From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from vm1.sequanux.org ([188.165.36.56]) by merlin.infradead.org with esmtp (Exim 4.80.1 #2 (Red Hat Linux)) id 1WUvlO-0003k0-NK for linux-mtd@lists.infradead.org; Tue, 01 Apr 2014 10:17:52 +0000 Date: Tue, 1 Apr 2014 12:04:17 +0200 From: Simon Guinot To: Ezequiel Garcia Subject: Re: [PATCH v2 0/4] Allow to specify an ECC scheme through DT Message-ID: <20140401100417.GI27979@kw.sim.vm.gnt> References: <1395401690-25221-1-git-send-email-ezequiel.garcia@free-electrons.com> MIME-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="YPJ8CVbwFUtL7OFW" Content-Disposition: inline In-Reply-To: <1395401690-25221-1-git-send-email-ezequiel.garcia@free-electrons.com> Cc: Lior Amsalem , devicetree@vger.kernel.org, Tawfik Bayouk , Seif Mazareeb , Willy Tarreau , Simon Guinot , linux-mtd@lists.infradead.org, Gregory Clement , Brian Norris , Thomas Petazzoni List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , --YPJ8CVbwFUtL7OFW Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Fri, Mar 21, 2014 at 08:34:46AM -0300, Ezequiel Garcia wrote: > Now that we have agreed on the ECC devicetree binding, here's a patchset > that makes pxa3xx-nand driver use it. >=20 > This work is required on platforms where the flash device reports a requi= red > ECC strength, but that are capable of handling a stronger one. Simon Guin= ot > reported a situation where the vendor provided a bootloader configured wi= th > a stronger than required ECC scheme. Without this patches, such a situati= on > would require a hack to make the driver use a specific ECC scheme. >=20 > (Simon: it would be great if you can confirm this allows you to remove an= y hacks). Hi Ezequiel, I confirm this allows me to configure the ECC strength and step size =66rom DT. Tested-by: Simon Guinot Thanks. Simon >=20 > The first two patches clean the pxa_ecc_init(), preparing for the real > work to be done. >=20 > The third patch makes use of the ECC strength and ECC step size parameter= s. > To ensure full backward compatibility, the patch *tries* to use DT-specif= ied > ECC parameters, falling back to using the ONFI-specified settings if the > former are missing. >=20 > Also, we use the ONFI 'datasheet' ECC strength requirement, to ensure the > DT-specified ECC strength is strong enough to meet the requirement. >=20 > Finally, the last patch updates the documentation binding, with the suppo= rted > ECC parameters. >=20 > Tested on Marvell's Armada 370 RD and also Compulab's CM-X300 (PXA3xx SoC= ), > which is a non-DT platform. Based on today's l2-mtd.git. >=20 > Changes from v1: >=20 > * Fixed typo in ECC strength normalization calculus. >=20 > Ezequiel Garcia (4): > mtd: nand: pxa3xx: Normalize ECC strength for ECC scheme selection > mtd: nand: pxa3xx: Clean pxa_ecc_init() error handling > mtd: nand: pxa3xx: Use ECC strength and step size devicetree binding > mtd: nand: pxa3xx: Add supported ECC strength and step size to the DT > binding >=20 > .../devicetree/bindings/mtd/pxa3xx-nand.txt | 8 +++ > drivers/mtd/nand/pxa3xx_nand.c | 78 +++++++++++++---= ------ > include/linux/platform_data/mtd-nand-pxa3xx.h | 3 + > 3 files changed, 59 insertions(+), 30 deletions(-) >=20 > --=20 > 1.9.0 >=20 >=20 > ______________________________________________________ > Linux MTD discussion mailing list > http://lists.infradead.org/mailman/listinfo/linux-mtd/ --YPJ8CVbwFUtL7OFW Content-Type: application/pgp-signature; name="signature.asc" Content-Description: Digital signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iEYEARECAAYFAlM6jyEACgkQgtp0PDeOcDrJYwCbB37olW/nb3Be40qffuROAMUa vMwAn3rZ9C8K4VHc2o76KoLzzENN3KIZ =RmDT -----END PGP SIGNATURE----- --YPJ8CVbwFUtL7OFW--