Date: Mon, 5 Jun 2017 13:06:51 -0700 From: Navdeep Parhar <np@FreeBSD.org> To: Mike Tancsa <mike@sentex.net>, "pholes@sentex.net" <pholes@sentex.net>, svn-src-stable-11@freebsd.org Subject: Re: svn commit: r319269 - in stable/11/sys: conf dev/cxgbe/firmware modules/cxgbe/t4_firmware modules/cxgbe/t5_firmware modules/cxgbe/t6_firmware Message-ID: <20a81b3e-17c7-df59-dda8-370d93eec899@FreeBSD.org> In-Reply-To: <ffa2d0ed-0306-e203-2b23-b9e0ac7ca927@sentex.net> References: <201705310016.v4V0G5n7012861@repo.freebsd.org> <ffa2d0ed-0306-e203-2b23-b9e0ac7ca927@sentex.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On 06/02/2017 14:16, Mike Tancsa wrote: > > Hi, > Not sure if its the firmware for sure, but a number of DAC cables I > have been using without issue all of a sudden are failing. On the > switch port side, the link does not go down, but on the NIC side, the > link is bouncing up and down once a bit of traffic is place on it. I The firmware handles all transceiver setup and link so it must be the firmware, but passing traffic should not affect the link. Does the link really stay up steadily as long as there's no traffic? Once the link has bounced at least once (with or without traffic), run this and provide its output: # sysctl -n dev.t5nex.0.misc.devlog > swapped out the DAC cable for a plain old fiber SFP+ and its OK. The > same DAC cable works fine prior to this update > > Jun 2 17:07:26 intels12004u kernel: cxl1: link state changed to DOWN > Jun 2 17:07:27 intels12004u kernel: cxl1: link state changed to UP > Jun 2 17:08:50 intels12004u kernel: cxl1: link state changed to DOWN > Jun 2 17:08:50 intels12004u kernel: cxl1: link state changed to UP > Jun 2 17:09:08 intels12004u kernel: cxl1: link state changed to DOWN > Jun 2 17:09:08 intels12004u kernel: cxl1: link state changed to UP > Jun 2 17:10:52 intels12004u kernel: cxl1: link state changed to DOWN > Jun 2 17:10:53 intels12004u kernel: cxl1: link state changed to UP > Jun 2 17:12:15 intels12004u kernel: cxl1: link state changed to DOWN > Jun 2 17:12:15 intels12004u kernel: cxl1: link state changed to UP > 0(intels12004u)# ifconfig cxl1 > cxl1: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500 > > options=ec07bb<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,JUMBO_MTU,VLAN_HWCSUM,TSO4,TSO6,LRO,VLAN_HWTSO,LINKSTATE,RXCSUM_IPV6,TXCSUM_IPV6> > ether 00:07:43:28:95:28 > hwaddr 00:07:43:28:95:28 > inet 3.3.3.1 netmask 0xffffff00 broadcast 3.3.3.255 > nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL> > media: Ethernet 10Gbase-SR <full-duplex> > status: active > plugged: SFP/SFP+/SFP28 10G Base-SR (LC) > vendor: SOURCEPHOTONICS PN: SPP10ESRCDFB SN: G0611090002 DATE: > 2010-09-19 > module temperature: 20.35 C Voltage: 3.37 Volts > RX: 0.00 mW (-40.00 dBm) TX: 6.55 mW (8.16 dBm) > 0(intels12004u)# > > it for whatever reason shows it as fibre, but its twinax Please provide the full dump of the module's eeprom? # ifconfig -vvv cxl0/1 It is possible that the transceiver misidentifies itself and the firmware sets it up incorrectly as a result. It may have worked by accident in the past. Regards, Navdeep
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20a81b3e-17c7-df59-dda8-370d93eec899>