On Mon, Jul 22, 2024 at 09:00:07AM -0400, Chris Wulff wrote:
> On Sun, Jul 21, 2024 at 9:07 PM Alan Stern <stern@xxxxxxxxxxxxxxxxxxx> wrote:
> >
> > On Sun, Jul 21, 2024 at 03:20:49PM -0400, crwulff@xxxxxxxxx wrote:
> > > From: Chris Wulff <crwulff@xxxxxxxxx>
> > >
> > > Make sure the descriptor has been set before looking at maxpacket.
> > > This fixes a null pointer panic in this case.
> > >
> > > This may happen if the gadget doesn't properly set up the endpoint
> > > for the current speed, or the gadget descriptors are malformed and
> > > the descriptor for the speed/endpoint are not found.
> >
> > If that happens, doesn't it mean there's a bug in the gadget driver?
> > And if there's a bug, don't we want to be told about it by a big
> > impossible-to-miss error message, so the bug can be fixed?
>
> Yes, this is an indicator of a problem in a gadget driver as was the
> previous check for a zero max packet size. In this case, the panic
> is in an interrupt context and it doesn't make it out to the console.
> This just results in a system freeze without this fix.
>
> >
> > > Fixes: 54f83b8c8ea9 ("USB: gadget: Reject endpoints with 0 maxpacket value")
> > > Cc: stable@xxxxxxxxxxxxxxx
> > > Signed-off-by: Chris Wulff <crwulff@xxxxxxxxx>
> > > ---
> > > drivers/usb/gadget/udc/core.c | 2 +-
> > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/usb/gadget/udc/core.c b/drivers/usb/gadget/udc/core.c
> > > index 2dfae7a17b3f..36a5d5935889 100644
> > > --- a/drivers/usb/gadget/udc/core.c
> > > +++ b/drivers/usb/gadget/udc/core.c
> > > @@ -118,7 +118,7 @@ int usb_ep_enable(struct usb_ep *ep)
> > > goto out;
> > >
> > > /* UDC drivers can't handle endpoints with maxpacket size 0 */
> > > - if (usb_endpoint_maxp(ep->desc) == 0) {
> > > + if (!ep->desc || usb_endpoint_maxp(ep->desc) == 0) {
> > > /*
> > > * We should log an error message here, but we can't call
> > > * dev_err() because there's no way to find the gadget
> >
> > This will just hide the error. That's not good.
>
> The previous check was also hiding the error, and introduced a panic.
> I could add a printk to that error case, though it would be unassociated
> with the gadget that caused the problem. This function does also return
> an error code when it fails, so the calling function can check that and
> print an error.
Okay. It wouldn't hurt to print out an error message, even if there's
no way to tell which gadget it refers to. A dump_stack() would help in
that regard, but it won't be needed if the guilty party will always be
pretty obvious.
By the way, how did you manage to trigger this error? None of the
in-kernel gadget drivers are known to have this bug, and both the
gadgetfs and raw_gadget drivers prevent userspace from doing it. Were
you testing a gadget driver that was under development?
Alan Stern
