From: Kent Overstreet <kent.overstreet@linux.dev>
Date: Tue, 22 Apr 2025 08:54:54 +0000 (-0400)
Subject: bcachefs: Fix btree_iter_peek_prev() at end of inode
X-Git-Tag: v6.15-rc4~27^2~5
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=353739f1d1675692d9de01483c75c15b314710ac;p=thirdparty%2Flinux.git

bcachefs: Fix btree_iter_peek_prev() at end of inode

At the end of the inode, on an extents iterator, peek_slot() has to
advance to the next position to avoid returning a 0 size extent, which
is not allowed.

Changing iter->pos confuses peek_prev(), but we don't need to call
peek_slot() in this case.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
---

diff --git a/fs/bcachefs/btree_iter.c b/fs/bcachefs/btree_iter.c
index 400ff650553b4..59fa527ac6854 100644
--- a/fs/bcachefs/btree_iter.c
+++ b/fs/bcachefs/btree_iter.c
@@ -2577,7 +2577,10 @@ struct bkey_s_c bch2_btree_iter_peek_prev_min(struct btree_trans *trans, struct
 					      struct bpos end)
 {
 	if ((iter->flags & (BTREE_ITER_is_extents|BTREE_ITER_filter_snapshots)) &&
-	   !bkey_eq(iter->pos, POS_MAX)) {
+	   !bkey_eq(iter->pos, POS_MAX) &&
+	   !((iter->flags & BTREE_ITER_is_extents) &&
+	     iter->pos.offset == U64_MAX)) {
+
 		/*
 		 * bkey_start_pos(), for extents, is not monotonically
 		 * increasing until after filtering for snapshots: