mirror of
https://github.com/sqlite/sqlite.git
synced 2024-11-21 11:19:14 +01:00
064b681e9b
FossilOrigin-Name: 6e5bb48a74d63fb8c30528f0005d1763cd2dbb882abf86baf1565721e6bfcf84
1370 lines
45 KiB
C
1370 lines
45 KiB
C
/*
|
|
** 2010 October 28
|
|
**
|
|
** 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 a VFS "shim" - a layer that sits in between the
|
|
** pager and the real VFS - that breaks up a very large database file
|
|
** into two or more smaller files on disk. This is useful, for example,
|
|
** in order to support large, multi-gigabyte databases on older filesystems
|
|
** that limit the maximum file size to 2 GiB.
|
|
**
|
|
** USAGE:
|
|
**
|
|
** Compile this source file and link it with your application. Then
|
|
** at start-time, invoke the following procedure:
|
|
**
|
|
** int sqlite3_multiplex_initialize(
|
|
** const char *zOrigVfsName, // The underlying real VFS
|
|
** int makeDefault // True to make multiplex the default VFS
|
|
** );
|
|
**
|
|
** The procedure call above will create and register a new VFS shim named
|
|
** "multiplex". The multiplex VFS will use the VFS named by zOrigVfsName to
|
|
** do the actual disk I/O. (The zOrigVfsName parameter may be NULL, in
|
|
** which case the default VFS at the moment sqlite3_multiplex_initialize()
|
|
** is called will be used as the underlying real VFS.)
|
|
**
|
|
** If the makeDefault parameter is TRUE then multiplex becomes the new
|
|
** default VFS. Otherwise, you can use the multiplex VFS by specifying
|
|
** "multiplex" as the 4th parameter to sqlite3_open_v2() or by employing
|
|
** URI filenames and adding "vfs=multiplex" as a parameter to the filename
|
|
** URI.
|
|
**
|
|
** The multiplex VFS allows databases up to 32 GiB in size. But it splits
|
|
** the files up into smaller pieces, so that they will work even on
|
|
** filesystems that do not support large files. The default chunk size
|
|
** is 2147418112 bytes (which is 64KiB less than 2GiB) but this can be
|
|
** changed at compile-time by defining the SQLITE_MULTIPLEX_CHUNK_SIZE
|
|
** macro. Use the "chunksize=NNNN" query parameter with a URI filename
|
|
** in order to select an alternative chunk size for individual connections
|
|
** at run-time.
|
|
*/
|
|
#include "sqlite3.h"
|
|
#include <string.h>
|
|
#include <assert.h>
|
|
#include <stdlib.h>
|
|
#include "test_multiplex.h"
|
|
|
|
#ifndef SQLITE_CORE
|
|
#define SQLITE_CORE 1 /* Disable the API redefinition in sqlite3ext.h */
|
|
#endif
|
|
#include "sqlite3ext.h"
|
|
|
|
/*
|
|
** These should be defined to be the same as the values in
|
|
** sqliteInt.h. They are defined separately here so that
|
|
** the multiplex VFS shim can be built as a loadable
|
|
** module.
|
|
*/
|
|
#define UNUSED_PARAMETER(x) (void)(x)
|
|
#define MAX_PAGE_SIZE 0x10000
|
|
#define DEFAULT_SECTOR_SIZE 0x1000
|
|
|
|
/* Maximum chunk number */
|
|
#define MX_CHUNK_NUMBER 299
|
|
|
|
/* First chunk for rollback journal files */
|
|
#define SQLITE_MULTIPLEX_JOURNAL_8_3_OFFSET 400
|
|
#define SQLITE_MULTIPLEX_WAL_8_3_OFFSET 700
|
|
|
|
|
|
/************************ Shim Definitions ******************************/
|
|
|
|
#ifndef SQLITE_MULTIPLEX_VFS_NAME
|
|
# define SQLITE_MULTIPLEX_VFS_NAME "multiplex"
|
|
#endif
|
|
|
|
/* This is the limit on the chunk size. It may be changed by calling
|
|
** the xFileControl() interface. It will be rounded up to a
|
|
** multiple of MAX_PAGE_SIZE. We default it here to 2GiB less 64KiB.
|
|
*/
|
|
#ifndef SQLITE_MULTIPLEX_CHUNK_SIZE
|
|
# define SQLITE_MULTIPLEX_CHUNK_SIZE 2147418112
|
|
#endif
|
|
|
|
/* This used to be the default limit on number of chunks, but
|
|
** it is no longer enforced. There is currently no limit to the
|
|
** number of chunks.
|
|
**
|
|
** May be changed by calling the xFileControl() interface.
|
|
*/
|
|
#ifndef SQLITE_MULTIPLEX_MAX_CHUNKS
|
|
# define SQLITE_MULTIPLEX_MAX_CHUNKS 12
|
|
#endif
|
|
|
|
/************************ Object Definitions ******************************/
|
|
|
|
/* Forward declaration of all object types */
|
|
typedef struct multiplexGroup multiplexGroup;
|
|
typedef struct multiplexConn multiplexConn;
|
|
|
|
/*
|
|
** A "multiplex group" is a collection of files that collectively
|
|
** makeup a single SQLite DB file. This allows the size of the DB
|
|
** to exceed the limits imposed by the file system.
|
|
**
|
|
** There is an instance of the following object for each defined multiplex
|
|
** group.
|
|
*/
|
|
struct multiplexGroup {
|
|
struct multiplexReal { /* For each chunk */
|
|
sqlite3_file *p; /* Handle for the chunk */
|
|
char *z; /* Name of this chunk */
|
|
} *aReal; /* list of all chunks */
|
|
int nReal; /* Number of chunks */
|
|
char *zName; /* Base filename of this group */
|
|
int nName; /* Length of base filename */
|
|
int flags; /* Flags used for original opening */
|
|
unsigned int szChunk; /* Chunk size used for this group */
|
|
unsigned char bEnabled; /* TRUE to use Multiplex VFS for this file */
|
|
unsigned char bTruncate; /* TRUE to enable truncation of databases */
|
|
};
|
|
|
|
/*
|
|
** An instance of the following object represents each open connection
|
|
** to a file that is multiplex'ed. This object is a
|
|
** subclass of sqlite3_file. The sqlite3_file object for the underlying
|
|
** VFS is appended to this structure.
|
|
*/
|
|
struct multiplexConn {
|
|
sqlite3_file base; /* Base class - must be first */
|
|
multiplexGroup *pGroup; /* The underlying group of files */
|
|
};
|
|
|
|
/************************* Global Variables **********************************/
|
|
/*
|
|
** All global variables used by this file are containing within the following
|
|
** gMultiplex structure.
|
|
*/
|
|
static struct {
|
|
/* The pOrigVfs is the real, original underlying VFS implementation.
|
|
** Most operations pass-through to the real VFS. This value is read-only
|
|
** during operation. It is only modified at start-time and thus does not
|
|
** require a mutex.
|
|
*/
|
|
sqlite3_vfs *pOrigVfs;
|
|
|
|
/* The sThisVfs is the VFS structure used by this shim. It is initialized
|
|
** at start-time and thus does not require a mutex
|
|
*/
|
|
sqlite3_vfs sThisVfs;
|
|
|
|
/* The sIoMethods defines the methods used by sqlite3_file objects
|
|
** associated with this shim. It is initialized at start-time and does
|
|
** not require a mutex.
|
|
**
|
|
** When the underlying VFS is called to open a file, it might return
|
|
** either a version 1 or a version 2 sqlite3_file object. This shim
|
|
** has to create a wrapper sqlite3_file of the same version. Hence
|
|
** there are two I/O method structures, one for version 1 and the other
|
|
** for version 2.
|
|
*/
|
|
sqlite3_io_methods sIoMethodsV1;
|
|
sqlite3_io_methods sIoMethodsV2;
|
|
|
|
/* True when this shim has been initialized.
