Re: [PATCH] memsw: handle swapaccount kernel parameter correctly

[KAMEZAWA Hiroyuki]

On Thu, 27 Jan 2011 10:29:51 +0100
Michal Hocko <mhocko@xxxxxxx> wrote:

> On Thu 27-01-11 18:03:30, KAMEZAWA Hiroyuki wrote:
> > On Thu, 27 Jan 2011 09:23:20 +0100
> > Michal Hocko <mhocko@xxxxxxx> wrote:
> [...]
> > > diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> > > index db76ef7..cea2be48 100644
> > > --- a/mm/memcontrol.c
> > > +++ b/mm/memcontrol.c
> > > @@ -5013,9 +5013,9 @@ struct cgroup_subsys mem_cgroup_subsys = {
> > >  static int __init enable_swap_account(char *s)
> > >  {
> > >  	/* consider enabled if no parameter or 1 is given */
> > > -	if (!s || !strcmp(s, "1"))
> > > +	if (!(*s) || !strcmp(s, "=1"))
> > >  		really_do_swap_account = 1;
> > > -	else if (!strcmp(s, "0"))
> > > +	else if (!strcmp(s, "=0"))
> > >  		really_do_swap_account = 0;
> > >  	return 1;
> > >  }
> > 
> > Hmm, usual callser of __setup() includes '=' to parameter name, as
> > 
> > mm/hugetlb.c:__setup("hugepages=", hugetlb_nrpages_setup);
> > mm/hugetlb.c:__setup("default_hugepagesz=", hugetlb_default_setup);
> > 
> > How about moving "=" to __setup() ?
> 
> I have considered that as well but then we couldn't use swapaccount
> parameter without any value because the parameter parsing matches the
> whole string. 
> I found it better to have consistent [no]swapaccount with the =0|1
> extension rather than keeping = in the setup like other users.
> 
> Sounds reasonable?

Hmm. ok for this time.

Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>


Could you try to write a patch for feature-removal-schedule.txt
and tries to remove noswapaccount and do clean up all ?
(And add warning to noswapaccount will be removed.....in 2.6.40)


Thanks,
-Kame




--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/