From mboxrd@z Thu Jan 1 00:00:00 1970 From: Florian Fainelli Subject: Re: [PATCH] bgmac: Remove all offloading features, including GRO. Date: Fri, 15 Sep 2017 23:04:29 -0700 Message-ID: <0CEBE536-2C18-4ED2-90CB-3836CEF9A70E@gmail.com> References: <20170916002305.2794-1-rosenp@gmail.com> <1505522096.29839.13.camel@edumazet-glaptop3.roam.corp.google.com> <1505522322.3726.0.camel@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Cc: netdev@vger.kernel.org To: rosenp@gmail.com, Eric Dumazet Return-path: Received: from mail-oi0-f68.google.com ([209.85.218.68]:34683 "EHLO mail-oi0-f68.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751182AbdIPGEe (ORCPT ); Sat, 16 Sep 2017 02:04:34 -0400 Received: by mail-oi0-f68.google.com with SMTP id v66so521088oig.1 for ; Fri, 15 Sep 2017 23:04:34 -0700 (PDT) In-Reply-To: <1505522322.3726.0.camel@gmail.com> Sender: netdev-owner@vger.kernel.org List-ID: On September 15, 2017 5:38:42 PM PDT, rosenp@gmail=2Ecom wrote: >I have not=2E Unfortunately I own no gigabit hardware to test this on=2E >The MIPS CPU runs at 300MHz on my unit=2E > bgmac is used on Gigabit capable hardware, like Northstar and Northstar Pl= us, and others too, so unless you can get access to such HW or get confirma= tion from someone that your patches changes something, I would just drop th= is change and not bother=2E This is already not 100mbits/sec linerate=2E=2E= =2E >On Fri, 2017-09-15 at 17:34 -0700, Eric Dumazet wrote: >> On Fri, 2017-09-15 at 17:23 -0700, Rosen Penev wrote: >> > On a linksys E1200v1 (actually a crossflashed E1000v2), the >> > offloading features give no measurable benefit to speed or latency=2E >> > Furthermore, disabling GRO actually improves iperf performance by a >> > whoppimg 3mbps=2E Results: >> >=20 >> > Currently: >> >=20 >> > v2: Changed napi_gro_receive to netif_receive_skb=2E Seems to have an >> > identical result=2E >> >=20 >> > Signed-off-by: Rosen Penev >> > --- >> > drivers/net/ethernet/broadcom/bgmac=2Ec | 2 +- >> > 1 file changed, 1 insertion(+), 1 deletion(-) >> >=20 >> > diff --git a/drivers/net/ethernet/broadcom/bgmac=2Ec >> > b/drivers/net/ethernet/broadcom/bgmac=2Ec >> > index 48d672b204a4=2E=2E1fb0053aeee7 100644 >> > --- a/drivers/net/ethernet/broadcom/bgmac=2Ec >> > +++ b/drivers/net/ethernet/broadcom/bgmac=2Ec >> > @@ -483,7 +483,7 @@ static int bgmac_dma_rx_read(struct bgmac >> > *bgmac, struct bgmac_dma_ring *ring, >> > skb->protocol =3D eth_type_trans(skb, bgmac- >> > >net_dev); >> > bgmac->net_dev->stats=2Erx_bytes +=3D len; >> > bgmac->net_dev->stats=2Erx_packets++; >> > - napi_gro_receive(&bgmac->napi, skb); >> > + netif_receive_skb(skb); >> > handled++; >> > } while (0); >> > =20 >>=20 >> And have you tested 1Gbit link speed ? >> ( Or 2=2E5 Gbit link speed ) >>=20 >> If you want to disable GRO on your host, fine : you can use ethtool >> -K >>=20 >>=20 >>=20 (please don't top-post) --=20 Florian