database/postgres
database/postgres
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2025-04-22 23:02:54 +03:00
Code

Complete TODO item:

Browse Source 
* -HOLDER/HOLDERTAB rename to PROCLOCK/PROCLOCKTAG
This commit is contained in:
Bruce Momjian 2002-07-19 00:17:40 +00:00
parent 97377048b4
commit b75fcf9326
6 changed files with 196 additions and 190 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

210
src/backend/storage/lmgr/README
View File

@ -1,4 +1,4 @@
$Header: /cvsroot/pgsql/src/backend/storage/lmgr/README,v 1.10 2002/04/15 23:46:13 momjian Exp $
$Header: /cvsroot/pgsql/src/backend/storage/lmgr/README,v 1.11 2002/07/19 00:17:40 momjian Exp $
LOCKING OVERVIEW
@ -7,38 +7,40 @@ Postgres uses three types of interprocess locks:
* Spinlocks. These are intended for *very* short-term locks. If a lock
is to be held more than a few dozen instructions, or across any sort of
kernel call (or even a call to a nontrivial subroutine), don't use a spinlock.
Spinlocks are primarily used as infrastructure for lightweight locks.
They are implemented using a hardware atomic-test-and-set instruction,
if available. Waiting processes busy-loop until they can get the lock.
There is no provision for deadlock detection, automatic release on error,
or any other nicety. There is a timeout if the lock cannot be gotten after
a minute or so (which is approximately forever in comparison to the intended
lock hold time, so this is certainly an error condition).
kernel call (or even a call to a nontrivial subroutine), don't use a
spinlock. Spinlocks are primarily used as infrastructure for lightweight
locks. They are implemented using a hardware atomic-test-and-set
instruction, if available. Waiting processes busy-loop until they can
get the lock. There is no provision for deadlock detection, automatic
release on error, or any other nicety. There is a timeout if the lock
cannot be gotten after a minute or so (which is approximately forever in
comparison to the intended lock hold time, so this is certainly an error
condition).
* Lightweight locks (LWLocks). These locks are typically used to interlock
access to datastructures in shared memory. LWLocks support both exclusive
and shared lock modes (for read/write and read-only access to a shared object).
There is no provision for deadlock detection, but the LWLock manager will
automatically release held LWLocks during elog() recovery, so it is safe to
raise an error while holding LWLocks. Obtaining or releasing an LWLock is
quite fast (a few dozen instructions) when there is no contention for the
lock. When a process has to wait for an LWLock, it blocks on a SysV semaphore
so as to not consume CPU time. Waiting processes will be granted the lock
in arrival order. There is no timeout.
* Lightweight locks (LWLocks). These locks are typically used to
interlock access to datastructures in shared memory. LWLocks support
both exclusive and shared lock modes (for read/write and read-only
access to a shared object). There is no provision for deadlock
detection, but the LWLock manager will automatically release held
LWLocks during elog() recovery, so it is safe to raise an error while
holding LWLocks. Obtaining or releasing an LWLock is quite fast (a few
dozen instructions) when there is no contention for the lock. When a
process has to wait for an LWLock, it blocks on a SysV semaphore so as
to not consume CPU time. Waiting processes will be granted the lock in
arrival order. There is no timeout.
* Regular locks (a/k/a heavyweight locks). The regular lock manager supports
a variety of lock modes with table-driven semantics, and it has full deadlock
detection and automatic release at transaction end. Regular locks should be
used for all user-driven lock requests.
* Regular locks (a/k/a heavyweight locks). The regular lock manager
supports a variety of lock modes with table-driven semantics, and it has
full deadlock detection and automatic release at transaction end.
Regular locks should be used for all user-driven lock requests.
Acquisition of either a spinlock or a lightweight lock causes query cancel
and die() interrupts to be held off until all such locks are released.
No such restriction exists for regular locks, however. Also note that we
can accept query cancel and die() interrupts while waiting for a regular
lock, but we will not accept them while waiting for spinlocks or LW locks.
It is therefore not a good idea to use LW locks when the wait time might
exceed a few seconds.
Acquisition of either a spinlock or a lightweight lock causes query
cancel and die() interrupts to be held off until all such locks are
released. No such restriction exists for regular locks, however. Also
note that we can accept query cancel and die() interrupts while waiting
for a regular lock, but we will not accept them while waiting for
spinlocks or LW locks. It is therefore not a good idea to use LW locks
when the wait time might exceed a few seconds.
The rest of this README file discusses the regular lock manager in detail.
@ -46,9 +48,9 @@ The rest of this README file discusses the regular lock manager in detail.
LOCK DATA STRUCTURES
There are two fundamental lock structures: the per-lockable-object LOCK
struct, and the per-lock-holder HOLDER struct. A LOCK object exists
struct, and the per-lock-holder PROCLOCK struct. A LOCK object exists
for each lockable object that currently has locks held or requested on it.
A HOLDER struct exists for each transaction that is holding or requesting
A PROCLOCK struct exists for each transaction that is holding or requesting
lock(s) on each LOCK object.
Lock methods describe the overall locking behavior. Currently there are
@ -102,9 +104,9 @@ waitMask -
is 1 if and only if requested[i] > granted[i].
lockHolders -
This is a shared memory queue of all the HOLDER structs associated with
the lock object. Note that both granted and waiting HOLDERs are in this
list (indeed, the same HOLDER might have some already-granted locks and
This is a shared memory queue of all the PROCLOCK structs associated with
the lock object. Note that both granted and waiting PROCLOCKs are in this
list (indeed, the same PROCLOCK might have some already-granted locks and
be waiting for more!).
waitProcs -
@ -144,22 +146,22 @@ zero, the lock object is no longer needed and can be freed.
