SQLite

Check-in [bd8740d1ae]
Login

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

Overview
Comment:Restructure the OS interface yet again. This time make the OsFile object a virtual base class which is subclassed for unix, windows, and the crash test simulator. Add the new file "os.c" for common os layer code. Move all OS-specific routines into the sqlite3Os structure. (CVS 2795)
Downloads: Tarball | ZIP archive
Timelines: family | ancestors | descendants | both | trunk
Files: files | file ages | folders
SHA1: bd8740d1aecba69e1b5d64d43db07e8ad8841f07
User & Date: drh 2005-11-30 03:20:31.000
Context
2005-12-02
01:57
Documentation updates. (CVS 2796) (check-in: 2ffb90c39d user: drh tags: trunk)
2005-11-30
03:20
Restructure the OS interface yet again. This time make the OsFile object a virtual base class which is subclassed for unix, windows, and the crash test simulator. Add the new file "os.c" for common os layer code. Move all OS-specific routines into the sqlite3Os structure. (CVS 2795) (check-in: bd8740d1ae user: drh tags: trunk)
2005-11-29
19:56
The crash test works now, at least on unix. Untested on windows. But it never worked on windows before so if it still does not there is no big loss. I am still troubled by the current design of the OS overloading mechanism. Expect to see more changes. (CVS 2794) (check-in: fa1d7ecfcc user: drh tags: trunk)
Changes
Unified Diff Ignore Whitespace Patch
Changes to Makefile.in.
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
TCC += -DSQLITE_OMIT_CURSOR

# Object files for the SQLite library.
#
LIBOBJ = alter.lo analyze.lo attach.lo auth.lo btree.lo build.lo \
         callback.lo complete.lo date.lo \
         delete.lo expr.lo func.lo hash.lo insert.lo \
         main.lo opcodes.lo os_unix.lo os_win.lo \
         pager.lo parse.lo pragma.lo prepare.lo printf.lo random.lo \
         select.lo table.lo tokenize.lo trigger.lo update.lo \
         util.lo vacuum.lo \
         vdbe.lo vdbeapi.lo vdbeaux.lo vdbefifo.lo vdbemem.lo \
         where.lo utf.lo legacy.lo

# All of the source code files.







|







111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
TCC += -DSQLITE_OMIT_CURSOR

# Object files for the SQLite library.
#
LIBOBJ = alter.lo analyze.lo attach.lo auth.lo btree.lo build.lo \
         callback.lo complete.lo date.lo \
         delete.lo expr.lo func.lo hash.lo insert.lo \
         main.lo opcodes.lo os.lo os_unix.lo os_win.lo \
         pager.lo parse.lo pragma.lo prepare.lo printf.lo random.lo \
         select.lo table.lo tokenize.lo trigger.lo update.lo \
         util.lo vacuum.lo \
         vdbe.lo vdbeapi.lo vdbeaux.lo vdbefifo.lo vdbemem.lo \
         where.lo utf.lo legacy.lo

# All of the source code files.
139
140
141
142
143
144
145

146
147
148
149
150
151
152
  $(TOP)/src/expr.c \
  $(TOP)/src/func.c \
  $(TOP)/src/hash.c \
  $(TOP)/src/hash.h \
  $(TOP)/src/insert.c \
  $(TOP)/src/legacy.c \
  $(TOP)/src/main.c \

  $(TOP)/src/os_unix.c \
  $(TOP)/src/os_win.c \
  $(TOP)/src/pager.c \
  $(TOP)/src/pager.h \
  $(TOP)/src/parse.y \
  $(TOP)/src/pragma.c \
  $(TOP)/src/prepare.c \







>







139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
  $(TOP)/src/expr.c \
  $(TOP)/src/func.c \
  $(TOP)/src/hash.c \
  $(TOP)/src/hash.h \
  $(TOP)/src/insert.c \
  $(TOP)/src/legacy.c \
  $(TOP)/src/main.c \
  $(TOP)/src/os.c \
  $(TOP)/src/os_unix.c \
  $(TOP)/src/os_win.c \
  $(TOP)/src/pager.c \
  $(TOP)/src/pager.h \
  $(TOP)/src/parse.y \
  $(TOP)/src/pragma.c \
  $(TOP)/src/prepare.c \
323
324
325
326
327
328
329



330
331
332
333
334
335
336

opcodes.c:	opcodes.h $(TOP)/mkopcodec.awk
	sort -n -b +2 opcodes.h | awk -f $(TOP)/mkopcodec.awk >opcodes.c

opcodes.h:	parse.h $(TOP)/src/vdbe.c $(TOP)/mkopcodeh.awk
	cat parse.h $(TOP)/src/vdbe.c | awk -f $(TOP)/mkopcodeh.awk >opcodes.h




os_unix.lo:	$(TOP)/src/os_unix.c $(HDR)
	$(LTCOMPILE) -c $(TOP)/src/os_unix.c

os_win.lo:	$(TOP)/src/os_win.c $(HDR)
	$(LTCOMPILE) -c $(TOP)/src/os_win.c

parse.lo:	parse.c $(HDR)







>
>
>







324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340

opcodes.c:	opcodes.h $(TOP)/mkopcodec.awk
	sort -n -b +2 opcodes.h | awk -f $(TOP)/mkopcodec.awk >opcodes.c

opcodes.h:	parse.h $(TOP)/src/vdbe.c $(TOP)/mkopcodeh.awk
	cat parse.h $(TOP)/src/vdbe.c | awk -f $(TOP)/mkopcodeh.awk >opcodes.h

os.lo:	$(TOP)/src/os.c $(HDR)
	$(LTCOMPILE) -c $(TOP)/src/os.c

os_unix.lo:	$(TOP)/src/os_unix.c $(HDR)
	$(LTCOMPILE) -c $(TOP)/src/os_unix.c

os_win.lo:	$(TOP)/src/os_win.c $(HDR)
	$(LTCOMPILE) -c $(TOP)/src/os_win.c

parse.lo:	parse.c $(HDR)
Changes to main.mk.
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
TCCX = $(TCC) $(OPTS) $(THREADSAFE) $(USLEEP) -I. -I$(TOP)/src

# Object files for the SQLite library.
#
LIBOBJ+= alter.o analyze.o attach.o auth.o btree.o build.o \
         callback.o complete.o date.o delete.o \
         expr.o func.o hash.o insert.o \
         main.o opcodes.o os_unix.o os_win.o \
         pager.o parse.o pragma.o prepare.o printf.o random.o \
         select.o table.o tclsqlite.o tokenize.o trigger.o \
         update.o util.o vacuum.o \
         vdbe.o vdbeapi.o vdbeaux.o vdbefifo.o vdbemem.o \
         where.o utf.o legacy.o

# All of the source code files.







|







53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
TCCX = $(TCC) $(OPTS) $(THREADSAFE) $(USLEEP) -I. -I$(TOP)/src

# Object files for the SQLite library.
#
LIBOBJ+= alter.o analyze.o attach.o auth.o btree.o build.o \
         callback.o complete.o date.o delete.o \
         expr.o func.o hash.o insert.o \
         main.o opcodes.o os.o os_unix.o os_win.o \
         pager.o parse.o pragma.o prepare.o printf.o random.o \
         select.o table.o tclsqlite.o tokenize.o trigger.o \
         update.o util.o vacuum.o \
         vdbe.o vdbeapi.o vdbeaux.o vdbefifo.o vdbemem.o \
         where.o utf.o legacy.o

# All of the source code files.
81
82
83
84
85
86
87

88
89
90
91
92
93
94
  $(TOP)/src/expr.c \
  $(TOP)/src/func.c \
  $(TOP)/src/hash.c \
  $(TOP)/src/hash.h \
  $(TOP)/src/insert.c \
  $(TOP)/src/legacy.c \
  $(TOP)/src/main.c \

  $(TOP)/src/os_unix.c \
  $(TOP)/src/os_win.c \
  $(TOP)/src/pager.c \
  $(TOP)/src/pager.h \
  $(TOP)/src/parse.y \
  $(TOP)/src/pragma.c \
  $(TOP)/src/prepare.c \







>







81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
  $(TOP)/src/expr.c \
  $(TOP)/src/func.c \
  $(TOP)/src/hash.c \
  $(TOP)/src/hash.h \
  $(TOP)/src/insert.c \
  $(TOP)/src/legacy.c \
  $(TOP)/src/main.c \
  $(TOP)/src/os.c \
  $(TOP)/src/os_unix.c \
  $(TOP)/src/os_win.c \
  $(TOP)/src/pager.c \
  $(TOP)/src/pager.h \
  $(TOP)/src/parse.y \
  $(TOP)/src/pragma.c \
  $(TOP)/src/prepare.c \
256
257
258
259
260
261
262



263
264
265
266
267
268
269
	$(TCCX) -c opcodes.c

opcodes.c:	opcodes.h $(TOP)/mkopcodec.awk
	sort -n -b +2 opcodes.h | awk -f $(TOP)/mkopcodec.awk >opcodes.c

opcodes.h:	parse.h $(TOP)/src/vdbe.c $(TOP)/mkopcodeh.awk
	cat parse.h $(TOP)/src/vdbe.c | awk -f $(TOP)/mkopcodeh.awk >opcodes.h




os_unix.o:	$(TOP)/src/os_unix.c $(HDR)
	$(TCCX) -c $(TOP)/src/os_unix.c

os_win.o:	$(TOP)/src/os_win.c $(HDR)
	$(TCCX) -c $(TOP)/src/os_win.c








>
>
>







257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
	$(TCCX) -c opcodes.c

opcodes.c:	opcodes.h $(TOP)/mkopcodec.awk
	sort -n -b +2 opcodes.h | awk -f $(TOP)/mkopcodec.awk >opcodes.c

opcodes.h:	parse.h $(TOP)/src/vdbe.c $(TOP)/mkopcodeh.awk
	cat parse.h $(TOP)/src/vdbe.c | awk -f $(TOP)/mkopcodeh.awk >opcodes.h

os.o:	$(TOP)/src/os.c $(HDR)
	$(TCCX) -c $(TOP)/src/os.c

os_unix.o:	$(TOP)/src/os_unix.c $(HDR)
	$(TCCX) -c $(TOP)/src/os_unix.c

os_win.o:	$(TOP)/src/os_win.c $(HDR)
	$(TCCX) -c $(TOP)/src/os_win.c

Changes to src/date.c.
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
** This file contains the C functions that implement date and time
** functions for SQLite.  
**
** There is only one exported symbol in this file - the function
** sqlite3RegisterDateTimeFunctions() found at the bottom of the file.
** All other code has file scope.
**
** $Id: date.c,v 1.45 2005/06/25 18:42:14 drh Exp $
**
** NOTES:
**
** SQLite processes all times and dates as Julian Day numbers.  The
** dates and times are stored as the number of days since noon
** in Greenwich on November 24, 4714 B.C. according to the Gregorian
** calendar system.







|







12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
** This file contains the C functions that implement date and time
** functions for SQLite.  
**
** There is only one exported symbol in this file - the function
** sqlite3RegisterDateTimeFunctions() found at the bottom of the file.
** All other code has file scope.
**
** $Id: date.c,v 1.46 2005/11/30 03:20:31 drh Exp $
**
** NOTES:
**
** SQLite processes all times and dates as Julian Day numbers.  The
** dates and times are stored as the number of days since noon
** in Greenwich on November 24, 4714 B.C. according to the Gregorian
** calendar system.
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
  memset(p, 0, sizeof(*p));
  if( parseYyyyMmDd(zDate,p)==0 ){
    return 0;
  }else if( parseHhMmSs(zDate, p)==0 ){
    return 0;
  }else if( sqlite3StrICmp(zDate,"now")==0){
    double r;
    sqlite3OsCurrentTime(&r);
    p->rJD = r;
    p->validJD = 1;
    return 0;
  }else if( sqlite3IsNumber(zDate, 0, SQLITE_UTF8) ){
    getValue(zDate, &p->rJD);
    p->validJD = 1;
    return 0;







|







307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
  memset(p, 0, sizeof(*p));
  if( parseYyyyMmDd(zDate,p)==0 ){
    return 0;
  }else if( parseHhMmSs(zDate, p)==0 ){
    return 0;
  }else if( sqlite3StrICmp(zDate,"now")==0){
    double r;
    sqlite3Os.xCurrentTime(&r);
    p->rJD = r;
    p->validJD = 1;
    return 0;
  }else if( sqlite3IsNumber(zDate, 0, SQLITE_UTF8) ){
    getValue(zDate, &p->rJD);
    p->validJD = 1;
    return 0;
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
    int s = x.s + 0.5;
    x.s = s;
  }
  x.tz = 0;
  x.validJD = 0;
  computeJD(&x);
  t = (x.rJD-2440587.5)*86400.0 + 0.5;
  sqlite3OsEnterMutex();
  pTm = localtime(&t);
  y.Y = pTm->tm_year + 1900;
  y.M = pTm->tm_mon + 1;
  y.D = pTm->tm_mday;
  y.h = pTm->tm_hour;
  y.m = pTm->tm_min;
  y.s = pTm->tm_sec;
  sqlite3OsLeaveMutex();
  y.validYMD = 1;
  y.validHMS = 1;
  y.validJD = 0;
  y.validTZ = 0;
  computeJD(&y);
  return y.rJD - x.rJD;
}







|







|







405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
    int s = x.s + 0.5;
    x.s = s;
  }
  x.tz = 0;
  x.validJD = 0;
  computeJD(&x);
  t = (x.rJD-2440587.5)*86400.0 + 0.5;
  sqlite3Os.xEnterMutex();
  pTm = localtime(&t);
  y.Y = pTm->tm_year + 1900;
  y.M = pTm->tm_mon + 1;
  y.D = pTm->tm_mday;
  y.h = pTm->tm_hour;
  y.m = pTm->tm_min;
  y.s = pTm->tm_sec;
  sqlite3Os.xLeaveMutex();
  y.validYMD = 1;
  y.validHMS = 1;
  y.validJD = 0;
  y.validTZ = 0;
  computeJD(&y);
  return y.rJD - x.rJD;
}
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
    extern int sqlite3_current_time;  /* See os_XXX.c */
    if( sqlite3_current_time ){
      t = sqlite3_current_time;
    }
  }
#endif

  sqlite3OsEnterMutex();
  strftime(zBuf, 20, zFormat, gmtime(&t));
  sqlite3OsLeaveMutex();

  sqlite3_result_text(context, zBuf, -1, SQLITE_TRANSIENT);
}
#endif

/*
** This function registered all of the above C functions as SQL







|

|







938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
    extern int sqlite3_current_time;  /* See os_XXX.c */
    if( sqlite3_current_time ){
      t = sqlite3_current_time;
    }
  }
#endif

  sqlite3Os.xEnterMutex();
  strftime(zBuf, 20, zFormat, gmtime(&t));
  sqlite3Os.xLeaveMutex();

  sqlite3_result_text(context, zBuf, -1, SQLITE_TRANSIENT);
}
#endif

/*
** This function registered all of the above C functions as SQL
Changes to src/main.c.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
**
*************************************************************************
** Main file for the SQLite library.  The routines in this file
** implement the programmer interface to the library.  Routines in
** other files are for internal use by SQLite and should not be
** accessed by users of the library.
**
** $Id: main.c,v 1.303 2005/10/20 07:28:18 drh Exp $
*/
#include "sqliteInt.h"
#include "os.h"
#include <ctype.h>

/*
** The following constant value is used by the SQLITE_BIGENDIAN and







|







10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
**
*************************************************************************
** Main file for the SQLite library.  The routines in this file
** implement the programmer interface to the library.  Routines in
** other files are for internal use by SQLite and should not be
** accessed by users of the library.
**
** $Id: main.c,v 1.304 2005/11/30 03:20:31 drh Exp $
*/
#include "sqliteInt.h"
#include "os.h"
#include <ctype.h>

/*
** The following constant value is used by the SQLITE_BIGENDIAN and
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
  if( db->pErr ){
    sqlite3ValueFree(db->pErr);
  }

#ifndef SQLITE_OMIT_GLOBALRECOVER
  {
    sqlite3 *pPrev;
    sqlite3OsEnterMutex();
    pPrev = pDbList;
    while( pPrev && pPrev->pNext!=db ){
      pPrev = pPrev->pNext;
    }
    if( pPrev ){
      pPrev->pNext = db->pNext;
    }else{
      assert( pDbList==db );
      pDbList = db->pNext;
    }
    sqlite3OsLeaveMutex();
  }
#endif

  db->magic = SQLITE_MAGIC_ERROR;
  sqliteFree(db);
  return SQLITE_OK;
}







|










|







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
  if( db->pErr ){
    sqlite3ValueFree(db->pErr);
  }

#ifndef SQLITE_OMIT_GLOBALRECOVER
  {
    sqlite3 *pPrev;
    sqlite3Os.xEnterMutex();
    pPrev = pDbList;
    while( pPrev && pPrev->pNext!=db ){
      pPrev = pPrev->pNext;
    }
    if( pPrev ){
      pPrev->pNext = db->pNext;
    }else{
      assert( pDbList==db );
      pDbList = db->pNext;
    }
    sqlite3Os.xLeaveMutex();
  }
#endif

  db->magic = SQLITE_MAGIC_ERROR;
  sqliteFree(db);
  return SQLITE_OK;
}
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
    delay = delays[NDELAY-1];
    prior = totals[NDELAY-1] + delay*(count-(NDELAY-1));
  }
  if( prior + delay > timeout ){
    delay = timeout - prior;
    if( delay<=0 ) return 0;
  }
  sqlite3OsSleep(delay);
  return 1;
#else
  int timeout = ((sqlite3 *)ptr)->busyTimeout;
  if( (count+1)*1000 > timeout ){
    return 0;
  }
  sqlite3OsSleep(1000);
  return 1;
#endif
}

/*
** Invoke the given busy handler.
**







|






|







288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
    delay = delays[NDELAY-1];
    prior = totals[NDELAY-1] + delay*(count-(NDELAY-1));
  }
  if( prior + delay > timeout ){
    delay = timeout - prior;
    if( delay<=0 ) return 0;
  }
  sqlite3Os.xSleep(delay);
  return 1;
#else
  int timeout = ((sqlite3 *)ptr)->busyTimeout;
  if( (count+1)*1000 > timeout ){
    return 0;
  }
  sqlite3Os.xSleep(1000);
  return 1;
#endif
}

/*
** Invoke the given busy handler.
**
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
opendb_out:
  if( sqlite3_errcode(db)==SQLITE_OK && sqlite3_malloc_failed ){
    sqlite3Error(db, SQLITE_NOMEM, 0);
  }
  *ppDb = db;
#ifndef SQLITE_OMIT_GLOBALRECOVER
  if( db ){
    sqlite3OsEnterMutex();
    db->pNext = pDbList;
    pDbList = db;
    sqlite3OsLeaveMutex();
  }
#endif
  return sqlite3_errcode(db);
}

/*
** Open a new database handle.







|


|







788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
opendb_out:
  if( sqlite3_errcode(db)==SQLITE_OK && sqlite3_malloc_failed ){
    sqlite3Error(db, SQLITE_NOMEM, 0);
  }
  *ppDb = db;
#ifndef SQLITE_OMIT_GLOBALRECOVER
  if( db ){
    sqlite3Os.xEnterMutex();
    db->pNext = pDbList;
    pDbList = db;
    sqlite3Os.xLeaveMutex();
  }
#endif
  return sqlite3_errcode(db);
}

/*
** Open a new database handle.
Added src/os.c.














































































































































>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
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
/*
** 2005 November 29
**
** The author disclaims copyright to this source code.  In place of
** a legal notice, here is a blessing:
**
**    May you do good and not evil.
**    May you find forgiveness for yourself and forgive others.
**    May you share freely, never taking more than you give.
**
******************************************************************************
**
** This file contains OS interface code that is common to all
** architectures.
*/
#include "sqliteInt.h"
#include "os.h"

/*
** 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 sqlite3OsFileHandle(OsFile *id){
  return id->pMethod->xFileHandle(id);
}
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 sqlite3OsLockState(OsFile *id){
  return id->pMethod->xLockState(id);
}
int sqlite3OsCheckReservedLock(OsFile *id){
  return id->pMethod->xCheckReservedLock(id);
}
Changes to src/os.h.
37
38
39
40
41
42
43
44

45



























46



47
48


49
50
51
52
53
54
55
#  define OS_UNIX 0
# endif
#else
# ifndef OS_WIN
#  define OS_WIN 0
# endif
#endif


/*



























** The OsFile object describes an open disk file in an OS-dependent way.



*/
typedef struct OsFile OsFile;



/*
** Define the maximum size of a temporary filename
*/
#if OS_WIN
# include <windows.h>
# define SQLITE_TEMPNAME_SIZE (MAX_PATH+50)








>

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

>
>
>

|
>
>







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 OS_UNIX 0
# endif
#else
# ifndef OS_WIN
#  define OS_WIN 0
# endif
#endif


/*
** Forward declarations
*/
typedef struct OsFile OsFile;
typedef struct IoMethod IoMethod;

/*
** An instance of the following structure contains pointers to all
** methods on an OsFile object.
*/
struct IoMethod {
  int (*xClose)(OsFile**);
  int (*xOpenDirectory)(OsFile*, const char*);
  int (*xRead)(OsFile*, void*, int amt);
  int (*xWrite)(OsFile*, const void*, int amt);
  int (*xSeek)(OsFile*, i64 offset);
  int (*xTruncate)(OsFile*, i64 size);
  int (*xSync)(OsFile*, int);
  void (*xSetFullSync)(OsFile *id, int setting);
  int (*xFileHandle)(OsFile *id);
  int (*xFileSize)(OsFile*, i64 *pSize);
  int (*xLock)(OsFile*, int);
  int (*xUnlock)(OsFile*, int);
  int (*xLockState)(OsFile *id);
  int (*xCheckReservedLock)(OsFile *id);
};

/*
** The OsFile object describes an open disk file in an OS-dependent way.
** The version of OsFile defined here is a generic versions.  Each Os
** implementation defines its own subclass of this structure that contains
** additional information needed to handle file I/O.
*/
struct OsFile {
  IoMethod const *pMethod;
};

/*
** Define the maximum size of a temporary filename
*/
#if OS_WIN
# include <windows.h>
# define SQLITE_TEMPNAME_SIZE (MAX_PATH+50)
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

#define RESERVED_BYTE     (PENDING_BYTE+1)
#define SHARED_FIRST      (PENDING_BYTE+2)
#define SHARED_SIZE       510

/*
** A single global instance of the following structure holds pointers to the
** various disk I/O routines.
*/
extern struct sqlite3IoVtbl {
  int (*xDelete)(const char*);
  int (*xFileExists)(const char*);
  int (*xOpenReadWrite)(const char*, OsFile**, int*);
  int (*xOpenExclusive)(const char*, OsFile**, int);
  int (*xOpenReadOnly)(const char*, OsFile**);

  int (*xOpenDirectory)(const char*, OsFile*);



  int (*xSyncDirectory)(const char*);
  int (*xTempFileName)(char*);
  int (*xIsDirWritable)(char*);










  int (*xClose)(OsFile**);

  int (*xRead)(OsFile*, void*, int amt);
  int (*xWrite)(OsFile*, const void*, int amt);
  int (*xSeek)(OsFile*, i64 offset);
  int (*xSync)(OsFile*, int);
  int (*xTruncate)(OsFile*, i64 size);
  int (*xFileSize)(OsFile*, i64 *pSize);
  char *(*xFullPathname)(const char*);
  int (*xLock)(OsFile*, int);
  int (*xUnlock)(OsFile*, int);
  int (*xCheckReservedLock)(OsFile *id);
  void (*xSetFullSync)(OsFile *id, int setting);
  int (*xFileHandle)(OsFile *id);
  int (*xLockState)(OsFile *id);
} sqlite3Io;

/* The interface for file I/O is above.  Other miscellaneous functions
** are below */

int sqlite3OsRandomSeed(char*);
int sqlite3OsSleep(int ms);
int sqlite3OsCurrentTime(double*);
void sqlite3OsEnterMutex(void);
void sqlite3OsLeaveMutex(void);


#endif /* _SQLITE_OS_H_ */







|

|
<
<



>
|
>
>
>


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



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

#define RESERVED_BYTE     (PENDING_BYTE+1)
#define SHARED_FIRST      (PENDING_BYTE+2)
#define SHARED_SIZE       510

/*
** A single global instance of the following structure holds pointers to the
** various system-specific interface routines.
*/
extern 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*);

  int  (*xRandomSeed)(char*);
  int  (*xSleep)(int ms);
  int  (*xCurrentTime)(double*);
  void (*xEnterMutex)(void);
  void (*xLeaveMutex)(void);
} sqlite3Os;

/*
** Prototypes for routines found in os.c
*/
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 sqlite3OsFileHandle(OsFile *id);


int sqlite3OsFileSize(OsFile*, i64 *pSize);


int sqlite3OsLock(OsFile*, int);
int sqlite3OsUnlock(OsFile*, int);
int sqlite3OsLockState(OsFile *id);
int sqlite3OsCheckReservedLock(OsFile *id);




#endif /* _SQLITE_OS_H_ */
Changes to src/os_unix.c.
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80

81

82
83
84
85
86
87
88
#ifndef SQLITE_DEFAULT_FILE_PERMISSIONS
# define SQLITE_DEFAULT_FILE_PERMISSIONS 0644
#endif



/*
** The OsFile structure is a operating-system dependent representation
** of an open file handle.  It is defined differently for each architecture.
**
** This is the definition for Unix.
**
** OsFile.locktype takes one of the values SHARED_LOCK, RESERVED_LOCK,
** PENDING_LOCK or EXCLUSIVE_LOCK.
*/

struct OsFile {

  struct openCnt *pOpen;    /* Info about all open fd's on this inode */
  struct lockInfo *pLock;   /* Info about locks on this inode */
  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 */







|
<
|
<
<
<
<

>
|
>







66
67
68
69
70
71
72
73

74




75
76
77
78
79
80
81
82
83
84
85
#ifndef SQLITE_DEFAULT_FILE_PERMISSIONS
# define SQLITE_DEFAULT_FILE_PERMISSIONS 0644
#endif



/*
** 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 */
  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 */
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
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
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
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747

748

749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
/*
** Return TRUE if the named file exists.
*/
static int unixFileExists(const char *zFilename){
  return access(zFilename, 0)==0;
}

/*
** Allocate memory for an OsFile.  Initialize the new OsFile
** to the value given in pInit and return a pointer to the new
** OsFile.  If we run out of memory, close the file and return NULL.
*/
static OsFile *allocateOsFile(OsFile *pInit){
  OsFile *pNew;
  pNew = sqliteMalloc( sizeof(OsFile) );
  if( pNew==0 ){
    close(pInit->h);
  }else{
    *pNew = *pInit;
  }
  return pNew;
}

