trap 12: page fault while in kernel mode on 8.0-RELEASE (possibly bge(4) related)

[Pyun YongHyeon]

On Fri, Feb 19, 2010 at 10:11:03PM +0300, Slawa Olhovchenkov wrote:
> On Fri, Feb 19, 2010 at 11:03:59AM -0800, Pyun YongHyeon wrote:
> 
> > On Fri, Feb 19, 2010 at 03:24:15PM +0300, Slawa Olhovchenkov wrote:
> > > On Fri, Feb 19, 2010 at 08:51:29AM +0300, Slawa Olhovchenkov wrote:
> > > 
> > > > On Thu, Feb 18, 2010 at 04:19:13PM -0800, Pyun YongHyeon wrote:
> > > > 
> > > > > 
> > > > > I'm still not sure whether the panic is related with bge(4) but
> > > > > there are a couple of missing workaround for PCIX BCM5704 silicon
> > > > > bug in bge(4). Did you also see the panic before updating to
> > > > > stable/8?
> > > > 
> > > > Before updating to stable/8 2010-Feb-16 I see network freez on stable/8
> > > > 2009-Sep -- bge stop receiving packets (by tcpdump), after aprox. 40-50
> > > > days uptime.
> > > > 
> > > > 
> > > > > Anyway, try attached patch and let me know how it works.
> > > > 
> > > > Thanks, I try.
> > > > 
> > > 
> > > I don't get trap after 2 hour, but already see next trouble:
> > > 
> > > ===
> > > PING 10.200.0.1 (10.200.0.1): 56 data bytes
> > > 
> > > --- 10.200.0.1 ping statistics ---
> > > 100 packets transmitted, 97 packets received, 3.0% packet loss
> > > round-trip min/avg/max/stddev = 0.188/0.268/0.356/0.044 ms
> > > ===
> > > 
> > > w/o patch, but witch fresh source I see same trouble: after 12 hour 7% lost.
> > > netstat -i don't show any errors.
> > 
> > I think BCM5704 supports HW MAC statistics counter. Try extract it
> > with "sysctl dev.bge.0.stats". It will give you much more
> > information.
> 

[...]

> dev.bge.1.stats.rx.Fragments: 1

You received a frame that is less than 64 bytes with a bad FCS.

> dev.bge.1.stats.rx.UcastPkts: 2956515
> dev.bge.1.stats.rx.MulticastPkts: 0
> dev.bge.1.stats.rx.FCSErrors: 18

You have a lot of FCS errors here.
Please double check cabling. If the statistics counter is right,
sender is guilty or you have bad cabling issues here.