blob: 0a14e847a2c7f79a1747c251b43a9c982244610b [file] [log] [blame]
/*
** 2008 October 7
**
** 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 use to implement an in-memory rollback journal.
** The in-memory rollback journal is used to journal transactions for
** ":memory:" databases and when the journal_mode=MEMORY pragma is used.
**
** Update: The in-memory journal is also used to temporarily cache
** smaller journals that are not critical for power-loss recovery.
** For example, statement journals that are not too big will be held
** entirely in memory, thus reducing the number of file I/O calls, and
** more importantly, reducing temporary file creation events. If these
** journals become too large for memory, they are spilled to disk. But
** in the common case, they are usually small and no file I/O needs to
** occur.
*/
#include "sqliteInt.h"
/* Forward references to internal structures */
typedef struct MemJournal MemJournal;
typedef struct FilePoint FilePoint;
typedef struct FileChunk FileChunk;
/*
** The rollback journal is composed of a linked list of these structures.
**
** The zChunk array is always at least 8 bytes in size - usually much more.
** Its actual size is stored in the MemJournal.nChunkSize variable.
*/
struct FileChunk {
FileChunk *pNext; /* Next chunk in the journal */
u8 zChunk[8]; /* Content of this chunk */
};
/*
** By default, allocate this many bytes of memory for each FileChunk object.
*/
#define MEMJOURNAL_DFLT_FILECHUNKSIZE 1024
/*
** For chunk size nChunkSize, return the number of bytes that should
** be allocated for each FileChunk structure.
*/
#define fileChunkSize(nChunkSize) (sizeof(FileChunk) + ((nChunkSize)-8))
/*
** An instance of this object serves as a cursor into the rollback journal.
** The cursor can be either for reading or writing.
*/
struct FilePoint {
sqlite3_int64 iOffset; /* Offset from the beginning of the file */
FileChunk *pChunk; /* Specific chunk into which cursor points */
};
/*
** This structure is a subclass of sqlite3_file. Each open memory-journal
** is an instance of this class.
*/
struct MemJournal {
const sqlite3_io_methods *pMethod; /* Parent class. MUST BE FIRST */
int nChunkSize; /* In-memory chunk-size */
int nSpill; /* Bytes of data before flushing */
int nSize; /* Bytes of data currently in memory */
FileChunk *pFirst; /* Head of in-memory chunk-list */
FilePoint endpoint; /* Pointer to the end of the file */
FilePoint readpoint; /* Pointer to the end of the last xRead() */
int flags; /* xOpen flags */
sqlite3_vfs *pVfs; /* The "real" underlying VFS */
const char *zJournal; /* Name of the journal file */
};
/*
** Read data from the in-memory journal file. This is the implementation
** of the sqlite3_vfs.xRead method.
*/
static int memjrnlRead(
sqlite3_file *pJfd, /* The journal file from which to read */
void *zBuf, /* Put the results here */
int iAmt, /* Number of bytes to read */
sqlite_int64 iOfst /* Begin reading at this offset */
){
MemJournal *p = (MemJournal *)pJfd;
u8 *zOut = zBuf;
int nRead = iAmt;
int iChunkOffset;
FileChunk *pChunk;
if( (iAmt+iOfst)>p->endpoint.iOffset ){
return SQLITE_IOERR_SHORT_READ;
}
assert( p->readpoint.iOffset==0 || p->readpoint.pChunk!=0 );
if( p->readpoint.iOffset!=iOfst || iOfst==0 ){
sqlite3_int64 iOff = 0;
for(pChunk=p->pFirst;
ALWAYS(pChunk) && (iOff+p->nChunkSize)<=iOfst;
pChunk=pChunk->pNext
){
iOff += p->nChunkSize;
}
}else{
pChunk = p->readpoint.pChunk;
assert( pChunk!=0 );
}
iChunkOffset = (int)(iOfst%p->nChunkSize);
do {
int iSpace = p->nChunkSize - iChunkOffset;
int nCopy = MIN(nRead, (p->nChunkSize - iChunkOffset));
memcpy(zOut, (u8*)pChunk->zChunk + iChunkOffset, nCopy);
zOut += nCopy;
nRead -= iSpace;
iChunkOffset = 0;
} while( nRead>=0 && (pChunk=pChunk->pNext)!=0 && nRead>0 );
p->readpoint.iOffset = pChunk ? iOfst+iAmt : 0;
p->readpoint.pChunk = pChunk;
return SQLITE_OK;
}
/*
** Free the list of FileChunk structures headed at MemJournal.pFirst.
*/
static void memjrnlFreeChunks(MemJournal *p){
FileChunk *pIter;
FileChunk *pNext;
for(pIter=p->pFirst; pIter; pIter=pNext){
pNext = pIter->pNext;
sqlite3_free(pIter);
}
p->pFirst = 0;
