Re: Voice calls over qmi was Re: Incoming sms problem on Motorola Droid 4

Re: Voice calls over qmi was Re: Incoming sms problem on Motorola Droid 4

[Harald Welte]

[-- Attachment #1: Type: text/plain, Size: 669 bytes --]

Hi Stanislav,

On Fri, May 18, 2018 at 10:07:47PM +0200, Stanislav Sinyagin wrote:
> Simcom seems to be quite consistent in their support for this feature:
> it's available in 3 latest models, the 7100, 7500, and 7600 series,
> and they share the same AT command set for this.

Thanks a lot for pointing this out.  This is definitely good to know.

-- 
- Harald Welte <laforge@gnumonks.org>           http://laforge.gnumonks.org/
============================================================================
"Privacy in residential applications is a desirable marketing option."
                                                  (ETSI EN 300 175-7 Ch. A6)

Re: Voice calls over qmi was Re: Incoming sms problem on Motorola Droid 4

[Harald Welte]

[-- Attachment #1: Type: text/plain, Size: 3314 bytes --]

Hi Bob,

On Wed, May 16, 2018 at 04:10:55PM +0100, Bob Ham wrote:
> On 12/05/18 08:37, Harald Welte wrote:
> 
> > As there are no modems that reliably can deliver the actual audio data via USB
> > (aside some non-supported outdated experimental firmwares), we're only testing
> > the signaling plane so far.
> 
> I'm alarmed to see this statement.  Can I ask what you mean here when you
> say "no modems"? :-)  Do you mean on the market in general?

I mean on the market in general, yes.  Sure you can find the odd outlier here or
there, but overall, I would estimate 90% or more modems do not support audio in the
first place, and 98% do not support it over USB.

The only "proper" approach I've ever seen was by Sierra Wireless, whose Qualcomm
based 3G modems for some time had an experimental firmware branch you could ask for,
which would then offer a composite USB device that had an USB audio class device
next to the usual other (cdc-ethernet, AT-command, QMI, MBIM, ...) interfaces.

However, that is discontinued and not supported for years, and there is no replacement.
FAEs are still handing it out on request, though.

> I've been using the SIMCom SIM7100E which is specified as supporting audio
> data over a USB serial connection.  Unfortunately I've had a lot of trouble
> getting it to work but I've been collaborating with Stanislav Sinyagin
> (CC'd) who has a SIM7100E with an earlier firmware and who can reliably get
> audio by dd'ing /dev/ttyUSBX.  I'm waiting on our supplier to get the same
> firmware version.

Which exactly reflects my statement:  It's very hard to get reliable,
working audio with commitment from a vendor.  Sure, you may find it more
or less accidentially working for some products in some firmware
version, but I yet have to find any vendor who would actually say "this
is a supported feature, we test it before every release, and we will
maintain it for every future firmware update".

> There are earlier SIMCom modems with application notes describing how to get
> audio over USB too, for example the SIM5360.

Yes, there have always been some (few) modules that offered audio codec
frames over (usb) serial.  You could find some Gemalto 3G modules that
also did this.  I find this a horrible hack, and at least on those
modules wehre I remember it always had some kind of side-effect, like
you had to give up the [virtual] UART that was normally used for GPS and
re-purpose that for the codec frames over serial.

With most of those modern 4G modems running Linux inside the modem, I
really cannot understand why none of them goes the extra mile and
enables a USB audio gadget so it simply shows up as usb-audio device on
the host.  I mean, USB audio profile is 1990ies... and we have 2018.

While doing reverse engineering on some Qualcomm LE based modems before,
I saw that the Qualcomm reference source code even appeared to include
usb-audio composite device capabilities.

