From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from down.free-electrons.com ([37.187.137.238] helo=mail.free-electrons.com) by bombadil.infradead.org with esmtp (Exim 4.80.1 #2 (Red Hat Linux)) id 1bAsRk-0004pC-H5 for linux-mtd@lists.infradead.org; Thu, 09 Jun 2016 05:24:01 +0000 Date: Thu, 9 Jun 2016 07:23:37 +0200 From: Boris Brezillon To: Kamal Dasu Cc: linux-mtd@lists.infradead.org, computersforpeace@gmail.com, f.fainelli@gmail.com, bcm-kernel-feedback-list@broadcom.com Subject: Re: [V2, 1/2] mtd: brcmnand: Add check for erased page bitflips Message-ID: <20160609072337.26f16f6a@bbrezillon> In-Reply-To: <1465419681-6162-1-git-send-email-kdasu.kdev@gmail.com> References: <1465419681-6162-1-git-send-email-kdasu.kdev@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Wed, 8 Jun 2016 17:01:20 -0400 Kamal Dasu wrote: > Check for erased page bitflips in a page. And if well within > threshold return data as all 0xff. Apply sw check for controller > version < 7.2. Controller vesion >= 7.2 has hw support. > > Signed-off-by: Kamal Dasu > --- > V2 changes > Added use of nand_check_erased_ecc_chunk > Restrict change to older controller < 7.2 > --- > drivers/mtd/nand/brcmnand/brcmnand.c | 61 ++++++++++++++++++++++++++++++++++++ > 1 file changed, 61 insertions(+) > > diff --git a/drivers/mtd/nand/brcmnand/brcmnand.c b/drivers/mtd/nand/brcmnand/brcmnand.c > index b76ad7c..212acb4 100644 > --- a/drivers/mtd/nand/brcmnand/brcmnand.c > +++ b/drivers/mtd/nand/brcmnand/brcmnand.c > @@ -1545,6 +1545,55 @@ static int brcmnand_read_by_pio(struct mtd_info *mtd, struct nand_chip *chip, > return ret; > } > > +/* > + * Check a page to see if it is erased (w/ bitflips) after an uncorrectable ECC > + * error > + * > + * Because the HW ECC signals an ECC error if an erase paged has even a single > + * bitflip, we must check each ECC error to see if it is actually an erased > + * page with bitflips, not a truly corrupted page. > + * > + * On a real error, return a negative error code (-EBADMSG for ECC error), and > + * buf will contain raw data. > + * Otherwise, buf gets filled with 0xffs and return the maximum number of > + * bitflips-per-ECC-sector to the caller. > + * > + */ > +static int brcmstb_nand_verify_erased_page(struct mtd_info *mtd, > + struct nand_chip *chip, void *buf, u64 addr) > +{ > + int i, sas; > + void *oob = chip->oob_poi; > + int bitflips = 0; > + int page = addr >> chip->page_shift; > + int ret; > + > + if (!buf) { > + buf = chip->buffers->databuf; > + /* Invalidate page cache */ > + chip->pagebuf = -1; > + } > + > + sas = mtd->oobsize / chip->ecc.steps; > + > + /* read without ecc for verification */ > + chip->cmdfunc(mtd, NAND_CMD_READ0, 0x00, page); > + ret = chip->ecc.read_page_raw(mtd, chip, buf, true, page); > + if (ret) > + return ret; > + > + for (i = 0; i < chip->ecc.steps; i++, oob += sas) { > + bitflips = nand_check_erased_ecc_chunk(buf, chip->ecc.size, > + oob, sas, NULL, 0, > + chip->ecc.strength); > + /* Too many bitflips */ > + if (bitflips < 0) > + break; Hm, should be ret = nand_check_erased_ecc_chunk(buf, chip->ecc.size, oob, sas, NULL, 0, chip->ecc.strength); if (ret < 0) return ret; bitflips = max(bitflips, ret); > + } > + > + return bitflips; > +} > + > static int brcmnand_read(struct mtd_info *mtd, struct nand_chip *chip, > u64 addr, unsigned int trans, u32 *buf, u8 *oob) > { > @@ -1575,6 +1624,18 @@ static int brcmnand_read(struct mtd_info *mtd, struct nand_chip *chip, > } > > if (mtd_is_eccerr(err)) { > + /* > + * Controller version 7.2 has hw encoder to detect erased page > + * bitflips, apply sw verification for older controllers only > + */ > + if (ctrl->nand_version < 0x0702) { > + err = brcmstb_nand_verify_erased_page(mtd, chip, buf, > + addr); > + /* erased page bitflips corrected */ > + if (err > 0) > + return err; > + } > + > dev_dbg(ctrl->dev, "uncorrectable error at 0x%llx\n", > (unsigned long long)err_addr); > mtd->ecc_stats.failed++; -- Boris Brezillon, Free Electrons Embedded Linux and Kernel engineering http://free-electrons.com