Re: [PATCH] mm: shmem: avoid overflowing in shmem_fallocate

[Naoya Horiguchi]

On Wed, Dec 03, 2014 at 07:24:07PM -0500, Sasha Levin wrote:
> "offset + len" has the potential of overflowing. Validate this user input
> first to avoid undefined behaviour.
> 
> Signed-off-by: Sasha Levin <sasha.levin@xxxxxxxxxx>
> ---
>  mm/shmem.c |    3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/mm/shmem.c b/mm/shmem.c
> index 185836b..5a0e344 100644
> --- a/mm/shmem.c
> +++ b/mm/shmem.c
> @@ -2098,6 +2098,9 @@ static long shmem_fallocate(struct file *file, int mode, loff_t offset,
>  	}
>  
>  	/* We need to check rlimit even when FALLOC_FL_KEEP_SIZE */
> +	error = -EOVERFLOW;
> +	if ((u64)len + offset < (u64)len)
> +		goto out;

Hi Sasha,

It seems to me that we already do some overflow check in common path,
do_fallocate():

        /* Check for wrap through zero too */
        if (((offset + len) > inode->i_sb->s_maxbytes) || ((offset + len) < 0))
                return -EFBIG;

Do we really need another check?

And this patch changes the return value of fallocate(2), so you need
update man document.

BTW, when I'm reading your patch, I noticed that returning -EOVERFLOW
(rather than -EFBIG) looks better when ((offset + len) < 0) in
do_fallocate() is true.

Thanks,
Naoya Horiguchi--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/