From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: [PATCH 2/2] IB/ipoib: fix GRO merge failure for IPoIB originated TCP streams Date: Mon, 06 Feb 2012 16:21:46 +0100 Message-ID: <1328541706.2220.50.camel@edumazet-HP-Compaq-6005-Pro-SFF-PC> References: <1828884A29C6694DAF28B7E6B8A823733349A5D2@ORSMSX101.amr.corp.intel.com> <1328259640.2480.43.camel@edumazet-laptop> <1828884A29C6694DAF28B7E6B8A823733349A6C9@ORSMSX101.amr.corp.intel.com> <4F2FEC36.6090800@mellanox.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: "Hefty, Sean" , Herbert Xu , linux-rdma , Shlomo Pongratz , Roland Dreier , "netdev-u79uwXL29TY76Z2rM5mHXA@public.gmane.org" To: Or Gerlitz Return-path: In-Reply-To: <4F2FEC36.6090800-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org> Sender: linux-rdma-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-Id: netdev.vger.kernel.org Le lundi 06 f=C3=A9vrier 2012 =C3=A0 17:05 +0200, Or Gerlitz a =C3=A9cr= it : > On 2/3/2012 10:24 PM, Hefty, Sean wrote: > >> Hmm... a tcpdump would help to understand what is going on (for ex= ample if GRO kicks in) > > > > I should stop trying to apply patches so late. I must have done so= mething wrong with my quick test. Re-applying the patch on 3.2, the ip= oib UD performance looks okay to me with Eric's patch. Sorry about the= flub. > > >=20 > Sean, same here, with both 3.2 and 3.3-rc1 I see with tcpdump that no= =20 > GRO aggregation is done in the receiving side, when applying the patc= h I=20 > sent (Eric's patch enhanced to change ipoib code such that it sets=20 > gro_mac_header_lento the ipoib header len), I can see GRO aggregation= =20 > and the performance goes well up. So how do we go from here? Herbert,= =20 > Roland, are you okay with the patch? >=20 > Eric, if this we're going on that patch, could you please add=20 > Reported-by: Shlomo Pongratz to the change-log= =2E Sure, I suggest two patches then. One from me (the one I sent earlier), introducing the core infrastructure, and one from you adding the "dev->gro_mac_header_len =3D IPOIB_ENCAP_LEN;" in drivers/infiniband/ulp/ipoib/ipoib_main.c If you agree, I'll resend my part with official submission, and you'll send your patch after ? -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" i= n the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html