[PATCH] dfu: mtd: mark bad the MTD block on erase error

[Patrick DELAUNAY]

Hi,

On 11/28/22 10:22, Patrick Delaunay wrote:
> In the MTD DFU backend, it is needed to mark the NAND block bad when the
> erase failed with the -EIO error, as it is done in UBI and JFFS2 code.
>
> This operation is not done in the MTD framework, but the bad block
> tag (in BBM or in BBT) is required to avoid to write data on this block
> in the next DFU_OP_WRITE loop in mtd_block_op(): the code skip the bad
> blocks, tested by mtd_block_isbad().
>
> Without this patch, when the NAND block become bad on DFU write operation
> - low probability on new NAND - the DFU write operation will always failed
> because the failing block is never marked bad.
>
> This patch also adds a test to avoid to request an erase operation on a
> block already marked bad; this test is not performed in MTD framework
> in mtd_erase().
>
> Signed-off-by: Patrick Delaunay <patrick.delaunay at foss.st.com>
> ---
>
>   drivers/dfu/dfu_mtd.c | 26 ++++++++++++++++++--------
>   1 file changed, 18 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/dfu/dfu_mtd.c b/drivers/dfu/dfu_mtd.c
> index c7075f12eca9..4fb02c4c806c 100644
> --- a/drivers/dfu/dfu_mtd.c
> +++ b/drivers/dfu/dfu_mtd.c
> @@ -91,22 +91,32 @@ static int mtd_block_op(enum dfu_op op, struct dfu_entity *dfu,
>   				return -EIO;
>   			}
>   
> +			/* Skip the block if it is bad, don't erase it again */
> +			if (mtd_block_isbad(mtd, off)) {


"off" is the not correct offset here

=> need to be replace to "erase_op.addr"


> +				erase_op.addr += mtd->erasesize;
> +				continue;
> +			}
> +
>   			ret = mtd_erase(mtd, &erase_op);
>   
>   			if (ret) {
> -				/* Abort if its not a bad block error */
> -				if (ret != -EIO) {
> -					printf("Failure while erasing at offset 0x%llx\n",
> -					       erase_op.fail_addr);
> -					return 0;
> +				/* If this is not -EIO, we have no idea what to do. */
> +				if (ret == -EIO) {
> +					printf("Marking bad block at 0x%08llx (%d)\n",
> +					       erase_op.fail_addr, ret);
> +					ret = mtd_block_markbad(mtd, erase_op.addr);
> +				}
> +				/* Abort if it is not -EIO or can't mark bad */
> +				if (ret) {
> +					printf("Failure while erasing at offset 0x%llx (%d)\n",
> +					       erase_op.fail_addr, ret);
> +					return ret;
>   				}
> -				printf("Skipping bad block at 0x%08llx\n",
> -				       erase_op.addr);
>   			} else {
>   				remaining -= mtd->erasesize;
>   			}
>   
> -			/* Continue erase behind bad block */
> +			/* Continue erase behind the current block */
>   			erase_op.addr += mtd->erasesize;
>   		}
>   	}


Regards

Patrick