From mboxrd@z Thu Jan 1 00:00:00 1970 From: Martin Lau Subject: Re: [Question] net/mlx4_en: Memory consumption issue with mlx4_en driver Date: Thu, 12 Mar 2015 09:56:47 -0700 Message-ID: <20150312165647.GD1032293@devbig242.prn2.facebook.com> References: <20150311185146.GA1032293@devbig242.prn2.facebook.com> <1426105262.11398.66.camel@edumazet-glaptop2.roam.corp.google.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Cc: Amir Vadai , Or Gerlitz , , To: Eric Dumazet Return-path: Received: from mx0b-00082601.pphosted.com ([67.231.153.30]:58375 "EHLO mx0b-00082601.pphosted.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932487AbbCLQ4z (ORCPT ); Thu, 12 Mar 2015 12:56:55 -0400 Content-Disposition: inline In-Reply-To: <1426105262.11398.66.camel@edumazet-glaptop2.roam.corp.google.com> Sender: netdev-owner@vger.kernel.org List-ID: On Wed, Mar 11, 2015 at 01:21:02PM -0700, Eric Dumazet wrote: > You know, even the order-3 allocations done for regular skb allocations > will hurt you : a single copybreaked skb stored a long time in a tcp > receive queue will hold 32KB of memory. > > Even 4KB can lead to disasters. Thanks for the pointer. I look a little deeper at the allocation in skbuff.c. I can see your point. > You could lower tcp_rmem so that collapsing happens sooner. It is what we did. However, a buggy process accommodated enough stalled sockets (stop reading from it but not closing it) will re-surface the problem. Thanks, --Martin