Re: [MM] Boost Option/HSO probings

[Dan Williams <dcbw redhat com>]

On Wed, 2012-10-31 at 11:51 +0100, Aleksander Morgado wrote:
> Hey Dan,
> 
> I prepared a patch which would make the option/hso modem probing almost
> instant, by moving the port type hint logic (the one reading the
> per-port 'hsotype' file) to a custom init operation done for each tty
> port when the probing sequence starts, instead of doing it once probing
> is finished. With this patch we really rely on the information given by
> the files; so e.g. if the 'hsotype' file for a given port tells us it's
> "Control" we do assume the port is the primary AT port (we won't do any
> AT probing in the port). Of course, if no 'hsotype' file is found for a
> given port, we just launch the normal port probing sequence.
> 
> Another approach would be to grab the port type hint and then only
> request to probe for the specific port type suggested, e.g. if the
> 'hsotype' file tells us the port is "Control", remove QCDM probing; or,
> if it tells us the port is "GPS", remove both AT and QCDM probings.
> 
> What do you think? Can we really always rely on the contents of the
> 'hsotype' files?

At the moment I think we can rely on it, and if we *can't*, then we need
to fix the kernel drivers anyway.  If we find some device that lies
about its ports like Huawei firmware sometimes does, then we can always
just do what you suggest and just filter the probing.

Patch seems to work fine for my Option devices.

Dan

> We could also use the same logic for Huawei and the GETPORTMODE result,
> btw, once we fix the already known issues with the NDIS port reported.
> 
> _______________________________________________
> networkmanager-list mailing list
> networkmanager-list gnome org
> https://mail.gnome.org/mailman/listinfo/networkmanager-list