[v1] exfat: fix timing of synchronizing bitmap and inode

Commit Message

[e-mail] March 27, 2024, 9:03 a.m. UTC

Commit(f55c096f62f1 exfat: do not zero the extended part) changed
the timing of synchronizing bitmap and inode in exfat_cont_expand().
The change caused xfstests generic/013 to fail if 'dirsync' or 'sync'
is enabled. So this commit restores the timing.

Fixes: f55c096f62f1 ("exfat: do not zero the extended part")
Signed-off-by: Yuezhang Mo <[e-mail]>
---
 fs/exfat/file.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

Comments

Namjae Jeon March 31, 2024, 2:43 p.m. UTC | #1

2024년 3월 27일 (수) 오후 6:04, [e-mail] <[e-mail]>님이 작성:
>
> Commit(f55c096f62f1 exfat: do not zero the extended part) changed
> the timing of synchronizing bitmap and inode in exfat_cont_expand().
> The change caused xfstests generic/013 to fail if 'dirsync' or 'sync'
> is enabled. So this commit restores the timing.
>
> Fixes: f55c096f62f1 ("exfat: do not zero the extended part")
> Signed-off-by: Yuezhang Mo <[e-mail]>
Applied it to #dev. Thanks for your patch!

Patch

From c155bda0ba189b3d2bf873e64086b23a07c78ac4 Mon Sep 17 00:00:00 2001
From: Yuezhang Mo <[e-mail]>
Date: Fri, 22 Mar 2024 17:52:19 +0800
Subject: [PATCH v1] exfat: fix timing of synchronizing bitmap and inode

Commit(f55c096f62f1 exfat: do not zero the extended part) changed
the timing of synchronizing bitmap and inode in exfat_cont_expand().
The change caused xfstests generic/013 to fail if 'dirsync' or 'sync'
is enabled. So this commit restores the timing.

Fixes: f55c096f62f1 ("exfat: do not zero the extended part")
Signed-off-by: Yuezhang Mo <[e-mail]>
---
 fs/exfat/file.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/fs/exfat/file.c b/fs/exfat/file.c
index 40aa134ad4cb..e067a82ec62a 100644
--- a/fs/exfat/file.c
+++ b/fs/exfat/file.c
@@ -51,7 +51,7 @@  static int exfat_cont_expand(struct inode *inode, loff_t size)
 	clu.flags = ei->flags;
 
 	ret = exfat_alloc_cluster(inode, new_num_clusters - num_clusters,
-			&clu, IS_DIRSYNC(inode));
+			&clu, inode_needs_sync(inode));
 	if (ret)
 		return ret;
 
@@ -75,12 +75,11 @@  static int exfat_cont_expand(struct inode *inode, loff_t size)
 	i_size_write(inode, size);
 
 	inode->i_blocks = round_up(size, sbi->cluster_size) >> 9;
+	mark_inode_dirty(inode);
 
-	if (IS_DIRSYNC(inode))
+	if (IS_SYNC(inode))
 		return write_inode_now(inode, 1);
 
-	mark_inode_dirty(inode);
-
 	return 0;
 
 free_clu:
-- 
2.34.1