl, *pzMaster, len);
  if( rc!=SQLITE_OK ){
    sqliteFree(*pzMaster);
    *pzMaster = 0;
    return rc;
  }

  /* See if the checksum matches the master journal name */







|












<
<
<
|


|


|


<
<
<




|







623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642



643
644
645
646
647
648
649
650
651



652
653
654
655
656
657
658
659
660
661
662
663
** written into memory obtained from sqliteMalloc(). *pzMaster is
** set to point at the memory and SQLITE_OK returned. The caller must
** sqliteFree() *pzMaster.
**
** If no master journal file name is present *pzMaster is set to 0 and
** SQLITE_OK returned.
*/
static int readMasterJournal(sqlite3_file *pJrnl, char **pzMaster){
  int rc;
  u32 len;
  i64 szJ;
  u32 cksum;
  int i;
  unsigned char aMagic[8]; /* A buffer to hold the magic header */

  *pzMaster = 0;

  rc = sqlite3OsFileSize(pJrnl, &szJ);
  if( rc!=SQLITE_OK || szJ<16 ) return rc;




  rc = read32bits(pJrnl, szJ-16, &len);
  if( rc!=SQLITE_OK ) return rc;

  rc = read32bits(pJrnl, szJ-12, &cksum);
  if( rc!=SQLITE_OK ) return rc;

  rc = sqlite3OsRead(pJrnl, aMagic, 8, szJ-8);
  if( rc!=SQLITE_OK || memcmp(aMagic, aJournalMagic, 8) ) return rc;




  *pzMaster = (char *)sqliteMalloc(len+1);
  if( !*pzMaster ){
    return SQLITE_NOMEM;
  }
  rc = sqlite3OsRead(pJrnl, *pzMaster, len, szJ-16-len);
  if( rc!=SQLITE_OK ){
    sqliteFree(*pzMaster);
    *pzMaster = 0;
    return rc;
  }

  /* See if the checksum matches the master journal name */
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
** ---------------------------------------
** 0                         0
** 512                       512
** 100                       512
** 2000                      2048
** 
*/
static int seekJournalHdr(Pager *pPager){
  i64 offset = 0;
  i64 c = pPager->journalOff;
  if( c ){
    offset = ((c-1)/JOURNAL_HDR_SZ(pPager) + 1) * JOURNAL_HDR_SZ(pPager);
  }
  assert( offset%JOURNAL_HDR_SZ(pPager)==0 );
  assert( offset>=c );
  assert( (offset-c)<JOURNAL_HDR_SZ(pPager) );
  pPager->journalOff = offset;
  return sqlite3OsSeek(pPager->jfd, pPager->journalOff);
}

/*
** The journal file must be open when this routine is called. A journal
** header (JOURNAL_HDR_SZ bytes) is written into the journal file at the
** current location.
**







|









<







690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706

707
708
709
710
711
712
713
** ---------------------------------------
** 0                         0
** 512                       512
** 100                       512
** 2000                      2048
** 
*/
static void seekJournalHdr(Pager *pPager){
  i64 offset = 0;
  i64 c = pPager->journalOff;
  if( c ){
    offset = ((c-1)/JOURNAL_HDR_SZ(pPager) + 1) * JOURNAL_HDR_SZ(pPager);
  }
  assert( offset%JOURNAL_HDR_SZ(pPager)==0 );
  assert( offset>=c );
  assert( (offset-c)<JOURNAL_HDR_SZ(pPager) );
  pPager->journalOff = offset;

}

/*
** The journal file must be open when this routine is called. A journal
** header (JOURNAL_HDR_SZ bytes) is written into the journal file at the
** current location.
**
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763

764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
  char zHeader[sizeof(aJournalMagic)+16];
  int rc;

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

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

  pPager->journalHdr = pPager->journalOff;
  pPager->journalOff += JOURNAL_HDR_SZ(pPager);

  /* FIX ME: 
  **
  ** Possibly for a pager not in no-sync mode, the journal magic should not
  ** be written until nRec is filled in as part of next syncJournal(). 
  **
  ** Actually maybe the whole journal header should be delayed until that
  ** point. Think about this.
  */
  memcpy(zHeader, aJournalMagic, sizeof(aJournalMagic));
  /* The nRec Field. 0xFFFFFFFF for no-sync journals. */
  put32bits(&zHeader[sizeof(aJournalMagic)], pPager->noSync ? 0xffffffff : 0);
  /* The random check-hash initialiser */ 
  sqlite3Randomness(sizeof(pPager->cksumInit), &pPager->cksumInit);
  put32bits(&zHeader[sizeof(aJournalMagic)+4], pPager->cksumInit);
  /* The initial database size */
  put32bits(&zHeader[sizeof(aJournalMagic)+8], pPager->dbSize);
  /* The assumed sector size for this process */
  put32bits(&zHeader[sizeof(aJournalMagic)+12], pPager->sectorSize);
  IOTRACE(("JHDR %p %lld %d\n", pPager, pPager->journalHdr, sizeof(zHeader)))
  rc = sqlite3OsWrite(pPager->jfd, zHeader, sizeof(zHeader));


  /* The journal header has been written successfully. Seek the journal
  ** file descriptor to the end of the journal header sector.
  */
  if( rc==SQLITE_OK ){
    IOTRACE(("JTAIL %p %lld\n", pPager, pPager->journalOff-1))
    rc = sqlite3OsSeek(pPager->jfd, pPager->journalOff-1);
    if( rc==SQLITE_OK ){
      rc = sqlite3OsWrite(pPager->jfd, "\000", 1);
    }
  }
  return rc;
}

/*
** The journal file must be open when this is called. A journal header file
** (JOURNAL_HDR_SZ bytes) is read from the current location in the journal







|
<
<

<




















|
>






<
<
|
<







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
753
754
755
756
757
758
759
760


761

762
763
764
765
766
767
768
  char zHeader[sizeof(aJournalMagic)+16];
  int rc;

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

  seekJournalHdr(pPager);


  pPager->journalHdr = pPager->journalOff;


  /* FIX ME: 
  **
  ** Possibly for a pager not in no-sync mode, the journal magic should not
  ** be written until nRec is filled in as part of next syncJournal(). 
  **
  ** Actually maybe the whole journal header should be delayed until that
  ** point. Think about this.
  */
  memcpy(zHeader, aJournalMagic, sizeof(aJournalMagic));
  /* The nRec Field. 0xFFFFFFFF for no-sync journals. */
  put32bits(&zHeader[sizeof(aJournalMagic)], pPager->noSync ? 0xffffffff : 0);
  /* The random check-hash initialiser */ 
  sqlite3Randomness(sizeof(pPager->cksumInit), &pPager->cksumInit);
  put32bits(&zHeader[sizeof(aJournalMagic)+4], pPager->cksumInit);
  /* The initial database size */
  put32bits(&zHeader[sizeof(aJournalMagic)+8], pPager->dbSize);
  /* The assumed sector size for this process */
  put32bits(&zHeader[sizeof(aJournalMagic)+12], pPager->sectorSize);
  IOTRACE(("JHDR %p %lld %d\n", pPager, pPager->journalHdr, sizeof(zHeader)))
  rc = sqlite3OsWrite(pPager->jfd, zHeader, sizeof(zHeader),pPager->journalOff);
  pPager->journalOff += JOURNAL_HDR_SZ(pPager);

  /* The journal header has been written successfully. Seek the journal
  ** file descriptor to the end of the journal header sector.
  */
  if( rc==SQLITE_OK ){
    IOTRACE(("JTAIL %p %lld\n", pPager, pPager->journalOff-1))


    rc = sqlite3OsWrite(pPager->jfd, "\000", 1, pPager->journalOff-1);

  }
  return rc;
}

/*
** The journal file must be open when this is called. A journal header file
** (JOURNAL_HDR_SZ bytes) is read from the current location in the journal
795
796
797
798
799
800
801

802
803
804
805
806
807
808

809
810
811

812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
  Pager *pPager, 
  i64 journalSize,
  u32 *pNRec, 
  u32 *pDbSize
){
  int rc;
  unsigned char aMagic[8]; /* A buffer to hold the magic header */


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

  if( pPager->journalOff+JOURNAL_HDR_SZ(pPager) > journalSize ){
    return SQLITE_DONE;
  }


  rc = sqlite3OsRead(pPager->jfd, aMagic, sizeof(aMagic));
  if( rc ) return rc;


  if( memcmp(aMagic, aJournalMagic, sizeof(aMagic))!=0 ){
    return SQLITE_DONE;
  }

  rc = read32bits(pPager->jfd, pNRec);
  if( rc ) return rc;

  rc = read32bits(pPager->jfd, &pPager->cksumInit);
  if( rc ) return rc;

  rc = read32bits(pPager->jfd, pDbSize);
  if( rc ) return rc;

  /* Update the assumed sector-size to match the value used by 
  ** the process that created this journal. If this journal was
  ** created by a process other than this one, then this routine
  ** is being called from within pager_playback(). The local value
  ** of Pager.sectorSize is restored at the end of that routine.
  */
  rc = read32bits(pPager->jfd, (u32 *)&pPager->sectorSize);
  if( rc ) return rc;

  pPager->journalOff += JOURNAL_HDR_SZ(pPager);
  rc = sqlite3OsSeek(pPager->jfd, pPager->journalOff);
  return rc;
}


/*
** Write the supplied master journal name into the journal file for pager
** pPager at the current location. The master journal name must be the last
** thing written to a journal file. If the pager is in full-sync mode, the







>

|
<
<



>

|

>





|


|


|








|



<
|







783
784
785
786
787
788
789
790
791
792


793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824

825
826
827
828
829
830
831
832
  Pager *pPager, 
  i64 journalSize,
  u32 *pNRec, 
  u32 *pDbSize
){
  int rc;
  unsigned char aMagic[8]; /* A buffer to hold the magic header */
  i64 jrnlOff;

  seekJournalHdr(pPager);


  if( pPager->journalOff+JOURNAL_HDR_SZ(pPager) > journalSize ){
    return SQLITE_DONE;
  }
  jrnlOff = pPager->journalOff;

  rc = sqlite3OsRead(pPager->jfd, aMagic, sizeof(aMagic), jrnlOff);
  if( rc ) return rc;
  jrnlOff += sizeof(aMagic);

  if( memcmp(aMagic, aJournalMagic, sizeof(aMagic))!=0 ){
    return SQLITE_DONE;
  }

  rc = read32bits(pPager->jfd, jrnlOff, pNRec);
  if( rc ) return rc;

  rc = read32bits(pPager->jfd, jrnlOff+4, &pPager->cksumInit);
  if( rc ) return rc;

  rc = read32bits(pPager->jfd, jrnlOff+8, pDbSize);
  if( rc ) return rc;

  /* Update the assumed sector-size to match the value used by 
  ** the process that created this journal. If this journal was
  ** created by a process other than this one, then this routine
  ** is being called from within pager_playback(). The local value
  ** of Pager.sectorSize is restored at the end of that routine.
  */
  rc = read32bits(pPager->jfd, jrnlOff+12, (u32 *)&pPager->sectorSize);
  if( rc ) return rc;

  pPager->journalOff += JOURNAL_HDR_SZ(pPager);

  return SQLITE_OK;
}


/*
** Write the supplied master journal name into the journal file for pager
** pPager at the current location. The master journal name must be the last
** thing written to a journal file. If the pager is in full-sync mode, the
857
858
859
860
861
862
863

864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882

883
884
885
886

887
888
889

890
891
892
893
894
895
896
897
898
899
900
901
** If zMaster is a NULL pointer (occurs for a single database transaction), 
** this call is a no-op.
*/
static int writeMasterJournal(Pager *pPager, const char *zMaster){
  int rc;
  int len; 
  int i; 

  u32 cksum = 0;
  char zBuf[sizeof(aJournalMagic)+2*4];

  if( !zMaster || pPager->setMaster) return SQLITE_OK;
  pPager->setMaster = 1;

  len = strlen(zMaster);
  for(i=0; i<len; i++){
    cksum += zMaster[i];
  }

  /* If in full-sync mode, advance to the next disk sector before writing
  ** the master journal name. This is in case the previous page written to
  ** the journal has already been synced.
  */
  if( pPager->fullSync ){
    rc = seekJournalHdr(pPager);
    if( rc!=SQLITE_OK ) return rc;
  }

  pPager->journalOff += (len+20);

  rc = write32bits(pPager->jfd, PAGER_MJ_PGNO(pPager));
  if( rc!=SQLITE_OK ) return rc;


  rc = sqlite3OsWrite(pPager->jfd, zMaster, len);
  if( rc!=SQLITE_OK ) return rc;


  put32bits(zBuf, len);
  put32bits(&zBuf[4], cksum);
  memcpy(&zBuf[8], aJournalMagic, sizeof(aJournalMagic));
  rc = sqlite3OsWrite(pPager->jfd, zBuf, 8+sizeof(aJournalMagic));
  pPager->needSync = !pPager->noSync;
  return rc;
}

/*
** Add or remove a page from the list of all pages that are in the
** statement journal.