---------------------------------------------------------------------------
The lock manager's HOLDER objects contain:
The lock manager's PROCLOCK objects contain:
tag -
The key fields that are used for hashing entries in the shared memory
holder hash table. This is declared as a separate struct to ensure that
PROCLOCK hash table. This is declared as a separate struct to ensure that
we always zero out the correct number of bytes.
tag.lock
SHMEM offset of the LOCK object this holder is for.
SHMEM offset of the LOCK object this PROCLOCK is for.
tag.proc
SHMEM offset of PROC of backend process that owns this holder.
SHMEM offset of PROC of backend process that owns this PROCLOCK.
tag.xid
XID of transaction this holder is for, or InvalidTransactionId
if the holder is for session-level locking.
XID of transaction this PROCLOCK is for, or InvalidTransactionId
if the PROCLOCK is for session-level locking.
Note that this structure will support multiple transactions running
concurrently in one backend, which may be handy if we someday decide
@ -169,18 +171,18 @@ tag -
transaction operations like VACUUM.
holding -
The number of successfully acquired locks of each type for this holder.
The number of successfully acquired locks of each type for this PROCLOCK.
This should be <= the corresponding granted[] value of the lock object!
nHolding -
Sum of the holding[] array.
lockLink -
List link for shared memory queue of all the HOLDER objects for the
List link for shared memory queue of all the PROCLOCK objects for the
same LOCK.
procLink -
List link for shared memory queue of all the HOLDER objects for the
List link for shared memory queue of all the PROCLOCK objects for the
same backend.
---------------------------------------------------------------------------
@ -193,47 +195,48 @@ fairly standard in essence, but there are many special considerations
needed to deal with Postgres' generalized locking model.
A key design consideration is that we want to make routine operations
(lock grant and release) run quickly when there is no deadlock, and avoid
the overhead of deadlock handling as much as possible. We do this using
an "optimistic waiting" approach: if a process cannot acquire the lock
it wants immediately, it goes to sleep without any deadlock check. But
it also sets a delay timer, with a delay of DeadlockTimeout milliseconds
(typically set to one second). If the delay expires before the process is
granted the lock it wants, it runs the deadlock detection/breaking code.
Normally this code will determine that there is no deadlock condition,
and then the process will go back to sleep and wait quietly until it is
granted the lock. But if a deadlock condition does exist, it will be
resolved, usually by aborting the detecting process' transaction. In this
way, we avoid deadlock handling overhead whenever the wait time for a lock
is less than DeadlockTimeout, while not imposing an unreasonable delay of
detection when there is an error.
(lock grant and release) run quickly when there is no deadlock, and
avoid the overhead of deadlock handling as much as possible. We do this
using an "optimistic waiting" approach: if a process cannot acquire the
lock it wants immediately, it goes to sleep without any deadlock check.
But it also sets a delay timer, with a delay of DeadlockTimeout
milliseconds (typically set to one second). If the delay expires before
the process is granted the lock it wants, it runs the deadlock
detection/breaking code. Normally this code will determine that there is
no deadlock condition, and then the process will go back to sleep and
wait quietly until it is granted the lock. But if a deadlock condition
does exist, it will be resolved, usually by aborting the detecting
process' transaction. In this way, we avoid deadlock handling overhead
whenever the wait time for a lock is less than DeadlockTimeout, while
not imposing an unreasonable delay of detection when there is an error.
Lock acquisition (routines LockAcquire and ProcSleep) follows these rules:
1. A lock request is granted immediately if it does not conflict with any
existing or waiting lock request, or if the process already holds an
1. A lock request is granted immediately if it does not conflict with
any existing or waiting lock request, or if the process already holds an
instance of the same lock type (eg, there's no penalty to acquire a read
lock twice). Note that a process never conflicts with itself, eg one can
obtain read lock when one already holds exclusive lock.
lock twice). Note that a process never conflicts with itself, eg one
can obtain read lock when one already holds exclusive lock.
2. Otherwise the process joins the lock's wait queue. Normally it will be
added to the end of the queue, but there is an exception: if the process
already holds locks on this same lockable object that conflict with the
request of any pending waiter, then the process will be inserted in the
wait queue just ahead of the first such waiter. (If we did not make this
check, the deadlock detection code would adjust the queue order to resolve
the conflict, but it's relatively cheap to make the check in ProcSleep and
avoid a deadlock timeout delay in this case.) Note special case when
inserting before the end of the queue: if the process's request does not
conflict with any existing lock nor any waiting request before its insertion
point, then go ahead and grant the lock without waiting.
2. Otherwise the process joins the lock's wait queue. Normally it will
be added to the end of the queue, but there is an exception: if the
process already holds locks on this same lockable object that conflict
with the request of any pending waiter, then the process will be
inserted in the wait queue just ahead of the first such waiter. (If we
did not make this check, the deadlock detection code would adjust the
queue order to resolve the conflict, but it's relatively cheap to make
the check in ProcSleep and avoid a deadlock timeout delay in this case.)
Note special case when inserting before the end of the queue: if the
process's request does not conflict with any existing lock nor any
waiting request before its insertion point, then go ahead and grant the
lock without waiting.
When a lock is released, the lock release routine (ProcLockWakeup) scans
the lock object's wait queue. Each waiter is awoken if (a) its request
does not conflict with already-granted locks, and (b) its request does
not conflict with the requests of prior un-wakable waiters. Rule (b)
ensures that conflicting requests are granted in order of arrival.
