Re: [PATCH net-next v4 3/5] net: hdlc_fr: Improve the initial checks when we receive an skb

[Willem de Bruijn]

On Fri, Oct 30, 2020 at 3:21 PM Xie He <xie.he.0141@xxxxxxxxx> wrote:
>
> On Fri, Oct 30, 2020 at 9:31 AM Willem de Bruijn
> <willemdebruijn.kernel@xxxxxxxxx> wrote:
> >
> > > Add an fh->ea2 check to the initial checks in fr_rx. fh->ea2 == 1 means
> > > the second address byte is the final address byte. We only support the
> > > case where the address length is 2 bytes.
> >
> > Can you elaborate a bit for readers not intimately familiar with the codebase?
> >
> > Is there something in the following code that has this implicit
> > assumption on 2-byte address lengths?
>
> Yes, the address length must be 2 bytes, otherwise the 3rd and 4th
> bytes would not be the control and protocol fields as we assumed in
> the code.
>
> The frame format is specified in RFC 2427
> (https://tools.ietf.org/html/rfc2427). We can see the overall frame
> format on Page 3. If the address length is longer than 2 bytes, all
> the following fields will be shifted behind.

Thanks for that context. If it's not captured in the code, it would be
great to include in the commit message.

>From a quick scan, RFC 2427 does not appear to actually define the
Q.922 address. For that I ended up reading ITU-T doc "Q.922 : ISDN
data link layer specification for frame mode bearer services", section
3.2.