em(4) cannot maintain link in 12.0-STABLE

2019-04-22 Thread R. Tyler Croy


I have two NICs in a machine which acts as a gateway, and with the upgrade from
11-STABLE to the latest 12.0-STABLE tree (r346473), neither em(4) based NIC
will maintain a connection.

The behavior is similar to what is described in the last couple comments of
this ticket: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=219428 

The behavior is identical with both devices. Link is established, dhclient
starts securing a DHCP lease, and then once traffic starts to flow, the link
switches to DOWN. Then back to UP, and back and forth.

A syslog snippet:

Apr 21 20:03:42 strawberry kernel: em0: Link is up 100 Mbps Full Duplex
Apr 21 20:03:42 strawberry kernel: em0: link state changed to UP
Apr 21 20:03:46 strawberry dhclient[15534]: New IP Address (em0): 
173.228.83.96
Apr 21 20:03:46 strawberry dhclient[16069]: New Subnet Mask (em0): 
255.255.255.0
Apr 21 20:03:46 strawberry dhclient[18914]: New Broadcast Address (em0): 
173.228.83.255
Apr 21 20:03:46 strawberry dhclient[21483]: New Routers (em0): 173.228.83.1
Apr 21 20:03:48 strawberry kernel: em0: link state changed to DOWN
Apr 21 20:03:50 strawberry kernel: em0: Link is up 100 Mbps Full Duplex
Apr 21 20:03:50 strawberry kernel: em0: link state changed to UP
Apr 21 20:03:57 strawberry kernel: em0: link state changed to DOWN
Apr 21 20:03:57 strawberry ntpd[75785]: error resolving pool 
0.freebsd.pool.ntp.org: Name does not resolve (8)
Apr 21 20:03:58 strawberry dhclient[38043]: send_packet: Network is down
Apr 21 20:03:59 strawberry kernel: em0: Link is up 100 Mbps Full Duplex
Apr 21 20:03:59 strawberry kernel: em0: link state changed to UP
Apr 21 20:04:05 strawberry kernel: em0: link state changed to DOWN
Apr 21 20:04:06 strawberry dhclient[38043]: send_packet: Network is down
Apr 21 20:04:07 strawberry kernel: em0: Link is up 100 Mbps Full Duplex
Apr 21 20:04:07 strawberry kernel: em0: link state changed to UP
Apr 21 20:04:15 strawberry kernel: em0: link state changed to DOWN
Apr 21 20:04:15 strawberry dhclient[38043]: send_packet: Network is down
Apr 21 20:04:16 strawberry kernel: em0: Link is up 100 Mbps Full Duplex

THe version/hardware details:

FreeBSD strawberry 12.0-STABLE FreeBSD 12.0-STABLE r346473 GENERIC  amd64


em0@pci0:7:0:0: class=0x02 card=0xa01f8086 chip=0x10d38086 rev=0x00 
hdr=0x00
vendor = 'Intel Corporation'
device = '82574L Gigabit Network Connection'
class  = network
subclass   = ethernet
em1@pci0:8:0:0: class=0x02 card=0xa01f8086 chip=0x10d38086 rev=0x00 
hdr=0x00
vendor = 'Intel Corporation'
device = '82574L Gigabit Network Connection'
class  = network
subclass   = ethernet


My temporary workaround has been to use a USB-based NIC, but I am wondering if
there are any tunables I could tweak to get this misbehaving driver to mind its
manners?



Cheers

--
GitHub:  https://github.com/rtyler

GPG Key ID: 0F2298A980EE31ACCA0A7825E5C92681BEF6CEA2
___
freebsd-stable@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"


Peculiar failure of re(4) to active a link after a RAM upgrade

2018-10-10 Thread R. Tyler Croy
I recently upgraded my 11.2-STABLE machine from 4 to 16GB of memory. Upon
reboot the dual-NIC re(4) card which had been previously working failed to
either negotiate or activate a link (i.e. lights up :)) with my DSL modem.
ifcnofig lists the interface as 'no carrier'

I verified that both ports (re0 and re1) were working by plugging network
cables into my switch, and found that both were able to activate a link
properly.

When I plugin the DSL modem's cable into my em(4) based NIC, that activates a
link properly. Regardless of which re(4) NIC I stick the cable into, it
fails to create a link (ifconfig lists em0 as 'active'), while the same cable
works properly in an em(4) NIC.


I'm running r339079 right now, and if I remember correctly, had upgraded the
world prior to the RAM upgrade, so it _seems_ like the RAM upgrade is the only
variable which changed.

I'm really clueless as to what would even be happening underneath the covers
here, and would appreciate any pointers in the right direction for debugging
here :)


Cheers,
-R Tyler Croy

--
GitHub:  https://github.com/rtyler
Twitter: https://twitter.com/agentdero


signature.asc
Description: PGP signature