[PATCH] fix root cause of NAND trouble
ben-linux at fluff.org
Sun Nov 2 13:31:53 CET 2008
On Sat, Nov 01, 2008 at 07:50:54PM -0200, Werner Almesberger wrote:
> Ben Dooks wrote:
> > My only comment would be
> > that a slightly better solution would be to do:
> Oh, that's much better, thanks ! I hadn't realized at first that
> once could read the FIFO one byte at a time.
> The revised patch is below.
> - Werner
> ---------------------------------- cut here -----------------------------------
> With the introduction of optimized OOB reads in nand_read_subpage,
> the length of the data requested may not be a multiple of four bytes.
> This caused a partial read on the 2440, leading to false ECC errors
> and, worse, attempts to "correct" them.
> Note that there is a similar issue in s3c2440_nand_write_buf, which
> doesn't seem to cause trouble yet.
> Signed-off-by: Werner Almesberger <werner at openmoko.org>
> Index: ktrack/drivers/mtd/nand/s3c2410.c
> --- ktrack.orig/drivers/mtd/nand/s3c2410.c 2008-11-01 19:25:11.000000000 -0200
> +++ ktrack/drivers/mtd/nand/s3c2410.c 2008-11-01 19:38:07.000000000 -0200
> @@ -530,7 +530,12 @@
> static void s3c2440_nand_read_buf(struct mtd_info *mtd, u_char *buf, int len)
> struct s3c2410_nand_info *info = s3c2410_nand_mtd_toinfo(mtd);
> + u8 *ptr = buf + (len & ~3);
> + int i;
> readsl(info->regs + S3C2440_NFDATA, buf, len / 4);
> + for (i = 0; i != (len & 3); i++)
> + ptr[i] = readb(info->regs + S3C2440_NFDATA);
actually, thinking about it, we can probably get better code by doing:
/* mop up any non-word aligned length reads. */
for (i = (len & ~3); i != len; i++)
ptr[i] = readb(info->regs + S3C2440_NFDATA);
this removes the need for the extra ptr variable, the comment is also
> static void s3c2410_nand_write_buf(struct mtd_info *mtd, const u_char *buf, int len)
Q: What's a light-year?
A: One-third less calories than a regular year.
More information about the openmoko-kernel