}
/*
** Flush the contents of memory to a real file on disk.
*/
static int memjrnlCreateFile(MemJournal *p){
int rc;
sqlite3_file *pReal = (sqlite3_file*)p;
MemJournal copy = *p;
memset(p, 0, sizeof(MemJournal));
rc = sqlite3OsOpen(copy.pVfs, copy.zJournal, pReal, copy.flags, 0);
if( rc==SQLITE_OK ){
int nChunk = copy.nChunkSize;
i64 iOff = 0;
FileChunk *pIter;
for(pIter=copy.pFirst; pIter; pIter=pIter->pNext){
if( iOff + nChunk > copy.endpoint.iOffset ){
nChunk = copy.endpoint.iOffset - iOff;
}
rc = sqlite3OsWrite(pReal, (u8*)pIter->zChunk, nChunk, iOff);
if( rc ) break;
iOff += nChunk;
}
if( rc==SQLITE_OK ){
/* No error has occurred. Free the in-memory buffers. */
memjrnlFreeChunks(&copy);
}
}
if( rc!=SQLITE_OK ){
/* If an error occurred while creating or writing to the file, restore
** the original before returning. This way, SQLite uses the in-memory
** journal data to roll back changes made to the internal page-cache
** before this function was called. */
sqlite3OsClose(pReal);
*p = copy;
}
return rc;
}
/*
** Write data to the file.
*/
static int memjrnlWrite(
sqlite3_file *pJfd, /* The journal file into which to write */
const void *zBuf, /* Take data to be written from here */
int iAmt, /* Number of bytes to write */
sqlite_int64 iOfst /* Begin writing at this offset into the file */
){
MemJournal *p = (MemJournal *)pJfd;
int nWrite = iAmt;
u8 *zWrite = (u8 *)zBuf;
/* If the file should be created now, create it and write the new data
** into the file on disk. */
if( p->nSpill>0 && (iAmt+iOfst)>p->nSpill ){
int rc = memjrnlCreateFile(p);
if( rc==SQLITE_OK ){
rc = sqlite3OsWrite(pJfd, zBuf, iAmt, iOfst);
}
return rc;
}
/* If the contents of this write should be stored in memory */
else{
/* An in-memory journal file should only ever be appended to. Random
** access writes are not required. The only exception to this is when
** the in-memory journal is being used by a connection using the
** atomic-write optimization. In this case the first 28 bytes of the
** journal file may be written as part of committing the transaction. */
assert( iOfst==p->endpoint.iOffset || iOfst==0 );
#if defined(SQLITE_ENABLE_ATOMIC_WRITE) \
|| defined(SQLITE_ENABLE_BATCH_ATOMIC_WRITE)
if( iOfst==0 && p->pFirst ){
assert( p->nChunkSize>iAmt );
memcpy((u8*)p->pFirst->zChunk, zBuf, iAmt);
}else
#else
assert( iOfst>0 || p->pFirst==0 );
#endif
{
while( nWrite>0 ){
FileChunk *pChunk = p->endpoint.pChunk;
int iChunkOffset = (int)(p->endpoint.iOffset%p->nChunkSize);
int iSpace = MIN(nWrite, p->nChunkSize - iChunkOffset);
if( iChunkOffset==0 ){
/* New chunk is required to extend the file. */
FileChunk *pNew = sqlite3_malloc(fileChunkSize(p->nChunkSize));
if( !pNew ){
return SQLITE_IOERR_NOMEM_BKPT;
}
pNew->pNext = 0;
if( pChunk ){
assert( p->pFirst );
pChunk->pNext = pNew;
}else{
assert( !p->pFirst );
p->pFirst = pNew;
}
p->endpoint.pChunk = pNew;
}
memcpy((u8*)p->endpoint.pChunk->zChunk + iChunkOffset, zWrite, iSpace);
zWrite += iSpace;
nWrite -= iSpace;
p->endpoint.iOffset += iSpace;
}
p->nSize = iAmt + iOfst;
}
}
return SQLITE_OK;
}
/*
** Truncate the file.
**
** If the journal file is already on disk, truncate it there. Or, if it
** is still in main memory but is being truncated to zero bytes in size,
** ignore
*/
static int memjrnlTruncate(sqlite3_file *pJfd, sqlite_int64 size){
MemJournal *p = (MemJournal *)pJfd;
if( ALWAYS(size==0) ){
memjrnlFreeChunks(p);
p->nSize = 0;
p->endpoint.pChunk = 0;
p->endpoint.iOffset = 0;
p->readpoint.pChunk = 0;
p->readpoint.iOffset = 0;
}
return SQLITE_OK;
}
/*
** Close the file.
*/
static int memjrnlClose(sqlite3_file *pJfd){
MemJournal *p = (MemJournal *)pJfd;
memjrnlFreeChunks(p);
return SQLITE_OK;
}
/*
** Sync the file.
**
** If the real file has been created, call its xSync method. Otherwise,
** syncing an in-memory journal is a no-op.
*/
static int memjrnlSync(sqlite3_file *pJfd, int flags){
UNUSED_PARAMETER2(pJfd, flags);
return SQLITE_OK;
