Re: [PATCH] ntfs: handle a failed size rollback in the write error path

[Hongling Zeng]

在 2026年10月06日 13:53, Hyunchul Lee 写道:
> On Sun, Oct 04, 2026 at 10:48:22AM +0800, Hongling Zeng wrote:
> > ntfs_file_write_iter() extends the data and initialized sizes before
> > it copies the user data and restores them when the write then fails.
> > The restoration is best-effort: both results are discarded, even
> > though the extension is already recorded in the dirty MFT record.
> > When the restoration fails too, the extended sizes reach the disk
> > while the VFS inode keeps reporting the old ones: after a remount the
> > file appears grown again, with an uninitialized tail where the failed
> > write never put any data, and nothing records the failure.
> >
> > Keep both undo results: on failure, set NVolSetErrors(), which the next
> > persistence point persists as VOLUME_IS_DIRTY, and log a message for
> > the errors= policy.
> >
> > Fixes: 9c87959601e8 ("ntfs: update file operations")
> > Signed-off-by: Hongling Zeng <zenghongling@xxxxxxxxxx>
> > ---
> >   fs/ntfs/file.c | 29 +++++++++++++++++++++++++++--
> >   1 file changed, 27 insertions(+), 2 deletions(-)
> >
> > diff --git a/fs/ntfs/file.c b/fs/ntfs/file.c
> > index 3ec82715a588..b5b30e44ac51 100644
> > --- a/fs/ntfs/file.c
> > +++ b/fs/ntfs/file.c
> > @@ -678,14 +678,39 @@ static ssize_t ntfs_file_write_iter(struct kiocb *iocb, struct iov_iter *from)
> >   				&ntfs_iomap_folio_ops, NULL);
> >   out:
> >   	if (ret < 0 && ret != -EIOCBQUEUED) {
> > +		int undo_err = 0;
> > +
> >   		if (ni->initialized_size != old_init_size) {
> > +			int init_err;
> > +
> >   			mutex_lock(&ni->mrec_lock);
> > -			ntfs_attr_set_initialized_size(ni, old_init_size);
> > +			init_err = ntfs_attr_set_initialized_size(ni,
> > +					old_init_size);
> I think the rollback needs to distinguish resident and compressed
> attributes.
> If the data attribute is still resident,
> ntfs_attr_set_initialized_size() always return -EINVAL.
Hi  Hyunchul：
you are right, thanks.  The initialized size rollback now skips a
  resident attribute, where the data size rollback restores both coupled
  sizes, and ntfs_attr_truncate() is skipped on a compressed attribute
  with a warning.
> >   			mutex_unlock(&ni->mrec_lock);
> > +			if (init_err && !undo_err)
> > +				undo_err = init_err;
> >   		}
> >   		if (ni->data_size != old_data_size) {
> > +			int data_err;
> > +
> >   			truncate_setsize(vi, old_data_size);
> > -			ntfs_attr_truncate(ni, old_data_size);
> > +			data_err = ntfs_attr_truncate(ni, old_data_size);
> And ntfs_attr_truncate() always return -EOPNOTSUPP for a compressed
> attribute.
>
> > +			if (data_err && !undo_err)
> > +				undo_err = data_err;
> > +		}
> > +		if (undo_err) {
> > +			/*
> > +			 * The extension is recorded in the dirty MFT
> > +			 * record already, so a failed rollback leaves the
> > +			 * new sizes on disk while the VFS inode keeps
> > +			 * reporting the old ones.  Record the error: the
> > +			 * next persistence point persists it as
> > +			 * VOLUME_IS_DIRTY.  Run chkdsk.
> > +			 */
> > +			NVolSetErrors(vol);
> > +			ntfs_error(vi->i_sb,
> > +				   "Failed to roll back the size update: %d",
> > +				   undo_err);
> >   		}
> >   	}
> >   out_lock:
> > --
> > 2.25.1