Re: [RFC] crypto: ccree - protect against short scatterlists

From: Gilad Ben-Yossef
Date: Mon Jan 27 2020 - 10:10:06 EST


On Mon, Jan 27, 2020 at 2:52 PM Geert Uytterhoeven <geert@xxxxxxxxxxxxxx> wrote:
>
> Hi Gilad,
>
> On Mon, Jan 27, 2020 at 1:29 PM Gilad Ben-Yossef <gilad@xxxxxxxxxxxxx> wrote:
> > Deal gracefully with the event of being handed a scatterlist
> > which is shorter than expected.
> >
> > This mitigates a crash in some cases of Crypto API calls due with
> > scatterlists with a NULL first buffer, despite the aead.h
> > forbidding doing so.
> >
> > Signed-off-by: Gilad Ben-Yossef <gilad@xxxxxxxxxxxxx>
> > Reported-by: Geert Uytterhoeven <geert@xxxxxxxxxxxxxx>
>
> Thanks for your patch!
>
> Unable to handle kernel paging request at virtual address fffeffffc0000000

OK, this is a progress of a sort.
We now crash during unmap, not map.

Sent another go. If this doesn't work I'll wait till I reunite with the board.
Blind debugging is hard...

Thanks again!
Gilad


values of Î will give rise to dom!