modem mode switching

modem mode switching

[Ozan Çağlayan]

Hi,

I wonder why there are two different userspace utilities for mode
switching on HSDPA modems. modem-modeswitch is shipped with udev,
usb_modeswitch is another project which is maintained as well.

Do they diverge in some point? If not, are there any plan to merge them
together? I'm not sure but relevant kernel drivers for those modems
seems to also have some sort of mode switching. When I plug my Huawei
device which isn't listed in neither udev rules for the userspace, it
first gets probed as an unknown 0xffffffff device and then it gets
reconnected and probbed as a serial modem. If kernel-space is able to
handle those kind of quirks, why the user space mode switching tools exist?

Thanks,
Ozan

Re: modem mode switching

[Kay Sievers]

On Thu, Aug 13, 2009 at 17:40, Ozan Çağlayan<ozan@pardus.org.tr> wrote:
> I wonder why there are two different userspace utilities for mode
> switching on HSDPA modems. modem-modeswitch is shipped with udev,
> usb_modeswitch is another project which is maintained as well.
>
> Do they diverge in some point? If not, are there any plan to merge them
> together? I'm not sure but relevant kernel drivers for those modems
> seems to also have some sort of mode switching. When I plug my Huawei
> device which isn't listed in neither udev rules for the userspace, it
> first gets probed as an unknown 0xffffffff device and then it gets
> reconnected and probbed as a serial modem. If kernel-space is able to
> handle those kind of quirks, why the user space mode switching tools exist?

Dan should know.

Thanks,
Kay

Re: modem mode switching

[Dan Williams]

On Thu, 2009-08-13 at 17:59 +0200, Kay Sievers wrote:
> On Thu, Aug 13, 2009 at 17:40, Ozan Çağlayan<ozan@pardus.org.tr> wrote:
> > I wonder why there are two different userspace utilities for mode
> > switching on HSDPA modems. modem-modeswitch is shipped with udev,
> > usb_modeswitch is another project which is maintained as well.
> >
> > Do they diverge in some point? If not, are there any plan to merge them
> > together? I'm not sure but relevant kernel drivers for those modems
> > seems to also have some sort of mode switching. When I plug my Huawei
> > device which isn't listed in neither udev rules for the userspace, it
> > first gets probed as an unknown 0xffffffff device and then it gets
> > reconnected and probbed as a serial modem. If kernel-space is able to
> > handle those kind of quirks, why the user space mode switching tools exist?
> 
> Dan should know.

Various reasons, none of which I care about any more, since the tool
works and as long as there's *something* that does the job, that's
great.

We could move the mobile action code into a separate binary since it
doesn't look like it would be that easy to fit into usb_modeswitch's
current architecture.

It also looks like it doesn't distribute a udev rules file by default,
that would be good to keep upstream, since they are clearly not thinking
about distributions, but individual users, which is pretty
short-sighted:

"You should have a folder named "/etc/udev" or similar. Somewhere in
there (I have a folder "rules.d") you find some files with the extension
".rules". Create a new one (or edit an existing one, but by convention
not the default "50-something.rules"). I chose one named
"45-hotplug.rules" since hotplugging is what USB is about after all.
In the chosen/new file add the line

SUBSYSTEM="usb", SYSFS{idProduct}="<YourDefaultProdID>", SYSFS{idVendor}="<YourDefaultVendID>", RUN+="<YourPathToUSB_ModeSwitch>"
"

Obviously for a distro, we need this to run for any USB devices that are
specified in usb_modeswitch's config file, and that means we also need
to uncomment everything in that file when actually putting it in a
distro.

Dan

Re: modem mode switching

[Ozan Çağlayan]

Dan Williams wrote:
> On Thu, 2009-08-13 at 17:59 +0200, Kay Sievers wrote:
>   
>> On Thu, Aug 13, 2009 at 17:40, Ozan Çağlayan<ozan@pardus.org.tr> wrote:
>>     
>>> I wonder why there are two different userspace utilities for mode
>>> switching on HSDPA modems. modem-modeswitch is shipped with udev,
>>> usb_modeswitch is another project which is maintained as well.
>>>
>>> Do they diverge in some point? If not, are there any plan to merge them
>>> together? I'm not sure but relevant kernel drivers for those modems
>>> seems to also have some sort of mode switching. When I plug my Huawei
>>> device which isn't listed in neither udev rules for the userspace, it
>>> first gets probed as an unknown 0xffffffff device and then it gets
>>> reconnected and probbed as a serial modem. If kernel-space is able to
>>> handle those kind of quirks, why the user space mode switching tools exist?
>>>       
>> Dan should know.
>>     
>
> Various reasons, none of which I care about any more, since the tool
> works and as long as there's *something* that does the job, that's
> great.
>
> We could move the mobile action code into a separate binary since it
> doesn't look like it would be that easy to fit into usb_modeswitch's
> current architecture.
>   

So what I understand from here is that the tools are not mutually
exclusive in terms of functionality. Is there a thing like "mode
switching on X can only be done with usb_modeswitch/modem-modeswitch" ?

I'm trying to understand if I have to package and ship usb_modeswitch
tool along with the in-udev modem-modeswitch for making various devices
detected out-of-the-box.

Thanks,

Re: modem mode switching

[Dan Williams]

On Thu, 2009-08-13 at 19:57 +0300, Ozan Çağlayan wrote:
> Dan Williams wrote:
> > On Thu, 2009-08-13 at 17:59 +0200, Kay Sievers wrote:
> >   
> >> On Thu, Aug 13, 2009 at 17:40, Ozan Çağlayan<ozan@pardus.org.tr> wrote:
> >>     
> >>> I wonder why there are two different userspace utilities for mode
> >>> switching on HSDPA modems. modem-modeswitch is shipped with udev,
> >>> usb_modeswitch is another project which is maintained as well.
> >>>
> >>> Do they diverge in some point? If not, are there any plan to merge them
> >>> together? I'm not sure but relevant kernel drivers for those modems
> >>> seems to also have some sort of mode switching. When I plug my Huawei
> >>> device which isn't listed in neither udev rules for the userspace, it
> >>> first gets probed as an unknown 0xffffffff device and then it gets
> >>> reconnected and probbed as a serial modem. If kernel-space is able to
> >>> handle those kind of quirks, why the user space mode switching tools exist?
> >>>       
> >> Dan should know.
> >>     
> >
> > Various reasons, none of which I care about any more, since the tool
> > works and as long as there's *something* that does the job, that's
> > great.
> >
> > We could move the mobile action code into a separate binary since it
> > doesn't look like it would be that easy to fit into usb_modeswitch's
> > current architecture.
> >   
> 
> So what I understand from here is that the tools are not mutually
> exclusive in terms of functionality. Is there a thing like "mode
> switching on X can only be done with usb_modeswitch/modem-modeswitch" ?
> 
> I'm trying to understand if I have to package and ship usb_modeswitch
> tool along with the in-udev modem-modeswitch for making various devices
> detected out-of-the-box.

Now I remember why I disliked usb_modeswitch: it's architecture doesn't
allow out-of-the-box operation.  You apparently have to uncomment your
device in the config file, or tell it exactly what you want it to do,
which pretty much duplicates what's in the config file.

What usb_modeswitch *should* be doing here is to just run, find all
devices that match something in the config file, and modeswitch them
automatically.

Ideally, we just fix usb_modeswitch to not be so unhelpful for
packagers/distributions, because it's an active project.

Dan