Don't call PageGetSpecialPointer() on page until it's been initialized.

After calling XLogInitBufferForRedo(), the page might be all-zeros if it was
not in page cache already. btree_xlog_unlink_page initialized the page
correctly, but it called PageGetSpecialPointer before initializing it, which
would lead to a corrupt page at WAL replay, if the unlinked page is not in
page cache.

Backpatch to 9.4, the bug came with the rewrite of B-tree page deletion.

src/backend/access/nbtree/nbtxlog.c

@@ -864,9 +864,10 @@ btree_xlog_unlink_page(uint8 info, XLogReaderState *record)
 
 		buffer = XLogInitBufferForRedo(record, 3);
 		page = (Page) BufferGetPage(buffer);
-		pageop = (BTPageOpaque) PageGetSpecialPointer(page);
 
 		_bt_pageinit(page, BufferGetPageSize(buffer));
+		pageop = (BTPageOpaque) PageGetSpecialPointer(page);
+
 		pageop->btpo_flags = BTP_HALF_DEAD | BTP_LEAF;
 		pageop->btpo_prev = xlrec->leafleftsib;
 		pageop->btpo_next = xlrec->leafrightsib;