Re: [Qemu-devel] [PATCH 1/2] usb: rearrange usb_ep_get()

[Gerd Hoffmann]

On Wed, Jan 30, 2019 at 02:37:01PM +0000, Liam Merwick wrote:
> There is no need to calculate the 'eps' variable in usb_ep_get()
> if 'ep' is the control endpoint.  Instead the calculation should
> be done after validating the input and the resulting pointer also
> validated before returning an entry indexed on the endpoint 'ep'.
> 
> Signed-off-by: Liam Merwick <address@hidden>
> Reviewed-by: Darren Kenny <address@hidden>
> Reviewed-by: Mark Kanda <address@hidden>
> Reviewed-by: Ameya More <address@hidden>
> ---
>  hw/usb/core.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/hw/usb/core.c b/hw/usb/core.c
> index 241ae66b1505..1aa0051b2b2d 100644
> --- a/hw/usb/core.c
> +++ b/hw/usb/core.c
> @@ -720,12 +720,13 @@ struct USBEndpoint *usb_ep_get(USBDevice *dev, int pid, 
> int ep)
>      if (dev == NULL) {
>          return NULL;
>      }
> -    eps = (pid == USB_TOKEN_IN) ? dev->ep_in : dev->ep_out;
>      if (ep == 0) {
>          return &dev->ep_ctl;
>      }
>      assert(pid == USB_TOKEN_IN || pid == USB_TOKEN_OUT);
>      assert(ep > 0 && ep <= USB_MAX_ENDPOINTS);
> +    eps = (pid == USB_TOKEN_IN) ? dev->ep_in : dev->ep_out;
> +    assert(eps != NULL);

That assert is rather pointless.  It's impossible for eps to be NULL at
this point.

cheers,
  Gerd