From mboxrd@z Thu Jan 1 00:00:00 1970 From: Bart Van Assche Subject: Re: [PATCH-v2 2/4] ib_srpt: Convert acl lookup to modern get_initiator_node_acl usage Date: Mon, 11 Jan 2016 13:34:46 -0800 Message-ID: <56941FF6.5090003@sandisk.com> References: <1452457724-10629-1-git-send-email-nab@daterainc.com> <1452457724-10629-3-git-send-email-nab@daterainc.com> Mime-Version: 1.0 Content-Type: text/plain; charset="windows-1252" Content-Transfer-Encoding: 7bit Return-path: Received: from mail-bl2on0091.outbound.protection.outlook.com ([65.55.169.91]:63456 "EHLO na01-bl2-obe.outbound.protection.outlook.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1759729AbcAKVgE (ORCPT ); Mon, 11 Jan 2016 16:36:04 -0500 In-Reply-To: <1452457724-10629-3-git-send-email-nab@daterainc.com> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: "Nicholas A. Bellinger" , target-devel Cc: linux-scsi , lkml , Sagi Grimberg , Christoph Hellwig , Hannes Reinecke , Andy Grover , Vasu Dev , Vu Pham , Nicholas Bellinger On 01/10/2016 12:28 PM, Nicholas A. Bellinger wrote: > This patch does a simple conversion of ib_srpt code to use > proper modern core_tpg_get_initiator_node_acl() lookup using > se_node_acl->acl_kref, and drops the legacy internal list > usage from srpt_lookup_acl(). > > This involves doing transport_init_session() earlier, and > making sure transport_free_session() is called during > a se_node_acl lookup failure to drop the last ->acl_kref. > > Also, it adds a minor backwards-compat hack to avoid the > potential for user-space wrt node-acl WWPN formatting by > simply stripping off '0x' prefix from ch->sess_name, and > retrying once if core_tpg_get_initiator_node_acl() fails. > > Finally, go ahead and drop port_acl_list port_acl_lock > since they are no longer used. Hello Nic, I think that you promised last week to remove the srpt_node_acl structure (see also http://thread.gmane.org/gmane.linux.scsi/109275/focus=2121617). However, this patch doesn't remove that structure. > - ch->sess = transport_init_session(TARGET_PROT_NORMAL); > - if (IS_ERR(ch->sess)) { > +try_again: > + se_acl = core_tpg_get_initiator_node_acl(&sport->port_tpg_1, p); > + if (!se_acl) { > + pr_info("Rejected login because no ACL has been" > + " configured yet for initiator %s.\n", ch->sess_name); > + /* > + * XXX: Hack to retry of ch->i_port_id without leading '0x' > + */ > + if (p == &ch->sess_name[0]) { > + p += 2; > + goto try_again; > + } > rej->reason = cpu_to_be32( > - SRP_LOGIN_REJ_INSUFFICIENT_RESOURCES); > - pr_debug("Failed to create session\n"); > - goto deregister_session; > + SRP_LOGIN_REJ_CHANNEL_LIMIT_REACHED); > + transport_free_session(ch->sess); > + goto destroy_ib; > } The "goto try_again" statement is executed at most once. Since the above loop can be unfolded with only a very minor code duplication I think it should be unfolded, e.g. as follows: se_acl = core_tpg_get_initiator_node_acl(&sport->port_tpg_1, ch->sess_name); if (!se_acl) se_acl = core_tpg_get_initiator_node_acl(&sport->port_tpg_1, ch->sess_name + strlen("0x")); if (!se_acl) { pr_info("Rejected login because no ACL has been" " configured yet for initiator %s.\n", ch->sess_name); [ ... ] } Bart.