Re: [PATCH] scsi: integer overflow in megadev_ioctl()

[Måns Rullgård]

"Chen.Yu" <chyyuu@xxxxxxxxx> writes:

> From: "Chen.Yu" <chyyuu@xxxxxxxxx>
>
> There is a potential integer overflow in megadev_ioctl() if
> userspace passes in a large u32 variable uioc.adapno.
> The int variable adapno would < 0, leading to an error
> array access for hdb_soft_state[adapno], or an error
> copy_to_user(uioc.uioc_uaddr, mcontroller+adapno,..)
>
> Reported-by: Wenliang Fan <fanwlexca@xxxxxxxxx>
> Suggested-by: Qixue Xiao <xiaoqixue_1@xxxxxxx>
> Signed-off-by: Yu Chen <chyyuu@xxxxxxxxx>
> Reviewed-by: Levente Kurusa <levex@xxxxxxxxx>
> ---
>  drivers/scsi/megaraid.c | 15 ++++++++++++---
>  1 file changed, 12 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/scsi/megaraid.c b/drivers/scsi/megaraid.c
> index 816db12..0b90c54 100644
> --- a/drivers/scsi/megaraid.c
> +++ b/drivers/scsi/megaraid.c
> @@ -3099,7 +3099,10 @@ megadev_ioctl(struct file *filep, unsigned int cmd, unsigned long arg)
>  		/*
>  		 * Which adapter
>  		 */
> -		if( (adapno = GETADAP(uioc.adapno)) >= hba_count )
> +		adapno = GETADAP(uioc.adapno);
> +		if( adapno < 0 )
> +			return (-EINVAL);
> +		if( adapno >= hba_count )
>  			return (-ENODEV);

This relies on implementation-defined behaviour when converting an
unsigned integer to signed integer.  A simpler and more robust fix is to
make the local variable 'adapno' unsigned.

-- 
Måns Rullgård
mans@xxxxxxxxx
--
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/