/*
** 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.
*/
static int unixOpenReadWrite(
  const char *zFilename,
  OsFile **pId,
  int *pReadonly
){
  int rc;
  OsFile f;

  assert( 0==*pId );
  f.dirfd = -1;
  SET_THREADID(&f);
  f.h = open(zFilename, O_RDWR|O_CREAT|O_LARGEFILE|O_BINARY,
                          SQLITE_DEFAULT_FILE_PERMISSIONS);
  if( f.h<0 ){
#ifdef EISDIR
    if( errno==EISDIR ){
      return SQLITE_CANTOPEN;
    }
#endif
    f.h = open(zFilename, O_RDONLY|O_LARGEFILE|O_BINARY);
    if( f.h<0 ){
      return SQLITE_CANTOPEN; 
    }
    *pReadonly = 1;
  }else{
    *pReadonly = 0;
  }
  sqlite3OsEnterMutex();
  rc = findLockInfo(f.h, &f.pLock, &f.pOpen);
  sqlite3OsLeaveMutex();
  if( rc ){
    close(f.h);
    return SQLITE_NOMEM;
  }
  f.locktype = 0;
  TRACE3("OPEN    %-3d %s\n", f.h, zFilename);
  *pId = allocateOsFile(&f);
  if( *pId==0 ){
    return SQLITE_NOMEM;
  }else{
    OpenCounter(+1);
    return SQLITE_OK;
  }
}


/*
** 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.
*/
static int unixOpenExclusive(const char *zFilename, OsFile **pId, int delFlag){
  int rc;
  OsFile f;

  assert( 0==*pId );
  if( access(zFilename, 0)==0 ){
    return SQLITE_CANTOPEN;
  }
  SET_THREADID(&f);
  f.dirfd = -1;
  f.h = open(zFilename,
                O_RDWR|O_CREAT|O_EXCL|O_NOFOLLOW|O_LARGEFILE|O_BINARY,
                SQLITE_DEFAULT_FILE_PERMISSIONS);
  if( f.h<0 ){
    return SQLITE_CANTOPEN;
  }
  sqlite3OsEnterMutex();
  rc = findLockInfo(f.h, &f.pLock, &f.pOpen);
  sqlite3OsLeaveMutex();
  if( rc ){
    close(f.h);
    unlink(zFilename);
    return SQLITE_NOMEM;
  }
  f.locktype = 0;
  if( delFlag ){
    unlink(zFilename);
  }
  TRACE3("OPEN-EX %-3d %s\n", f.h, zFilename);
  *pId = allocateOsFile(&f);
  if( *pId==0 ){
    return SQLITE_NOMEM;
  }else{
    OpenCounter(+1);
    return SQLITE_OK;
  }
}

/*
** 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.
*/
static int unixOpenReadOnly(const char *zFilename, OsFile **pId){
  int rc;
  OsFile f;

  assert( 0==*pId );
  SET_THREADID(&f);
  f.dirfd = -1;
  f.h = open(zFilename, O_RDONLY|O_LARGEFILE|O_BINARY);
  if( f.h<0 ){
    return SQLITE_CANTOPEN;
  }
  sqlite3OsEnterMutex();
  rc = findLockInfo(f.h, &f.pLock, &f.pOpen);
  sqlite3OsLeaveMutex();
  if( rc ){
    close(f.h);
    return SQLITE_NOMEM;
  }
  f.locktype = 0;
  TRACE3("OPEN-RO %-3d %s\n", f.h, zFilename);
  *pId = allocateOsFile(&f);
  if( *pId==0 ){
    return SQLITE_NOMEM;
  }else{
    OpenCounter(+1);
    return SQLITE_OK;
  }
}

/*
** 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.
**
** This routine is only meaningful for Unix.  It is a no-op under
** windows since windows does not support hard links.
**
** On success, a handle for a previously open file at *id is
** updated with the new directory file descriptor and SQLITE_OK is
** returned.
**
** On failure, the function returns SQLITE_CANTOPEN and leaves
** *id unchanged.
*/
static int unixOpenDirectory(
  const char *zDirname,
  OsFile *id

){

  if( id==0 ){
    /* Do not open the directory if the corresponding file is not already
    ** open. */
    return SQLITE_CANTOPEN;
  }
  SET_THREADID(id);
  assert( id->dirfd<0 );
  id->dirfd = open(zDirname, O_RDONLY|O_BINARY, 0);
  if( id->dirfd<0 ){
    return SQLITE_CANTOPEN; 
  }
  TRACE3("OPENDIR %-3d %s\n", id->dirfd, zDirname);
  return SQLITE_OK;
}

/*
** If the following global variable points to a string which is the
** name of a directory, then that directory will be used to store
** temporary files.







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




















|




















|

|






|
<
<
<
<
<
<



















|













|

|










|
<
<
<
<
<
<











|








|

|






|
<
<
<
<
<
<



















<
|
>

>
|




|
|
|
|


|







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
662






663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692






693
694
695
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
721
722
723
724
725
726
727
728
729
730
731
732
733
734
/*
** Return TRUE if the named file exists.
*/
static int unixFileExists(const char *zFilename){
  return access(zFilename, 0)==0;
}

/* Forward declaration */





static int allocateUnixFile(unixFile *pInit, OsFile **pId);









/*
** 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.
*/
static int unixOpenReadWrite(
  const char *zFilename,
  OsFile **pId,
  int *pReadonly
){
  int rc;
  unixFile f;

  assert( 0==*pId );
  f.dirfd = -1;
  SET_THREADID(&f);
  f.h = open(zFilename, O_RDWR|O_CREAT|O_LARGEFILE|O_BINARY,
                          SQLITE_DEFAULT_FILE_PERMISSIONS);
  if( f.h<0 ){
#ifdef EISDIR
    if( errno==EISDIR ){
      return SQLITE_CANTOPEN;
    }
#endif
    f.h = open(zFilename, O_RDONLY|O_LARGEFILE|O_BINARY);
    if( f.h<0 ){
      return SQLITE_CANTOPEN; 
    }
    *pReadonly = 1;
  }else{
    *pReadonly = 0;
  }
  sqlite3Os.xEnterMutex();
  rc = findLockInfo(f.h, &f.pLock, &f.pOpen);
  sqlite3Os.xLeaveMutex();
  if( rc ){
    close(f.h);
    return SQLITE_NOMEM;
  }
  f.locktype = 0;
  TRACE3("OPEN    %-3d %s\n", f.h, zFilename);
  return allocateUnixFile(&f, pId);






}


/*
** 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.
*/
static int unixOpenExclusive(const char *zFilename, OsFile **pId, int delFlag){
  int rc;
  unixFile f;

  assert( 0==*pId );
  if( access(zFilename, 0)==0 ){
    return SQLITE_CANTOPEN;
  }
  SET_THREADID(&f);
  f.dirfd = -1;
  f.h = open(zFilename,
                O_RDWR|O_CREAT|O_EXCL|O_NOFOLLOW|O_LARGEFILE|O_BINARY,
                SQLITE_DEFAULT_FILE_PERMISSIONS);
  if( f.h<0 ){
    return SQLITE_CANTOPEN;
  }
  sqlite3Os.xEnterMutex();
  rc = findLockInfo(f.h, &f.pLock, &f.pOpen);
  sqlite3Os.xLeaveMutex();
  if( rc ){
    close(f.h);
    unlink(zFilename);
    return SQLITE_NOMEM;
  }
  f.locktype = 0;
  if( delFlag ){
    unlink(zFilename);
  }
  TRACE3("OPEN-EX %-3d %s\n", f.h, zFilename);
  return allocateUnixFile(&f, pId);






}

/*
** 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.
*/
static int unixOpenReadOnly(const char *zFilename, OsFile **pId){
  int rc;
  unixFile f;

  assert( 0==*pId );
  SET_THREADID(&f);
  f.dirfd = -1;
  f.h = open(zFilename, O_RDONLY|O_LARGEFILE|O_BINARY);
  if( f.h<0 ){
    return SQLITE_CANTOPEN;
  }
  sqlite3Os.xEnterMutex();
  rc = findLockInfo(f.h, &f.pLock, &f.pOpen);
  sqlite3Os.xLeaveMutex();
  if( rc ){
    close(f.h);
    return SQLITE_NOMEM;
  }
  f.locktype = 0;
  TRACE3("OPEN-RO %-3d %s\n", f.h, zFilename);
  return allocateUnixFile(&f, pId);






}

/*
** 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.
**
** This routine is only meaningful for Unix.  It is a no-op under
** windows since windows does not support hard links.
**
** On success, a handle for a previously open file at *id is
** updated with the new directory file descriptor and SQLITE_OK is
** returned.
**
** On failure, the function returns SQLITE_CANTOPEN and leaves
** *id unchanged.
*/
static int unixOpenDirectory(

  OsFile *id,
  const char *zDirname
){
  unixFile *pFile = (unixFile*)id;
  if( pFile==0 ){
    /* Do not open the directory if the corresponding file is not already
    ** open. */
    return SQLITE_CANTOPEN;
  }
  SET_THREADID(pFile);
  assert( pFile->dirfd<0 );
  pFile->dirfd = open(zDirname, O_RDONLY|O_BINARY, 0);
  if( pFile->dirfd<0 ){
    return SQLITE_CANTOPEN; 
  }
  TRACE3("OPENDIR %-3d %s\n", pFile->dirfd, zDirname);
  return SQLITE_OK;
}

/*
** If the following global variable points to a string which is the
** name of a directory, then that directory will be used to store
** temporary files.
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
** wrong.
*/
static int unixRead(OsFile *id, void *pBuf, int amt){
  int got;
  assert( id );
  SimulateIOError(SQLITE_IOERR);
  TIMER_START;
  got = read(id->h, pBuf, amt);
  TIMER_END;
  TRACE5("READ    %-3d %5d %7d %d\n", id->h, got, last_page, TIMER_ELAPSED);

  SEEK(0);
  /* if( got<0 ) got = 0; */
  if( got==amt ){
    return SQLITE_OK;
  }else{
    return SQLITE_IOERR;
  }
}

/*
** 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 );
  SimulateIOError(SQLITE_IOERR);
  SimulateDiskfullError;
  TIMER_START;
  while( amt>0 && (wrote = write(id->h, pBuf, amt))>0 ){
    amt -= wrote;
    pBuf = &((char*)pBuf)[wrote];
  }
  TIMER_END;
  TRACE5("WRITE   %-3d %5d %7d %d\n", id->h, wrote, last_page, TIMER_ELAPSED);

  SEEK(0);
  if( amt>0 ){
    return SQLITE_FULL;
  }
  return SQLITE_OK;
}

/*
** Move the read/write pointer in a file.
*/
static int unixSeek(OsFile *id, i64 offset){
  assert( id );
  SEEK(offset/1024 + 1);
#ifdef SQLITE_TEST
  if( offset ) SimulateDiskfullError
#endif
  lseek(id->h, offset, SEEK_SET);
  return SQLITE_OK;
}

#ifdef SQLITE_TEST
/*
** Count the number of fullsyncs and normal syncs.  This is used to test
** that syncs and fullsyncs are occuring at the right times.







|

|
>




















|




|
>
















|







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
** wrong.
*/
static int unixRead(OsFile *id, void *pBuf, int amt){
  int got;
  assert( id );
  SimulateIOError(SQLITE_IOERR);
  TIMER_START;
  got = read(((unixFile*)id)->h, pBuf, amt);
  TIMER_END;
  TRACE5("READ    %-3d %5d %7d %d\n", ((unixFile*)id)->h, got,
          last_page, TIMER_ELAPSED);
  SEEK(0);
  /* if( got<0 ) got = 0; */
  if( got==amt ){
    return SQLITE_OK;
  }else{
    return SQLITE_IOERR;
  }
}

/*
** 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 );
  SimulateIOError(SQLITE_IOERR);
  SimulateDiskfullError;
  TIMER_START;
  while( amt>0 && (wrote = write(((unixFile*)id)->h, pBuf, amt))>0 ){
    amt -= wrote;
    pBuf = &((char*)pBuf)[wrote];
  }
  TIMER_END;
  TRACE5("WRITE   %-3d %5d %7d %d\n", ((unixFile*)id)->h, wrote,
          last_page, TIMER_ELAPSED);
  SEEK(0);
  if( amt>0 ){
    return SQLITE_FULL;
  }
  return SQLITE_OK;
}

/*
** Move the read/write pointer in a file.
*/
static int unixSeek(OsFile *id, i64 offset){
  assert( id );
  SEEK(offset/1024 + 1);
#ifdef SQLITE_TEST
  if( offset ) SimulateDiskfullError
#endif
  lseek(((unixFile*)id)->h, offset, SEEK_SET);
  return SQLITE_OK;
}

#ifdef SQLITE_TEST
/*
** Count the number of fullsyncs and normal syncs.  This is used to test
** that syncs and fullsyncs are occuring at the right times.
964
965
966
967
968
969
970

971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
** 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){

  assert( id );
  SimulateIOError(SQLITE_IOERR);
  TRACE2("SYNC    %-3d\n", id->h);
  if( full_fsync(id->h, id->fullSync, dataOnly) ){
    return SQLITE_IOERR;
  }
  if( id->dirfd>=0 ){
    TRACE2("DIRSYNC %-3d\n", id->dirfd);
#ifndef SQLITE_DISABLE_DIRSYNC
    if( full_fsync(id->dirfd, id->fullSync, 0) ){
        return SQLITE_IOERR;
    }
#endif
    close(id->dirfd);  /* Only need to sync once, so close the directory */
    id->dirfd = -1;    /* when we are done. */
  }
  return SQLITE_OK;
}

/*
** Sync the directory zDirname. This is a no-op on operating systems other
** than UNIX.







>
|

|
|


|
|

|



|
|







933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
** 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){
  unixFile *pFile = (unixFile*)id;
  assert( pFile );
  SimulateIOError(SQLITE_IOERR);
  TRACE2("SYNC    %-3d\n", pFile->h);
  if( full_fsync(pFile->h, pFile->fullSync, dataOnly) ){
    return SQLITE_IOERR;
  }
  if( pFile->dirfd>=0 ){
    TRACE2("DIRSYNC %-3d\n", pFile->dirfd);
#ifndef SQLITE_DISABLE_DIRSYNC
    if( full_fsync(pFile->dirfd, pFile->fullSync, 0) ){
        return SQLITE_IOERR;
    }
#endif
    close(pFile->dirfd);  /* Only need to sync once, so close the directory */
    pFile->dirfd = -1;    /* when we are done. */
  }
  return SQLITE_OK;
}

/*
** Sync the directory zDirname. This is a no-op on operating systems other
** than UNIX.
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

/*
** Truncate an open file to a specified size
*/
static int unixTruncate(OsFile *id, i64 nByte){
  assert( id );
  SimulateIOError(SQLITE_IOERR);
  return ftruncate(id->h, nByte)==0 ? SQLITE_OK : SQLITE_IOERR;
}

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


  assert( id );
  if( CHECK_THREADID(id) ) return SQLITE_MISUSE;
  sqlite3OsEnterMutex(); /* Needed because id->pLock is shared across threads */

  /* Check if a thread in this process holds such a lock */
  if( id->pLock->locktype>SHARED_LOCK ){
    r = 1;
  }

  /* Otherwise see if some other process holds it.
  */
  if( !r ){
    struct flock lock;
    lock.l_whence = SEEK_SET;
    lock.l_start = RESERVED_BYTE;
    lock.l_len = 1;
    lock.l_type = F_WRLCK;
    fcntl(id->h, F_GETLK, &lock);
    if( lock.l_type!=F_UNLCK ){
      r = 1;
    }
  }
  
  sqlite3OsLeaveMutex();
  TRACE3("TEST WR-LOCK %d %d\n", id->h, r);

  return r;
}

#ifdef SQLITE_DEBUG
/*
** Helper function for printing out trace information from debugging
** binaries. This returns the string represetation of the supplied
** integer lock-type.
*/
static const char * locktypeName(int locktype){
  switch( locktype ){
  case NO_LOCK: return "NONE";
  case SHARED_LOCK: return "SHARED";
  case RESERVED_LOCK: return "RESERVED";
  case PENDING_LOCK: return "PENDING";
  case EXCLUSIVE_LOCK: return "EXCLUSIVE";
  }







|









|














>

|
|
|


|











|





|
|










|







985
986
987
988
989
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

/*
** Truncate an open file to a specified size
*/
static int unixTruncate(OsFile *id, i64 nByte){
  assert( id );
  SimulateIOError(SQLITE_IOERR);
  return ftruncate(((unixFile*)id)->h, nByte)==0 ? SQLITE_OK : SQLITE_IOERR;
}

/*
** Determine the current size of a file in bytes
*/
static int unixFileSize(OsFile *id, i64 *pSize){
  struct stat buf;
  assert( id );
  SimulateIOError(SQLITE_IOERR);
  if( fstat(((unixFile*)id)->h, &buf)!=0 ){
    return SQLITE_IOERR;
  }
  *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 );
  if( CHECK_THREADID(pFile) ) return SQLITE_MISUSE;
  sqlite3Os.xEnterMutex(); /* Because pFile->pLock is shared across threads */

  /* Check if a thread in this process holds such a lock */
  if( pFile->pLock->locktype>SHARED_LOCK ){
    r = 1;
  }

  /* Otherwise see if some other process holds it.
  */
  if( !r ){
    struct flock lock;
    lock.l_whence = SEEK_SET;
    lock.l_start = RESERVED_BYTE;
    lock.l_len = 1;
    lock.l_type = F_WRLCK;
    fcntl(pFile->h, F_GETLK, &lock);
    if( lock.l_type!=F_UNLCK ){
      r = 1;
    }
  }
  
  sqlite3Os.xLeaveMutex();
  TRACE3("TEST WR-LOCK %d %d\n", pFile->h, r);

  return r;
}

#ifdef SQLITE_DEBUG
/*
** Helper function for printing out trace information from debugging
** binaries. This returns the string represetation of the supplied
** integer lock-type.
*/
static const char *locktypeName(int locktype){
  switch( locktype ){
  case NO_LOCK: return "NONE";
  case SHARED_LOCK: return "SHARED";
  case RESERVED_LOCK: return "RESERVED";
  case PENDING_LOCK: return "PENDING";
  case EXCLUSIVE_LOCK: return "EXCLUSIVE";
  }
1152
1153
1154
1155
1156
1157
1158

1159
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
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
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
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
1317
1318

1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
  **
  ** The reason a single byte cannot be used instead of the 'shared byte
  ** range' is that some versions of windows do not support read-locks. By
  ** locking a random byte from a range, concurrent SHARED locks may exist
  ** even if the locking primitive used is always a write-lock.
  */
  int rc = SQLITE_OK;

  struct lockInfo *pLock = id->pLock;
  struct flock lock;
  int s;

  assert( id );
  TRACE7("LOCK    %d %s was %s(%s,%d) pid=%d\n", id->h, locktypeName(locktype), 

      locktypeName(id->locktype), locktypeName(pLock->locktype), pLock->cnt
      ,getpid() );
  if( CHECK_THREADID(id) ) return SQLITE_MISUSE;

  /* If there is already a lock of this type or more restrictive on the
  ** OsFile, do nothing. Don't use the end_lock: exit path, as
  ** sqlite3OsEnterMutex() hasn't been called yet.
  */
  if( id->locktype>=locktype ){
    TRACE3("LOCK    %d %s ok (already held)\n", id->h, locktypeName(locktype));

    return SQLITE_OK;
  }

  /* Make sure the locking sequence is correct
  */
  assert( id->locktype!=NO_LOCK || locktype==SHARED_LOCK );
  assert( locktype!=PENDING_LOCK );
  assert( locktype!=RESERVED_LOCK || id->locktype==SHARED_LOCK );

  /* This mutex is needed because id->pLock is shared across threads
  */
  sqlite3OsEnterMutex();

  /* If some thread using this PID has a lock via a different OsFile*
  ** handle that precludes the requested lock, return BUSY.
  */
  if( (id->locktype!=pLock->locktype && 
          (pLock->locktype>=PENDING_LOCK || locktype>SHARED_LOCK))
  ){
    rc = SQLITE_BUSY;
    goto end_lock;
  }

  /* If a SHARED lock is requested, and some thread using this PID already
  ** has a SHARED or RESERVED lock, then increment reference counts and
  ** return SQLITE_OK.
  */
  if( locktype==SHARED_LOCK && 
      (pLock->locktype==SHARED_LOCK || pLock->locktype==RESERVED_LOCK) ){
    assert( locktype==SHARED_LOCK );
    assert( id->locktype==0 );
    assert( pLock->cnt>0 );
    id->locktype = SHARED_LOCK;
    pLock->cnt++;
    id->pOpen->nLock++;
    goto end_lock;
  }

  lock.l_len = 1L;

  lock.l_whence = SEEK_SET;

  /* A PENDING lock is needed before acquiring a SHARED lock and before
  ** acquiring an EXCLUSIVE lock.  For the SHARED lock, the PENDING will
  ** be released.
  */
  if( locktype==SHARED_LOCK 
      || (locktype==EXCLUSIVE_LOCK && id->locktype<PENDING_LOCK)
  ){
    lock.l_type = (locktype==SHARED_LOCK?F_RDLCK:F_WRLCK);
    lock.l_start = PENDING_BYTE;
    s = fcntl(id->h, F_SETLK, &lock);
    if( s ){
      rc = (errno==EINVAL) ? SQLITE_NOLFS : SQLITE_BUSY;
      goto end_lock;
    }
  }


  /* If control gets to this point, then actually go ahead and make
  ** operating system calls for the specified lock.
  */
  if( locktype==SHARED_LOCK ){
    assert( pLock->cnt==0 );
    assert( pLock->locktype==0 );

    /* Now get the read-lock */
    lock.l_start = SHARED_FIRST;
    lock.l_len = SHARED_SIZE;
    s = fcntl(id->h, F_SETLK, &lock);

    /* Drop the temporary PENDING lock */
    lock.l_start = PENDING_BYTE;
    lock.l_len = 1L;
    lock.l_type = F_UNLCK;
    if( fcntl(id->h, F_SETLK, &lock)!=0 ){
      rc = SQLITE_IOERR;  /* This should never happen */
      goto end_lock;
    }
    if( s ){
      rc = (errno==EINVAL) ? SQLITE_NOLFS : SQLITE_BUSY;
    }else{
      id->locktype = SHARED_LOCK;
      id->pOpen->nLock++;
      pLock->cnt = 1;
    }
  }else if( locktype==EXCLUSIVE_LOCK && pLock->cnt>1 ){
    /* We are trying for an exclusive lock but another thread in this
    ** same process is still holding a shared lock. */
    rc = SQLITE_BUSY;
  }else{
    /* The request was for a RESERVED or EXCLUSIVE lock.  It is
    ** assumed that there is a SHARED or greater lock on the file
    ** already.
    */
    assert( 0!=id->locktype );
    lock.l_type = F_WRLCK;
    switch( locktype ){
      case RESERVED_LOCK:
        lock.l_start = RESERVED_BYTE;
        break;
      case EXCLUSIVE_LOCK:
        lock.l_start = SHARED_FIRST;
        lock.l_len = SHARED_SIZE;
        break;
      default:
        assert(0);
    }
    s = fcntl(id->h, F_SETLK, &lock);
    if( s ){
      rc = (errno==EINVAL) ? SQLITE_NOLFS : SQLITE_BUSY;
    }
  }
  
  if( rc==SQLITE_OK ){
    id->locktype = locktype;
    pLock->locktype = locktype;
  }else if( locktype==EXCLUSIVE_LOCK ){
    id->locktype = PENDING_LOCK;
    pLock->locktype = PENDING_LOCK;
  }

end_lock:
  sqlite3OsLeaveMutex();
  TRACE4("LOCK    %d %s %s\n", id->h, locktypeName(locktype), 
      rc==SQLITE_OK ? "ok" : "failed");
  return rc;
}

/*
** Lower the locking level on file descriptor id 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.
**
** It is not possible for this routine to fail if the second argument
** is NO_LOCK.  If the second argument is SHARED_LOCK, this routine
** might return SQLITE_IOERR instead of SQLITE_OK.
*/
static int unixUnlock(OsFile *id, int locktype){
  struct lockInfo *pLock;
  struct flock lock;
  int rc = SQLITE_OK;


  assert( id );
  TRACE7("UNLOCK  %d %d was %d(%d,%d) pid=%d\n", id->h, locktype, id->locktype, 
      id->pLock->locktype, id->pLock->cnt, getpid());
  if( CHECK_THREADID(id) ) return SQLITE_MISUSE;

  assert( locktype<=SHARED_LOCK );
  if( id->locktype<=locktype ){
    return SQLITE_OK;
  }
  sqlite3OsEnterMutex();
  pLock = id->pLock;
  assert( pLock->cnt!=0 );
  if( id->locktype>SHARED_LOCK ){
    assert( pLock->locktype==id->locktype );
    if( locktype==SHARED_LOCK ){
      lock.l_type = F_RDLCK;
      lock.l_whence = SEEK_SET;
      lock.l_start = SHARED_FIRST;
      lock.l_len = SHARED_SIZE;
      if( fcntl(id->h, F_SETLK, &lock)!=0 ){
        /* This should never happen */
        rc = SQLITE_IOERR;
      }
    }
    lock.l_type = F_UNLCK;
    lock.l_whence = SEEK_SET;
    lock.l_start = PENDING_BYTE;
    lock.l_len = 2L;  assert( PENDING_BYTE+1==RESERVED_BYTE );
    if( fcntl(id->h, F_SETLK, &lock)==0 ){
      pLock->locktype = SHARED_LOCK;
    }else{
      rc = SQLITE_IOERR;  /* This should never happen */
    }
  }
  if( locktype==NO_LOCK ){
    struct openCnt *pOpen;

    /* Decrement the shared lock counter.  Release the lock using an
    ** OS call only when all threads in this same process have released
    ** the lock.
    */
    pLock->cnt--;
    if( pLock->cnt==0 ){
      lock.l_type = F_UNLCK;
      lock.l_whence = SEEK_SET;
      lock.l_start = lock.l_len = 0L;
      if( fcntl(id->h, F_SETLK, &lock)==0 ){
        pLock->locktype = NO_LOCK;
      }else{
        rc = SQLITE_IOERR;  /* This should never happen */
      }
    }

    /* Decrement the count of locks against this same file.  When the
    ** count reaches zero, close any other file descriptors whose close
    ** was deferred because of outstanding locks.
    */
    pOpen = id->pOpen;
    pOpen->nLock--;
    assert( pOpen->nLock>=0 );
    if( pOpen->nLock==0 && pOpen->nPending>0 ){
      int i;
      for(i=0; i<pOpen->nPending; i++){
        close(pOpen->aPending[i]);
      }
      sqliteFree(pOpen->aPending);
      pOpen->nPending = 0;
      pOpen->aPending = 0;
    }
  }
  sqlite3OsLeaveMutex();
  id->locktype = locktype;
  return rc;
}

/*
** Close a file.
*/
static int unixClose(OsFile **pId){
  OsFile *id = *pId;
  if( !id ) return SQLITE_OK;
  if( CHECK_THREADID(id) ) return SQLITE_MISUSE;
  unixUnlock(id, 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;







