[ntp:questions] Re: ACTS - too many recvbufs allocated (40) (Correct the Version of ntp-dev)

Ronan Flood ronan at noc.ulcc.ac.uk
Fri Feb 11 15:32:49 UTC 2005


cipo <cseplo_l at netlock.net> wrote:

> >Oh, I missed that. I have nowhere to test this. Do you know what the
> >error is? I'll try and take another look at it.
> >
> >Dannny
> >  
> >
> For my slackware linux system the error is: errno=4, Interrupted system 
> call.

Does read() set errno when it returns 0?  The Solaris manpage says

     In addition, read() and readv() will fail if the STREAM head
     had  processed  an  asynchronous  error before the call.  In
     this case, the value of errno does not reflect the result of
     read()  or readv() but reflects the prior error. If a hangup
     occurs on the STREAM being read, read() continues to operate
     normally   until  the  STREAM  head  read  queue  is  empty.
     Thereafter, it returns 0.

BTW I'm concentrating on the hangup behaviour as that's what cipo
is seeing here and I had seen before.  There might be other, less
drastic reasons for read() to return 0 from a refclock, so checking
for that and logging it would be reasonable anyway.

-- 
                      Ronan Flood <R.Flood at noc.ulcc.ac.uk>
                        working for but not speaking for
             Network Services, University of London Computer Centre
     (which means: don't bother ULCC if I've said something you don't like)



More information about the questions mailing list