Re: [PATCH 2/2] MTD: Add nand_ecc test module

[Artem Bityutskiy <Artem.Bityutskiy@nokia.com>]

On Wed, 2009-10-21 at 06:46 +0200, ext Akinobu Mita wrote:
> On Tue, Oct 20, 2009 at 02:50:01PM +0300, Artem Bityutskiy wrote:
> > > diff --git a/drivers/mtd/Kconfig b/drivers/mtd/Kconfig
> > > index ecf90f5..abe5c7f 100644
> > > --- a/drivers/mtd/Kconfig
> > > +++ b/drivers/mtd/Kconfig
> > > @@ -33,6 +33,12 @@ config MTD_TESTS
> > >  	  should normally be compiled as kernel modules. The modules perform
> > >  	  various checks and verifications when loaded.
> > >  
> > > +config MTD_NAND_TESTS
> > > +	tristate "MTD NAND tests support"
> > > +	depends on MTD_TESTS && MTD_NAND
> > > +	help
> > > +	  This option enables MTD tests which require NAND Device support.
> > > +
> > >  config MTD_CONCAT
> > >  	tristate "MTD concatenating support"
> > >  	help
> > 
> > Could this please be a separate patch? Also, some of the existing tests
> > are NAND only as well, so could do corresponding Makefile changes in the
> > same patch and move the to the MTD_NAND_TESTS set? The tests are:
> > 
> > mtd_oobtest.c
> > mtd_pagetest.c
> > mtd_subpagetest.c
> 
> Should these MTD NAND tests support OneNAND device, too?
> 
> If so, the config with !CONFIG_MTD_NAND && MTD_ONENAND=m cannot select
> this new MTD_NAND_TESTS. So Kconfig dependency should be:
> 
> config MTD_NAND_TESTS
>         tristate "MTD NAND tests support"
>         depends on MTD_TEST
>         depends on MTD_NAND || MTD_ONENAND
> 
> But nand_ecc-test obviously needs CONFIG_MTD_NAND and cannot exist
> in the same group.

Oh, I see. May be your original idea with ifdefs in the code was not
bad at all :-) ?

-- 
Best Regards,
Artem Bityutskiy (Артём Битюцкий)