There are cases where a later waiter must be allowed to go in front of
ensures that conflicting requests are granted in order of arrival. There
are cases where a later waiter must be allowed to go in front of
conflicting earlier waiters to avoid deadlock, but it is not
ProcLockWakeup's responsibility to recognize these cases; instead, the
deadlock detection code will re-order the wait queue when necessary.
@ -242,35 +245,36 @@ To perform deadlock checking, we use the standard method of viewing the
various processes as nodes in a directed graph (the waits-for graph or
WFG). There is a graph edge leading from process A to process B if A
waits for B, ie, A is waiting for some lock and B holds a conflicting
lock. There is a deadlock condition if and only if the WFG contains
a cycle. We detect cycles by searching outward along waits-for edges
to see if we return to our starting point. There are three possible
lock. There is a deadlock condition if and only if the WFG contains a
cycle. We detect cycles by searching outward along waits-for edges to
see if we return to our starting point. There are three possible
outcomes:
1. All outgoing paths terminate at a running process (which has no
outgoing edge).
2. A deadlock is detected by looping back to the start point. We resolve
such a deadlock by canceling the start point's lock request and reporting
an error in that transaction, which normally leads to transaction abort
and release of that transaction's held locks. Note that it's sufficient
to cancel one request to remove the cycle; we don't need to kill all the
transactions involved.
2. A deadlock is detected by looping back to the start point. We
resolve such a deadlock by canceling the start point's lock request and
reporting an error in that transaction, which normally leads to
transaction abort and release of that transaction's held locks. Note
that it's sufficient to cancel one request to remove the cycle; we don't
need to kill all the transactions involved.
3. Some path(s) loop back to a node other than the start point. This
indicates a deadlock, but one that does not involve our starting process.
We ignore this condition on the grounds that resolving such a deadlock
is the responsibility of the processes involved --- killing our start-
point process would not resolve the deadlock. So, cases 1 and 3 both
report "no deadlock".
indicates a deadlock, but one that does not involve our starting
process. We ignore this condition on the grounds that resolving such a
deadlock is the responsibility of the processes involved --- killing our
start- point process would not resolve the deadlock. So, cases 1 and 3
both report "no deadlock".
Postgres' situation is a little more complex than the standard discussion
of deadlock detection, for two reasons:
1. A process can be waiting for more than one other process, since there
might be multiple holders of (non-conflicting) lock types that all conflict
with the waiter's request. This creates no real difficulty however; we
simply need to be prepared to trace more than one outgoing edge.
might be multiple PROCLOCKs of (non-conflicting) lock types that all
conflict with the waiter's request. This creates no real difficulty
however; we simply need to be prepared to trace more than one outgoing
edge.
2. If a process A is behind a process B in some lock's wait queue, and
their requested locks conflict, then we must say that A waits for B, since
@ -409,16 +413,18 @@ LockWaitCancel (abort a waiter due to outside factors) must run
ProcLockWakeup, in case the canceled waiter was soft-blocking other
waiters.
4. We can minimize excess rearrangement-trial work by being careful to scan
the wait queue from the front when looking for soft edges. For example,
if we have queue order A,B,C and C has deadlock conflicts with both A and B,
we want to generate the "C before A" constraint first, rather than wasting
time with "C before B", which won't move C far enough up. So we look for
soft edges outgoing from C starting at the front of the wait queue.
4. We can minimize excess rearrangement-trial work by being careful to
scan the wait queue from the front when looking for soft edges. For
example, if we have queue order A,B,C and C has deadlock conflicts with
both A and B, we want to generate the "C before A" constraint first,
rather than wasting time with "C before B", which won't move C far
enough up. So we look for soft edges outgoing from C starting at the
front of the wait queue.
5. The working data structures needed by the deadlock detection code can
be limited to numbers of entries computed from MaxBackends. Therefore,
we can allocate the worst-case space needed during backend startup.
This seems a safer approach than trying to allocate workspace on the fly;
we don't want to risk having the deadlock detector run out of memory,
else we really have no guarantees at all that deadlock will be detected.
we can allocate the worst-case space needed during backend startup. This
seems a safer approach than trying to allocate workspace on the fly; we
don't want to risk having the deadlock detector run out of memory, else
we really have no guarantees at all that deadlock will be detected.

