proc.c

postgresql8.3.4源码,开源数据库

/*-------------------------------------------------------------------------
 *
 * proc.c
 *	  routines to manage per-process shared memory data structure
 *
 * Portions Copyright (c) 1996-2008, PostgreSQL Global Development Group
 * Portions Copyright (c) 1994, Regents of the University of California
 *
 *
 * IDENTIFICATION
 *	  $PostgreSQL: pgsql/src/backend/storage/lmgr/proc.c,v 1.199 2008/01/26 19:55:08 tgl Exp $
 *
 *-------------------------------------------------------------------------
 */
/*
 * Interface (a):
 *		ProcSleep(), ProcWakeup(),
 *		ProcQueueAlloc() -- create a shm queue for sleeping processes
 *		ProcQueueInit() -- create a queue without allocing memory
 *
 * Waiting for a lock causes the backend to be put to sleep.  Whoever releases
 * the lock wakes the process up again (and gives it an error code so it knows
 * whether it was awoken on an error condition).
 *
 * Interface (b):
 *
 * ProcReleaseLocks -- frees the locks associated with current transaction
 *
 * ProcKill -- destroys the shared memory state (and locks)
 * associated with the process.
 */
#include "postgres.h"

#include <signal.h>
#include <unistd.h>
#include <sys/time.h>

#include "access/transam.h"
#include "access/xact.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
#include "storage/ipc.h"
#include "storage/lmgr.h"
#include "storage/proc.h"
#include "storage/procarray.h"
#include "storage/spin.h"


/* GUC variables */
int			DeadlockTimeout = 1000;
int			StatementTimeout = 0;
bool		log_lock_waits = false;

/* Pointer to this process's PGPROC struct, if any */
PGPROC	   *MyProc = NULL;

/*
 * This spinlock protects the freelist of recycled PGPROC structures.
 * We cannot use an LWLock because the LWLock manager depends on already
 * having a PGPROC and a wait semaphore!  But these structures are touched
 * relatively infrequently (only at backend startup or shutdown) and not for
 * very long, so a spinlock is okay.
 */
NON_EXEC_STATIC slock_t *ProcStructLock = NULL;

/* Pointers to shared-memory structures */
NON_EXEC_STATIC PROC_HDR *ProcGlobal = NULL;
NON_EXEC_STATIC PGPROC *AuxiliaryProcs = NULL;

/* If we are waiting for a lock, this points to the associated LOCALLOCK */
static LOCALLOCK *lockAwaited = NULL;

/* Mark these volatile because they can be changed by signal handler */
static volatile bool statement_timeout_active = false;
static volatile bool deadlock_timeout_active = false;
static volatile DeadLockState deadlock_state = DS_NOT_YET_CHECKED;
volatile bool cancel_from_timeout = false;

/* timeout_start_time is set when log_lock_waits is true */
static TimestampTz timeout_start_time;

/* statement_fin_time is valid only if statement_timeout_active is true */
static TimestampTz statement_fin_time;


static void RemoveProcFromArray(int code, Datum arg);
static void ProcKill(int code, Datum arg);
static void AuxiliaryProcKill(int code, Datum arg);
static bool CheckStatementTimeout(void);


/*
 * Report shared-memory space needed by InitProcGlobal.
 */
Size
ProcGlobalShmemSize(void)
{
	Size		size = 0;

	/* ProcGlobal */
	size = add_size(size, sizeof(PROC_HDR));
	/* AuxiliaryProcs */
	size = add_size(size, mul_size(NUM_AUXILIARY_PROCS, sizeof(PGPROC)));
	/* MyProcs, including autovacuum */
	size = add_size(size, mul_size(MaxBackends, sizeof(PGPROC)));
	/* ProcStructLock */
	size = add_size(size, sizeof(slock_t));

	return size;
}

/*
 * Report number of semaphores needed by InitProcGlobal.
 */
int
ProcGlobalSemas(void)
{
	/*
	 * We need a sema per backend (including autovacuum), plus one for each
	 * auxiliary process.
	 */
	return MaxBackends + NUM_AUXILIARY_PROCS;
}

