In ubifs_jnl_update() we sync parent and child inodes to the flash, in case of xattrs, the parent inode (AKA host inode) has a non-zero data_len. Therefore we need to adjust synced_i_size too.
This issue was reported by ubifs self tests unter a xattr related work load. UBIFS error (ubi0:0 pid 1896): dbg_check_synced_i_size: ui_size is 4, synced_i_size is 0, but inode is clean UBIFS error (ubi0:0 pid 1896): dbg_check_synced_i_size: i_ino 65, i_mode 0x81a4, i_size 4
Cc: stable@vger.kernel.org Fixes: 1e51764a3c2a ("UBIFS: add new flash file system") Signed-off-by: Richard Weinberger richard@nod.at --- fs/ubifs/journal.c | 2 ++ 1 file changed, 2 insertions(+)
diff --git a/fs/ubifs/journal.c b/fs/ubifs/journal.c index 04c4ec6483e5..e3e1e093db81 100644 --- a/fs/ubifs/journal.c +++ b/fs/ubifs/journal.c @@ -664,6 +664,8 @@ int ubifs_jnl_update(struct ubifs_info *c, const struct inode *dir, finish_reservation(c); spin_lock(&ui->ui_lock); ui->synced_i_size = ui->ui_size; + if (xent) + host_ui->synced_i_size = host_ui->ui_size; spin_unlock(&ui->ui_lock); mark_inode_clean(c, ui); mark_inode_clean(c, host_ui);
Am Mittwoch, 14. Februar 2018, 12:02:48 CET schrieb Richard Weinberger:
In ubifs_jnl_update() we sync parent and child inodes to the flash, in case of xattrs, the parent inode (AKA host inode) has a non-zero data_len. Therefore we need to adjust synced_i_size too.
This issue was reported by ubifs self tests unter a xattr related work load. UBIFS error (ubi0:0 pid 1896): dbg_check_synced_i_size: ui_size is 4, synced_i_size is 0, but inode is clean UBIFS error (ubi0:0 pid 1896): dbg_check_synced_i_size: i_ino 65, i_mode 0x81a4, i_size 4
Cc: stable@vger.kernel.org Fixes: 1e51764a3c2a ("UBIFS: add new flash file system") Signed-off-by: Richard Weinberger richard@nod.at
fs/ubifs/journal.c | 2 ++ 1 file changed, 2 insertions(+)
diff --git a/fs/ubifs/journal.c b/fs/ubifs/journal.c index 04c4ec6483e5..e3e1e093db81 100644 --- a/fs/ubifs/journal.c +++ b/fs/ubifs/journal.c @@ -664,6 +664,8 @@ int ubifs_jnl_update(struct ubifs_info *c, const struct inode *dir, finish_reservation(c); spin_lock(&ui->ui_lock); ui->synced_i_size = ui->ui_size;
- if (xent)
spin_unlock(&ui->ui_lock);host_ui->synced_i_size = host_ui->ui_size;
Reviewing my own patch on LKML helps a lot. We need to take host_ui->ui_lock too...
Thanks, //richard
linux-stable-mirror@lists.linaro.org