>
















|
<

>


|

>

|

>




|







845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869

870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
** If zMaster is a NULL pointer (occurs for a single database transaction), 
** this call is a no-op.
*/
static int writeMasterJournal(Pager *pPager, const char *zMaster){
  int rc;
  int len; 
  int i; 
  i64 jrnlOff;
  u32 cksum = 0;
  char zBuf[sizeof(aJournalMagic)+2*4];

  if( !zMaster || pPager->setMaster) return SQLITE_OK;
  pPager->setMaster = 1;

  len = strlen(zMaster);
  for(i=0; i<len; i++){
    cksum += zMaster[i];
  }

  /* If in full-sync mode, advance to the next disk sector before writing
  ** the master journal name. This is in case the previous page written to
  ** the journal has already been synced.
  */
  if( pPager->fullSync ){
    seekJournalHdr(pPager);

  }
  jrnlOff = pPager->journalOff;
  pPager->journalOff += (len+20);

  rc = write32bits(pPager->jfd, jrnlOff, PAGER_MJ_PGNO(pPager));
  if( rc!=SQLITE_OK ) return rc;
  jrnlOff += 4;

  rc = sqlite3OsWrite(pPager->jfd, zMaster, len, jrnlOff);
  if( rc!=SQLITE_OK ) return rc;
  jrnlOff += len;

  put32bits(zBuf, len);
  put32bits(&zBuf[4], cksum);
  memcpy(&zBuf[8], aJournalMagic, sizeof(aJournalMagic));
  rc = sqlite3OsWrite(pPager->jfd, zBuf, 8+sizeof(aJournalMagic), jrnlOff);
  pPager->needSync = !pPager->noSync;
  return rc;
}

/*
** Add or remove a page from the list of all pages that are in the
** statement journal.
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
  if( pPager->stmtOpen && !pPager->exclusiveMode ){
    sqlite3OsClose(&pPager->stfd);
    pPager->stmtOpen = 0;
  }
  if( pPager->journalOpen ){
    if( pPager->exclusiveMode 
          && (rc = sqlite3OsTruncate(pPager->jfd, 0))==SQLITE_OK ){;
      sqlite3OsSeek(pPager->jfd, 0);
      pPager->journalOff = 0;
      pPager->journalStarted = 0;
    }else{
      sqlite3OsClose(&pPager->jfd);
      pPager->journalOpen = 0;
      if( rc==SQLITE_OK ){
        rc = sqlite3OsDelete(pPager->zJournal);







<







1013
1014
1015
1016
1017
1018
1019

1020
1021
1022
1023
1024
1025
1026
  if( pPager->stmtOpen && !pPager->exclusiveMode ){
    sqlite3OsClose(&pPager->stfd);
    pPager->stmtOpen = 0;
  }
  if( pPager->journalOpen ){
    if( pPager->exclusiveMode 
          && (rc = sqlite3OsTruncate(pPager->jfd, 0))==SQLITE_OK ){;

      pPager->journalOff = 0;
      pPager->journalStarted = 0;
    }else{
      sqlite3OsClose(&pPager->jfd);
      pPager->journalOpen = 0;
      if( rc==SQLITE_OK ){
        rc = sqlite3OsDelete(pPager->zJournal);
1107
1108
1109
1110
1111
1112
1113
1114





1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
** Read a single page from the journal file opened on file descriptor
** jfd.  Playback this one page.
**
** If useCksum==0 it means this journal does not use checksums.  Checksums
** are not used in statement journals because statement journals do not
** need to survive power failures.
*/
static int pager_playback_one_page(Pager *pPager, OsFile *jfd, int useCksum){





  int rc;
  PgHdr *pPg;                   /* An existing page in the cache */
  Pgno pgno;                    /* The page number of a page in journal */
  u32 cksum;                    /* Checksum used for sanity checking */
  u8 *aData = (u8 *)pPager->pTmpSpace;   /* Temp storage for a page */

  /* useCksum should be true for the main journal and false for
  ** statement journals.  Verify that this is always the case
  */
  assert( jfd == (useCksum ? pPager->jfd : pPager->stfd) );
  assert( aData );

  rc = read32bits(jfd, &pgno);
  if( rc!=SQLITE_OK ) return rc;
  rc = sqlite3OsRead(jfd, aData, pPager->pageSize);
  if( rc!=SQLITE_OK ) return rc;
  pPager->journalOff += pPager->pageSize + 4;

  /* Sanity checking on the page.  This is more important that I originally
  ** thought.  If a power failure occurs while the journal is being written,
  ** it could cause invalid data to be written into the journal.  We need to
  ** detect this invalid data (with high probability) and ignore it.
  */
  if( pgno==0 || pgno==PAGER_MJ_PGNO(pPager) ){
    return SQLITE_DONE;
  }
  if( pgno>(unsigned)pPager->dbSize ){
    return SQLITE_OK;
  }
  if( useCksum ){
    rc = read32bits(jfd, &cksum);
    if( rc ) return rc;
    pPager->journalOff += 4;
    if( pager_cksum(pPager, aData)!=cksum ){
      return SQLITE_DONE;
    }
  }








|
>
>
>
>
>












|

|















|







1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
** Read a single page from the journal file opened on file descriptor
** jfd.  Playback this one page.
**
** If useCksum==0 it means this journal does not use checksums.  Checksums
** are not used in statement journals because statement journals do not
** need to survive power failures.
*/
static int pager_playback_one_page(
  Pager *pPager, 
  sqlite3_file *jfd,
  i64 offset,
  int useCksum
){
  int rc;
  PgHdr *pPg;                   /* An existing page in the cache */
  Pgno pgno;                    /* The page number of a page in journal */
  u32 cksum;                    /* Checksum used for sanity checking */
  u8 *aData = (u8 *)pPager->pTmpSpace;   /* Temp storage for a page */

  /* useCksum should be true for the main journal and false for
  ** statement journals.  Verify that this is always the case
  */
  assert( jfd == (useCksum ? pPager->jfd : pPager->stfd) );
  assert( aData );

  rc = read32bits(jfd, offset, &pgno);
  if( rc!=SQLITE_OK ) return rc;
  rc = sqlite3OsRead(jfd, aData, pPager->pageSize, offset+4);
  if( rc!=SQLITE_OK ) return rc;
  pPager->journalOff += pPager->pageSize + 4;

  /* Sanity checking on the page.  This is more important that I originally
  ** thought.  If a power failure occurs while the journal is being written,
  ** it could cause invalid data to be written into the journal.  We need to
  ** detect this invalid data (with high probability) and ignore it.
  */
  if( pgno==0 || pgno==PAGER_MJ_PGNO(pPager) ){
    return SQLITE_DONE;
  }
  if( pgno>(unsigned)pPager->dbSize ){
    return SQLITE_OK;
  }
  if( useCksum ){
    rc = read32bits(jfd, offset+pPager->pageSize+4, &cksum);
    if( rc ) return rc;
    pPager->journalOff += 4;
    if( pager_cksum(pPager, aData)!=cksum ){
      return SQLITE_DONE;
    }
  }

1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
  ** in the main journal either because the page is not in cache or else
  ** the page is marked as needSync==0.
  */
  pPg = pager_lookup(pPager, pgno);
  PAGERTRACE4("PLAYBACK %d page %d hash(%08x)\n",
               PAGERID(pPager), pgno, pager_datahash(pPager->pageSize, aData));
  if( pPager->state>=PAGER_EXCLUSIVE && (pPg==0 || pPg->needSync==0) ){
    rc = sqlite3OsSeek(pPager->fd, (pgno-1)*(i64)pPager->pageSize);
    if( rc==SQLITE_OK ){
      rc = sqlite3OsWrite(pPager->fd, aData, pPager->pageSize);
    }
    if( pPg ){
      makeClean(pPg);
    }
  }
  if( pPg ){
    /* No page should ever be explicitly rolled back that is in use, except
    ** for page 1 which is held in use in order to keep the lock on the







|
<
|
<







1175
1176
1177
1178
1179
1180
1181
1182

1183

1184
1185
1186
1187
1188
1189
1190
  ** in the main journal either because the page is not in cache or else
  ** the page is marked as needSync==0.
  */
  pPg = pager_lookup(pPager, pgno);
  PAGERTRACE4("PLAYBACK %d page %d hash(%08x)\n",
               PAGERID(pPager), pgno, pager_datahash(pPager->pageSize, aData));
  if( pPager->state>=PAGER_EXCLUSIVE && (pPg==0 || pPg->needSync==0) ){
    i64 offset = (pgno-1)*(i64)pPager->pageSize;

    rc = sqlite3OsWrite(pPager->fd, aData, pPager->pageSize, offset);

    if( pPg ){
      makeClean(pPg);
    }
  }
  if( pPg ){
    /* No page should ever be explicitly rolled back that is in use, except
    ** for page 1 which is held in use in order to keep the lock on the
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
** To tell if a master journal can be deleted, check to each of the
** children.  If all children are either missing or do not refer to
** a different master journal, then this master journal can be deleted.
*/
static int pager_delmaster(const char *zMaster){
  int rc;
  int master_open = 0;
  OsFile *master = 0;
  char *zMasterJournal = 0; /* Contents of master journal file */
  i64 nMasterJournal;       /* Size of master journal file */

  /* Open the master journal file exclusively in case some other process
  ** is running this routine also. Not that it makes too much difference.
  */
  rc = sqlite3OsOpenReadOnly(zMaster, &master);







|







1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
** To tell if a master journal can be deleted, check to each of the
** children.  If all children are either missing or do not refer to
** a different master journal, then this master journal can be deleted.
*/
static int pager_delmaster(const char *zMaster){
  int rc;
  int master_open = 0;
  sqlite3_file *master = 0;
  char *zMasterJournal = 0; /* Contents of master journal file */
  i64 nMasterJournal;       /* Size of master journal file */

  /* Open the master journal file exclusively in case some other process
  ** is running this routine also. Not that it makes too much difference.
  */
  rc = sqlite3OsOpenReadOnly(zMaster, &master);
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
    ** sqliteMalloc() and pointed to by zMasterJournal. 
    */
    zMasterJournal = (char *)sqliteMalloc(nMasterJournal);
    if( !zMasterJournal ){
      rc = SQLITE_NOMEM;
      goto delmaster_out;
    }
    rc = sqlite3OsRead(master, zMasterJournal, nMasterJournal);
    if( rc!=SQLITE_OK ) goto delmaster_out;

    zJournal = zMasterJournal;
    while( (zJournal-zMasterJournal)<nMasterJournal ){
      if( sqlite3OsFileExists(zJournal) ){
        /* One of the journals pointed to by the master journal exists.
        ** Open it and check if it points at the master journal. If
        ** so, return without deleting the master journal file.
        */
        OsFile *journal = 0;
        int c;

        rc = sqlite3OsOpenReadOnly(zJournal, &journal);
        assert( rc!=SQLITE_OK || journal );
        if( rc!=SQLITE_OK ){
          goto delmaster_out;
        }







|









|







1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
    ** sqliteMalloc() and pointed to by zMasterJournal. 
    */
    zMasterJournal = (char *)sqliteMalloc(nMasterJournal);
    if( !zMasterJournal ){
      rc = SQLITE_NOMEM;
      goto delmaster_out;
    }
    rc = sqlite3OsRead(master, zMasterJournal, nMasterJournal, 0);
    if( rc!=SQLITE_OK ) goto delmaster_out;

    zJournal = zMasterJournal;
    while( (zJournal-zMasterJournal)<nMasterJournal ){
      if( sqlite3OsFileExists(zJournal) ){
        /* One of the journals pointed to by the master journal exists.
        ** Open it and check if it points at the master journal. If
        ** so, return without deleting the master journal file.
        */
        sqlite3_file *journal = 0;
        int c;

        rc = sqlite3OsOpenReadOnly(zJournal, &journal);
        assert( rc!=SQLITE_OK || journal );
        if( rc!=SQLITE_OK ){
          goto delmaster_out;
        }
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
  assert( rc!=SQLITE_DONE );
  if( rc!=SQLITE_OK || (zMaster && !sqlite3OsFileExists(zMaster)) ){
    sqliteFree(zMaster);
    zMaster = 0;
    if( rc==SQLITE_DONE ) rc = SQLITE_OK;
    goto end_playback;
  }
  sqlite3OsSeek(pPager->jfd, 0);
  pPager->journalOff = 0;