/*
 * InitProcGlobal -
 *	  Initialize the global process table during postmaster or standalone
 *	  backend startup.
 *
 *	  We also create all the per-process semaphores we will need to support
 *	  the requested number of backends.  We used to allocate semaphores
 *	  only when backends were actually started up, but that is bad because
 *	  it lets Postgres fail under load --- a lot of Unix systems are
 *	  (mis)configured with small limits on the number of semaphores, and
 *	  running out when trying to start another backend is a common failure.
 *	  So, now we grab enough semaphores to support the desired max number
 *	  of backends immediately at initialization --- if the sysadmin has set
 *	  MaxConnections or autovacuum_max_workers higher than his kernel will
 *	  support, he'll find out sooner rather than later.
 *
 *	  Another reason for creating semaphores here is that the semaphore
 *	  implementation typically requires us to create semaphores in the
 *	  postmaster, not in backends.
 *
 * Note: this is NOT called by individual backends under a postmaster,
 * not even in the EXEC_BACKEND case.  The ProcGlobal and AuxiliaryProcs
 * pointers must be propagated specially for EXEC_BACKEND operation.
 */
void
InitProcGlobal(void)
{
	PGPROC	   *procs;
	int			i;
	bool		found;

	/* Create the ProcGlobal shared structure */
	ProcGlobal = (PROC_HDR *)
		ShmemInitStruct("Proc Header", sizeof(PROC_HDR), &found);
	Assert(!found);

	/*
	 * Create the PGPROC structures for auxiliary (bgwriter) processes, too.
	 * These do not get linked into the freeProcs list.
	 */
	AuxiliaryProcs = (PGPROC *)
		ShmemInitStruct("AuxiliaryProcs", NUM_AUXILIARY_PROCS * sizeof(PGPROC),
						&found);
	Assert(!found);

	/*
	 * Initialize the data structures.
	 */
	ProcGlobal->freeProcs = INVALID_OFFSET;
	ProcGlobal->autovacFreeProcs = INVALID_OFFSET;

	ProcGlobal->spins_per_delay = DEFAULT_SPINS_PER_DELAY;

	/*
	 * Pre-create the PGPROC structures and create a semaphore for each.
	 */
	procs = (PGPROC *) ShmemAlloc((MaxConnections) * sizeof(PGPROC));
	if (!procs)
		ereport(FATAL,
				(errcode(ERRCODE_OUT_OF_MEMORY),
				 errmsg("out of shared memory")));
	MemSet(procs, 0, MaxConnections * sizeof(PGPROC));
	for (i = 0; i < MaxConnections; i++)
	{
		PGSemaphoreCreate(&(procs[i].sem));
		procs[i].links.next = ProcGlobal->freeProcs;
		ProcGlobal->freeProcs = MAKE_OFFSET(&procs[i]);
	}

	procs = (PGPROC *) ShmemAlloc((autovacuum_max_workers) * sizeof(PGPROC));
	if (!procs)
		ereport(FATAL,
				(errcode(ERRCODE_OUT_OF_MEMORY),
				 errmsg("out of shared memory")));
	MemSet(procs, 0, autovacuum_max_workers * sizeof(PGPROC));
	for (i = 0; i < autovacuum_max_workers; i++)
	{
		PGSemaphoreCreate(&(procs[i].sem));
		procs[i].links.next = ProcGlobal->autovacFreeProcs;
		ProcGlobal->autovacFreeProcs = MAKE_OFFSET(&procs[i]);
	}

	MemSet(AuxiliaryProcs, 0, NUM_AUXILIARY_PROCS * sizeof(PGPROC));
	for (i = 0; i < NUM_AUXILIARY_PROCS; i++)
	{
		AuxiliaryProcs[i].pid = 0;		/* marks auxiliary proc as not in use */
		PGSemaphoreCreate(&(AuxiliaryProcs[i].sem));
	}

	/* Create ProcStructLock spinlock, too */
	ProcStructLock = (slock_t *) ShmemAlloc(sizeof(slock_t));
	SpinLockInit(ProcStructLock);
}

/*
 * InitProcess -- initialize a per-process data structure for this backend
 */
void
InitProcess(void)
{
	/* use volatile pointer to prevent code rearrangement */
	volatile PROC_HDR *procglobal = ProcGlobal;
	SHMEM_OFFSET myOffset;
	int			i;

	/*
	 * ProcGlobal should be set up already (if we are a backend, we inherit
	 * this by fork() or EXEC_BACKEND mechanism from the postmaster).
	 */
	if (procglobal == NULL)
		elog(PANIC, "proc header uninitialized");

	if (MyProc != NULL)
		elog(ERROR, "you already exist");

