msk link problems on 8.0

[Chris Cowart]

Hello,

I recently upgraded my home router box to 8.0. Since doing so, I've had
some problems with the msk driver.

On startup, I'm experiencing the behavior that seems to be documented in
kern/139093. Specifically, during network initialization, rc displays
the message that it's "Waiting 30s for the default route interface". If
I physically flap the link at this point, the system continues booting
(when missed, the rest of the system doesn't do so well starting up, but
I can flap the link at any future point to get it working).

Before and after flapping the link, ifconfig reports the same thing:

| % ifconfig msk0
| msk0: flags=8943<UP,BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICAST> metric 0 mtu 2290
|         options=118<VLAN_MTU,VLAN_HWTAGGING,TSO4>
|         ether 00:16:cb:ae:5b:1f
|         inet6 fe80::216:cbff:feae:5b1f%gige0 prefixlen 64 scopeid 0x1 
|         media: Ethernet autoselect (1000baseT <full-duplex,flag0,flag1>)
|         status: active

This interface is configured as a trunk for several VLANs.

| % dmesg | grep msk
| mskc0: <Marvell Yukon 88E8053 Gigabit Ethernet> port 0x1000-0x10ff mem
| 0x50200000-0x50203fff irq 16 at device 0.0 on pci1
| msk0: <Marvell Technology Group Ltd. Yukon EC Id 0xb6 Rev 0x02> on mskc0
| msk0: Ethernet address: 00:16:cb:ae:5b:1f
| miibus0: <MII bus> on msk0
| mskc0: [FILTER]
| % dmesg | grep e1000
| e1000phy0: <Marvell 88E1111 Gigabit PHY> PHY 0 on miibus0
| e1000phy0:  10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-FDX, auto

| % pciconf -lv
| mskc0 at pci0:1:0:0:       class=0x020000 card=0x532111ab chip=0x436211ab rev=0x22 hdr=0x00
|     vendor     = 'Marvell Semiconductor (Was: Galileo Technology Ltd)'
|     device     = 'Marvell Yukon 88E8053 PCI-E Gigabit Ethernet Controller (88E8053)'
|     class      = network
|     subclass   = ethernet

| % uname -a
| FreeBSD gw.ccowart.com 8.0-RELEASE-p1 FreeBSD 8.0-RELEASE-p1 #0: Sun
| Dec  6 14:43:41 PST 2009 root at gw.ccowart.com:/usr/obj/usr/src/sys/ROUTER
| amd64

I have the following in my /boot/loader.conf as a workaround that I
believe dates back to 7.0:

| hw.msk.msi_disable="1"

I also have "-txcsum" in the interface configuration to address an
earlier bug (both of these may have been resolved at this point; I just
haven't been daring enough to test it).

Having read the PR, I copied sys/dev/{msk,e1000} from HEAD into the
8.0-p1 source tree and installed the resulting kernel. The behavior did
not change. If anyone has any alternative patching I could do, either
against 8.0 or HEAD for these drivers, I'd be more than willing to try
them out.

On a related note, last night, when the system did boot, I would also
run into a problem where the following message would be logged: "msk0:
Rx FIFO overrun!". Once logged, the NIC seemed to be completely wedged
and unusable. Doing ifconfig down/up did not help things. At the time, I
hadn't discovered the physical down/up workaround, so I can't speak to
whether that would have helped (and this error condition hasn't
recurred (knock on wood)). I don't know if the issues are related or
separate, but if you have any ideas, I'd like to be prepared if it comes
up again.

Thanks for the help,

-- 
Chris Cowart
  http://www.timesinks.net/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 834 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-net/attachments/20091207/63e8b0aa/attachment.pgp