From: Kent Overstreet <kent.overstreet@linux.dev>
Date: Mon, 9 Dec 2024 11:00:33 +0000 (-0500)
Subject: bcachefs: Don't start rewriting btree nodes until after journal replay
X-Git-Tag: io_uring-6.14-20240131~27^2~82
X-Git-Url: https://git.kernel.dk/?a=commitdiff_plain;h=54dacdada6de8d97e7ca7e51eadc96c61032bdb4;p=linux-block.git

bcachefs: Don't start rewriting btree nodes until after journal replay

This fixes a deadlock during journal replay when btree node read errors
kick off a ton of rewrites: we don't want them competing with journal
replay.

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

diff --git a/fs/bcachefs/btree_update_interior.c b/fs/bcachefs/btree_update_interior.c
index 7d9dab95bdcf..03a6eba7403d 100644
--- a/fs/bcachefs/btree_update_interior.c
+++ b/fs/bcachefs/btree_update_interior.c
@@ -2291,7 +2291,8 @@ void bch2_btree_node_rewrite_async(struct bch_fs *c, struct btree *b)
 	bool now = false, pending = false;
 
 	spin_lock(&c->btree_node_rewrites_lock);
-	if (bch2_write_ref_tryget(c, BCH_WRITE_REF_node_rewrite)) {
+	if (c->curr_recovery_pass > BCH_RECOVERY_PASS_journal_replay &&
+	    bch2_write_ref_tryget(c, BCH_WRITE_REF_node_rewrite)) {
 		list_add(&a->list, &c->btree_node_rewrites);
 		now = true;
 	} else if (!test_bit(BCH_FS_may_go_rw, &c->flags)) {