From owner-freebsd-stable@freebsd.org Tue Apr 9 11:33:10 2019 Return-Path: Delivered-To: freebsd-stable@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2610:1c1:1:606c::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id B4F20157E016 for ; Tue, 9 Apr 2019 11:33:10 +0000 (UTC) (envelope-from mach@swishmail.com) Received: from vorlon.swishmail.com (vorlon.swishmail.com [208.72.56.19]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (Client did not present a certificate) by mx1.freebsd.org (Postfix) with ESMTPS id 6E959897DE for ; Tue, 9 Apr 2019 11:33:09 +0000 (UTC) (envelope-from mach@swishmail.com) Received: (qmail 15244 invoked by uid 89); 9 Apr 2019 11:33:02 -0000 Received: from unknown (HELO ?IPv6:2001:b030:14e:100:959b:801a:9ec8:14a1?) (mach@swishmail.com@2001:b030:14e:100:959b:801a:9ec8:14a1) by 2602:ffb8::208:72:56:19 with ESMTPSA (ECDHE-RSA-AES128-GCM-SHA256 encrypted, authenticated); 9 Apr 2019 11:33:02 -0000 Subject: Re: em performs worse than igb (latency wise) in 12? To: freebsd-stable@freebsd.org References: <7673edad-1e50-7e9b-961e-f28ab7a0f41e@ingresso.co.uk> <4f9b9259-f5a1-ecc6-366e-4a26de0ca3dc@protected-networks.net> From: Kris von Mach Message-ID: <2af4d1fe-1581-baa7-e38e-bba631639deb@swishmail.com> Date: Tue, 9 Apr 2019 19:33:00 +0800 User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:60.0) Gecko/20100101 Thunderbird/60.6.1 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: 7bit Content-Language: en-US X-Rspamd-Queue-Id: 6E959897DE X-Spamd-Bar: + Authentication-Results: mx1.freebsd.org; spf=pass (mx1.freebsd.org: domain of mach@swishmail.com designates 208.72.56.19 as permitted sender) smtp.mailfrom=mach@swishmail.com X-Spamd-Result: default: False [1.29 / 15.00]; ARC_NA(0.00)[]; FROM_HAS_DN(0.00)[]; R_SPF_ALLOW(-0.20)[+ip4:208.72.56.0/22]; TO_MATCH_ENVRCPT_ALL(0.00)[]; MIME_GOOD(-0.10)[text/plain]; MIME_TRACE(0.00)[0:+]; TO_DN_NONE(0.00)[]; NEURAL_SPAM_MEDIUM(0.44)[0.444,0]; RCPT_COUNT_ONE(0.00)[1]; NEURAL_HAM_LONG(-0.30)[-0.305,0]; RCVD_COUNT_THREE(0.00)[3]; DMARC_NA(0.00)[swishmail.com]; NEURAL_SPAM_SHORT(0.57)[0.571,0]; MX_GOOD(-0.01)[mxfilter1.nyc.swishmail.com,mxfilter2.nyc.swishmail.com]; IP_SCORE(-0.01)[country: US(-0.06)]; RCVD_IN_DNSWL_LOW(-0.10)[19.56.72.208.list.dnswl.org : 127.0.5.1]; R_DKIM_NA(0.00)[]; SUBJECT_ENDS_QUESTION(1.00)[]; ASN(0.00)[asn:14469, ipnet:208.72.56.0/22, country:US]; MID_RHS_MATCH_FROM(0.00)[]; RCVD_TLS_ALL(0.00)[]; FROM_EQ_ENVFROM(0.00)[] X-BeenThere: freebsd-stable@freebsd.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Production branch of FreeBSD source code List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 09 Apr 2019 11:33:10 -0000 On 4/7/2019 6:49 AM, Matthew Macy wrote: > On Sat, Apr 6, 2019 at 1:23 PM Michael Butler > wrote: >> I'd be interested to see if substituting the port net/intel-em-kmod >> has any effect on the issue, > I would as well. igb, em, and lem are all the same driver in 12. This > makes maintenance a lot easier. However, the older NICs have a lot of > errata workarounds that aren't explicitly commented as such. My first > guess is this card suffers from one such errata workaround that has > been dropped in the update. I've tried net/intel-em-kmod, it actually became worse went from 100 requests/sec to about 90. That makes sense about maintenance. Though I believe i350 is less than 5 years old, HP's 366FLR version is 4. So it's not that old, and at least in gigabit level nic, is one of the best afaik. Is there some other 1gig nic that is recommended for 12? Or is it time to switch to 10gbit? I've heard good things about Chelsio for 10gbit. I went back to 11-Stable for now.