  /* This loop terminates either when the readJournalHdr() call returns
  ** SQLITE_DONE or an IO error occurs. */
  while( 1 ){

    /* Read the next journal header from the journal file.  If there are







<







1413
1414
1415
1416
1417
1418
1419

1420
1421
1422
1423
1424
1425
1426
  assert( rc!=SQLITE_DONE );
  if( rc!=SQLITE_OK || (zMaster && !sqlite3OsFileExists(zMaster)) ){
    sqliteFree(zMaster);
    zMaster = 0;
    if( rc==SQLITE_DONE ) rc = SQLITE_OK;
    goto end_playback;
  }

  pPager->journalOff = 0;

  /* This loop terminates either when the readJournalHdr() call returns
  ** SQLITE_DONE or an IO error occurs. */
  while( 1 ){

    /* Read the next journal header from the journal file.  If there are
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
        goto end_playback;
      }
    }

    /* Copy original pages out of the journal and back into the database file.
    */
    for(i=0; i<nRec; i++){
      rc = pager_playback_one_page(pPager, pPager->jfd, 1);
      if( rc!=SQLITE_OK ){
        if( rc==SQLITE_DONE ){
          rc = SQLITE_OK;
          pPager->journalOff = szJ;
          break;
        }else{
          goto end_playback;







|







1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
        goto end_playback;
      }
    }

    /* Copy original pages out of the journal and back into the database file.
    */
    for(i=0; i<nRec; i++){
      rc = pager_playback_one_page(pPager, pPager->jfd, pPager->journalOff, 1);
      if( rc!=SQLITE_OK ){
        if( rc==SQLITE_DONE ){
          rc = SQLITE_OK;
          pPager->journalOff = szJ;
          break;
        }else{
          goto end_playback;
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578

1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
  */
  rc = pager_truncate(pPager, pPager->stmtSize);
  assert( pPager->state>=PAGER_SHARED );

  /* Figure out how many records are in the statement journal.
  */
  assert( pPager->stmtInUse && pPager->journalOpen );
  sqlite3OsSeek(pPager->stfd, 0);
  nRec = pPager->stmtNRec;
  
  /* Copy original pages out of the statement journal and back into the
  ** database file.  Note that the statement journal omits checksums from
  ** each record since power-failure recovery is not important to statement
  ** journals.
  */
  for(i=nRec-1; i>=0; i--){

    rc = pager_playback_one_page(pPager, pPager->stfd, 0);
    assert( rc!=SQLITE_DONE );
    if( rc!=SQLITE_OK ) goto end_stmt_playback;
  }

  /* Now roll some pages back from the transaction journal. Pager.stmtJSize
  ** was the size of the journal file when this statement was started, so
  ** everything after that needs to be rolled back, either into the
  ** database, the memory cache, or both.
  **
  ** If it is not zero, then Pager.stmtHdrOff is the offset to the start
  ** of the first journal header written during this statement transaction.
  */
  rc = sqlite3OsSeek(pPager->jfd, pPager->stmtJSize);
  if( rc!=SQLITE_OK ){
    goto end_stmt_playback;
  }
  pPager->journalOff = pPager->stmtJSize;
  pPager->cksumInit = pPager->stmtCksum;
  while( pPager->journalOff < hdrOff ){
    rc = pager_playback_one_page(pPager, pPager->jfd, 1);
    assert( rc!=SQLITE_DONE );
    if( rc!=SQLITE_OK ) goto end_stmt_playback;
  }

  while( pPager->journalOff < szJ ){
    u32 nJRec;         /* Number of Journal Records */
    u32 dummy;
    rc = readJournalHdr(pPager, szJ, &nJRec, &dummy);
    if( rc!=SQLITE_OK ){
      assert( rc!=SQLITE_DONE );
      goto end_stmt_playback;
    }
    if( nJRec==0 ){
      nJRec = (szJ - pPager->journalOff) / (pPager->pageSize+8);
    }
    for(i=nJRec-1; i>=0 && pPager->journalOff < szJ; i--){
      rc = pager_playback_one_page(pPager, pPager->jfd, 1);
      assert( rc!=SQLITE_DONE );
      if( rc!=SQLITE_OK ) goto end_stmt_playback;
    }
  }

  pPager->journalOff = szJ;
  







<







|
>
|












<
<
<
<



|
















|







1555
1556
1557
1558
1559
1560
1561

1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583




1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
  */
  rc = pager_truncate(pPager, pPager->stmtSize);
  assert( pPager->state>=PAGER_SHARED );

  /* Figure out how many records are in the statement journal.
  */
  assert( pPager->stmtInUse && pPager->journalOpen );

  nRec = pPager->stmtNRec;
  
  /* Copy original pages out of the statement journal and back into the
  ** database file.  Note that the statement journal omits checksums from
  ** each record since power-failure recovery is not important to statement
  ** journals.
  */
  for(i=0; i<nRec; i++){
    i64 offset = i*(4+pPager->pageSize);
    rc = pager_playback_one_page(pPager, pPager->stfd, offset, 0);
    assert( rc!=SQLITE_DONE );
    if( rc!=SQLITE_OK ) goto end_stmt_playback;
  }

  /* Now roll some pages back from the transaction journal. Pager.stmtJSize
  ** was the size of the journal file when this statement was started, so
  ** everything after that needs to be rolled back, either into the
  ** database, the memory cache, or both.
  **
  ** If it is not zero, then Pager.stmtHdrOff is the offset to the start
  ** of the first journal header written during this statement transaction.
  */




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

  while( pPager->journalOff < szJ ){
    u32 nJRec;         /* Number of Journal Records */
    u32 dummy;
    rc = readJournalHdr(pPager, szJ, &nJRec, &dummy);
    if( rc!=SQLITE_OK ){
      assert( rc!=SQLITE_DONE );
      goto end_stmt_playback;
    }
    if( nJRec==0 ){
      nJRec = (szJ - pPager->journalOff) / (pPager->pageSize+8);
    }
    for(i=nJRec-1; i>=0 && pPager->journalOff < szJ; i--){
      rc = pager_playback_one_page(pPager, pPager->jfd, pPager->journalOff, 1);
      assert( rc!=SQLITE_DONE );
      if( rc!=SQLITE_OK ) goto end_stmt_playback;
    }
  }

  pPager->journalOff = szJ;
  
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
**
** Write the file descriptor into *fd.  Return SQLITE_OK on success or some
** other error code if we fail.
**
** The OS will automatically delete the temporary file when it is
** closed.
*/
static int sqlite3PagerOpentemp(OsFile **pFd){
  int cnt = 8;
  int rc;
  char zFile[SQLITE_TEMPNAME_SIZE];

#ifdef SQLITE_TEST
  sqlite3_opentemp_count++;  /* Used for testing and analysis only */
#endif







|







1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
**
** Write the file descriptor into *fd.  Return SQLITE_OK on success or some
** other error code if we fail.
**
** The OS will automatically delete the temporary file when it is
** closed.
*/
static int sqlite3PagerOpentemp(sqlite3_file **pFd){
  int cnt = 8;
  int rc;
  char zFile[SQLITE_TEMPNAME_SIZE];

#ifdef SQLITE_TEST
  sqlite3_opentemp_count++;  /* Used for testing and analysis only */
#endif
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
  const char *zFilename,   /* Name of the database file to open */
  int nExtra,              /* Extra bytes append to each in-memory page */
  int flags                /* flags controlling this file */
){
  Pager *pPager = 0;
  char *zFullPathname = 0;
  int nameLen;  /* Compiler is wrong. This is always initialized before use */
  OsFile *fd = 0;
  int rc = SQLITE_OK;
  int i;
  int tempFile = 0;
  int memDb = 0;
  int readOnly = 0;
  int useJournal = (flags & PAGER_OMIT_JOURNAL)==0;
  int noReadlock = (flags & PAGER_NO_READLOCK)!=0;







|







1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
  const char *zFilename,   /* Name of the database file to open */
  int nExtra,              /* Extra bytes append to each in-memory page */
  int flags                /* flags controlling this file */
){
  Pager *pPager = 0;
  char *zFullPathname = 0;
  int nameLen;  /* Compiler is wrong. This is always initialized before use */
  sqlite3_file *fd = 0;
  int rc = SQLITE_OK;
  int i;
  int tempFile = 0;
  int memDb = 0;
  int readOnly = 0;
  int useJournal = (flags & PAGER_OMIT_JOURNAL)==0;
  int noReadlock = (flags & PAGER_NO_READLOCK)!=0;
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
** response is to zero the memory at pDest and continue.  A real IO error 
** will presumably recur and be picked up later (Todo: Think about this).
*/
int sqlite3PagerReadFileheader(Pager *pPager, int N, unsigned char *pDest){
  int rc = SQLITE_OK;
  memset(pDest, 0, N);
  if( MEMDB==0 ){
    disable_simulated_io_errors();
    sqlite3OsSeek(pPager->fd, 0);
    enable_simulated_io_errors();
    IOTRACE(("DBHDR %p 0 %d\n", pPager, N))
    rc = sqlite3OsRead(pPager->fd, pDest, N);
    if( rc==SQLITE_IOERR_SHORT_READ ){
      rc = SQLITE_OK;
    }
  }
  return rc;
}








<
<
<

|







1954
1955
1956
1957
1958
1959
1960



1961
1962
1963
1964
1965
1966
1967
1968
1969
** response is to zero the memory at pDest and continue.  A real IO error 
** will presumably recur and be picked up later (Todo: Think about this).
*/
int sqlite3PagerReadFileheader(Pager *pPager, int N, unsigned char *pDest){
  int rc = SQLITE_OK;
  memset(pDest, 0, N);
  if( MEMDB==0 ){



    IOTRACE(("DBHDR %p 0 %d\n", pPager, N))
    rc = sqlite3OsRead(pPager->fd, pDest, N, 0);
    if( rc==SQLITE_IOERR_SHORT_READ ){
      rc = SQLITE_OK;
    }
  }
  return rc;
}

2377
2378
2379
2380
2381
2382
2383


2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
        i64 jSz;
        rc = sqlite3OsFileSize(pPager->jfd, &jSz);
        if( rc!=0 ) return rc;
        assert( pPager->journalOff==jSz );
      }
#endif
      {


        /* Write the nRec value into the journal file header. If in
        ** full-synchronous mode, sync the journal first. This ensures that
        ** all data has really hit the disk before nRec is updated to mark
        ** it as a candidate for rollback. 
        */
        if( pPager->fullSync ){
          PAGERTRACE2("SYNC journal of %d\n", PAGERID(pPager));
          IOTRACE(("JSYNC %p\n", pPager))
          rc = sqlite3OsSync(pPager->jfd, 0);
          if( rc!=0 ) return rc;
        }
        rc = sqlite3OsSeek(pPager->jfd,
                           pPager->journalHdr + sizeof(aJournalMagic));
        if( rc ) return rc;
        IOTRACE(("JHDR %p %lld %d\n", pPager,
                  pPager->journalHdr + sizeof(aJournalMagic), 4))
        rc = write32bits(pPager->jfd, pPager->nRec);
        if( rc ) return rc;

        rc = sqlite3OsSeek(pPager->jfd, pPager->journalOff);
        if( rc ) return rc;
      }
      PAGERTRACE2("SYNC journal of %d\n", PAGERID(pPager));
      IOTRACE(("JSYNC %p\n", pPager))
      rc = sqlite3OsSync(pPager->jfd, pPager->full_fsync);
      if( rc!=0 ) return rc;
      pPager->journalStarted = 1;







>
>











|
|
<
|
<
|
<
<
<







2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383

2384

2385



2386
2387
2388
2389
2390
2391
2392
        i64 jSz;
        rc = sqlite3OsFileSize(pPager->jfd, &jSz);
        if( rc!=0 ) return rc;
        assert( pPager->journalOff==jSz );
      }
#endif
      {
        i64 jrnlOff;

        /* Write the nRec value into the journal file header. If in
        ** full-synchronous mode, sync the journal first. This ensures that
        ** all data has really hit the disk before nRec is updated to mark
        ** it as a candidate for rollback. 
        */
        if( pPager->fullSync ){
          PAGERTRACE2("SYNC journal of %d\n", PAGERID(pPager));
          IOTRACE(("JSYNC %p\n", pPager))
          rc = sqlite3OsSync(pPager->jfd, 0);
          if( rc!=0 ) return rc;
        }

        jrnlOff = pPager->journalHdr + sizeof(aJournalMagic);

        IOTRACE(("JHDR %p %lld %d\n", pPager, jrnlOff, 4));

        rc = write32bits(pPager->jfd, jrnlOff, pPager->nRec);



        if( rc ) return rc;
      }
      PAGERTRACE2("SYNC journal of %d\n", PAGERID(pPager));
      IOTRACE(("JSYNC %p\n", pPager))
      rc = sqlite3OsSync(pPager->jfd, pPager->full_fsync);
      if( rc!=0 ) return rc;
      pPager->journalStarted = 1;
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555

