Re: [RFA][PATCH 2/8] netfilter: Remove return values for print_conntrack callbacks

[Steven Rostedt <rostedt@goodmis.org>]

On Tue, 04 Nov 2014 06:46:19 -0800
Joe Perches <joe@perches.com> wrote:

> On Tue, 2014-11-04 at 09:31 -0500, Steven Rostedt wrote:
> > On Tue, 4 Nov 2014 15:22:36 +0100
> > Pablo Neira Ayuso <pablo@netfilter.org> wrote:
> > > On Tue, Nov 04, 2014 at 08:05:35AM -0500, Steven Rostedt wrote:
> > > > On Wed, 29 Oct 2014 17:56:04 -0400 Steven Rostedt <rostedt@goodmis.org> wrote:
> > > > > From: "Steven Rostedt (Red Hat)" <rostedt@goodmis.org>
> > > > > [ REQUEST FOR ACKS ]
> > > > Can any of the netfilter folks give me an Acked-by for this?
> > > If Florian's concern were addressed, then:
> > Yeah, the change he mentioned was done is 3/8. As that was written by
> > Joe Perches, I did some work that he missed and put it before his
> > patch, which showed a discrepancy between the two functions. After all
> > patches are applied, it should be consistent to his liking.
> 
> I think seq_has_overflowed does not need
> to be used after every seq_<put/print> call.
> 
> It interrupts reading code flow and just
> isn't alll that necessary as every operation
> before it will be redone anyway.
> 
> It should be used before or after a large
> blocks though.
> 

It's not used in every occurrence. The problem that Florian had was that
there were two almost identical functions, and you changed one to have
the seq_has_overflowed() check, but the other one was left without it.

It wasn't about checking multiple times, it was about consistency
between two similar functions.

-- Steve