Re: [PATCH] cifs: Fix reacquisition of volume cookie on still-live connection

[David Howells <dhowells@redhat.com>]

Paulo Alcantara <pc@manguebit.com> wrote:

> > [!] Note: Looking at cifs_mount_get_tcon(), a more general solution may
> > actually be required.  Reacquiring the volume cookie isn't the only thing
> > that function does: it also partially reinitialises the tcon record without
> > any locking - which may cause live filesystem ops already using the tcon
> > through a previous mount to malfunction.
> 
> Agreed.

Looking over the code again, I'm not sure whether is actually necessary - or
whether it is necessary and will be a bit nasty to implement as it will
require read locking also.

Firstly, reset_cifs_unix_caps() seems to re-set tcon->fsUnixInfo.Capability
and tcon->unix_ext, which it would presumably set to the same things - which
is probably fine.

However, cifs_qfs_tcon() makes RPC operations that reloads tcon->fsDevInfo and
tcon->fsAttrInfo - both of which may be being accessed without locks.

smb2_qfs_tcon() and smb3_qfs_tcon() alters everything cifs_qfs_tcon() does,
plus a bunch of extra tcon members.  Can this locally cached information
change over time on the server whilst we have a connection to it?

David