2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
  if( rc!=SQLITE_OK ){
    return rc;
  }

  pList = sort_pagelist(pList);
  while( pList ){
    assert( pList->dirty );
    rc = sqlite3OsSeek(pPager->fd, (pList->pgno-1)*(i64)pPager->pageSize);
    if( rc ) return rc;
    /* If there are dirty pages in the page cache with page numbers greater
    ** than Pager.dbSize, this means sqlite3PagerTruncate() was called to
    ** make the file smaller (presumably by auto-vacuum code). Do not write
    ** any such pages to the file.
    */
    if( pList->pgno<=pPager->dbSize ){

      char *pData = CODEC2(pPager, PGHDR_TO_DATA(pList), pList->pgno, 6);
      PAGERTRACE4("STORE %d page %d hash(%08x)\n",
                   PAGERID(pPager), pList->pgno, pager_pagehash(pList));
      IOTRACE(("PGOUT %p %d\n", pPager, pList->pgno));
      rc = sqlite3OsWrite(pPager->fd, pData, pPager->pageSize);
      PAGER_INCR(sqlite3_pager_writedb_count);
      PAGER_INCR(pPager->nWrite);
      if( pList->pgno==1 ){
        memcpy(&pPager->dbFileVers, &pData[24], sizeof(pPager->dbFileVers));
      }
    }
#ifndef NDEBUG







<
<






>




|







2523
2524
2525
2526
2527
2528
2529


2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
  if( rc!=SQLITE_OK ){
    return rc;
  }

  pList = sort_pagelist(pList);
  while( pList ){
    assert( pList->dirty );


    /* If there are dirty pages in the page cache with page numbers greater
    ** than Pager.dbSize, this means sqlite3PagerTruncate() was called to
    ** make the file smaller (presumably by auto-vacuum code). Do not write
    ** any such pages to the file.
    */
    if( pList->pgno<=pPager->dbSize ){
      i64 offset = (pList->pgno-1)*(i64)pPager->pageSize;
      char *pData = CODEC2(pPager, PGHDR_TO_DATA(pList), pList->pgno, 6);
      PAGERTRACE4("STORE %d page %d hash(%08x)\n",
                   PAGERID(pPager), pList->pgno, pager_pagehash(pList));
      IOTRACE(("PGOUT %p %d\n", pPager, pList->pgno));
      rc = sqlite3OsWrite(pPager->fd, pData, pPager->pageSize, offset);
      PAGER_INCR(sqlite3_pager_writedb_count);
      PAGER_INCR(pPager->nWrite);
      if( pList->pgno==1 ){
        memcpy(&pPager->dbFileVers, &pData[24], sizeof(pPager->dbFileVers));
      }
    }
#ifndef NDEBUG
2792
2793
2794
2795
2796
2797
2798

2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
#endif /* SQLITE_ENABLE_MEMORY_MANAGEMENT && !SQLITE_OMIT_DISKIO */

/*
** Read the content of page pPg out of the database file.
*/
static int readDbPage(Pager *pPager, PgHdr *pPg, Pgno pgno){
  int rc;

  assert( MEMDB==0 );
  rc = sqlite3OsSeek(pPager->fd, (pgno-1)*(i64)pPager->pageSize);
  if( rc==SQLITE_OK ){
    rc = sqlite3OsRead(pPager->fd, PGHDR_TO_DATA(pPg),
                          pPager->pageSize);
  }
  PAGER_INCR(sqlite3_pager_readdb_count);
  PAGER_INCR(pPager->nRead);
  IOTRACE(("PGIN %p %d\n", pPager, pgno));
  if( pgno==1 ){
    memcpy(&pPager->dbFileVers, &((u8*)PGHDR_TO_DATA(pPg))[24],
                                              sizeof(pPager->dbFileVers));
  }







>

|
<
|
<
<







2773
2774
2775
2776
2777
2778
2779
2780
2781
2782

2783


2784
2785
2786
2787
2788
2789
2790
#endif /* SQLITE_ENABLE_MEMORY_MANAGEMENT && !SQLITE_OMIT_DISKIO */

/*
** Read the content of page pPg out of the database file.
*/
static int readDbPage(Pager *pPager, PgHdr *pPg, Pgno pgno){
  int rc;
  i64 offset;
  assert( MEMDB==0 );
  offset = (pgno-1)*(i64)pPager->pageSize;

  rc = sqlite3OsRead(pPager->fd, PGHDR_TO_DATA(pPg), pPager->pageSize, offset);


  PAGER_INCR(sqlite3_pager_readdb_count);
  PAGER_INCR(pPager->nRead);
  IOTRACE(("PGIN %p %d\n", pPager, pgno));
  if( pgno==1 ){
    memcpy(&pPager->dbFileVers, &((u8*)PGHDR_TO_DATA(pPg))[24],
                                              sizeof(pPager->dbFileVers));
  }
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948

        if( pPager->errCode ){
          return pPager->errCode;
        }

        if( pPager->dbSize>0 ){
          IOTRACE(("CKVERS %p %d\n", pPager, sizeof(dbFileVers)));
          rc = sqlite3OsSeek(pPager->fd, 24);
          if( rc!=SQLITE_OK ){
            return rc;
          }
          rc = sqlite3OsRead(pPager->fd, &dbFileVers, sizeof(dbFileVers));
          if( rc!=SQLITE_OK ){
            return rc;
          }
        }else{
          memset(dbFileVers, 0, sizeof(dbFileVers));
        }








<
<
<
<
|







2909
2910
2911
2912
2913
2914
2915




2916
2917
2918
2919
2920
2921
2922
2923

        if( pPager->errCode ){
          return pPager->errCode;
        }

        if( pPager->dbSize>0 ){
          IOTRACE(("CKVERS %p %d\n", pPager, sizeof(dbFileVers)));




          rc = sqlite3OsRead(pPager->fd, &dbFileVers, sizeof(dbFileVers), 24);
          if( rc!=SQLITE_OK ){
            return rc;
          }
        }else{
          memset(dbFileVers, 0, sizeof(dbFileVers));
        }

3328
3329
3330
3331
3332
3333
3334

3335
3336
3337

3338
3339
3340
3341
3342
3343
3344
  pPager->journalHdr = 0;
  if( rc!=SQLITE_OK ){
    if( rc==SQLITE_NOMEM ){
      sqlite3OsDelete(pPager->zJournal);
    }
    goto failed_to_open_journal;
  }

  sqlite3OsSetFullSync(pPager->jfd, pPager->full_fsync);
  sqlite3OsSetFullSync(pPager->fd, pPager->full_fsync);
  sqlite3OsOpenDirectory(pPager->jfd, pPager->zDirectory);

  pPager->journalOpen = 1;
  pPager->journalStarted = 0;
  pPager->needSync = 0;
  pPager->alwaysRollback = 0;
  pPager->nRec = 0;
  if( pPager->errCode ){
    rc = pPager->errCode;







>



>







3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
  pPager->journalHdr = 0;
  if( rc!=SQLITE_OK ){
    if( rc==SQLITE_NOMEM ){
      sqlite3OsDelete(pPager->zJournal);
    }
    goto failed_to_open_journal;
  }
#if 0
  sqlite3OsSetFullSync(pPager->jfd, pPager->full_fsync);
  sqlite3OsSetFullSync(pPager->fd, pPager->full_fsync);
  sqlite3OsOpenDirectory(pPager->jfd, pPager->zDirectory);
#endif
  pPager->journalOpen = 1;
  pPager->journalStarted = 0;
  pPager->needSync = 0;
  pPager->alwaysRollback = 0;
  pPager->nRec = 0;
  if( pPager->errCode ){
    rc = pPager->errCode;
3579
3580
3581
3582
3583
3584
3585
3586
3587
3588
3589
3590
3591
3592
3593
          cksum = pager_cksum(pPager, (u8*)pData2);
          pEnd = pData2 + pPager->pageSize;
          pData2 -= 4;
          saved = *(u32*)pEnd;
          put32bits(pEnd, cksum);
          szPg = pPager->pageSize+8;
          put32bits(pData2, pPg->pgno);
          rc = sqlite3OsWrite(pPager->jfd, pData2, szPg);
          IOTRACE(("JOUT %p %d %lld %d\n", pPager, pPg->pgno,
                   pPager->journalOff, szPg));
          PAGER_INCR(sqlite3_pager_writej_count);
          pPager->journalOff += szPg;
          PAGERTRACE5("JOURNAL %d page %d needSync=%d hash(%08x)\n",
               PAGERID(pPager), pPg->pgno, pPg->needSync, pager_pagehash(pPg));
          *(u32*)pEnd = saved;







|







3556
3557
3558
3559
3560
3561
3562
3563
3564
3565
3566
3567
3568
3569
3570
          cksum = pager_cksum(pPager, (u8*)pData2);
          pEnd = pData2 + pPager->pageSize;
          pData2 -= 4;
          saved = *(u32*)pEnd;
          put32bits(pEnd, cksum);
          szPg = pPager->pageSize+8;
          put32bits(pData2, pPg->pgno);
          rc = sqlite3OsWrite(pPager->jfd, pData2, szPg, pPager->journalOff);
          IOTRACE(("JOUT %p %d %lld %d\n", pPager, pPg->pgno,
                   pPager->journalOff, szPg));
          PAGER_INCR(sqlite3_pager_writej_count);
          pPager->journalOff += szPg;
          PAGERTRACE5("JOURNAL %d page %d needSync=%d hash(%08x)\n",
               PAGERID(pPager), pPg->pgno, pPg->needSync, pager_pagehash(pPg));
          *(u32*)pEnd = saved;
3634
3635
3636
3637
3638
3639
3640

3641
3642
3643
3644
3645
3646
3647
3648
3649
3650
        pHist->pStmt = sqliteMallocRaw( pPager->pageSize );
        if( pHist->pStmt ){
          memcpy(pHist->pStmt, PGHDR_TO_DATA(pPg), pPager->pageSize);
        }
        PAGERTRACE3("STMT-JOURNAL %d page %d\n", PAGERID(pPager), pPg->pgno);
        page_add_to_stmt_list(pPg);
      }else{

        char *pData2 = CODEC2(pPager, pData, pPg->pgno, 7)-4;
        put32bits(pData2, pPg->pgno);
        rc = sqlite3OsWrite(pPager->stfd, pData2, pPager->pageSize+4);
        PAGERTRACE3("STMT-JOURNAL %d page %d\n", PAGERID(pPager), pPg->pgno);
        if( rc!=SQLITE_OK ){
          return rc;
        }
        pPager->stmtNRec++;
        assert( pPager->aInStmt!=0 );
        pPager->aInStmt[pPg->pgno/8] |= 1<<(pPg->pgno&7);







>


|







3611
3612
3613
3614
3615
3616
3617
3618
3619
3620
3621
3622
3623
3624
3625
3626
3627
3628
        pHist->pStmt = sqliteMallocRaw( pPager->pageSize );
        if( pHist->pStmt ){
          memcpy(pHist->pStmt, PGHDR_TO_DATA(pPg), pPager->pageSize);
        }
        PAGERTRACE3("STMT-JOURNAL %d page %d\n", PAGERID(pPager), pPg->pgno);
        page_add_to_stmt_list(pPg);
      }else{
        i64 offset = pPager->stmtNRec*(4+pPager->pageSize);
        char *pData2 = CODEC2(pPager, pData, pPg->pgno, 7)-4;
        put32bits(pData2, pPg->pgno);
        rc = sqlite3OsWrite(pPager->stfd, pData2, pPager->pageSize+4, offset);
        PAGERTRACE3("STMT-JOURNAL %d page %d\n", PAGERID(pPager), pPg->pgno);
        if( rc!=SQLITE_OK ){
          return rc;
        }
        pPager->stmtNRec++;
        assert( pPager->aInStmt!=0 );
        pPager->aInStmt[pPg->pgno/8] |= 1<<(pPg->pgno&7);
4211
4212
4213
4214
4215
4216
4217
4218
4219
4220
4221
4222
4223
4224
4225
** Commit a statement.
*/
int sqlite3PagerStmtCommit(Pager *pPager){
  if( pPager->stmtInUse ){
    PgHdr *pPg, *pNext;
    PAGERTRACE2("STMT-COMMIT %d\n", PAGERID(pPager));
    if( !MEMDB ){
      sqlite3OsSeek(pPager->stfd, 0);
      /* sqlite3OsTruncate(pPager->stfd, 0); */
      sqliteFree( pPager->aInStmt );
      pPager->aInStmt = 0;
    }else{
      for(pPg=pPager->pStmt; pPg; pPg=pNext){
        PgHistory *pHist = PGHDR_TO_HIST(pPg, pPager);
        pNext = pHist->pNextStmt;







