diff --git a/src/backend/replication/slot.c b/src/backend/replication/slot.c
index b496b11b5a9..77e022dfb0a 100644
--- a/src/backend/replication/slot.c
+++ b/src/backend/replication/slot.c
@@ -1580,16 +1580,19 @@ ReplicationSlotReserveWal(void)
 	Assert(!XLogRecPtrIsValid(slot->last_saved_restart_lsn));
 
 	/*
+	 * The replication slot mechanism is used to prevent the removal of
+	 * required WAL.
+	 *
 	 * Acquire an exclusive lock to prevent the checkpoint process from
-	 * concurrently calculating the minimum slot LSN (see
-	 * CheckPointReplicationSlots), ensuring that the reserved WAL cannot be
-	 * removed during a checkpoint.
+	 * concurrently computing the minimum slot LSN (see
+	 * CheckPointReplicationSlots). This ensures that the WAL reserved for
+	 * replication cannot be removed during a checkpoint.
 	 *
 	 * The mechanism is reliable because if WAL reservation occurs first, the
 	 * checkpoint must wait for the restart_lsn update before determining the
-	 * minimum non-removable LSN. On the other hand, if the checkpoint occurs
-	 * first, subsequent WAL reservations must choose positions beyond or
-	 * equal to the redo pointer of checkpoint.
+	 * minimum non-removable LSN. On the other hand, if the checkpoint happens
+	 * first, subsequent WAL reservations will select positions at or beyond
+	 * the redo pointer of that checkpoint.
 	 */
 	LWLockAcquire(ReplicationSlotAllocationLock, LW_EXCLUSIVE);
 
@@ -1603,10 +1606,9 @@ ReplicationSlotReserveWal(void)
 	 * record.
 	 *
 	 * None of this is needed (or indeed helpful) for physical slots as
-	 * they'll start replay at the last logged checkpoint anyway. Instead
-	 * return the location of the last redo LSN. While that slightly increases
-	 * the chance that we have to retry, it's where a base backup has to start
-	 * replay at.
+	 * they'll start replay at the last logged checkpoint anyway. Instead,
+	 * return the location of the last redo LSN, where a base backup has to
+	 * start replay at.
 	 */
 	if (SlotIsPhysical(slot))
 		restart_lsn = GetRedoRecPtr();
@@ -1622,6 +1624,7 @@ ReplicationSlotReserveWal(void)
 	/* prevent WAL removal as fast as possible */
 	ReplicationSlotsComputeRequiredLSN();
 
+	/* Checkpoint shouldn't remove the required WAL. */
 	XLByteToSeg(slot->data.restart_lsn, segno, wal_segment_size);
 	if (XLogGetLastRemovedSegno() >= segno)
 		elog(ERROR, "WAL required by replication slot %s has been removed concurrently",
@@ -2139,8 +2142,8 @@ CheckPointReplicationSlots(bool is_shutdown)
 	 * Additionally, acquiring the Allocation lock is necessary to serialize
 	 * the slot flush process with concurrent slot WAL reservation. This
 	 * ensures that the WAL position being reserved is either flushed to disk
-	 * or beyond or equal to the redo pointer (See ReplicationSlotReserveWal
-	 * for details).
+	 * or is beyond or equal to the redo pointer of the current checkpoint (See
+	 * ReplicationSlotReserveWal for details).
 	 */
 	LWLockAcquire(ReplicationSlotAllocationLock, LW_SHARED);