Re: [PATCH v3 3/8] mm/debug-pagealloc: make debug-pagealloc boottime configurable

[Joonsoo Kim]

On Wed, Nov 26, 2014 at 12:49:36PM -0800, Andrew Morton wrote:
> On Tue, 25 Nov 2014 08:42:37 +0900 Joonsoo Kim <iamjoonsoo.kim@xxxxxxx> wrote:
> 
> > On Mon, Nov 24, 2014 at 02:55:42PM -0800, Andrew Morton wrote:
> > > On Mon, 24 Nov 2014 17:15:21 +0900 Joonsoo Kim <iamjoonsoo.kim@xxxxxxx> wrote:
> > > 
> > > > Now, we have prepared to avoid using debug-pagealloc in boottime. So
> > > > introduce new kernel-parameter to disable debug-pagealloc in boottime,
> > > > and makes related functions to be disabled in this case.
> > > > 
> > > > Only non-intuitive part is change of guard page functions. Because
> > > > guard page is effective only if debug-pagealloc is enabled, turning off
> > > > according to debug-pagealloc is reasonable thing to do.
> > > > 
> > > > ...
> > > >
> > > > --- a/Documentation/kernel-parameters.txt
> > > > +++ b/Documentation/kernel-parameters.txt
> > > > @@ -858,6 +858,14 @@ bytes respectively. Such letter suffixes can also be entirely omitted.
> > > >  			causing system reset or hang due to sending
> > > >  			INIT from AP to BSP.
> > > >  
> > > > +	disable_debug_pagealloc
> > > > +			[KNL] When CONFIG_DEBUG_PAGEALLOC is set, this
> > > > +			parameter allows user to disable it at boot time.
> > > > +			With this parameter, we can avoid allocating huge
> > > > +			chunk of memory for debug pagealloc and then
> > > > +			the system will work mostly same with the kernel
> > > > +			built without CONFIG_DEBUG_PAGEALLOC.
> > > > +
> > > 
> > > Weren't we going to make this default to "off", require a boot option
> > > to turn debug_pagealloc on?
> > 
> > Hello, Andrew.
> > 
> > I'm afraid that changing default to "off" confuses some old users.
> > They would expect that it is default "on". But, it is just debug
> > feature, so, it may be no problem. If you prefer to change default, I
> > will rework this patch. Please let me know your decision.
> 
> I suspect the number of "old users" is one ;)
> 
> I think it would be better to default to off - that's the typical
> behaviour for debug features, for good reasons.

Okay.
Here goes the patch.

------------->8----------------