Re: [PATCH] Revert "ubifs: Fix memory leak bug in alloc_ubifs_info() error path"

[Romain Izard]

On Thu, Oct 24, 2019 at 10:25:35AM +0200, Richard Weinberger wrote:
> This reverts commit 9163e0184bd7d5f779934d34581843f699ad2ffd.
> 
> At the point when ubifs_fill_super() runs, we have already a reference
> to the super block. So upon deactivate_locked_super() c will get
> free()'ed via ->kill_sb().

And without a revert, trying to mount a UBI volume with a squashfs image
on it as a UBIFS partition will lead to kernel panics due to the double
free.

> 
> Cc: Wenwen Wang <wenwen@xxxxxxxxxx>
> Fixes: 9163e0184bd7 ("ubifs: Fix memory leak bug in alloc_ubifs_info() error path")
> Reported-by: https://twitter.com/grsecurity/status/1180609139359277056
> Signed-off-by: Richard Weinberger <richard@xxxxxx>

Tested-by: Romain Izard <romain.izard.pro@xxxxxxxxx>

> ---
>  fs/ubifs/super.c | 4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)
> 
> diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c
> index 7d4547e5202d..5e1e8ec0589e 100644
> --- a/fs/ubifs/super.c
> +++ b/fs/ubifs/super.c
> @@ -2267,10 +2267,8 @@ static struct dentry *ubifs_mount(struct file_system_type *fs_type, int flags,
>  		}
>  	} else {
>  		err = ubifs_fill_super(sb, data, flags & SB_SILENT ? 1 : 0);
> -		if (err) {
> -			kfree(c);
> +		if (err)
>  			goto out_deact;
> -		}
>  		/* We do not support atime */
>  		sb->s_flags |= SB_ACTIVE;
>  		if (IS_ENABLED(CONFIG_UBIFS_ATIME_SUPPORT))
> -- 
> 2.16.4
>