On Thu, Jun 09, 2022 at 09:42:15AM +0800, Dongliang Mu wrote:
> On Thu, Jun 2, 2022 at 2:45 PM Johan Hovold <johan@xxxxxxxxxx> wrote:
> >
> > On Thu, Jun 02, 2022 at 01:50:24PM +0800, Dongliang Mu wrote:
> > > From: Dongliang Mu <mudongliangabcd@xxxxxxxxx>
> > >
> > > Commit d725d20e81c2 ("media: flexcop-usb: sanity checking of endpoint type
> > > ") adds a sanity check for endpoint[1], but fails to modify the sanity
> > > check of bNumEndpoints.
> > >
> > > Fix this by modifying the sanity check of bNumEndpoints to 2.
> >
> > Are you sure that d725d20e81c2 ("media: flexcop-usb: sanity checking of
> > endpoint type) was correct?
>
> I am not sure if fc_usb->uintf->cur_altsetting->endpoint[1] is right or not.
>
> But the sanity check of bNumEndpoints does not match the following dereference.
I should know better than try to be subtle on the internet: Your fix is
wrong.
You need to try to understand the code you're changing. Not just do
pattern matching.
> > > Fixes: d725d20e81c2 ("media: flexcop-usb: sanity checking of endpoint type")
> > > Signed-off-by: Dongliang Mu <mudongliangabcd@xxxxxxxxx>
> > > ---
> > > drivers/media/usb/b2c2/flexcop-usb.c | 2 +-
> > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/media/usb/b2c2/flexcop-usb.c b/drivers/media/usb/b2c2/flexcop-usb.c
> > > index 7835bb0f32fc..eb54bfdda01f 100644
> > > --- a/drivers/media/usb/b2c2/flexcop-usb.c
> > > +++ b/drivers/media/usb/b2c2/flexcop-usb.c
> > > @@ -509,7 +509,7 @@ static int flexcop_usb_init(struct flexcop_usb *fc_usb)
> > > return ret;
> > > }
> > >
> > > - if (fc_usb->uintf->cur_altsetting->desc.bNumEndpoints < 1)
> > > + if (fc_usb->uintf->cur_altsetting->desc.bNumEndpoints < 2)
> > > return -ENODEV;
> > > if (!usb_endpoint_is_isoc_in(&fc_usb->uintf->cur_altsetting->endpoint[1].desc))
> > > return -ENODEV;
> >
> > It looks like it should have verified the type of endpoint[0] here
> > instead.
>
> This sanity check originally makes sure the
> fc_usb->uintf->cur_altsetting->endpoint[0] in
> flexcop_usb_transfer_init.
Yes, and that is the only endpoint that this driver uses.
I've posted a fix here:
https://lore.kernel.org/r/20220609135341.19941-1-johan@xxxxxxxxxx
Johan
