httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jim Jagielski <>
Subject Re: systems without accept_mutex, bug?
Date Sat, 19 Apr 1997 20:03:20 GMT
Dean Gaudet wrote:
> On Sat, 19 Apr 1997, Ben Laurie wrote:
> > Surely the blocked children will be scoreboarded as busy, so new children will
> > be brought up. But it does mean that there will be more children than there
> > should be.
> Nope, the BUSY setting doesn't happen until after a socket has been
> accept()ed.  So they're all recorded as being READY and not enough servers
> are started.
> It looks like this might be a problem we should solve.  The serializing
> options are a workaround.  The only solution I can think of without
> serialize is to make the listening sockets non-blocking... I wonder what
> is actually faster, maybe serializing is a good thing.

I think the speed depends on the locking method, for some OSs
FLOCK might be much faster than FCNTL

      Jim Jagielski            |       jaguNET Access Services           |
                  "Not the Craw... the CRAW!"

View raw message