>
|



|
|
>
|
<
|



|

|
|
>





|

|

|

|




|













|

|

|












|



|

















|





|






|
|











|












|






|


|




|
|





|













>

|
|
|
|


|


|
|

|
|





|








|

















|










|












|
|







|


|


|







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
1153
1154
1155
1156
1157
1158
1159
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
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
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
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
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
  **
  ** The reason a single byte cannot be used instead of the 'shared byte
  ** range' is that some versions of windows do not support read-locks. By
  ** locking a random byte from a range, concurrent SHARED locks may exist
  ** even if the locking primitive used is always a write-lock.
  */
  int rc = SQLITE_OK;
  unixFile *pFile = (unixFile*)id;
  struct lockInfo *pLock = pFile->pLock;
  struct flock lock;
  int s;

  assert( pFile );
  TRACE7("LOCK    %d %s was %s(%s,%d) pid=%d\n", pFile->h,
      locktypeName(locktype), locktypeName(pFile->locktype),
      locktypeName(pLock->locktype), pLock->cnt , getpid());

  if( CHECK_THREADID(pFile) ) return SQLITE_MISUSE;

  /* If there is already a lock of this type or more restrictive on the
  ** OsFile, do nothing. Don't use the end_lock: exit path, as
  ** sqlite3Os.xEnterMutex() hasn't been called yet.
  */
  if( pFile->locktype>=locktype ){
    TRACE3("LOCK    %d %s ok (already held)\n", pFile->h,
            locktypeName(locktype));
    return SQLITE_OK;
  }

  /* Make sure the locking sequence is correct
  */
  assert( pFile->locktype!=NO_LOCK || locktype==SHARED_LOCK );
  assert( locktype!=PENDING_LOCK );
  assert( locktype!=RESERVED_LOCK || pFile->locktype==SHARED_LOCK );

  /* This mutex is needed because pFile->pLock is shared across threads
  */
  sqlite3Os.xEnterMutex();

  /* If some thread using this PID has a lock via a different OsFile*
  ** handle that precludes the requested lock, return BUSY.
  */
  if( (pFile->locktype!=pLock->locktype && 
          (pLock->locktype>=PENDING_LOCK || locktype>SHARED_LOCK))
  ){
    rc = SQLITE_BUSY;
    goto end_lock;
  }

  /* If a SHARED lock is requested, and some thread using this PID already
  ** has a SHARED or RESERVED lock, then increment reference counts and
  ** return SQLITE_OK.
  */
  if( locktype==SHARED_LOCK && 
      (pLock->locktype==SHARED_LOCK || pLock->locktype==RESERVED_LOCK) ){
    assert( locktype==SHARED_LOCK );
    assert( pFile->locktype==0 );
    assert( pLock->cnt>0 );
    pFile->locktype = SHARED_LOCK;
    pLock->cnt++;
    pFile->pOpen->nLock++;
    goto end_lock;
  }

  lock.l_len = 1L;

  lock.l_whence = SEEK_SET;

  /* A PENDING lock is needed before acquiring a SHARED lock and before
  ** acquiring an EXCLUSIVE lock.  For the SHARED lock, the PENDING will
  ** be released.
  */
  if( locktype==SHARED_LOCK 
      || (locktype==EXCLUSIVE_LOCK && pFile->locktype<PENDING_LOCK)
  ){
    lock.l_type = (locktype==SHARED_LOCK?F_RDLCK:F_WRLCK);
    lock.l_start = PENDING_BYTE;
    s = fcntl(pFile->h, F_SETLK, &lock);
    if( s ){
      rc = (errno==EINVAL) ? SQLITE_NOLFS : SQLITE_BUSY;
      goto end_lock;
    }
  }


  /* If control gets to this point, then actually go ahead and make
  ** operating system calls for the specified lock.
  */
  if( locktype==SHARED_LOCK ){
    assert( pLock->cnt==0 );
    assert( pLock->locktype==0 );

    /* Now get the read-lock */
    lock.l_start = SHARED_FIRST;
    lock.l_len = SHARED_SIZE;
    s = fcntl(pFile->h, F_SETLK, &lock);

    /* Drop the temporary PENDING lock */
    lock.l_start = PENDING_BYTE;
    lock.l_len = 1L;
    lock.l_type = F_UNLCK;
    if( fcntl(pFile->h, F_SETLK, &lock)!=0 ){
      rc = SQLITE_IOERR;  /* This should never happen */
      goto end_lock;
    }
    if( s ){
      rc = (errno==EINVAL) ? SQLITE_NOLFS : SQLITE_BUSY;
    }else{
      pFile->locktype = SHARED_LOCK;
      pFile->pOpen->nLock++;
      pLock->cnt = 1;
    }
  }else if( locktype==EXCLUSIVE_LOCK && pLock->cnt>1 ){
    /* We are trying for an exclusive lock but another thread in this
    ** same process is still holding a shared lock. */
    rc = SQLITE_BUSY;
  }else{
    /* The request was for a RESERVED or EXCLUSIVE lock.  It is
    ** assumed that there is a SHARED or greater lock on the file
    ** already.
    */
    assert( 0!=pFile->locktype );
    lock.l_type = F_WRLCK;
    switch( locktype ){
      case RESERVED_LOCK:
        lock.l_start = RESERVED_BYTE;
        break;
      case EXCLUSIVE_LOCK:
        lock.l_start = SHARED_FIRST;
        lock.l_len = SHARED_SIZE;
        break;
      default:
        assert(0);
    }
    s = fcntl(pFile->h, F_SETLK, &lock);
    if( s ){
      rc = (errno==EINVAL) ? SQLITE_NOLFS : SQLITE_BUSY;
    }
  }
  
  if( rc==SQLITE_OK ){
    pFile->locktype = locktype;
    pLock->locktype = locktype;
  }else if( locktype==EXCLUSIVE_LOCK ){
    pFile->locktype = PENDING_LOCK;
    pLock->locktype = PENDING_LOCK;
  }

end_lock:
  sqlite3Os.xLeaveMutex();
  TRACE4("LOCK    %d %s %s\n", pFile->h, locktypeName(locktype), 
      rc==SQLITE_OK ? "ok" : "failed");
  return rc;
}

/*
** 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.
**
** It is not possible for this routine to fail if the second argument
** is NO_LOCK.  If the second argument is SHARED_LOCK, this routine
** might return SQLITE_IOERR instead of SQLITE_OK.
*/
static int unixUnlock(OsFile *id, int locktype){
  struct lockInfo *pLock;
  struct flock lock;
  int rc = SQLITE_OK;
  unixFile *pFile = (unixFile*)id;

  assert( pFile );
  TRACE7("UNLOCK  %d %d was %d(%d,%d) pid=%d\n", pFile->h, locktype,
      pFile->locktype, pFile->pLock->locktype, pFile->pLock->cnt, getpid());
  if( CHECK_THREADID(pFile) ) return SQLITE_MISUSE;

  assert( locktype<=SHARED_LOCK );
  if( pFile->locktype<=locktype ){
    return SQLITE_OK;
  }
  sqlite3Os.xEnterMutex();
  pLock = pFile->pLock;
  assert( pLock->cnt!=0 );
  if( pFile->locktype>SHARED_LOCK ){
    assert( pLock->locktype==pFile->locktype );
    if( locktype==SHARED_LOCK ){
      lock.l_type = F_RDLCK;
      lock.l_whence = SEEK_SET;
      lock.l_start = SHARED_FIRST;
      lock.l_len = SHARED_SIZE;
      if( fcntl(pFile->h, F_SETLK, &lock)!=0 ){
        /* This should never happen */
        rc = SQLITE_IOERR;
      }
    }
    lock.l_type = F_UNLCK;
    lock.l_whence = SEEK_SET;
    lock.l_start = PENDING_BYTE;
    lock.l_len = 2L;  assert( PENDING_BYTE+1==RESERVED_BYTE );
    if( fcntl(pFile->h, F_SETLK, &lock)==0 ){
      pLock->locktype = SHARED_LOCK;
    }else{
      rc = SQLITE_IOERR;  /* This should never happen */
    }
  }
  if( locktype==NO_LOCK ){
    struct openCnt *pOpen;

    /* Decrement the shared lock counter.  Release the lock using an
    ** OS call only when all threads in this same process have released
    ** the lock.
    */
    pLock->cnt--;
    if( pLock->cnt==0 ){
      lock.l_type = F_UNLCK;
      lock.l_whence = SEEK_SET;
      lock.l_start = lock.l_len = 0L;
      if( fcntl(pFile->h, F_SETLK, &lock)==0 ){
        pLock->locktype = NO_LOCK;
      }else{
        rc = SQLITE_IOERR;  /* This should never happen */
      }
    }

    /* Decrement the count of locks against this same file.  When the
    ** count reaches zero, close any other file descriptors whose close
    ** was deferred because of outstanding locks.
    */
    pOpen = pFile->pOpen;
    pOpen->nLock--;
    assert( pOpen->nLock>=0 );
    if( pOpen->nLock==0 && pOpen->nPending>0 ){
      int i;
      for(i=0; i<pOpen->nPending; i++){
        close(pOpen->aPending[i]);
      }
      sqliteFree(pOpen->aPending);
      pOpen->nPending = 0;
      pOpen->aPending = 0;
    }
  }
  sqlite3Os.xLeaveMutex();
  pFile->locktype = locktype;
  return rc;
}

/*
** Close a file.
*/
static int unixClose(OsFile **pId){
  unixFile *id = (unixFile*)*pId;
  if( !id ) return SQLITE_OK;
  if( CHECK_THREADID(id) ) return SQLITE_MISUSE;
  unixUnlock(*pId, NO_LOCK);
  if( id->dirfd>=0 ) close(id->dirfd);
  id->dirfd = -1;
  sqlite3Os.xEnterMutex();
  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;
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
    }
  }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;
  TRACE2("CLOSE   %-3d\n", id->h);
  OpenCounter(-1);
  sqliteFree(id);
  *pId = 0;
  return SQLITE_OK;
}







|







1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
    }
  }else{
    /* There are no outstanding locks so we can close the file immediately */
    close(id->h);
  }
  releaseLockInfo(id->pLock);
  releaseOpenCnt(id->pOpen);
  sqlite3Os.xLeaveMutex();
  id->isOpen = 0;
  TRACE2("CLOSE   %-3d\n", id->h);
  OpenCounter(-1);
  sqliteFree(id);
  *pId = 0;
  return SQLITE_OK;
}
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483

1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498

1499
1500
1501
1502
1503
1504

1505

1506






1507
1508




1509







1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
  return zFull;
}

/*
** Change the value of the fullsync flag in the given file descriptor.
*/
static void unixSetFullSync(OsFile *id, int v){
  id->fullSync = v;
}

/*
** Return the underlying file handle for an OsFile
*/
static int unixFileHandle(OsFile *id){
  return id->h;
}

/*
** Return an integer that indices the type of lock currently held
** by this handle.  (Used for testing and analysis only.)
*/
static int unixLockState(OsFile *id){
  return id->locktype;
}

/*
** This is the structure that defines all of the I/O routines.

*/
struct sqlite3IoVtbl sqlite3Io = {
  unixDelete,
  unixFileExists,
  unixOpenReadWrite,
  unixOpenExclusive,
  unixOpenReadOnly,
  unixOpenDirectory,
  unixSyncDirectory,
  unixTempFileName,
  unixIsDirWritable,
  unixClose,
  unixRead,
  unixWrite,
  unixSeek,

  unixSync,
  unixTruncate,
  unixFileSize,
  unixFullPathname,
  unixLock,
  unixUnlock,

  unixCheckReservedLock,

  unixSetFullSync,






  unixFileHandle,
  unixLockState,




};









#endif /* SQLITE_OMIT_DISKIO */
/***************************************************************************
** Everything above deals with file I/O.  Everything that follows deals
** with other miscellanous aspects of the operating system interface
****************************************************************************/


/*
** Get information to seed the random number generator.  The seed
** is written into the buffer zBuf[256].  The calling function must
** supply a sufficiently large buffer.
*/
int sqlite3OsRandomSeed(char *zBuf){
  /* We have to initialize zBuf to prevent valgrind from reporting
  ** errors.  The reports issued by valgrind are incorrect - we would
  ** prefer that the randomness be increased by making use of the
  ** uninitialized space in zBuf - but valgrind errors tend to worry
  ** some users.  Rather than argue, it seems easier just to initialize
  ** the whole array and silence valgrind, even if that means less randomness
  ** in the random seed.







|






|







|



|
>

|
|
<
<
<
<

<
<
<
<



>

|
|
|


>

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














|







1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461




1462




1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
  return zFull;
}

/*
** Change the value of the fullsync flag in the given file descriptor.
*/
static void unixSetFullSync(OsFile *id, int v){
  ((unixFile*)id)->fullSync = v;
}

/*
** Return the underlying file handle for an OsFile
*/
static int unixFileHandle(OsFile *id){
  return ((unixFile*)id)->h;
}

/*
** Return an integer that indices the type of lock currently held
** by this handle.  (Used for testing and analysis only.)
*/
static int unixLockState(OsFile *id){
  return ((unixFile*)id)->locktype;
}

/*
** 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,
};

/*
** Allocate memory for a unixFile.  Initialize the new unixFile
** to the value given in pInit and return a pointer to the new
** OsFile.  If we run out of memory, close the file and return NULL.
*/
static int allocateUnixFile(unixFile *pInit, OsFile **pId){
  unixFile *pNew;
  pNew = sqliteMalloc( sizeof(unixFile) );
  if( pNew==0 ){
    close(pInit->h);
    *pId = 0;
    return SQLITE_NOMEM;
  }else{
    *pNew = *pInit;
    pNew->pMethod = &sqlite3UnixIoMethod;
    *pId = (OsFile*)pNew;
    OpenCounter(+1);
    return SQLITE_OK;
  }
}


#endif /* SQLITE_OMIT_DISKIO */
/***************************************************************************
** Everything above deals with file I/O.  Everything that follows deals
** with other miscellanous aspects of the operating system interface
****************************************************************************/


/*
** Get information to seed the random number generator.  The seed
** is written into the buffer zBuf[256].  The calling function must
** supply a sufficiently large buffer.
*/
static int unixRandomSeed(char *zBuf){
  /* We have to initialize zBuf to prevent valgrind from reporting
  ** errors.  The reports issued by valgrind are incorrect - we would
  ** prefer that the randomness be increased by making use of the
  ** uninitialized space in zBuf - but valgrind errors tend to worry
  ** some users.  Rather than argue, it seems easier just to initialize
  ** the whole array and silence valgrind, even if that means less randomness
  ** in the random seed.
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
#endif
  return SQLITE_OK;
}

/*
** Sleep for a little while.  Return the amount of time slept.
*/
int sqlite3OsSleep(int ms){
#if defined(HAVE_USLEEP) && HAVE_USLEEP
  usleep(ms*1000);
  return ms;
#else
  sleep((ms+999)/1000);
  return 1000*((ms+999)/1000);
#endif







|







1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
#endif
  return SQLITE_OK;
}

/*
** Sleep for a little while.  Return the amount of time slept.
*/
static int unixSleep(int ms){
#if defined(HAVE_USLEEP) && HAVE_USLEEP
  usleep(ms*1000);
  return ms;
#else
  sleep((ms+999)/1000);
  return 1000*((ms+999)/1000);
#endif
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
1624
1625
1626
1627
1628
1629
1630






























1631
** The following pair of routine implement mutual exclusion for
** multi-threaded processes.  Only a single thread is allowed to
** executed code that is surrounded by EnterMutex() and LeaveMutex().
**
** SQLite uses only a single Mutex.  There is not much critical
** code and what little there is executes quickly and without blocking.
*/
void sqlite3OsEnterMutex(){
#ifdef SQLITE_UNIX_THREADS
  pthread_mutex_lock(&mutex);
#endif
  assert( !inMutex );
  inMutex = 1;
}
void sqlite3OsLeaveMutex(){
  assert( inMutex );
  inMutex = 0;
#ifdef SQLITE_UNIX_THREADS
  pthread_mutex_unlock(&mutex);
#endif
}

/*
** The following variable, if set to a non-zero value, becomes the result
** returned from sqlite3OsCurrentTime().  This is used for testing.
*/
#ifdef SQLITE_TEST
int sqlite3_current_time = 0;
#endif

/*
** Find the current time (in Universal Coordinated Time).  Write the
** current time and date as a Julian Day number into *prNow and
** return 0.  Return 1 if the time and date cannot be found.
*/
int sqlite3OsCurrentTime(double *prNow){
#ifdef NO_GETTOD
  time_t t;
  time(&t);
  *prNow = t/86400.0 + 2440587.5;
#else
  struct timeval sNow;
  struct timezone sTz;  /* Not used */
  gettimeofday(&sNow, &sTz);
  *prNow = 2440587.5 + sNow.tv_sec/86400.0 + sNow.tv_usec/86400000000.0;
#endif
#ifdef SQLITE_TEST
  if( sqlite3_current_time ){
    *prNow = sqlite3_current_time/86400.0 + 2440587.5;
  }
#endif
  return 0;
}































#endif /* OS_UNIX */







|






|









|










|


















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

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
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
** The following pair of routine implement mutual exclusion for
** multi-threaded processes.  Only a single thread is allowed to
** executed code that is surrounded by EnterMutex() and LeaveMutex().
**
** SQLite uses only a single Mutex.  There is not much critical
** code and what little there is executes quickly and without blocking.
*/
static void unixEnterMutex(){
#ifdef SQLITE_UNIX_THREADS
  pthread_mutex_lock(&mutex);
#endif
  assert( !inMutex );
  inMutex = 1;
}
static void unixLeaveMutex(){
  assert( inMutex );
  inMutex = 0;
#ifdef SQLITE_UNIX_THREADS
  pthread_mutex_unlock(&mutex);
#endif
}

/*
** The following variable, if set to a non-zero value, becomes the result
** returned from sqlite3Os.xCurrentTime().  This is used for testing.
*/
#ifdef SQLITE_TEST
int sqlite3_current_time = 0;
#endif

/*
** Find the current time (in Universal Coordinated Time).  Write the
** current time and date as a Julian Day number into *prNow and
** return 0.  Return 1 if the time and date cannot be found.
*/
static int unixCurrentTime(double *prNow){
#ifdef NO_GETTOD
  time_t t;
  time(&t);
  *prNow = t/86400.0 + 2440587.5;
#else
  struct timeval sNow;
  struct timezone sTz;  /* Not used */
  gettimeofday(&sNow, &sTz);
  *prNow = 2440587.5 + sNow.tv_sec/86400.0 + sNow.tv_usec/86400000000.0;
#endif
#ifdef SQLITE_TEST
  if( sqlite3_current_time ){
    *prNow = sqlite3_current_time/86400.0 + 2440587.5;
  }
#endif
  return 0;
}

/* Macro used to comment out routines that do not exists when there is
** no disk I/O */
#ifdef SQLITE_OMIT_DISKIO
# define IF_DISKIO(X)  0
#else
# define IF_DISKIO(X)  X
#endif

/*
** This is the structure that defines all of the I/O routines.
*/
struct sqlite3OsVtbl sqlite3Os = {
  IF_DISKIO( unixOpenReadWrite ),
  IF_DISKIO( unixOpenExclusive ),
  IF_DISKIO( unixOpenReadOnly ),
  IF_DISKIO( unixDelete ),
  IF_DISKIO( unixFileExists ),
  IF_DISKIO( unixFullPathname ),
  IF_DISKIO( unixIsDirWritable ),
  IF_DISKIO( unixSyncDirectory ),
  IF_DISKIO( unixTempFileName ),
  unixRandomSeed,
  unixSleep,
  unixCurrentTime,
  unixEnterMutex,
  unixLeaveMutex,
};



#endif /* OS_UNIX */
Changes to src/os_win.c.
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

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

/*
** The OsFile structure is a operating-system independing representation
** of an open file handle.  It is defined differently for each architecture.
**
** This is the definition for Win32.
*/

struct OsFile {

  HANDLE h;               /* Handle for accessing the file */
  unsigned char locktype; /* Type of lock currently held on this file */
  short sharedLockByte;   /* Randomly chosen byte used as a shared lock */
};



/*
** Do not include any of the File I/O interface procedures if the
** SQLITE_OMIT_DISKIO macro is defined (indicating that there database
** will be in-memory only)
*/







|
<
|
<

>
|
>




<







31
32
33
34
35
36
37
38

39

40
41
42
43
44
45
46
47

48
49
50
51
52
53
54

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

/*
** The winFile structure is a subclass of OsFile specific to the win32

** portability layer.

*/
typedef struct winFile winFile;
struct winFile {
  IoMethod const *pMethod;/* Must be first */
  HANDLE h;               /* Handle for accessing the file */
  unsigned char locktype; /* Type of lock currently held on this file */
  short sharedLockByte;   /* Randomly chosen byte used as a shared lock */
};



/*
** Do not include any of the File I/O interface procedures if the
** SQLITE_OMIT_DISKIO macro is defined (indicating that there database
** will be in-memory only)
*/
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
    sqliteFree(zWide);
  }else{
    exists = GetFileAttributesA(zFilename) != 0xffffffff;
  }
  return exists;
}


/*
** Allocate memory for an OsFile.  Initialize the new OsFile
** to the value given in pInit and return a pointer to the new
** OsFile.  If we run out of memory, close the file and return NULL.
*/
static OsFile *allocateOsFile(OsFile *pInit){
  OsFile *pNew;
  pNew = sqliteMalloc( sizeof(OsFile) );
  if( pNew==0 ){
    CloseHandle(pInit->h);
  }else{
    *pNew = *pInit;
  }
  return pNew;
}

/*
** 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.
*/
static int winOpenReadWrite(
  const char *zFilename,
  OsFile **pId,
  int *pReadonly
){
  OsFile f;
  HANDLE h;
  WCHAR *zWide = utf8ToUnicode(zFilename);
  assert( *pId==0 );
  if( zWide ){
    h = CreateFileW(zWide,
       GENERIC_READ | GENERIC_WRITE,
       FILE_SHARE_READ | FILE_SHARE_WRITE,







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



















|







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
    sqliteFree(zWide);
  }else{
    exists = GetFileAttributesA(zFilename) != 0xffffffff;
  }
  return exists;
}

/* Forward declaration */






int allocateWinFile(winFile *pInit, OsFile **pId);









/*
** 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.
*/
static int winOpenReadWrite(
  const char *zFilename,
  OsFile **pId,
  int *pReadonly
){
  winFile f;
  HANDLE h;
  WCHAR *zWide = utf8ToUnicode(zFilename);
  assert( *pId==0 );
  if( zWide ){
    h = CreateFileW(zWide,
       GENERIC_READ | GENERIC_WRITE,
       FILE_SHARE_READ | FILE_SHARE_WRITE,
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
      *pReadonly = 0;
    }
  }
  f.h = h;
  f.locktype = NO_LOCK;
  f.sharedLockByte = 0;
  TRACE3("OPEN R/W %d \"%s\"\n", h, zFilename);
  *pId = allocateOsFile(&f);
  if( *pId==0 ){
    return SQLITE_NOMEM;
  }else{
    OpenCounter(+1);
    return SQLITE_OK;
  }
}


/*
** 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.
*/
static int winOpenExclusive(const char *zFilename, OsFile **pId, int delFlag){
  OsFile f;
  HANDLE h;
  int fileflags;
  WCHAR *zWide = utf8ToUnicode(zFilename);
  assert( *pId == 0 );
  if( delFlag ){
    fileflags = FILE_ATTRIBUTE_TEMPORARY | FILE_FLAG_RANDOM_ACCESS 
                     | FILE_FLAG_DELETE_ON_CLOSE;







|
<
<
<
<
<
<


















|







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
      *pReadonly = 0;
    }
  }
  f.h = h;
  f.locktype = NO_LOCK;
  f.sharedLockByte = 0;
  TRACE3("OPEN R/W %d \"%s\"\n", h, zFilename);
  return allocateWinFile(&f, pId);






}


/*
** 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.
*/
static int winOpenExclusive(const char *zFilename, OsFile **pId, int delFlag){
  winFile f;
  HANDLE h;
  int fileflags;
  WCHAR *zWide = utf8ToUnicode(zFilename);
  assert( *pId == 0 );
  if( delFlag ){
    fileflags = FILE_ATTRIBUTE_TEMPORARY | FILE_FLAG_RANDOM_ACCESS 
                     | FILE_FLAG_DELETE_ON_CLOSE;
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
  if( h==INVALID_HANDLE_VALUE ){
    return SQLITE_CANTOPEN;
  }
  f.h = h;
  f.locktype = NO_LOCK;
  f.sharedLockByte = 0;
  TRACE3("OPEN EX %d \"%s\"\n", h, zFilename);
  *pId = allocateOsFile(&f);
  if( *pId==0 ){
    return SQLITE_NOMEM;
  }else{
    OpenCounter(+1);
    return SQLITE_OK;
  }
}

/*
** 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.
*/
static int winOpenReadOnly(const char *zFilename, OsFile **pId){
  OsFile f;
  HANDLE h;
  WCHAR *zWide = utf8ToUnicode(zFilename);
  assert( *pId==0 );
  if( zWide ){
    h = CreateFileW(zWide,
       GENERIC_READ,
       0,







|
<
<
<
<
<
<










|







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
  if( h==INVALID_HANDLE_VALUE ){
    return SQLITE_CANTOPEN;
  }
  f.h = h;
  f.locktype = NO_LOCK;
  f.sharedLockByte = 0;
  TRACE3("OPEN EX %d \"%s\"\n", h, zFilename);
  return allocateWinFile(&f, pId);






}

/*
** 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.
*/
static int winOpenReadOnly(const char *zFilename, OsFile **pId){
  winFile f;
  HANDLE h;
  WCHAR *zWide = utf8ToUnicode(zFilename);
  assert( *pId==0 );
  if( zWide ){
    h = CreateFileW(zWide,
       GENERIC_READ,
       0,
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
  if( h==INVALID_HANDLE_VALUE ){
    return SQLITE_CANTOPEN;
  }
  f.h = h;
  f.locktype = NO_LOCK;
  f.sharedLockByte = 0;
  TRACE3("OPEN RO %d \"%s\"\n", h, zFilename);
  *pId = allocateOsFile(&f);
  if( *pId==0 ){
    return SQLITE_NOMEM;
  }else{
    OpenCounter(+1);
    return SQLITE_OK;
  }
}

/*
** 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.
**
** This routine is only meaningful for Unix.  It is a no-op under
** windows since windows does not support hard links.
**
** On success, a handle for a previously open file is at *id is
** updated with the new directory file descriptor and SQLITE_OK is
** returned.
**
** On failure, the function returns SQLITE_CANTOPEN and leaves
** *id unchanged.
*/
static int winOpenDirectory(
  const char *zDirname,
  OsFile *id

){
  return SQLITE_OK;
}

/*
** If the following global variable points to a string which is the
** name of a directory, then that directory will be used to store







|
<
<
<
<
<
<



















<
|
>







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
  if( h==INVALID_HANDLE_VALUE ){
    return SQLITE_CANTOPEN;
  }
  f.h = h;
  f.locktype = NO_LOCK;
  f.sharedLockByte = 0;
  TRACE3("OPEN RO %d \"%s\"\n", h, zFilename);
  return allocateWinFile(&f, pId);