|
|
*/
|
|
int isInitialized;
|
|
} gMultiplex;
|
|
|
|
/************************* Utility Routines *********************************/
|
|
/*
|
|
** Compute a string length that is limited to what can be stored in
|
|
** lower 30 bits of a 32-bit signed integer.
|
|
**
|
|
** The value returned will never be negative. Nor will it ever be greater
|
|
** than the actual length of the string. For very long strings (greater
|
|
** than 1GiB) the value returned might be less than the true string length.
|
|
*/
|
|
static int multiplexStrlen30(const char *z){
|
|
const char *z2 = z;
|
|
if( z==0 ) return 0;
|
|
while( *z2 ){ z2++; }
|
|
return 0x3fffffff & (int)(z2 - z);
|
|
}
|
|
|
|
/*
|
|
** Generate the file-name for chunk iChunk of the group with base name
|
|
** zBase. The file-name is written to buffer zOut before returning. Buffer
|
|
** zOut must be allocated by the caller so that it is at least (nBase+5)
|
|
** bytes in size, where nBase is the length of zBase, not including the
|
|
** nul-terminator.
|
|
**
|
|
** If iChunk is 0 (or 400 - the number for the first journal file chunk),
|
|
** the output is a copy of the input string. Otherwise, if
|
|
** SQLITE_ENABLE_8_3_NAMES is not defined or the input buffer does not contain
|
|
** a "." character, then the output is a copy of the input string with the
|
|
** three-digit zero-padded decimal representation if iChunk appended to it.
|
|
** For example:
|
|
**
|
|
** zBase="test.db", iChunk=4 -> zOut="test.db004"
|
|
**
|
|
** Or, if SQLITE_ENABLE_8_3_NAMES is defined and the input buffer contains
|
|
** a "." character, then everything after the "." is replaced by the
|
|
** three-digit representation of iChunk.
|
|
**
|
|
** zBase="test.db", iChunk=4 -> zOut="test.004"
|
|
**
|
|
** The output buffer string is terminated by 2 0x00 bytes. This makes it safe
|
|
** to pass to sqlite3_uri_parameter() and similar.
|
|
*/
|
|
static void multiplexFilename(
|
|
const char *zBase, /* Filename for chunk 0 */
|
|
int nBase, /* Size of zBase in bytes (without \0) */
|
|
int flags, /* Flags used to open file */
|
|
int iChunk, /* Chunk to generate filename for */
|
|
char *zOut /* Buffer to write generated name to */
|
|
){
|
|
int n = nBase;
|
|
memcpy(zOut, zBase, n+1);
|
|
if( iChunk!=0 && iChunk<=MX_CHUNK_NUMBER ){
|
|
#ifdef SQLITE_ENABLE_8_3_NAMES
|
|
int i;
|
|
for(i=n-1; i>0 && i>=n-4 && zOut[i]!='.'; i--){}
|
|
if( i>=n-4 ) n = i+1;
|
|
if( flags & SQLITE_OPEN_MAIN_JOURNAL ){
|
|
/* The extensions on overflow files for main databases are 001, 002,
|
|
** 003 and so forth. To avoid name collisions, add 400 to the
|
|
** extensions of journal files so that they are 401, 402, 403, ....
|
|
*/
|
|
iChunk += SQLITE_MULTIPLEX_JOURNAL_8_3_OFFSET;
|
|
}else if( flags & SQLITE_OPEN_WAL ){
|
|
/* To avoid name collisions, add 700 to the
|
|
** extensions of WAL files so that they are 701, 702, 703, ....
|
|
*/
|
|
iChunk += SQLITE_MULTIPLEX_WAL_8_3_OFFSET;
|
|
}
|
|
#endif
|
|
sqlite3_snprintf(4,&zOut[n],"%03d",iChunk);
|
|
n += 3;
|
|
}
|
|
|
|
assert( zOut[n]=='\0' );
|
|
zOut[n+1] = '\0';
|
|
}
|
|
|
|
/* Compute the filename for the iChunk-th chunk
|
|
*/
|
|
static int multiplexSubFilename(multiplexGroup *pGroup, int iChunk){
|
|
if( iChunk>=pGroup->nReal ){
|
|
struct multiplexReal *p;
|
|
p = sqlite3_realloc64(pGroup->aReal, (iChunk+1)*sizeof(*p));
|
|
if( p==0 ){
|
|
return SQLITE_NOMEM;
|
|
}
|
|
memset(&p[pGroup->nReal], 0, sizeof(p[0])*(iChunk+1-pGroup->nReal));
|
|
pGroup->aReal = p;
|
|
pGroup->nReal = iChunk+1;
|
|
}
|
|
if( pGroup->zName && pGroup->aReal[iChunk].z==0 ){
|
|
char *z;
|
|
int n = pGroup->nName;
|
|
z = sqlite3_malloc64( n+5 );
|
|
if( z==0 ){
|
|
return SQLITE_NOMEM;
|
|
}
|
|
multiplexFilename(pGroup->zName, pGroup->nName, pGroup->flags, iChunk, z);
|
|
pGroup->aReal[iChunk].z = (char*)sqlite3_create_filename(z,"","",0,0);
|
|
sqlite3_free(z);
|
|
if( pGroup->aReal[iChunk].z==0 ) return SQLITE_NOMEM;
|
|
}
|
|
return SQLITE_OK;
|
|
}
|
|
|
|
/* Translate an sqlite3_file* that is really a multiplexGroup* into
|
|
** the sqlite3_file* for the underlying original VFS.
|
|
**
|
|
** For chunk 0, the pGroup->flags determines whether or not a new file
|
|
** is created if it does not already exist. For chunks 1 and higher, the
|
|
** file is created only if createFlag is 1.
|
|
*/
|
|
static sqlite3_file *multiplexSubOpen(
|
|
multiplexGroup *pGroup, /* The multiplexor group */
|
|
int iChunk, /* Which chunk to open. 0==original file */
|
|
int *rc, /* Result code in and out */
|
|
int *pOutFlags, /* Output flags */
|
|
int createFlag /* True to create if iChunk>0 */
|
|
){
|
|
sqlite3_file *pSubOpen = 0;
|
|
sqlite3_vfs *pOrigVfs = gMultiplex.pOrigVfs; /* Real VFS */
|
|
|
|
#ifdef SQLITE_ENABLE_8_3_NAMES
|
|
/* If JOURNAL_8_3_OFFSET is set to (say) 400, then any overflow files are
|
|
** part of a database journal are named db.401, db.402, and so on. A
|
|
** database may therefore not grow to larger than 400 chunks. Attempting
|
|
** to open chunk 401 indicates the database is full. */
|
|
if( iChunk>=SQLITE_MULTIPLEX_JOURNAL_8_3_OFFSET ){
|
|
sqlite3_log(SQLITE_FULL, "multiplexed chunk overflow: %s", pGroup->zName);
|
|
*rc = SQLITE_FULL;
|
|
return 0;
|
|
}
|
|
#endif
|
|
|
|
*rc = multiplexSubFilename(pGroup, iChunk);
|
|
if( (*rc)==SQLITE_OK && (pSubOpen = pGroup->aReal[iChunk].p)==0 ){
|
|
int flags, bExists;
|
|
flags = pGroup->flags;
|
|
if( createFlag ){
|
|
flags |= SQLITE_OPEN_CREATE;
|
|
}else if( iChunk==0 ){
|
|
/* Fall through */
|
|
}else if( pGroup->aReal[iChunk].z==0 ){
|
|
return 0;
|
|
}else{
|
|
*rc = pOrigVfs->xAccess(pOrigVfs, pGroup->aReal[iChunk].z,
|
|
SQLITE_ACCESS_EXISTS, &bExists);
|
|
if( *rc || !bExists ){
|
|
if( *rc ){
|
|
sqlite3_log(*rc, "multiplexor.xAccess failure on %s",
|
|
pGroup->aReal[iChunk].z);
|
|
}
|
|
return 0;
|
|
}
|
|
flags &= ~SQLITE_OPEN_CREATE;
|
|
}
|
|
pSubOpen = sqlite3_malloc64( pOrigVfs->szOsFile );
|
|
if( pSubOpen==0 ){
|
|
*rc = SQLITE_IOERR_NOMEM;
|
|
return 0;
|
|
}
|
|
pGroup->aReal[iChunk].p = pSubOpen;
|
|
*rc = pOrigVfs->xOpen(pOrigVfs, pGroup->aReal[iChunk].z, pSubOpen,
|
|
flags, pOutFlags);
|
|
if( (*rc)!=SQLITE_OK ){
|
|
sqlite3_log(*rc, "multiplexor.xOpen failure on %s",
|
|
pGroup->aReal[iChunk].z);
|
|
sqlite3_free(pSubOpen);
|
|
pGroup->aReal[iChunk].p = 0;
|
|
return 0;
|
|
}
|
|
}
|
|
return pSubOpen;