<







4189
4190
4191
4192
4193
4194
4195

4196
4197
4198
4199
4200
4201
4202
** Commit a statement.
*/
int sqlite3PagerStmtCommit(Pager *pPager){
  if( pPager->stmtInUse ){
    PgHdr *pPg, *pNext;
    PAGERTRACE2("STMT-COMMIT %d\n", PAGERID(pPager));
    if( !MEMDB ){

      /* sqlite3OsTruncate(pPager->stfd, 0); */
      sqliteFree( pPager->aInStmt );
      pPager->aInStmt = 0;
    }else{
      for(pPg=pPager->pStmt; pPg; pPg=pNext){
        PgHistory *pHist = PGHDR_TO_HIST(pPg, pPager);
        pNext = pHist->pNextStmt;
Changes to src/sqlite.h.in.
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
** on how SQLite interfaces are suppose to operate.
**
** The name of this file under configuration management is "sqlite.h.in".
** The makefile makes some minor changes to this file (such as inserting
** the version number) and changes its name to "sqlite3.h" as
** part of the build process.
**
** @(#) $Id: sqlite.h.in,v 1.223 2007/08/15 13:04:54 drh Exp $
*/
#ifndef _SQLITE3_H_
#define _SQLITE3_H_
#include <stdarg.h>     /* Needed for the definition of va_list */

/*
** Make sure we can call this stuff from C++.







|







26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
** on how SQLite interfaces are suppose to operate.
**
** The name of this file under configuration management is "sqlite.h.in".
** The makefile makes some minor changes to this file (such as inserting
** the version number) and changes its name to "sqlite3.h" as
** part of the build process.
**
** @(#) $Id: sqlite.h.in,v 1.224 2007/08/15 17:08:46 danielk1977 Exp $
*/
#ifndef _SQLITE3_H_
#define _SQLITE3_H_
#include <stdarg.h>     /* Needed for the definition of va_list */

/*
** Make sure we can call this stuff from C++.
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
** to xWrite().
*/
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, sqlite_int64 iOfst);
  int (*xWrite)(sqlite3_file*, void*, int iAmt, sqlite_int64 iOfst);
  int (*xTruncate)(sqlite3_file*, sqlite_int64 size);
  int (*xSync)(sqlite3_file*, int flags);
  int (*xFileSize)(sqlite3_file*, sqlite_int64 *pSize);
  int (*xLock)(sqlite3_file*, int);
  int (*xUnlock)(sqlite3_file*, int);
  int (*xCheckReservedLock)(sqlite3_file*);
  int (*xBreakLock)(sqlite3_file*);







|







484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
** to xWrite().
*/
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, sqlite_int64 iOfst);
  int (*xWrite)(sqlite3_file*, const void*, int iAmt, sqlite_int64 iOfst);
  int (*xTruncate)(sqlite3_file*, sqlite_int64 size);
  int (*xSync)(sqlite3_file*, int flags);
  int (*xFileSize)(sqlite3_file*, sqlite_int64 *pSize);
  int (*xLock)(sqlite3_file*, int);
  int (*xUnlock)(sqlite3_file*, int);
  int (*xCheckReservedLock)(sqlite3_file*);
  int (*xBreakLock)(sqlite3_file*);
Changes to src/test2.c.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
**    May you share freely, never taking more than you give.
**
*************************************************************************
** Code for testing the pager.c module in SQLite.  This code
** is not included in the SQLite library.  It is used for automated
** testing of the SQLite library.
**
** $Id: test2.c,v 1.43 2007/04/02 05:07:47 danielk1977 Exp $
*/
#include "sqliteInt.h"
#include "os.h"
#include "pager.h"
#include "tcl.h"
#include <stdlib.h>
#include <string.h>







|







9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
**    May you share freely, never taking more than you give.
**
*************************************************************************
** Code for testing the pager.c module in SQLite.  This code
** is not included in the SQLite library.  It is used for automated
** testing of the SQLite library.
**
** $Id: test2.c,v 1.44 2007/08/15 17:08:46 danielk1977 Exp $
*/
#include "sqliteInt.h"
#include "os.h"
#include "pager.h"
#include "tcl.h"
#include <stdlib.h>
#include <string.h>
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
  rc = sqlite3OsOpenReadWrite(argv[2], &fd, &readOnly);
  if( rc ){
    Tcl_AppendResult(interp, "open failed: ", errorName(rc), 0);
    return TCL_ERROR;
  }
  offset = n;
  offset *= 1024*1024;
  rc = sqlite3OsSeek(fd, offset);
  if( rc ){
    Tcl_AppendResult(interp, "seek failed: ", errorName(rc), 0);
    return TCL_ERROR;
  }
  rc = sqlite3OsWrite(fd, "Hello, World!", 14);
  sqlite3OsClose(&fd);
  if( rc ){
    Tcl_AppendResult(interp, "write failed: ", errorName(rc), 0);
    return TCL_ERROR;
  }
  return TCL_OK;
}







<
<
<
<
<
|







542
543
544
545
546
547
548





549
550
551
552
553
554
555
556
  rc = sqlite3OsOpenReadWrite(argv[2], &fd, &readOnly);
  if( rc ){
    Tcl_AppendResult(interp, "open failed: ", errorName(rc), 0);
    return TCL_ERROR;
  }
  offset = n;
  offset *= 1024*1024;





  rc = sqlite3OsWrite(fd, "Hello, World!", 14, offset);
  sqlite3OsClose(&fd);
  if( rc ){
    Tcl_AppendResult(interp, "write failed: ", errorName(rc), 0);
    return TCL_ERROR;
  }
  return TCL_OK;
}
Changes to src/test6.c.
17
18
19
20
21
22
23
24

25
26
27
28
29
30
31
#if SQLITE_TEST          /* This file is used for the testing only */
#include "sqliteInt.h"
#include "os.h"
#include "tcl.h"

#ifndef SQLITE_OMIT_DISKIO  /* This file is a no-op if disk I/O is disabled */

typedef struct crashFile crashFile;

typedef struct WriteBuffer WriteBuffer;

/*
** Method:
**
**   This layer is implemented as a wrapper around the "real" 
**   sqlite3_file object for the host system. Each time data is 







|
>







17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
#if SQLITE_TEST          /* This file is used for the testing only */
#include "sqliteInt.h"
#include "os.h"
#include "tcl.h"

#ifndef SQLITE_OMIT_DISKIO  /* This file is a no-op if disk I/O is disabled */

typedef struct CrashFile CrashFile;
typedef struct CrashGlobal CrashGlobal;
typedef struct WriteBuffer WriteBuffer;

/*
** Method:
**
**   This layer is implemented as a wrapper around the "real" 
**   sqlite3_file object for the host system. Each time data is 
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154

155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
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
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491





492


493


494

495

496
497
498


499


500




501
502







503


504
505




506

507






508




509
510








511





512



513

514

515


516
517


518
519
520
521
522
523
524
525
526
527
528

529
530
531
532
533

534







535



















536
537









538

539
540
541
542

543
544


545

546

547
548
549

550
551
552


553






554


555


556



557







558



















559
560
561

562

563



564
565
566
567
568
569
570
571

572

573



574
575




576
577
578
579
580

581
582
583
584
585

586
587

588

589



590



591


592



593





594



595

596







597


598

















599

600
601


602
603



604
605
606
607

608
609
610
611
612

613
614
615

616
617




























618
619
620
621
622












623
624
625
626
627
628
629


630
631









































632










633
634
635
636
637


638
639
640
641
642
643
644


645
646

647







648
649

650



651
652
653

654













655
656
657
658
659
660
661
**   that occur before at least one write() on the file-handle specified
**   as part of the xSync() are written to their associated real files.
**
**   If IOCAP_SAFEAPPEND is set and the first byte written by the write()
**   operation is one byte past the current end of the file, then option
**   (1) is always selected.
*/
struct WriteBuffer {
  i64 iOffset;             /* Byte offset of the start of this write() */
  int nBuf;                /* Number of bytes written */
  u8 *zBuf;                /* Pointer to copy of written data */
  crashFile *pFile;        /* File this write() applies to */
};

/*
** crashFile is a subclass of OsFile that is taylored for 
** the crash test module.
*/
struct crashFile {
  IoMethod const *pMethod; /* Must be first */
  u8 **apBlk;              /* Array of blocks that have been written to. */
  int nBlk;                /* Size of apBlock. */
  i64 offset;              /* Next character to be read from the file */
  int nMaxWrite;           /* Largest offset written to. */
  char *zName;             /* File name */
  OsFile *pBase;           /* The real file */
  crashFile *pNext;        /* Next in a list of them all */
};

/*
** Size of a simulated disk block. Default is 512 bytes.
*/
static int BLOCKSIZE = 512;
#define BLOCK_OFFSET(x) ((x) * BLOCKSIZE)


/*
** The following variables control when a simulated crash occurs.
**
** If iCrashDelay is non-zero, then zCrashFile contains (full path) name of
** a file that SQLite will call sqlite3OsSync() on. Each time this happens
** iCrashDelay is decremented. If iCrashDelay is zero after being
** decremented, a "crash" occurs during the sync() operation.
**
** In other words, a crash occurs the iCrashDelay'th time zCrashFile is
** synced.
*/
static int iCrashDelay = 0;
static char zCrashFile[500];

/*
** A list of all open files.
*/
static crashFile *pAllFiles = 0;

/*

** Set the value of the two crash parameters.
*/
static void setCrashParams(int iDelay, char const *zFile){
  sqlite3OsEnterMutex();
  assert( strlen(zFile)<sizeof(zCrashFile) );
  strcpy(zCrashFile, zFile);
  iCrashDelay = iDelay;
  sqlite3OsLeaveMutex();
}

/*
** Set the value of the simulated disk block size.
*/
static void setBlocksize(int iBlockSize){
  sqlite3OsEnterMutex();
  assert( !pAllFiles );
  BLOCKSIZE = iBlockSize;
  sqlite3OsLeaveMutex();
}

/*
** File zPath is being sync()ed. Return non-zero if this should
** cause a crash.
*/
static int crashRequired(char const *zPath){
  int r;
  int n;
  sqlite3OsEnterMutex();
  n = strlen(zCrashFile);
  if( zCrashFile[n-1]=='*' ){
    n--;
  }else if( strlen(zPath)>n ){
    n = strlen(zPath);
  }
  r = 0;
  if( iCrashDelay>0 && strncmp(zPath, zCrashFile, n)==0 ){
    iCrashDelay--;
    if( iCrashDelay<=0 ){
      r = 1;
    }
  }
  sqlite3OsLeaveMutex();
  return r;
}


/* Forward reference */
static void initFile(OsFile **pId, char const *zName, OsFile *pBase);

/*
** Undo the work done by initFile. Delete the OsFile structure
** and unlink the structure from the pAllFiles list.
*/
static void closeFile(crashFile **pId){
  crashFile *pFile = *pId;
  if( pFile==pAllFiles ){
    pAllFiles = pFile->pNext;
  }else{
    crashFile *p;
    for(p=pAllFiles; p->pNext!=pFile; p=p->pNext ){
      assert( p );
    }
    p->pNext = pFile->pNext;
  }
  sqliteFree(*pId);
  *pId = 0;
}