Regards,
	Harald
-- 
- Harald Welte <laforge@gnumonks.org>           http://laforge.gnumonks.org/
============================================================================
"Privacy in residential applications is a desirable marketing option."
                                                  (ETSI EN 300 175-7 Ch. A6)

Re: Voice calls over qmi was Re: Incoming sms problem on Motorola Droid 4

[Harald Welte]

[-- Attachment #1: Type: text/plain, Size: 2693 bytes --]

Hi Stanislav,

On Wed, May 16, 2018 at 05:15:07PM +0200, Stanislav Sinyagin wrote:
> also most mPCIE modems deliver PCI voice over dedicated wires on their
> mPCIE socket. These wires are standardized as "reserved" and most
> standard boards with mPCIE don't connect them anywhere.

This is the "PCM" interface I was referring-to.  This means you need
a PCM slave interface to interface with it, as the modems typically all
insist on being a PCM master.  The intention here is that you attach
some external audio codec IC that converts from the PCM to analog audio.

This means you cannot interface this PCM interface e.g. with standard
USB-Audio bridge ICs, as all those USB-Audi bridge ICs (basically "USB
soundcard" ICs) all also only implement the "master" of the PCM
interface and not the slave.

See http://laforge.gnumonks.org/blog/20170902-cellular_modems-voice/

In the end, you have to use something like the SSC (synchronous serial)
peripheal of Atmel SAM3/SAM4/... devices, or an XMOS device in order to
interface with that audio.

Needless to say, the lack of a standard for where PCM lines are on mPCIe
slots means that you cannot build any base board that will interoperate
with mPCIe modules from different vendors.

> But if you develop your own PCB, you can actually retrieve the voice
> signal. It just needs a bit of hacking.

The fact that a PCM bus is present on the hardware pins of the mPCIe socket
or the pads of a LGA module also still doesn't mean that you actually will
have working audio.

Many modem module maker do not obtain patent licenses for audio/voice, as
they know/expect their modems are typically only used in machine2machine
or other data-only applications.

Finally, even if the firmware and hardware interface is present, in many
cases the modem manufacturers make you sign a declaration that you will
only use the voice interface as some kind of "emergency communication"
only, and not as part of your normal product.  Once again, patent
licensing differences for voice vs. data-only use cases are to be blamed
for that.

Oh, and I'm not even touching the question on whether audio will work in
circuit-switched 2G/3G only, or whether it will also work with VoLTE,
and particularly not whether they will do SRVCC, etc.  That adds yet
another dimension to the problem.

Regards,
	Harald
-- 
- Harald Welte <laforge@gnumonks.org>           http://laforge.gnumonks.org/
============================================================================
"Privacy in residential applications is a desirable marketing option."
                                                  (ETSI EN 300 175-7 Ch. A6)

Re: Voice calls over qmi was Re: Incoming sms problem on Motorola Droid 4

[Bob Ham]

[-- Attachment #1: Type: text/plain, Size: 940 bytes --]

On 12/05/18 08:37, Harald Welte wrote:

> As there are no modems that reliably can deliver the actual audio data via USB
> (aside some non-supported outdated experimental firmwares), we're only testing
> the signaling plane so far.

Hi Harald,

I'm alarmed to see this statement.  Can I ask what you mean here when 
you say "no modems"? :-)  Do you mean on the market in general?

I've been using the SIMCom SIM7100E which is specified as supporting 
audio data over a USB serial connection.  Unfortunately I've had a lot 
of trouble getting it to work but I've been collaborating with Stanislav 
Sinyagin (CC'd) who has a SIM7100E with an earlier firmware and who can 
reliably get audio by dd'ing /dev/ttyUSBX.  I'm waiting on our supplier 
to get the same firmware version.

There are earlier SIMCom modems with application notes describing how to 
get audio over USB too, for example the SIM5360.

Regards,

Bob

Re: Voice calls over qmi was Re: Incoming sms problem on Motorola Droid 4

[Denis Kenzior]

[-- Attachment #1: Type: text/plain, Size: 1188 bytes --]

Hi Alexander, Joey,

On 05/12/2018 09:19 AM, Alexander Couzens wrote:
> Hi Joey,
> 
> nice to hear, somebody else is using it.
> 
> The voicecall driver needs to be refactored as Denis already pointed
> out on the mailinglist before it can go upstream.

