Date: Sun, 2 Feb 2014 21:56:44 -0800 From: hiren panchasara <hiren.panchasara@gmail.com> To: "freebsd-net@freebsd.org" <freebsd-net@freebsd.org>, Jack F Vogel <jfv@freebsd.org> Subject: Re: Errors using span interface on if_bridge(4) Message-ID: <CALCpEUEZmMvVRAf8DtRBpPtYgZDB3ufNaAhDpmrnVaiPXv3veg@mail.gmail.com> In-Reply-To: <CALCpEUH-5vWoU-Q-fWsmUGLrS3H--4WXcbejf1TQc=ZNmFOXvQ@mail.gmail.com> References: <CALCpEUH7YHhG3gw0wrObsJux=PWmJu2K6_tzCbCfXQ21ENYHdw@mail.gmail.com> <CALCpEUEuXR7_Q8_zt2TjnsOqt_rYMH5do=UEjssqkj684Xdimg@mail.gmail.com> <CALCpEUFLwZ6vismFKgyo1Q_P4iwzxS%2BjBozTV%2BO1kvap6U8brQ@mail.gmail.com> <CALCpEUE4V-CXkT3R0tzr1K075RBX_HKHPF%2BP-WTT15FkH5cuQA@mail.gmail.com> <CALCpEUH-5vWoU-Q-fWsmUGLrS3H--4WXcbejf1TQc=ZNmFOXvQ@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, Feb 2, 2014 at 7:32 PM, hiren panchasara <hiren.panchasara@gmail.com> wrote: > On Sun, Feb 2, 2014 at 11:19 AM, hiren panchasara > <hiren.panchasara@gmail.com> wrote: >> On Sat, Feb 1, 2014 at 5:16 PM, hiren panchasara >> <hiren.panchasara@gmail.com> wrote: >> >>> >>> -bash-4.2$ sysctl -a | grep mac_stats.checksum_errs >>> dev.ix.0.mac_stats.checksum_errs: 0 >>> dev.ix.1.mac_stats.checksum_errs: 0 >>> dev.ix.2.mac_stats.checksum_errs: 0 >>> dev.ix.3.mac_stats.checksum_errs: 3371119 <-- most of them fail checksum >> >> In ixgbe.c : >> ixgbe_add_hw_stats() has >> SYSCTL_ADD_UQUAD(ctx, stat_list, OID_AUTO, "checksum_errs", >> CTLFLAG_RD, &stats->xec, >> "Checksum Errors"); >> >> which is updated in ixgbe_update_stats_counters() >> adapter->stats.xec += IXGBE_READ_REG(hw, IXGBE_XEC); >> >> ixgbe_type.h defines >> #define IXGBE_XEC 0x04120 >> >> So, this is something that firmware of the card updates? We just read >> that register from drivers. > > ix3@pci0:66:0:1: class=0x020000 card=0x7a118086 chip=0x10fb8086 > rev=0x01 hdr=0x00 > vendor = 'Intel Corporation' > device = '82599EB 10-Gigabit SFI/SFP+ Network Connection' > class = network > subclass = ethernet > > I found following in 82599 controller spec update: > http://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/82599-10-gbe-controller-spec-update.pdf#G2.1474004 > http://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/82599-10-gbe-controller-spec-update.pdf#G2.2301787 > > Thought here I am sending unicast tcp traffic with iperf3. I tried to use netperf to generate smaller size packets (64 bytes) and with that I saw only 0.004% packets reporting checksum errors. $ netperf -H 10.73.149.91 -t TCP_STREAM -- -m 64 -D TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 10.73.149.91 () port 0 AF_INET : nodelay : histogram : interval : dirty data : demo Recv Send Send Socket Socket Message Elapsed Size Size Size Time Throughput bytes bytes bytes secs. 10^6bits/sec 65536 65536 64 10.00 56.92 Anything larger was showing 93% packets with checksum error. (as I've been reporting in this thread). cheers, Hiren
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CALCpEUEZmMvVRAf8DtRBpPtYgZDB3ufNaAhDpmrnVaiPXv3veg>