Don't destroy SMgrRelations at relcache invalidation

With commit 21d9c3ee4e, SMgrRelations remain valid until end of
transaction (or longer if they're "pinned"). Relcache invalidation can
happen in the middle of a transaction, so we must not destroy them at
relcache invalidation anymore.

This was revealed by failures in the 'constraints' test in buildfarm
animals using -DCLOBBER_CACHE_ALWAYS. That started failing with commit
8af2565248, which was the first commit that started to rely on an
SMgrRelation living until end of transaction.

Diagnosed-by: Tomas Vondra, Thomas Munro
Reviewed-by: Thomas Munro
Discussion: https://www.postgresql.org/message-id/CA%2BhUKGK%2B5DOmLaBp3Z7C4S-Yv6yoROvr1UncjH2S1ZbPT8D%2BZg%40mail.gmail.com

src/backend/storage/smgr/smgr.c

@@ -348,8 +348,6 @@ smgrdestroyall(void)
 
 /*
  * smgrreleaseall() -- Release resources used by all objects.
- *
- * This is called for PROCSIGNAL_BARRIER_SMGRRELEASE.
  */
 void
 smgrreleaseall(void)

src/backend/utils/cache/relcache.c

@@ -2985,9 +2985,6 @@ RelationCacheInvalidate(bool debug_discard)
 	{
 		relation = idhentry->reldesc;
 
-		/* Must close all smgr references to avoid leaving dangling ptrs */
-		RelationCloseSmgr(relation);
-
 		/*
 		 * Ignore new relations; no other backend will manipulate them before
 		 * we commit.  Likewise, before replacing a relation's relfilelocator,
@@ -3039,11 +3036,10 @@ RelationCacheInvalidate(bool debug_discard)
 	}
 
 	/*
-	 * Now zap any remaining smgr cache entries.  This must happen before we
-	 * start to rebuild entries, since that may involve catalog fetches which
-	 * will re-open catalog files.
+	 * We cannot destroy the SMgrRelations as there might still be references
+	 * to them, but close the underlying file descriptors.
 	 */
-	smgrdestroyall();
+	smgrreleaseall();
 
 	/* Phase 2: rebuild the items found to need rebuild in phase 1 */
 	foreach(l, rebuildFirstList)