Date: Sun, 2 Oct 2011 21:39:53 -0400 From: Arnaud Lacombe <lacombar@gmail.com> To: Jack Vogel <jfvogel@gmail.com> Cc: "freebsd-net@freebsd.org" <freebsd-net@freebsd.org>, Hooman Fazaeli <fazaeli@sepehrs.com> Subject: Re: em(4) high latency w/o msix Message-ID: <CACqU3MU3=5o8v7WbSaPEYzkhu8vxjypfwyLA7frU6Z77WAUP_A@mail.gmail.com> In-Reply-To: <CACqU3MWundT7h4xT9SSGzLGDc6QN_6FyOfy=juG4A1NKH0SGMw@mail.gmail.com> References: <4E886A3B.7000502@sepehrs.com> <CAFOYbcnUbdrCYyDdSeYOWVNL_vhH_zA1d5B=mgrkoUijHXj-SQ@mail.gmail.com> <CACqU3MWundT7h4xT9SSGzLGDc6QN_6FyOfy=juG4A1NKH0SGMw@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi, On Sun, Oct 2, 2011 at 8:29 PM, Arnaud Lacombe <lacombar@gmail.com> wrote: > Hi, > > On Sun, Oct 2, 2011 at 12:08 PM, Jack Vogel <jfvogel@gmail.com> wrote: >> On what hardware? >> > Only the 82574 is using em(4)'s MSI-X. > FWIW, I'm not seeing this issue on a 82574 box I've access to: result of the following command: # ping -nqt 30 -f <ip> The two box are connected on an old 10/100 switch MSI-X enabled: 239361 packets transmitted, 239358 packets received, 0.0% packet loss round-trip min/avg/max/stddev =3D 0.090/0.117/2.488/0.027 ms MSI-X disabled: 233122 packets transmitted, 233119 packets received, 0.0% packet loss round-trip min/avg/max/stddev =3D 0.096/0.120/2.589/0.027 ms On another interface (same boxes involved, but through a 10/100/1000 switch= ): MSI-X enabled: 240080 packets transmitted, 240079 packets received, 0.0% packet loss round-trip min/avg/max/stddev =3D 0.090/0.116/0.175/0.008 ms MSI-X disabled: 207310 packets transmitted, 207309 packets received, 0.0% packet loss round-trip min/avg/max/stddev =3D 0.095/0.136/0.310/0.044 ms However, there seem to be a performance drop with MSI-X disabled and Gigabit link. - Arnaud >> Jack >> >> >> On Sun, Oct 2, 2011 at 6:42 AM, Hooman Fazaeli <fazaeli@sepehrs.com> wro= te: >> >>> >>> Latest em(4) driver from HEAD seems to have high latency >>> when MSIX is disabled. >>> >>> With MSIX enabled (hw.em.enable_msix=3D1): >>> >>> # ping -c5 192.168.1.83 >>> PING 192.168.1.83 (192.168.1.83): 56 data bytes >>> 64 bytes from 192.168.1.83: icmp_seq=3D0 ttl=3D64 time=3D0.055 ms >>> 64 bytes from 192.168.1.83: icmp_seq=3D1 ttl=3D64 time=3D0.076 ms >>> 64 bytes from 192.168.1.83: icmp_seq=3D2 ttl=3D64 time=3D0.066 ms >>> 64 bytes from 192.168.1.83: icmp_seq=3D3 ttl=3D64 time=3D0.051 ms >>> 64 bytes from 192.168.1.83: icmp_seq=3D4 ttl=3D64 time=3D0.063 ms >>> >>> --- 192.168.1.83 ping statistics --- >>> 5 packets transmitted, 5 packets received, 0.0% packet loss >>> round-trip min/avg/max/stddev =3D 0.051/0.062/0.076/0.009 ms >>> >>> With MSIX disabled: >>> >>> # ping -c5 192.168.1.83 >>> PING 192.168.1.83 (192.168.1.83): 56 data bytes >>> 64 bytes from 192.168.1.83: icmp_seq=3D0 ttl=3D64 time=3D0.180 ms >>> 64 bytes from 192.168.1.83: icmp_seq=3D1 ttl=3D64 time=3D0.164 ms >>> 64 bytes from 192.168.1.83: icmp_seq=3D2 ttl=3D64 time=3D0.169 ms >>> 64 bytes from 192.168.1.83: icmp_seq=3D3 ttl=3D64 time=3D0.172 ms >>> 64 bytes from 192.168.1.83: icmp_seq=3D4 ttl=3D64 time=3D0.167 ms >>> >>> --- 192.168.1.83 ping statistics --- >>> 5 packets transmitted, 5 packets received, 0.0% packet loss >>> round-trip min/avg/max/stddev =3D 0.164/0.170/0.180/0.005 ms >>> >>> As you see, w/o MSIX, RTT increases by a factor of 3. >>> >>> I also tested the following drivers: >>> =A0 =A0- igb(4) from HEAD: OK. >>> =A0 =A0- Stock 7.3-RELEASE: OK. >>> =A0 =A0- Stock 7.4-RELEASE: problem exist. >>> >>> Any ideas? >>> >>> >>> >>> >>> >> _______________________________________________ >> freebsd-net@freebsd.org mailing list >> http://lists.freebsd.org/mailman/listinfo/freebsd-net >> To unsubscribe, send any mail to "freebsd-net-unsubscribe@freebsd.org" >> >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CACqU3MU3=5o8v7WbSaPEYzkhu8vxjypfwyLA7frU6Z77WAUP_A>