|
|
}
|
|
|
|
/*
|
|
** Return the size, in bytes, of chunk number iChunk. If that chunk
|
|
** does not exist, then return 0. This function does not distinguish between
|
|
** non-existent files and zero-length files.
|
|
*/
|
|
static sqlite3_int64 multiplexSubSize(
|
|
multiplexGroup *pGroup, /* The multiplexor group */
|
|
int iChunk, /* Which chunk to open. 0==original file */
|
|
int *rc /* Result code in and out */
|
|
){
|
|
sqlite3_file *pSub;
|
|
sqlite3_int64 sz = 0;
|
|
|
|
if( *rc ) return 0;
|
|
pSub = multiplexSubOpen(pGroup, iChunk, rc, NULL, 0);
|
|
if( pSub==0 ) return 0;
|
|
*rc = pSub->pMethods->xFileSize(pSub, &sz);
|
|
return sz;
|
|
}
|
|
|
|
/*
|
|
** This is the implementation of the multiplex_control() SQL function.
|
|
*/
|
|
static void multiplexControlFunc(
|
|
sqlite3_context *context,
|
|
int argc,
|
|
sqlite3_value **argv
|
|
){
|
|
int rc = SQLITE_OK;
|
|
sqlite3 *db = sqlite3_context_db_handle(context);
|
|
int op = 0;
|
|
int iVal;
|
|
|
|
if( !db || argc!=2 ){
|
|
rc = SQLITE_ERROR;
|
|
}else{
|
|
/* extract params */
|
|
op = sqlite3_value_int(argv[0]);
|
|
iVal = sqlite3_value_int(argv[1]);
|
|
/* map function op to file_control op */
|
|
switch( op ){
|
|
case 1:
|
|
op = MULTIPLEX_CTRL_ENABLE;
|
|
break;
|
|
case 2:
|
|
op = MULTIPLEX_CTRL_SET_CHUNK_SIZE;
|
|
break;
|
|
case 3:
|
|
op = MULTIPLEX_CTRL_SET_MAX_CHUNKS;
|
|
break;
|
|
default:
|
|
rc = SQLITE_NOTFOUND;
|
|
break;
|
|
}
|
|
}
|
|
if( rc==SQLITE_OK ){
|
|
rc = sqlite3_file_control(db, 0, op, &iVal);
|
|
}
|
|
sqlite3_result_error_code(context, rc);
|
|
}
|
|
|
|
/*
|
|
** This is the entry point to register the auto-extension for the
|
|
** multiplex_control() function.
|
|
*/
|
|
static int multiplexFuncInit(
|
|
sqlite3 *db,
|
|
char **pzErrMsg,
|
|
const sqlite3_api_routines *pApi
|
|
){
|
|
int rc;
|
|
rc = sqlite3_create_function(db, "multiplex_control", 2, SQLITE_ANY,
|
|
0, multiplexControlFunc, 0, 0);
|
|
return rc;
|
|
}
|
|
|
|
/*
|
|
** Close a single sub-file in the connection group.
|
|
*/
|
|
static void multiplexSubClose(
|
|
multiplexGroup *pGroup,
|
|
int iChunk,
|
|
sqlite3_vfs *pOrigVfs
|
|
){
|
|
sqlite3_file *pSubOpen = pGroup->aReal[iChunk].p;
|
|
if( pSubOpen ){
|
|
pSubOpen->pMethods->xClose(pSubOpen);
|
|
if( pOrigVfs && pGroup->aReal[iChunk].z ){
|
|
pOrigVfs->xDelete(pOrigVfs, pGroup->aReal[iChunk].z, 0);
|
|
}
|
|
sqlite3_free(pGroup->aReal[iChunk].p);
|
|
}
|
|
sqlite3_free_filename(pGroup->aReal[iChunk].z);
|
|
memset(&pGroup->aReal[iChunk], 0, sizeof(pGroup->aReal[iChunk]));
|
|
}
|
|
|
|
/*
|
|
** Deallocate memory held by a multiplexGroup
|
|
*/
|
|
static void multiplexFreeComponents(multiplexGroup *pGroup){
|
|
int i;
|
|
for(i=0; i<pGroup->nReal; i++){ multiplexSubClose(pGroup, i, 0); }
|
|
sqlite3_free(pGroup->aReal);
|
|
pGroup->aReal = 0;
|
|
pGroup->nReal = 0;
|
|
}
|
|
|
|
|
|
/************************* VFS Method Wrappers *****************************/
|
|
|
|
/*
|
|
** This is the xOpen method used for the "multiplex" VFS.
|
|
**
|
|
** Most of the work is done by the underlying original VFS. This method
|
|
** simply links the new file into the appropriate multiplex group if it is a
|
|
** file that needs to be tracked.
|
|
*/
|
|
static int multiplexOpen(
|
|
sqlite3_vfs *pVfs, /* The multiplex VFS */
|
|
const char *zName, /* Name of file to be opened */
|
|
sqlite3_file *pConn, /* Fill in this file descriptor */
|
|
int flags, /* Flags to control the opening */
|
|
int *pOutFlags /* Flags showing results of opening */
|
|
){
|
|
int rc = SQLITE_OK; /* Result code */
|
|
multiplexConn *pMultiplexOpen; /* The new multiplex file descriptor */
|
|
multiplexGroup *pGroup = 0; /* Corresponding multiplexGroup object */
|
|
sqlite3_file *pSubOpen = 0; /* Real file descriptor */
|
|
sqlite3_vfs *pOrigVfs = gMultiplex.pOrigVfs; /* Real VFS */
|
|
int nName = 0;
|
|
int sz = 0;
|
|
char *zToFree = 0;
|
|
|
|
UNUSED_PARAMETER(pVfs);
|
|
memset(pConn, 0, pVfs->szOsFile);
|
|
assert( zName || (flags & SQLITE_OPEN_DELETEONCLOSE) );
|
|
|
|
/* We need to create a group structure and manage
|
|
** access to this group of files.
|
|
*/
|
|
pMultiplexOpen = (multiplexConn*)pConn;
|
|
|
|
if( rc==SQLITE_OK ){
|
|
/* allocate space for group */
|
|
nName = zName ? multiplexStrlen30(zName) : 0;
|
|
sz = sizeof(multiplexGroup) /* multiplexGroup */
|
|
+ nName + 1; /* zName */
|
|
pGroup = sqlite3_malloc64( sz );
|
|
if( pGroup==0 ){
|
|
rc = SQLITE_NOMEM;
|
|
}
|
|
}
|
|
|
|
if( rc==SQLITE_OK ){
|
|
const char *zUri = (flags & SQLITE_OPEN_URI) ? zName : 0;
|
|
/* assign pointers to extra space allocated */
|
|
memset(pGroup, 0, sz);
|
|
pMultiplexOpen->pGroup = pGroup;
|
|
pGroup->bEnabled = (unsigned char)-1;
|
|
pGroup->bTruncate = (unsigned char)sqlite3_uri_boolean(zUri, "truncate",
|
|
(flags & SQLITE_OPEN_MAIN_DB)==0);
|
|
pGroup->szChunk = (int)sqlite3_uri_int64(zUri, "chunksize",
|
|
SQLITE_MULTIPLEX_CHUNK_SIZE);
|
|
pGroup->szChunk = (pGroup->szChunk+0xffff)&~0xffff;
|
|
if( zName ){
|
|
char *p = (char *)&pGroup[1];
|
|
pGroup->zName = p;
|
|
memcpy(pGroup->zName, zName, nName+1);
|
|
pGroup->nName = nName;
|
|
}
|
|
if( pGroup->bEnabled ){
|
|
/* Make sure that the chunksize is such that the pending byte does not
|
|
** falls at the end of a chunk. A region of up to 64K following
|
|
** the pending byte is never written, so if the pending byte occurs
|
|
** near the end of a chunk, that chunk will be too small. */
|
|
#ifndef SQLITE_OMIT_WSD
|
|
extern int sqlite3PendingByte;
|
|
#else
|
|
int sqlite3PendingByte = 0x40000000;
|
|
#endif
|
|
while( (sqlite3PendingByte % pGroup->szChunk)>=(pGroup->szChunk-65536) ){
|
|
pGroup->szChunk += 65536;
|
|
}
|
|
}
|
|
pGroup->flags = (flags & ~SQLITE_OPEN_URI);
|
|
rc = multiplexSubFilename(pGroup, 1);
|
|
if( rc==SQLITE_OK ){
|
|
pSubOpen = multiplexSubOpen(pGroup, 0, &rc, pOutFlags, 0);
|
|
if( pSubOpen==0 && rc==SQLITE_OK ) rc = SQLITE_CANTOPEN;
|
|
}
|
|
if( rc==SQLITE_OK ){
|
|
sqlite3_int64 sz64;
|
|
|
|
rc = pSubOpen->pMethods->xFileSize(pSubOpen, &sz64);
|
|
if( rc==SQLITE_OK && zName ){
|
|
int bExists;
|
|
if( flags & SQLITE_OPEN_SUPER_JOURNAL ){
|
|
pGroup->bEnabled = 0;
|
|
}else
|
|
if( sz64==0 ){
|
|
if( flags & SQLITE_OPEN_MAIN_JOURNAL ){
|
|
/* If opening a main journal file and the first chunk is zero
|
|
** bytes in size, delete any subsequent chunks from the
|
|
** file-system. */
|
|
int iChunk = 1;
|
|
do {
|
|
rc = pOrigVfs->xAccess(pOrigVfs,
|
|
pGroup->aReal[iChunk].z, SQLITE_ACCESS_EXISTS, &bExists
|
|
);
|
|
if( rc==SQLITE_OK && bExists ){
|
|
rc = pOrigVfs->xDelete(pOrigVfs, pGroup->aReal[iChunk].z, 0);
|
|
if( rc==SQLITE_OK ){
|
|
rc = multiplexSubFilename(pGroup, ++iChunk);
|
|
}
|
|
}
|
|
}while( rc==SQLITE_OK && bExists );
