On Thu, Feb 13, 2025 at 01:17:41AM +0000, Thinh Nguyen wrote:
> On Wed, Feb 12, 2025, Andy Shevchenko wrote:
> > On Wed, Feb 12, 2025 at 12:36:04PM +0200, Andy Shevchenko wrote:
> > > On Wed, Feb 12, 2025 at 01:10:17AM +0000, Thinh Nguyen wrote:
> > > > On Mon, Feb 03, 2025, Andy Shevchenko wrote:
...
> > > > > static bool dwc3_gadget_endpoint_trbs_complete(struct dwc3_ep *dep,
> > >
> > > > > for (i = 0; i < DWC3_ENDPOINTS_NUM; i++) {
> > > > > dep = dwc->eps[i];
> > > > > + if (!dep)
> > > > > + continue;
> > > >
> > > > It should be fine to ignore this check here. Something must be really
> > > > wrong if there's an interrupt pointing to an endpoint that we shouldn't
> > > > be touching. If we do add a check, we should print a warn or something
> > > > here. But that should be a patch separate from this.
> > >
> > > Theoretically everything is possible as it may be HW integration bug,
> > > for example. But are you asking about separate patch even from the rest
> > > of the checks? Please, elaborate what do you want to see.
> >
> > Re-reading the code again, I don't understand. If we get to this loop
> > ever (theoretically it might be an old IP with the reserved endpoints),
> > we crash the kernel on the first gap in the array. And since the function
> > is called on an endpoint, it doesn't mean that all endpoints are allocated,
> > so I do not see the justification to issue a warning here.
> > Or do you imply that DWC3_VER_IS_PRIOR(DWC3, 183A) may not have an HW
> > integration similar to what we have on Intel Merrifield?
> >
> > For now I'm going to leave this check as is.
>
> Oops, you are correct. I read this as the same logic as below.
NP. Thank you for the review, and thanks for acking the next version!
--
With Best Regards,
Andy Shevchenko