12
src/backend/storage/lmgr/deadlock.c
View File

@ -12,7 +12,7 @@
*
*
* IDENTIFICATION
* $Header: /cvsroot/pgsql/src/backend/storage/lmgr/deadlock.c,v 1.11 2002/07/18 23:06:19 momjian Exp $
* $Header: /cvsroot/pgsql/src/backend/storage/lmgr/deadlock.c,v 1.12 2002/07/19 00:17:40 momjian Exp $
*
* Interface:
*
@ -377,7 +377,7 @@ FindLockCycleRecurse(PGPROC *checkProc,
{
PGPROC *proc;
LOCK *lock;
HOLDER *holder;
PROCLOCK *holder;
SHM_QUEUE *lockHolders;
LOCKMETHODTABLE *lockMethodTable;
PROC_QUEUE *waitQueue;
@ -427,8 +427,8 @@ FindLockCycleRecurse(PGPROC *checkProc,
*/
lockHolders = &(lock->lockHolders);
holder = (HOLDER *) SHMQueueNext(lockHolders, lockHolders,
offsetof(HOLDER, lockLink));
holder = (PROCLOCK *) SHMQueueNext(lockHolders, lockHolders,
offsetof(PROCLOCK, lockLink));
while (holder)
{
@ -451,8 +451,8 @@ FindLockCycleRecurse(PGPROC *checkProc,
}
}
holder = (HOLDER *) SHMQueueNext(lockHolders, &holder->lockLink,
offsetof(HOLDER, lockLink));
holder = (PROCLOCK *) SHMQueueNext(lockHolders, &holder->lockLink,
offsetof(PROCLOCK, lockLink));
}
/*

122
src/backend/storage/lmgr/lock.c
View File

@ -8,7 +8,7 @@
*
*
* IDENTIFICATION
* $Header: /cvsroot/pgsql/src/backend/storage/lmgr/lock.c,v 1.109 2002/07/18 23:06:19 momjian Exp $
* $Header: /cvsroot/pgsql/src/backend/storage/lmgr/lock.c,v 1.110 2002/07/19 00:17:40 momjian Exp $
*
* NOTES
* Outside modules can create a lock table and acquire/release
@ -48,7 +48,7 @@ int max_locks_per_xact; /* set by guc.c */
static int WaitOnLock(LOCKMETHOD lockmethod, LOCKMODE lockmode,
LOCK *lock, HOLDER *holder);
LOCK *lock, PROCLOCK *holder);
static void LockCountMyLocks(SHMEM_OFFSET lockOffset, PGPROC *proc,
int *myHolding);
@ -125,18 +125,18 @@ LOCK_PRINT(const char *where, const LOCK *lock, LOCKMODE type)
inline static void
HOLDER_PRINT(const char *where, const HOLDER *holderP)
PROCLOCK_PRINT(const char *where, const PROCLOCK *holderP)
{
if (
(((HOLDER_LOCKMETHOD(*holderP) == DEFAULT_LOCKMETHOD && Trace_locks)
|| (HOLDER_LOCKMETHOD(*holderP) == USER_LOCKMETHOD && Trace_userlocks))
(((PROCLOCK_LOCKMETHOD(*holderP) == DEFAULT_LOCKMETHOD && Trace_locks)
|| (PROCLOCK_LOCKMETHOD(*holderP) == USER_LOCKMETHOD && Trace_userlocks))
&& (((LOCK *) MAKE_PTR(holderP->tag.lock))->tag.relId >= (Oid) Trace_lock_oidmin))
|| (Trace_lock_table && (((LOCK *) MAKE_PTR(holderP->tag.lock))->tag.relId == Trace_lock_table))
)
elog(LOG,
"%s: holder(%lx) lock(%lx) tbl(%d) proc(%lx) xid(%u) hold(%d,%d,%d,%d,%d,%d,%d)=%d",
where, MAKE_OFFSET(holderP), holderP->tag.lock,
HOLDER_LOCKMETHOD(*(holderP)),
PROCLOCK_LOCKMETHOD(*(holderP)),
holderP->tag.proc, holderP->tag.xid,
holderP->holding[1], holderP->holding[2], holderP->holding[3],
holderP->holding[4], holderP->holding[5], holderP->holding[6],
@ -146,7 +146,7 @@ HOLDER_PRINT(const char *where, const HOLDER *holderP)
#else /* not LOCK_DEBUG */
#define LOCK_PRINT(where, lock, type)
#define HOLDER_PRINT(where, holderP)
#define PROCLOCK_PRINT(where, holderP)
#endif /* not LOCK_DEBUG */
@ -316,11 +316,11 @@ LockMethodTableInit(char *tabName,
Assert(lockMethodTable->lockHash->hash == tag_hash);
/*
* allocate a hash table for HOLDER structs. This is used to store
* allocate a hash table for PROCLOCK structs. This is used to store
* per-lock-holder information.
*/
info.keysize = sizeof(HOLDERTAG);
info.entrysize = sizeof(HOLDER);
info.keysize = sizeof(PROCLOCKTAG);
info.entrysize = sizeof(PROCLOCK);
info.hash = tag_hash;
hash_flags = (HASH_ELEM | HASH_FUNCTION);
@ -440,8 +440,8 @@ bool
LockAcquire(LOCKMETHOD lockmethod, LOCKTAG *locktag,
TransactionId xid, LOCKMODE lockmode, bool dontWait)
{
HOLDER *holder;
HOLDERTAG holdertag;
PROCLOCK *holder;
PROCLOCKTAG holdertag;
HTAB *holderTable;
bool found;
LOCK *lock;
@ -513,7 +513,7 @@ LockAcquire(LOCKMETHOD lockmethod, LOCKTAG *locktag,
/*
* Create the hash key for the holder table.
