Date: Wed, 9 Sep 2009 04:40:26 -0700 (PDT) From: Barney Cordoba <barney_cordoba@yahoo.com> To: Mike Tancsa <mike@sentex.net> Cc: freebsd-net@freebsd.org Subject: Re: em driver input errors Message-ID: <992693.15985.qm@web63902.mail.re1.yahoo.com> In-Reply-To: <200909082218.n88MI7TH073975@lava.sentex.ca>
next in thread | previous in thread | raw e-mail | index | archive | help
=0A=0A--- On Tue, 9/8/09, Mike Tancsa <mike@sentex.net> wrote:=0A=0A> From:= Mike Tancsa <mike@sentex.net>=0A> Subject: Re: em driver input errors=0A> = To: "Barney Cordoba" <barney_cordoba@yahoo.com>=0A> Cc: freebsd-net@freebsd= .org=0A> Date: Tuesday, September 8, 2009, 6:21 PM=0A> At 05:42 PM 9/8/2009= , Barney Cordoba=0A> wrote:=0A> > Manish What specific kinds of input error= s are you=0A> getting? How many PPS are you doing, what is the size of the= =0A> ring, and the interrupt modulation rate? Are the NICs PCIe=0A> or PCIx= ? Barney=0A> =0A> In my case, our backup server (mix of dump via nfs and so= me=0A> dumps through ssh as well as writes via samba mounts) has a=0A> fair= ly low pps rate and starts to see input errors at about=0A> 100Mb.=A0 Addin= g=0A> =0A> hw.em.rxd=3D4096=0A> hw.em.txd=3D4096=0A> =0A> and=0A> =0A> net.= inet.tcp.recvbuf_max=3D16777216=0A> net.inet.tcp.recvspace=3D131072=0A> net= .inet.tcp.sendbuf_max=3D16777216=0A> net.inet.tcp.sendspace=3D131072=0A> ne= t.inet.udp.recvspace=3D65536=0A> kern.ipc.somaxconn=3D1024=0A> kern.ipc.max= sockbuf=3D4194304=0A> net.inet.ip.redirect=3D0=0A> net.inet.ip.intr_queue_m= axlen=3D4096=0A> net.route.netisr_maxqlen=3D1024=0A> kern.ipc.nmbclusters= =3D655360=0A> =0A> didnt seem to make a difference in the amount of errors = I=0A> was seeing=0A> =0A> =0A> em1@pci0:7:5:0: class=3D0x020000 card=3D0x34= 8f8086=0A> chip=3D0x10768086 rev=3D0x05 hdr=3D0x00=0A> =A0 =A0 vendor=A0 = =A0=A0=A0=3D 'Intel=0A> Corporation'=0A> =A0 =A0 device=A0 =A0=A0=A0=3D 'Gi= gabit=0A> Ethernet Controller (82541EI)'=0A> =A0 =A0 class=A0 =A0 =A0 =3D n= etwork=0A> =A0 =A0 subclass=A0=A0=A0=3D ethernet=0A> =A0 =A0 cap 01[dc] =3D= powerspec 2=A0 supports D0=0A> D3=A0 current D0=0A> =A0 =A0 cap 07[e4] =3D= PCI-X supports 2048 burst read,=0A> 1 split transaction=0A> =0A> Core(TM)= 2 Quad CPU=A0 =A0 Q6600=A0 @ 2.40GHz,=0A> AMD64, RELENG_7 from Jun 18th=0A>= =0A> Plenty of disk IO left and the CPUs doent seem to be taxed=0A> that m= uch.=0A> =0A> Sep=A0 8 00:02:01 backup3 kernel: em1: Excessive=0A> collisio= ns =3D 0=0A> Sep=A0 8 00:02:01 backup3 kernel: em1: Sequence errors =3D=0A>= 0=0A> Sep=A0 8 00:02:01 backup3 kernel: em1: Defer count =3D 0=0A> Sep=A0 = 8 00:02:01 backup3 kernel: em1: Missed Packets =3D=0A> 61316=0A> Sep=A0 8 0= 0:02:01 backup3 kernel: em1: Receive No=0A> Buffers =3D 0=0A> Sep=A0 8 00:0= 2:01 backup3 kernel: em1: Receive Length=0A> Errors =3D 0=0A> Sep=A0 8 00:0= 2:01 backup3 kernel: em1: Receive errors =3D=0A> 0=0A> Sep=A0 8 00:02:01 ba= ckup3 kernel: em1: Crc errors =3D 0=0A> Sep=A0 8 00:02:01 backup3 kernel: e= m1: Alignment errors=0A> =3D 0=0A> Sep=A0 8 00:02:01 backup3 kernel: em1: C= ollision/Carrier=0A> extension errors =3D 0=0A> Sep=A0 8 00:02:01 backup3 k= ernel: em1: RX overruns =3D=0A> 22397=0A> Sep=A0 8 00:02:01 backup3 kernel:= em1: watchdog timeouts=0A> =3D 0=0A> Sep=A0 8 00:02:01 backup3 kernel: em1= : RX MSIX IRQ =3D 0=0A> TX MSIX IRQ =3D 0 LINK MSIX IRQ =3D 0=0A> Sep=A0 8 = 00:02:01 backup3 kernel: em1: XON Rcvd =3D 0=0A> Sep=A0 8 00:02:01 backup3 = kernel: em1: XON Xmtd =3D 0=0A> Sep=A0 8 00:02:01 backup3 kernel: em1: XOFF= Rcvd =3D 0=0A> Sep=A0 8 00:02:01 backup3 kernel: em1: XOFF Xmtd =3D 0=0A> = Sep=A0 8 00:02:01 backup3 kernel: em1: Good Packets Rcvd=0A> =3D 544276980= =0A> Sep=A0 8 00:02:01 backup3 kernel: em1: Good Packets Xmtd=0A> =3D 49047= 5071=0A> Sep=A0 8 00:02:01 backup3 kernel: em1: TSO Contexts Xmtd=0A> =3D 0= =0A> Sep=A0 8 00:02:01 backup3 kernel: em1: TSO Contexts=0A> Failed =3D 0= =0A> =0A> =0A> pf is in the kernel as well=0A> =0A> > _____________________= __________________________=0A> > freebsd-net@freebsd.org=0A> mailing list= =0A> > http://lists.freebsd.org/mailman/listinfo/freebsd-net=0A>; > To unsub= scribe, send any mail to "freebsd-net-unsubscribe@freebsd.org"=0A> =0A> ---= -----------------------------------------------------------------=0A> Mike = Tancsa,=A0 =A0 =A0 =A0 =A0 =A0=0A> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=0A> =A0 = =A0 =A0 =A0 =A0 tel +1 519 651 3400=0A> Sentex Communications,=A0 =A0 =A0 = =A0 =A0=0A> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=0A> =A0 mike@sentex.net=0A> Pro= viding Internet since 1994=A0 =A0 =A0 =A0=0A> =A0 =A0 =A0 =A0 =A0 =A0 www.s= entex.net=0A> Cambridge, Ontario Canada=A0 =A0 =A0 =A0 =A0=0A> =A0 =A0 =A0 = =A0 =A0 =A0=0A> =A0=A0=A0www.sentex.net/mike=0A> =0A> =0A=0AThe 8241GI may = not be able to handle full gigabit flows if its only=0Awired at 32-bit 33Mh= z, which is only capable of bursting to 1Gb/s. With=0Aa single NIC it likel= y just fine, but it a bridged or firewall type =0Aconfig you may just be se= eing bus failures.=0A=0ABarney=0A=0A=0A=0A
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?992693.15985.qm>