linux-kernel - [usb-gadget-udc] question about null check after calling phys_to

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-next>] [day] [month] [year] [list]

Date:   Tue, 02 May 2017 15:24:40 -0500
From:   "Gustavo A. R. Silva" <garsilva@...eddedor.com>
To:     Felipe Balbi <balbi@...nel.org>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Michal Nazarewicz <mina86@...a86.com>
Cc:     linux-geode@...ts.infradead.org, linux-usb@...r.kernel.org,
        linux-kernel@...r.kernel.org,
        Peter Senna Tschudin <peter.senna@...il.com>
Subject: [usb-gadget-udc] question about null check after calling
 phys_to_virt() function


Hello everybody,

While looking into Coverity ID 145958 I ran into the following piece  
of code at drivers/usb/gadget/udc/amd5536udc.c:852:

} else if (i == buf_len) {
         /* first td */
         td = (struct udc_data_dma *)phys_to_virt(
                                 req->td_data->next);
         td->status = 0;
} else {
         td = (struct udc_data_dma *)phys_to_virt(last->next);
         td->status = 0;
}

if (td)
         td->bufptr = req->req.dma + i; /* assign buffer */
else
         break;

The issue here is that _td_ pointer is being dereferenced before null check.

After searching for calls to phys_to_virt() function, I've noticed  
that is not common at all to test the returned address value.

So either the null check at line 862 is not needed or a null check  
before each td->status = 0; needs to be added.

I think it would be good to apply a patch like the following one:

-                       td->status = 0;
+
+                       if (td)
+                               td->status = 0;
+                       else
+                               break;
                 } else {
                         td = (struct udc_data_dma *)phys_to_virt(last->next);
-                       td->status = 0;
+
+                       if (td)
+                               td->status = 0;
+                       else
+                               break;
                 }

-               if (td)
-                       td->bufptr = req->req.dma + i; /* assign buffer */
-               else
-                       break;
+               td->bufptr = req->req.dma + i; /* assign buffer */


What do you think?

Thanks in advance for your comments
--
Gustavo A. R. Silva