*/
MemSet(&holdertag, 0, sizeof(HOLDERTAG)); /* must clear padding,
MemSet(&holdertag, 0, sizeof(PROCLOCKTAG)); /* must clear padding,
* needed */
holdertag.lock = MAKE_OFFSET(lock);
holdertag.proc = MAKE_OFFSET(MyProc);
@ -523,7 +523,7 @@ LockAcquire(LOCKMETHOD lockmethod, LOCKTAG *locktag,
* Find or create a holder entry with this tag
*/
holderTable = lockMethodTable->holderHash;
holder = (HOLDER *) hash_search(holderTable,
holder = (PROCLOCK *) hash_search(holderTable,
(void *) &holdertag,
HASH_ENTER, &found);
if (!holder)
@ -543,11 +543,11 @@ LockAcquire(LOCKMETHOD lockmethod, LOCKTAG *locktag,
/* Add holder to appropriate lists */
SHMQueueInsertBefore(&lock->lockHolders, &holder->lockLink);
SHMQueueInsertBefore(&MyProc->procHolders, &holder->procLink);
HOLDER_PRINT("LockAcquire: new", holder);
PROCLOCK_PRINT("LockAcquire: new", holder);
}
else
{
HOLDER_PRINT("LockAcquire: found", holder);
PROCLOCK_PRINT("LockAcquire: found", holder);
Assert((holder->nHolding >= 0) && (holder->holding[lockmode] >= 0));
Assert(holder->nHolding <= lock->nGranted);
@ -600,7 +600,7 @@ LockAcquire(LOCKMETHOD lockmethod, LOCKTAG *locktag,
if (holder->holding[lockmode] > 0)
{
GrantLock(lock, holder, lockmode);
HOLDER_PRINT("LockAcquire: owning", holder);
PROCLOCK_PRINT("LockAcquire: owning", holder);
LWLockRelease(masterLock);
return TRUE;
}
@ -613,7 +613,7 @@ LockAcquire(LOCKMETHOD lockmethod, LOCKTAG *locktag,
if (myHolding[lockmode] > 0)
{
GrantLock(lock, holder, lockmode);
HOLDER_PRINT("LockAcquire: my other XID owning", holder);
PROCLOCK_PRINT("LockAcquire: my other XID owning", holder);
LWLockRelease(masterLock);
return TRUE;
}
@ -650,14 +650,14 @@ LockAcquire(LOCKMETHOD lockmethod, LOCKTAG *locktag,
{
SHMQueueDelete(&holder->lockLink);
SHMQueueDelete(&holder->procLink);
holder = (HOLDER *) hash_search(holderTable,
holder = (PROCLOCK *) hash_search(holderTable,
(void *) holder,
HASH_REMOVE, NULL);
if (!holder)
elog(WARNING, "LockAcquire: remove holder, table corrupted");
}
else
HOLDER_PRINT("LockAcquire: NHOLDING", holder);
PROCLOCK_PRINT("LockAcquire: NHOLDING", holder);
lock->nRequested--;
lock->requested[lockmode]--;
LOCK_PRINT("LockAcquire: conditional lock failed", lock, lockmode);
@ -702,13 +702,13 @@ LockAcquire(LOCKMETHOD lockmethod, LOCKTAG *locktag,
*/
if (!((holder->nHolding > 0) && (holder->holding[lockmode] > 0)))
{
HOLDER_PRINT("LockAcquire: INCONSISTENT", holder);
PROCLOCK_PRINT("LockAcquire: INCONSISTENT", holder);
LOCK_PRINT("LockAcquire: INCONSISTENT", lock, lockmode);
/* Should we retry ? */
LWLockRelease(masterLock);
return FALSE;
}
HOLDER_PRINT("LockAcquire: granted", holder);
PROCLOCK_PRINT("LockAcquire: granted", holder);
LOCK_PRINT("LockAcquire: granted", lock, lockmode);
}
@ -737,7 +737,7 @@ int
LockCheckConflicts(LOCKMETHODTABLE *lockMethodTable,
LOCKMODE lockmode,
LOCK *lock,
HOLDER *holder,
PROCLOCK *holder,
PGPROC *proc,
int *myHolding) /* myHolding[] array or NULL */
{
@ -758,7 +758,7 @@ LockCheckConflicts(LOCKMETHODTABLE *lockMethodTable,
*/
if (!(lockMethodTable->conflictTab[lockmode] & lock->grantMask))
{
HOLDER_PRINT("LockCheckConflicts: no conflict", holder);
PROCLOCK_PRINT("LockCheckConflicts: no conflict", holder);
return STATUS_OK;
}
@ -792,11 +792,11 @@ LockCheckConflicts(LOCKMETHODTABLE *lockMethodTable,
if (!(lockMethodTable->conflictTab[lockmode] & bitmask))
{
/* no conflict. OK to get the lock */
HOLDER_PRINT("LockCheckConflicts: resolved", holder);
PROCLOCK_PRINT("LockCheckConflicts: resolved", holder);
return STATUS_OK;
}
HOLDER_PRINT("LockCheckConflicts: conflicting", holder);
PROCLOCK_PRINT("LockCheckConflicts: conflicting", holder);
return STATUS_FOUND;
}
@ -814,13 +814,13 @@ static void
LockCountMyLocks(SHMEM_OFFSET lockOffset, PGPROC *proc, int *myHolding)
{
SHM_QUEUE *procHolders = &(proc->procHolders);
HOLDER *holder;
PROCLOCK *holder;
int i;
MemSet(myHolding, 0, MAX_LOCKMODES * sizeof(int));
holder = (HOLDER *) SHMQueueNext(procHolders, procHolders,
offsetof(HOLDER, procLink));
holder = (PROCLOCK *) SHMQueueNext(procHolders, procHolders,
offsetof(PROCLOCK, procLink));
while (holder)
{
@ -830,8 +830,8 @@ LockCountMyLocks(SHMEM_OFFSET lockOffset, PGPROC *proc, int *myHolding)
myHolding[i] += holder->holding[i];
}
holder = (HOLDER *) SHMQueueNext(procHolders, &holder->procLink,
offsetof(HOLDER, procLink));
holder = (PROCLOCK *) SHMQueueNext(procHolders, &holder->procLink,
offsetof(PROCLOCK, procLink));
}
}
@ -843,7 +843,7 @@ LockCountMyLocks(SHMEM_OFFSET lockOffset, PGPROC *proc, int *myHolding)
* and have its waitLock/waitHolder fields cleared. That's not done here.
