4 ** The author disclaims copyright to this source code. In place of
5 ** a legal notice, here is a blessing:
7 ** May you do good and not evil.
8 ** May you find forgiveness for yourself and forgive others.
9 ** May you share freely, never taking more than you give.
11 *************************************************************************
13 ** This file contains code used to implement mutexes on Btree objects.
14 ** This code really belongs in btree.c. But btree.c is getting too
15 ** big and we want to break it down some. This packaged seemed like
19 #ifndef SQLITE_OMIT_SHARED_CACHE
23 ** Obtain the BtShared mutex associated with B-Tree handle p. Also,
24 ** set BtShared.db to the database handle associated with p and the
25 ** p->locked boolean to true.
27 static void lockBtreeMutex(Btree
*p
){
28 assert( p
->locked
==0 );
29 assert( sqlite3_mutex_notheld(p
->pBt
->mutex
) );
30 assert( sqlite3_mutex_held(p
->db
->mutex
) );
32 sqlite3_mutex_enter(p
->pBt
->mutex
);
38 ** Release the BtShared mutex associated with B-Tree handle p and
39 ** clear the p->locked boolean.
41 static void unlockBtreeMutex(Btree
*p
){
42 BtShared
*pBt
= p
->pBt
;
43 assert( p
->locked
==1 );
44 assert( sqlite3_mutex_held(pBt
->mutex
) );
45 assert( sqlite3_mutex_held(p
->db
->mutex
) );
46 assert( p
->db
==pBt
->db
);
48 sqlite3_mutex_leave(pBt
->mutex
);
53 ** Enter a mutex on the given BTree object.
55 ** If the object is not sharable, then no mutex is ever required
56 ** and this routine is a no-op. The underlying mutex is non-recursive.
57 ** But we keep a reference count in Btree.wantToLock so the behavior
58 ** of this interface is recursive.
60 ** To avoid deadlocks, multiple Btrees are locked in the same order
61 ** by all database connections. The p->pNext is a list of other
62 ** Btrees belonging to the same database connection as the p Btree
63 ** which need to be locked after p. If we cannot get a lock on
64 ** p, then first unlock all of the others on p->pNext, then wait
65 ** for the lock to become available on p, then relock all of the
66 ** subsequent Btrees that desire a lock.
68 void sqlite3BtreeEnter(Btree
*p
){
71 /* Some basic sanity checking on the Btree. The list of Btrees
72 ** connected by pNext and pPrev should be in sorted order by
73 ** Btree.pBt value. All elements of the list should belong to
74 ** the same connection. Only shared Btrees are on the list. */
75 assert( p
->pNext
==0 || p
->pNext
->pBt
>p
->pBt
);
76 assert( p
->pPrev
==0 || p
->pPrev
->pBt
<p
->pBt
);
77 assert( p
->pNext
==0 || p
->pNext
->db
==p
->db
);
78 assert( p
->pPrev
==0 || p
->pPrev
->db
==p
->db
);
79 assert( p
->sharable
|| (p
->pNext
==0 && p
->pPrev
==0) );
81 /* Check for locking consistency */
82 assert( !p
->locked
|| p
->wantToLock
>0 );
83 assert( p
->sharable
|| p
->wantToLock
==0 );
85 /* We should already hold a lock on the database connection */
86 assert( sqlite3_mutex_held(p
->db
->mutex
) );
88 /* Unless the database is sharable and unlocked, then BtShared.db
89 ** should already be set correctly. */
90 assert( (p
->locked
==0 && p
->sharable
) || p
->pBt
->db
==p
->db
);
92 if( !p
->sharable
) return;
94 if( p
->locked
) return;
96 /* In most cases, we should be able to acquire the lock we
97 ** want without having to go throught the ascending lock
98 ** procedure that follows. Just be sure not to block.
100 if( sqlite3_mutex_try(p
->pBt
->mutex
)==SQLITE_OK
){
106 /* To avoid deadlock, first release all locks with a larger
107 ** BtShared address. Then acquire our lock. Then reacquire
108 ** the other BtShared locks that we used to hold in ascending
111 for(pLater
=p
->pNext
; pLater
; pLater
=pLater
->pNext
){
112 assert( pLater
->sharable
);
113 assert( pLater
->pNext
==0 || pLater
->pNext
->pBt
>pLater
->pBt
);
114 assert( !pLater
->locked
|| pLater
->wantToLock
>0 );
115 if( pLater
->locked
){
116 unlockBtreeMutex(pLater
);
120 for(pLater
=p
->pNext
; pLater
; pLater
=pLater
->pNext
){
121 if( pLater
->wantToLock
){
122 lockBtreeMutex(pLater
);
128 ** Exit the recursive mutex on a Btree.
130 void sqlite3BtreeLeave(Btree
*p
){
132 assert( p
->wantToLock
>0 );
134 if( p
->wantToLock
==0 ){
142 ** Return true if the BtShared mutex is held on the btree, or if the
143 ** B-Tree is not marked as sharable.
