Re: [PATCH v3 1/5] ovl: do not open non-data lower file for fsync

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Mon, Oct 07, 2024 at 05:55:40PM +0100, Al Viro wrote:

> > +static int ovl_upper_fdget(const struct file *file, struct fd *real, bool data)
> > +{
> > +	struct dentry *dentry = file_dentry(file);
> > +	struct path realpath;
> > +	enum ovl_path_type type;
> > +
> > +	if (data)
> > +		type = ovl_path_realdata(dentry, &realpath);
> 
> ... but not here.
> 
> I can see the point of not doing that in ->fsync() after we'd already
> done ovl_verify_lowerdata() at open time, but what's different about
> ->read_iter() and friends that also come only after ->open()?
> IOW, why is fdatasync() different from other data-access cases?

Nevermind that one - the answer is that ovl_path_realdata()
calls ovl_path_lowerdata() only in case when it sees
!OVL_TYPE_UPPER(type) || OVL_TYPE_MERGE(type), which guarantees
that the type check below that if (data) will fail anyway
(check being (!OVL_TYPE_UPPER(type) || (data && OVL_TYPE_MERGE(type))).

So this reduces the fdatasync case to
	if (!OVL_TYPE_UPPER(type) || OVL_TYPE_MERGE(type))
		fail;
	ovl_path_upper(dentry, &realpath);
just as the fsync case reduces to
	if (!OVL_TYPE_UPPER(type))
		fail;
	ovl_path_upper(dentry, &realpath);

making any lowerpath-related stuff irrelevant.




[Index of Archives]     [Linux Ext4 Filesystem]     [Union Filesystem]     [Filesystem Testing]     [Ceph Users]     [Ecryptfs]     [NTFS 3]     [AutoFS]     [Kernel Newbies]     [Share Photos]     [Security]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux Cachefs]     [Reiser Filesystem]     [Linux RAID]     [NTFS 3]     [Samba]     [Device Mapper]     [CEPH Development]

  Powered by Linux