*/
void
GrantLock(LOCK *lock, HOLDER *holder, LOCKMODE lockmode)
GrantLock(LOCK *lock, PROCLOCK *holder, LOCKMODE lockmode)
{
lock->nGranted++;
lock->granted[lockmode]++;
@ -868,7 +868,7 @@ GrantLock(LOCK *lock, HOLDER *holder, LOCKMODE lockmode)
*/
static int
WaitOnLock(LOCKMETHOD lockmethod, LOCKMODE lockmode,
LOCK *lock, HOLDER *holder)
LOCK *lock, PROCLOCK *holder)
{
LOCKMETHODTABLE *lockMethodTable = LockMethodTable[lockmethod];
char *new_status,
@ -984,8 +984,8 @@ LockRelease(LOCKMETHOD lockmethod, LOCKTAG *locktag,
LOCK *lock;
LWLockId masterLock;
LOCKMETHODTABLE *lockMethodTable;
HOLDER *holder;
HOLDERTAG holdertag;
PROCLOCK *holder;
PROCLOCKTAG holdertag;
HTAB *holderTable;
bool wakeupNeeded = false;
@ -1031,14 +1031,14 @@ LockRelease(LOCKMETHOD lockmethod, LOCKTAG *locktag,
/*
* Find the holder entry for this holder.
*/
MemSet(&holdertag, 0, sizeof(HOLDERTAG)); /* must clear padding,
MemSet(&holdertag, 0, sizeof(PROCLOCKTAG)); /* must clear padding,
* needed */
holdertag.lock = MAKE_OFFSET(lock);
holdertag.proc = MAKE_OFFSET(MyProc);
TransactionIdStore(xid, &holdertag.xid);
holderTable = lockMethodTable->holderHash;
holder = (HOLDER *) hash_search(holderTable,
holder = (PROCLOCK *) hash_search(holderTable,
(void *) &holdertag,
HASH_FIND_SAVE, NULL);
if (!holder)
@ -1052,7 +1052,7 @@ LockRelease(LOCKMETHOD lockmethod, LOCKTAG *locktag,
elog(WARNING, "LockRelease: holder table corrupted");
return FALSE;
}
HOLDER_PRINT("LockRelease: found", holder);
PROCLOCK_PRINT("LockRelease: found", holder);
/*
* Check that we are actually holding a lock of the type we want to
@ -1060,7 +1060,7 @@ LockRelease(LOCKMETHOD lockmethod, LOCKTAG *locktag,
*/
if (!(holder->holding[lockmode] > 0))
{
HOLDER_PRINT("LockRelease: WRONGTYPE", holder);
PROCLOCK_PRINT("LockRelease: WRONGTYPE", holder);
Assert(holder->holding[lockmode] >= 0);
LWLockRelease(masterLock);
elog(WARNING, "LockRelease: you don't own a lock of type %s",
@ -1128,7 +1128,7 @@ LockRelease(LOCKMETHOD lockmethod, LOCKTAG *locktag,
*/
holder->holding[lockmode]--;
holder->nHolding--;
HOLDER_PRINT("LockRelease: updated", holder);
PROCLOCK_PRINT("LockRelease: updated", holder);
Assert((holder->nHolding >= 0) && (holder->holding[lockmode] >= 0));
/*
@ -1137,10 +1137,10 @@ LockRelease(LOCKMETHOD lockmethod, LOCKTAG *locktag,
*/
if (holder->nHolding == 0)
{
HOLDER_PRINT("LockRelease: deleting", holder);
PROCLOCK_PRINT("LockRelease: deleting", holder);
SHMQueueDelete(&holder->lockLink);
SHMQueueDelete(&holder->procLink);
holder = (HOLDER *) hash_search(holderTable,
holder = (PROCLOCK *) hash_search(holderTable,
(void *) &holder,
HASH_REMOVE_SAVED, NULL);
if (!holder)
@ -1177,8 +1177,8 @@ LockReleaseAll(LOCKMETHOD lockmethod, PGPROC *proc,
bool allxids, TransactionId xid)
{
SHM_QUEUE *procHolders = &(proc->procHolders);
HOLDER *holder;
HOLDER *nextHolder;
PROCLOCK *holder;
PROCLOCK *nextHolder;
LWLockId masterLock;
LOCKMETHODTABLE *lockMethodTable;
int i,
@ -1204,16 +1204,16 @@ LockReleaseAll(LOCKMETHOD lockmethod, PGPROC *proc,
LWLockAcquire(masterLock, LW_EXCLUSIVE);
holder = (HOLDER *) SHMQueueNext(procHolders, procHolders,
offsetof(HOLDER, procLink));
holder = (PROCLOCK *) SHMQueueNext(procHolders, procHolders,
offsetof(PROCLOCK, procLink));
while (holder)
{
bool wakeupNeeded = false;
/* Get link first, since we may unlink/delete this holder */
nextHolder = (HOLDER *) SHMQueueNext(procHolders, &holder->procLink,
offsetof(HOLDER, procLink));
nextHolder = (PROCLOCK *) SHMQueueNext(procHolders, &holder->procLink,
offsetof(PROCLOCK, procLink));
Assert(holder->tag.proc == MAKE_OFFSET(proc));
@ -1227,7 +1227,7 @@ LockReleaseAll(LOCKMETHOD lockmethod, PGPROC *proc,
if (!allxids && !TransactionIdEquals(xid, holder->tag.xid))
goto next_item;
HOLDER_PRINT("LockReleaseAll", holder);
PROCLOCK_PRINT("LockReleaseAll", holder);
