[PATCH v3 0/3] mtd: Add sysfs entries and account BBT blocks properly

[Gupta, Pekon]

>From: Ezequiel Garcia [mailto:ezequiel.garcia at free-electrons.com]
>
>Here's a new round of the patchset to add sysfs entries for the ECC stats
>struct fields. See the discussion around previous version for context:
>
>http://lists.infradead.org/pipermail/linux-mtd/2014-March/052804.html
>
>Changes from v2:
>
>  * As suggested by Pekon Gupta, use "ecc_failures" instead of
>    "uncorrectable_errors".
>
>  * Added documentation for the new sysfs ABI as requested by Greg KH.
>
>Changes from v1:
>
>  * As agreed I've added a sysfs file per field to replace the ecc_stats
>    sysfs file. The current proposal follows the "one file per value" sysfs
>    rule, adding the following device attributes:
>    - corrected_bits (ecc_stats.corrected)
>    - uncorrectable_errors (ecc_stats.failed)
>    - bad_blocks (ecc_stats.badblocks)
>    - bbt_blocks (ecc_stats.bbtblocks)
>
>  * Following Brian's request, added small wrapper nand_block_isreserved()
>    that checks if the NAND has a bad block table before querying the number
>    of reserved blocks.
>
>  * Finally, this new series drops patch: "nand: Account the blocks used by
>    the BBT in the ecc_stats". It's not yet clear how to best deal with the
>    current inconsistencies in MTD master and partition handling.
>
>The first patch adds the sysfs entries. The rest of the patches fixes the
>bbtblock accounting and allows to keep track of the reserved blocks correctly.
>
>This is a very old issue, probably around since the dawn of time,
>and nobody seemed to care much about it. However, with the introduction of
>the sysfs access, it's easier to access this stats and easier to
>see the issue.
>
>Ezequiel Garcia (3):
>  mtd: Add sysfs attributes to expose the ECC stats fields
>  mtd: Introduce mtd_block_isreserved()
>  mtd: Account for BBT blocks when a partition is being allocated
>
> Documentation/ABI/testing/sysfs-class-mtd | 38 ++++++++++++++++++++
> drivers/mtd/mtdcore.c                     | 59 +++++++++++++++++++++++++++++--
> drivers/mtd/mtdpart.c                     | 13 ++++++-
> drivers/mtd/nand/nand_base.c              | 18 ++++++++++
> drivers/mtd/nand/nand_bbt.c               | 14 ++++++++
> include/linux/mtd/mtd.h                   |  2 ++
> include/linux/mtd/nand.h                  |  1 +
> 7 files changed, 142 insertions(+), 3 deletions(-)
>
>--
>1.9.1

- Thanks Ezequiel, This will be helpful for NAND Flash users while debugging issues,
   And checking NAND flash quality (bit-flips numbers) during production.
- Tested on Beaglebone-Black using 4K NAND cape..
- Needed following fix in OMAP NAND driver to enable BBT creation and scanning.
------------- 
diff --git a/drivers/mtd/nand/omap2.c b/drivers/mtd/nand/omap2.c
index f0ed92e..2ad7451 100644
--- a/drivers/mtd/nand/omap2.c
+++ b/drivers/mtd/nand/omap2.c
@@ -1663,7 +1663,7 @@ static int omap_nand_probe(struct platform_device *pdev)
        mtd->owner              = THIS_MODULE;
        nand_chip               = &info->nand;
        nand_chip->ecc.priv     = NULL;
-       nand_chip->options      |= NAND_SKIP_BBTSCAN;
+       nand_chip->bbt_options  |= NAND_BBT_USE_FLASH;

        res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
        nand_chip->IO_ADDR_R = devm_ioremap_resource(&pdev->dev, res);
------------- 

So for whole series
Tested-by: Pekon Gupta <pekon at ti.com>


with regards, pekon