aleksander at aleksander.es
Thu Apr 10 09:07:58 PDT 2014
On Thu, Apr 10, 2014 at 5:59 PM, Dan Williams <dcbw at redhat.com> wrote:
>> Thanks Aleksander for the quick response.
>> I tend to agree that DSS data should not be forwarded via VLAN interfaces as it's not a natural approach for such data pipes.
>> Also userspace has to deal with bogus MAC address inserted by mbim driver.
>> My request to handle this in userspace was based on fear that it would be much harder to integrate(or convince the vendors to integrate) modifications into usbnet and mbim kernel drivers.
>> I'm afraid that it would take ages to hit the upstream kernel and finally the end distros.
>> Of course as the long term solution a DSS channel should be visible as a character device.
> If there are known/quasi-standardized DSS protocols that eventually show
> up, I think libmbim or something else should probably speak those
> protocols too. I guess we'll have to see how this all falls out when
> DSS starts showing up (if it hasn't already?).
I don't think that will ever happen. DSS is the 'standardized
backdoor' that enables access to proprietary protocols to be used by
vendors; think of firmware upgrades for the easiest case. My feeling
is that from libmbim's perspective we'll always end up dealing with
just channel creations and destructions, not with the actual DSS
More information about the libmbim-devel