From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga02.intel.com (mga02.intel.com [134.134.136.20]) by dpdk.org (Postfix) with ESMTP id 658FD1B5EE; Fri, 23 Nov 2018 23:43:31 +0100 (CET) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from fmsmga007.fm.intel.com ([10.253.24.52]) by orsmga101.jf.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 23 Nov 2018 14:43:30 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.56,271,1539673200"; d="scan'208";a="88217230" Received: from fmsmsx104.amr.corp.intel.com ([10.18.124.202]) by fmsmga007.fm.intel.com with ESMTP; 23 Nov 2018 14:43:30 -0800 Received: from fmsmsx157.amr.corp.intel.com (10.18.116.73) by fmsmsx104.amr.corp.intel.com (10.18.124.202) with Microsoft SMTP Server (TLS) id 14.3.408.0; Fri, 23 Nov 2018 14:43:29 -0800 Received: from fmsmsx118.amr.corp.intel.com ([169.254.1.160]) by FMSMSX157.amr.corp.intel.com ([169.254.14.68]) with mapi id 14.03.0415.000; Fri, 23 Nov 2018 14:43:29 -0800 From: "Wiles, Keith" To: "Zhang, Qi Z" CC: "Richardson, Bruce" , "dev@dpdk.org" , "Lu, Wenzhuo" , "Iremonger, Bernard" , "stable@dpdk.org" Thread-Topic: [PATCH] app/testpmd: improve MAC swap performance Thread-Index: AQHUgIu2WDX9Iwc3dkmCSNxM+XXwtKVegLqA Date: Fri, 23 Nov 2018 22:43:28 +0000 Message-ID: <8DB80D71-6025-4FDF-B039-D898A0522279@intel.com> References: <20181120044537.9495-1-qi.z.zhang@intel.com> In-Reply-To: <20181120044537.9495-1-qi.z.zhang@intel.com> Accept-Language: en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [10.255.228.125] Content-Type: text/plain; charset="us-ascii" Content-ID: <72E092556B07574C805B32772EF2104B@intel.com> Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Subject: Re: [dpdk-dev] [PATCH] app/testpmd: improve MAC swap performance X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 23 Nov 2018 22:43:32 -0000 > On Nov 19, 2018, at 10:45 PM, Zhang, Qi Z wrote: >=20 > The patch optimizes the mac swap operation by taking advantage > of SSE instructions, it only impacts x86 platform. >=20 > Cc: stable@dpdk.org >=20 > Signed-off-by: Qi Zhang > --- > app/test-pmd/macswap.c | 16 +++++++++++++++- > 1 file changed, 15 insertions(+), 1 deletion(-) >=20 > diff --git a/app/test-pmd/macswap.c b/app/test-pmd/macswap.c > index a8384d5b8..0722782b0 100644 > --- a/app/test-pmd/macswap.c > +++ b/app/test-pmd/macswap.c > @@ -78,7 +78,6 @@ pkt_burst_mac_swap(struct fwd_stream *fs) > struct rte_port *txp; > struct rte_mbuf *mb; > struct ether_hdr *eth_hdr; > - struct ether_addr addr; > uint16_t nb_rx; > uint16_t nb_tx; > uint16_t i; > @@ -95,6 +94,15 @@ pkt_burst_mac_swap(struct fwd_stream *fs) > start_tsc =3D rte_rdtsc(); > #endif >=20 > +#ifdef RTE_ARCH_X86 > + __m128i addr; > + __m128i shfl_msk =3D _mm_set_epi8(15, 14, 13, 12, > + 5, 4, 3, 2, > + 1, 0, 11, 10, > + 9, 8, 7, 6); I was playing around with these mask values and I was not able to make it w= ork as I expected. I ended up with different values in the mask. _mm_set_epi8(15, 14, 13, 12, 5, 4, 3, 2, 1, 0, 11, 10, 9, 8, 7, 6); After dumping the memory for a large number of tests this one seems correct= , can you verify your mask is correct? > +#else > + struct ether_addr addr; > +#endif > /* > * Receive a burst of packets and forward them. > */ > @@ -123,9 +131,15 @@ pkt_burst_mac_swap(struct fwd_stream *fs) > eth_hdr =3D rte_pktmbuf_mtod(mb, struct ether_hdr *); >=20 > /* Swap dest and src mac addresses. */ > +#ifdef RTE_ARCH_X86 > + addr =3D _mm_loadu_si128((__m128i *)eth_hdr); > + addr =3D _mm_shuffle_epi8(addr, shfl_msk); > + _mm_storeu_si128((__m128i *)eth_hdr, addr); > +#else > ether_addr_copy(ð_hdr->d_addr, &addr); > ether_addr_copy(ð_hdr->s_addr, ð_hdr->d_addr); > ether_addr_copy(&addr, ð_hdr->s_addr); > +#endif >=20 > mb->ol_flags &=3D IND_ATTACHED_MBUF | EXT_ATTACHED_MBUF; > mb->ol_flags |=3D ol_flags; > --=20 > 2.13.6 >=20 Regards, Keith