Port Down State Diagnostics
The driver is able to extract information from the hardware in order to easily detect the reason for the port link to be in inactive state.
The driver supports providing a more detailed, human-readable reason for "port down" state as it is held by the firmware.
This information is available to administrators as the extended media report in normal ifconfig(8) status output.
# ifconfig mce0
mce0: flags=8843
<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0
mtu 1500
options=3ed07bb<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,JUMBO_MTU,VLAN_HWCSUM,TSO4,TSO6,LRO,VLAN_HWFILTER,VLAN_HWTSO,LINKSTATE,RXCSUM_IPV6,TXCSUM_IPV6,TXRTLMT,HWRXTSTMP>
ether 24
:8a:07
:a5:28
:70
inet6 fe80::268a:7ff:fea5:2870
%mce0 prefixlen 64
scopeid 0x6
media: Ethernet autoselect <full-duplex,rxpause,txpause> (100GBase-CR4 <full-duplex,rxpause,txpause>)
status: no carrier (FEC sync loss)
nd6 options=21
<PERFORMNUD,AUTO_LINKLOCAL>
Events report on module plugging and unplugging, as well as related error conditions. Unlike the process of "port down" state cause detection, events are reported upon change by emitting a kernel message.
For instance, when a successful detection of the newly inserted module occurs, the following message is printed.
mlx5_core0: INFO: mlx5_port_module_event:710:(pid 11): Module 0, status: plugged and enabled
Statistics for each supported module event type is available under the sysctl OID dev.mlx5_core.N.pme_stats.errors per type.