svn commit: r274583 - head/usr.sbin/watchdogd

[Ian Lepore]

On Mon, 2014-11-17 at 13:09 -0500, Benjamin Kaduk wrote:
> On Mon, 17 Nov 2014, Xin Li wrote:
> 
> > On 11/16/14 13:46, Benjamin Kaduk wrote:
> > > On Sun, Nov 16, 2014 at 4:44 AM, Xin LI <delphij at freebsd.org
> > > <mailto:delphij at freebsd.org>> wrote:
> > >
> > > Author: delphij Date: Sun Nov 16 09:44:30 2014 New Revision:
> > > 274583 URL: https://svnweb.freebsd.org/changeset/base/274583
> > >
> > > Log: Default to use 10 seconds as nap interval instead of 1.
> > >
> > > relnotes: yes?
> >
> > I'm not sure -- no objection but is this noteworthy?
> 
> Well, I'm not sure, either.  It seems borderline to me, but IIRC Glen had
> asked us to err on the side of using relnotes: yes too often instead of
> too little.  It seems like this could fall under "a change of default
> configuration that would result in a noticeable behavior change" for some
> systems, but probably not a particularly large number.
> 
> I don't have strong feelings either way.
> 
> -Ben
> 

I think the only way you would see a behavior change is if we had a
system whose max timeout was under 10 seconds.  The shortest I know of
are some arm systems whose max is 16 seconds.

-- Ian