Re: [PATCH 3/8] usb: typec: ucsi: glink: check message data sizes

[Dmitry Baryshkov]

On Tue, 16 Apr 2024 at 17:33, Konrad Dybcio <konrad.dybcio@xxxxxxxxxx> wrote:
>
>
>
> On 4/16/24 04:20, Dmitry Baryshkov wrote:
> > The driver gets data from the DSP firmware. Sanitize data size before
> > reading corresponding message structures.
> >
> > Fixes: 62b5412b1f4a ("usb: typec: ucsi: add PMIC Glink UCSI driver")
> > Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@xxxxxxxxxx>
> > ---
>
> I think more backstory would be beneficial here.. Does this happen often?
> What are the consequences? What are the causes? Can there be one-off invalid
> messages, or does that mean the firwmare has entered some unstable state?
>
> And I suppose, if answer to the last question is "unstable state", are we
> doing something incorrectly in Linux that causes it to happen?

No. I haven't seen such cases. However as we are getting the data from
external entity which don't control, we should check that the messages
conform to what we are expecting.



-- 
With best wishes
Dmitry