}

/*
** 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.
**
** This routine is only meaningful for Unix.  It is a no-op under
** windows since windows does not support hard links.
**
** On success, a handle for a previously open file is at *id is
** updated with the new directory file descriptor and SQLITE_OK is
** returned.
**
** On failure, the function returns SQLITE_CANTOPEN and leaves
** *id unchanged.
*/
static int winOpenDirectory(

  OsFile *id,
  const char *zDirname
){
  return SQLITE_OK;
}

/*
** If the following global variable points to a string which is the
** name of a directory, then that directory will be used to store
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
    sprintf(zBuf, "%s\\"TEMP_FILE_PREFIX, zTempPath);
    j = strlen(zBuf);
    sqlite3Randomness(15, &zBuf[j]);
    for(i=0; i<15; i++, j++){
      zBuf[j] = (char)zChars[ ((unsigned char)zBuf[j])%(sizeof(zChars)-1) ];
    }
    zBuf[j] = 0;
    if( !sqlite3Io.xFileExists(zBuf) ) break;
  }
  TRACE2("TEMP FILENAME: %s\n", zBuf);
  return SQLITE_OK; 
}

/*
** Close a file.
*/
static int winClose(OsFile **pId){
  if( *pId ){

    TRACE2("CLOSE %d\n", (*pId)->h);
    CloseHandle((*pId)->h);
    OpenCounter(-1);
    sqliteFree(*pId);
    *pId = 0;
  }
  return SQLITE_OK;
}

/*
** 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 winRead(OsFile *id, void *pBuf, int amt){
  DWORD got;
  assert( id!=0 );
  SimulateIOError(SQLITE_IOERR);
  TRACE3("READ %d lock=%d\n", id->h, id->locktype);
  if( !ReadFile(id->h, pBuf, amt, &got, 0) ){
    got = 0;
  }
  if( got==(DWORD)amt ){
    return SQLITE_OK;
  }else{
    return SQLITE_IOERR;
  }
}

/*
** Write data from a buffer into a file.  Return SQLITE_OK on success
** or some other error code on failure.
*/
static int winWrite(OsFile *id, const void *pBuf, int amt){
  int rc = 0;
  DWORD wrote;
  assert( id!=0 );
  SimulateIOError(SQLITE_IOERR);
  SimulateDiskfullError;
  TRACE3("WRITE %d lock=%d\n", id->h, id->locktype);
  assert( amt>0 );
  while( amt>0 && (rc = WriteFile(id->h, pBuf, amt, &wrote, 0))!=0 && wrote>0 ){

    amt -= wrote;
    pBuf = &((char*)pBuf)[wrote];
  }
  if( !rc || amt>(int)wrote ){
    return SQLITE_FULL;
  }
  return SQLITE_OK;







|









|
>
|
|

|














|
|



















|

|
>







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
    sprintf(zBuf, "%s\\"TEMP_FILE_PREFIX, zTempPath);
    j = strlen(zBuf);
    sqlite3Randomness(15, &zBuf[j]);
    for(i=0; i<15; i++, j++){
      zBuf[j] = (char)zChars[ ((unsigned char)zBuf[j])%(sizeof(zChars)-1) ];
    }
    zBuf[j] = 0;
    if( !sqlite3Os.xFileExists(zBuf) ) break;
  }
  TRACE2("TEMP FILENAME: %s\n", zBuf);
  return SQLITE_OK; 
}

/*
** Close a file.
*/
static int winClose(OsFile **pId){
  winFile *pFile;
  if( pId && (pFile = (winFile*)*pId)!=0 ){
    TRACE2("CLOSE %d\n", pFile->h);
    CloseHandle(pFile->h);
    OpenCounter(-1);
    sqliteFree(pFile);
    *pId = 0;
  }
  return SQLITE_OK;
}

/*
** 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 winRead(OsFile *id, void *pBuf, int amt){
  DWORD got;
  assert( id!=0 );
  SimulateIOError(SQLITE_IOERR);
  TRACE3("READ %d lock=%d\n", ((winFile*)id)->h, ((winFile*)id)->locktype);
  if( !ReadFile(((winFile*)id)->h, pBuf, amt, &got, 0) ){
    got = 0;
  }
  if( got==(DWORD)amt ){
    return SQLITE_OK;
  }else{
    return SQLITE_IOERR;
  }
}

/*
** Write data from a buffer into a file.  Return SQLITE_OK on success
** or some other error code on failure.
*/
static int winWrite(OsFile *id, const void *pBuf, int amt){
  int rc = 0;
  DWORD wrote;
  assert( id!=0 );
  SimulateIOError(SQLITE_IOERR);
  SimulateDiskfullError;
  TRACE3("WRITE %d lock=%d\n", ((winFile*)id)->h, ((winFile*)id)->locktype);
  assert( amt>0 );
  while( amt>0 && (rc = WriteFile(((winFile*)id)->h, pBuf, amt, &wrote, 0))!=0
         && wrote>0 ){
    amt -= wrote;
    pBuf = &((char*)pBuf)[wrote];
  }
  if( !rc || amt>(int)wrote ){
    return SQLITE_FULL;
  }
  return SQLITE_OK;
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
  LONG lowerBits = offset & 0xffffffff;
  DWORD rc;
  assert( id!=0 );
#ifdef SQLITE_TEST
  if( offset ) SimulateDiskfullError
#endif
  SEEK(offset/1024 + 1);
  rc = SetFilePointer(id->h, lowerBits, &upperBits, FILE_BEGIN);
  TRACE3("SEEK %d %lld\n", id->h, offset);
  if( rc==INVALID_SET_FILE_POINTER && GetLastError()!=NO_ERROR ){
    return SQLITE_FULL;
  }
  return SQLITE_OK;
}

/*
** Make sure all writes to a particular file are committed to disk.
*/
static int winSync(OsFile *id, int dataOnly){
  assert( id!=0 );
  TRACE3("SYNC %d lock=%d\n", id->h, id->locktype);
  if( FlushFileBuffers(id->h) ){
    return SQLITE_OK;
  }else{
    return SQLITE_IOERR;
  }
}

/*







|
|











|
|







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
  LONG lowerBits = offset & 0xffffffff;
  DWORD rc;
  assert( id!=0 );
#ifdef SQLITE_TEST
  if( offset ) SimulateDiskfullError
#endif
  SEEK(offset/1024 + 1);
  rc = SetFilePointer(((winFile*)id)->h, lowerBits, &upperBits, FILE_BEGIN);
  TRACE3("SEEK %d %lld\n", ((winFile*)id)->h, offset);
  if( rc==INVALID_SET_FILE_POINTER && GetLastError()!=NO_ERROR ){
    return SQLITE_FULL;
  }
  return SQLITE_OK;
}

/*
** Make sure all writes to a particular file are committed to disk.
*/
static int winSync(OsFile *id, int dataOnly){
  assert( id!=0 );
  TRACE3("SYNC %d lock=%d\n", ((winFile*)id)->h, ((winFile*)id)->locktype);
  if( FlushFileBuffers(((winFile*)id)->h) ){
    return SQLITE_OK;
  }else{
    return SQLITE_IOERR;
  }
}

/*
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

/*
** Truncate an open file to a specified size
*/
static int winTruncate(OsFile *id, i64 nByte){
  LONG upperBits = nByte>>32;
  assert( id!=0 );
  TRACE3("TRUNCATE %d %lld\n", id->h, nByte);
  SimulateIOError(SQLITE_IOERR);
  SetFilePointer(id->h, nByte, &upperBits, FILE_BEGIN);
  SetEndOfFile(id->h);
  return SQLITE_OK;
}

/*
** Determine the current size of a file in bytes
*/
static int winFileSize(OsFile *id, i64 *pSize){
  DWORD upperBits, lowerBits;
  assert( id!=0 );
  SimulateIOError(SQLITE_IOERR);
  lowerBits = GetFileSize(id->h, &upperBits);
  *pSize = (((i64)upperBits)<<32) + lowerBits;
  return SQLITE_OK;
}

/*
** Acquire a reader lock.
** Different API routines are called depending on whether or not this
** is Win95 or WinNT.
*/
static int getReadLock(OsFile *id){
  int res;
  if( isNT() ){
    OVERLAPPED ovlp;
    ovlp.Offset = SHARED_FIRST;
    ovlp.OffsetHigh = 0;
    ovlp.hEvent = 0;
    res = LockFileEx(id->h, LOCKFILE_FAIL_IMMEDIATELY, 0, SHARED_SIZE,0,&ovlp);
  }else{
    int lk;
    sqlite3Randomness(sizeof(lk), &lk);
    id->sharedLockByte = (lk & 0x7fffffff)%(SHARED_SIZE - 1);
    res = LockFile(id->h, SHARED_FIRST+id->sharedLockByte, 0, 1, 0);
  }
  return res;
}

/*
** Undo a readlock
*/
static int unlockReadLock(OsFile *id){
  int res;
  if( isNT() ){
    res = UnlockFile(id->h, SHARED_FIRST, 0, SHARED_SIZE, 0);
  }else{
    res = UnlockFile(id->h, SHARED_FIRST + id->sharedLockByte, 0, 1, 0);
  }
  return res;
}

#ifndef SQLITE_OMIT_PAGER_PRAGMAS
/*
** Check that a given pathname is a directory and is writable 







|

|
|










|









|



















|


|

|







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

/*
** Truncate an open file to a specified size
*/
static int winTruncate(OsFile *id, i64 nByte){
  LONG upperBits = nByte>>32;
  assert( id!=0 );
  TRACE3("TRUNCATE %d %lld\n", ((winFile*)id)->h, nByte);
  SimulateIOError(SQLITE_IOERR);
  SetFilePointer(((winFile*)id)->h, nByte, &upperBits, FILE_BEGIN);
  SetEndOfFile(((winFile*)id)->h);
  return SQLITE_OK;
}

/*
** Determine the current size of a file in bytes
*/
static int winFileSize(OsFile *id, i64 *pSize){
  DWORD upperBits, lowerBits;
  assert( id!=0 );
  SimulateIOError(SQLITE_IOERR);
  lowerBits = GetFileSize(((winFile*)id)->h, &upperBits);
  *pSize = (((i64)upperBits)<<32) + lowerBits;
  return SQLITE_OK;
}

/*
** Acquire a reader lock.
** Different API routines are called depending on whether or not this
** is Win95 or WinNT.
*/
static int getReadLock(winFile *id){
  int res;
  if( isNT() ){
    OVERLAPPED ovlp;
    ovlp.Offset = SHARED_FIRST;
    ovlp.OffsetHigh = 0;
    ovlp.hEvent = 0;
    res = LockFileEx(id->h, LOCKFILE_FAIL_IMMEDIATELY, 0, SHARED_SIZE,0,&ovlp);
  }else{
    int lk;
    sqlite3Randomness(sizeof(lk), &lk);
    id->sharedLockByte = (lk & 0x7fffffff)%(SHARED_SIZE - 1);
    res = LockFile(id->h, SHARED_FIRST+id->sharedLockByte, 0, 1, 0);
  }
  return res;
}

/*
** Undo a readlock
*/
static int unlockReadLock(winFile *pFile){
  int res;
  if( isNT() ){
    res = UnlockFile(pFile->h, SHARED_FIRST, 0, SHARED_SIZE, 0);
  }else{
    res = UnlockFile(pFile->h, SHARED_FIRST + pFile->sharedLockByte, 0, 1, 0);
  }
  return res;
}

#ifndef SQLITE_OMIT_PAGER_PRAGMAS
/*
** Check that a given pathname is a directory and is writable 
673
674
675
676
677
678
679

680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
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
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
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
781
782
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
833
834
835
836
837
838
839
840
841
842
843
844
845
** must go straight to locking level 0.
*/
static int winLock(OsFile *id, int locktype){
  int rc = SQLITE_OK;    /* Return code from subroutines */
  int res = 1;           /* Result of a windows lock call */
  int newLocktype;       /* Set id->locktype to this value before exiting */
  int gotPendingLock = 0;/* True if we acquired a PENDING lock this time */


  assert( id!=0 );
  TRACE5("LOCK %d %d was %d(%d)\n",
          id->h, locktype, id->locktype, id->sharedLockByte);

  /* If there is already a lock of this type or more restrictive on the
  ** OsFile, do nothing. Don't use the end_lock: exit path, as
  ** sqlite3OsEnterMutex() hasn't been called yet.
  */
  if( id->locktype>=locktype ){
    return SQLITE_OK;
  }

  /* Make sure the locking sequence is correct
  */
  assert( id->locktype!=NO_LOCK || locktype==SHARED_LOCK );
  assert( locktype!=PENDING_LOCK );
  assert( locktype!=RESERVED_LOCK || id->locktype==SHARED_LOCK );

  /* Lock the PENDING_LOCK byte if we need to acquire a PENDING lock or
  ** a SHARED lock.  If we are acquiring a SHARED lock, the acquisition of
  ** the PENDING_LOCK byte is temporary.
  */
  newLocktype = id->locktype;
  if( id->locktype==NO_LOCK
   || (locktype==EXCLUSIVE_LOCK && id->locktype==RESERVED_LOCK)
  ){
    int cnt = 3;
    while( cnt-->0 && (res = LockFile(id->h, PENDING_BYTE, 0, 1, 0))==0 ){
      /* Try 3 times to get the pending lock.  The pending lock might be
      ** held by another reader process who will release it momentarily.
      */
      TRACE2("could not get a PENDING lock. cnt=%d\n", cnt);
      Sleep(1);
    }
    gotPendingLock = res;
  }

  /* Acquire a shared lock
  */
  if( locktype==SHARED_LOCK && res ){
    assert( id->locktype==NO_LOCK );
    res = getReadLock(id);
    if( res ){
      newLocktype = SHARED_LOCK;
    }
  }

  /* Acquire a RESERVED lock
  */
  if( locktype==RESERVED_LOCK && res ){
    assert( id->locktype==SHARED_LOCK );
    res = LockFile(id->h, RESERVED_BYTE, 0, 1, 0);
    if( res ){
      newLocktype = RESERVED_LOCK;
    }
  }

  /* Acquire a PENDING lock
  */
  if( locktype==EXCLUSIVE_LOCK && res ){
    newLocktype = PENDING_LOCK;
    gotPendingLock = 0;
  }

  /* Acquire an EXCLUSIVE lock
  */
  if( locktype==EXCLUSIVE_LOCK && res ){
    assert( id->locktype>=SHARED_LOCK );
    res = unlockReadLock(id);
    TRACE2("unreadlock = %d\n", res);
    res = LockFile(id->h, SHARED_FIRST, 0, SHARED_SIZE, 0);
    if( res ){
      newLocktype = EXCLUSIVE_LOCK;
    }else{
      TRACE2("error-code = %d\n", GetLastError());
    }
  }

  /* If we are holding a PENDING lock that ought to be released, then
  ** release it now.
  */
  if( gotPendingLock && locktype==SHARED_LOCK ){
    UnlockFile(id->h, PENDING_BYTE, 0, 1, 0);
  }

  /* Update the state of the lock has held in the file descriptor then
  ** return the appropriate result code.
  */
  if( res ){
    rc = SQLITE_OK;
  }else{
    TRACE4("LOCK FAILED %d trying for %d but got %d\n", id->h,
           locktype, newLocktype);
    rc = SQLITE_BUSY;
  }
  id->locktype = newLocktype;
  return rc;
}

/*
** 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, otherwise zero.
*/
static int winCheckReservedLock(OsFile *id){
  int rc;

  assert( id!=0 );
  if( id->locktype>=RESERVED_LOCK ){
    rc = 1;
    TRACE3("TEST WR-LOCK %d %d (local)\n", id->h, rc);
  }else{
    rc = LockFile(id->h, RESERVED_BYTE, 0, 1, 0);
    if( rc ){
      UnlockFile(id->h, RESERVED_BYTE, 0, 1, 0);
    }
    rc = !rc;
    TRACE3("TEST WR-LOCK %d %d (remote)\n", id->h, rc);
  }
  return rc;
}

/*
** Lower the locking level on file descriptor id 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.
**
** It is not possible for this routine to fail if the second argument
** is NO_LOCK.  If the second argument is SHARED_LOCK then this routine
** might return SQLITE_IOERR;
*/
static int winUnlock(OsFile *id, int locktype){
  int type;
  int rc = SQLITE_OK;

  assert( id!=0 );
  assert( locktype<=SHARED_LOCK );
  TRACE5("UNLOCK %d to %d was %d(%d)\n", id->h, locktype,
          id->locktype, id->sharedLockByte);
  type = id->locktype;
  if( type>=EXCLUSIVE_LOCK ){
    UnlockFile(id->h, SHARED_FIRST, 0, SHARED_SIZE, 0);
    if( locktype==SHARED_LOCK && !getReadLock(id) ){
      /* This should never happen.  We should always be able to
      ** reacquire the read lock */
      rc = SQLITE_IOERR;
    }
  }
  if( type>=RESERVED_LOCK ){
    UnlockFile(id->h, RESERVED_BYTE, 0, 1, 0);
  }
  if( locktype==NO_LOCK && type>=SHARED_LOCK ){
    unlockReadLock(id);
  }
  if( type>=PENDING_LOCK ){
    UnlockFile(id->h, PENDING_BYTE, 0, 1, 0);
  }
  id->locktype = locktype;
  return rc;
}

/*
** Turn a relative pathname into a full pathname.  Return a pointer
** to the full pathname stored in space obtained from sqliteMalloc().
** The calling function is responsible for freeing this space once it







>

|

|





|





|

|





|
|
|


|












|
|








|
|















|
|

|











|








|



|










>
|
|

|

|

|


|


















>
|

|
|
|

|
|






|


|


|

|







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
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
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
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
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
781
782
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
** must go straight to locking level 0.
*/
static int winLock(OsFile *id, int locktype){
  int rc = SQLITE_OK;    /* Return code from subroutines */
  int res = 1;           /* Result of a windows lock call */
  int newLocktype;       /* Set id->locktype to this value before exiting */
  int gotPendingLock = 0;/* True if we acquired a PENDING lock this time */
  winFile *pFile = (winFile*)id;

  assert( pFile!=0 );
  TRACE5("LOCK %d %d was %d(%d)\n",
          pFile->h, locktype, pFile->locktype, pFile->sharedLockByte);

  /* If there is already a lock of this type or more restrictive on the
  ** OsFile, do nothing. Don't use the end_lock: exit path, as
  ** sqlite3OsEnterMutex() hasn't been called yet.
  */
  if( pFile->locktype>=locktype ){
    return SQLITE_OK;
  }

  /* Make sure the locking sequence is correct
  */
  assert( pFile->locktype!=NO_LOCK || locktype==SHARED_LOCK );
  assert( locktype!=PENDING_LOCK );
  assert( locktype!=RESERVED_LOCK || pFile->locktype==SHARED_LOCK );

  /* Lock the PENDING_LOCK byte if we need to acquire a PENDING lock or
  ** a SHARED lock.  If we are acquiring a SHARED lock, the acquisition of
  ** the PENDING_LOCK byte is temporary.
  */
  newLocktype = pFile->locktype;
  if( pFile->locktype==NO_LOCK
   || (locktype==EXCLUSIVE_LOCK && pFile->locktype==RESERVED_LOCK)
  ){
    int cnt = 3;
    while( cnt-->0 && (res = LockFile(pFile->h, PENDING_BYTE, 0, 1, 0))==0 ){
      /* Try 3 times to get the pending lock.  The pending lock might be
      ** held by another reader process who will release it momentarily.
      */
      TRACE2("could not get a PENDING lock. cnt=%d\n", cnt);
      Sleep(1);
    }
    gotPendingLock = res;
  }

  /* Acquire a shared lock
  */
  if( locktype==SHARED_LOCK && res ){
    assert( pFile->locktype==NO_LOCK );
    res = getReadLock(pFile);
    if( res ){
      newLocktype = SHARED_LOCK;
    }
  }

  /* Acquire a RESERVED lock
  */
  if( locktype==RESERVED_LOCK && res ){
    assert( pFile->locktype==SHARED_LOCK );
    res = LockFile(pFile->h, RESERVED_BYTE, 0, 1, 0);
    if( res ){
      newLocktype = RESERVED_LOCK;
    }
  }

  /* Acquire a PENDING lock
  */
  if( locktype==EXCLUSIVE_LOCK && res ){
    newLocktype = PENDING_LOCK;
    gotPendingLock = 0;
  }

  /* Acquire an EXCLUSIVE lock
  */
  if( locktype==EXCLUSIVE_LOCK && res ){
    assert( pFile->locktype>=SHARED_LOCK );
    res = unlockReadLock(pFile);
    TRACE2("unreadlock = %d\n", res);
    res = LockFile(pFile->h, SHARED_FIRST, 0, SHARED_SIZE, 0);
    if( res ){
      newLocktype = EXCLUSIVE_LOCK;
    }else{
      TRACE2("error-code = %d\n", GetLastError());
    }
  }

  /* If we are holding a PENDING lock that ought to be released, then
  ** release it now.
  */
  if( gotPendingLock && locktype==SHARED_LOCK ){
    UnlockFile(pFile->h, PENDING_BYTE, 0, 1, 0);
  }

  /* Update the state of the lock has held in the file descriptor then
  ** return the appropriate result code.
  */
  if( res ){
    rc = SQLITE_OK;
  }else{
    TRACE4("LOCK FAILED %d trying for %d but got %d\n", pFile->h,
           locktype, newLocktype);
    rc = SQLITE_BUSY;
  }
  pFile->locktype = newLocktype;
  return rc;
}

/*
** 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, otherwise zero.
*/
static int winCheckReservedLock(OsFile *id){
  int rc;
  winFile *pFile = (winFile*)id;
  assert( pFile!=0 );
  if( pFile->locktype>=RESERVED_LOCK ){
    rc = 1;
    TRACE3("TEST WR-LOCK %d %d (local)\n", pFile->h, rc);
  }else{
    rc = LockFile(pFile->h, RESERVED_BYTE, 0, 1, 0);
    if( rc ){
      UnlockFile(pFile->h, RESERVED_BYTE, 0, 1, 0);
    }
    rc = !rc;
    TRACE3("TEST WR-LOCK %d %d (remote)\n", pFile->h, rc);
  }
  return rc;
}

/*
** Lower the locking level on file descriptor id 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.
**
** It is not possible for this routine to fail if the second argument
** is NO_LOCK.  If the second argument is SHARED_LOCK then this routine
** might return SQLITE_IOERR;
*/
static int winUnlock(OsFile *id, int locktype){
  int type;
  int rc = SQLITE_OK;
  winFile *pFile = (winFile*)id;
  assert( pFile!=0 );
  assert( locktype<=SHARED_LOCK );
  TRACE5("UNLOCK %d to %d was %d(%d)\n", pFile->h, locktype,
          pFile->locktype, pFile->sharedLockByte);
  type = pFile->locktype;
  if( type>=EXCLUSIVE_LOCK ){
    UnlockFile(pFile->h, SHARED_FIRST, 0, SHARED_SIZE, 0);
    if( locktype==SHARED_LOCK && !getReadLock(pFile) ){
      /* This should never happen.  We should always be able to
      ** reacquire the read lock */
      rc = SQLITE_IOERR;
    }
  }
  if( type>=RESERVED_LOCK ){
    UnlockFile(pFile->h, RESERVED_BYTE, 0, 1, 0);
  }
  if( locktype==NO_LOCK && type>=SHARED_LOCK ){
    unlockReadLock(pFile);
  }
  if( type>=PENDING_LOCK ){
    UnlockFile(pFile->h, PENDING_BYTE, 0, 1, 0);
  }
  pFile->locktype = locktype;
  return rc;
}

/*
** Turn a relative pathname into a full pathname.  Return a pointer
** to the full pathname stored in space obtained from sqliteMalloc().
** The calling function is responsible for freeing this space once it
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903

904
905
906
907
908
909
910
911
912
913
914
915
916
917
918

919
920
921
922
923
924

925

926






927





928






929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
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
  return;
}

/*
** Return the underlying file handle for an OsFile
*/
static int winFileHandle(OsFile *id){
  return (int)id->h;
}

/*
** Return an integer that indices the type of lock currently held
** by this handle.  (Used for testing and analysis only.)
*/
static int winLockState(OsFile *id){
  return id->locktype;
}


/*
** This is the structure that defines all of the I/O routines.

