Hello Alexander,
On 2024-Apr-03, Alexander Korotkov wrote:
> Regarding the shmem data structure for LSN waiters. I didn't pick
> LWLock or ConditionVariable, because I needed the ability to wake up
> only those waiters whose LSN is already replayed. In my experience
> waking up a process is way slower than scanning a short flat array.
I agree, but I think that's unrelated to what I was saying, which is
just the patch I attach here.
> However, I agree that when the number of waiters is very high and flat
> array may become a problem. It seems that the pairing heap is not
> hard to use for shmem structures. The only memory allocation call in
> paritingheap.c is in pairingheap_allocate(). So, it's only needed to
> be able to initialize the pairing heap in-place, and it will be fine
> for shmem.
Ok.
With the code as it stands today, everything in WaitLSNState apart from
the pairing heap is accessed without any locking. I think this is at
least partly OK because each backend only accesses its own entry; but it
deserves a comment. Or maybe something more, because WaitLSNSetLatches
does modify the entry for other backends. (Admittedly, this could only
happens for backends that are already sleeping, and it only happens
with the lock acquired, so it's probably okay. But clearly it deserves
a comment.)
Don't we need to WaitLSNCleanup() during error recovery or something?
--
Álvaro Herrera 48°01'N 7°57'E — https://www.EnterpriseDB.com/
"World domination is proceeding according to plan" (Andrew Morton)
>From 4079dc6a6a6893055b32dee75f178b324bbaef77 Mon Sep 17 00:00:00 2001
From: Alvaro Herrera <alvhe...@alvh.no-ip.org>
Date: Wed, 3 Apr 2024 18:35:15 +0200
Subject: [PATCH] Use an LWLock instead of a spinlock in waitlsn.c
---
src/backend/commands/waitlsn.c | 15 +++++++--------
src/backend/utils/activity/wait_event_names.txt | 1 +
src/include/commands/waitlsn.h | 5 +----
src/include/storage/lwlocklist.h | 1 +
4 files changed, 10 insertions(+), 12 deletions(-)
diff --git a/src/backend/commands/waitlsn.c b/src/backend/commands/waitlsn.c
index 51a34d422e..a57b818a2d 100644
--- a/src/backend/commands/waitlsn.c
+++ b/src/backend/commands/waitlsn.c
@@ -58,7 +58,6 @@ WaitLSNShmemInit(void)
&found);
if (!found)
{
- SpinLockInit(&waitLSN->waitersHeapMutex);
pg_atomic_init_u64(&waitLSN->minWaitedLSN, PG_UINT64_MAX);
pairingheap_initialize(&waitLSN->waitersHeap, lsn_cmp, NULL);
memset(&waitLSN->procInfos, 0, MaxBackends * sizeof(WaitLSNProcInfo));
@@ -115,13 +114,13 @@ addLSNWaiter(XLogRecPtr lsn)
procInfo->procnum = MyProcNumber;
procInfo->waitLSN = lsn;
- SpinLockAcquire(&waitLSN->waitersHeapMutex);
+ LWLockAcquire(WaitLSNLock, LW_EXCLUSIVE);
pairingheap_add(&waitLSN->waitersHeap, &procInfo->phNode);
procInfo->inHeap = true;
updateMinWaitedLSN();
- SpinLockRelease(&waitLSN->waitersHeapMutex);
+ LWLockRelease(WaitLSNLock);
}
/*
@@ -132,11 +131,11 @@ deleteLSNWaiter(void)
{
WaitLSNProcInfo *procInfo = &waitLSN->procInfos[MyProcNumber];
- SpinLockAcquire(&waitLSN->waitersHeapMutex);
+ LWLockAcquire(WaitLSNLock, LW_EXCLUSIVE);
if (!procInfo->inHeap)
{
- SpinLockRelease(&waitLSN->waitersHeapMutex);
+ LWLockRelease(WaitLSNLock);
return;
}
@@ -144,7 +143,7 @@ deleteLSNWaiter(void)
procInfo->inHeap = false;
updateMinWaitedLSN();
- SpinLockRelease(&waitLSN->waitersHeapMutex);
+ LWLockRelease(WaitLSNLock);
}
/*
@@ -160,7 +159,7 @@ WaitLSNSetLatches(XLogRecPtr currentLSN)
wakeUpProcNums = palloc(sizeof(int) * MaxBackends);
- SpinLockAcquire(&waitLSN->waitersHeapMutex);
+ LWLockAcquire(WaitLSNLock, LW_EXCLUSIVE);
/*
* Iterate the pairing heap of waiting processes till we find LSN not yet
@@ -182,7 +181,7 @@ WaitLSNSetLatches(XLogRecPtr currentLSN)
updateMinWaitedLSN();
- SpinLockRelease(&waitLSN->waitersHeapMutex);
+ LWLockRelease(WaitLSNLock);
/*
* Set latches for processes, whose waited LSNs are already replayed. This
diff --git a/src/backend/utils/activity/wait_event_names.txt b/src/backend/utils/activity/wait_event_names.txt
index 0d288d6b3d..ec43a78d60 100644
--- a/src/backend/utils/activity/wait_event_names.txt
+++ b/src/backend/utils/activity/wait_event_names.txt
@@ -330,6 +330,7 @@ WALSummarizer "Waiting to read or update WAL summarization state."
DSMRegistry "Waiting to read or update the dynamic shared memory registry."
InjectionPoint "Waiting to read or update information related to injection points."
SerialControl "Waiting to read or update shared <filename>pg_serial</filename> state."
+WaitLSN "Waiting to read or update shared Wait-for-LSN state."
#
# END OF PREDEFINED LWLOCKS (DO NOT CHANGE THIS LINE)
diff --git a/src/include/commands/waitlsn.h b/src/include/commands/waitlsn.h
index 0d80248682..b3d9eed64d 100644
--- a/src/include/commands/waitlsn.h
+++ b/src/include/commands/waitlsn.h
@@ -55,13 +55,10 @@ typedef struct WaitLSNState
/*
* A pairing heap of waiting processes order by LSN values (least LSN is
- * on top).
+ * on top). Protected by WaitLSNLock.
*/
pairingheap waitersHeap;
- /* A mutex protecting the pairing heap above */
- slock_t waitersHeapMutex;
-
/* An array with per-process information, indexed by the process number */
WaitLSNProcInfo procInfos[FLEXIBLE_ARRAY_MEMBER];
} WaitLSNState;
diff --git a/src/include/storage/lwlocklist.h b/src/include/storage/lwlocklist.h
index 85f6568b9e..c2bab5a794 100644
--- a/src/include/storage/lwlocklist.h
+++ b/src/include/storage/lwlocklist.h
@@ -83,3 +83,4 @@ PG_LWLOCK(49, WALSummarizer)
PG_LWLOCK(50, DSMRegistry)
PG_LWLOCK(51, InjectionPoint)
PG_LWLOCK(52, SerialControl)
+PG_LWLOCK(53, WaitLSN)
--
2.39.2
