From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jesper Dangaard Brouer via iovisor-dev Subject: Re: README: [PATCH RFC 11/11] net/mlx5e: XDP TX xmit more Date: Tue, 13 Sep 2016 18:47:38 +0200 Message-ID: <20160913184738.706e2d91@redhat.com> References: <1473252152-11379-1-git-send-email-saeedm@mellanox.com> <1473252152-11379-12-git-send-email-saeedm@mellanox.com> <20160908101147.1b351432@redhat.com> <20160909032202.GA62966@ast-mbp.thefacebook.com> <20160912121530.4b4f0ad7@redhat.com> <1473782310.18970.138.camel@edumazet-glaptop3.roam.corp.google.com> Reply-To: Jesper Dangaard Brouer Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Cc: Tom Herbert , iovisor-dev , Jamal Hadi Salim , Saeed Mahameed , Eric Dumazet , Linux Netdev List , Edward Cree To: Eric Dumazet Return-path: In-Reply-To: <1473782310.18970.138.camel-XN9IlZ5yJG9HTL0Zs8A6p+yfmBU6pStAUsxypvmhUTTZJqsBc5GL+g@public.gmane.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: iovisor-dev-bounces-9jONkmmOlFHEE9lA1F8Ukti2O/JbrIOy@public.gmane.org Errors-To: iovisor-dev-bounces-9jONkmmOlFHEE9lA1F8Ukti2O/JbrIOy@public.gmane.org List-Id: netdev.vger.kernel.org On Tue, 13 Sep 2016 08:58:30 -0700 Eric Dumazet wrote: > We also care about icache pressure, and GRO/TSO already provides > bundling where it is applicable, without adding insane complexity in > the stacks. Sorry, I cannot resist. The GRO code is really bad regarding icache pressure/usage, due to how everything is function pointers calling function pointers, even if the general case is calling the function defined just next to it in the same C-file (which usually cause inlining). I can easily get 10% more performance for UDP use-cases by simply disabling the GRO code, and I measure a significant drop in icache-misses. Edward's solution should lower icache pressure. -- Best regards, Jesper Dangaard Brouer MSc.CS, Principal Kernel Engineer at Red Hat Author of http://www.iptv-analyzer.org LinkedIn: http://www.linkedin.com/in/brouer