*/
struct sqlite3IoVtbl sqlite3Io = {
  winDelete,
  winFileExists,
  winOpenReadWrite,
  winOpenExclusive,
  winOpenReadOnly,
  winOpenDirectory,
  winSyncDirectory,
  winTempFileName,
  winIsDirWritable,
  winClose,
  winRead,
  winWrite,
  winSeek,

  winSync,
  winTruncate,
  winFileSize,
  winFullPathname,
  winLock,
  winUnlock,

  winCheckReservedLock,

  winSetFullSync,






  winFileHandle,





  winLockState,






};

#endif /* SQLITE_OMIT_DISKIO */
/***************************************************************************
** Everything above deals with file I/O.  Everything that follows deals
** with other miscellanous aspects of the operating system interface
****************************************************************************/

/*
** Get information to seed the random number generator.  The seed
** is written into the buffer zBuf[256].  The calling function must
** supply a sufficiently large buffer.
*/
int sqlite3OsRandomSeed(char *zBuf){
  /* We have to initialize zBuf to prevent valgrind from reporting
  ** errors.  The reports issued by valgrind are incorrect - we would
  ** prefer that the randomness be increased by making use of the
  ** uninitialized space in zBuf - but valgrind errors tend to worry
  ** some users.  Rather than argue, it seems easier just to initialize
  ** the whole array and silence valgrind, even if that means less randomness
  ** in the random seed.
  **
  ** When testing, initializing zBuf[] to zero is all we do.  That means
  ** that we always use the same random number sequence.* This makes the
  ** tests repeatable.
  */
  memset(zBuf, 0, 256);
  GetSystemTime((LPSYSTEMTIME)zBuf);
  return SQLITE_OK;
}

/*
** Sleep for a little while.  Return the amount of time slept.
*/
int sqlite3OsSleep(int ms){
  Sleep(ms);
  return ms;
}

/*
** Static variables used for thread synchronization
*/
static int inMutex = 0;
#ifdef SQLITE_W32_THREADS
  static CRITICAL_SECTION cs;
#endif

/*
** The following pair of routine implement mutual exclusion for
** multi-threaded processes.  Only a single thread is allowed to
** executed code that is surrounded by EnterMutex() and LeaveMutex().
**
** SQLite uses only a single Mutex.  There is not much critical
** code and what little there is executes quickly and without blocking.
*/
void sqlite3OsEnterMutex(){
#ifdef SQLITE_W32_THREADS
  static int isInit = 0;
  while( !isInit ){
    static long lock = 0;
    if( InterlockedIncrement(&lock)==1 ){
      InitializeCriticalSection(&cs);
      isInit = 1;
    }else{
      Sleep(1);
    }
  }
  EnterCriticalSection(&cs);
#endif
  assert( !inMutex );
  inMutex = 1;
}
void sqlite3OsLeaveMutex(){
  assert( inMutex );
  inMutex = 0;
#ifdef SQLITE_W32_THREADS
  LeaveCriticalSection(&cs);
#endif
}

/*
** The following variable, if set to a non-zero value, becomes the result
** returned from sqlite3OsCurrentTime().  This is used for testing.
*/
#ifdef SQLITE_TEST
int sqlite3_current_time = 0;
#endif

/*
** Find the current time (in Universal Coordinated Time).  Write the
** current time and date as a Julian Day number into *prNow and
** return 0.  Return 1 if the time and date cannot be found.
*/
int sqlite3OsCurrentTime(double *prNow){
  FILETIME ft;
  /* FILETIME structure is a 64-bit value representing the number of 
     100-nanosecond intervals since January 1, 1601 (= JD 2305813.5). 
  */
  double now;
  GetSystemTimeAsFileTime( &ft );
  now = ((double)ft.dwHighDateTime) * 4294967296.0; 
  *prNow = (now + ft.dwLowDateTime)/864000000000.0 + 2305813.5;
#ifdef SQLITE_TEST
  if( sqlite3_current_time ){
    *prNow = sqlite3_current_time/86400.0 + 2440587.5;
  }
#endif
  return 0;
}






























#endif /* OS_WIN */







|







|


<

|
>

|
|
<
<
<
<

<
<
<
<



>

|
|
|


>

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












|




















|




















|
















|




















|
















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

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
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
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
  return;
}

/*
** Return the underlying file handle for an OsFile
*/
static int winFileHandle(OsFile *id){
  return (int)((winFile*)id)->h;
}

/*
** Return an integer that indices the type of lock currently held
** by this handle.  (Used for testing and analysis only.)
*/
static int winLockState(OsFile *id){
  return ((winFile*)id)->locktype;
}


/*
** This vector defines all the methods that can operate on an OsFile
** for win32.
*/
static const IoMethod sqlite3WinIoMethod = {
  winClose,




  winOpenDirectory,




  winRead,
  winWrite,
  winSeek,
  winTruncate,
  winSync,
  winSetFullSync,
  winFileHandle,
  winFileSize,
  winLock,
  winUnlock,
  winLockState,
  winCheckReservedLock,
};

/*
** Allocate memory for an OsFile.  Initialize the new OsFile
** to the value given in pInit and return a pointer to the new
** OsFile.  If we run out of memory, close the file and return NULL.
*/
int allocateWinFile(winFile *pInit, OsFile **pId){
  winFile *pNew;
  pNew = sqliteMalloc( sizeof(*pNew) );
  if( pNew==0 ){
    CloseHandle(pInit->h);
    *pId = 0;
    return SQLITE_NOMEM;
  }else{
    *pNew = *pInit;
    pNew->pMethod = &sqlite3WinIoMethod;
    *pId = pNew;
    return SQLITE_OK;
  }
}


#endif /* SQLITE_OMIT_DISKIO */
/***************************************************************************
** Everything above deals with file I/O.  Everything that follows deals
** with other miscellanous aspects of the operating system interface
****************************************************************************/

/*
** Get information to seed the random number generator.  The seed
** is written into the buffer zBuf[256].  The calling function must
** supply a sufficiently large buffer.
*/
static int winRandomSeed(char *zBuf){
  /* We have to initialize zBuf to prevent valgrind from reporting
  ** errors.  The reports issued by valgrind are incorrect - we would
  ** prefer that the randomness be increased by making use of the
  ** uninitialized space in zBuf - but valgrind errors tend to worry
  ** some users.  Rather than argue, it seems easier just to initialize
  ** the whole array and silence valgrind, even if that means less randomness
  ** in the random seed.
  **
  ** When testing, initializing zBuf[] to zero is all we do.  That means
  ** that we always use the same random number sequence.* This makes the
  ** tests repeatable.
  */
  memset(zBuf, 0, 256);
  GetSystemTime((LPSYSTEMTIME)zBuf);
  return SQLITE_OK;
}

/*
** Sleep for a little while.  Return the amount of time slept.
*/
static int winSleep(int ms){
  Sleep(ms);
  return ms;
}

/*
** Static variables used for thread synchronization
*/
static int inMutex = 0;
#ifdef SQLITE_W32_THREADS
  static CRITICAL_SECTION cs;
#endif

/*
** The following pair of routine implement mutual exclusion for
** multi-threaded processes.  Only a single thread is allowed to
** executed code that is surrounded by EnterMutex() and LeaveMutex().
**
** SQLite uses only a single Mutex.  There is not much critical
** code and what little there is executes quickly and without blocking.
*/
static void winEnterMutex(){
#ifdef SQLITE_W32_THREADS
  static int isInit = 0;
  while( !isInit ){
    static long lock = 0;
    if( InterlockedIncrement(&lock)==1 ){
      InitializeCriticalSection(&cs);
      isInit = 1;
    }else{
      Sleep(1);
    }
  }
  EnterCriticalSection(&cs);
#endif
  assert( !inMutex );
  inMutex = 1;
}
static void winLeaveMutex(){
  assert( inMutex );
  inMutex = 0;
#ifdef SQLITE_W32_THREADS
  LeaveCriticalSection(&cs);
#endif
}

/*
** The following variable, if set to a non-zero value, becomes the result
** returned from sqlite3OsCurrentTime().  This is used for testing.
*/
#ifdef SQLITE_TEST
int sqlite3_current_time = 0;
#endif

/*
** Find the current time (in Universal Coordinated Time).  Write the
** current time and date as a Julian Day number into *prNow and
** return 0.  Return 1 if the time and date cannot be found.
*/
static int winCurrentTime(double *prNow){
  FILETIME ft;
  /* FILETIME structure is a 64-bit value representing the number of 
     100-nanosecond intervals since January 1, 1601 (= JD 2305813.5). 
  */
  double now;
  GetSystemTimeAsFileTime( &ft );
  now = ((double)ft.dwHighDateTime) * 4294967296.0; 
  *prNow = (now + ft.dwLowDateTime)/864000000000.0 + 2305813.5;
#ifdef SQLITE_TEST
  if( sqlite3_current_time ){
    *prNow = sqlite3_current_time/86400.0 + 2440587.5;
  }
#endif
  return 0;
}

/* Macro used to comment out routines that do not exists when there is
** no disk I/O
*/
#ifdef SQLITE_OMIT_DISKIO
# define IF_DISKIO(X)  0
#else
# define IF_DISKIO(X)  X
#endif

/*
** This is the structure that defines all of the I/O routines.
*/
struct sqlite3OsVtbl sqlite3Os = {
  IF_DISKIO( winOpenReadWrite ),
  IF_DISKIO( winOpenExclusive ),
  IF_DISKIO( winOpenReadOnly ),
  IF_DISKIO( winDelete ),
  IF_DISKIO( winFileExists ),
  IF_DISKIO( winFullPathname ),
  IF_DISKIO( winIsDirWritable ),
  IF_DISKIO( winSyncDirectory ),
  IF_DISKIO( winTempFileName ),
  winRandomSeed,
  winSleep,
  winCurrentTime,
  winEnterMutex,
  winLeaveMutex,
};

#endif /* OS_WIN */
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 1.221 2005/11/29 03:13:22 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 1.222 2005/11/30 03:20:31 drh Exp $
*/
#ifndef SQLITE_OMIT_DISKIO
#include "sqliteInt.h"
#include "os.h"
#include "pager.h"
#include <assert.h>
#include <string.h>
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
** 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) FILEHANDLEID(&(p)->fd)
#define FILEHANDLEID(fd) (sqlite3Io.xFileHandle(&fd))

/*
** The page cache as a whole is always in one of the following
** states:
**
**   PAGER_UNLOCK        The page cache is not currently reading or 
**                       writing the database file.  There is no







|







49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
** 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) FILEHANDLEID(&(p)->fd)
#define FILEHANDLEID(fd) (sqlite3OsFileHandle(&fd))

/*
** The page cache as a whole is always in one of the following
** states:
**
**   PAGER_UNLOCK        The page cache is not currently reading or 
**                       writing the database file.  There is no
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
** error code is something goes wrong.
**
** All values are stored on disk as big-endian.
*/
static int read32bits(OsFile *fd, u32 *pRes){
  u32 res;
  int rc;
  rc = sqlite3Io.xRead(fd, &res, sizeof(res));
  if( rc==SQLITE_OK ){
    unsigned char ac[4];
    memcpy(ac, &res, 4);
    res = (ac[0]<<24) | (ac[1]<<16) | (ac[2]<<8) | ac[3];
  }
  *pRes = res;
  return rc;
}

/*
** 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){
  unsigned char ac[4];
  ac[0] = (val>>24) & 0xff;
  ac[1] = (val>>16) & 0xff;
  ac[2] = (val>>8) & 0xff;
  ac[3] = val & 0xff;
  return sqlite3Io.xWrite(fd, ac, 4);
}

/*
** Write the 32-bit integer 'val' into the page identified by page header
** 'p' at offset 'offset'.
*/
static void store32bits(u32 val, PgHdr *p, int offset){







|



















|







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
** error code is something goes wrong.
**
** All values are stored on disk as big-endian.
*/
static int read32bits(OsFile *fd, u32 *pRes){
  u32 res;
  int rc;
  rc = sqlite3OsRead(fd, &res, sizeof(res));
  if( rc==SQLITE_OK ){
    unsigned char ac[4];
    memcpy(ac, &res, 4);
    res = (ac[0]<<24) | (ac[1]<<16) | (ac[2]<<8) | ac[3];
  }
  *pRes = res;
  return rc;
}

/*
** 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){
  unsigned char ac[4];
  ac[0] = (val>>24) & 0xff;
  ac[1] = (val>>16) & 0xff;
  ac[2] = (val>>8) & 0xff;
  ac[3] = val & 0xff;
  return sqlite3OsWrite(fd, ac, 4);
}

/*
** Write the 32-bit integer 'val' into the page identified by page header
** 'p' at offset 'offset'.
*/
static void store32bits(u32 val, PgHdr *p, int offset){
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
  i64 szJ;
  u32 cksum;
  int i;
  unsigned char aMagic[8]; /* A buffer to hold the magic header */

  *pzMaster = 0;

  rc = sqlite3Io.xFileSize(pJrnl, &szJ);
  if( rc!=SQLITE_OK || szJ<16 ) return rc;

  rc = sqlite3Io.xSeek(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 = sqlite3Io.xRead(pJrnl, aMagic, 8);
  if( rc!=SQLITE_OK || memcmp(aMagic, aJournalMagic, 8) ) return rc;

  rc = sqlite3Io.xSeek(pJrnl, szJ-16-len);
  if( rc!=SQLITE_OK ) return rc;

  *pzMaster = (char *)sqliteMalloc(len+1);
  if( !*pzMaster ){
    return SQLITE_NOMEM;
  }
  rc = sqlite3Io.xRead(pJrnl, *pzMaster, len);
  if( rc!=SQLITE_OK ){
    sqliteFree(*pzMaster);
    *pzMaster = 0;
    return rc;
  }

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







|


|








|


|






|







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
  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 */
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
  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 sqlite3Io.xSeek(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.
**







|







586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
  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.
**
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
  **
  ** 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.
  */
  rc = sqlite3Io.xWrite(pPager->jfd, aJournalMagic, sizeof(aJournalMagic));

  if( rc==SQLITE_OK ){
    /* The nRec Field. 0xFFFFFFFF for no-sync journals. */
    rc = write32bits(pPager->jfd, pPager->noSync ? 0xffffffff : 0);
  }
  if( rc==SQLITE_OK ){
    /* The random check-hash initialiser */ 







|







622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
  **
  ** 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.
  */
  rc = sqlite3OsWrite(pPager->jfd, aJournalMagic, sizeof(aJournalMagic));

  if( rc==SQLITE_OK ){
    /* The nRec Field. 0xFFFFFFFF for no-sync journals. */
    rc = write32bits(pPager->jfd, pPager->noSync ? 0xffffffff : 0);
  }
  if( rc==SQLITE_OK ){
    /* The random check-hash initialiser */ 
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
    rc = write32bits(pPager->jfd, pPager->sectorSize);
  }

  /* The journal header has been written successfully. Seek the journal
  ** file descriptor to the end of the journal header sector.
  */
  if( rc==SQLITE_OK ){
    rc = sqlite3Io.xSeek(pPager->jfd, pPager->journalOff-1);
    if( rc==SQLITE_OK ){
      rc = sqlite3Io.xWrite(pPager->jfd, "\000", 1);
    }
  }
  return rc;
}

/*
** The journal file must be open when this is called. A journal header file







|

|







646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
    rc = write32bits(pPager->jfd, pPager->sectorSize);
  }

  /* The journal header has been written successfully. Seek the journal
  ** file descriptor to the end of the journal header sector.
  */
  if( rc==SQLITE_OK ){
    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
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
  rc = seekJournalHdr(pPager);
  if( rc ) return rc;

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

  rc = sqlite3Io.xRead(pPager->jfd, aMagic, sizeof(aMagic));
  if( rc ) return rc;

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

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







|







686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
  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);
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
  ** 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 = sqlite3Io.xSeek(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







|







712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
  ** 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
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
    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 = sqlite3Io.xWrite(pPager->jfd, zMaster, len);
  if( rc!=SQLITE_OK ) return rc;

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

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

  rc = sqlite3Io.xWrite(pPager->jfd, aJournalMagic, 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.







|








|







760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
    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;

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

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

  rc = sqlite3OsWrite(pPager->jfd, aJournalMagic, 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.
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
  pPager->pLast = 0;
  pPager->pAll = 0;
  memset(pPager->aHash, 0, sizeof(pPager->aHash));
  pPager->nPage = 0;
  if( pPager->state>=PAGER_RESERVED ){
    sqlite3pager_rollback(pPager);
  }
  sqlite3Io.xUnlock(pPager->fd, NO_LOCK);
  pPager->state = PAGER_UNLOCK;
  pPager->dbSize = -1;
  pPager->nRef = 0;
  assert( pPager->journalOpen==0 );
}

/*







|







847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
  pPager->pLast = 0;
  pPager->pAll = 0;
  memset(pPager->aHash, 0, sizeof(pPager->aHash));
  pPager->nPage = 0;
  if( pPager->state>=PAGER_RESERVED ){
    sqlite3pager_rollback(pPager);
  }
  sqlite3OsUnlock(pPager->fd, NO_LOCK);
  pPager->state = PAGER_UNLOCK;
  pPager->dbSize = -1;
  pPager->nRef = 0;
  assert( pPager->journalOpen==0 );
}

/*
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
  int rc;
  assert( !MEMDB );
  if( pPager->state<PAGER_RESERVED ){
    return SQLITE_OK;
  }
  sqlite3pager_stmt_commit(pPager);
  if( pPager->stmtOpen ){
    sqlite3Io.xClose(&pPager->stfd);
    pPager->stmtOpen = 0;
  }
  if( pPager->journalOpen ){
    sqlite3Io.xClose(&pPager->jfd);
    pPager->journalOpen = 0;
    sqlite3Io.xDelete(pPager->zJournal);
    sqliteFree( pPager->aInJournal );
    pPager->aInJournal = 0;
    for(pPg=pPager->pAll; pPg; pPg=pPg->pNextAll){
      pPg->inJournal = 0;
      pPg->dirty = 0;
      pPg->needSync = 0;
#ifdef SQLITE_CHECK_PAGES
      pPg->pageHash = pager_pagehash(pPg);
#endif
    }
    pPager->dirtyCache = 0;
    pPager->nRec = 0;
  }else{
    assert( pPager->aInJournal==0 );
    assert( pPager->dirtyCache==0 || pPager->useJournal==0 );
  }
  rc = sqlite3Io.xUnlock(pPager->fd, SHARED_LOCK);
  pPager->state = PAGER_SHARED;
  pPager->origDbSize = 0;
  pPager->setMaster = 0;
  return rc;
}

/*







|



|

|
















|







896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
  int rc;
  assert( !MEMDB );
  if( pPager->state<PAGER_RESERVED ){
    return SQLITE_OK;
  }
  sqlite3pager_stmt_commit(pPager);
  if( pPager->stmtOpen ){
    sqlite3OsClose(&pPager->stfd);
    pPager->stmtOpen = 0;
  }
  if( pPager->journalOpen ){
    sqlite3OsClose(&pPager->jfd);
    pPager->journalOpen = 0;
    sqlite3Os.xDelete(pPager->zJournal);
    sqliteFree( pPager->aInJournal );
    pPager->aInJournal = 0;
    for(pPg=pPager->pAll; pPg; pPg=pPg->pNextAll){
      pPg->inJournal = 0;
      pPg->dirty = 0;
      pPg->needSync = 0;
#ifdef SQLITE_CHECK_PAGES
      pPg->pageHash = pager_pagehash(pPg);
#endif
    }
    pPager->dirtyCache = 0;
    pPager->nRec = 0;
  }else{
    assert( pPager->aInJournal==0 );
    assert( pPager->dirtyCache==0 || pPager->useJournal==0 );
  }
  rc = sqlite3OsUnlock(pPager->fd, SHARED_LOCK);
  pPager->state = PAGER_SHARED;
  pPager->origDbSize = 0;
  pPager->setMaster = 0;
  return rc;
}

/*
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
  ** statement journals.  Verify that this is always the case
  */
  assert( jfd == (useCksum ? pPager->jfd : pPager->stfd) );


  rc = read32bits(jfd, &pgno);
  if( rc!=SQLITE_OK ) return rc;
  rc = sqlite3Io.xRead(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.







|







979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
  ** statement journals.  Verify that this is always the case
  */
  assert( jfd == (useCksum ? pPager->jfd : pPager->stfd) );


  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.
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
  ** page content is in the main journal either because the page is not in
  ** cache or else it is marked as needSync==0.
  */
  pPg = pager_lookup(pPager, pgno);
  assert( pPager->state>=PAGER_EXCLUSIVE || pPg!=0 );
  TRACE3("PLAYBACK %d page %d\n", PAGERID(pPager), pgno);
  if( pPager->state>=PAGER_EXCLUSIVE && (pPg==0 || pPg->needSync==0) ){
    rc = sqlite3Io.xSeek(pPager->fd, (pgno-1)*(i64)pPager->pageSize);
    if( rc==SQLITE_OK ){
      rc = sqlite3Io.xWrite(pPager->fd, aData, pPager->pageSize);
    }
    if( pPg ) pPg->dirty = 0;
  }
  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
    ** database active. However such a page may be rolled back as a result







|

|







1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
  ** page content is in the main journal either because the page is not in
  ** cache or else it is marked as needSync==0.
  */
  pPg = pager_lookup(pPager, pgno);
  assert( pPager->state>=PAGER_EXCLUSIVE || pPg!=0 );
  TRACE3("PLAYBACK %d page %d\n", PAGERID(pPager), pgno);
  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 ) pPg->dirty = 0;
  }
  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
    ** database active. However such a page may be rolled back as a result
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
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
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
  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 = sqlite3Io.xOpenReadOnly(zMaster, &master);
  if( rc!=SQLITE_OK ) goto delmaster_out;
  master_open = 1;
  rc = sqlite3Io.xFileSize(master, &nMasterJournal);
  if( rc!=SQLITE_OK ) goto delmaster_out;

  if( nMasterJournal>0 ){
    char *zJournal;
    char *zMasterPtr = 0;

    /* Load the entire master journal file into space obtained from
    ** sqliteMalloc() and pointed to by zMasterJournal. 
    */
    zMasterJournal = (char *)sqliteMalloc(nMasterJournal);
    if( !zMasterJournal ){
      rc = SQLITE_NOMEM;
      goto delmaster_out;
    }
    rc = sqlite3Io.xRead(master, zMasterJournal, nMasterJournal);
    if( rc!=SQLITE_OK ) goto delmaster_out;

    zJournal = zMasterJournal;
    while( (zJournal-zMasterJournal)<nMasterJournal ){
      if( sqlite3Io.xFileExists(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 = sqlite3Io.xOpenReadOnly(zJournal, &journal);
        if( rc!=SQLITE_OK ){
          goto delmaster_out;
        }

        rc = readMasterJournal(journal, &zMasterPtr);
        sqlite3Io.xClose(&journal);
        if( rc!=SQLITE_OK ){
          goto delmaster_out;
        }

        c = zMasterPtr!=0 && strcmp(zMasterPtr, zMaster)==0;
        sqliteFree(zMasterPtr);
        if( c ){
          /* We have a match. Do not delete the master journal file. */
          goto delmaster_out;
        }
      }
      zJournal += (strlen(zJournal)+1);
    }
  }
  
  sqlite3Io.xDelete(zMaster);

delmaster_out:
  if( zMasterJournal ){
    sqliteFree(zMasterJournal);
  }  
  if( master_open ){
    sqlite3Io.xClose(&master);
  }
  return rc;
}

/*
** Make every page in the cache agree with what is on disk.  In other words,
** reread the disk to reset the state of the cache.
**
** This routine is called after a rollback in which some of the dirty cache
** pages had never been written out to disk.  We need to roll back the
** cache content and the easiest way to do that is to reread the old content
** back from the disk.
*/
static int pager_reload_cache(Pager *pPager){
  PgHdr *pPg;
  int rc = SQLITE_OK;
  for(pPg=pPager->pAll; pPg; pPg=pPg->pNextAll){
    char zBuf[SQLITE_MAX_PAGE_SIZE];
    if( !pPg->dirty ) continue;
    if( (int)pPg->pgno <= pPager->origDbSize ){
      rc = sqlite3Io.xSeek(pPager->fd, pPager->pageSize*(i64)(pPg->pgno-1));
      if( rc==SQLITE_OK ){
        rc = sqlite3Io.xRead(pPager->fd, zBuf, pPager->pageSize);
      }
      TRACE3("REFETCH %d page %d\n", PAGERID(pPager), pPg->pgno);
      if( rc ) break;
      CODEC(pPager, zBuf, pPg->pgno, 2);
    }else{
      memset(zBuf, 0, pPager->pageSize);
    }







|


|














|




|







|





|















|






|




















|

|







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
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
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
  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 = sqlite3Os.xOpenReadOnly(zMaster, &master);
  if( rc!=SQLITE_OK ) goto delmaster_out;
  master_open = 1;
  rc = sqlite3OsFileSize(master, &nMasterJournal);
  if( rc!=SQLITE_OK ) goto delmaster_out;

  if( nMasterJournal>0 ){
    char *zJournal;
    char *zMasterPtr = 0;

    /* Load the entire master journal file into space obtained from
    ** 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( sqlite3Os.xFileExists(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 = sqlite3Os.xOpenReadOnly(zJournal, &journal);
        if( rc!=SQLITE_OK ){
          goto delmaster_out;
        }

        rc = readMasterJournal(journal, &zMasterPtr);
        sqlite3OsClose(&journal);
        if( rc!=SQLITE_OK ){
          goto delmaster_out;
        }

        c = zMasterPtr!=0 && strcmp(zMasterPtr, zMaster)==0;
        sqliteFree(zMasterPtr);
        if( c ){
          /* We have a match. Do not delete the master journal file. */
          goto delmaster_out;
        }
      }
      zJournal += (strlen(zJournal)+1);
    }
  }
  
  sqlite3Os.xDelete(zMaster);

delmaster_out:
  if( zMasterJournal ){
    sqliteFree(zMasterJournal);
  }  
  if( master_open ){
    sqlite3OsClose(&master);
  }
  return rc;
}

/*
** Make every page in the cache agree with what is on disk.  In other words,
** reread the disk to reset the state of the cache.
**
** This routine is called after a rollback in which some of the dirty cache
** pages had never been written out to disk.  We need to roll back the
** cache content and the easiest way to do that is to reread the old content
** back from the disk.
*/
static int pager_reload_cache(Pager *pPager){
  PgHdr *pPg;
  int rc = SQLITE_OK;
  for(pPg=pPager->pAll; pPg; pPg=pPg->pNextAll){
    char zBuf[SQLITE_MAX_PAGE_SIZE];
    if( !pPg->dirty ) continue;
    if( (int)pPg->pgno <= pPager->origDbSize ){
      rc = sqlite3OsSeek(pPager->fd, pPager->pageSize*(i64)(pPg->pgno-1));
      if( rc==SQLITE_OK ){
        rc = sqlite3OsRead(pPager->fd, zBuf, pPager->pageSize);
      }
      TRACE3("REFETCH %d page %d\n", PAGERID(pPager), pPg->pgno);
      if( rc ) break;
      CODEC(pPager, zBuf, pPg->pgno, 2);
    }else{
      memset(zBuf, 0, pPager->pageSize);
    }
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206

/*
** Truncate the main file of the given pager to the number of pages
** indicated.
*/
static int pager_truncate(Pager *pPager, int nPage){
  assert( pPager->state>=PAGER_EXCLUSIVE );
  return sqlite3Io.xTruncate(pPager->fd, pPager->pageSize*(i64)nPage);
}

/*
** Playback the journal and thus restore the database file to
** the state it was in before we started making changes.  
**
** The journal file format is as follows: 







|







1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206

/*
** Truncate the main file of the given pager to the number of pages
** indicated.
*/
static int pager_truncate(Pager *pPager, int nPage){
  assert( pPager->state>=PAGER_EXCLUSIVE );
  return sqlite3OsTruncate(pPager->fd, pPager->pageSize*(i64)nPage);
}

/*
** Playback the journal and thus restore the database file to
** the state it was in before we started making changes.  
**
** The journal file format is as follows: 
1260
1261
1262
1263
1264
1265
1266
1267
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
  int rc;                  /* Result code of a subroutine */
  char *zMaster = 0;       /* Name of master journal file if any */

  /* Figure out how many records are in the journal.  Abort early if
  ** the journal is empty.
  */
  assert( pPager->journalOpen );
  rc = sqlite3Io.xFileSize(pPager->jfd, &szJ);
  if( rc!=SQLITE_OK ){
    goto end_playback;
  }

  /* Read the master journal name from the journal, if it is present.
  ** If a master journal file name is specified, but the file is not
  ** present on disk, then the journal is not hot and does not need to be
  ** played back.
  */
  rc = readMasterJournal(pPager->jfd, &zMaster);
  assert( rc!=SQLITE_DONE );
  if( rc!=SQLITE_OK || (zMaster && !sqlite3Io.xFileExists(zMaster)) ){
    sqliteFree(zMaster);
    zMaster = 0;
    if( rc==SQLITE_DONE ) rc = SQLITE_OK;
    goto end_playback;
  }
  sqlite3Io.xSeek(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







|











|





|







1260
1261
1262
1263
1264
1265
1266
1267
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
  int rc;                  /* Result code of a subroutine */
  char *zMaster = 0;       /* Name of master journal file if any */

  /* Figure out how many records are in the journal.  Abort early if
  ** the journal is empty.
  */
  assert( pPager->journalOpen );
  rc = sqlite3OsFileSize(pPager->jfd, &szJ);
  if( rc!=SQLITE_OK ){
    goto end_playback;
  }