	/*
	 * Try to get a proc struct from the free list.  If this fails, we must be
	 * out of PGPROC structures (not to mention semaphores).
	 *
	 * While we are holding the ProcStructLock, also copy the current shared
	 * estimate of spins_per_delay to local storage.
	 */
	SpinLockAcquire(ProcStructLock);

	set_spins_per_delay(procglobal->spins_per_delay);

	if (IsAutoVacuumWorkerProcess())
		myOffset = procglobal->autovacFreeProcs;
	else
		myOffset = procglobal->freeProcs;

	if (myOffset != INVALID_OFFSET)
	{
		MyProc = (PGPROC *) MAKE_PTR(myOffset);
		if (IsAutoVacuumWorkerProcess())
			procglobal->autovacFreeProcs = MyProc->links.next;
		else
			procglobal->freeProcs = MyProc->links.next;
		SpinLockRelease(ProcStructLock);
	}
	else
	{
		/*
		 * If we reach here, all the PGPROCs are in use.  This is one of the
		 * possible places to detect "too many backends", so give the standard
		 * error message.  XXX do we need to give a different failure message
		 * in the autovacuum case?
		 */
		SpinLockRelease(ProcStructLock);
		ereport(FATAL,
				(errcode(ERRCODE_TOO_MANY_CONNECTIONS),
				 errmsg("sorry, too many clients already")));
	}

	/*
	 * Initialize all fields of MyProc, except for the semaphore which was
	 * prepared for us by InitProcGlobal.
	 */
	SHMQueueElemInit(&(MyProc->links));
	MyProc->waitStatus = STATUS_OK;
	MyProc->lxid = InvalidLocalTransactionId;
	MyProc->xid = InvalidTransactionId;
	MyProc->xmin = InvalidTransactionId;
	MyProc->pid = MyProcPid;
	/* backendId, databaseId and roleId will be filled in later */
	MyProc->backendId = InvalidBackendId;
	MyProc->databaseId = InvalidOid;
	MyProc->roleId = InvalidOid;
	MyProc->inCommit = false;
	MyProc->vacuumFlags = 0;
	if (IsAutoVacuumWorkerProcess())
		MyProc->vacuumFlags |= PROC_IS_AUTOVACUUM;
	MyProc->lwWaiting = false;
	MyProc->lwExclusive = false;
	MyProc->lwWaitLink = NULL;
	MyProc->waitLock = NULL;
	MyProc->waitProcLock = NULL;
	for (i = 0; i < NUM_LOCK_PARTITIONS; i++)
		SHMQueueInit(&(MyProc->myProcLocks[i]));

	/*
	 * We might be reusing a semaphore that belonged to a failed process. So
	 * be careful and reinitialize its value here.	(This is not strictly
	 * necessary anymore, but seems like a good idea for cleanliness.)
	 */
	PGSemaphoreReset(&MyProc->sem);

	/*
	 * Arrange to clean up at backend exit.
	 */
	on_shmem_exit(ProcKill, 0);

	/*
	 * Now that we have a PGPROC, we could try to acquire locks, so initialize
	 * the deadlock checker.
	 */
	InitDeadLockChecking();
}

/*
 * InitProcessPhase2 -- make MyProc visible in the shared ProcArray.
 *
 * This is separate from InitProcess because we can't acquire LWLocks until
 * we've created a PGPROC, but in the EXEC_BACKEND case there is a good deal
 * of stuff to be done before this step that will require LWLock access.
 */
void
InitProcessPhase2(void)
{
	Assert(MyProc != NULL);

	/*
	 * We should now know what database we're in, so advertise that.  (We need
	 * not do any locking here, since no other backend can yet see our
	 * PGPROC.)
	 */
	Assert(OidIsValid(MyDatabaseId));
	MyProc->databaseId = MyDatabaseId;

	/*
	 * Add our PGPROC to the PGPROC array in shared memory.
	 */
	ProcArrayAdd(MyProc);

	/*
	 * Arrange to clean that up at backend exit.
	 */
	on_shmem_exit(RemoveProcFromArray, 0);
}

/*
 * InitAuxiliaryProcess -- create a per-auxiliary-process data structure
 *
 * This is called by bgwriter and similar processes so that they will have a
 * MyProc value that's real enough to let them wait for LWLocks.  The PGPROC
 * and sema that are assigned are one of the extra ones created during
 * InitProcGlobal.
 *
 * Auxiliary processes are presently not expected to wait for real (lockmgr)
 * locks, so we need not set up the deadlock checker.  They are never added
 * to the ProcArray or the sinval messaging mechanism, either.	They also
 * don't get a VXID assigned, since this is only useful when we actually
 * hold lockmgr locks.
 */
void
InitAuxiliaryProcess(void)
{
	PGPROC	   *auxproc;
	int			proctype;
	int			i;

