if_sk patch to get more info from people with problems

Bjoern A. Zeeb bzeeb-lists at lists.zabbadoz.net
Mon Nov 1 03:40:10 PST 2004


a lot of people seem to have problems with if_sk. A lot of people
consider it simply broken on some boards.

Let's see if we can find out what problems happen for which people.
- watchdog timeouts
- rx/tx "hangs"
- ...

At [1] you can find a quick, dirty and ugly hack[2] (unverified to be
implemented correctly in the last 5 minutes) to dump more VPD (vital
product data) from those cards. It will print out sth like this
(on may amd64 with ASUS K8V SE Deluxe):

! ...
! skc0: <Marvell Gigabit Ethernet> port 0xa800-0xa8ff mem 0xf7b00000-0xf7b03fff irq 17 at device 10.0 on pci0
! skc0: Yukon Gigabit Ethernet 10/100/1000Base-T Adapter
! skc0: PN: Yukon 88E8001
! skc0: EC: Rev. 1.3
! skc0: MN: Marvell
! skc0: SN: AbCdEfG334454
! sk0: <Marvell Semiconductor, Inc. Yukon> on skc0
! ...

PN is Part Number
EC is Engineering Change level
MN is the Manufacture ID
SN is Serial Number

would be interesting to see if we can find more similarities.

[1] http://sources.zabbadoz.net/freebsd/patchset/EXPERIMENTAL/if_sk_vpd.diff
[2] I am still not checking VPD checksum; nor should this handling be
    done in the driver because it's a PCI thing and should be cleanly
    implemented somewhere else.

Bjoern A. Zeeb				bzeeb at Zabbadoz dot NeT

More information about the freebsd-amd64 mailing list