/*
** Read block 'blk' off of the real disk file and into the cache of pFile.
*/
static int readBlockIntoCache(crashFile *pFile, int blk){
  if( blk>=pFile->nBlk ){
    int n = ((pFile->nBlk * 2) + 100 + blk);
    /* if( pFile->nBlk==0 ){ printf("DIRTY %s\n", pFile->zName); } */
    pFile->apBlk = (u8 **)sqliteRealloc(pFile->apBlk, n * sizeof(u8*));
    if( !pFile->apBlk ) return SQLITE_NOMEM;
    memset(&pFile->apBlk[pFile->nBlk], 0, (n - pFile->nBlk)*sizeof(u8*));
    pFile->nBlk = n;
  }

  if( !pFile->apBlk[blk] ){
    i64 filesize;
    int rc;

    u8 *p = sqliteMalloc(BLOCKSIZE);
    if( !p ) return SQLITE_NOMEM;
    pFile->apBlk[blk] = p;

    rc = sqlite3OsFileSize(pFile->pBase, &filesize);
    if( rc!=SQLITE_OK ) return rc;

    if( BLOCK_OFFSET(blk)<filesize ){
      int len = BLOCKSIZE;
      rc = sqlite3OsSeek(pFile->pBase, blk*BLOCKSIZE);
      if( BLOCK_OFFSET(blk+1)>filesize ){
        len = filesize - BLOCK_OFFSET(blk);
      }
      if( rc!=SQLITE_OK ) return rc;
      rc = sqlite3OsRead(pFile->pBase, p, len);
      if( rc!=SQLITE_OK ) return rc;
    }
  }

  return SQLITE_OK;
}

/*
** Write the cache of pFile to disk. If crash is non-zero, randomly
** skip blocks when writing. The cache is deleted before returning.
*/
static int writeCache2(crashFile *pFile, int crash){
  int i;
  int nMax = pFile->nMaxWrite;
  int rc = SQLITE_OK;

  for(i=0; i<pFile->nBlk; i++){
    u8 *p = pFile->apBlk[i];
    if( p ){
      int skip = 0;
      int trash = 0;
      if( crash ){
        char random;
        sqlite3Randomness(1, &random);
        if( random & 0x01 ){
          if( random & 0x02 ){
            trash = 1;
#ifdef TRACE_WRITECACHE
printf("Trashing block %d of %s\n", i, pFile->zName); 
#endif
          }else{
            skip = 1;
#ifdef TRACE_WRITECACHE
printf("Skiping block %d of %s\n", i, pFile->zName); 
#endif
          }
        }else{
#ifdef TRACE_WRITECACHE
printf("Writing block %d of %s\n", i, pFile->zName); 
#endif
        }
      }
      if( rc==SQLITE_OK ){
        rc = sqlite3OsSeek(pFile->pBase, BLOCK_OFFSET(i));
      }
      if( rc==SQLITE_OK && !skip ){
        int len = BLOCKSIZE;
        if( BLOCK_OFFSET(i+1)>nMax ){
          len = nMax-BLOCK_OFFSET(i);
        }
        if( len>0 ){
          if( trash ){
            sqlite3Randomness(len, p);
          }
          rc = sqlite3OsWrite(pFile->pBase, p, len);
        }
      }
      sqliteFree(p);
    }
  }
  sqliteFree(pFile->apBlk);
  pFile->nBlk = 0;
  pFile->apBlk = 0;
  pFile->nMaxWrite = 0;
  return rc;
}

/*
** Write the cache to disk.
*/
static int writeCache(crashFile *pFile){
  if( pFile->apBlk ){
    int c = crashRequired(pFile->zName);
    if( c ){
      crashFile *p;
#ifdef TRACE_WRITECACHE
      printf("\nCrash during sync of %s\n", pFile->zName);
#endif
      for(p=pAllFiles; p; p=p->pNext){
        writeCache2(p, 1);
      }
      exit(-1);
    }else{
      return writeCache2(pFile, 0);
    }
  }
  return SQLITE_OK;
}

/*
** Close the file.
*/
static int crashClose(OsFile **pId){
  crashFile *pFile = (crashFile*)*pId;
  if( pFile ){
    /* printf("CLOSE %s (%d blocks)\n", pFile->zName, pFile->nBlk); */
    writeCache(pFile);
    sqlite3OsClose(&pFile->pBase);
  }
  closeFile(&pFile);
  *pId = 0;
  return SQLITE_OK;
}

static int crashSeek(OsFile *id, i64 offset){
  ((crashFile*)id)->offset = offset;
  return SQLITE_OK;
}

static int crashRead(OsFile *id, void *pBuf, int amt){
  i64 offset;       /* The current offset from the start of the file */
  i64 end;          /* The byte just past the last byte read */
  int blk;            /* Block number the read starts on */
  int i;
  u8 *zCsr;
  int rc = SQLITE_OK;
  crashFile *pFile = (crashFile*)id;

  offset = pFile->offset;
  end = offset+amt;
  blk = (offset/BLOCKSIZE);

  zCsr = (u8 *)pBuf;
  for(i=blk; i*BLOCKSIZE<end; i++){
    int off = 0;
    int len = 0;


    if( BLOCK_OFFSET(i) < offset ){
      off = offset-BLOCK_OFFSET(i);
    }
    len = BLOCKSIZE - off;
    if( BLOCK_OFFSET(i+1) > end ){
      len = len - (BLOCK_OFFSET(i+1)-end);
    }

    if( i<pFile->nBlk && pFile->apBlk[i]){
      u8 *pBlk = pFile->apBlk[i];
      memcpy(zCsr, &pBlk[off], len);
    }else{
      rc = sqlite3OsSeek(pFile->pBase, BLOCK_OFFSET(i) + off);
      if( rc!=SQLITE_OK ) return rc;
      rc = sqlite3OsRead(pFile->pBase, zCsr, len);
      if( rc!=SQLITE_OK ) return rc;
    }

    zCsr += len;
  }
  assert( zCsr==&((u8 *)pBuf)[amt] );

  pFile->offset = end;
  return rc;
}

static int crashWrite(OsFile *id, const void *pBuf, int amt){
  i64 offset;       /* The current offset from the start of the file */
  i64 end;          /* The byte just past the last byte written */
  int blk;            /* Block number the write starts on */
  int i;
  const u8 *zCsr;
  int rc = SQLITE_OK;
  crashFile *pFile = (crashFile*)id;

  offset = pFile->offset;
  end = offset+amt;
  blk = (offset/BLOCKSIZE);

  zCsr = (u8 *)pBuf;
  for(i=blk; i*BLOCKSIZE<end; i++){
    u8 *pBlk;
    int off = 0;
    int len = 0;

    /* Make sure the block is in the cache */
    rc = readBlockIntoCache(pFile, i);
    if( rc!=SQLITE_OK ) return rc;

    /* Write into the cache */
    pBlk = pFile->apBlk[i];
    assert( pBlk );

    if( BLOCK_OFFSET(i) < offset ){
      off = offset-BLOCK_OFFSET(i);
    }
    len = BLOCKSIZE - off;
    if( BLOCK_OFFSET(i+1) > end ){
      len = len - (BLOCK_OFFSET(i+1)-end);
    }
    memcpy(&pBlk[off], zCsr, len);
    zCsr += len;
  }
  if( pFile->nMaxWrite<end ){
    pFile->nMaxWrite = end;
  }
  assert( zCsr==&((u8 *)pBuf)[amt] );
  pFile->offset = end;
  return rc;
}

/*
** Sync the file. First flush the write-cache to disk, then call the
** real sync() function.
*/
static int crashSync(OsFile *id, int dataOnly){
  return writeCache((crashFile*)id);
}

/*
** Truncate the file. Set the internal OsFile.nMaxWrite variable to the new
** file size to ensure that nothing in the write-cache past this point
** is written to disk.
*/
static int crashTruncate(OsFile *id, i64 nByte){
  crashFile *pFile = (crashFile*)id;
  pFile->nMaxWrite = nByte;
  return sqlite3OsTruncate(pFile->pBase, nByte);
}

/*
** Return the size of the file. If the cache contains a write that extended
** the file, then return this size instead of the on-disk size.
*/
static int crashFileSize(OsFile *id, i64 *pSize){
  crashFile *pFile = (crashFile*)id;
  int rc = sqlite3OsFileSize(pFile->pBase, pSize);
  if( rc==SQLITE_OK && pSize && *pSize<pFile->nMaxWrite ){
    *pSize = pFile->nMaxWrite;
  }
  return rc;
}

/*
** Set this global variable to 1 to enable crash testing.
*/
int sqlite3CrashTestEnable = 0;

/*
** The three functions used to open files. All that is required is to





** initialise the os_test.c specific fields and then call the corresponding


** os_unix.c function to really open the file.


*/

int sqlite3CrashOpenReadWrite(const char *zFilename, OsFile **pId,int *pRdonly){

  OsFile *pBase = 0;
  int rc;



  sqlite3CrashTestEnable = 0;


  rc = sqlite3OsOpenReadWrite(zFilename, &pBase, pRdonly);




  sqlite3CrashTestEnable = 1;
  if( !rc ){







    initFile(pId, zFilename, pBase);


  }
  return rc;




}

int sqlite3CrashOpenExclusive(const char *zFilename, OsFile **pId, int delFlag){






  OsFile *pBase = 0;




  int rc;









  sqlite3CrashTestEnable = 0;





  rc = sqlite3OsOpenExclusive(zFilename, &pBase, delFlag);



  sqlite3CrashTestEnable = 1;

  if( !rc ){

    initFile(pId, zFilename, pBase);


  }
  return rc;


}
int sqlite3CrashOpenReadOnly(const char *zFilename, OsFile **pId, int NotUsed){
  OsFile *pBase = 0;
  int rc;

  sqlite3CrashTestEnable = 0;
  rc = sqlite3OsOpenReadOnly(zFilename, &pBase);
  sqlite3CrashTestEnable = 1;
  if( !rc ){
    initFile(pId, zFilename, pBase);
  }

  return rc;
}

/*
** OpenDirectory is a no-op

*/







static int crashOpenDir(OsFile *id, const char *zName){



















  return SQLITE_OK;
}











/*
** Locking primitives are passed through into the underlying
** file descriptor.
*/

int crashLock(OsFile *id, int lockType){
  return sqlite3OsLock(((crashFile*)id)->pBase, lockType);


}

int crashUnlock(OsFile *id, int lockType){

  return sqlite3OsUnlock(((crashFile*)id)->pBase, lockType);
}
int crashCheckReservedLock(OsFile *id){

  return sqlite3OsCheckReservedLock(((crashFile*)id)->pBase);
}
void crashSetFullSync(OsFile *id, int setting){


  return;  /* This is a no-op */






}


int crashLockState(OsFile *id){


  return sqlite3OsLockState(((crashFile*)id)->pBase);



}



























/*
** Return the underlying file handle.
*/

int crashFileHandle(OsFile *id){

#if defined(SQLITE_TEST) || defined(SQLITE_DEBUG)



  return sqlite3OsFileHandle(((crashFile*)id)->pBase);
#endif
  return 0;
}

/*
** Return the simulated file-system sector size.
*/

int crashSectorSize(OsFile *id){

  return BLOCKSIZE;



}





/*
** This vector defines all the methods that can operate on an OsFile
** for the crash tester.
*/
static const IoMethod crashIoMethod = {

  crashClose,
  crashOpenDir,
  crashRead,
  crashWrite,
  crashSeek,

  crashTruncate,
  crashSync,

  crashSetFullSync,

  crashFileHandle,



  crashFileSize,



  crashLock,


  crashUnlock,



  crashLockState,





  crashCheckReservedLock,



  crashSectorSize,

};




























/*

** Initialise the os_test.c specific fields of pFile.
*/


static void initFile(OsFile **pId, char const *zName, OsFile *pBase){
  crashFile *pFile = sqliteMalloc(sizeof(crashFile) + strlen(zName)+1);



  pFile->pMethod = &crashIoMethod;
  pFile->nMaxWrite = 0; 
  pFile->offset = 0;
  pFile->nBlk = 0; 

  pFile->apBlk = 0; 
  pFile->zName = (char *)(&pFile[1]);
  strcpy(pFile->zName, zName);
  pFile->pBase = pBase;
  pFile->pNext = pAllFiles;

  pAllFiles = pFile;
  *pId = (OsFile*)pFile;
}































/*
** tclcmd:   sqlite_crashparams DELAY CRASHFILE ?BLOCKSIZE?
**
** This procedure implements a TCL command that enables crash testing
** in testfixture.  Once enabled, crash testing cannot be disabled.












