Re: [PATCH] mtd: super: don't reply on mtdblock device minor

From: Miquel Raynal <miquel.raynal@bootlin.com>
To: Daniel Golle <daniel@makrotopia.org>
Cc: linux-mtd@lists.infradead.org,
	Vignesh Raghavendra <vigneshr@ti.com>,
	Richard Weinberger <richard@nod.at>,
	David Woodhouse <dwmw2@infradead.org>
Subject: Re: [PATCH] mtd: super: don't reply on mtdblock device minor
Date: Thu, 8 Apr 2021 08:15:08 +0200	[thread overview]
Message-ID: <20210408081508.1d04eb31@xps13> (raw)
In-Reply-To: <YG4qVPz68WAwPtB+@makrotopia.org>

Hi Daniel,

Daniel Golle <daniel@makrotopia.org> wrote on Wed, 7 Apr 2021 22:55:32
+0100:

> For mtdblock devices with partitions (ie. part_bits != 0) the
> assumption that the minor number of the mtdblock device matches
> the mtdnum doesn't hold true.
> Properly resolve mtd device from blktrans layer instead.
> 
> Signed-off-by: Daniel Golle <daniel@makrotopia.org>
> ---
>  drivers/mtd/mtdsuper.c | 33 ++++++++++++++++++++++++++-------
>  1 file changed, 26 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/mtd/mtdsuper.c b/drivers/mtd/mtdsuper.c
> index c3e2098372f2e..dc52f9929c8fb 100644
> --- a/drivers/mtd/mtdsuper.c
> +++ b/drivers/mtd/mtdsuper.c
> @@ -9,6 +9,7 @@
>   */
>  
>  #include <linux/mtd/super.h>
> +#include <linux/mtd/blktrans.h>
>  #include <linux/namei.h>
>  #include <linux/export.h>
>  #include <linux/ctype.h>
> @@ -121,7 +122,8 @@ int get_tree_mtd(struct fs_context *fc,
>  {
>  #ifdef CONFIG_BLOCK
>  	struct block_device *bdev;
> -	int ret, major;
> +	struct mtd_blktrans_dev *blktrans_dev;
> +	int ret, major, part_bits;
>  #endif
>  	int mtdnr;
>  
> @@ -169,21 +171,38 @@ int get_tree_mtd(struct fs_context *fc,
>  	/* try the old way - the hack where we allowed users to mount
>  	 * /dev/mtdblock$(n) but didn't actually _use_ the blockdev
>  	 */
> -	bdev = lookup_bdev(fc->source);
> +	bdev = blkdev_get_by_path(fc->source, FMODE_READ, NULL);
>  	if (IS_ERR(bdev)) {
>  		ret = PTR_ERR(bdev);
>  		errorf(fc, "MTD: Couldn't look up '%s': %d", fc->source, ret);
>  		return ret;
>  	}
> -	pr_debug("MTDSB: lookup_bdev() returned 0\n");
> +	pr_debug("MTDSB: blkdev_get_by_path() returned 0\n");
>  
>  	major = MAJOR(bdev->bd_dev);
> -	mtdnr = MINOR(bdev->bd_dev);
> -	bdput(bdev);
>  
> -	if (major == MTD_BLOCK_MAJOR)
> -		return mtd_get_sb_by_nr(fc, mtdnr, fill_super);
> +	if (major == MTD_BLOCK_MAJOR) {
> +		if (!bdev->bd_disk) {
> +			blkdev_put(bdev, FMODE_READ);
> +			BUG();
> +			return -EINVAL;
> +		}
> +
> +		blktrans_dev = (struct mtd_blktrans_dev *)(bdev->bd_disk->private_data);
> +		if (!blktrans_dev || !blktrans_dev->tr) {
> +			blkdev_put(bdev, FMODE_READ);
> +			BUG();

Could you avoid these calls? I doubt they bring anything useful and
returning -EINVAL should be enough anyway.

> +			return -EINVAL;
> +		}
> +		mtdnr = blktrans_dev->devnum;
> +		part_bits = blktrans_dev->tr->part_bits;
> +		blkdev_put(bdev, FMODE_READ);
> +		if (MINOR(bdev->bd_dev) != (mtdnr << part_bits))
> +			return -EINVAL;
>  
> +		return mtd_get_sb_by_nr(fc, mtdnr, fill_super);
> +	}
> +	blkdev_put(bdev, FMODE_READ);
>  #endif /* CONFIG_BLOCK */
>  
>  	if (!(fc->sb_flags & SB_SILENT))

Otherwise lgtm.

Thanks,
Miquèl

______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/