|
|
}
|
|
}else{
|
|
/* If the first overflow file exists and if the size of the main file
|
|
** is different from the chunk size, that means the chunk size is set
|
|
** set incorrectly. So fix it.
|
|
**
|
|
** Or, if the first overflow file does not exist and the main file is
|
|
** larger than the chunk size, that means the chunk size is too small.
|
|
** But we have no way of determining the intended chunk size, so
|
|
** just disable the multiplexor all together.
|
|
*/
|
|
rc = pOrigVfs->xAccess(pOrigVfs, pGroup->aReal[1].z,
|
|
SQLITE_ACCESS_EXISTS, &bExists);
|
|
bExists = multiplexSubSize(pGroup, 1, &rc)>0;
|
|
if( rc==SQLITE_OK && bExists && sz64==(sz64&0xffff0000) && sz64>0
|
|
&& sz64!=pGroup->szChunk ){
|
|
pGroup->szChunk = (int)sz64;
|
|
}else if( rc==SQLITE_OK && !bExists && sz64>pGroup->szChunk ){
|
|
pGroup->bEnabled = 0;
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
if( rc==SQLITE_OK ){
|
|
if( pSubOpen->pMethods->iVersion==1 ){
|
|
pConn->pMethods = &gMultiplex.sIoMethodsV1;
|
|
}else{
|
|
pConn->pMethods = &gMultiplex.sIoMethodsV2;
|
|
}
|
|
}else{
|
|
multiplexFreeComponents(pGroup);
|
|
sqlite3_free(pGroup);
|
|
}
|
|
}
|
|
sqlite3_free(zToFree);
|
|
return rc;
|
|
}
|
|
|
|
/*
|
|
** This is the xDelete method used for the "multiplex" VFS.
|
|
** It attempts to delete the filename specified.
|
|
*/
|
|
static int multiplexDelete(
|
|
sqlite3_vfs *pVfs, /* The multiplex VFS */
|
|
const char *zName, /* Name of file to delete */
|
|
int syncDir
|
|
){
|
|
int rc;
|
|
sqlite3_vfs *pOrigVfs = gMultiplex.pOrigVfs; /* Real VFS */
|
|
rc = pOrigVfs->xDelete(pOrigVfs, zName, syncDir);
|
|
if( rc==SQLITE_OK ){
|
|
/* If the main chunk was deleted successfully, also delete any subsequent
|
|
** chunks - starting with the last (highest numbered).
|
|
*/
|
|
int nName = (int)strlen(zName);
|
|
char *z;
|
|
z = sqlite3_malloc64(nName + 5);
|
|
if( z==0 ){
|
|
rc = SQLITE_IOERR_NOMEM;
|
|
}else{
|
|
int iChunk = 0;
|
|
int bExists;
|
|
do{
|
|
multiplexFilename(zName, nName, SQLITE_OPEN_MAIN_JOURNAL, ++iChunk, z);
|
|
rc = pOrigVfs->xAccess(pOrigVfs, z, SQLITE_ACCESS_EXISTS, &bExists);
|
|
}while( rc==SQLITE_OK && bExists );
|
|
while( rc==SQLITE_OK && iChunk>1 ){
|
|
multiplexFilename(zName, nName, SQLITE_OPEN_MAIN_JOURNAL, --iChunk, z);
|
|
rc = pOrigVfs->xDelete(pOrigVfs, z, syncDir);
|
|
}
|
|
if( rc==SQLITE_OK ){
|
|
iChunk = 0;
|
|
do{
|
|
multiplexFilename(zName, nName, SQLITE_OPEN_WAL, ++iChunk, z);
|
|
rc = pOrigVfs->xAccess(pOrigVfs, z, SQLITE_ACCESS_EXISTS, &bExists);
|
|
}while( rc==SQLITE_OK && bExists );
|
|
while( rc==SQLITE_OK && iChunk>1 ){
|
|
multiplexFilename(zName, nName, SQLITE_OPEN_WAL, --iChunk, z);
|
|
rc = pOrigVfs->xDelete(pOrigVfs, z, syncDir);
|
|
}
|
|
}
|
|
}
|
|
sqlite3_free(z);
|
|
}
|
|
return rc;
|
|
}
|
|
|
|
static int multiplexAccess(sqlite3_vfs *a, const char *b, int c, int *d){
|
|
return gMultiplex.pOrigVfs->xAccess(gMultiplex.pOrigVfs, b, c, d);
|
|
}
|
|
static int multiplexFullPathname(sqlite3_vfs *a, const char *b, int c, char *d){
|
|
return gMultiplex.pOrigVfs->xFullPathname(gMultiplex.pOrigVfs, b, c, d);
|
|
}
|
|
static void *multiplexDlOpen(sqlite3_vfs *a, const char *b){
|
|
return gMultiplex.pOrigVfs->xDlOpen(gMultiplex.pOrigVfs, b);
|
|
}
|
|
static void multiplexDlError(sqlite3_vfs *a, int b, char *c){
|
|
gMultiplex.pOrigVfs->xDlError(gMultiplex.pOrigVfs, b, c);
|
|
}
|
|
static void (*multiplexDlSym(sqlite3_vfs *a, void *b, const char *c))(void){
|
|
return gMultiplex.pOrigVfs->xDlSym(gMultiplex.pOrigVfs, b, c);
|
|
}
|
|
static void multiplexDlClose(sqlite3_vfs *a, void *b){
|
|
gMultiplex.pOrigVfs->xDlClose(gMultiplex.pOrigVfs, b);
|
|
}
|
|
static int multiplexRandomness(sqlite3_vfs *a, int b, char *c){
|
|
return gMultiplex.pOrigVfs->xRandomness(gMultiplex.pOrigVfs, b, c);
|
|
}
|
|
static int multiplexSleep(sqlite3_vfs *a, int b){
|
|
return gMultiplex.pOrigVfs->xSleep(gMultiplex.pOrigVfs, b);
|
|
}
|
|
static int multiplexCurrentTime(sqlite3_vfs *a, double *b){
|
|
return gMultiplex.pOrigVfs->xCurrentTime(gMultiplex.pOrigVfs, b);
|
|
}
|
|
static int multiplexGetLastError(sqlite3_vfs *a, int b, char *c){
|
|
if( gMultiplex.pOrigVfs->xGetLastError ){
|
|
return gMultiplex.pOrigVfs->xGetLastError(gMultiplex.pOrigVfs, b, c);
|
|
}else{
|
|
return 0;
|
|
}
|
|
}
|
|
static int multiplexCurrentTimeInt64(sqlite3_vfs *a, sqlite3_int64 *b){
|
|
return gMultiplex.pOrigVfs->xCurrentTimeInt64(gMultiplex.pOrigVfs, b);
|
|
}
|
|
|
|
/************************ I/O Method Wrappers *******************************/
|
|
|
|
/* xClose requests get passed through to the original VFS.
