[PATCH v4 0/4] mtd:nand:omap2: clean-up of supported ECC schemes

Gupta, Pekon pekon at ti.com
Fri Jul 5 00:12:15 EDT 2013


> 
> On 7/3/2013 11:34 PM, Gupta, Pekon wrote:
> >> On Wednesday 03 July 2013, Artem Bityutskiy wrote:
> >>> On Wed, 2013-07-03 at 13:16 +0000, Gupta, Pekon wrote:
> >>>> [Pekon]: Yes, I'm not seeing these build issues, as I'm cleanly
> >>>> returning from probe with pr_err(), if the required libraries (/lib/bch.c)
> >>>> are not build-in the system.
> >>>> -------------------
> >>>> [Patch v4 1/4]: mtd:nand:omap2: clean-up BCHx_HW and BCHx_SW
> ECC..
> >>>> @@static int omap_nand_probe(struct platform_device *pdev)
> >>>> +	default:
> >>>> +		pr_err("selected ECC scheme not supported or not
> >> enabled\n");
> >>>> +		err = -EINVAL;
> >>>> +		goto out_release_mem_region;
> >>>> +	}
> >>>> -------------------
> >>>> However, if you are still seeing this, could you please send me your
> >> config?
> >>> I compile tested your patches too, and did not see any issues with my
> >>> omap2_defconfig.
> >>
> >> To clarify: I'm getting this error with randconfig on today's linux-next.
> >> I think it only happens with MTD_NAND_ECC_BCH enabled, which is not
> >> the default in omap2_defconfig.
> >>
> >> 	Arnd
> > Hi Arnd, Artem, Stephen,
> >
> > I see the issue in linux-next/master tree..
> > linux-next/master has only the first [Patch v4 1/4] from the series.
> > Other patches [Patch v4 2/4.. 4/4] of this series were dropped,
> > as 'Olof Johansson' had some apprehensions about DT bindings.
> > http://permalink.gmane.org/gmane.linux.ports.arm.kernel/249662
> >
> > As these patches are inter-dependent, so please drop the whole series
> > including the commit below.
> > fb1585b  [Patch v4 1/4] mtd: nand: omap2: clean-up BCHx_HW and
> BCHx_SW
> >
> > I'm awaiting feedbacks from 'Olof Johansson' or other DT maintainers,
> > once approved, I'll resend the whole series.
> >
> > "[PATCH v4 2/4] ARM: OMAP2+: cleaned-up DT support " had the
> > declarations which were required for above build errors.
> > ---------------
> > include/linux/platform_data/mtd-nand-omap2.h
> > enum omap_ecc {
> > +	/* 4-bit  ECC calculation by GPMC, Error detection by Software */
> > +	OMAP_ECC_BCH4_CODE_HW_DETECTION_SW,
> > ...
> > +	/* 8-bit  ECC calculation by GPMC, Error detection by Software */
> > +	OMAP_ECC_BCH8_CODE_HW_DETECTION_SW
> > ---------------
> >
> But patches should not have inter dependency for compilation. It will affect
> the bisect process.
> 
[Pekon]: They don't, If you read through the email, this issue was hit
with randconfig, and not with omap2plus_defconfig. So you would not see
build issue if you are bisecting using correct config (omap2plus_defconfig).
Also if I merge all the relevant changes in single patch, then if would
become difficult for review, and it would be conflict your earlier feedback
of breaking the patch sets.

However, I leave it to Arnd and Artem if they want me to re-shuffle 
changes and re-submit.


with regards, pekon


More information about the linux-mtd mailing list