LOCK_PRINT("LockReleaseAll", lock, 0);
Assert(lock->nRequested >= 0);
Assert(lock->nGranted >= 0);
@ -1281,7 +1281,7 @@ LockReleaseAll(LOCKMETHOD lockmethod, PGPROC *proc,
}
LOCK_PRINT("LockReleaseAll: updated", lock, 0);
HOLDER_PRINT("LockReleaseAll: deleting", holder);
PROCLOCK_PRINT("LockReleaseAll: deleting", holder);
/*
* Remove the holder entry from the linked lists
@ -1292,7 +1292,7 @@ LockReleaseAll(LOCKMETHOD lockmethod, PGPROC *proc,
/*
* remove the holder entry from the hashtable
*/
holder = (HOLDER *) hash_search(lockMethodTable->holderHash,
holder = (PROCLOCK *) hash_search(lockMethodTable->holderHash,
(void *) holder,
HASH_REMOVE,
NULL);
@ -1353,7 +1353,7 @@ LockShmemSize(int maxBackends)
size += hash_estimate_size(max_table_size, sizeof(LOCK));
/* holderHash table */
size += hash_estimate_size(max_table_size, sizeof(HOLDER));
size += hash_estimate_size(max_table_size, sizeof(PROCLOCK));
/*
* Since the lockHash entry count above is only an estimate, add 10%
@ -1376,7 +1376,7 @@ DumpLocks(void)
{
PGPROC *proc;
SHM_QUEUE *procHolders;
HOLDER *holder;
PROCLOCK *holder;
LOCK *lock;
int lockmethod = DEFAULT_LOCKMETHOD;
LOCKMETHODTABLE *lockMethodTable;
@ -1395,8 +1395,8 @@ DumpLocks(void)
if (proc->waitLock)
LOCK_PRINT("DumpLocks: waiting on", proc->waitLock, 0);
holder = (HOLDER *) SHMQueueNext(procHolders, procHolders,
offsetof(HOLDER, procLink));
holder = (PROCLOCK *) SHMQueueNext(procHolders, procHolders,
offsetof(PROCLOCK, procLink));
while (holder)
{
@ -1404,11 +1404,11 @@ DumpLocks(void)
lock = (LOCK *) MAKE_PTR(holder->tag.lock);
HOLDER_PRINT("DumpLocks", holder);
PROCLOCK_PRINT("DumpLocks", holder);
LOCK_PRINT("DumpLocks", lock, 0);
holder = (HOLDER *) SHMQueueNext(procHolders, &holder->procLink,
offsetof(HOLDER, procLink));
holder = (PROCLOCK *) SHMQueueNext(procHolders, &holder->procLink,
offsetof(PROCLOCK, procLink));
}
}
@ -1419,7 +1419,7 @@ void
DumpAllLocks(void)
{
PGPROC *proc;
HOLDER *holder;
PROCLOCK *holder;
LOCK *lock;
int lockmethod = DEFAULT_LOCKMETHOD;
LOCKMETHODTABLE *lockMethodTable;
@ -1441,9 +1441,9 @@ DumpAllLocks(void)
LOCK_PRINT("DumpAllLocks: waiting on", proc->waitLock, 0);
hash_seq_init(&status, holderTable);
while ((holder = (HOLDER *) hash_seq_search(&status)) != NULL)
while ((holder = (PROCLOCK *) hash_seq_search(&status)) != NULL)
{
HOLDER_PRINT("DumpAllLocks", holder);
PROCLOCK_PRINT("DumpAllLocks", holder);
if (holder->tag.lock)
{

4
src/backend/storage/lmgr/proc.c
View File

@ -8,7 +8,7 @@
*
*
* IDENTIFICATION
* $Header: /cvsroot/pgsql/src/backend/storage/lmgr/proc.c,v 1.123 2002/07/18 23:06:20 momjian Exp $
* $Header: /cvsroot/pgsql/src/backend/storage/lmgr/proc.c,v 1.124 2002/07/19 00:17:40 momjian Exp $
*
*-------------------------------------------------------------------------
*/
@ -501,7 +501,7 @@ int
ProcSleep(LOCKMETHODTABLE *lockMethodTable,
LOCKMODE lockmode,
LOCK *lock,
HOLDER *holder)
PROCLOCK *holder)
{
LWLockId masterLock = lockMethodTable->masterLock;
PROC_QUEUE *waitQueue = &(lock->waitProcs);

30
src/include/storage/lock.h
View File

@ -7,7 +7,7 @@
* Portions Copyright (c) 1996-2002, PostgreSQL Global Development Group
* Portions Copyright (c) 1994, Regents of the University of California
*
* $Id: lock.h,v 1.62 2002/07/18 23:06:20 momjian Exp $
* $Id: lock.h,v 1.63 2002/07/19 00:17:40 momjian Exp $
*
*-------------------------------------------------------------------------
*/
@ -130,7 +130,7 @@ typedef struct LOCKTAG
* tag -- uniquely identifies the object being locked
* grantMask -- bitmask for all lock types currently granted on this object.
* waitMask -- bitmask for all lock types currently awaited on this object.
* lockHolders -- list of HOLDER objects for this lock.
* lockHolders -- list of PROCLOCK objects for this lock.
* waitProcs -- queue of processes waiting for this lock.
