Re: "le0: dropping chained buffer", OpenBSD 3.1 on Sparc 5 (sun4m)

On 16 Aug 2003, Rich Kulawiec wrote:
[...]
| Yesterday afternoon around 13:00, both of them simultaneously became
[...]
| So I went to the source. I have traced this to the following bit of
[...]
| Trying to find the cause brings me back to that bit of code above and
[...]

The code is right: if (bits != (STP | ENP)) then it is a chained buffer. Anyway, the troubleshooting you have done raises some important points:

  (1) both of your hosts became unresponsive at the same time   (2) the am7990 driver does not handle chained buffers

Maybe your 2 hosts are receiving some weird ethernet frames and your le0 NICs store these frames in chained buffers, but unfortunately the am7990 driver does not implement reception of chained buffers and eventually shut down the interface. These guys have some possible explanations for the weird network packets:

        http://www.geocrawler.com/archives/3/260/1998/9/0/1390514/

You said your have ran tcpdump, but have you succeeded to capture some traffic at the same time that the "le0: dropping chained buffer" msg occured ?

-- 
Marc Bevand                          
http://www.epita.fr/~bevand_m
Computer Science School EPITA - System, Network and Security Dept.