On Wed, Sep 17, 2025 at 11:35:56AM +0900, Michael Paquier wrote:
> Since this commit has been merged, batta has kept failing. Here is
> the first failure:
> https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=batta&dt=2025-09-12%2002%3A05%3A01
Thanks for bringing this to my attention.
> And here is the backtrace:
> #0 0x000055fcdf6bc97a in NumLWLocksForNamedTranches () at lwlock.c:385
> 385 numLocks += NamedLWLockTrancheRequestArray[i].num_lwlocks;
> (gdb) bt
> #0 0x000055fcdf6bc97a in NumLWLocksForNamedTranches () at lwlock.c:385
> #1 0x000055fcdf6bc9b3 in LWLockShmemSize () at lwlock.c:400
> #2 0x000055fcdf65bda5 in CalculateShmemSize (num_semaphores=0x7ffcaf7a78e4)
> at ipci.c:130
> #3 0x000055fcdf65c0b1 in CreateSharedMemoryAndSemaphores () at ipci.c:210
> #4 0x000055fcdf42830c in PostmasterStateMachine () at postmaster.c:3223
> #5 0x000055fcdf42703f in process_pm_child_exit () at postmaster.c:2558
> #6 0x000055fcdf425729 in ServerLoop () at postmaster.c:1696
> #7 0x000055fcdf424be1 in PostmasterMain (argc=4, argv=0x55fd0a8faa10) at
> postmaster.c:1403
> #8 0x000055fcdef80a19 in main (argc=4, argv=0x55fd0a8faa10) at main.c:231
> (gdb) p i
> $3 = 0
> (gdb) p NamedLWLockTrancheRequestArray[0]
> Cannot access memory at address 0x7f15ee4ccc08
It looks like the postmaster is trying to access the request array after
re-initializing shared memory, which of course fails. So, we need to keep
the request array in postmaster's local memory, too. Attached is a quick
attempt at a fix.
--
nathan
>From ddc39754763edc82a9eb76164559addbcbced3c4 Mon Sep 17 00:00:00 2001
From: Nathan Bossart <nat...@postgresql.org>
Date: Tue, 16 Sep 2025 22:11:47 -0500
Subject: [PATCH v1 1/1] fix LWLock shmem reinitialization
---
src/backend/storage/lmgr/lwlock.c | 18 ++++++++++--------
1 file changed, 10 insertions(+), 8 deletions(-)
diff --git a/src/backend/storage/lmgr/lwlock.c
b/src/backend/storage/lmgr/lwlock.c
index 46c82c63ca5..694e9e33510 100644
--- a/src/backend/storage/lmgr/lwlock.c
+++ b/src/backend/storage/lmgr/lwlock.c
@@ -192,6 +192,9 @@ typedef struct NamedLWLockTrancheRequest
int NamedLWLockTrancheRequests = 0;
NamedLWLockTrancheRequest *NamedLWLockTrancheRequestArray = NULL;
+/* postmaster's local copy of the request array */
+static NamedLWLockTrancheRequest *LocalNamedLWLockTrancheRequestArray;
+
/* shared memory counter of registered tranches */
int *LWLockCounter = NULL;
@@ -382,7 +385,7 @@ NumLWLocksForNamedTranches(void)
int i;
for (i = 0; i < NamedLWLockTrancheRequests; i++)
- numLocks += NamedLWLockTrancheRequestArray[i].num_lwlocks;
+ numLocks += LocalNamedLWLockTrancheRequestArray[i].num_lwlocks;
return numLocks;
}
@@ -457,9 +460,8 @@ CreateLWLocks(void)
*/
if (NamedLWLockTrancheRequests > 0)
{
- memcpy(ptr, NamedLWLockTrancheRequestArray,
+ memcpy(ptr, LocalNamedLWLockTrancheRequestArray,
NamedLWLockTrancheRequests *
sizeof(NamedLWLockTrancheRequest));
- pfree(NamedLWLockTrancheRequestArray);
NamedLWLockTrancheRequestArray =
(NamedLWLockTrancheRequest *) ptr;
ptr += NamedLWLockTrancheRequests *
sizeof(NamedLWLockTrancheRequest);
}
@@ -648,10 +650,10 @@ RequestNamedLWLockTranche(const char *tranche_name, int
num_lwlocks)
errdetail("LWLock tranche names must be no
longer than %d bytes.",
NAMEDATALEN - 1)));
- if (NamedLWLockTrancheRequestArray == NULL)
+ if (LocalNamedLWLockTrancheRequestArray == NULL)
{
NamedLWLockTrancheRequestsAllocated = 16;
- NamedLWLockTrancheRequestArray = (NamedLWLockTrancheRequest *)
+ LocalNamedLWLockTrancheRequestArray =
(NamedLWLockTrancheRequest *)
MemoryContextAlloc(TopMemoryContext,
NamedLWLockTrancheRequestsAllocated
*
sizeof(NamedLWLockTrancheRequest));
@@ -661,13 +663,13 @@ RequestNamedLWLockTranche(const char *tranche_name, int
num_lwlocks)
{
int i =
pg_nextpower2_32(NamedLWLockTrancheRequests + 1);
- NamedLWLockTrancheRequestArray = (NamedLWLockTrancheRequest *)
- repalloc(NamedLWLockTrancheRequestArray,
+ LocalNamedLWLockTrancheRequestArray =
(NamedLWLockTrancheRequest *)
+ repalloc(LocalNamedLWLockTrancheRequestArray,
i * sizeof(NamedLWLockTrancheRequest));
NamedLWLockTrancheRequestsAllocated = i;
}
- request = &NamedLWLockTrancheRequestArray[NamedLWLockTrancheRequests];
+ request =
&LocalNamedLWLockTrancheRequestArray[NamedLWLockTrancheRequests];
strlcpy(request->tranche_name, tranche_name, NAMEDATALEN);
request->num_lwlocks = num_lwlocks;
NamedLWLockTrancheRequests++;
--
2.39.5 (Apple Git-154)
