Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S967091AbdLSMCg (ORCPT ); Tue, 19 Dec 2017 07:02:36 -0500 Received: from mo1501.tsb.2iij.net ([210.149.48.173]:49885 "EHLO mo.tsb.2iij.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1762502AbdLSMCc (ORCPT ); Tue, 19 Dec 2017 07:02:32 -0500 X-MXL-Hash: 5a38ffae52ca08ca-47c55b83386a22fbc7665fcfa771b16a396d9277 Subject: Re: [PATCH -next v3 2/2] mtd: nand: toshiba: Add support for Toshiba Memory BENAND (Built-in ECC NAND) To: Boris Brezillon Cc: richard@nod.at, dwmw2@infradead.org, computersforpeace@gmail.com, marek.vasut@gmail.com, cyrille.pitchen@wedev4u.fr, linux-mtd@lists.infradead.org, linux-kernel@vger.kernel.org References: <1512569098-30038-1-git-send-email-yoshitake.kobayashi@toshiba.co.jp> <1512569098-30038-3-git-send-email-yoshitake.kobayashi@toshiba.co.jp> <20171206162404.3a1d83b9@bbrezillon> From: KOBAYASHI Yoshitake Message-ID: Date: Tue, 19 Dec 2017 21:01:47 +0900 User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:52.0) Gecko/20100101 Thunderbird/52.5.0 MIME-Version: 1.0 In-Reply-To: <20171206162404.3a1d83b9@bbrezillon> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 7bit X-MAIL-FROM: X-SOURCE-IP: [172.27.153.190] X-Spam: exempt Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 5068 Lines: 166 Thanks for reviewing the patches and taht is a good news for me about nand_exec_op(). I'll change this patch as you suggested. I would like to make sure the following one. >> @@ -70,6 +155,10 @@ static int toshiba_nand_init(struct nand_chip *chip) >> if (nand_is_slc(chip)) >> chip->bbt_options |= NAND_BBT_SCAN2NDPAGE; >> >> + if (nand_is_slc(chip) && (chip->id.data[4] & 0x80) /* BENAND */ && >> + (chip->ecc.mode == NAND_ECC_ON_DIE)) >> + toshiba_nand_benand_init(chip); >> + > > Please move this block to toshiba_nand_init(). I think it's already in toshiba_nand_init(). -- YOSHI On 2017/12/07 0:24, Boris Brezillon wrote: > On Wed, 6 Dec 2017 23:04:58 +0900 > KOBAYASHI Yoshitake wrote: > >> This patch enables support for Toshiba Memory BENAND. This checks >> internal ECC status in read operation when using BENAND and selecting >> ECC mode as on-die. >> >> Signed-off-by: KOBAYASHI Yoshitake >> --- >> drivers/mtd/nand/nand_toshiba.c | 89 +++++++++++++++++++++++++++++++++++++++++ >> 1 file changed, 89 insertions(+) >> >> diff --git a/drivers/mtd/nand/nand_toshiba.c b/drivers/mtd/nand/nand_toshiba.c >> index c2c141b..35c0ddf 100644 >> --- a/drivers/mtd/nand/nand_toshiba.c >> +++ b/drivers/mtd/nand/nand_toshiba.c >> @@ -17,6 +17,91 @@ >> >> #include >> >> +/* ECC Status Read Command for BENAND */ >> +#define TOSHIBA_NAND_CMD_ECC_STATUS 0x7A >> + >> +/* Recommended to rewrite for BENAND */ >> +#define TOSHIBA_NAND_STATUS_REWRITE_RECOMMENDED BIT(3) >> + >> +static int toshiba_nand_benand_status_chk(struct mtd_info *mtd, >> + struct nand_chip *chip) >> +{ >> + unsigned int max_bitflips = 0; >> + u8 status; >> + >> + /* Check Read Status */ >> + chip->cmdfunc(mtd, NAND_CMD_STATUS, -1, -1); >> + status = chip->read_byte(mtd); > > Please use the recently introduced