linux-kernel - Re: [PATCH] usb: gadget: mv_u3d: add check for dma mapping error

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <871sytndwo.fsf@linux.intel.com>
Date:   Thu, 03 Nov 2016 10:52:39 +0200
From:   Felipe Balbi <balbi@...nel.org>
To:     Alexey Khoroshilov <khoroshilov@...ras.ru>
Cc:     Alexey Khoroshilov <khoroshilov@...ras.ru>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        linux-usb@...r.kernel.org, linux-kernel@...r.kernel.org,
        ldv-project@...uxtesting.org
Subject: Re: [PATCH] usb: gadget: mv_u3d: add check for dma mapping error


Hi,

Alexey Khoroshilov <khoroshilov@...ras.ru> writes:
> mv_u3d_req_to_trb() does not check for dma mapping errors.
>
> By the way, the patch improves readability of mv_u3d_start_queue()
> by rearranging its code with two semantic modifications:
> - assignment zero to ep->processing if usb_gadget_map_request() fails;
> - propagation of error code from mv_u3d_req_to_trb() instead of 
>   hardcoded -ENOMEM.

cleanups and fixes should be done separately.

> Found by Linux Driver Verification project (linuxtesting.org).
>
> Signed-off-by: Alexey Khoroshilov <khoroshilov@...ras.ru>
> ---
>  drivers/usb/gadget/udc/mv_u3d_core.c | 34 +++++++++++++++++++++-------------
>  1 file changed, 21 insertions(+), 13 deletions(-)
>
> diff --git a/drivers/usb/gadget/udc/mv_u3d_core.c b/drivers/usb/gadget/udc/mv_u3d_core.c
> index b9e19a591322..8d726bd767fd 100644
> --- a/drivers/usb/gadget/udc/mv_u3d_core.c
> +++ b/drivers/usb/gadget/udc/mv_u3d_core.c
> @@ -462,6 +462,12 @@ static int mv_u3d_req_to_trb(struct mv_u3d_req *req)
>  					req->trb_head->trb_hw,
>  					trb_num * sizeof(*trb_hw),
>  					DMA_BIDIRECTIONAL);
> +		if (dma_mapping_error(u3d->gadget.dev.parent,
> +					req->trb_head->trb_dma)) {
> +			kfree(req->trb_head->trb_hw);
> +			kfree(req->trb_head);
> +			return -EFAULT;
> +		}
>  
>  		req->chain = 1;
>  	}

this is one patch: add dma_mapping_error() check

AKA $subject :-p

> @@ -487,30 +493,32 @@ mv_u3d_start_queue(struct mv_u3d_ep *ep)
>  	ret = usb_gadget_map_request(&u3d->gadget, &req->req,
>  					mv_u3d_ep_dir(ep));
>  	if (ret)
> -		return ret;
> +		goto break_processing;
>  
>  	req->req.status = -EINPROGRESS;
>  	req->req.actual = 0;
>  	req->trb_count = 0;
>  
> -	/* build trbs and push them to device queue */
> -	if (!mv_u3d_req_to_trb(req)) {
> -		ret = mv_u3d_queue_trb(ep, req);
> -		if (ret) {
> -			ep->processing = 0;
> -			return ret;
> -		}
> -	} else {
> -		ep->processing = 0;
> +	/* build trbs */
> +	ret = mv_u3d_req_to_trb(req);
> +	if (ret) {
>  		dev_err(u3d->dev, "%s, mv_u3d_req_to_trb fail\n", __func__);
> -		return -ENOMEM;
> +		goto break_processing;
>  	}
>  
> +	/* and push them to device queue */
> +	ret = mv_u3d_queue_trb(ep, req);
> +	if (ret)
> +		goto break_processing;
> +
>  	/* irq handler advances the queue */
> -	if (req)
> -		list_add_tail(&req->queue, &ep->queue);
> +	list_add_tail(&req->queue, &ep->queue);
>  
>  	return 0;
> +
> +break_processing:
> +	ep->processing = 0;
> +	return ret;
>  }
>  
>  static int mv_u3d_ep_enable(struct usb_ep *_ep,

this is another, unrelated patch. Please split

-- 
balbi

Download attachment "signature.asc" of type "application/pgp-signature" (801 bytes)