*/
static int crashParamsObjCmd(
  void * clientData,
  Tcl_Interp *interp,
  int objc,
  Tcl_Obj *CONST objv[]
){


  int iDelay;
  const char *zFile;









































  int nFile;











  if( objc!=3 && objc!=4 ){
    Tcl_WrongNumArgs(interp, 1, objv, "DELAY CRASHFILE ?BLOCKSIZE?");
    return TCL_ERROR;
  }


  if( Tcl_GetIntFromObj(interp, objv[1], &iDelay) ) return TCL_ERROR;
  zFile = Tcl_GetStringFromObj(objv[2], &nFile);
  if( nFile>=sizeof(zCrashFile)-1 ){
    Tcl_AppendResult(interp, "crash file name too big", 0);
    return TCL_ERROR;
  }
  setCrashParams(iDelay, zFile);


  if( objc==4 ){
    int iBlockSize = 0;

    if( Tcl_GetIntFromObj(interp, objv[3], &iBlockSize) ) return TCL_ERROR;







    if( pAllFiles ){
      char *zErr = "Cannot modify blocksize after opening files";

      Tcl_SetResult(interp, zErr, TCL_STATIC);



      return TCL_ERROR;
    }
    setBlocksize(iBlockSize);

  }













  sqlite3CrashTestEnable = 1;
  return TCL_OK;
}

#endif /* SQLITE_OMIT_DISKIO */

/*







<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<


>
|
|
<
|
<
<
<
<
<
|
<
|

<
|
<
<
<
<
|
<
<
<
<
<
<
|
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
<
|
>
<
<

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

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

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







|
>
>
>
>
>
|
>
>
|
>
>
|
>
|
>
|
|
|
>
>
|
>
>
|
>
>
>
>
|
|
>
>
>
>
>
>
>
|
>
>
|
|
>
>
>
>
|
>
|
>
>
>
>
>
>
|
>
>
>
>
|
|
>
>
>
>
>
>
>
>
|
>
>
>
>
>
|
>
>
>
|
>
|
>
|
>
>
|
|
>
>
|
|
<
|
|
|
<
<
|
|

>




<
>

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


>
>
>
>
>
>
>
>
>
|
>

<
|

>
|
|
>
>
|
>
|
>
|
|
<
>
|
|
|
>
>
|
>
>
>
>
>
>
|
>
>
|
>
>
|
>
>
>
|
>
>
>
>
>
>
>
|
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>

|

>
|
>
|
>
>
>
|
<
|



|

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

<
|

<
>
|
<
<
|
|
>
|
|
>
|
>
|
>
>
>
|
>
>
>
|
>
>
|
>
>
>
|
>
>
>
>
>
|
>
>
>
|
>
|
>
>
>
>
>
>
>
|
>
>
|
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>

>
|

>
>
|
|
>
>
>
|
|
|
|
>
|
|
|
<
<
>
|
<
|
>
|
|
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>

|



>
>
>
>
>
>
>
>
>
>
>
>







>
>

|
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
|
>
>
>
>
>
>
>
>
>
>
|
|
|
|
|
>
>
|
<
<
<
|
|
|
>
>
|
|
>
|
>
>
>
>
>
>
>
|
<
>
|
>
>
>
|
|
|
>
|
>
>
>
>
>
>
>
>
>
>
>
>
>







100
101
102
103
104
105
106















































107
108
109
110
111

112





113

114
115

116




117






118

















119
120


121


















122











123





























































124
































125













126








127





128



129
130








131



132













133









134



135


136
137








138




















139



140










141































142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
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
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242

243
244
245


246
247
248
249
250
251
252
253

254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296

297
298
299
300
301
302
303
304
305
306
307
308
309

310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371

372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391

392
393

394
395


396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478


479
480

481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600



601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617

618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
**   that occur before at least one write() on the file-handle specified
**   as part of the xSync() are written to their associated real files.
**
**   If IOCAP_SAFEAPPEND is set and the first byte written by the write()
**   operation is one byte past the current end of the file, then option
**   (1) is always selected.
*/
















































/*
** Each write operation in the write-list is represented by an instance
** of the following structure.
**

** If zBuf is 0, then this structure represents a call to xTruncate(), 





** not xWrite(). In that case, iOffset is the size that the file is

** truncated to.
*/

struct WriteBuffer {




  i64 iOffset;                 /* Byte offset of the start of this write() */






  int nBuf;                    /* Number of bytes written */

















  u8 *zBuf;                    /* Pointer to copy of written data */
  CrashFile *pFile;            /* File this write() applies to */





















  WriteBuffer *pNext;          /* Next in CrashGlobal.pWriteList */











};






























































































struct CrashFile {













  const sqlite3_io_methods *pMethod;   /* Must be first */








  sqlite3_file *pRealFile;             /* Underlying "real" file handle */





  const char *zName;



};









struct CrashGlobal {



  WriteBuffer *pWriteList;     /* Head of write-list */























  int iSectorSize;             /* Value of simulated sector size */



  int iDeviceCharacteristics;  /* Value of simulated device characteristics */



  int iCrash;                  /* Crash on the iCrash'th call to xSync() */








  char zCrashFile[500];        /* Crash during an xSync() on this file */ 




















};














static CrashGlobal g = {0, SQLITE_DEFAULT_SECTOR_SIZE, 0, 0};
































/*
** Set this global variable to 1 to enable crash testing.
*/
int sqlite3CrashTestEnable = 0;

/*
** Flush the write-list as if xSync() had been called on file handle
** pFile. If isCrash is true, simulate a crash.
*/
static int writeListSync(CrashFile *pFile, int isCrash){
  int rc = SQLITE_OK;
  int iDc = g.iDeviceCharacteristics;

  WriteBuffer *pWrite;
  WriteBuffer **ppPtr;

  /* Set pFinal to point to the last element of the write-list that
  ** is associated with file handle pFile.
  */
  WriteBuffer *pFinal = 0;
  if( !isCrash ){
    for(pWrite=g.pWriteList; pWrite; pWrite=pWrite->pNext){
      if( pWrite->pFile==pFile ){
        pFinal = pWrite;
      }
    }
  }

  ppPtr = &g.pWriteList;
  for(pWrite=*ppPtr; rc==SQLITE_OK && pWrite; pWrite=*ppPtr){

    /* (eAction==1)      -> write block out normally,
    ** (eAction==2)      -> do nothing,
    ** (eAction==3)      -> trash sectors.
    */
    int eAction = 0;
    if( !isCrash ){
      eAction = 2;
      if( (pWrite->pFile==pFile || iDc&SQLITE_IOCAP_SEQUENTIAL) ){
        eAction = 1;
      }
    }else{
      char random;
      sqlite3Randomness(1, &random);

      if( iDc&SQLITE_IOCAP_ATOMIC || pWrite->zBuf==0 ){
        random &= 0x01;
      }

      if( (random&0x06)==0x06 ){
        eAction = 3;
      }else{
        eAction = ((random&0x01)?2:1);
      }
    }

    switch( eAction ){
      case 1: {               /* Write out correctly */
        if( pWrite->zBuf ){
          rc = sqlite3OsWrite(
              pFile->pRealFile, pWrite->zBuf, pWrite->nBuf, pWrite->iOffset
          );
        }else{
          rc = sqlite3OsTruncate(pFile->pRealFile, pWrite->iOffset);
        }
        *ppPtr = pWrite->pNext;
        sqliteFree(pWrite);
        break;
      }
      case 2: {               /* Do nothing */
        ppPtr = &pWrite->pNext;
        break;
      }
      case 3: {               /* Trash sectors */
        u8 *zGarbage;
        sqlite3_int64 iFirst = (pWrite->iOffset%g.iSectorSize);
        sqlite3_int64 iLast = (pWrite->iOffset+pWrite->nBuf-1)%g.iSectorSize;

        zGarbage = sqliteMalloc(g.iSectorSize);
        if( zGarbage ){
          sqlite3_int64 i;
          for(i=iFirst; rc==SQLITE_OK && i<=iLast; i++){
            sqlite3Randomness(g.iSectorSize, zGarbage); 
            rc = sqlite3OsWrite(
              pFile->pRealFile, i*g.iSectorSize, zGarbage, g.iSectorSize
            );
          }
          sqliteFree(zGarbage);
        }else{
          rc = SQLITE_NOMEM;
        }

        ppPtr = &pWrite->pNext;
        break;
      }

      default:
        assert(!"Cannot happen");
    }


    if( pWrite==pFinal ) break;
  }



  if( rc==SQLITE_OK && isCrash ){
    exit(-1);
  }

  return rc;
}

/*

** Add an entry to the end of the write-list.
*/
static int writeListAppend(
  sqlite3_file *pFile,
  sqlite3_int64 iOffset,
  const u8 *zBuf,
  int nBuf
){
  WriteBuffer *pNew;

  assert((zBuf && nBuf) || (!nBuf && !zBuf));

  pNew = (WriteBuffer *)sqliteMalloc(sizeof(WriteBuffer) + nBuf);
  pNew->iOffset = iOffset;
  pNew->nBuf = nBuf;
  pNew->pFile = (CrashFile *)pFile;
  if( zBuf ){
    pNew->zBuf = (u8 *)&pNew[1];
    memcpy(pNew->zBuf, zBuf, nBuf);
  }

  if( g.pWriteList ){
    WriteBuffer *pEnd;
    for(pEnd=g.pWriteList; pEnd->pNext; pEnd=pEnd->pNext);
    pEnd->pNext = pNew;
  }else{
    g.pWriteList = pNew;
  }
  
  return SQLITE_OK;
}

/*
** Close a crash-file.
*/
int cfClose(sqlite3_file *pFile){
  CrashFile *pCrash = (CrashFile *)pFile;
  writeListSync(pCrash, 0);
  sqlite3OsClose(&pCrash->pRealFile);
  return SQLITE_OK;
}

/*

** Read data from a crash-file.
*/
int cfRead(sqlite3_file *pFile, void *zBuf, int iAmt, sqlite_int64 iOfst){
  CrashFile *pCrash = (CrashFile *)pFile;
  sqlite3_int64 iSize;
  int rc;
  WriteBuffer *pWrite;

  /* Check the file-size to see if this is a short-read */
  rc = sqlite3OsFileSize(pFile, &iSize);
  if( rc!=SQLITE_OK ){
    return rc;
  }

  if( iSize<(iOfst+iAmt) ){
    return SQLITE_IOERR_SHORT_READ;
  }

  /* Zero the output buffer */
  memset(zBuf, 0, iAmt);

  /* Read some data from the real file */
  rc = sqlite3OsFileSize(pCrash->pRealFile, &iSize);
  if( rc==SQLITE_OK && iSize>iOfst ){
    int nRead = iAmt;
    if( iSize<(iOfst+iAmt) ){
      nRead = iSize - iOfst;
    }
    rc = sqlite3OsRead(pCrash->pRealFile, zBuf, nRead, iOfst);
  }

  /* Fill in the buffer by traversing the write-list */
  for(pWrite=g.pWriteList; rc==SQLITE_OK && pWrite; pWrite=pWrite->pNext){
    if( pWrite->pFile==pCrash ){
      int iWriteOffset;
      int nWriteBuf;
      u8 *zWriteBuf;

      iWriteOffset = pWrite->iOffset - iOfst;
      nWriteBuf = pWrite->nBuf;
      zWriteBuf = pWrite->zBuf;
      if( iWriteOffset<0 ){
        nWriteBuf += iWriteOffset;
        zWriteBuf -= iWriteOffset;
        iWriteOffset = 0;
      }
      if( (iWriteOffset+nWriteBuf)>iAmt ){
        nWriteBuf = iAmt - iWriteOffset;
      }
      
      if( pWrite->zBuf && nWriteBuf>0){
        /* Copy data to the buffer */
        memcpy(&((u8 *)zBuf)[iWriteOffset], zWriteBuf, nWriteBuf);
      }

      if( pWrite->zBuf==0 && iWriteOffset<iAmt ){
        /* Zero part of the buffer to simulate a truncate */
        memset(&((u8 *)zBuf)[iWriteOffset], 0, iAmt-iWriteOffset);
      }
    }
  }

  return rc;
}

/*
** Write data to a crash-file.
*/
int cfWrite(sqlite3_file *pFile, void *zBuf, int iAmt, sqlite_int64 iOfst){
  return writeListAppend(pFile, iOfst, zBuf, iAmt);
}

/*
** Truncate a crash-file.
*/
int cfTruncate(sqlite3_file *pFile, sqlite_int64 size){

  return writeListAppend(pFile, size, 0, 0);
}

/*
** Sync a crash-file.
*/
int cfSync(sqlite3_file *pFile, int flags){
  CrashFile *pCrash = (CrashFile *)pFile;
  int isCrash = 0;

  if( 0==strcmp(pCrash->zName, g.zCrashFile) ){
    if( (--g.iCrash==0) ){
      isCrash = 1;
    }
  }

  return writeListSync(pCrash, isCrash);
}

/*

** Return the current file-size of the crash-file.
*/

int cfFileSize(sqlite3_file *pFile, sqlite_int64 *pSize){
  CrashFile *pCrash = (CrashFile *)pFile;


  WriteBuffer *pWrite;
  int rc;
  sqlite_int64 iSize;

  rc = sqlite3OsFileSize(pCrash->pRealFile, &iSize);
  if( rc!=SQLITE_OK ){
    return rc;
  }

  for(pWrite=g.pWriteList; pWrite; pWrite=pWrite->pNext){
    sqlite_int64 iEnd = pWrite->nBuf+pWrite->iOffset;
    if( pWrite->pFile==pCrash && (pWrite->zBuf==0 || iEnd>iSize) ){
      iSize = iEnd;
    }
  }
  *pSize = iSize;

  return SQLITE_OK;
}

/*
** Calls related to file-locks are passed on to the real file handle.
*/
int cfLock(sqlite3_file *pFile, int eLock){
  return sqlite3OsLock(((CrashFile *)pFile)->pRealFile, eLock);
}
int cfUnlock(sqlite3_file *pFile, int eLock){
  return sqlite3OsUnlock(((CrashFile *)pFile)->pRealFile, eLock);
}
int cfCheckReservedLock(sqlite3_file *pFile){
  return sqlite3OsCheckReservedLock(((CrashFile *)pFile)->pRealFile);
}
int cfBreakLock(sqlite3_file *pFile){
  return sqlite3OsBreakLock(((CrashFile *)pFile)->pRealFile);
}

