I'm running 1.4.2 on my AlphaStation. It runs just fine just about all
the time. Once in a rare while I have a problem where the ppp interface
locks up and I get this error (demonstrated with ping):

ping: sendto: no buffer space available

A kill -1 on pppd seems to clear the problem up (normally). No other
interface seems affected (de0, my 10base-T network).

Tonight I can't seem to get the problem to stop happening. I've lost
my ppp interface three times now (so far). The ppp line stays up, ppp0
stays configured, but no traffic comes in or out. 

Could this be a DoS? 

What sort of statistics should I gather when the problem happens? I've
never been able to reproduce this problem before now, but tonight I
have just about everything netstat and vmstat can give. I'd post them
but they tend to be rather large. (Oh, and pppstats and nfsstat also
gave numbers, but I'm not quite sure what they mean.)

Suffice to say that all ping packets get dropped, the error count goes
up, and slowly the amount of kernel memory in use rises. 

Some excerpts taken while ping was running reporting errors:

Name  Mtu   Network       Address              Ipkts Ierrs    Opkts Oerrs Colls Drops
ppp0  1524  <Link>                            197089     1   217692  6180     0  6180
ppp0  1524  209.42.217.8/ client209-42-217-   197089     1   217692  6180     0  6180

2c2
< 	400764 total packets received
---
> 	400789 total packets received
15c15
< 	390129 packets for this host
---
> 	390154 packets for this host
20,21c20,21
< 	393207 packets sent from this host
< 	259 packets sent with fabricated ip header
---
> 	393239 packets sent from this host
> 	278 packets sent with fabricated ip header
39c39
< 		echo reply: 222
---
> 		echo reply: 241
110c110
< 	102076 datagrams received
---
> 	102082 datagrams received
117,119c117,119
< 	101509 delivered
< 	92867 PCB hash misses
< 	101931 datagrams output
---
> 	101515 delivered
> 	92873 PCB hash misses
> 	101944 datagrams output


Thanks for any help in fixing this. 
-- 
Kevin P. Neal                                http://www.pobox.com/~kpn/

Seen on bottom of IBM part number 1887724:
DO NOT EXPOSE MOUSE PAD TO DIRECT SUNLIGHT FOR EXTENDED PERIODS OF TIME.