2.2.1a on Solaris; runaway fcntl

[Tristan Ball]

I read this as saying you think you have the problem fixed?

If so, how well tested is it? I'd be happy to run a cvs snapshot for a
while...

T.

----- Original Message -----
From: "Gerald Carter" <gcarter at valinux.com>
To: "Michael Gerdts" <Michael.Gerdts at usa.alcatel.com>
Cc: "David Lee" <t.d.lee at durham.ac.uk>; <samba-technical at lists.samba.org>
Sent: Sunday, September 02, 2001 2:09 PM
Subject: Re: 2.2.1a on Solaris; runaway fcntl


> On Fri, 31 Aug 2001, Michael Gerdts wrote:
>
> > On Fri, Aug 31, 2001 at 02:43:55PM +0100, David Lee wrote:
> > > I realise there there has been some discussion of the problem of smbd
> > > going into a CPU-gobbling loop with "fcntl(...)", particularly with
> > > version 2.2.1a on Solaris.
> > >
> > > Was there any resolution to this, for sites such as ours which must
run
> > > basically production, not CVS, versions of Samba?
> >
> > As an AOL user once said, "Me too!"
> >
> > I would be quite interested in seeing either a minimal patch or some
> > indication as to when 2.2.1b or 2.2.2 will be released with this fix.
> > It is somewhat important to the perception of Samba in my environment.
>
> Mike,
>
> We are working towards a 2.2.2 release maybe by the first of
> October.  Currently our infrastructure doesn't really handle these
> intermediate patches very well.  We are working on correcting
> this for later releases.
>
> The main change was in smbd/oplock.c:open_mode_check() IIRC and few
> additional functions added to support this.
>
>
>