From 6d62b5e5ee4899fd7f979063f471be079951cfa4 Mon Sep 17 00:00:00 2001 From: Nathan Bossart Date: Tue, 8 Sep 2026 12:40:05 -0500 Subject: [PATCH v3 2/6] Convert SISeg->maxMsgNum to an atomic variable. Currently, this variable is a uint32 protected by a spinlock. The spinlock exists only to provide memory barriers, so by converting the variable to an atomic and using the barrier-providing accessors in the spinlock's place, we can remove the spinlock. --- src/backend/storage/ipc/sinvaladt.c | 51 ++++++++++------------------- 1 file changed, 17 insertions(+), 34 deletions(-) diff --git a/src/backend/storage/ipc/sinvaladt.c b/src/backend/storage/ipc/sinvaladt.c index 4e6f9a84375..5c0cb317101 100644 --- a/src/backend/storage/ipc/sinvaladt.c +++ b/src/backend/storage/ipc/sinvaladt.c @@ -24,7 +24,6 @@ #include "storage/procsignal.h" #include "storage/shmem.h" #include "storage/sinvaladt.h" -#include "storage/spin.h" #include "storage/subsystems.h" /* @@ -87,19 +86,10 @@ * has no need to touch anyone's ProcState, except in the infrequent cases * when SICleanupQueue is needed. The only point of overlap is that * the writer wants to change maxMsgNum while readers need to read it. - * We deal with that by having a spinlock that readers must take for just - * long enough to read maxMsgNum, while writers take it for just long enough - * to write maxMsgNum. (The exact rule is that you need the spinlock to - * read maxMsgNum if you are not holding SInvalWriteLock, and you need the - * spinlock to write maxMsgNum unless you are holding both locks.) - * - * Note: since maxMsgNum is a uint32 and hence presumably atomically readable/ - * writable, the spinlock might seem unnecessary. The reason it is needed - * is to provide a memory barrier: we need to be sure that messages written - * to the array are actually there before maxMsgNum is increased, and that - * readers will see that data after fetching maxMsgNum. Multiprocessors - * that have weak memory-ordering guarantees can fail without the memory - * barrier instructions that are included in the spinlock sequences. + * We deal with that by making maxMsgNum an atomic variable. (The exact rule + * is that you need to use a barrier-providing accessor to read maxMsgNum if + * you are not holding SInvalWriteLock, and you need a barrier-providing + * accessor to write maxMsgNum unless you are holding both locks.) */ @@ -169,11 +159,9 @@ typedef struct SISeg * General state information */ uint32 minMsgNum; /* oldest message still needed */ - uint32 maxMsgNum; /* next message number to be assigned */ + pg_atomic_uint32 maxMsgNum; /* next message number to be assigned */ uint32 nextThreshold; /* # of messages to call SICleanupQueue */ - slock_t msgnumLock; /* spinlock protecting maxMsgNum */ - /* * Circular buffer holding shared-inval messages */ @@ -244,11 +232,10 @@ SharedInvalShmemInit(void *arg) { int i; - /* Clear message counters, init spinlock */ + /* Clear message counters */ shmInvalBuffer->minMsgNum = 0; - shmInvalBuffer->maxMsgNum = 0; + pg_atomic_init_u32(&shmInvalBuffer->maxMsgNum, 0); shmInvalBuffer->nextThreshold = CLEANUP_MIN; - SpinLockInit(&shmInvalBuffer->msgnumLock); /* The buffer[] array is initially all unused, so we need not fill it */ @@ -306,7 +293,7 @@ SharedInvalBackendInit(bool sendOnly) /* mark myself active, with all extant messages already read */ stateP->procPid = MyProcPid; - stateP->nextMsgNum = segP->maxMsgNum; + stateP->nextMsgNum = pg_atomic_read_u32(&segP->maxMsgNum); stateP->resetState = false; stateP->signaled = false; stateP->hasMessages = false; @@ -402,7 +389,7 @@ SIInsertDataEntries(const SharedInvalidationMessage *data, int n) */ for (;;) { - numMsgs = segP->maxMsgNum - segP->minMsgNum; + numMsgs = pg_atomic_read_u32(&segP->maxMsgNum) - segP->minMsgNum; if (numMsgs + nthistime > MAXNUMMESSAGES || numMsgs >= segP->nextThreshold) SICleanupQueue(true, nthistime); @@ -413,17 +400,15 @@ SIInsertDataEntries(const SharedInvalidationMessage *data, int n) /* * Insert new message(s) into proper slot of circular buffer */ - max = segP->maxMsgNum; + max = pg_atomic_read_u32(&segP->maxMsgNum); while (nthistime-- > 0) { segP->buffer[max % MAXNUMMESSAGES] = *data++; max++; } - /* Update current value of maxMsgNum using spinlock */ - SpinLockAcquire(&segP->msgnumLock); - segP->maxMsgNum = max; - SpinLockRelease(&segP->msgnumLock); + /* Update current value of maxMsgNum using barrier */ + pg_atomic_write_membarrier_u32(&segP->maxMsgNum, max); /* * Now that the maxMsgNum change is globally visible, we give everyone @@ -509,10 +494,8 @@ SIGetDataEntries(SharedInvalidationMessage *data, int datasize) */ stateP->hasMessages = false; - /* Fetch current value of maxMsgNum using spinlock */ - SpinLockAcquire(&segP->msgnumLock); - max = segP->maxMsgNum; - SpinLockRelease(&segP->msgnumLock); + /* Fetch current value of maxMsgNum using barrier */ + max = pg_atomic_read_membarrier_u32(&segP->maxMsgNum); if (stateP->resetState) { @@ -602,7 +585,7 @@ SICleanupQueue(bool callerHasWriteLock, int minFree) * wrap around. A threshold of zero disables its check, just as a * negative one would. */ - min = segP->maxMsgNum; + min = pg_atomic_read_u32(&segP->maxMsgNum); minsig = (min > SIG_THRESHOLD) ? min - SIG_THRESHOLD : 0; lowbound = (min + minFree > MAXNUMMESSAGES) ? min + minFree - MAXNUMMESSAGES : 0; @@ -649,7 +632,7 @@ SICleanupQueue(bool callerHasWriteLock, int minFree) if (min >= MSGNUMWRAPAROUND) { segP->minMsgNum -= MSGNUMWRAPAROUND; - segP->maxMsgNum -= MSGNUMWRAPAROUND; + pg_atomic_fetch_sub_u32(&segP->maxMsgNum, MSGNUMWRAPAROUND); for (i = 0; i < segP->numProcs; i++) segP->procState[segP->pgprocnos[i]].nextMsgNum -= MSGNUMWRAPAROUND; } @@ -658,7 +641,7 @@ SICleanupQueue(bool callerHasWriteLock, int minFree) * Determine how many messages are still in the queue, and set the * threshold at which we should repeat SICleanupQueue(). */ - numMsgs = segP->maxMsgNum - segP->minMsgNum; + numMsgs = pg_atomic_read_u32(&segP->maxMsgNum) - segP->minMsgNum; if (numMsgs < CLEANUP_MIN) segP->nextThreshold = CLEANUP_MIN; else -- 2.55.0