Re: [PATCH v4 2/4] ataflop: address add_disk() error handling on probe

[Christoph Hellwig <hch@xxxxxx>]

On Wed, Nov 03, 2021 at 11:12:56AM -0700, Luis Chamberlain wrote:
> We need to cleanup resources on the probe() callback registered
> with __register_blkdev(), now that add_disk() error handling is
> supported. Address this.
> 
> Signed-off-by: Luis Chamberlain <mcgrof@xxxxxxxxxx>
> ---
>  drivers/block/ataflop.c | 16 +++++++++++++---
>  1 file changed, 13 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/block/ataflop.c b/drivers/block/ataflop.c
> index 170dd193cef6..e981b351f37e 100644
> --- a/drivers/block/ataflop.c
> +++ b/drivers/block/ataflop.c
> @@ -2012,6 +2012,7 @@ static void ataflop_probe(dev_t dev)
>  {
>  	int drive = MINOR(dev) & 3;
>  	int type  = MINOR(dev) >> 2;
> +	int err = 0;
>  
>  	if (type)
>  		type--;
> @@ -2019,11 +2020,20 @@ static void ataflop_probe(dev_t dev)
>  	if (drive >= FD_MAX_UNITS || type >= NUM_DISK_MINORS)
>  		return;
>  	if (!unit[drive].disk[type]) {
> +		err = ataflop_alloc_disk(drive, type);
> +		if (err == 0) {
> +			err = add_disk(unit[drive].disk[type]);
> +			if (err)
> +				goto err_out;
> +			else
> +				unit[drive].registered[type] = true;

This looks weird.  Why not:

 	if (unit[drive].disk[type])
		return;

	if (ataflop_alloc_disk(drive, type))
		return;
	if (add_disk(unit[drive].disk[type]))
		goto cleanup_disk;
	unit[drive].registered[type] = true;
	return;

cleanup_disk:
	blk_cleanup_disk(unit[drive].disk[type]);
	unit[drive].disk[type] = NULL;