Improve scalability of WAL insertions.

This patch replaces WALInsertLock with a number of WAL insertion slots,
allowing multiple backends to insert WAL records to the WAL buffers
concurrently. This is particularly useful for parallel loading large amounts
of data on a system with many CPUs.

This has one user-visible change: switching to a new WAL segment with
pg_switch_xlog() now fills the remaining unused portion of the segment with
zeros. This potentially adds some overhead, but it has been a very common
practice by DBA's to clear the "tail" of the segment with an external
pg_clearxlogtail utility anyway, to make the WAL files compress better.
With this patch, it's no longer necessary to do that.

This patch adds a new GUC, xloginsert_slots, to tune the number of WAL
insertion slots. Performance testing suggests that the default, 8, works
pretty well for all kinds of worklods, but I left the GUC in place to allow
others with different hardware to test that easily. We might want to remove
that before release.

Reviewed by Andres Freund.

src/backend/storage/lmgr/spin.c

@@ -22,6 +22,7 @@
  */
 #include "postgres.h"
 
+#include "access/xlog.h"
 #include "miscadmin.h"
 #include "replication/walsender.h"
 #include "storage/lwlock.h"
@@ -63,6 +64,7 @@ SpinlockSemas(void)
 	nsemas = NumLWLocks();		/* one for each lwlock */
 	nsemas += NBuffers;			/* one for each buffer header */
 	nsemas += max_wal_senders;	/* one for each wal sender process */
+	nsemas += num_xloginsert_slots; /* one for each WAL insertion slot */
 	nsemas += 30;				/* plus a bunch for other small-scale use */
 
 	return nsemas;

src/backend/utils/misc/guc.c

@@ -2037,6 +2037,17 @@ static struct config_int ConfigureNamesInt[] =
 		NULL, NULL, NULL
 	},
 
+	{
+		{"xloginsert_slots", PGC_POSTMASTER, WAL_SETTINGS,
+			gettext_noop("Sets the number of slots for concurrent xlog insertions."),
+			NULL,
+			GUC_NOT_IN_SAMPLE
+		},
+		&num_xloginsert_slots,
+		8, 1, 1000,
+		NULL, NULL, NULL
+	},
+
 	{
 		/* see max_connections */
 		{"max_wal_senders", PGC_POSTMASTER, REPLICATION_SENDING,

src/include/access/xlog.h

@@ -190,6 +190,7 @@ extern char *XLogArchiveCommand;
 extern bool EnableHotStandby;
 extern bool fullPageWrites;
 extern bool log_checkpoints;
+extern int	num_xloginsert_slots;
 
 /* WAL levels */
 typedef enum WalLevel

src/include/access/xlogdefs.h

@@ -93,14 +93,4 @@ typedef uint32 TimeLineID;
 #define DEFAULT_SYNC_METHOD		SYNC_METHOD_FSYNC
 #endif
 
-/*
- * Limitation of buffer-alignment for direct IO depends on OS and filesystem,
- * but XLOG_BLCKSZ is assumed to be enough for it.
- */
-#ifdef O_DIRECT
-#define ALIGNOF_XLOG_BUFFER		XLOG_BLCKSZ
-#else
-#define ALIGNOF_XLOG_BUFFER		ALIGNOF_BUFFER
-#endif
-
 #endif   /* XLOG_DEFS_H */

src/include/storage/lwlock.h

@@ -53,7 +53,7 @@ typedef enum LWLockId
 	ProcArrayLock,
 	SInvalReadLock,
 	SInvalWriteLock,
-	WALInsertLock,
+	WALBufMappingLock,
 	WALWriteLock,
 	ControlFileLock,
 	CheckpointLock,