On 2/8/16, 9:49 PM, "Nicholas A. Bellinger" <n...@linux-iscsi.org> wrote:
>On Mon, 2016-02-08 at 17:43 +0000, Himanshu Madhani wrote: >> > >> >So looking at this patch beyond the debugfs part, it does change where >> >->check_initiator_node_acl() gets call during qlt_create_sess(). >> > >> >I assume this is related to new debugfs attribute, and these changes >> >(plus others in qlt_del_sess_work_fn) are not bug-fixes on their own, >> >correct..? >> > >> >Aside from that, I don't have an objection to merge as v4.6 for-next >> >code if QLogic finds it useful for debugging. >> >> Yes. The changes in this patch are strictly for making information >> available via debugFS and not bug-fixes on their own. > >Thanks for clarifying. > >> This information has been requested by few customer who >> finds this information useful for debugging and in addition they use >>this >> information in their configuration scripts as well. > >So I've got no particular objection to people adding debugfs >informational attributes for whatever to aid debugging of their drivers. > >That said, the same information can too be exposed as a tcm_qla2xxx >struct se_node_acl->acl_group configfs attribute, and for demo-mode >sessions via a tcm_qla2xxx provided dynamic_sessions configfs attribute >handler. > >I'd be happy to add expose this under tcm_qla2xxx configfs for v4.6 code >if you'd like, otherwise I'm OK with the patch in queue-next as-is if >you find it useful. > I really like your idea of exposing this information via configFS. Please go ahead and make updates on exposing this information via se_node_acl->acl_group Configfs attribute. We can ask our customer to explore configfs to retrieve information. However, We would also like to keep this information in debugFS for now. >
<<attachment: winmail.dat>>
