linux-kernel - Re: [PATCH] Prevent buffer overflow in UVC Gadget setup handler

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 for Android: free password hash cracker in your pocket

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

Message-ID: <abe0cc82-2fb8-2e29-e7b9-90fe6ae4d203@gmail.com>
Date:   Thu, 1 Dec 2022 15:22:54 +0100
From:   Szymon Heidrich <szymon.heidrich@...il.com>
To:     Dan Scally <dan.scally@...asonboard.com>,
        laurent.pinchart@...asonboard.com
Cc:     Felipe Balbi <balbi@...nel.org>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        linux-usb@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] Prevent buffer overflow in UVC Gadget setup handler

On 01/12/2022 14:49, Dan Scally wrote:
> Hello - thanks for the patch
> 
> On 01/12/2022 12:21, Szymon Heidrich wrote:
>> Setup function uvc_function_setup
> 
> 
> You've written uvc_function_setup here, but the code changes uvc_function_ep0_complete.

Yes, this was intentional as uvc_function_setup prevents handling of control
transfer requests with wLength grater than UVC_MAX_REQUEST_SIZE.
The uvc_function_ep0_complete function handles data phase thus was modified.

> 
>>   permits control transfer
>> requests with up to 64 bytes of payload (UVC_MAX_REQUEST_SIZE),
>> data stage handler for OUT transfer uses memcpy to copy req->actual
>> bytes to uvc_event->data.data array of size 60. This may result
>> in an overflow of 4 bytes.
>>
>> Signed-off-by: Szymon Heidrich <szymon.heidrich@...il.com>
> 
> 
> Good catch
> 
>> ---
>>   drivers/usb/gadget/function/f_uvc.c | 5 +++--
>>   1 file changed, 3 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/usb/gadget/function/f_uvc.c b/drivers/usb/gadget/function/f_uvc.c
>> index 6e196e061..69c5eb3a3 100644
>> --- a/drivers/usb/gadget/function/f_uvc.c
>> +++ b/drivers/usb/gadget/function/f_uvc.c
>> @@ -216,8 +216,9 @@ uvc_function_ep0_complete(struct usb_ep *ep, struct usb_request *req)
>>             memset(&v4l2_event, 0, sizeof(v4l2_event));
>>           v4l2_event.type = UVC_EVENT_DATA;
>> -        uvc_event->data.length = req->actual;
>> -        memcpy(&uvc_event->data.data, req->buf, req->actual);
>> +        uvc_event->data.length = (req->actual > sizeof(uvc_event->data.data) ?
>> +            sizeof(uvc_event->data.data) : req->actual);
> 
> 
> There's a clamp() macro in f_uvc.c, can we use that?
> 
>> +        memcpy(&uvc_event->data.data, req->buf, uvc_event->data.length);
>>           v4l2_event_queue(&uvc->vdev, &v4l2_event);
>>       }
>>   }

If it is more appropriate I will use min_t(unsigned int, req->actual, sizeof(uvc_event->data.data)).