[PATCH v1] mtd: nandbiterrs: Have init function return 0 on success

Marc Gonzalez marc_gonzalez at sigmadesigns.com
Mon Dec 19 02:30:54 PST 2016


On 13/12/2016 15:36, Marc Gonzalez wrote:

> The init function currently returns -EIO on success. This behavior
> was probably chosen in order to avoid a subsequent rmmod, but this
> complicates failure detection from user-space.
> 
> Signed-off-by: Marc Gonzalez <marc_gonzalez at sigmadesigns.com>
> ---
> I'm not sure failures are reported as expected. I would expect
> the test to report a failure if the driver cannot fix less than
> $STRENGTH bit flips, but it doesn't, AFAICT.
> cf. incremental_errors_test which sets err to 0 in the 
> "After %d biterrors per subpage, read reported error %d\n"
> code path.
> ---
>  drivers/mtd/tests/nandbiterrs.c | 1 -
>  1 file changed, 1 deletion(-)
> 
> diff --git a/drivers/mtd/tests/nandbiterrs.c b/drivers/mtd/tests/nandbiterrs.c
> index f26dec896afa..41050bcae9f1 100644
> --- a/drivers/mtd/tests/nandbiterrs.c
> +++ b/drivers/mtd/tests/nandbiterrs.c
> @@ -403,7 +403,6 @@ static int __init mtd_nandbiterrs_init(void)
>  	if (err)
>  		goto exit_error;
>  
> -	err = -EIO;
>  	pr_info("finished successfully.\n");
>  	printk(KERN_INFO "==================================================\n");
>  

What are your thoughts on this patch?

Regards.




More information about the linux-mtd mailing list