Kernel memory corruption(?) with age(4)

[YongHyeon PYUN]

On Wed, Mar 30, 2011 at 09:50:12PM +0200, Yamagi Burmeister wrote:
> On Wed, 30 Mar 2011, YongHyeon PYUN wrote:
> 
> >On Wed, Mar 30, 2011 at 04:22:23PM +0200, Yamagi Burmeister wrote:
> >
> >>All for boxes are unstable if the Attansic NIC is in use, no one of them
> >>survived more than 60 minutes of ~20mb/s network traffic. I managed to
> >>get some coredumps and extracted the backtraces. Since everytime one of
> >>the boxes paniced I got different panic message and a different backtrace
> >>with a different subsystem involved I suspected broken hardware. I
> >>plugged a em(4) NIC into the PCI slot and wasn't able to reproduce the
> >>problem, in fact the boxes run rock solid for several days. Next I set
> >>up a Windows 7, installed the Attansic vendor driver and did another
> >>run. All went smooth, no crash for nearly 24 hours.
> >>
> >>My guess is kernel memory corruption by age(4), which would explain all
> >>the different backtraces and the different panic messages. This problem
> >>is reproducible in at least FreeBSD 7.4 and 8.2 and with TSO4 enabled
> >>and disabled. I'm willing to debug this, but I really don't know how. So
> >>any help or a pointer into the right direction would be appreciated.
> >>
> >
> >AFAIK this is the first report for possible memory corruption
> >triggered by age(4). I'm still not sure whether it's caused by
> >age(4) but you can disable RX checksum offloading and see whether
> >that makes any difference.
> >Since I have no longer access to the hardware it would be even
> >better if you can tell me which traffic pattern triggered the
> >issue.
> 
> Okay, I did a test run with RX checksum, TX checksum and both disabled. 
> In all three cases the crash occurs within about 20 minutes. I'm either
> not sure that age(4) is the problem but it has definedly something to do
> with the problem, since with another nic driver the same scenario is
> rock solid...
> 

OK.

> The workload: It's a NFS3 server (FreeBSDs non-experimental
> implementation), serving and receiving file with about 250 to 500
> megabytes at about 20mb/s. The clients are FreeBSD 7 and 8 systems and
> are mounting the shares via TCP. The connection is 1000mbit/s via a
> "dumb" gigabit switch.
> 

That's too broad to narrow down the issue. :-(
I'm not sure but your box seem to have more than 4GB memory. Could
you limit the available memory to 3GB via loader.conf and test it
again?