|
|
** We loop over all open chunk handles and close them.
|
|
** The group structure for this file is unlinked from
|
|
** our list of groups and freed.
|
|
*/
|
|
static int multiplexClose(sqlite3_file *pConn){
|
|
multiplexConn *p = (multiplexConn*)pConn;
|
|
multiplexGroup *pGroup = p->pGroup;
|
|
int rc = SQLITE_OK;
|
|
multiplexFreeComponents(pGroup);
|
|
sqlite3_free(pGroup);
|
|
return rc;
|
|
}
|
|
|
|
/* Pass xRead requests thru to the original VFS after
|
|
** determining the correct chunk to operate on.
|
|
** Break up reads across chunk boundaries.
|
|
*/
|
|
static int multiplexRead(
|
|
sqlite3_file *pConn,
|
|
void *pBuf,
|
|
int iAmt,
|
|
sqlite3_int64 iOfst
|
|
){
|
|
multiplexConn *p = (multiplexConn*)pConn;
|
|
multiplexGroup *pGroup = p->pGroup;
|
|
int rc = SQLITE_OK;
|
|
if( !pGroup->bEnabled ){
|
|
sqlite3_file *pSubOpen = multiplexSubOpen(pGroup, 0, &rc, NULL, 0);
|
|
if( pSubOpen==0 ){
|
|
rc = SQLITE_IOERR_READ;
|
|
}else{
|
|
rc = pSubOpen->pMethods->xRead(pSubOpen, pBuf, iAmt, iOfst);
|
|
}
|
|
}else{
|
|
while( iAmt > 0 ){
|
|
int i = (int)(iOfst / pGroup->szChunk);
|
|
sqlite3_file *pSubOpen;
|
|
pSubOpen = multiplexSubOpen(pGroup, i, &rc, NULL, 1);
|
|
if( pSubOpen ){
|
|
int extra = ((int)(iOfst % pGroup->szChunk) + iAmt) - pGroup->szChunk;
|
|
if( extra<0 ) extra = 0;
|
|
iAmt -= extra;
|
|
rc = pSubOpen->pMethods->xRead(pSubOpen, pBuf, iAmt,
|
|
iOfst % pGroup->szChunk);
|
|
if( rc!=SQLITE_OK ) break;
|
|
pBuf = (char *)pBuf + iAmt;
|
|
iOfst += iAmt;
|
|
iAmt = extra;
|
|
}else{
|
|
rc = SQLITE_IOERR_READ;
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
|
|
return rc;
|
|
}
|
|
|
|
/* Pass xWrite requests thru to the original VFS after
|
|
** determining the correct chunk to operate on.
|
|
** Break up writes across chunk boundaries.
|
|
*/
|
|
static int multiplexWrite(
|
|
sqlite3_file *pConn,
|
|
const void *pBuf,
|
|
int iAmt,
|
|
sqlite3_int64 iOfst
|
|
){
|
|
multiplexConn *p = (multiplexConn*)pConn;
|
|
multiplexGroup *pGroup = p->pGroup;
|
|
int rc = SQLITE_OK;
|
|
if( !pGroup->bEnabled ){
|
|
sqlite3_file *pSubOpen = multiplexSubOpen(pGroup, 0, &rc, NULL, 0);
|
|
if( pSubOpen==0 ){
|
|
rc = SQLITE_IOERR_WRITE;
|
|
}else{
|
|
rc = pSubOpen->pMethods->xWrite(pSubOpen, pBuf, iAmt, iOfst);
|
|
}
|
|
}else{
|
|
while( rc==SQLITE_OK && iAmt>0 ){
|
|
int i = (int)(iOfst / pGroup->szChunk);
|
|
sqlite3_file *pSubOpen = multiplexSubOpen(pGroup, i, &rc, NULL, 1);
|
|
if( pSubOpen ){
|
|
int extra = ((int)(iOfst % pGroup->szChunk) + iAmt) -
|
|
pGroup->szChunk;
|
|
if( extra<0 ) extra = 0;
|
|
iAmt -= extra;
|
|
rc = pSubOpen->pMethods->xWrite(pSubOpen, pBuf, iAmt,
|
|
iOfst % pGroup->szChunk);
|
|
pBuf = (char *)pBuf + iAmt;
|
|
iOfst += iAmt;
|
|
iAmt = extra;
|
|
}
|
|
}
|
|
}
|
|
return rc;
|
|
}
|
|
|
|
/* Pass xTruncate requests thru to the original VFS after
|
|
** determining the correct chunk to operate on. Delete any
|
|
** chunks above the truncate mark.
|
|
*/
|
|
static int multiplexTruncate(sqlite3_file *pConn, sqlite3_int64 size){
|
|
multiplexConn *p = (multiplexConn*)pConn;
|
|
multiplexGroup *pGroup = p->pGroup;
|
|
int rc = SQLITE_OK;
|
|
if( !pGroup->bEnabled ){
|
|
sqlite3_file *pSubOpen = multiplexSubOpen(pGroup, 0, &rc, NULL, 0);
|
|
if( pSubOpen==0 ){
|
|
rc = SQLITE_IOERR_TRUNCATE;
|
|
}else{
|
|
rc = pSubOpen->pMethods->xTruncate(pSubOpen, size);
|
|
}
|
|
}else{
|
|
int i;
|
|
int iBaseGroup = (int)(size / pGroup->szChunk);
|
|
sqlite3_file *pSubOpen;
|
|
sqlite3_vfs *pOrigVfs = gMultiplex.pOrigVfs; /* Real VFS */
|
|
/* delete the chunks above the truncate limit */
|
|
for(i = pGroup->nReal-1; i>iBaseGroup && rc==SQLITE_OK; i--){
|
|
if( pGroup->bTruncate ){
|
|
multiplexSubClose(pGroup, i, pOrigVfs);
|
|
}else{
|
|
pSubOpen = multiplexSubOpen(pGroup, i, &rc, 0, 0);
|
|
if( pSubOpen ){
|
|
rc = pSubOpen->pMethods->xTruncate(pSubOpen, 0);
|
|
}
|
|
}
|
|
}
|
|
if( rc==SQLITE_OK ){
|
|
pSubOpen = multiplexSubOpen(pGroup, iBaseGroup, &rc, 0, 0);
|
|
if( pSubOpen ){
|
|
rc = pSubOpen->pMethods->xTruncate(pSubOpen, size % pGroup->szChunk);
|
|
}
|
|
}
|
|
if( rc ) rc = SQLITE_IOERR_TRUNCATE;
|
|
}
|
|
return rc;
|
|
}
|
|
|
|
/* Pass xSync requests through to the original VFS without change
|
|
*/
|
|
static int multiplexSync(sqlite3_file *pConn, int flags){
|
|
multiplexConn *p = (multiplexConn*)pConn;
|
|
multiplexGroup *pGroup = p->pGroup;
|
|
int rc = SQLITE_OK;
|
|
int i;
|
|
for(i=0; i<pGroup->nReal; i++){
|
|
sqlite3_file *pSubOpen = pGroup->aReal[i].p;
|
|
if( pSubOpen ){
|
|
int rc2 = pSubOpen->pMethods->xSync(pSubOpen, flags);
|
|
if( rc2!=SQLITE_OK ) rc = rc2;
|
|
}
|
|
}
|
|
return rc;
|
|
}
|
|
|
|
/* Pass xFileSize requests through to the original VFS.
|
|
** Aggregate the size of all the chunks before returning.
|
|
*/
|
|
static int multiplexFileSize(sqlite3_file *pConn, sqlite3_int64 *pSize){
|
|
multiplexConn *p = (multiplexConn*)pConn;
|
|
multiplexGroup *pGroup = p->pGroup;
|
|
int rc = SQLITE_OK;
|
|
int i;
|
|
if( !pGroup->bEnabled ){
|
|
sqlite3_file *pSubOpen = multiplexSubOpen(pGroup, 0, &rc, NULL, 0);
|
|
if( pSubOpen==0 ){
|
|
rc = SQLITE_IOERR_FSTAT;
|
|
}else{
|
|
rc = pSubOpen->pMethods->xFileSize(pSubOpen, pSize);
|
|
}
|
|
}else{
|
|
*pSize = 0;
|
|
for(i=0; rc==SQLITE_OK; i++){
|
|
sqlite3_int64 sz = multiplexSubSize(pGroup, i, &rc);
|
|
if( sz==0 ) break;
|
|
*pSize = i*(sqlite3_int64)pGroup->szChunk + sz;
|
|
}
|
|
}
|
|
return rc;
|
|
}
|
|
|
|
/* Pass xLock requests through to the original VFS unchanged.
