From: Kent Overstreet <kent.overstreet@linux.dev>
Date: Sat, 18 Jan 2025 07:05:57 +0000 (-0500)
Subject: bcachefs: Self healing writes are BCH_WRITE_alloc_nowait
X-Git-Tag: v6.15-rc1~146^2~157
X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=29ad31c780d1e4e37244140442aaec41c3efb7d6;p=thirdparty%2Fkernel%2Flinux.git

bcachefs: Self healing writes are BCH_WRITE_alloc_nowait

If a drive is failing and we're moving data off of it, we can't
necessairly depend on capacity/disk reservation calculations to avoid
deadlocking/blocking on the allocator.

And, we don't want to queue up infinite self healing moves anyways.

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

diff --git a/fs/bcachefs/io_read.c b/fs/bcachefs/io_read.c
index 7a66feb1c011e..bdb554f6db8c6 100644
--- a/fs/bcachefs/io_read.c
+++ b/fs/bcachefs/io_read.c
@@ -188,12 +188,12 @@ static struct bch_read_bio *__promote_alloc(struct btree_trans *trans,
 		goto err;
 	}
 
-	struct data_update_opts update_opts = {};
+	struct data_update_opts update_opts = { .write_flags = BCH_WRITE_alloc_nowait };
 
 	if (!have_io_error(failed)) {
 		update_opts.target = orig->opts.promote_target;
 		update_opts.extra_replicas = 1;
-		update_opts.write_flags = BCH_WRITE_alloc_nowait|BCH_WRITE_cached;
+		update_opts.write_flags |= BCH_WRITE_cached;
 		update_opts.write_flags |= BCH_WRITE_only_specified_devs;
 	} else {
 		update_opts.target = orig->opts.foreground_target;