From mboxrd@z Thu Jan 1 00:00:00 1970 From: Akhil Goyal Subject: Re: [PATCH v2 2/2] examples/ipsec-secgw: attach session-qp Date: Thu, 23 Mar 2017 17:48:14 +0530 Message-ID: References: <20170317084510.2120-1-akhil.goyal@nxp.com> <20170323080648.7149-1-akhil.goyal@nxp.com> <20170323080648.7149-2-akhil.goyal@nxp.com> <1db2a8d1-e59c-b4ae-7496-6b59e28adba3@intel.com> <05d41e32-1ec0-8532-c8a5-a0eea852e1f5@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset="windows-1252"; format=flowed Content-Transfer-Encoding: 7bit Cc: , , , To: Sergio Gonzalez Monroy , Return-path: Received: from NAM01-BY2-obe.outbound.protection.outlook.com (mail-by2nam01on0073.outbound.protection.outlook.com [104.47.34.73]) by dpdk.org (Postfix) with ESMTP id 5B78AD1BB for ; Thu, 23 Mar 2017 13:18:24 +0100 (CET) In-Reply-To: <05d41e32-1ec0-8532-c8a5-a0eea852e1f5@intel.com> List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" On 3/23/2017 4:13 PM, Sergio Gonzalez Monroy wrote: > On 23/03/2017 10:38, Akhil Goyal wrote: >> On 3/23/2017 4:00 PM, Sergio Gonzalez Monroy wrote: >>> That was simpler than I thought. >>> >>> For some reason I understood that the device will support thousands of >>> queues and single session per queue which would have needed more app >>> changes. >>> >>> On 23/03/2017 08:06, akhil.goyal@nxp.com wrote: >>>> From: Akhil Goyal >>>> >>>> adding support for attaching session to queue pairs. >>>> This is required as underlying crypto driver may only >>>> support limited number of sessions per queue pair >>>> if max_nb_sessions_per_qp > 0, session should be >>>> attached to a particular qp. >>>> >>>> Signed-off-by: Akhil Goyal >>>> --- >>>> examples/ipsec-secgw/ipsec.c | 12 ++++++++++++ >>>> 1 file changed, 12 insertions(+) >>>> >>>> diff --git a/examples/ipsec-secgw/ipsec.c >>>> b/examples/ipsec-secgw/ipsec.c >>>> index 144f0aa..b35b30f 100644 >>>> --- a/examples/ipsec-secgw/ipsec.c >>>> +++ b/examples/ipsec-secgw/ipsec.c >>>> @@ -47,6 +47,7 @@ >>>> static inline int >>>> create_session(struct ipsec_ctx *ipsec_ctx __rte_unused, struct >>>> ipsec_sa *sa) >>>> { >>>> + struct rte_cryptodev_info cdev_info; >>>> unsigned long cdev_id_qp = 0; >>>> int32_t ret; >>>> struct cdev_key key = { 0 }; >>>> @@ -73,6 +74,17 @@ create_session(struct ipsec_ctx *ipsec_ctx >>>> __rte_unused, struct ipsec_sa *sa) >>>> sa->crypto_session = rte_cryptodev_sym_session_create( >>>> ipsec_ctx->tbl[cdev_id_qp].id, sa->xforms); >>>> + rte_cryptodev_info_get(ipsec_ctx->tbl[cdev_id_qp].id, &cdev_info); >>>> + if (cdev_info.sym.max_nb_sessions_per_qp > 0) { >>>> + ret = rte_cryptodev_queue_pair_attach_sym_session( >>>> + ipsec_ctx->tbl[cdev_id_qp].qp, >>>> + sa->crypto_session); >>>> + if (ret < 0) { >>>> + RTE_LOG(ERR, IPSEC, "Session cannot be attached" >>>> + " to qp %u ", ipsec_ctx->tbl[cdev_id_qp].qp); >>> >>> Guideline is to keep error strings in single line to facilitate grep. >>> Other than that: >>> >>> Acked-by: Sergio Gonzalez Monroy >>> >>>> + return -1; >>>> + } >>>> + } >>>> sa->cdev_id_qp = cdev_id_qp; >>>> return 0; >>> >>> >>> >> Hi Sergio, >> >> Similar error string is mentioned above my change also in the same >> function. I deliberately did that as per the error strings in the file. >> >> Thanks, >> Akhil >> > > That means that we need to update those strings too, but new code should > try to keep error string in single line. > > Thanks, > Sergio > Ok I would correct the string. Will send the v3 with this change only if there is no more comments from Fiona. Thanks, Akhil