145 ** This routine is used only from within assert() statements.
147 int sqlite3BtreeHoldsMutex(Btree
*p
){
148 assert( p
->sharable
==0 || p
->locked
==0 || p
->wantToLock
>0 );
149 assert( p
->sharable
==0 || p
->locked
==0 || p
->db
==p
->pBt
->db
);
150 assert( p
->sharable
==0 || p
->locked
==0 || sqlite3_mutex_held(p
->pBt
->mutex
) );
151 assert( p
->sharable
==0 || p
->locked
==0 || sqlite3_mutex_held(p
->db
->mutex
) );
153 return (p
->sharable
==0 || p
->locked
);
158 #ifndef SQLITE_OMIT_INCRBLOB
160 ** Enter and leave a mutex on a Btree given a cursor owned by that
161 ** Btree. These entry points are used by incremental I/O and can be
162 ** omitted if that module is not used.
164 void sqlite3BtreeEnterCursor(BtCursor
*pCur
){
165 sqlite3BtreeEnter(pCur
->pBtree
);
167 void sqlite3BtreeLeaveCursor(BtCursor
*pCur
){
168 sqlite3BtreeLeave(pCur
->pBtree
);
170 #endif /* SQLITE_OMIT_INCRBLOB */
174 ** Enter the mutex on every Btree associated with a database
175 ** connection. This is needed (for example) prior to parsing
176 ** a statement since we will be comparing table and column names
177 ** against all schemas and we do not want those schemas being
178 ** reset out from under us.
180 ** There is a corresponding leave-all procedures.
182 ** Enter the mutexes in accending order by BtShared pointer address
183 ** to avoid the possibility of deadlock when two threads with
184 ** two or more btrees in common both try to lock all their btrees
185 ** at the same instant.
187 void sqlite3BtreeEnterAll(sqlite3
*db
){
190 assert( sqlite3_mutex_held(db
->mutex
) );
191 for(i
=0; i
<db
->nDb
; i
++){
193 if( p
) sqlite3BtreeEnter(p
);
196 void sqlite3BtreeLeaveAll(sqlite3
*db
){
199 assert( sqlite3_mutex_held(db
->mutex
) );
200 for(i
=0; i
<db
->nDb
; i
++){
202 if( p
) sqlite3BtreeLeave(p
);
207 ** Return true if a particular Btree requires a lock. Return FALSE if
208 ** no lock is ever required since it is not sharable.
210 int sqlite3BtreeSharable(Btree
*p
){
216 ** Return true if the current thread holds the database connection
217 ** mutex and all required BtShared mutexes.
219 ** This routine is used inside assert() statements only.
221 int sqlite3BtreeHoldsAllMutexes(sqlite3
*db
){
223 if( !sqlite3_mutex_held(db
->mutex
) ){
226 for(i
=0; i
<db
->nDb
; i
++){
229 if( p
&& p
->sharable
&&
230 (p
->wantToLock
==0 || !sqlite3_mutex_held(p
->pBt
->mutex
)) ){
240 ** Return true if the correct mutexes are held for accessing the
241 ** db->aDb[iDb].pSchema structure. The mutexes required for schema
244 ** (1) The mutex on db
245 ** (2) if iDb!=1, then the mutex on db->aDb[iDb].pBt.
247 ** If pSchema is not NULL, then iDb is computed from pSchema and
248 ** db using sqlite3SchemaToIndex().
250 int sqlite3SchemaMutexHeld(sqlite3
*db
, int iDb
, Schema
*pSchema
){
253 if( pSchema
) iDb
= sqlite3SchemaToIndex(db
, pSchema
);
254 assert( iDb
>=0 && iDb
<db
->nDb
);
255 if( !sqlite3_mutex_held(db
->mutex
) ) return 0;
256 if( iDb
==1 ) return 1;
257 p
= db
->aDb
[iDb
].pBt
;
259 return p
->sharable
==0 || p
->locked
==1;
263 #else /* SQLITE_THREADSAFE>0 above. SQLITE_THREADSAFE==0 below */
265 ** The following are special cases for mutex enter routines for use
266 ** in single threaded applications that use shared cache. Except for
267 ** these two routines, all mutex operations are no-ops in that case and
268 ** are null #defines in btree.h.
270 ** If shared cache is disabled, then all btree mutex routines, including
271 ** the ones below, are no-ops and are null #defines in btree.h.
274 void sqlite3BtreeEnter(Btree
*p
){
277 void sqlite3BtreeEnterAll(sqlite3
*db
){
279 for(i
=0; i
<db
->nDb
; i
++){
280 Btree
*p
= db
->aDb
[i
].pBt
;
286 #endif /* if SQLITE_THREADSAFE */
287 #endif /* ifndef SQLITE_OMIT_SHARED_CACHE */