Re: Call for bge(4) testers

On Mon, Sep 17, 2012 at 5:27 AM, John Baldwin <jhb_at_freebsd.org> wrote:
> On Monday, September 17, 2012 6:46:31 am Kurt Jaeger wrote:
>> Hi!
>>
>> > > box. I'm especially interested in whether there is any ASF/IPMI
>> > > regression on BCM570x/571x.
>> >
>> > Any ipmi-specific tests I should make ?
>>
>> This comes with kldload ipmi:
>>
>> ipmi0: <IPMI System Interface> port 0xca2-0xca3 on acpi0
>> ipmi0: KCS mode found at io 0xca2 on acpi
>> ipmi0: KCS error: ff
>> ipmi0: IPMI device rev. 1, firmware rev. 1.10, version 2.0
>> ipmi0: Number of channels 1
>> ipmi0: Attached watchdog
>> ipmi1: <IPMI System Interface> on isa0
>> device_attach: ipmi1 attach returned 16
>> ppc0: cannot reserve I/O port range
>
> That shows you have a working BMC (just ignore the ipmi1 warning).  I think to
> test you will want to use ipmitool from a remote machine to access the BMC
> over the network.  (Some BMC's have web UI's as well that support remote KVM,
> etc.  That would be a better test than just using ipmitool.)

    Also, just to see what the behavior is like, try rebooting the box
while connected via remote KVM/SOL and if you have the gear to do so,
try being connected via serial in parallel. Broadcoms have had issues
in the past (personal experience) with ipmi on older versions of
FreeBSD (well, ok.. bce, not bge) because of the song and dance done
when initializing the BMC/NIC at boot where if you're connected via
ipmitool/the remote KVM software while booting, it might kick you out
(which is ok depending on the phase of the moon), but not let you back
in (this is a problem if this didn't occur before). A similar issue
with the NIC song and dance exists with Supermicros and the shared
em(4)/igb(4) port and on pre-8.x there were bugs that prevent you from
accessing the BMC after boot (but that was another driver interaction
probably on the system).
Thanks!
-Garrett

PS I'll give the patch a shot on my Lenovo workstation running
9-STABLE when I get a chance.