|
|
*/
|
|
static int multiplexLock(sqlite3_file *pConn, int lock){
|
|
multiplexConn *p = (multiplexConn*)pConn;
|
|
int rc;
|
|
sqlite3_file *pSubOpen = multiplexSubOpen(p->pGroup, 0, &rc, NULL, 0);
|
|
if( pSubOpen ){
|
|
return pSubOpen->pMethods->xLock(pSubOpen, lock);
|
|
}
|
|
return SQLITE_BUSY;
|
|
}
|
|
|
|
/* Pass xUnlock requests through to the original VFS unchanged.
|
|
*/
|
|
static int multiplexUnlock(sqlite3_file *pConn, int lock){
|
|
multiplexConn *p = (multiplexConn*)pConn;
|
|
int rc;
|
|
sqlite3_file *pSubOpen = multiplexSubOpen(p->pGroup, 0, &rc, NULL, 0);
|
|
if( pSubOpen ){
|
|
return pSubOpen->pMethods->xUnlock(pSubOpen, lock);
|
|
}
|
|
return SQLITE_IOERR_UNLOCK;
|
|
}
|
|
|
|
/* Pass xCheckReservedLock requests through to the original VFS unchanged.
|
|
*/
|
|
static int multiplexCheckReservedLock(sqlite3_file *pConn, int *pResOut){
|
|
multiplexConn *p = (multiplexConn*)pConn;
|
|
int rc;
|
|
sqlite3_file *pSubOpen = multiplexSubOpen(p->pGroup, 0, &rc, NULL, 0);
|
|
if( pSubOpen ){
|
|
return pSubOpen->pMethods->xCheckReservedLock(pSubOpen, pResOut);
|
|
}
|
|
return SQLITE_IOERR_CHECKRESERVEDLOCK;
|
|
}
|
|
|
|
/* Pass xFileControl requests through to the original VFS unchanged,
|
|
** except for any MULTIPLEX_CTRL_* requests here.
|
|
*/
|
|
static int multiplexFileControl(sqlite3_file *pConn, int op, void *pArg){
|
|
multiplexConn *p = (multiplexConn*)pConn;
|
|
multiplexGroup *pGroup = p->pGroup;
|
|
int rc = SQLITE_ERROR;
|
|
sqlite3_file *pSubOpen;
|
|
|
|
if( !gMultiplex.isInitialized ) return SQLITE_MISUSE;
|
|
switch( op ){
|
|
case MULTIPLEX_CTRL_ENABLE:
|
|
if( pArg ) {
|
|
int bEnabled = *(int *)pArg;
|
|
pGroup->bEnabled = (unsigned char)bEnabled;
|
|
rc = SQLITE_OK;
|
|
}
|
|
break;
|
|
case MULTIPLEX_CTRL_SET_CHUNK_SIZE:
|
|
if( pArg ) {
|
|
unsigned int szChunk = *(unsigned*)pArg;
|
|
if( szChunk<1 ){
|
|
rc = SQLITE_MISUSE;
|
|
}else{
|
|
/* Round up to nearest multiple of MAX_PAGE_SIZE. */
|
|
szChunk = (szChunk + (MAX_PAGE_SIZE-1));
|
|
szChunk &= ~(MAX_PAGE_SIZE-1);
|
|
pGroup->szChunk = szChunk;
|
|
rc = SQLITE_OK;
|
|
}
|
|
}
|
|
break;
|
|
case MULTIPLEX_CTRL_SET_MAX_CHUNKS:
|
|
rc = SQLITE_OK;
|
|
break;
|
|
case SQLITE_FCNTL_SIZE_HINT:
|
|
case SQLITE_FCNTL_CHUNK_SIZE:
|
|
/* no-op these */
|
|
rc = SQLITE_OK;
|
|
break;
|
|
case SQLITE_FCNTL_PRAGMA: {
|
|
char **aFcntl = (char**)pArg;
|
|
/*
|
|
** EVIDENCE-OF: R-29875-31678 The argument to the SQLITE_FCNTL_PRAGMA
|
|
** file control is an array of pointers to strings (char**) in which the
|
|
** second element of the array is the name of the pragma and the third
|
|
** element is the argument to the pragma or NULL if the pragma has no
|
|
** argument.
|
|
*/
|
|
if( aFcntl[1] && sqlite3_strnicmp(aFcntl[1],"multiplex_",10)==0 ){
|
|
sqlite3_int64 sz = 0;
|
|
(void)multiplexFileSize(pConn, &sz);
|
|
/*
|
|
** PRAGMA multiplex_truncate=BOOLEAN;
|
|
** PRAGMA multiplex_truncate;
|
|
**
|
|
** Turn the multiplexor truncate feature on or off. Return either
|
|
** "on" or "off" to indicate the new setting. If the BOOLEAN argument
|
|
** is omitted, just return the current value for the truncate setting.
|
|
*/
|
|
if( sqlite3_stricmp(aFcntl[1],"multiplex_truncate")==0 ){
|
|
if( aFcntl[2] && aFcntl[2][0] ){
|
|
if( sqlite3_stricmp(aFcntl[2], "on")==0
|
|
|| sqlite3_stricmp(aFcntl[2], "1")==0 ){
|
|
pGroup->bTruncate = 1;
|
|
}else
|
|
if( sqlite3_stricmp(aFcntl[2], "off")==0
|
|
|| sqlite3_stricmp(aFcntl[2], "0")==0 ){
|
|
pGroup->bTruncate = 0;
|
|
}
|
|
}
|
|
/* EVIDENCE-OF: R-27806-26076 The handler for an SQLITE_FCNTL_PRAGMA
|
|
** file control can optionally make the first element of the char**
|
|
** argument point to a string obtained from sqlite3_mprintf() or the
|
|
** equivalent and that string will become the result of the pragma
|
|
** or the error message if the pragma fails.
|
|
*/
|
|
aFcntl[0] = sqlite3_mprintf(pGroup->bTruncate ? "on" : "off");
|
|
rc = SQLITE_OK;
|
|
break;
|
|
}
|
|
/*
|
|
** PRAGMA multiplex_enabled;
|
|
**
|
|
** Return 0 or 1 depending on whether the multiplexor is enabled or
|
|
** disabled, respectively.
|
|
*/
|
|
if( sqlite3_stricmp(aFcntl[1],"multiplex_enabled")==0 ){
|
|
aFcntl[0] = sqlite3_mprintf("%d", pGroup->bEnabled!=0);
|
|
rc = SQLITE_OK;
|
|
break;
|
|
}
|
|
/*
|
|
** PRAGMA multiplex_chunksize;
|
|
**
|
|
** Return the chunksize for the multiplexor, or no-op if the
|
|
** multiplexor is not active.
|
|
*/
|
|
if( sqlite3_stricmp(aFcntl[1],"multiplex_chunksize")==0
|
|
&& pGroup->bEnabled
|
|
){
|
|
aFcntl[0] = sqlite3_mprintf("%u", pGroup->szChunk);
|
|
rc = SQLITE_OK;
|
|
break;
|
|
}
|
|
/*
|
|
** PRAGMA multiplex_filecount;
|
|
**
|
|
** Return the number of disk files currently in use by the
|
|
** multiplexor. This should be the total database size size
|
|
** divided by the chunksize and rounded up.
|
|
*/
|
|
if( sqlite3_stricmp(aFcntl[1],"multiplex_filecount")==0 ){
|
|
int n = 0;
|
|
int ii;
|
|
for(ii=0; ii<pGroup->nReal; ii++){
|
|
if( pGroup->aReal[ii].p!=0 ) n++;
|
|
}
|
|
aFcntl[0] = sqlite3_mprintf("%d", n);
|
|
rc = SQLITE_OK;
|
|
break;
|
|
}
|
|
}
|
|
/* If the multiplexor does not handle the pragma, pass it through
|
|
** into the default case. */
|
|
}
|
|
default:
|
|
pSubOpen = multiplexSubOpen(pGroup, 0, &rc, NULL, 0);
|
|
if( pSubOpen ){
|
|
rc = pSubOpen->pMethods->xFileControl(pSubOpen, op, pArg);
|
|
if( op==SQLITE_FCNTL_VFSNAME && rc==SQLITE_OK ){
|
|
*(char**)pArg = sqlite3_mprintf("multiplex/%z", *(char**)pArg);
|
|
}
|
|
}
|
|
break;
|
|
}
|
|
return rc;
|
|
}
|
|
|
|
/* Pass xSectorSize requests through to the original VFS unchanged.