}
/*
** Query the size of the file in bytes.
*/
static int memjrnlFileSize(sqlite3_file *pJfd, sqlite_int64 *pSize){
MemJournal *p = (MemJournal *)pJfd;
*pSize = (sqlite_int64) p->endpoint.iOffset;
return SQLITE_OK;
}
/*
** Table of methods for MemJournal sqlite3_file object.
*/
static const struct sqlite3_io_methods MemJournalMethods = {
1, /* iVersion */
memjrnlClose, /* xClose */
memjrnlRead, /* xRead */
memjrnlWrite, /* xWrite */
memjrnlTruncate, /* xTruncate */
memjrnlSync, /* xSync */
memjrnlFileSize, /* xFileSize */
0, /* xLock */
0, /* xUnlock */
0, /* xCheckReservedLock */
0, /* xFileControl */
0, /* xSectorSize */
0, /* xDeviceCharacteristics */
0, /* xShmMap */
0, /* xShmLock */
0, /* xShmBarrier */
0, /* xShmUnmap */
0, /* xFetch */
0 /* xUnfetch */
};
/*
** Open a journal file.
**
** The behaviour of the journal file depends on the value of parameter
** nSpill. If nSpill is 0, then the journal file is always create and
** accessed using the underlying VFS. If nSpill is less than zero, then
** all content is always stored in main-memory. Finally, if nSpill is a
** positive value, then the journal file is initially created in-memory
** but may be flushed to disk later on. In this case the journal file is
** flushed to disk either when it grows larger than nSpill bytes in size,
** or when sqlite3JournalCreate() is called.
*/
int sqlite3JournalOpen(
sqlite3_vfs *pVfs, /* The VFS to use for actual file I/O */
const char *zName, /* Name of the journal file */
sqlite3_file *pJfd, /* Preallocated, blank file handle */
int flags, /* Opening flags */
int nSpill /* Bytes buffered before opening the file */
){
MemJournal *p = (MemJournal*)pJfd;
/* Zero the file-handle object. If nSpill was passed zero, initialize
** it using the sqlite3OsOpen() function of the underlying VFS. In this
** case none of the code in this module is executed as a result of calls
** made on the journal file-handle. */
memset(p, 0, sizeof(MemJournal));
if( nSpill==0 ){
return sqlite3OsOpen(pVfs, zName, pJfd, flags, 0);
}
if( nSpill>0 ){
p->nChunkSize = nSpill;
}else{
p->nChunkSize = 8 + MEMJOURNAL_DFLT_FILECHUNKSIZE - sizeof(FileChunk);
assert( MEMJOURNAL_DFLT_FILECHUNKSIZE==fileChunkSize(p->nChunkSize) );
}
p->pMethod = (const sqlite3_io_methods*)&MemJournalMethods;
p->nSpill = nSpill;
p->flags = flags;
p->zJournal = zName;
p->pVfs = pVfs;
return SQLITE_OK;
}
/*
** Open an in-memory journal file.
*/
void sqlite3MemJournalOpen(sqlite3_file *pJfd){
sqlite3JournalOpen(0, 0, pJfd, 0, -1);
}
#if defined(SQLITE_ENABLE_ATOMIC_WRITE) \
|| defined(SQLITE_ENABLE_BATCH_ATOMIC_WRITE)
/*
** If the argument p points to a MemJournal structure that is not an
** in-memory-only journal file (i.e. is one that was opened with a +ve
** nSpill parameter or as SQLITE_OPEN_MAIN_JOURNAL), and the underlying
** file has not yet been created, create it now.
*/
int sqlite3JournalCreate(sqlite3_file *pJfd){
int rc = SQLITE_OK;
MemJournal *p = (MemJournal*)pJfd;
if( p->pMethod==&MemJournalMethods && (
#ifdef SQLITE_ENABLE_ATOMIC_WRITE
p->nSpill>0
#else
/* While this appears to not be possible without ATOMIC_WRITE, the
** paths are complex, so it seems prudent to leave the test in as
** a NEVER(), in case our analysis is subtly flawed. */
NEVER(p->nSpill>0)
#endif
#ifdef SQLITE_ENABLE_BATCH_ATOMIC_WRITE
|| (p->flags & SQLITE_OPEN_MAIN_JOURNAL)
#endif
)){
rc = memjrnlCreateFile(p);
}
return rc;
}
#endif
/*
** The file-handle passed as the only argument is open on a journal file.
** Return true if this "journal file" is currently stored in heap memory,
** or false otherwise.
*/
int sqlite3JournalIsInMemory(sqlite3_file *p){
return p->pMethods==&MemJournalMethods;
}
/*
** Return the number of bytes required to store a JournalFile that uses vfs
** pVfs to create the underlying on-disk files.
*/
int sqlite3JournalSize(sqlite3_vfs *pVfs){
return MAX(pVfs->szOsFile, (int)sizeof(MemJournal));
}