From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from huawei.com (lhrrgout.huawei.com [194.213.3.17]) by dpdk.org (Postfix) with ESMTP id F069AA49F for ; Thu, 25 Jan 2018 15:47:58 +0100 (CET) Received: from lhreml708-cah.china.huawei.com (unknown [172.18.7.106]) by Forcepoint Email with ESMTP id 4A04FCB11AEF2 for ; Thu, 25 Jan 2018 14:47:57 +0000 (GMT) Received: from FRAEMA707-CHM.china.huawei.com (10.206.14.56) by lhreml708-cah.china.huawei.com (10.201.108.49) with Microsoft SMTP Server (TLS) id 14.3.361.1; Thu, 25 Jan 2018 14:47:58 +0000 Received: from FRAEML521-MBX.china.huawei.com ([169.254.1.156]) by FRAEMA707-CHM.china.huawei.com ([169.254.7.54]) with mapi id 14.03.0361.001; Thu, 25 Jan 2018 15:47:56 +0100 From: "Avi Cohen (A)" To: "users@dpdk.org" Thread-Topic: IPSEC-SECGW - type no-offload huge latency Thread-Index: AdOV615cTtrcXnPgQiaPpNVIsBEVRA== Date: Thu, 25 Jan 2018 14:47:56 +0000 Message-ID: Accept-Language: en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [10.200.202.108] Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 X-CFilter-Loop: Reflected Subject: Re: [dpdk-users] IPSEC-SECGW - type no-offload huge latency X-BeenThere: users@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK usage discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 25 Jan 2018 14:47:59 -0000 I see that the delay is derived from the MAX_PKT_BURST=20 If I decrease it, the delay decreases accordingly. And we always transmit= a burst of pkts (=3D=3DMAX_PKT_BURST) But this behavior is wrong because we should transmit if there are ant pkt= s ready for tx. I suspect that the delay is in the ipsec_dequeue logic ... Any idea ? Best Regards Avi > -----Original Message----- > From: Avi Cohen (A) > Sent: Tuesday, 23 January, 2018 5:25 PM > To: 'users@dpdk.org' > Subject: IPSEC-SECGW - type no-oofload huge latency >=20 > Hi > I'm running ipsec-secgw with type no-offload (default mode) Clear text pk= ts > (ping) are recived , encrypted and send to the end tunnel , then it is de= crypted > and send to the other port at the peer. > So everything looks OK - except that I get huge latency of 10s of seconds= . > Btw - this application in mode inline-crypto - is working perfect Any id= ea why is > the latency ? > Best Regards > Avi