Fake /dev/ed0 on VM corrupted

[Sean Bruno]

Not sure what's going on here, but I've installed an updated 7-stable on
a VM in my Fedora kvm enabled laptop and I see that /dev/ed0 is having
issues when doing network operations.  I could have sworn that we
resolved this, but can't remember what what the fix was (probably I
switched to a fake /dev/em0).  Thoughts?

<err log snip>
ed0: NIC memory corrupt - invalid packet length 13905
ed0: NIC memory corrupt - invalid packet length 8263
ed0: NIC memory corrupt - invalid packet length 58250
ed0: NIC memory corrupt - invalid packet length 7531
ed0: NIC memory corrupt - invalid packet length 51706
ed0: NIC memory corrupt - invalid packet length 16108
ed0: NIC memory corrupt - invalid packet length 28242
ed0: NIC memory corrupt - invalid packet length 13162

<dmesg snip>
ed0: <RealTek 8029> port 0xc100-0xc1ff irq 11 at device 3.0 on pci0
ed0: [ITHREAD]
ed0: WARNING: using obsoleted if_watchdog interface
ed0: Ethernet address: 52:54:00:1c:b8:82

<pciconf -lv snip>
ed0 at pci0:0:3:0:	class=0x020000 card=0x11001af4 chip=0x802910ec rev=0x00
hdr=0x00
    vendor     = 'Realtek Semiconductor'
    device     = 'PCI Full-Duplex Ethernet Controller with PnP Function
(RTL8029)'
    class      = network
    subclass   = ethernet

<sysctl -a |grep dev.ed.0 snop>
dev.ed.0.type: RTL8029
dev.ed.0.TxMem: 3072
dev.ed.0.RxMem: 13312
dev.ed.0.Mem: 16384
dev.ed.0.%desc: RealTek 8029
dev.ed.0.%driver: ed
dev.ed.0.%location: slot=3 function=0 handle=\_SB_.PCI0.S3__
dev.ed.0.%pnpinfo: vendor=0x10ec device=0x8029 subvendor=0x1af4
subdevice=0x1100 class=0x020000
dev.ed.0.%parent: pci0