  /* Read the master journal name from the journal, if it is present.
  ** If a master journal file name is specified, but the file is not
  ** present on disk, then the journal is not hot and does not need to be
  ** played back.
  */
  rc = readMasterJournal(pPager->jfd, &zMaster);
  assert( rc!=SQLITE_DONE );
  if( rc!=SQLITE_OK || (zMaster && !sqlite3Os.xFileExists(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
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
      rc = pager_truncate(pPager, mxPg);
      if( rc!=SQLITE_OK ){
        goto end_playback;
      }
      pPager->dbSize = mxPg;
    }

    /* rc = sqlite3Io.xSeek(pPager->jfd, JOURNAL_HDR_SZ(pPager)); */
    if( rc!=SQLITE_OK ) 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 ){







|







1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
      rc = pager_truncate(pPager, mxPg);
      if( rc!=SQLITE_OK ){
        goto end_playback;
      }
      pPager->dbSize = mxPg;
    }

    /* rc = sqlite3OsSeek(pPager->jfd, JOURNAL_HDR_SZ(pPager)); */
    if( rc!=SQLITE_OK ) 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 ){
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
  int i;                   /* Loop counter */
  int rc;

  szJ = pPager->journalOff;
#ifndef NDEBUG 
  {
    i64 os_szJ;
    rc = sqlite3Io.xFileSize(pPager->jfd, &os_szJ);
    if( rc!=SQLITE_OK ) return rc;
    assert( szJ==os_szJ );
  }
#endif

  /* Set hdrOff to be the offset to the first journal header written
  ** this statement transaction, or the end of the file if no journal







|







1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
  int i;                   /* Loop counter */
  int rc;

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

  /* Set hdrOff to be the offset to the first journal header written
  ** this statement transaction, or the end of the file if no journal
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
    rc = pager_truncate(pPager, pPager->stmtSize);
  }
  pPager->dbSize = pPager->stmtSize;

  /* Figure out how many records are in the statement journal.
  */
  assert( pPager->stmtInUse && pPager->journalOpen );
  sqlite3Io.xSeek(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.
  */







|







1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
    rc = pager_truncate(pPager, pPager->stmtSize);
  }
  pPager->dbSize = pPager->stmtSize;

  /* 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.
  */
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
  ** 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 = sqlite3Io.xSeek(pPager->jfd, pPager->stmtJSize);
  if( rc!=SQLITE_OK ){
    goto end_stmt_playback;
  }
  pPager->journalOff = pPager->stmtJSize;
  pPager->cksumInit = pPager->stmtCksum;
  assert( JOURNAL_HDR_SZ(pPager)<(pPager->pageSize+8) );
  while( pPager->journalOff <= (hdrOff-(pPager->pageSize+8)) ){







|







1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
  ** 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;
  assert( JOURNAL_HDR_SZ(pPager)<(pPager->pageSize+8) );
  while( pPager->journalOff <= (hdrOff-(pPager->pageSize+8)) ){
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
}

/*
** Adjust the robustness of the database to damage due to OS crashes
** or power failures by changing the number of syncs()s when writing
** the rollback journal.  There are three levels:
**
**    OFF       sqlite3Io.xSync() is never called.  This is the default
**              for temporary and transient files.
**
**    NORMAL    The journal is synced once before writes begin on the
**              database.  This is normally adequate protection, but
**              it is theoretically possible, though very unlikely,
**              that an inopertune power failure could leave the journal
**              in a state which would cause damage to the database







|







1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
}

/*
** Adjust the robustness of the database to damage due to OS crashes
** or power failures by changing the number of syncs()s when writing
** the rollback journal.  There are three levels:
**
**    OFF       sqlite3OsSync() is never called.  This is the default
**              for temporary and transient files.
**
**    NORMAL    The journal is synced once before writes begin on the
**              database.  This is normally adequate protection, but
**              it is theoretically possible, though very unlikely,
**              that an inopertune power failure could leave the journal
**              in a state which would cause damage to the database
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
*/
static int sqlite3pager_opentemp(char *zFile, OsFile **pFd){
  int cnt = 8;
  int rc;
  sqlite3_opentemp_count++;  /* Used for testing and analysis only */
  do{
    cnt--;
    sqlite3Io.xTempFileName(zFile);
    rc = sqlite3Io.xOpenExclusive(zFile, pFd, 1);
  }while( cnt>0 && rc!=SQLITE_OK && rc!=SQLITE_NOMEM );
  return rc;
}

/*
** Create a new page cache and put a pointer to the page cache in *ppPager.
** The file to be cached need not exist.  The file is not locked until







|
|







1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
*/
static int sqlite3pager_opentemp(char *zFile, OsFile **pFd){
  int cnt = 8;
  int rc;
  sqlite3_opentemp_count++;  /* Used for testing and analysis only */
  do{
    cnt--;
    sqlite3Os.xTempFileName(zFile);
    rc = sqlite3Os.xOpenExclusive(zFile, pFd, 1);
  }while( cnt>0 && rc!=SQLITE_OK && rc!=SQLITE_NOMEM );
  return rc;
}

/*
** Create a new page cache and put a pointer to the page cache in *ppPager.
** The file to be cached need not exist.  The file is not locked until
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
    if( strcmp(zFilename,":memory:")==0 ){
      memDb = 1;
      zFullPathname = sqliteStrDup("");
      rc = SQLITE_OK;
    }else
#endif
    {
      zFullPathname = sqlite3Io.xFullPathname(zFilename);
      if( zFullPathname ){
        rc = sqlite3Io.xOpenReadWrite(zFullPathname, &fd, &readOnly);
      }
    }
  }else{
    rc = sqlite3pager_opentemp(zTemp, &fd);
    zFilename = zTemp;
    zFullPathname = sqlite3Io.xFullPathname(zFilename);
    if( rc==SQLITE_OK ){
      tempFile = 1;
    }
  }
  if( !zFullPathname ){
    sqlite3Io.xClose(&fd);
    return SQLITE_NOMEM;
  }
  if( rc!=SQLITE_OK ){
    sqlite3Io.xClose(&fd);
    sqliteFree(zFullPathname);
    return rc;
  }
  nameLen = strlen(zFullPathname);
  pPager = sqliteMalloc( sizeof(*pPager) + nameLen*3 + 30 );
  if( pPager==0 ){
    sqlite3Io.xClose(&fd);
    sqliteFree(zFullPathname);
    return SQLITE_NOMEM;
  }
  TRACE3("OPEN %d %s\n", FILEHANDLEID(fd), zFullPathname);
  pPager->zFilename = (char*)&pPager[1];
  pPager->zDirectory = &pPager->zFilename[nameLen+1];
  pPager->zJournal = &pPager->zDirectory[nameLen+1];







|

|





|





|



|






|







1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
    if( strcmp(zFilename,":memory:")==0 ){
      memDb = 1;
      zFullPathname = sqliteStrDup("");
      rc = SQLITE_OK;
    }else
#endif
    {
      zFullPathname = sqlite3Os.xFullPathname(zFilename);
      if( zFullPathname ){
        rc = sqlite3Os.xOpenReadWrite(zFullPathname, &fd, &readOnly);
      }
    }
  }else{
    rc = sqlite3pager_opentemp(zTemp, &fd);
    zFilename = zTemp;
    zFullPathname = sqlite3Os.xFullPathname(zFilename);
    if( rc==SQLITE_OK ){
      tempFile = 1;
    }
  }
  if( !zFullPathname ){
    sqlite3OsClose(&fd);
    return SQLITE_NOMEM;
  }
  if( rc!=SQLITE_OK ){
    sqlite3OsClose(&fd);
    sqliteFree(zFullPathname);
    return rc;
  }
  nameLen = strlen(zFullPathname);
  pPager = sqliteMalloc( sizeof(*pPager) + nameLen*3 + 30 );
  if( pPager==0 ){
    sqlite3OsClose(&fd);
    sqliteFree(zFullPathname);
    return SQLITE_NOMEM;
  }
  TRACE3("OPEN %d %s\n", FILEHANDLEID(fd), zFullPathname);
  pPager->zFilename = (char*)&pPager[1];
  pPager->zDirectory = &pPager->zFilename[nameLen+1];
  pPager->zJournal = &pPager->zDirectory[nameLen+1];
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
/*
** Read the first N bytes from the beginning of the file into memory
** that pDest points to.  No error checking is done.
*/
void sqlite3pager_read_fileheader(Pager *pPager, int N, unsigned char *pDest){
  memset(pDest, 0, N);
  if( MEMDB==0 ){
    sqlite3Io.xSeek(pPager->fd, 0);
    sqlite3Io.xRead(pPager->fd, pDest, N);
    clear_simulated_io_error();
  }
}

/*
** Return the total number of pages in the disk file associated with
** pPager. 
**
** If the PENDING_BYTE lies on the page directly after the end of the
** file, then consider this page part of the file too. For example, if
** PENDING_BYTE is byte 4096 (the first byte of page 5) and the size of the
** file is 4096 bytes, 5 is returned instead of 4.
*/
int sqlite3pager_pagecount(Pager *pPager){
  i64 n;
  assert( pPager!=0 );
  if( pPager->dbSize>=0 ){
    n = pPager->dbSize;
  } else {
    if( sqlite3Io.xFileSize(pPager->fd, &n)!=SQLITE_OK ){
      pPager->errMask |= PAGER_ERR_DISK;
      return 0;
    }
    if( n>0 && n<pPager->pageSize ){
      n = 1;
    }else{
      n /= pPager->pageSize;







|
|



















|







1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
/*
** Read the first N bytes from the beginning of the file into memory
** that pDest points to.  No error checking is done.
*/
void sqlite3pager_read_fileheader(Pager *pPager, int N, unsigned char *pDest){
  memset(pDest, 0, N);
  if( MEMDB==0 ){
    sqlite3OsSeek(pPager->fd, 0);
    sqlite3OsRead(pPager->fd, pDest, N);
    clear_simulated_io_error();
  }
}

/*
** Return the total number of pages in the disk file associated with
** pPager. 
**
** If the PENDING_BYTE lies on the page directly after the end of the
** file, then consider this page part of the file too. For example, if
** PENDING_BYTE is byte 4096 (the first byte of page 5) and the size of the
** file is 4096 bytes, 5 is returned instead of 4.
*/
int sqlite3pager_pagecount(Pager *pPager){
  i64 n;
  assert( pPager!=0 );
  if( pPager->dbSize>=0 ){
    n = pPager->dbSize;
  } else {
    if( sqlite3OsFileSize(pPager->fd, &n)!=SQLITE_OK ){
      pPager->errMask |= PAGER_ERR_DISK;
      return 0;
    }
    if( n>0 && n<pPager->pageSize ){
      n = 1;
    }else{
      n /= pPager->pageSize;
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
  assert( PAGER_SHARED==SHARED_LOCK );
  assert( PAGER_RESERVED==RESERVED_LOCK );
  assert( PAGER_EXCLUSIVE==EXCLUSIVE_LOCK );
  if( pPager->state>=locktype ){
    rc = SQLITE_OK;
  }else{
    do {
      rc = sqlite3Io.xLock(pPager->fd, locktype);
    }while( rc==SQLITE_BUSY && sqlite3InvokeBusyHandler(pPager->pBusyHandler) );
    if( rc==SQLITE_OK ){
      pPager->state = locktype;
    }
  }
  return rc;
}







|







1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
  assert( PAGER_SHARED==SHARED_LOCK );
  assert( PAGER_RESERVED==RESERVED_LOCK );
  assert( PAGER_EXCLUSIVE==EXCLUSIVE_LOCK );
  if( pPager->state>=locktype ){
    rc = SQLITE_OK;
  }else{
    do {
      rc = sqlite3OsLock(pPager->fd, locktype);
    }while( rc==SQLITE_BUSY && sqlite3InvokeBusyHandler(pPager->pBusyHandler) );
    if( rc==SQLITE_OK ){
      pPager->state = locktype;
    }
  }
  return rc;
}
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
      ** operation. So disable IO error simulation so that testing
      ** works more easily.
      */
      disable_simulated_io_errors();
      sqlite3pager_rollback(pPager);
      enable_simulated_io_errors();
      if( !MEMDB ){
        sqlite3Io.xUnlock(pPager->fd, NO_LOCK);
      }
      assert( pPager->errMask || pPager->journalOpen==0 );
      break;
    }
    case PAGER_SHARED: {
      if( !MEMDB ){
        sqlite3Io.xUnlock(pPager->fd, NO_LOCK);
      }
      break;
    }
    default: {
      /* Do nothing */
      break;
    }







|






|







1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
      ** operation. So disable IO error simulation so that testing
      ** works more easily.
      */
      disable_simulated_io_errors();
      sqlite3pager_rollback(pPager);
      enable_simulated_io_errors();
      if( !MEMDB ){
        sqlite3OsUnlock(pPager->fd, NO_LOCK);
      }
      assert( pPager->errMask || pPager->journalOpen==0 );
      break;
    }
    case PAGER_SHARED: {
      if( !MEMDB ){
        sqlite3OsUnlock(pPager->fd, NO_LOCK);
      }
      break;
    }
    default: {
      /* Do nothing */
      break;
    }
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
#endif
    pNext = pPg->pNextAll;
    sqliteFree(pPg);
  }
  TRACE2("CLOSE %d\n", PAGERID(pPager));
  assert( pPager->errMask || (pPager->journalOpen==0 && pPager->stmtOpen==0) );
  if( pPager->journalOpen ){
    sqlite3Io.xClose(&pPager->jfd);
  }
  sqliteFree(pPager->aInJournal);
  if( pPager->stmtOpen ){
    sqlite3Io.xClose(&pPager->stfd);
  }
  sqlite3Io.xClose(&pPager->fd);
  /* Temp files are automatically deleted by the OS
  ** if( pPager->tempFile ){
  **   sqlite3Io.xDelete(pPager->zFilename);
  ** }
  */

  sqliteFree(pPager);
  return SQLITE_OK;
}








|



|

|


|







2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
#endif
    pNext = pPg->pNextAll;
    sqliteFree(pPg);
  }
  TRACE2("CLOSE %d\n", PAGERID(pPager));
  assert( pPager->errMask || (pPager->journalOpen==0 && pPager->stmtOpen==0) );
  if( pPager->journalOpen ){
    sqlite3OsClose(&pPager->jfd);
  }
  sqliteFree(pPager->aInJournal);
  if( pPager->stmtOpen ){
    sqlite3OsClose(&pPager->stfd);
  }
  sqlite3OsClose(&pPager->fd);
  /* Temp files are automatically deleted by the OS
  ** if( pPager->tempFile ){
  **   sqlite3Os.xDelete(pPager->zFilename);
  ** }
  */

  sqliteFree(pPager);
  return SQLITE_OK;
}

2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
      ** was turned off after the transaction was started.  Ticket #615 */
#ifndef NDEBUG
      {
        /* Make sure the pPager->nRec counter we are keeping agrees
        ** with the nRec computed from the size of the journal file.
        */
        i64 jSz;
        rc = sqlite3Io.xFileSize(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 ){
          TRACE2("SYNC journal of %d\n", PAGERID(pPager));
          rc = sqlite3Io.xSync(pPager->jfd, 0);
          if( rc!=0 ) return rc;
        }
        rc = sqlite3Io.xSeek(pPager->jfd,
                           pPager->journalHdr + sizeof(aJournalMagic));
        if( rc ) return rc;
        rc = write32bits(pPager->jfd, pPager->nRec);
        if( rc ) return rc;

        rc = sqlite3Io.xSeek(pPager->jfd, pPager->journalOff);
        if( rc ) return rc;
      }
      TRACE2("SYNC journal of %d\n", PAGERID(pPager));
      rc = sqlite3Io.xSync(pPager->jfd, pPager->fullSync);
      if( rc!=0 ) return rc;
      pPager->journalStarted = 1;
    }
    pPager->needSync = 0;

    /* Erase the needSync flag from every page.
    */







|












|


|





|



|







2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
      ** was turned off after the transaction was started.  Ticket #615 */
#ifndef NDEBUG
      {
        /* Make sure the pPager->nRec counter we are keeping agrees
        ** with the nRec computed from the size of the journal file.
        */
        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 ){
          TRACE2("SYNC journal of %d\n", PAGERID(pPager));
          rc = sqlite3OsSync(pPager->jfd, 0);
          if( rc!=0 ) return rc;
        }
        rc = sqlite3OsSeek(pPager->jfd,
                           pPager->journalHdr + sizeof(aJournalMagic));
        if( rc ) return rc;
        rc = write32bits(pPager->jfd, pPager->nRec);
        if( rc ) return rc;

        rc = sqlite3OsSeek(pPager->jfd, pPager->journalOff);
        if( rc ) return rc;
      }
      TRACE2("SYNC journal of %d\n", PAGERID(pPager));
      rc = sqlite3OsSync(pPager->jfd, pPager->fullSync);
      if( rc!=0 ) return rc;
      pPager->journalStarted = 1;
    }
    pPager->needSync = 0;

    /* Erase the needSync flag from every page.
    */
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
  int rc;

  if( pList==0 ) return SQLITE_OK;
  pPager = pList->pPager;

  /* At this point there may be either a RESERVED or EXCLUSIVE lock on the
  ** database file. If there is already an EXCLUSIVE lock, the following
  ** calls to sqlite3Io.xLock() are no-ops.
  **
  ** Moving the lock from RESERVED to EXCLUSIVE actually involves going
  ** through an intermediate state PENDING.   A PENDING lock prevents new
  ** readers from attaching to the database but is unsufficient for us to
  ** write.  The idea of a PENDING lock is to prevent new readers from
  ** coming in while we wait for existing readers to clear.
  **
  ** While the pager is in the RESERVED state, the original database file
  ** is unchanged and we can rollback without having to playback the
  ** journal into the original database file.  Once we transition to
  ** EXCLUSIVE, it means the database file has been changed and any rollback
  ** will require a journal playback.
  */
  rc = pager_wait_on_lock(pPager, EXCLUSIVE_LOCK);
  if( rc!=SQLITE_OK ){
    return rc;
  }

  while( pList ){
    assert( pList->dirty );
    rc = sqlite3Io.xSeek(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 sqlite3pager_truncate() 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 ){
      CODEC(pPager, PGHDR_TO_DATA(pList), pList->pgno, 6);
      TRACE3("STORE %d page %d\n", PAGERID(pPager), pList->pgno);
      rc = sqlite3Io.xWrite(pPager->fd, PGHDR_TO_DATA(pList),
                             pPager->pageSize);
      CODEC(pPager, PGHDR_TO_DATA(pList), pList->pgno, 0);
      TEST_INCR(pPager->nWrite);
    }
#ifndef NDEBUG
    else{
      TRACE3("NOSTORE %d page %d\n", PAGERID(pPager), pList->pgno);







|




















|









|







2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
  int rc;

  if( pList==0 ) return SQLITE_OK;
  pPager = pList->pPager;

  /* At this point there may be either a RESERVED or EXCLUSIVE lock on the
  ** database file. If there is already an EXCLUSIVE lock, the following
  ** calls to sqlite3OsLock() are no-ops.
  **
  ** Moving the lock from RESERVED to EXCLUSIVE actually involves going
  ** through an intermediate state PENDING.   A PENDING lock prevents new
  ** readers from attaching to the database but is unsufficient for us to
  ** write.  The idea of a PENDING lock is to prevent new readers from
  ** coming in while we wait for existing readers to clear.
  **
  ** While the pager is in the RESERVED state, the original database file
  ** is unchanged and we can rollback without having to playback the
  ** journal into the original database file.  Once we transition to
  ** EXCLUSIVE, it means the database file has been changed and any rollback
  ** will require a journal playback.
  */
  rc = pager_wait_on_lock(pPager, EXCLUSIVE_LOCK);
  if( rc!=SQLITE_OK ){
    return rc;
  }

  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 sqlite3pager_truncate() 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 ){
      CODEC(pPager, PGHDR_TO_DATA(pList), pList->pgno, 6);
      TRACE3("STORE %d page %d\n", PAGERID(pPager), pList->pgno);
      rc = sqlite3OsWrite(pPager->fd, PGHDR_TO_DATA(pList),
                             pPager->pageSize);
      CODEC(pPager, PGHDR_TO_DATA(pList), pList->pgno, 0);
      TEST_INCR(pPager->nWrite);
    }
#ifndef NDEBUG
    else{
      TRACE3("NOSTORE %d page %d\n", PAGERID(pPager), pList->pgno);
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
**
** If the current size of the database file is 0 but a journal file
** exists, that is probably an old journal left over from a prior
** database with the same name.  Just delete the journal.
*/
static int hasHotJournal(Pager *pPager){
  if( !pPager->useJournal ) return 0;
  if( !sqlite3Io.xFileExists(pPager->zJournal) ) return 0;
  if( sqlite3Io.xCheckReservedLock(pPager->fd) ) return 0;
  if( sqlite3pager_pagecount(pPager)==0 ){
    sqlite3Io.xDelete(pPager->zJournal);
    return 0;
  }else{
    return 1;
  }
}

/*







|
|

|







2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
**
** If the current size of the database file is 0 but a journal file
** exists, that is probably an old journal left over from a prior
** database with the same name.  Just delete the journal.
*/
static int hasHotJournal(Pager *pPager){
  if( !pPager->useJournal ) return 0;
  if( !sqlite3Os.xFileExists(pPager->zJournal) ) return 0;
  if( sqlite3OsCheckReservedLock(pPager->fd) ) return 0;
  if( sqlite3pager_pagecount(pPager)==0 ){
    sqlite3Os.xDelete(pPager->zJournal);
    return 0;
  }else{
    return 1;
  }
}

/*
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
2382
2383
2384
2385
2386
2387
       ** database is safe to read while this process is still rolling it 
       ** back.
       ** 
       ** Because the intermediate RESERVED lock is not requested, the
       ** second process will get to this point in the code and fail to
       ** obtain it's own EXCLUSIVE lock on the database file.
       */
       rc = sqlite3Io.xLock(pPager->fd, EXCLUSIVE_LOCK);
       if( rc!=SQLITE_OK ){
         sqlite3Io.xUnlock(pPager->fd, NO_LOCK);
         pPager->state = PAGER_UNLOCK;
         return rc;
       }
       pPager->state = PAGER_EXCLUSIVE;

       /* Open the journal for reading only.  Return SQLITE_BUSY if
       ** we are unable to open the journal file. 
       **
       ** The journal file does not need to be locked itself.  The
       ** journal file is never open unless the main database file holds
       ** a write lock, so there is never any chance of two or more
       ** processes opening the journal at the same time.
       */
       rc = sqlite3Io.xOpenReadOnly(pPager->zJournal, &pPager->jfd);
       if( rc!=SQLITE_OK ){
         sqlite3Io.xUnlock(pPager->fd, NO_LOCK);
         pPager->state = PAGER_UNLOCK;
         return SQLITE_BUSY;
       }
       pPager->journalOpen = 1;
       pPager->journalStarted = 0;
       pPager->journalOff = 0;
       pPager->setMaster = 0;







|

|













|

|







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
2382
2383
2384
2385
2386
2387
       ** database is safe to read while this process is still rolling it 
       ** back.
       ** 
       ** Because the intermediate RESERVED lock is not requested, the
       ** second process will get to this point in the code and fail to
       ** obtain it's own EXCLUSIVE lock on the database file.
       */
       rc = sqlite3OsLock(pPager->fd, EXCLUSIVE_LOCK);
       if( rc!=SQLITE_OK ){
         sqlite3OsUnlock(pPager->fd, NO_LOCK);
         pPager->state = PAGER_UNLOCK;
         return rc;
       }
       pPager->state = PAGER_EXCLUSIVE;

       /* Open the journal for reading only.  Return SQLITE_BUSY if
       ** we are unable to open the journal file. 
       **
       ** The journal file does not need to be locked itself.  The
       ** journal file is never open unless the main database file holds
       ** a write lock, so there is never any chance of two or more
       ** processes opening the journal at the same time.
       */
       rc = sqlite3Os.xOpenReadOnly(pPager->zJournal, &pPager->jfd);
       if( rc!=SQLITE_OK ){
         sqlite3OsUnlock(pPager->fd, NO_LOCK);
         pPager->state = PAGER_UNLOCK;
         return SQLITE_BUSY;
       }
       pPager->journalOpen = 1;
       pPager->journalStarted = 0;
       pPager->journalOff = 0;
       pPager->setMaster = 0;
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
      return rc;
    }
    if( sqlite3pager_pagecount(pPager)<(int)pgno ){
      memset(PGHDR_TO_DATA(pPg), 0, pPager->pageSize);
    }else{
      int rc;
      assert( MEMDB==0 );
      rc = sqlite3Io.xSeek(pPager->fd, (pgno-1)*(i64)pPager->pageSize);
      if( rc==SQLITE_OK ){
        rc = sqlite3Io.xRead(pPager->fd, PGHDR_TO_DATA(pPg),
                              pPager->pageSize);
      }
      TRACE3("FETCH %d page %d\n", PAGERID(pPager), pPg->pgno);
      CODEC(pPager, PGHDR_TO_DATA(pPg), pPg->pgno, 3);
      if( rc!=SQLITE_OK ){
        i64 fileSize;
        if( sqlite3Io.xFileSize(pPager->fd,&fileSize)!=SQLITE_OK
               || fileSize>=pgno*pPager->pageSize ){
          sqlite3pager_unref(PGHDR_TO_DATA(pPg));
          return rc;
        }else{
          clear_simulated_io_error();
          memset(PGHDR_TO_DATA(pPg), 0, pPager->pageSize);
        }







|

|






|







2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
      return rc;
    }
    if( sqlite3pager_pagecount(pPager)<(int)pgno ){
      memset(PGHDR_TO_DATA(pPg), 0, pPager->pageSize);
    }else{
      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);
      }
      TRACE3("FETCH %d page %d\n", PAGERID(pPager), pPg->pgno);
      CODEC(pPager, PGHDR_TO_DATA(pPg), pPg->pgno, 3);
      if( rc!=SQLITE_OK ){
        i64 fileSize;
        if( sqlite3OsFileSize(pPager->fd,&fileSize)!=SQLITE_OK
               || fileSize>=pgno*pPager->pageSize ){
          sqlite3pager_unref(PGHDR_TO_DATA(pPg));
          return rc;
        }else{
          clear_simulated_io_error();
          memset(PGHDR_TO_DATA(pPg), 0, pPager->pageSize);
        }
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
  assert( pPager->aInJournal==0 );
  sqlite3pager_pagecount(pPager);
  pPager->aInJournal = sqliteMalloc( pPager->dbSize/8 + 1 );
  if( pPager->aInJournal==0 ){
    rc = SQLITE_NOMEM;
    goto failed_to_open_journal;
  }
  rc = sqlite3Io.xOpenExclusive(pPager->zJournal, &pPager->jfd,
                                 pPager->tempFile);
  pPager->journalOff = 0;
  pPager->setMaster = 0;
  pPager->journalHdr = 0;
  if( rc!=SQLITE_OK ){
    goto failed_to_open_journal;
  }
  sqlite3Io.xSetFullSync(pPager->jfd, pPager->fullSync);
  sqlite3Io.xSetFullSync(pPager->fd, pPager->fullSync);
  sqlite3Io.xOpenDirectory(pPager->zDirectory, pPager->jfd);
  pPager->journalOpen = 1;
  pPager->journalStarted = 0;
  pPager->needSync = 0;
  pPager->alwaysRollback = 0;
  pPager->nRec = 0;
  if( pPager->errMask!=0 ){
    rc = pager_errcode(pPager);







|







|
|
|







2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
  assert( pPager->aInJournal==0 );
  sqlite3pager_pagecount(pPager);
  pPager->aInJournal = sqliteMalloc( pPager->dbSize/8 + 1 );
  if( pPager->aInJournal==0 ){
    rc = SQLITE_NOMEM;
    goto failed_to_open_journal;
  }
  rc = sqlite3Os.xOpenExclusive(pPager->zJournal, &pPager->jfd,
                                 pPager->tempFile);
  pPager->journalOff = 0;
  pPager->setMaster = 0;
  pPager->journalHdr = 0;
  if( rc!=SQLITE_OK ){
    goto failed_to_open_journal;
  }
  sqlite3OsSetFullSync(pPager->jfd, pPager->fullSync);
  sqlite3OsSetFullSync(pPager->fd, pPager->fullSync);
  sqlite3OsOpenDirectory(pPager->jfd, pPager->zDirectory);
  pPager->journalOpen = 1;
  pPager->journalStarted = 0;
  pPager->needSync = 0;
  pPager->alwaysRollback = 0;
  pPager->nRec = 0;
  if( pPager->errMask!=0 ){
    rc = pager_errcode(pPager);
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
    }
  }
  return rc;

failed_to_open_journal:
  sqliteFree(pPager->aInJournal);
  pPager->aInJournal = 0;
  sqlite3Io.xUnlock(pPager->fd, NO_LOCK);
  pPager->state = PAGER_UNLOCK;
  return rc;
}

