v1-0001-Do-not-lock-in-BufferGetLSNAtomic-on-archs-with-8.patch
text/x-patch
Filename: v1-0001-Do-not-lock-in-BufferGetLSNAtomic-on-archs-with-8.patch
Type: text/x-patch
Part: 0
Patch
Same data as JSON:
GET /api/v1/attachments/:id/patch
the parsed metadata as JSON — format, series position, per-file stats; never the diff bytes.
API reference →
Format: format-patch
Series: patch v1-0001
Subject: Do not lock in BufferGetLSNAtomic() on archs with 8 byte atomic reads
| File | + | − |
|---|---|---|
| src/backend/storage/buffer/bufmgr.c | 9 | 3 |
| src/include/storage/bufpage.h | 24 | 9 |
From dbf345ac5a857a11801d07291ecd71eb224050af Mon Sep 17 00:00:00 2001
From: Andreas Karlsson <andreas@proxel.se>
Date: Fri, 21 Nov 2025 10:15:06 +0100
Subject: [PATCH v1 1/2] Do not lock in BufferGetLSNAtomic() on archs with 8
byte atomic reads
On platforms where we can read or write the whole LSN to a buffer page
we do not need to lock the buffer header to be sure we do not get a torn
LSN. To ahcieve this we need to make sure we always read and write the
while LSN and do not write it field by field.
While working on this I converted PageXLogRecPtr from a struct with two
uint32 fields into an uint64 to make the intent more clear.
Idea by Andres Freund.
---
src/backend/storage/buffer/bufmgr.c | 12 ++++++++---
src/include/storage/bufpage.h | 33 +++++++++++++++++++++--------
2 files changed, 33 insertions(+), 12 deletions(-)
diff --git a/src/backend/storage/buffer/bufmgr.c b/src/backend/storage/buffer/bufmgr.c
index 327ddb7adc8..34c4c4bd925 100644
--- a/src/backend/storage/buffer/bufmgr.c
+++ b/src/backend/storage/buffer/bufmgr.c
@@ -4491,13 +4491,18 @@ BufferIsPermanent(Buffer buffer)
/*
* BufferGetLSNAtomic
- * Retrieves the LSN of the buffer atomically using a buffer header lock.
- * This is necessary for some callers who may not have an exclusive lock
- * on the buffer.
+ * Retrieves the LSN of the buffer atomically.
+ *
+ * On platforms without 8 byte atomic reads/write we need to take a
+ * header lock. This is necessary for some callers who may not have an
+ * exclusive lock on the buffer.
*/
XLogRecPtr
BufferGetLSNAtomic(Buffer buffer)
{
+#ifdef PG_HAVE_8BYTE_SINGLE_COPY_ATOMICITY
+ return PageGetLSN(BufferGetPage(buffer));
+#else
char *page = BufferGetPage(buffer);
BufferDesc *bufHdr;
XLogRecPtr lsn;
@@ -4518,6 +4523,7 @@ BufferGetLSNAtomic(Buffer buffer)
UnlockBufHdr(bufHdr);
return lsn;
+#endif
}
/* ---------------------------------------------------------------------
diff --git a/src/include/storage/bufpage.h b/src/include/storage/bufpage.h
index abc2cf2a020..1b54967c70c 100644
--- a/src/include/storage/bufpage.h
+++ b/src/include/storage/bufpage.h
@@ -91,23 +91,38 @@ typedef uint16 LocationIndex;
/*
- * For historical reasons, the 64-bit LSN value is stored as two 32-bit
- * values.
+ * For historical reasons, the storage of 64-bit LSN values depends on the
+ * endianess because the field used to be stored as two 32-bit values. When
+ * reading and writing the LSN we need to convert between the two formats.
+ *
+ * We are careful to try to treat the LSN as a single uint64 so callers like
+ * BufferGetLSNAtomic() can be sure there are no torn reads or writes.
*/
-typedef struct
-{
- uint32 xlogid; /* high bits */
- uint32 xrecoff; /* low bits */
-} PageXLogRecPtr;
+typedef uint64 PageXLogRecPtr;
+
+#ifdef WORDS_BIGENDIAN
static inline XLogRecPtr
PageXLogRecPtrGet(PageXLogRecPtr val)
{
- return (uint64) val.xlogid << 32 | val.xrecoff;
+ return val;
}
#define PageXLogRecPtrSet(ptr, lsn) \
- ((ptr).xlogid = (uint32) ((lsn) >> 32), (ptr).xrecoff = (uint32) (lsn))
+ ((ptr) = (lsn))
+
+#else
+
+static inline XLogRecPtr
+PageXLogRecPtrGet(volatile PageXLogRecPtr val)
+{
+ return (val << 32) | (val >> 32);
+}
+
+#define PageXLogRecPtrSet(ptr, lsn) \
+ ((ptr) = ((lsn) << 32) | ((lsn) >> 32))
+
+#endif
/*
* disk page organization
--
2.47.3