nand_status_op() helper. > >> + >> + /* >> + * TOSHIBA_NAND_CMD_ECC_STATUS is vendor specific command. >> + * Currently we have no way to send arbitrary sequences of >> + * CMD+ADDR+DATA cycles and thus cannot support this custom >> + * TOSHIBA_NAND_CMD_ECC_STATUS operation. > > That's about to change :-), if everything goes well, nand_exec_op() > should be available in 4.16. > >> + * For now, we set max_bitflips mtd->bitflip_threshold. >> + */ >> + if (status & NAND_STATUS_FAIL) { >> + /* uncorrectable */ >> + mtd->ecc_stats.failed++; >> + } else if (status & TOSHIBA_NAND_STATUS_REWRITE_RECOMMENDED) { >> + /* correctable */ >> + max_bitflips = mtd->bitflip_threshold; >> + mtd->ecc_stats.corrected += max_bitflips; >> + } >> + >> + return max_bitflips; >> +} >> + >> +static int >> +toshiba_nand_read_page_benand(struct mtd_info *mtd, >> + struct nand_chip *chip, uint8_t *buf, >> + int oob_required, int page) >> +{ >> + nand_read_page_raw(mtd, chip, buf, oob_required, page); > > Please check the return code of nand_read_page_raw(). > >> + >> + return toshiba_nand_benand_status_chk(mtd, chip); >> +} >> + >> +static int >> +toshiba_nand_read_subpage_benand(struct mtd_info *mtd, >> + struct nand_chip *chip, uint32_t data_offs, >> + uint32_t readlen, uint8_t *bufpoi, int page) >> +{ >> + uint8_t *p; >> + >> + if (data_offs != 0) >> + chip->cmdfunc(mtd, NAND_CMD_RNDOUT, data_offs, -1); >> + >> + p = bufpoi + data_offs; >> + chip->read_buf(mtd, p, readlen); > > The core is no longer sending the initial READ0 command, so this should > be turned into: > > ret = nand_read_page_op(chip, data_offs, bufpoi + data_offs, > readlen); > if (ret) > return ret; > > >> + >> + return toshiba_nand_benand_status_chk(mtd, chip); >> +} >> + >> +static void toshiba_nand_benand_init(struct nand_chip *chip) >> +{ >> + struct mtd_info *mtd = nand_to_mtd(chip); >> + >> + /* >> + * On BENAND, the entire OOB region can be used by the MTD user. >> + * The calculated ECC bytes are stored into other isolated >> + * area which is not accessible to users. >> + * This is why chip->ecc.bytes = 0. >> + */ >> + chip->ecc.bytes = 0; >> + chip->ecc.size = 512; >> + chip->ecc.strength = 8; >> + chip->ecc.read_page = toshiba_nand_read_page_benand; >> + chip->ecc.read_subpage = toshiba_nand_read_subpage_benand; >> + chip->ecc.write_page = nand_write_page_raw; >> + chip->ecc.read_page_raw = nand_read_page_raw; >> + chip->ecc.write_page_raw = nand_write_page_raw; >> + >> + chip->options |= NAND_SUBPAGE_READ; >> + >> + mtd_set_ooblayout(mtd, &nand_ooblayout_lp_ops); >> +} >> + >> static void toshiba_nand_decode_id(struct nand_chip *chip) >> { >> struct mtd_info *mtd = nand_to_mtd(chip); >> @@ -70,6 +155,10 @@ static int toshiba_nand_init(struct nand_chip *chip) >> if (nand_is_slc(chip)) >> chip->bbt_options |= NAND_BBT_SCAN2NDPAGE; >> >> + if (nand_is_slc(chip) && (chip->id.data[4] & 0x80) /* BENAND */ && >> + (chip->ecc.mode == NAND_ECC_ON_DIE)) >> + toshiba_nand_benand_init(chip); >> + > > Please move this block to toshiba_nand_init(). > >> return 0; >> } >> > > >