From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from n17.bullet.mail.mud.yahoo.com ([68.142.206.144]) by bombadil.infradead.org with smtp (Exim 4.69 #1 (Red Hat Linux)) id 1LctP4-0004Gg-HI for linux-mtd@lists.infradead.org; Fri, 27 Feb 2009 03:28:49 +0000 From: David Brownell To: Andrew Morton Subject: Re: [patch/RESEND 2.6.29-rc6] NAND: fix "raw" reads with ECC syndrome layouts Date: Thu, 26 Feb 2009 19:28:43 -0800 References: <200902241508.13835.david-b@pacbell.net> <20090226122543.797eac77.akpm@linux-foundation.org> In-Reply-To: <20090226122543.797eac77.akpm@linux-foundation.org> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 8bit Content-Disposition: inline Message-Id: <200902261928.43642.david-b@pacbell.net> Cc: linux-mtd@lists.infradead.org, linux-kernel@vger.kernel.org List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Thursday 26 February 2009, Andrew Morton wrote: > > +     temp = chip->ecc.steps; > > +     do { > > +             ... > > +     } while (--temp); > > It would be clearer to code this as a plain old up-counting for loop. Downside: GCC adds a pointless jump-to-loop-test instruction at the top of the loop ... but we know there's at least one step, so there's no need to do that. And up-counting needs an extra variable, with initialization etc. ======== CUT HERE From: David Brownell Cosmetic fixes to the patch fixing raw HW_SYNDROME read/write. Signed-off-by: David Brownell --- drivers/mtd/nand/nand_base.c | 28 +++++++++++++--------------- 1 file changed, 13 insertions(+), 15 deletions(-) --- a/drivers/mtd/nand/nand_base.c +++ b/drivers/mtd/nand/nand_base.c @@ -773,10 +773,9 @@ static int nand_read_page_raw_syndrome(s int eccsize = chip->ecc.size; int eccbytes = chip->ecc.bytes; uint8_t *oob = chip->oob_poi; - int temp; + int steps, size; - temp = chip->ecc.steps; - do { + for (steps = chip->ecc.steps; steps > 0; steps--) { chip->read_buf(mtd, buf, eccsize); buf += eccsize; @@ -792,11 +791,11 @@ static int nand_read_page_raw_syndrome(s chip->read_buf(mtd, oob, chip->ecc.postpad); oob += chip->ecc.postpad; } - } while (--temp); + } - temp = mtd->oobsize - (oob - chip->oob_poi); - if (temp) - chip->read_buf(mtd, oob, temp); + size = mtd->oobsize - (oob - chip->oob_poi); + if (size) + chip->read_buf(mtd, oob, size); return 0; } @@ -1542,7 +1541,7 @@ static void nand_write_page_raw(struct m * @chip: nand chip info structure * @buf: data buffer * - * We need a special oob layout and handling even when ECC isn't used. + * We need a special oob layout and handling even when ECC isn't checked. */ static void nand_write_page_raw_syndrome(struct mtd_info *mtd, struct nand_chip *chip, const uint8_t *buf) @@ -1550,10 +1549,9 @@ static void nand_write_page_raw_syndrome int eccsize = chip->ecc.size; int eccbytes = chip->ecc.bytes; uint8_t *oob = chip->oob_poi; - int temp; + int steps, size; - temp = chip->ecc.steps; - do { + for (steps = chip->ecc.steps; steps > 0; steps--) { chip->write_buf(mtd, buf, eccsize); buf += eccsize; @@ -1569,11 +1567,11 @@ static void nand_write_page_raw_syndrome chip->write_buf(mtd, oob, chip->ecc.postpad); oob += chip->ecc.postpad; } - } while (--temp); + } - temp = mtd->oobsize - (oob - chip->oob_poi); - if (temp) - chip->write_buf(mtd, oob, temp); + size = mtd->oobsize - (oob - chip->oob_poi); + if (size) + chip->write_buf(mtd, oob, size); } /** * nand_write_page_swecc - [REPLACABLE] software ecc based page write function