/*
** The xSectorSize() and xDeviceCharacteristics() functions return
** the global values configured by the [sqlite_crashparams] tcl
*  interface.
*/
int cfSectorSize(sqlite3_file *pFile){
  return g.iSectorSize;
}
int cfDeviceCharacteristics(sqlite3_file *pFile){
  return g.iDeviceCharacteristics;
}

static const sqlite3_io_methods CrashFileVtab = {
  1,                            /* iVersion */
  cfClose,                      /* xClose */
  cfRead,                       /* xRead */
  cfWrite,                      /* xWrite */
  cfTruncate,                   /* xTruncate */
  cfSync,                       /* xSync */
  cfFileSize,                   /* xFileSize */
  cfLock,                       /* xLock */
  cfUnlock,                     /* xUnlock */
  cfCheckReservedLock,          /* xCheckReservedLock */
  cfBreakLock,                  /* xBreakLock */
  cfSectorSize,                 /* xSectorSize */
  cfDeviceCharacteristics       /* xDeviceCharacteristics */
};

/*
** Open a crash-file file handle. The vfs pVfs is used to open
** the underlying real file.
*/
int sqlite3CrashFileOpen(
  sqlite3_vfs *pVfs,
  const char *zName,
  sqlite3_file *pFile,
  int flags,
  int *pOutFlags
){
  CrashFile *pWrapper = (CrashFile *)pFile;
  int rc = SQLITE_NOMEM;
  sqlite3_file *pReal;
  pReal = (sqlite3_file *)sqliteMalloc(pVfs->szOsFile);
  if( pReal ){
    pWrapper->pMethod = &CrashFileVtab;
    pWrapper->zName = zName;
    rc = pVfs->xOpen(pVfs->pAppData, zName, pReal, flags, pOutFlags);


    if( rc==SQLITE_OK ){
      pWrapper->pRealFile = pFile;

    }else{
      sqliteFree(pReal);
    }
  }
  return rc;
}

int sqlite3CrashFileWrap(
  sqlite3_file *pFile,
  const char *zName,
  sqlite3_file **ppWrapper
){
  CrashFile *pWrapper;
  pWrapper = (CrashFile *)sqliteMalloc(sizeof(CrashFile)+strlen(zName)+1);
  if( !pWrapper ){
    return SQLITE_NOMEM;
  }

  pWrapper->pMethod = &CrashFileVtab;
  pWrapper->pRealFile = pFile;
  pWrapper->zName = &pWrapper[1];
  memcpy(pWrapper->zName, zName, strlen(zName)+1);

  *ppWrapper = (sqlite3_file *)pWrapper;
  return SQLITE_OK;
}


int sqlite3CrashFileSize(){
  return (int)sizeof(CrashFile);
}

/*
** tclcmd:   sqlite_crashparams ?OPTIONS? DELAY CRASHFILE
**
** This procedure implements a TCL command that enables crash testing
** in testfixture.  Once enabled, crash testing cannot be disabled.
**
** Available options are "-characteristics" and "-sectorsize". Both require
** an argument. For -sectorsize, this is the simulated sector size in
** bytes. For -characteristics, the argument must be a list of io-capability
** flags to simulate. Valid flags are "atomic", "atomic512", "atomic1K",
** "atomic2K", "atomic4K", "atomic8K", "atomic16K", "atomic32K", 
** "atomic64K", "sequential" and "safe_append".
**
** Example:
**
**   sqlite_crashparams -sect 1024 -char {atomic sequential} ./test.db 1
**
*/
static int crashParamsObjCmd(
  void * clientData,
  Tcl_Interp *interp,
  int objc,
  Tcl_Obj *CONST objv[]
){
  int i;

  int iDelay;
  const char *zCrashFile;
  int nCrashFile;

  int iDc = 0;
  int iSectorSize = 0;
  int setSectorsize = 0;
  int setDeviceChar = 0;

  struct DeviceFlag {
    char *zName;
    int iValue;
  } aFlag[] = {
    { "atomic",      SQLITE_IOCAP_ATOMIC      },
    { "atomic512",   SQLITE_IOCAP_ATOMIC512   },
    { "atomic1k",    SQLITE_IOCAP_ATOMIC1K    },
    { "atomic2k",    SQLITE_IOCAP_ATOMIC2K    },
    { "atomic4k",    SQLITE_IOCAP_ATOMIC4K    },
    { "atomic8k",    SQLITE_IOCAP_ATOMIC8K    },
    { "atomic16k",   SQLITE_IOCAP_ATOMIC16K   },
    { "atomic32k",   SQLITE_IOCAP_ATOMIC32K   },
    { "atomic64k",   SQLITE_IOCAP_ATOMIC64K   },
    { "sequential",  SQLITE_IOCAP_SEQUENTIAL  },
    { "safe_append", SQLITE_IOCAP_SAFE_APPEND },
    { 0, 0 }
  };
  
  if( objc<3 ){
    Tcl_WrongNumArgs(interp, 1, objv, "?OPTIONS? DELAY CRASHFILE");
    return TCL_ERROR;
  }

  zCrashFile = sqlite3OsFullPathname(Tcl_GetString(objv[objc-1]));
  nCrashFile = strlen(zCrashFile);
  if( nCrashFile>=sizeof(g.zCrashFile) ){
    Tcl_AppendResult(interp, "Filename is too long: \"", zCrashFile, "\"", 0);
    return TCL_ERROR;
  }
  if( Tcl_GetIntFromObj(interp, objv[objc-2], &iDelay) ){
    return TCL_ERROR;
  }

  for(i=1; i<(objc-2); i+=2){
    int nOpt;
    char *zOpt = Tcl_GetStringFromObj(objv[i], &nOpt);

    if( (nOpt>11 || nOpt<2 || strncmp("-sectorsize", zOpt, nOpt)) 
     && (nOpt>16 || nOpt<2 || strncmp("-characteristics", zOpt, nOpt))
    ){
      Tcl_AppendResult(interp, 
        "Bad option: \"", zOpt, 
        "\" - must be \"-characteristics\" or \"-sectorsize\"", 0
      );
      return TCL_ERROR;
    }
    if( i==objc-3 ){
      Tcl_AppendResult(interp, "Option requires an argument: \"", zOpt, "\"",0);
      return TCL_ERROR;
    }

    if( zOpt[1]=='s' ){
      if( Tcl_GetIntFromObj(interp, objv[i+1], &iSectorSize) ){



        return TCL_ERROR;
      }
      setSectorsize = 1;
    }else{
      int j;
      Tcl_Obj **apObj;
      int nObj;
      if( Tcl_ListObjGetElements(interp, objv[i+1], &nObj, &apObj) ){
        return TCL_ERROR;
      }
      for(j=0; j<nObj; j++){
        int rc;
        int iChoice;
        Tcl_Obj *pFlag = Tcl_DuplicateObj(apObj[j]);
        Tcl_IncrRefCount(pFlag);
        Tcl_UtfToLower(Tcl_GetString(pFlag));
 

        rc = Tcl_GetIndexFromObjStruct(
            interp, pFlag, aFlag, sizeof(aFlag[0]), "no such flag", 0, &iChoice
        );
        Tcl_DecrRefCount(pFlag);
        if( rc ){
          return TCL_ERROR;
        }

        iDc |= aFlag[iChoice].iValue;
      }
      setDeviceChar = 1;
    }
  }

  if( setDeviceChar ){
    g.iDeviceCharacteristics = iDc;
  }
  if( setSectorsize ){
    g.iSectorSize = iSectorSize;
  }
  g.iCrash = iDelay;
  memcpy(g.zCrashFile, zCrashFile, nCrashFile+1);

  sqlite3CrashTestEnable = 1;
  return TCL_OK;
}

#endif /* SQLITE_OMIT_DISKIO */

/*
Changes to src/vdbeaux.c.
1125
1126
1127
1128
1129
1130
1131
1132

1133
1134
1135
1136
1137
1138
1139
  ** committed atomicly.
  */
#ifndef SQLITE_OMIT_DISKIO
  else{
    int needSync = 0;
    char *zMaster = 0;   /* File-name for the master journal */
    char const *zMainFile = sqlite3BtreeGetFilename(db->aDb[0].pBt);
    OsFile *master = 0;


    /* Select a master journal file name */
    do {
      u32 random;
      sqliteFree(zMaster);
      sqlite3Randomness(sizeof(random), &random);
      zMaster = sqlite3MPrintf("%s-mj%08X", zMainFile, random&0x7fffffff);







|
>







1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
  ** committed atomicly.
  */
#ifndef SQLITE_OMIT_DISKIO
  else{
    int needSync = 0;
    char *zMaster = 0;   /* File-name for the master journal */
    char const *zMainFile = sqlite3BtreeGetFilename(db->aDb[0].pBt);
    sqlite3_file *master = 0;
    i64 offset = 0;

    /* Select a master journal file name */
    do {
      u32 random;
      sqliteFree(zMaster);
      sqlite3Randomness(sizeof(random), &random);
      zMaster = sqlite3MPrintf("%s-mj%08X", zMainFile, random&0x7fffffff);
1160
1161
1162
1163
1164
1165
1166
1167

1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181

1182
1183
1184
1185
1186
1187
1188
1189

1190
1191
1192
1193
1194
1195
1196
      if( i==1 ) continue;   /* Ignore the TEMP database */
      if( pBt && sqlite3BtreeIsInTrans(pBt) ){
        char const *zFile = sqlite3BtreeGetJournalname(pBt);
        if( zFile[0]==0 ) continue;  /* Ignore :memory: databases */
        if( !needSync && !sqlite3BtreeSyncDisabled(pBt) ){
          needSync = 1;
        }
        rc = sqlite3OsWrite(master, zFile, strlen(zFile)+1);

        if( rc!=SQLITE_OK ){
          sqlite3OsClose(&master);
          sqlite3OsDelete(zMaster);
          sqliteFree(zMaster);
          return rc;
        }
      }
    }


    /* Sync the master journal file. Before doing this, open the directory
    ** the master journal file is store in so that it gets synced too.
    */
    zMainFile = sqlite3BtreeGetDirname(db->aDb[0].pBt);

    rc = sqlite3OsOpenDirectory(master, zMainFile);
    if( rc!=SQLITE_OK ||
          (needSync && (rc=sqlite3OsSync(master,0))!=SQLITE_OK) ){
      sqlite3OsClose(&master);
      sqlite3OsDelete(zMaster);
      sqliteFree(zMaster);
      return rc;
    }


    /* Sync all the db files involved in the transaction. The same call
    ** sets the master journal pointer in each individual journal. If
    ** an error occurs here, do not delete the master journal file.
    **
    ** If the error occurs during the first call to
    ** sqlite3BtreeCommitPhaseOne(), then there is a chance that the







|
>














>








>







1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
      if( i==1 ) continue;   /* Ignore the TEMP database */
      if( pBt && sqlite3BtreeIsInTrans(pBt) ){
        char const *zFile = sqlite3BtreeGetJournalname(pBt);
        if( zFile[0]==0 ) continue;  /* Ignore :memory: databases */
        if( !needSync && !sqlite3BtreeSyncDisabled(pBt) ){
          needSync = 1;
        }
        rc = sqlite3OsWrite(master, zFile, strlen(zFile)+1, offset);
        offset += strlen(zFile)+1;
        if( rc!=SQLITE_OK ){
          sqlite3OsClose(&master);
          sqlite3OsDelete(zMaster);
          sqliteFree(zMaster);
          return rc;
        }
      }
    }


    /* Sync the master journal file. Before doing this, open the directory
    ** the master journal file is store in so that it gets synced too.
    */
    zMainFile = sqlite3BtreeGetDirname(db->aDb[0].pBt);
#if 0
    rc = sqlite3OsOpenDirectory(master, zMainFile);
    if( rc!=SQLITE_OK ||
          (needSync && (rc=sqlite3OsSync(master,0))!=SQLITE_OK) ){
      sqlite3OsClose(&master);
      sqlite3OsDelete(zMaster);
      sqliteFree(zMaster);
      return rc;
    }
#endif

    /* Sync all the db files involved in the transaction. The same call
    ** sets the master journal pointer in each individual journal. If
    ** an error occurs here, do not delete the master journal file.
    **
    ** If the error occurs during the first call to
    ** sqlite3BtreeCommitPhaseOne(), then there is a chance that the