/*
** Acquire a write-lock on the database.  The lock is removed when
** the any of the following happen:







|







2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
    }
  }
  return rc;

failed_to_open_journal:
  sqliteFree(pPager->aInJournal);
  pPager->aInJournal = 0;
  sqlite3OsUnlock(pPager->fd, NO_LOCK);
  pPager->state = PAGER_UNLOCK;
  return rc;
}

/*
** Acquire a write-lock on the database.  The lock is removed when
** the any of the following happen:
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
  assert( pPager->state!=PAGER_UNLOCK );
  if( pPager->state==PAGER_SHARED ){
    assert( pPager->aInJournal==0 );
    if( MEMDB ){
      pPager->state = PAGER_EXCLUSIVE;
      pPager->origDbSize = pPager->dbSize;
    }else{
      rc = sqlite3Io.xLock(pPager->fd, RESERVED_LOCK);
      if( rc==SQLITE_OK ){
        pPager->state = PAGER_RESERVED;
        if( exFlag ){
          rc = pager_wait_on_lock(pPager, EXCLUSIVE_LOCK);
        }
      }
      if( rc!=SQLITE_OK ){







|







2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
  assert( pPager->state!=PAGER_UNLOCK );
  if( pPager->state==PAGER_SHARED ){
    assert( pPager->aInJournal==0 );
    if( MEMDB ){
      pPager->state = PAGER_EXCLUSIVE;
      pPager->origDbSize = pPager->dbSize;
    }else{
      rc = sqlite3OsLock(pPager->fd, RESERVED_LOCK);
      if( rc==SQLITE_OK ){
        pPager->state = PAGER_RESERVED;
        if( exFlag ){
          rc = pager_wait_on_lock(pPager, EXCLUSIVE_LOCK);
        }
      }
      if( rc!=SQLITE_OK ){
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
          assert( pPg->pgno!=PAGER_MJ_PGNO(pPager) );
          CODEC(pPager, pData, pPg->pgno, 7);
          cksum = pager_cksum(pPager, pPg->pgno, pData);
          saved = *(u32*)PGHDR_TO_EXTRA(pPg, pPager);
          store32bits(cksum, pPg, pPager->pageSize);
          szPg = pPager->pageSize+8;
          store32bits(pPg->pgno, pPg, -4);
          rc = sqlite3Io.xWrite(pPager->jfd, &((char*)pData)[-4], szPg);
          pPager->journalOff += szPg;
          TRACE4("JOURNAL %d page %d needSync=%d\n",
                  PAGERID(pPager), pPg->pgno, pPg->needSync);
          CODEC(pPager, pData, pPg->pgno, 0);
          *(u32*)PGHDR_TO_EXTRA(pPg, pPager) = saved;
          if( rc!=SQLITE_OK ){
            sqlite3pager_rollback(pPager);







|







2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
          assert( pPg->pgno!=PAGER_MJ_PGNO(pPager) );
          CODEC(pPager, pData, pPg->pgno, 7);
          cksum = pager_cksum(pPager, pPg->pgno, pData);
          saved = *(u32*)PGHDR_TO_EXTRA(pPg, pPager);
          store32bits(cksum, pPg, pPager->pageSize);
          szPg = pPager->pageSize+8;
          store32bits(pPg->pgno, pPg, -4);
          rc = sqlite3OsWrite(pPager->jfd, &((char*)pData)[-4], szPg);
          pPager->journalOff += szPg;
          TRACE4("JOURNAL %d page %d needSync=%d\n",
                  PAGERID(pPager), pPg->pgno, pPg->needSync);
          CODEC(pPager, pData, pPg->pgno, 0);
          *(u32*)PGHDR_TO_EXTRA(pPg, pPager) = saved;
          if( rc!=SQLITE_OK ){
            sqlite3pager_rollback(pPager);
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
        if( pHist->pStmt ){
          memcpy(pHist->pStmt, PGHDR_TO_DATA(pPg), pPager->pageSize);
        }
        TRACE3("STMT-JOURNAL %d page %d\n", PAGERID(pPager), pPg->pgno);
      }else{
        store32bits(pPg->pgno, pPg, -4);
        CODEC(pPager, pData, pPg->pgno, 7);
        rc = sqlite3Io.xWrite(pPager->stfd,((char*)pData)-4,
                               pPager->pageSize+4);
        TRACE3("STMT-JOURNAL %d page %d\n", PAGERID(pPager), pPg->pgno);
        CODEC(pPager, pData, pPg->pgno, 0);
        if( rc!=SQLITE_OK ){
          sqlite3pager_rollback(pPager);
          pPager->errMask |= PAGER_ERR_FULL;
          return rc;







|







2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
        if( pHist->pStmt ){
          memcpy(pHist->pStmt, PGHDR_TO_DATA(pPg), pPager->pageSize);
        }
        TRACE3("STMT-JOURNAL %d page %d\n", PAGERID(pPager), pPg->pgno);
      }else{
        store32bits(pPg->pgno, pPg, -4);
        CODEC(pPager, pData, pPg->pgno, 7);
        rc = sqlite3OsWrite(pPager->stfd,((char*)pData)-4,
                               pPager->pageSize+4);
        TRACE3("STMT-JOURNAL %d page %d\n", PAGERID(pPager), pPg->pgno);
        CODEC(pPager, pData, pPg->pgno, 0);
        if( rc!=SQLITE_OK ){
          sqlite3pager_rollback(pPager);
          pPager->errMask |= PAGER_ERR_FULL;
          return rc;
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
    }
#endif
    pPager->pStmt = 0;
    pPager->state = PAGER_SHARED;
    return SQLITE_OK;
  }
  if( pPager->dirtyCache==0 ){
    /* Exit early (without doing the time-consuming sqlite3Io.xSync() calls)
    ** if there have been no changes to the database file. */
    assert( pPager->needSync==0 );
    rc = pager_unwritelock(pPager);
    pPager->dbSize = -1;
    return rc;
  }
  assert( pPager->journalOpen );







|







3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
    }
#endif
    pPager->pStmt = 0;
    pPager->state = PAGER_SHARED;
    return SQLITE_OK;
  }
  if( pPager->dirtyCache==0 ){
    /* Exit early (without doing the time-consuming sqlite3OsSync() calls)
    ** if there have been no changes to the database file. */
    assert( pPager->needSync==0 );
    rc = pager_unwritelock(pPager);
    pPager->dbSize = -1;
    return rc;
  }
  assert( pPager->journalOpen );
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
  if( !pPager->journalOpen ){
    pPager->stmtAutoopen = 1;
    return SQLITE_OK;
  }
  assert( pPager->journalOpen );
  pPager->aInStmt = sqliteMalloc( pPager->dbSize/8 + 1 );
  if( pPager->aInStmt==0 ){
    sqlite3Io.xLock(pPager->fd, SHARED_LOCK);
    return SQLITE_NOMEM;
  }
#ifndef NDEBUG
  rc = sqlite3Io.xFileSize(pPager->jfd, &pPager->stmtJSize);
  if( rc ) goto stmt_begin_failed;
  assert( pPager->stmtJSize == pPager->journalOff );
#endif
  pPager->stmtJSize = pPager->journalOff;
  pPager->stmtSize = pPager->dbSize;
  pPager->stmtHdrOff = 0;
  pPager->stmtCksum = pPager->cksumInit;







|



|







3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
  if( !pPager->journalOpen ){
    pPager->stmtAutoopen = 1;
    return SQLITE_OK;
  }
  assert( pPager->journalOpen );
  pPager->aInStmt = sqliteMalloc( pPager->dbSize/8 + 1 );
  if( pPager->aInStmt==0 ){
    sqlite3OsLock(pPager->fd, SHARED_LOCK);
    return SQLITE_NOMEM;
  }
#ifndef NDEBUG
  rc = sqlite3OsFileSize(pPager->jfd, &pPager->stmtJSize);
  if( rc ) goto stmt_begin_failed;
  assert( pPager->stmtJSize == pPager->journalOff );
#endif
  pPager->stmtJSize = pPager->journalOff;
  pPager->stmtSize = pPager->dbSize;
  pPager->stmtHdrOff = 0;
  pPager->stmtCksum = pPager->cksumInit;
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
** Commit a statement.
*/
int sqlite3pager_stmt_commit(Pager *pPager){
  if( pPager->stmtInUse ){
    PgHdr *pPg, *pNext;
    TRACE2("STMT-COMMIT %d\n", PAGERID(pPager));
    if( !MEMDB ){
      sqlite3Io.xSeek(pPager->stfd, 0);
      /* sqlite3Io.xTruncate(pPager->stfd, 0); */
      sqliteFree( pPager->aInStmt );
      pPager->aInStmt = 0;
    }
    for(pPg=pPager->pStmt; pPg; pPg=pNext){
      pNext = pPg->pNextStmt;
      assert( pPg->inStmt );
      pPg->inStmt = 0;







|
|







3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
** Commit a statement.
*/
int sqlite3pager_stmt_commit(Pager *pPager){
  if( pPager->stmtInUse ){
    PgHdr *pPg, *pNext;
    TRACE2("STMT-COMMIT %d\n", PAGERID(pPager));
    if( !MEMDB ){
      sqlite3OsSeek(pPager->stfd, 0);
      /* sqlite3OsTruncate(pPager->stfd, 0); */
      sqliteFree( pPager->aInStmt );
      pPager->aInStmt = 0;
    }
    for(pPg=pPager->pStmt; pPg; pPg=pNext){
      pNext = pPg->pNextStmt;
      assert( pPg->inStmt );
      pPg->inStmt = 0;
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
    /* Write all dirty pages to the database file */
    pPg = pager_get_all_dirty_pages(pPager);
    rc = pager_write_pagelist(pPg);
    if( rc!=SQLITE_OK ) goto sync_exit;

    /* Sync the database file. */
    if( !pPager->noSync ){
      rc = sqlite3Io.xSync(pPager->fd, 0);
    }

    pPager->state = PAGER_SYNCED;
  }

sync_exit:
  return rc;







|







3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
    /* Write all dirty pages to the database file */
    pPg = pager_get_all_dirty_pages(pPager);
    rc = pager_write_pagelist(pPg);
    if( rc!=SQLITE_OK ) goto sync_exit;

    /* Sync the database file. */
    if( !pPager->noSync ){
      rc = sqlite3OsSync(pPager->fd, 0);
    }

    pPager->state = PAGER_SYNCED;
  }

sync_exit:
  return rc;
3621
3622
3623
3624
3625
3626
3627
3628
3629
3630
3631
3632
3633
3634
3635
#if defined(SQLITE_DEBUG) || defined(SQLITE_TEST)
/*
** Return the current state of the file lock for the given pager.
** The return value is one of NO_LOCK, SHARED_LOCK, RESERVED_LOCK,
** PENDING_LOCK, or EXCLUSIVE_LOCK.
*/
int sqlite3pager_lockstate(Pager *pPager){
  return sqlite3Io.xLockState(pPager->fd);
}
#endif

#ifdef SQLITE_DEBUG
/*
** Print a listing of all referenced pages and their ref count.
*/







|







3621
3622
3623
3624
3625
3626
3627
3628
3629
3630
3631
3632
3633
3634
3635
#if defined(SQLITE_DEBUG) || defined(SQLITE_TEST)
/*
** Return the current state of the file lock for the given pager.
** The return value is one of NO_LOCK, SHARED_LOCK, RESERVED_LOCK,
** PENDING_LOCK, or EXCLUSIVE_LOCK.
*/
int sqlite3pager_lockstate(Pager *pPager){
  return sqlite3OsLockState(pPager->fd);
}
#endif

#ifdef SQLITE_DEBUG
/*
** Print a listing of all referenced pages and their ref count.
*/
Changes to src/pragma.c.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
/*
** 2003 April 6
**
** The author disclaims copyright to this source code.  In place of
** a legal notice, here is a blessing:
**
**    May you do good and not evil.
**    May you find forgiveness for yourself and forgive others.
**    May you share freely, never taking more than you give.
**
*************************************************************************
** This file contains code used to implement the PRAGMA command.
**
** $Id: pragma.c,v 1.104 2005/11/26 00:25:03 drh Exp $
*/
#include "sqliteInt.h"
#include "os.h"
#include <ctype.h>

/* Ignore this whole file if pragmas are disabled
*/













|







1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
/*
** 2003 April 6
**
** The author disclaims copyright to this source code.  In place of
** a legal notice, here is a blessing:
**
**    May you do good and not evil.
**    May you find forgiveness for yourself and forgive others.
**    May you share freely, never taking more than you give.
**
*************************************************************************
** This file contains code used to implement the PRAGMA command.
**
** $Id: pragma.c,v 1.105 2005/11/30 03:20:31 drh Exp $
*/
#include "sqliteInt.h"
#include "os.h"
#include <ctype.h>

/* Ignore this whole file if pragmas are disabled
*/
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
      if( sqlite3_temp_directory ){
        sqlite3VdbeSetNumCols(v, 1);
        sqlite3VdbeSetColName(v, 0, "temp_store_directory", P3_STATIC);
        sqlite3VdbeOp3(v, OP_String8, 0, 0, sqlite3_temp_directory, 0);
        sqlite3VdbeAddOp(v, OP_Callback, 1, 0);
      }
    }else{
      if( zRight[0] && !sqlite3Io.xIsDirWritable(zRight) ){
        sqlite3ErrorMsg(pParse, "not a writable directory");
        goto pragma_out;
      }
      if( TEMP_STORE==0
       || (TEMP_STORE==1 && db->temp_store<=1)
       || (TEMP_STORE==2 && db->temp_store==1)
      ){







|







388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
      if( sqlite3_temp_directory ){
        sqlite3VdbeSetNumCols(v, 1);
        sqlite3VdbeSetColName(v, 0, "temp_store_directory", P3_STATIC);
        sqlite3VdbeOp3(v, OP_String8, 0, 0, sqlite3_temp_directory, 0);
        sqlite3VdbeAddOp(v, OP_Callback, 1, 0);
      }
    }else{
      if( zRight[0] && !sqlite3Os.xIsDirWritable(zRight) ){
        sqlite3ErrorMsg(pParse, "not a writable directory");
        goto pragma_out;
      }
      if( TEMP_STORE==0
       || (TEMP_STORE==1 && db->temp_store<=1)
       || (TEMP_STORE==2 && db->temp_store==1)
      ){
Changes to src/random.c.
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
*************************************************************************
** This file contains code to implement a pseudo-random number
** generator (PRNG) for SQLite.
**
** Random numbers are used by some of the database backends in order
** to generate random integer keys for tables or random filenames.
**
** $Id: random.c,v 1.13 2005/06/12 21:35:52 drh Exp $
*/
#include "sqliteInt.h"
#include "os.h"


/*
** Get a single 8-bit random value from the RC4 PRNG.  The Mutex







|







11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
*************************************************************************
** This file contains code to implement a pseudo-random number
** generator (PRNG) for SQLite.
**
** Random numbers are used by some of the database backends in order
** to generate random integer keys for tables or random filenames.
**
** $Id: random.c,v 1.14 2005/11/30 03:20:32 drh Exp $
*/
#include "sqliteInt.h"
#include "os.h"


/*
** Get a single 8-bit random value from the RC4 PRNG.  The Mutex
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
  ** number generator) not as an encryption device.
  */
  if( !prng.isInit ){
    int i;
    char k[256];
    prng.j = 0;
    prng.i = 0;
    sqlite3OsRandomSeed(k);
    for(i=0; i<256; i++){
      prng.s[i] = i;
    }
    for(i=0; i<256; i++){
      prng.j += prng.s[i] + k[i];
      t = prng.s[prng.j];
      prng.s[prng.j] = prng.s[i];







|







59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
  ** number generator) not as an encryption device.
  */
  if( !prng.isInit ){
    int i;
    char k[256];
    prng.j = 0;
    prng.i = 0;
    sqlite3Os.xRandomSeed(k);
    for(i=0; i<256; i++){
      prng.s[i] = i;
    }
    for(i=0; i<256; i++){
      prng.j += prng.s[i] + k[i];
      t = prng.s[prng.j];
      prng.s[prng.j] = prng.s[i];
88
89
90
91
92
93
94
95
96
97
98
99
100
}

/*
** Return N random bytes.
*/
void sqlite3Randomness(int N, void *pBuf){
  unsigned char *zBuf = pBuf;
  sqlite3OsEnterMutex();
  while( N-- ){
    *(zBuf++) = randomByte();
  }
  sqlite3OsLeaveMutex();
}







|



|

88
89
90
91
92
93
94
95
96
97
98
99
100
}

/*
** Return N random bytes.
*/
void sqlite3Randomness(int N, void *pBuf){
  unsigned char *zBuf = pBuf;
  sqlite3Os.xEnterMutex();
  while( N-- ){
    *(zBuf++) = randomByte();
  }
  sqlite3Os.xLeaveMutex();
}
Changes to src/test1.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 printf() interface to SQLite.  This code
** is not included in the SQLite library.  It is used for automated
** testing of the SQLite library.
**
** $Id: test1.c,v 1.167 2005/11/29 03:13:22 drh Exp $
*/
#include "sqliteInt.h"
#include "tcl.h"
#include "os.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 printf() interface to SQLite.  This code
** is not included in the SQLite library.  It is used for automated
** testing of the SQLite library.
**
** $Id: test1.c,v 1.168 2005/11/30 03:20:32 drh Exp $
*/
#include "sqliteInt.h"
#include "tcl.h"
#include "os.h"
#include <stdlib.h>
#include <string.h>

2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450

  if( objc!=2 ){
    Tcl_AppendResult(interp, "wrong # args: should be \"", 
       Tcl_GetString(objv[0]), " filename", 0);
    return TCL_ERROR;
  }

  rc = sqlite3Io.xOpenReadWrite(Tcl_GetString(objv[1]), &pFile, &dummy);
  if( rc!=SQLITE_OK ){
    Tcl_SetResult(interp, (char *)errorName(rc), TCL_STATIC);
    return TCL_ERROR;
  }
  makePointerStr(interp, zBuf, pFile);
  Tcl_SetResult(interp, zBuf, 0);
  return TCL_ERROR;







|







2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450

  if( objc!=2 ){
    Tcl_AppendResult(interp, "wrong # args: should be \"", 
       Tcl_GetString(objv[0]), " filename", 0);
    return TCL_ERROR;
  }

  rc = sqlite3Os.xOpenReadWrite(Tcl_GetString(objv[1]), &pFile, &dummy);
  if( rc!=SQLITE_OK ){
    Tcl_SetResult(interp, (char *)errorName(rc), TCL_STATIC);
    return TCL_ERROR;
  }
  makePointerStr(interp, zBuf, pFile);
  Tcl_SetResult(interp, zBuf, 0);
  return TCL_ERROR;
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
       Tcl_GetString(objv[0]), " filehandle", 0);
    return TCL_ERROR;
  }

  if( getFilePointer(interp, Tcl_GetString(objv[1]), &pFile) ){
    return TCL_ERROR;
  }
  rc = sqlite3Io.xClose(&pFile);
  if( rc!=SQLITE_OK ){
    Tcl_SetResult(interp, (char *)errorName(rc), TCL_STATIC);
    return TCL_ERROR;
  }
  return TCL_OK;
}








|







2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
       Tcl_GetString(objv[0]), " filehandle", 0);
    return TCL_ERROR;
  }

  if( getFilePointer(interp, Tcl_GetString(objv[1]), &pFile) ){
    return TCL_ERROR;
  }
  rc = sqlite3OsClose(&pFile);
  if( rc!=SQLITE_OK ){
    Tcl_SetResult(interp, (char *)errorName(rc), TCL_STATIC);
    return TCL_ERROR;
  }
  return TCL_OK;
}

2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
  }

  if( getFilePointer(interp, Tcl_GetString(objv[1]), &pFile) ){
    return TCL_ERROR;
  }

  if( 0==strcmp("SHARED", Tcl_GetString(objv[2])) ){
    rc = sqlite3Io.xLock(pFile, SHARED_LOCK);
  }
  else if( 0==strcmp("RESERVED", Tcl_GetString(objv[2])) ){
    rc = sqlite3Io.xLock(pFile, RESERVED_LOCK);
  }
  else if( 0==strcmp("PENDING", Tcl_GetString(objv[2])) ){
    rc = sqlite3Io.xLock(pFile, PENDING_LOCK);
  }
  else if( 0==strcmp("EXCLUSIVE", Tcl_GetString(objv[2])) ){
    rc = sqlite3Io.xLock(pFile, EXCLUSIVE_LOCK);
  }else{
    Tcl_AppendResult(interp, "wrong # args: should be \"", 
        Tcl_GetString(objv[0]), 
        " filehandle (SHARED|RESERVED|PENDING|EXCLUSIVE)", 0);
    return TCL_ERROR;
  }








|


|


|


|







2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
  }

  if( getFilePointer(interp, Tcl_GetString(objv[1]), &pFile) ){
    return TCL_ERROR;
  }

  if( 0==strcmp("SHARED", Tcl_GetString(objv[2])) ){
    rc = sqlite3OsLock(pFile, SHARED_LOCK);
  }
  else if( 0==strcmp("RESERVED", Tcl_GetString(objv[2])) ){
    rc = sqlite3OsLock(pFile, RESERVED_LOCK);
  }
  else if( 0==strcmp("PENDING", Tcl_GetString(objv[2])) ){
    rc = sqlite3OsLock(pFile, PENDING_LOCK);
  }
  else if( 0==strcmp("EXCLUSIVE", Tcl_GetString(objv[2])) ){
    rc = sqlite3OsLock(pFile, EXCLUSIVE_LOCK);
  }else{
    Tcl_AppendResult(interp, "wrong # args: should be \"", 
        Tcl_GetString(objv[0]), 
        " filehandle (SHARED|RESERVED|PENDING|EXCLUSIVE)", 0);
    return TCL_ERROR;
  }

2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
       Tcl_GetString(objv[0]), " filehandle", 0);
    return TCL_ERROR;
  }

  if( getFilePointer(interp, Tcl_GetString(objv[1]), &pFile) ){
    return TCL_ERROR;
  }
  rc = sqlite3Io.xUnlock(pFile, NO_LOCK);
  if( rc!=SQLITE_OK ){
    Tcl_SetResult(interp, (char *)errorName(rc), TCL_STATIC);
    return TCL_ERROR;
  }
  return TCL_OK;
}

/*
** Usage:  sqlite3OsTempFileName
*/
static int test_sqlite3OsTempFileName(
  void * clientData,
  Tcl_Interp *interp,
  int objc,
  Tcl_Obj *CONST objv[]
){
  char zFile[SQLITE_TEMPNAME_SIZE];
  int rc;

  rc = sqlite3Io.xTempFileName(zFile);
  if( rc!=SQLITE_OK ){
    Tcl_SetResult(interp, (char *)errorName(rc), TCL_STATIC);
    return TCL_ERROR;
  }
  Tcl_AppendResult(interp, zFile, 0);
  return TCL_OK;
}







|



















|







2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
       Tcl_GetString(objv[0]), " filehandle", 0);
    return TCL_ERROR;
  }

  if( getFilePointer(interp, Tcl_GetString(objv[1]), &pFile) ){
    return TCL_ERROR;
  }
  rc = sqlite3OsUnlock(pFile, NO_LOCK);
  if( rc!=SQLITE_OK ){
    Tcl_SetResult(interp, (char *)errorName(rc), TCL_STATIC);
    return TCL_ERROR;
  }
  return TCL_OK;
}

/*
** Usage:  sqlite3OsTempFileName
*/
static int test_sqlite3OsTempFileName(
  void * clientData,
  Tcl_Interp *interp,
  int objc,
  Tcl_Obj *CONST objv[]
){
  char zFile[SQLITE_TEMPNAME_SIZE];
  int rc;

  rc = sqlite3Os.xTempFileName(zFile);
  if( rc!=SQLITE_OK ){
    Tcl_SetResult(interp, (char *)errorName(rc), TCL_STATIC);
    return TCL_ERROR;
  }
  Tcl_AppendResult(interp, zFile, 0);
  return TCL_OK;
}
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 1.37 2005/11/29 03:13:22 drh 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 1.38 2005/11/30 03:20:32 drh Exp $
*/
#include "sqliteInt.h"
#include "os.h"
#include "pager.h"
#include "tcl.h"
#include <stdlib.h>
#include <string.h>
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
  int readOnly = 0;
  if( argc!=3 ){
    Tcl_AppendResult(interp, "wrong # args: should be \"", argv[0],
       " N-MEGABYTES FILE\"", 0);
    return TCL_ERROR;
  }
  if( Tcl_GetInt(interp, argv[1], &n) ) return TCL_ERROR;
  rc = sqlite3Io.xOpenReadWrite(argv[2], &fd, &readOnly);
  if( rc ){
    Tcl_AppendResult(interp, "open failed: ", errorName(rc), 0);
    return TCL_ERROR;
  }
  offset = n;
  offset *= 1024*1024;
  rc = sqlite3Io.xSeek(fd, offset);
  if( rc ){
    Tcl_AppendResult(interp, "seek failed: ", errorName(rc), 0);
    return TCL_ERROR;
  }
  rc = sqlite3Io.xWrite(fd, "Hello, World!", 14);
  sqlite3Io.xClose(&fd);
  if( rc ){
    Tcl_AppendResult(interp, "write failed: ", errorName(rc), 0);
    return TCL_ERROR;
  }
  return TCL_OK;
}
#endif







|






|




|
|







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
  int readOnly = 0;
  if( argc!=3 ){
    Tcl_AppendResult(interp, "wrong # args: should be \"", argv[0],
       " N-MEGABYTES FILE\"", 0);
    return TCL_ERROR;
  }
  if( Tcl_GetInt(interp, argv[1], &n) ) return TCL_ERROR;
  rc = sqlite3Os.xOpenReadWrite(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;
}
#endif
Changes to src/test6.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
*/
#if SQLITE_TEST          /* This file is used for the testing only */
#include "sqliteInt.h"
#include "os.h"
#include "tcl.h"

