Immediately after doing an upgrade from 11.1-release to 11.2-release the network connection using em0 constantly goes up and down. It's driving an Intel 82574 card on PCI-E. I've tried rebooting, power cycling, using different cables, switch ports and a different switch with no change. Prior everything was stable for many moons in the same configuration. Switched to using the onboard Realtek in the system and that is also stable with the same cable/switch.
Sep 4 20:37:20 itsirk kernel: em0: link state changed to UP
Sep 4 20:37:20 itsirk kernel: em0: link state changed to DOWN
Sep 4 20:37:27 itsirk kernel: em0: link state changed to UP
Sep 4 20:37:27 itsirk kernel: em0: link state changed to DOWN
Sep 4 20:37:31 itsirk kernel: em0: link state changed to UP
Sep 4 20:37:31 itsirk kernel: em0: link state changed to DOWN
em0: <Intel(R) PRO/1000 Network Connection 7.6.1-k> port 0xdf00-0xdf1f mem 0xfdcc0000-0xfdcdffff,0xfdc00000-0xfdc7ffff,0xfdcfc000-0xfdcfffff irq 19 at device 0.0 on pci2
em0: Using MSIX interrupts with 3 vectors
em0: Ethernet address: 00:1b:21:39:dc:5f
em0: netmap queues/slots: TX 1/1024, RX 1/1024
em0@pci0:3:0:0: class=0x020000 card=0xa01f8086 chip=0x10d38086 rev=0x00 hdr=0x00
vendor = 'Intel Corporation'
device = '82574L Gigabit Network Connection'
class = network
subclass = ethernet
cap 01[c8] = powerspec 2 supports D0 D3 current D0
cap 05[d0] = MSI supports 1 message, 64 bit
cap 10[e0] = PCI-Express 1 endpoint max data 128(256) RO NS
link x1(x1) speed 2.5(2.5) ASPM disabled(L0s/L1)
cap 11[a0] = MSI-X supports 5 messages, enabled
Table in map 0x1c[0x0], PBA in map 0x1c[0x2000]
ecap 0001 = AER 1 0 fatal 0 non-fatal 0 corrected
ecap 0003 = Serial 1 001b21ffff39dc5f
Hi, it seems to be, a problem looks at similar on my case: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=231169
from intel manual, try this:
Important system configuration changes:
When there is a choice run on a 64bit OS rather than 32, it makes a
significant difference in improvement.
The interface can generate high number of interrupts. To avoid running
into the limit set by the kernel, adjust hw.intr_storm_threshold
setting using sysctl:
sysctl hw.intr_storm_threshold=9000 (the default is 1000)
For this change to take effect on boot, edit /etc/sysctl.conf and add the
If you still see Interrupt Storm detected messages, increase the limit to a
Best throughput results are seen with a large MTU; use 9000 if possible.
Weird behaviour today. I left the cable unplugged over night. Plugged it in this morning and the interface worked properly. It worked fine for an hour while I watched and poked at it, passed IPv6 packets fine, picked up IPv4 DHCP, forgot to test otherwise. Then I downed the realtek interface and things seemed to stop working on the intel interface as well. Downed and upped the intel interface and it came back to life, but the link state flapping started again. Trying all the usual things isn't making it go away again either.
Per Compri, I am using 64bit OS, not seeing Interrupt Storm messages, don't think it's interrupt rate related as there is hardly any traffic flowing when this starts, but I did try upping the setting, which had no effect.
Tried a different 82574 card I had and the behaviour was the same.
As you use DHCP, this problem seems to be duplicate of already resolved problem https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=229432. Take a look for a workaround and solution.
Close per comment 5
@Steven Please re-open this issue with additional information if the workaround described in bug 229432 does not resolve the problem.
*** This bug has been marked as a duplicate of bug 229432 ***