On 1/21/2021 12:31 AM, Miroslav Lichvar wrote:
> On Wed, Jan 20, 2021 at 10:13:25PM +0000, Keller, Jacob E wrote:
>> It makes sense to remove forwarding, but I am not sure I understand the
>> justification for removing access to subscriptions.. if the subscription is
>> for read only data, why doesn't it make sense to allow that over the read
>> only socket?
>
> The subscription itself requires some state. If we say the new socket
> is safe to be accessed by untrusted applications, we need to be really
> sure they cannot do anything bad, e.g. create a large number of
> subscriptions to crash ptp4l or break subscriptions on the "rw"
> socket. Such issues would become security issues.
>
> There might be a way to provide subscriptions on the "ro" socket, but
> they need to be separate from the "rw" subscriptions and have some
> limiting implemented.
>
> I'd like to start with a minimal feature set that we can be
> comfortable with and maybe add other features later if there is a
> demand for them.
>
Right. This makes sense. We can obviously extend the RO sockets in the
future, but I think it makes sense to limit it. To me, it seems good to
have some condensed version of this explanation in the commit message or
somewhere, since it may not be obvious why it is limited to those on the
outside.
_______________________________________________
Linuxptp-devel mailing list
Linuxptp-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linuxptp-devel