|
|
*/
|
|
static int multiplexSectorSize(sqlite3_file *pConn){
|
|
multiplexConn *p = (multiplexConn*)pConn;
|
|
int rc;
|
|
sqlite3_file *pSubOpen = multiplexSubOpen(p->pGroup, 0, &rc, NULL, 0);
|
|
if( pSubOpen && pSubOpen->pMethods->xSectorSize ){
|
|
return pSubOpen->pMethods->xSectorSize(pSubOpen);
|
|
}
|
|
return DEFAULT_SECTOR_SIZE;
|
|
}
|
|
|
|
/* Pass xDeviceCharacteristics requests through to the original VFS unchanged.
|
|
*/
|
|
static int multiplexDeviceCharacteristics(sqlite3_file *pConn){
|
|
multiplexConn *p = (multiplexConn*)pConn;
|
|
int rc;
|
|
sqlite3_file *pSubOpen = multiplexSubOpen(p->pGroup, 0, &rc, NULL, 0);
|
|
if( pSubOpen ){
|
|
return pSubOpen->pMethods->xDeviceCharacteristics(pSubOpen);
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
/* Pass xShmMap requests through to the original VFS unchanged.
|
|
*/
|
|
static int multiplexShmMap(
|
|
sqlite3_file *pConn, /* Handle open on database file */
|
|
int iRegion, /* Region to retrieve */
|
|
int szRegion, /* Size of regions */
|
|
int bExtend, /* True to extend file if necessary */
|
|
void volatile **pp /* OUT: Mapped memory */
|
|
){
|
|
multiplexConn *p = (multiplexConn*)pConn;
|
|
int rc;
|
|
sqlite3_file *pSubOpen = multiplexSubOpen(p->pGroup, 0, &rc, NULL, 0);
|
|
if( pSubOpen ){
|
|
return pSubOpen->pMethods->xShmMap(pSubOpen, iRegion, szRegion, bExtend,pp);
|
|
}
|
|
return SQLITE_IOERR;
|
|
}
|
|
|
|
/* Pass xShmLock requests through to the original VFS unchanged.
|
|
*/
|
|
static int multiplexShmLock(
|
|
sqlite3_file *pConn, /* Database file holding the shared memory */
|
|
int ofst, /* First lock to acquire or release */
|
|
int n, /* Number of locks to acquire or release */
|
|
int flags /* What to do with the lock */
|
|
){
|
|
multiplexConn *p = (multiplexConn*)pConn;
|
|
int rc;
|
|
sqlite3_file *pSubOpen = multiplexSubOpen(p->pGroup, 0, &rc, NULL, 0);
|
|
if( pSubOpen ){
|
|
return pSubOpen->pMethods->xShmLock(pSubOpen, ofst, n, flags);
|
|
}
|
|
return SQLITE_BUSY;
|
|
}
|
|
|
|
/* Pass xShmBarrier requests through to the original VFS unchanged.
|
|
*/
|
|
static void multiplexShmBarrier(sqlite3_file *pConn){
|
|
multiplexConn *p = (multiplexConn*)pConn;
|
|
int rc;
|
|
sqlite3_file *pSubOpen = multiplexSubOpen(p->pGroup, 0, &rc, NULL, 0);
|
|
if( pSubOpen ){
|
|
pSubOpen->pMethods->xShmBarrier(pSubOpen);
|
|
}
|
|
}
|
|
|
|
/* Pass xShmUnmap requests through to the original VFS unchanged.
|
|
*/
|
|
static int multiplexShmUnmap(sqlite3_file *pConn, int deleteFlag){
|
|
multiplexConn *p = (multiplexConn*)pConn;
|
|
int rc;
|
|
sqlite3_file *pSubOpen = multiplexSubOpen(p->pGroup, 0, &rc, NULL, 0);
|
|
if( pSubOpen ){
|
|
return pSubOpen->pMethods->xShmUnmap(pSubOpen, deleteFlag);
|
|
}
|
|
return SQLITE_OK;
|
|
}
|
|
|
|
/************************** Public Interfaces *****************************/
|
|
/*
|
|
** CAPI: Initialize the multiplex VFS shim - sqlite3_multiplex_initialize()
|
|
**
|
|
** Use the VFS named zOrigVfsName as the VFS that does the actual work.
|
|
** Use the default if zOrigVfsName==NULL.
|
|
**
|
|
** The multiplex VFS shim is named "multiplex". It will become the default
|
|
** VFS if makeDefault is non-zero.
|
|
**
|
|
** THIS ROUTINE IS NOT THREADSAFE. Call this routine exactly once
|
|
** during start-up.
|
|
*/
|
|
int sqlite3_multiplex_initialize(const char *zOrigVfsName, int makeDefault){
|
|
sqlite3_vfs *pOrigVfs;
|
|
if( gMultiplex.isInitialized ) return SQLITE_MISUSE;
|
|
pOrigVfs = sqlite3_vfs_find(zOrigVfsName);
|
|
if( pOrigVfs==0 ) return SQLITE_ERROR;
|
|
assert( pOrigVfs!=&gMultiplex.sThisVfs );
|
|
gMultiplex.isInitialized = 1;
|
|
gMultiplex.pOrigVfs = pOrigVfs;
|
|
gMultiplex.sThisVfs = *pOrigVfs;
|
|
gMultiplex.sThisVfs.szOsFile += sizeof(multiplexConn);
|
|
gMultiplex.sThisVfs.zName = SQLITE_MULTIPLEX_VFS_NAME;
|
|
gMultiplex.sThisVfs.xOpen = multiplexOpen;
|
|
gMultiplex.sThisVfs.xDelete = multiplexDelete;
|
|
gMultiplex.sThisVfs.xAccess = multiplexAccess;
|
|
gMultiplex.sThisVfs.xFullPathname = multiplexFullPathname;
|
|
gMultiplex.sThisVfs.xDlOpen = multiplexDlOpen;
|
|
gMultiplex.sThisVfs.xDlError = multiplexDlError;
|
|
gMultiplex.sThisVfs.xDlSym = multiplexDlSym;
|
|
gMultiplex.sThisVfs.xDlClose = multiplexDlClose;
|
|
gMultiplex.sThisVfs.xRandomness = multiplexRandomness;
|
|
gMultiplex.sThisVfs.xSleep = multiplexSleep;
|
|
gMultiplex.sThisVfs.xCurrentTime = multiplexCurrentTime;
|
|
gMultiplex.sThisVfs.xGetLastError = multiplexGetLastError;
|
|
gMultiplex.sThisVfs.xCurrentTimeInt64 = multiplexCurrentTimeInt64;
|
|
|
|
gMultiplex.sIoMethodsV1.iVersion = 1;
|
|
gMultiplex.sIoMethodsV1.xClose = multiplexClose;
|
|
gMultiplex.sIoMethodsV1.xRead = multiplexRead;
|
|
gMultiplex.sIoMethodsV1.xWrite = multiplexWrite;
|
|
gMultiplex.sIoMethodsV1.xTruncate = multiplexTruncate;
|
|
gMultiplex.sIoMethodsV1.xSync = multiplexSync;
|
|
gMultiplex.sIoMethodsV1.xFileSize = multiplexFileSize;
|
|
gMultiplex.sIoMethodsV1.xLock = multiplexLock;
|
|
gMultiplex.sIoMethodsV1.xUnlock = multiplexUnlock;
|
|
gMultiplex.sIoMethodsV1.xCheckReservedLock = multiplexCheckReservedLock;
|
|
gMultiplex.sIoMethodsV1.xFileControl = multiplexFileControl;
|
|
gMultiplex.sIoMethodsV1.xSectorSize = multiplexSectorSize;
|
|
gMultiplex.sIoMethodsV1.xDeviceCharacteristics =
|
|
multiplexDeviceCharacteristics;
|
|
gMultiplex.sIoMethodsV2 = gMultiplex.sIoMethodsV1;
|
|
gMultiplex.sIoMethodsV2.iVersion = 2;
|
|
gMultiplex.sIoMethodsV2.xShmMap = multiplexShmMap;
|
|
gMultiplex.sIoMethodsV2.xShmLock = multiplexShmLock;
|
|
gMultiplex.sIoMethodsV2.xShmBarrier = multiplexShmBarrier;
|
|
gMultiplex.sIoMethodsV2.xShmUnmap = multiplexShmUnmap;
|
|
sqlite3_vfs_register(&gMultiplex.sThisVfs, makeDefault);
|
|
|
|
sqlite3_auto_extension((void(*)(void))multiplexFuncInit);
|
|
|
|
return SQLITE_OK;
