On Tue, 2016-01-26 at 18:19 +0100, Christoph Hellwig wrote:
> > +static bool __target_check_io_state(struct se_cmd *se_cmd)
> > +{
> > + struct se_session *sess = se_cmd->se_sess;
> > +
> > + assert_spin_locked(&se_session->sess_cmd_lock);
> > + WARN_ON_ONCE(!irqs_disabled());
>
> Btw, I looked a the code and can't really see what sess_cmd_lock is
> supposed to protect here.
>
> > + sess = cmd->se_sess;
> > + if (WARN_ON_ONCE(!sess))
> > + continue;
> > +
> > + spin_lock(&sess->sess_cmd_lock);
> > + rc = __target_check_io_state(cmd);
> > + spin_unlock(&sess->sess_cmd_lock);
> > + if (!rc) {
> > + printk("LUN_RESET I/O: non-zero kref_get_unless_zero\n");
> > + continue;
> > + }
>
> And thus why we care about taking it here.
I'm still working on -v3 series to handle se_session shutdown during
this specific multi-port LUN_RESET case, and considering using the
existing se_cmd->cmd_wait_set bit to signal when this special case
happens.
Currently ->sess_cmd_lock is held in target_release_cmd_kref() and
target_sess_cmd_list_set_waiting() while checking and setting this
value.
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
