I'm getting huge Ierrs through netstat -i on my em device.
I really never update the machine, but I guess since last kernel/world update this happened. I really can't be very descriptive in that area since I'm not exactly sure when. I just tried to do some portupgrades today and noticed the awful receive speed of 15kb/sec.
I have a custom kernel with just altq / hz=1000 / polling enabled. Everything else is generic.
I've disabled polling.. hasn't helped.
I've removed rxcsum / txcsum.. hasn't helped.
Disabling pf rules hasn't helped.
Unloading pf module hasn't helped.
I've removed all sysctl tweaks of buffers etc.. nada
I've tried :
Still hasn't helped.
I've just recently compiled the smp em module from http://people.yandex-team.ru/~wawa/
I really can't seem to debug this machine. It gets HORRIBLE receiving.. yet tx is perfectly fine. Unfortunatly I can't debug the machine in person its on a colo out of state, but I assume the switch / cable is fine.
em0@pci0:0:8:0: class=0x020000 card=0x004e8086 chip=0x100e8086 rev=0x02 hdr=0x00
vendor = 'Intel Corporation'
device = 'Gigabit Ethernet Controller (82540EM)'
class = network
subclass = ethernet
cap 01[dc] = powerspec 2 supports D0 D3 current D0
cap 07[e4] = PCI-X supports 2048 burst read, 1 split transaction
cap 05[f0] = MSI supports 1 message, 64 bit
em0: link state changed to UP
em0: Excessive collisions = 0
em0: Sequence errors = 0
em0: Defer count = 0
em0: Missed Packets = 0
em0: Receive No Buffers = 0
em0: Receive Length Errors = 0
em0: Receive errors = 240
em0: Crc errors = 289
em0: Alignment errors = 7
em0: Collision/Carrier extension errors = 0
em0: RX overruns = 0
em0: watchdog timeouts = 0
em0: RX MSIX IRQ = 0 TX MSIX IRQ = 0 LINK MSIX IRQ = 0
em0: XON Rcvd = 0
em0: XON Xmtd = 0
em0: XOFF Rcvd = 0
em0: XOFF Xmtd = 0
em0: Good Packets Rcvd = 10562
em0: Good Packets Xmtd = 10386
Over to maintainer(s).
Over to maintainer.
em(4) has been significantly updated in 10.2r and current. If this is still a problem in your case (might be using lem(4) instead of em(4)), reopen this issue for further research.