/*
** A copy of the original sqlite3Io structure
*/
static struct sqlite3IoVtbl origIo;

/*
** The OsFile structure for the crash-test backend.  The pBase field
** points to an OsFile structure for the native backend.
*/

struct OsFile {

  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;      /* Base class */
  OsFile *pNext;      /* Next in a list of them all */
};

/*
** Size of a simulated disk block
*/
#define BLOCKSIZE 512
#define BLOCK_OFFSET(x) ((x) * BLOCKSIZE)







|

|


|
|

>
|
>
|
|
|
|
|
|
|







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
*/
#if SQLITE_TEST          /* This file is used for the testing only */
#include "sqliteInt.h"
#include "os.h"
#include "tcl.h"

/*
** A copy of the original sqlite3Os structure
*/
static struct sqlite3OsVtbl origOs;

/*
** crashFile is a subclass of OsFile that is taylored for the
** crash test module.
*/
typedef struct crashFile crashFile;
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
*/
#define BLOCKSIZE 512
#define BLOCK_OFFSET(x) ((x) * BLOCKSIZE)
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
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
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
static int iCrashDelay = 0;
static char zCrashFile[500];

/*
** 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();
}

/*
** 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;
}

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

/*
** Initialise the os_test.c specific fields of pFile.
*/
static void initFile(OsFile **pId, char const *zName, OsFile *pBase){
  OsFile *pFile = *pId = sqliteMalloc(sizeof(OsFile) + strlen(zName)+1);
  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;
}

/*
** Undo the work done by initFile. Delete the OsFile structure
** and unlink the structure from the pAllFiles list.
*/
static void closeFile(OsFile **pId){
  OsFile *pFile = *pId;
  if( pFile==pAllFiles ){
    pAllFiles = pFile->pNext;
  }else{
    OsFile *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(OsFile *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 = origIo.xFileSize(pFile->pBase, &filesize);
    if( rc!=SQLITE_OK ) return rc;

    if( BLOCK_OFFSET(blk)<filesize ){
      int len = BLOCKSIZE;
      rc = origIo.xSeek(pFile->pBase, blk*BLOCKSIZE);
      if( BLOCK_OFFSET(blk+1)>filesize ){
        len = filesize - BLOCK_OFFSET(blk);
      }
      if( rc!=SQLITE_OK ) return rc;
      rc = origIo.xRead(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(OsFile *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 ){







|



|









|













|






|

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





|
|



|












|

















|




|




|











|







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
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
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
static int iCrashDelay = 0;
static char zCrashFile[500];

/*
** Set the value of the two crash parameters.
*/
static void setCrashParams(int iDelay, char const *zFile){
  sqlite3Os.xEnterMutex();
  assert( strlen(zFile)<sizeof(zCrashFile) );
  strcpy(zCrashFile, zFile);
  iCrashDelay = iDelay;
  sqlite3Os.xLeaveMutex();
}

/*
** 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;
  sqlite3Os.xEnterMutex();
  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;
    }
  }
  sqlite3Os.xLeaveMutex();
  return r;
}

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

/* 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 ){
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
        }else{
#ifdef TRACE_WRITECACHE
printf("Writing block %d of %s\n", i, pFile->zName); 
#endif
        }
      }
      if( rc==SQLITE_OK ){
        rc = origIo.xSeek(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 = origIo.xWrite(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(OsFile *pFile){
  if( pFile->apBlk ){
    int c = crashRequired(pFile->zName);
    if( c ){
      OsFile *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){
  OsFile *pFile = *pId;
  if( pFile ){
    /* printf("CLOSE %s (%d blocks)\n", pFile->zName, pFile->nBlk); */
    writeCache(pFile);
    origIo.xClose(&pFile->pBase);
  }
  closeFile(pId);

  return SQLITE_OK;
}

static int crashSeek(OsFile *id, i64 offset){
  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;
  OsFile *pFile = id;

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

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







|










|















|



|


















|



|

|
>




|










|







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
        }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++){
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
      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 = origIo.xSeek(id->pBase, BLOCK_OFFSET(i) + off);
      if( rc!=SQLITE_OK ) return rc;
      rc = origIo.xRead(id->pBase, zCsr, len);
      if( rc!=SQLITE_OK ) return rc;
    }

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

  id->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;


  offset = id->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(id, i);
    if( rc!=SQLITE_OK ) return rc;

    /* Write into the cache */
    pBlk = id->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( id->nMaxWrite<end ){
    id->nMaxWrite = end;
  }
  assert( zCsr==&((u8 *)pBuf)[amt] );
  id->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){
  int rc;
  /* printf("SYNC %s (%d blocks)\n", (*id)->zName, (*id)->nBlk); */
  rc = writeCache(id);
  /* if( rc!=SQLITE_OK ) return rc; rc = origIo.xSync(id->pBase, dataOnly); */
  return rc;
}

/*
** 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){

  id->nMaxWrite = nByte;
  return origIo.xTruncate(id->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){

  int rc = origIo.xFileSize(id->pBase, pSize);
  if( rc==SQLITE_OK && pSize && *pSize<id->nMaxWrite ){
    *pSize = id->nMaxWrite;
  }
  return rc;
}

/*
** 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.
*/
static int crashOpenReadWrite(const char *zFilename, OsFile **pId,int *pRdonly){
  OsFile *pBase = 0;
  int rc = origIo.xOpenReadWrite(zFilename, &pBase, pRdonly);
  if( !rc ){
    initFile(pId, zFilename, pBase);
  }
  return rc;
}
static int crashOpenExclusive(const char *zFilename, OsFile **pId, int delFlag){
  OsFile *pBase = 0;
  int rc = origIo.xOpenExclusive(zFilename, &pBase, delFlag);
  if( !rc ){
    initFile(pId, zFilename, pBase);
  }
  return rc;
}
static int crashOpenReadOnly(const char *zFilename, OsFile **pId){
  OsFile *pBase = 0;
  int rc = origIo.xOpenReadOnly(zFilename, &pBase);
  if( !rc ){
    initFile(pId, zFilename, pBase);
  }
  return rc;
}

/*
** OpenDirectory and SyncDirectory are no-ops
*/
static int crashOpenDir(const char *zName, OsFile *id){
  return SQLITE_OK;
}
static int crashSyncDir(const char *zName){
  return SQLITE_OK;
}

/*
** Locking primitives are passed through into the underlying
** file descriptor.
*/
int crashLock(OsFile *id, int lockType){
  return origIo.xLock(id->pBase, lockType);
}
int crashUnlock(OsFile *id, int lockType){
  return origIo.xUnlock(id->pBase, lockType);
}
int crashCheckReservedLock(OsFile *id){
  return origIo.xCheckReservedLock(id->pBase);
}
void crashSetFullSync(OsFile *id, int setting){
  return;  /* This is a no-op */
}
int crashLockState(OsFile *id){
  return origIo.xLockState(id->pBase);
}

/*
** Return the underlying file handle.
*/
int crashFileHandle(OsFile *id){
















  return origIo.xFileHandle(id->pBase);






}




















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







|

|







|










>

|










|



|












|
|


|








<
<
<
<
|








>
|
|







>
|
|
|











|







|







|







|

|
<
<
<








|


|


|





|






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







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
      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;
}

/*
** 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.
*/
static int crashOpenReadWrite(const char *zFilename, OsFile **pId,int *pRdonly){
  OsFile *pBase = 0;
  int rc = origOs.xOpenReadWrite(zFilename, &pBase, pRdonly);
  if( !rc ){
    initFile(pId, zFilename, pBase);
  }
  return rc;
}
static int crashOpenExclusive(const char *zFilename, OsFile **pId, int delFlag){
  OsFile *pBase = 0;
  int rc = origOs.xOpenExclusive(zFilename, &pBase, delFlag);
  if( !rc ){
    initFile(pId, zFilename, pBase);
  }
  return rc;
}
static int crashOpenReadOnly(const char *zFilename, OsFile **pId){
  OsFile *pBase = 0;
  int rc = origOs.xOpenReadOnly(zFilename, &pBase);
  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){
  return sqlite3OsFileHandle(((crashFile*)id)->pBase);
}

/*
** 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,
};


/*
** 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
**
** This procedure implements a TCL command that enables crash testing
** in testfixture.  Once enabled, crash testing cannot be disabled.
*/
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
  if( Tcl_GetIntFromObj(interp, objv[1], &delay) ) 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(delay, zFile);

  origIo = sqlite3Io;
  /* xDelete unchanged */
  /* xFileExists unchanged */
  sqlite3Io.xOpenReadWrite = crashOpenReadWrite;
  sqlite3Io.xOpenExclusive = crashOpenExclusive;
  sqlite3Io.xOpenReadOnly = crashOpenReadOnly;
  sqlite3Io.xOpenDirectory = crashOpenDir;
  sqlite3Io.xSyncDirectory = crashSyncDir;
  /* xTempFileName unchanged */
  /* xIsDirWritable unchanged */
  sqlite3Io.xClose = crashClose;
  sqlite3Io.xRead = crashRead;
  sqlite3Io.xWrite = crashWrite;
  sqlite3Io.xSeek = crashSeek;
  sqlite3Io.xSync = crashSync;
  sqlite3Io.xTruncate = crashTruncate;
  sqlite3Io.xFileSize = crashFileSize;
  /* xFullPathname unchanged */
  sqlite3Io.xLock = crashLock;
  sqlite3Io.xUnlock = crashUnlock;
  sqlite3Io.xCheckReservedLock = crashCheckReservedLock;
  sqlite3Io.xSetFullSync = crashSetFullSync;
  sqlite3Io.xFileHandle = crashFileHandle;
  sqlite3Io.xLockState = crashLockState;

  return TCL_OK;
}

/*
** This procedure registers the TCL procedures defined in this file.
*/
int Sqlitetest6_Init(Tcl_Interp *interp){
  Tcl_CreateObjCommand(interp, "sqlite3_crashparams", crashParamsObjCmd, 0, 0);
  return TCL_OK;
}

#endif /* SQLITE_TEST */







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












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
  if( Tcl_GetIntFromObj(interp, objv[1], &delay) ) 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(delay, zFile);
  if( origOs.xOpenReadWrite==0 ){
    origOs = sqlite3Os;


    sqlite3Os.xOpenReadWrite = crashOpenReadWrite;
    sqlite3Os.xOpenExclusive = crashOpenExclusive;
    sqlite3Os.xOpenReadOnly = crashOpenReadOnly;


















  }
  return TCL_OK;
}

/*
** This procedure registers the TCL procedures defined in this file.
*/
int Sqlitetest6_Init(Tcl_Interp *interp){
  Tcl_CreateObjCommand(interp, "sqlite3_crashparams", crashParamsObjCmd, 0, 0);
  return TCL_OK;
}

#endif /* SQLITE_TEST */
Changes to src/vacuum.c.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
**
*************************************************************************
** This file contains code used to implement the VACUUM command.
**
** Most of the code in this file may be omitted by defining the
** SQLITE_OMIT_VACUUM macro.
**
** $Id: vacuum.c,v 1.48 2005/11/26 00:25:04 drh Exp $
*/
#include "sqliteInt.h"
#include "os.h"

#ifndef SQLITE_OMIT_VACUUM
/*
** Generate a random name of 20 character in length.







|







10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
**
*************************************************************************
** This file contains code used to implement the VACUUM command.
**
** Most of the code in this file may be omitted by defining the
** SQLITE_OMIT_VACUUM macro.
**
** $Id: vacuum.c,v 1.49 2005/11/30 03:20:32 drh Exp $
*/
#include "sqliteInt.h"
#include "os.h"

#ifndef SQLITE_OMIT_VACUUM
/*
** Generate a random name of 20 character in length.
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
  ** run more than once or twice is vanishingly small.  We are certain
  ** enough that this loop will always terminate (and terminate quickly)
  ** that we don't even bother to set a maximum loop count.
  */
  do {
    zTemp[nFilename] = '-';
    randomName((unsigned char*)&zTemp[nFilename+1]);
  } while( sqlite3Io.xFileExists(zTemp) );

  /* Attach the temporary database as 'vacuum_db'. The synchronous pragma
  ** can be set to 'off' for this file, as it is not recovered if a crash
  ** occurs anyway. The integrity of the database is maintained by a
  ** (possibly synchronous) transaction opened on the main database before
  ** sqlite3BtreeCopyFile() is called.
  **







|







142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
  ** run more than once or twice is vanishingly small.  We are certain
  ** enough that this loop will always terminate (and terminate quickly)
  ** that we don't even bother to set a maximum loop count.
  */
  do {
    zTemp[nFilename] = '-';
    randomName((unsigned char*)&zTemp[nFilename+1]);
  } while( sqlite3Os.xFileExists(zTemp) );

  /* Attach the temporary database as 'vacuum_db'. The synchronous pragma
  ** can be set to 'off' for this file, as it is not recovered if a crash
  ** occurs anyway. The integrity of the database is maintained by a
  ** (possibly synchronous) transaction opened on the main database before
  ** sqlite3BtreeCopyFile() is called.
  **
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
  */
  db->autoCommit = 1;
  rc2 = execSql(db, "DETACH vacuum_db;");
  if( rc==SQLITE_OK ){
    rc = rc2;
  }
  if( zTemp ){
    sqlite3Io.xDelete(zTemp);
    sqliteFree(zTemp);
  }
  sqliteFree( zSql );
  sqlite3ResetInternalSchema(db, 0);
#endif

  return rc;
}







|








296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
  */
  db->autoCommit = 1;
  rc2 = execSql(db, "DETACH vacuum_db;");
  if( rc==SQLITE_OK ){
    rc = rc2;
  }
  if( zTemp ){
    sqlite3Os.xDelete(zTemp);
    sqliteFree(zTemp);
  }
  sqliteFree( zSql );
  sqlite3ResetInternalSchema(db, 0);
#endif

  return rc;
}
Changes to src/vdbe.c.
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
**
** Various scripts scan this source file in order to generate HTML
** documentation, headers files, or other derived files.  The formatting
** of the code in this file is, therefore, important.  See other comments
** in this file for details.  If in doubt, do not deviate from existing
** commenting and indentation practices when changing or adding code.
**
** $Id: vdbe.c,v 1.500 2005/11/26 00:25:04 drh Exp $
*/
#include "sqliteInt.h"
#include "os.h"
#include <ctype.h>
#include "vdbeInt.h"

/*







|







39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
**
** Various scripts scan this source file in order to generate HTML
** documentation, headers files, or other derived files.  The formatting
** of the code in this file is, therefore, important.  See other comments
** in this file for details.  If in doubt, do not deviate from existing
** commenting and indentation practices when changing or adding code.
**
** $Id: vdbe.c,v 1.501 2005/11/30 03:20:32 drh Exp $
*/
#include "sqliteInt.h"
#include "os.h"
#include <ctype.h>
#include "vdbeInt.h"

/*
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
    if( p->trace ){
      if( pc==0 ){
        printf("VDBE Execution Trace:\n");
        sqlite3VdbePrintSql(p);
      }
      sqlite3VdbePrintOp(p->trace, pc, pOp);
    }
    if( p->trace==0 && pc==0 && sqlite3Io.xFileExists("vdbe_sqltrace") ){
      sqlite3VdbePrintSql(p);
    }
#endif
      

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







|







418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
    if( p->trace ){
      if( pc==0 ){
        printf("VDBE Execution Trace:\n");
        sqlite3VdbePrintSql(p);
      }
      sqlite3VdbePrintOp(p->trace, pc, pOp);
    }
    if( p->trace==0 && pc==0 && sqlite3Os.xFileExists("vdbe_sqltrace") ){
      sqlite3VdbePrintSql(p);
    }
#endif
      

    /* Check to see if we need to simulate an interrupt.  This only happens
    ** if we have a special test build.
Changes to src/vdbeapi.c.
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
      if( sqlite3SafetyOn(db) ){
        p->rc = SQLITE_MISUSE;
        return SQLITE_MISUSE;
      }
    }
    if( db->xProfile && !db->init.busy ){
      double rNow;
      sqlite3OsCurrentTime(&rNow);
      p->startTime = (rNow - (int)rNow)*3600.0*24.0*1000000000.0;
    }
#endif

    /* Print a copy of SQL as it is executed if the SQL_TRACE pragma is turned
    ** on in debugging mode.
    */







|







187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
      if( sqlite3SafetyOn(db) ){
        p->rc = SQLITE_MISUSE;
        return SQLITE_MISUSE;
      }
    }
    if( db->xProfile && !db->init.busy ){
      double rNow;
      sqlite3Os.xCurrentTime(&rNow);
      p->startTime = (rNow - (int)rNow)*3600.0*24.0*1000000000.0;
    }
#endif

    /* Print a copy of SQL as it is executed if the SQL_TRACE pragma is turned
    ** on in debugging mode.
    */
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
#ifndef SQLITE_OMIT_TRACE
  /* Invoke the profile callback if there is one
  */
  if( rc!=SQLITE_ROW && db->xProfile && !db->init.busy ){
    double rNow;
    u64 elapseTime;

    sqlite3OsCurrentTime(&rNow);
    elapseTime = (rNow - (int)rNow)*3600.0*24.0*1000000000.0 - p->startTime;
    assert( p->nOp>0 );
    assert( p->aOp[p->nOp-1].opcode==OP_Noop );
    assert( p->aOp[p->nOp-1].p3!=0 );
    assert( p->aOp[p->nOp-1].p3type==P3_DYNAMIC );
    db->xProfile(db->pProfileArg, p->aOp[p->nOp-1].p3, elapseTime);
  }







|







224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
#ifndef SQLITE_OMIT_TRACE
  /* Invoke the profile callback if there is one
  */
  if( rc!=SQLITE_ROW && db->xProfile && !db->init.busy ){
    double rNow;
    u64 elapseTime;

    sqlite3Os.xCurrentTime(&rNow);
    elapseTime = (rNow - (int)rNow)*3600.0*24.0*1000000000.0 - p->startTime;
    assert( p->nOp>0 );
    assert( p->aOp[p->nOp-1].opcode==OP_Noop );
    assert( p->aOp[p->nOp-1].p3!=0 );
    assert( p->aOp[p->nOp-1].p3type==P3_DYNAMIC );
    db->xProfile(db->pProfileArg, p->aOp[p->nOp-1].p3, elapseTime);
  }
Changes to src/vdbeaux.c.
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
  p->aLabel = 0;

  *pMaxFuncArgs = nMaxArgs;
  *pMaxStack = nMaxStack;

  /* If we never rollback a statement transaction, then statement
  ** transactions are not needed.  So change every OP_Statement
  ** opcode into an OP_Noop.  This avoid a call to sqlite3Io.xOpenExclusive()
  ** which can be expensive on some platforms.
  */
  if( hasStatementBegin && !doesStatementRollback ){
    for(pOp=p->aOp, i=p->nOp-1; i>=0; i--, pOp++){
      if( pOp->opcode==OP_Statement ){
        pOp->opcode = OP_Noop;
      }







|







269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
  p->aLabel = 0;

  *pMaxFuncArgs = nMaxArgs;
  *pMaxStack = nMaxStack;

  /* If we never rollback a statement transaction, then statement
  ** transactions are not needed.  So change every OP_Statement
  ** opcode into an OP_Noop.  This avoid a call to sqlite3Os.xOpenExclusive()
  ** which can be expensive on some platforms.
  */
  if( hasStatementBegin && !doesStatementRollback ){
    for(pOp=p->aOp, i=p->nOp-1; i>=0; i--, pOp++){
      if( pOp->opcode==OP_Statement ){
        pOp->opcode = OP_Noop;
      }
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
  }
  for(n=0; n<p->nMem; n++){
    p->aMem[n].flags = MEM_Null;
  }

#ifdef SQLITE_DEBUG
  if( (p->db->flags & SQLITE_VdbeListing)!=0
    || sqlite3Io.xFileExists("vdbe_explain")
  ){
    int i;
    printf("VDBE Program Listing:\n");
    sqlite3VdbePrintSql(p);
    for(i=0; i<p->nOp; i++){
      sqlite3VdbePrintOp(stdout, i, &p->aOp[i]);
    }
  }
  if( sqlite3Io.xFileExists("vdbe_trace") ){
    p->trace = stdout;
  }
#endif
  p->pTos = &p->aStack[-1];
  p->pc = -1;
  p->rc = SQLITE_OK;
  p->uniqueCnt = 0;







|








|







751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
  }
  for(n=0; n<p->nMem; n++){
    p->aMem[n].flags = MEM_Null;
  }

#ifdef SQLITE_DEBUG
  if( (p->db->flags & SQLITE_VdbeListing)!=0
    || sqlite3Os.xFileExists("vdbe_explain")
  ){
    int i;
    printf("VDBE Program Listing:\n");
    sqlite3VdbePrintSql(p);
    for(i=0; i<p->nOp; i++){
      sqlite3VdbePrintOp(stdout, i, &p->aOp[i]);
    }
  }
  if( sqlite3Os.xFileExists("vdbe_trace") ){
    p->trace = stdout;
  }
#endif
  p->pTos = &p->aStack[-1];
  p->pc = -1;
  p->rc = SQLITE_OK;
  p->uniqueCnt = 0;
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
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
      u32 random;
      sqliteFree(zMaster);
      sqlite3Randomness(sizeof(random), &random);
      zMaster = sqlite3MPrintf("%s-mj%08X", zMainFile, random&0x7fffffff);
      if( !zMaster ){
        return SQLITE_NOMEM;
      }
    }while( sqlite3Io.xFileExists(zMaster) );

    /* Open the master journal. */
    rc = sqlite3Io.xOpenExclusive(zMaster, &master, 0);
    if( rc!=SQLITE_OK ){
      sqliteFree(zMaster);
      return rc;
    }
 
    /* Write the name of each database file in the transaction into the new
    ** master journal file. If an error occurs at this point close
    ** and delete the master journal file. All the individual journal files
    ** still have 'null' as the master journal pointer, so they will roll
    ** back independently if a failure occurs.
    */
    for(i=0; i<db->nDb; i++){ 
      Btree *pBt = db->aDb[i].pBt;
      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 = sqlite3Io.xWrite(master, zFile, strlen(zFile)+1);
        if( rc!=SQLITE_OK ){
          sqlite3Io.xClose(&master);
          sqlite3Io.xDelete(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 = sqlite3Io.xOpenDirectory(zMainFile, master);
    if( rc!=SQLITE_OK ||
          (needSync && (rc=sqlite3Io.xSync(master,0))!=SQLITE_OK) ){
      sqlite3Io.xClose(&master);
      sqlite3Io.xDelete(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 sqlite3BtreeSync(),
    ** then there is a chance that the master journal file will be
    ** orphaned. But we cannot delete it, in case the master journal
    ** file name was written into the journal file before the failure
    ** occured.
    */
    for(i=0; i<db->nDb; i++){ 
      Btree *pBt = db->aDb[i].pBt;
      if( pBt && sqlite3BtreeIsInTrans(pBt) ){
        rc = sqlite3BtreeSync(pBt, zMaster);
        if( rc!=SQLITE_OK ){
          sqlite3Io.xClose(&master);
          sqliteFree(zMaster);
          return rc;
        }
      }
    }
    sqlite3Io.xClose(&master);

    /* Delete the master journal file. This commits the transaction. After
    ** doing this the directory is synced again before any individual
    ** transaction files are deleted.
    */
    rc = sqlite3Io.xDelete(zMaster);
    assert( rc==SQLITE_OK );
    sqliteFree(zMaster);
    zMaster = 0;
    rc = sqlite3Io.xSyncDirectory(zMainFile);
    if( rc!=SQLITE_OK ){
      /* This is not good. The master journal file has been deleted, but
      ** the directory sync failed. There is no completely safe course of
      ** action from here. The individual journals contain the name of the
      ** master journal file, but there is no way of knowing if that
      ** master journal exists now or if it will exist after the operating
      ** system crash that may follow the fsync() failure.







|


|




















|

|
|











|

|
|
|



















|





|





|



|







974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
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
      u32 random;
      sqliteFree(zMaster);
      sqlite3Randomness(sizeof(random), &random);
      zMaster = sqlite3MPrintf("%s-mj%08X", zMainFile, random&0x7fffffff);
      if( !zMaster ){
        return SQLITE_NOMEM;
      }
    }while( sqlite3Os.xFileExists(zMaster) );

    /* Open the master journal. */
    rc = sqlite3Os.xOpenExclusive(zMaster, &master, 0);
    if( rc!=SQLITE_OK ){
      sqliteFree(zMaster);
      return rc;
    }
 
    /* Write the name of each database file in the transaction into the new
    ** master journal file. If an error occurs at this point close
    ** and delete the master journal file. All the individual journal files
    ** still have 'null' as the master journal pointer, so they will roll
    ** back independently if a failure occurs.
    */
    for(i=0; i<db->nDb; i++){ 
      Btree *pBt = db->aDb[i].pBt;
      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);
          sqlite3Os.xDelete(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);
      sqlite3Os.xDelete(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 sqlite3BtreeSync(),
    ** then there is a chance that the master journal file will be
    ** orphaned. But we cannot delete it, in case the master journal
    ** file name was written into the journal file before the failure
    ** occured.
    */
    for(i=0; i<db->nDb; i++){ 
      Btree *pBt = db->aDb[i].pBt;
      if( pBt && sqlite3BtreeIsInTrans(pBt) ){
        rc = sqlite3BtreeSync(pBt, zMaster);
        if( rc!=SQLITE_OK ){
          sqlite3OsClose(&master);
          sqliteFree(zMaster);
          return rc;
        }
      }
    }
    sqlite3OsClose(&master);

    /* Delete the master journal file. This commits the transaction. After
    ** doing this the directory is synced again before any individual
    ** transaction files are deleted.
    */
    rc = sqlite3Os.xDelete(zMaster);
    assert( rc==SQLITE_OK );
    sqliteFree(zMaster);
    zMaster = 0;
    rc = sqlite3Os.xSyncDirectory(zMainFile);
    if( rc!=SQLITE_OK ){
      /* This is not good. The master journal file has been deleted, but
      ** the directory sync failed. There is no completely safe course of
      ** action from here. The individual journals contain the name of the
      ** master journal file, but there is no way of knowing if that
      ** master journal exists now or if it will exist after the operating
      ** system crash that may follow the fsync() failure.