* requested -- count of each lock type currently requested on the lock
* (includes requests already granted!!).
@ -146,7 +146,7 @@ typedef struct LOCK
/* data */
int grantMask; /* bitmask for lock types already granted */
int waitMask; /* bitmask for lock types awaited */
SHM_QUEUE lockHolders; /* list of HOLDER objects assoc. with lock */
SHM_QUEUE lockHolders; /* list of PROCLOCK objects assoc. with lock */
PROC_QUEUE waitProcs; /* list of PGPROC objects waiting on lock */
int requested[MAX_LOCKMODES]; /* counts of requested
* locks */
@ -163,8 +163,8 @@ typedef struct LOCK
* on the same lockable object. We need to store some per-holder information
* for each such holder (or would-be holder).
*
* HOLDERTAG is the key information needed to look up a HOLDER item in the
* holder hashtable. A HOLDERTAG value uniquely identifies a lock holder.
* PROCLOCKTAG is the key information needed to look up a PROCLOCK item in the
* holder hashtable. A PROCLOCKTAG value uniquely identifies a lock holder.
*
* There are two possible kinds of holder tags: a transaction (identified
* both by the PGPROC of the backend running it, and the xact's own ID) and
@ -180,32 +180,32 @@ typedef struct LOCK
* Otherwise, holder objects whose counts have gone to zero are recycled
* as soon as convenient.
*
* Each HOLDER object is linked into lists for both the associated LOCK object
* and the owning PGPROC object. Note that the HOLDER is entered into these
* Each PROCLOCK object is linked into lists for both the associated LOCK object
* and the owning PGPROC object. Note that the PROCLOCK is entered into these
* lists as soon as it is created, even if no lock has yet been granted.
* A PGPROC that is waiting for a lock to be granted will also be linked into
* the lock's waitProcs queue.
*/
typedef struct HOLDERTAG
typedef struct PROCLOCKTAG
{
SHMEM_OFFSET lock; /* link to per-lockable-object information */
SHMEM_OFFSET proc; /* link to PGPROC of owning backend */
TransactionId xid; /* xact ID, or InvalidTransactionId */
} HOLDERTAG;
} PROCLOCKTAG;
typedef struct HOLDER
typedef struct PROCLOCK
{
/* tag */
HOLDERTAG tag; /* unique identifier of holder object */
PROCLOCKTAG tag; /* unique identifier of holder object */
/* data */
int holding[MAX_LOCKMODES]; /* count of locks currently held */
int nHolding; /* total of holding[] array */
SHM_QUEUE lockLink; /* list link for lock's list of holders */
SHM_QUEUE procLink; /* list link for process's list of holders */
} HOLDER;
} PROCLOCK;
#define HOLDER_LOCKMETHOD(holder) \
#define PROCLOCK_LOCKMETHOD(holder) \
(((LOCK *) MAKE_PTR((holder).tag.lock))->tag.lockmethod)
@ -225,9 +225,9 @@ extern bool LockReleaseAll(LOCKMETHOD lockmethod, PGPROC *proc,
bool allxids, TransactionId xid);
extern int LockCheckConflicts(LOCKMETHODTABLE *lockMethodTable,
LOCKMODE lockmode,
LOCK *lock, HOLDER *holder, PGPROC *proc,
LOCK *lock, PROCLOCK *holder, PGPROC *proc,
int *myHolding);
extern void GrantLock(LOCK *lock, HOLDER *holder, LOCKMODE lockmode);
extern void GrantLock(LOCK *lock, PROCLOCK *holder, LOCKMODE lockmode);
extern void RemoveFromWaitQueue(PGPROC *proc);
extern int LockShmemSize(int maxBackends);
extern bool DeadLockCheck(PGPROC *proc);

8
src/include/storage/proc.h
View File

@ -7,7 +7,7 @@
* Portions Copyright (c) 1996-2002, PostgreSQL Global Development Group
* Portions Copyright (c) 1994, Regents of the University of California
*
* $Id: proc.h,v 1.58 2002/07/13 01:02:14 momjian Exp $
* $Id: proc.h,v 1.59 2002/07/19 00:17:40 momjian Exp $
*
*-------------------------------------------------------------------------
*/
@ -61,12 +61,12 @@ struct PGPROC
/* Info about lock the process is currently waiting for, if any. */
/* waitLock and waitHolder are NULL if not currently waiting. */
LOCK *waitLock; /* Lock object we're sleeping on ... */
HOLDER *waitHolder; /* Per-holder info for awaited lock */
PROCLOCK *waitHolder; /* Per-holder info for awaited lock */
LOCKMODE waitLockMode; /* type of lock we're waiting for */
LOCKMASK heldLocks; /* bitmask for lock types already held on
* this lock object by this backend */
SHM_QUEUE procHolders; /* list of HOLDER objects for locks held
SHM_QUEUE procHolders; /* list of PROCLOCK objects for locks held
* or awaited by this backend */
};
@ -101,7 +101,7 @@ extern void ProcReleaseLocks(bool isCommit);
extern void ProcQueueInit(PROC_QUEUE *queue);
extern int ProcSleep(LOCKMETHODTABLE *lockMethodTable, LOCKMODE lockmode,
LOCK *lock, HOLDER *holder);
LOCK *lock, PROCLOCK *holder);
extern PGPROC *ProcWakeup(PGPROC *proc, int errType);
extern void ProcLockWakeup(LOCKMETHODTABLE *lockMethodTable, LOCK *lock);
extern bool LockWaitCancel(void);