Re: select() regression in libpcap-devel?

[Guy Harris <guy () alum mit edu>]

On Jun 23, 2009, at 7:34 PM, Mike Kershaw wrote:

> (This now actually hits my error catcher where 100 fd highs in a row
> with no packets triggers a shutdown of the source, since libpcap-1.0.0
> seems to not return errors in pcap_dispatch when a netdev is removed

There does not appear to be a way for the memory-mapped interface to  
directly return such an error.

> (ie
> usb unplugged or driver crash) and instead returns a 'hot' FD  
> constantly
> as well... this is a separate issue however :).

"hot" as in readable, or "hot" as in "exceptional condition"?

A quick look at datagram_poll() seems to indicate that it sets POLLERR  
if sk->sk_err is set, so, at least for poll, the fd will be marked as  
"hot" in some sense; I couldn't find any obvious indication of whether  
select() is a system call or a purely userland wrapper around poll(),  
but, in either case, I'd expect select() to say the descriptor is  
"hot" in that it'd be listed at least among the error/exceptional  
condition fds.

It might be that the descriptor will be "stuck" if you don't do a  
receive from the socket - even if it's memory-mapped - to get the  
error code.
-
This is the tcpdump-workers list.
Visit https://cod.sandelman.ca/ to unsubscribe.