From: Kent Overstreet <kent.overstreet@linux.dev>
Date: Mon, 5 May 2025 19:52:57 +0000 (-0400)
Subject: bcachefs: Call bch2_fs_start before getting vfs superblock
X-Git-Tag: v6.15-rc6~24^2~4
X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=9c618560994794d6f477e81f3b695fe799feec8a;p=thirdparty%2Fkernel%2Flinux.git

bcachefs: Call bch2_fs_start before getting vfs superblock

This reverts

1fdbe0b184c8 bcachefs: Make sure c->vfs_sb is set before starting fs

switched up bch2_fs_get_tree() so that we got a superblock before
calling bch2_fs_start, so that c->vfs_sb would always be initialized
while the filesystem was active.

This turned out not to be necessary, because blk_holder_ops were
implemented using our own locking, not vfs locking.

And this had the side effect of creating a super_block and doing our
full recovery (including potentially fsck) before setting SB_BORN, which
causes things like sync calls to hang until our recovery is finished.

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

diff --git a/fs/bcachefs/fs.c b/fs/bcachefs/fs.c
index 113db85b6ef9d..b6801861c66f1 100644
--- a/fs/bcachefs/fs.c
+++ b/fs/bcachefs/fs.c
@@ -2502,10 +2502,9 @@ static int bch2_fs_get_tree(struct fs_context *fc)
 
 	bch2_opts_apply(&c->opts, opts);
 
-	/*
-	 * need to initialise sb and set c->vfs_sb _before_ starting fs,
-	 * for blk_holder_ops
-	 */
+	ret = bch2_fs_start(c);
+	if (ret)
+		goto err_stop_fs;
 
 	sb = sget(fc->fs_type, NULL, bch2_set_super, fc->sb_flags|SB_NOSEC, c);
 	ret = PTR_ERR_OR_ZERO(sb);
@@ -2567,10 +2566,6 @@ got_sb:
 
 	sb->s_shrink->seeks = 0;
 
-	ret = bch2_fs_start(c);
-	if (ret)
-		goto err_put_super;
-
 #ifdef CONFIG_UNICODE
 	sb->s_encoding = c->cf_encoding;
 #endif