	/*
	 * ProcGlobal should be set up already (if we are a backend, we inherit
	 * this by fork() or EXEC_BACKEND mechanism from the postmaster).
	 */
	if (ProcGlobal == NULL || AuxiliaryProcs == NULL)
		elog(PANIC, "proc header uninitialized");

	if (MyProc != NULL)
		elog(ERROR, "you already exist");

	/*
	 * We use the ProcStructLock to protect assignment and releasing of
	 * AuxiliaryProcs entries.
	 *
	 * While we are holding the ProcStructLock, also copy the current shared
	 * estimate of spins_per_delay to local storage.
	 */
	SpinLockAcquire(ProcStructLock);

	set_spins_per_delay(ProcGlobal->spins_per_delay);

	/*
	 * Find a free auxproc ... *big* trouble if there isn't one ...
	 */
	for (proctype = 0; proctype < NUM_AUXILIARY_PROCS; proctype++)
	{
		auxproc = &AuxiliaryProcs[proctype];
		if (auxproc->pid == 0)
			break;
	}
	if (proctype >= NUM_AUXILIARY_PROCS)
	{
		SpinLockRelease(ProcStructLock);
		elog(FATAL, "all AuxiliaryProcs are in use");
	}

	/* Mark auxiliary proc as in use by me */
	/* use volatile pointer to prevent code rearrangement */
	((volatile PGPROC *) auxproc)->pid = MyProcPid;

	MyProc = auxproc;

	SpinLockRelease(ProcStructLock);

	/*
	 * Initialize all fields of MyProc, except for the semaphore which was
	 * prepared for us by InitProcGlobal.
	 */
	SHMQueueElemInit(&(MyProc->links));
	MyProc->waitStatus = STATUS_OK;
	MyProc->lxid = InvalidLocalTransactionId;
	MyProc->xid = InvalidTransactionId;
	MyProc->xmin = InvalidTransactionId;
	MyProc->backendId = InvalidBackendId;
	MyProc->databaseId = InvalidOid;
	MyProc->roleId = InvalidOid;
	MyProc->inCommit = false;
	/* we don't set the "is autovacuum" flag in the launcher */
	MyProc->vacuumFlags = 0;
	MyProc->lwWaiting = false;
	MyProc->lwExclusive = false;
	MyProc->lwWaitLink = NULL;
	MyProc->waitLock = NULL;
	MyProc->waitProcLock = NULL;
	for (i = 0; i < NUM_LOCK_PARTITIONS; i++)
		SHMQueueInit(&(MyProc->myProcLocks[i]));

	/*
	 * We might be reusing a semaphore that belonged to a failed process. So
	 * be careful and reinitialize its value here.	(This is not strictly
	 * necessary anymore, but seems like a good idea for cleanliness.)
	 */
	PGSemaphoreReset(&MyProc->sem);

	/*
	 * Arrange to clean up at process exit.
	 */
	on_shmem_exit(AuxiliaryProcKill, Int32GetDatum(proctype));
}

/*
 * Check whether there are at least N free PGPROC objects.
 *
 * Note: this is designed on the assumption that N will generally be small.
 */
bool
HaveNFreeProcs(int n)
{
	SHMEM_OFFSET offset;
	PGPROC	   *proc;

	/* use volatile pointer to prevent code rearrangement */
	volatile PROC_HDR *procglobal = ProcGlobal;

	SpinLockAcquire(ProcStructLock);

	offset = procglobal->freeProcs;

	while (n > 0 && offset != INVALID_OFFSET)
	{
		proc = (PGPROC *) MAKE_PTR(offset);
		offset = proc->links.next;
		n--;
	}

	SpinLockRelease(ProcStructLock);

	return (n <= 0);
}

/*
 * Cancel any pending wait for lock, when aborting a transaction.
 *
 * (Normally, this would only happen if we accept a cancel/die
 * interrupt while waiting; but an ereport(ERROR) while waiting is
 * within the realm of possibility, too.)
 */
void
LockWaitCancel(void)
{
	LWLockId	partitionLock;

	/* Nothing to do if we weren't waiting for a lock */
	if (lockAwaited == NULL)
		return;