|
|
}
|
|
|
|
/*
|
|
** CAPI: Shutdown the multiplex system - sqlite3_multiplex_shutdown()
|
|
**
|
|
** All SQLite database connections must be closed before calling this
|
|
** routine.
|
|
**
|
|
** THIS ROUTINE IS NOT THREADSAFE. Call this routine exactly once while
|
|
** shutting down in order to free all remaining multiplex groups.
|
|
*/
|
|
int sqlite3_multiplex_shutdown(int eForce){
|
|
int rc = SQLITE_OK;
|
|
if( gMultiplex.isInitialized==0 ) return SQLITE_MISUSE;
|
|
gMultiplex.isInitialized = 0;
|
|
sqlite3_vfs_unregister(&gMultiplex.sThisVfs);
|
|
memset(&gMultiplex, 0, sizeof(gMultiplex));
|
|
return rc;
|
|
}
|
|
|
|
/***************************** Test Code ***********************************/
|
|
#ifdef SQLITE_TEST
|
|
#include "tclsqlite.h"
|
|
extern const char *sqlite3ErrName(int);
|
|
|
|
|
|
/*
|
|
** tclcmd: sqlite3_multiplex_initialize NAME MAKEDEFAULT
|
|
*/
|
|
static int SQLITE_TCLAPI test_multiplex_initialize(
|
|
void * clientData,
|
|
Tcl_Interp *interp,
|
|
int objc,
|
|
Tcl_Obj *CONST objv[]
|
|
){
|
|
const char *zName; /* Name of new multiplex VFS */
|
|
int makeDefault; /* True to make the new VFS the default */
|
|
int rc; /* Value returned by multiplex_initialize() */
|
|
|
|
UNUSED_PARAMETER(clientData);
|
|
|
|
/* Process arguments */
|
|
if( objc!=3 ){
|
|
Tcl_WrongNumArgs(interp, 1, objv, "NAME MAKEDEFAULT");
|
|
return TCL_ERROR;
|
|
}
|
|
zName = Tcl_GetString(objv[1]);
|
|
if( Tcl_GetBooleanFromObj(interp, objv[2], &makeDefault) ) return TCL_ERROR;
|
|
if( zName[0]=='\0' ) zName = 0;
|
|
|
|
/* Call sqlite3_multiplex_initialize() */
|
|
rc = sqlite3_multiplex_initialize(zName, makeDefault);
|
|
Tcl_SetResult(interp, (char *)sqlite3ErrName(rc), TCL_STATIC);
|
|
|
|
return TCL_OK;
|
|
}
|
|
|
|
/*
|
|
** tclcmd: sqlite3_multiplex_shutdown
|
|
*/
|
|
static int SQLITE_TCLAPI test_multiplex_shutdown(
|
|
void * clientData,
|
|
Tcl_Interp *interp,
|
|
int objc,
|
|
Tcl_Obj *CONST objv[]
|
|
){
|
|
int rc; /* Value returned by multiplex_shutdown() */
|
|
|
|
UNUSED_PARAMETER(clientData);
|
|
|
|
if( objc==2 && strcmp(Tcl_GetString(objv[1]),"-force")!=0 ){
|
|
objc = 3;
|
|
}
|
|
if( (objc!=1 && objc!=2) ){
|
|
Tcl_WrongNumArgs(interp, 1, objv, "?-force?");
|
|
return TCL_ERROR;
|
|
}
|
|
|
|
/* Call sqlite3_multiplex_shutdown() */
|
|
rc = sqlite3_multiplex_shutdown(objc==2);
|
|
Tcl_SetResult(interp, (char *)sqlite3ErrName(rc), TCL_STATIC);
|
|
|
|
return TCL_OK;
|
|
}
|
|
|
|
/*
|
|
** Tclcmd: test_multiplex_control HANDLE DBNAME SUB-COMMAND ?INT-VALUE?
|
|
*/
|
|
static int SQLITE_TCLAPI test_multiplex_control(
|
|
ClientData cd,
|
|
Tcl_Interp *interp,
|
|
int objc,
|
|
Tcl_Obj *CONST objv[]
|
|
){
|
|
int rc; /* Return code from file_control() */
|
|
int idx; /* Index in aSub[] */
|
|
Tcl_CmdInfo cmdInfo; /* Command info structure for HANDLE */
|
|
sqlite3 *db; /* Underlying db handle for HANDLE */
|
|
int iValue = 0;
|
|
void *pArg = 0;
|
|
|
|
struct SubCommand {
|
|
const char *zName;
|
|
int op;
|
|
int argtype;
|
|
} aSub[] = {
|
|
{ "enable", MULTIPLEX_CTRL_ENABLE, 1 },
|
|
{ "chunk_size", MULTIPLEX_CTRL_SET_CHUNK_SIZE, 1 },
|
|
{ "max_chunks", MULTIPLEX_CTRL_SET_MAX_CHUNKS, 1 },
|
|
{ 0, 0, 0 }
|
|
};
|
|
|
|
if( objc!=5 ){
|
|
Tcl_WrongNumArgs(interp, 1, objv, "HANDLE DBNAME SUB-COMMAND INT-VALUE");
|
|
return TCL_ERROR;
|
|
}
|
|
|
|
if( 0==Tcl_GetCommandInfo(interp, Tcl_GetString(objv[1]), &cmdInfo) ){
|
|
Tcl_AppendResult(interp, "expected database handle, got \"", 0);
|
|
Tcl_AppendResult(interp, Tcl_GetString(objv[1]), "\"", 0);
|
|
return TCL_ERROR;
|
|
}else{
|
|
db = *(sqlite3 **)cmdInfo.objClientData;
|
|
}
|
|
|
|
rc = Tcl_GetIndexFromObjStruct(
|
|
interp, objv[3], aSub, sizeof(aSub[0]), "sub-command", 0, &idx
|
|
);
|
|
if( rc!=TCL_OK ) return rc;
|
|
|
|
switch( aSub[idx].argtype ){
|
|
case 1:
|
|
if( Tcl_GetIntFromObj(interp, objv[4], &iValue) ){
|
|
return TCL_ERROR;
|
|
}
|
|
pArg = (void *)&iValue;
|
|
break;
|
|
default:
|
|
Tcl_WrongNumArgs(interp, 4, objv, "SUB-COMMAND");
|
|
return TCL_ERROR;
|
|
}
|
|
|
|
rc = sqlite3_file_control(db, Tcl_GetString(objv[2]), aSub[idx].op, pArg);
|
|
Tcl_SetResult(interp, (char *)sqlite3ErrName(rc), TCL_STATIC);
|
|
return (rc==SQLITE_OK) ? TCL_OK : TCL_ERROR;
|
|
}
|
|
|
|
/*
|
|
** This routine registers the custom TCL commands defined in this
|
|
** module. This should be the only procedure visible from outside
|
|
** of this module.
|
|
*/
|
|
int Sqlitemultiplex_Init(Tcl_Interp *interp){
|
|
static struct {
|
|
char *zName;
|
|
Tcl_ObjCmdProc *xProc;
|
|
} aCmd[] = {
|
|
{ "sqlite3_multiplex_initialize", test_multiplex_initialize },
|
|
{ "sqlite3_multiplex_shutdown", test_multiplex_shutdown },
|
|
{ "sqlite3_multiplex_control", test_multiplex_control },
|
|
};
|
|
int i;
|
|
|
|
for(i=0; i<sizeof(aCmd)/sizeof(aCmd[0]); i++){
|
|
Tcl_CreateObjCommand(interp, aCmd[i].zName, aCmd[i].xProc, 0, 0);
|
|
}
|
|
|
|
return TCL_OK;
|
|
}
|
|
#endif
|