If it wasn't clear from my earlier mail, I would like to see these 
patches upstreamed.  Do you guys have plans to do so?  I'm happy to 
review/help these along...

> 
> The original idea of the voice_generated.c/h files was to create a
> prototype how qmi generated files should look like and then write a
> code generator for it, as the other qmi related projects already do.

As you can tell we're not really sold on the whole code generation thing 
here :)  Can we simply stuff all structure definitions into voice.h and 
the parsers into qmiutils.[ch].

In the past we have parsed fairly complex TLV based objects using a 
common parsing function that took care of the various funny rules and 
corner cases.  Adding support for additional objects was then quite 
trivial.  See src/stkutil.[ch], start at 'parse_dataobj()'.  Perhaps 
this approach might be one for you to consider.

Regards,
-Denis

Re: Voice calls over qmi was Re: Incoming sms problem on Motorola Droid 4

[Pavel Machek]

[-- Attachment #1: Type: text/plain, Size: 4513 bytes --]

Hi!

> nice to hear, somebody else is using it.
> 
> The voicecall driver needs to be refactored as Denis already pointed
> out on the mailinglist before it can go upstream.
> 
> The original idea of the voice_generated.c/h files was to create a
> prototype how qmi generated files should look like and then write a
> code generator for it, as the other qmi related projects already do.
> 
> I'ven't yet took a closer look on your patches together with modems and
> a gsm network. Hopefully I can do it next week together with the
> required refactoring for upstream.

I got it to build, but quick test did not result in working call. I'm
using AT commands simultaneously on another channel.

Ideas welcome.

Best regards,
								Pavel

ce64aadfc798225807c25
Author: Alexander Couzens <lynxis@fe80.eu>
Date:   Tue Oct 17 10:11:19 2017 +0200

    call-list: fix a race condition in ofono_call_list_dial_callback

    If ofono_call_list_dial_callback is called later than
        ofono_call_list_notify, the new call is added, removed, added
	    again.

ofonod[6708]: drivers/qmimodem/gprs.c:extract_ss_info() radio in use 5
ofonod[6708]: drivers/qmimodem/gprs.c:extract_ss_info() radio in use 2
ofonod[6708]: src/gprs.c:registration_status_cb() /gobi_0 error 0
status 1
ofonod[6708]: src/gprs.c:ofono_gprs_status_notify() /gobi_0 status
registered (1)
ofonod[6708]: src/gprs.c:gprs_netreg_update() attach: 1,
driver_attached: 1
ofonod[6708]: drivers/qmimodem/network-registration.c:get_rssi_cb()
ofonod[6708]: drivers/qmimodem/network-registration.c:get_rssi_cb()
signal with -56 dBm on 5
ofonod[6708]: drivers/qmimodem/network-registration.c:get_rssi_cb()
ofonod[6708]: drivers/qmimodem/network-registration.c:get_rssi_cb()
signal with -56 dBm on 5
ofonod[6708]: drivers/qmimodem/gprs.c:ss_info_notify()
ofonod[6708]: drivers/qmimodem/gprs.c:handle_ss_info()
ofonod[6708]: drivers/qmimodem/gprs.c:extract_ss_info()
ofonod[6708]: drivers/qmimodem/gprs.c:extract_ss_info() radio in use 5
ofonod[6708]: drivers/qmimodem/gprs.c:extract_ss_info() radio in use 2
ofonod[6708]: src/gprs.c:ofono_gprs_status_notify() /gobi_0 status
registered (1)
ofonod[6708]: drivers/qmimodem/network-registration.c:ss_info_notify()
ofonod[6708]:
drivers/qmimodem/network-registration.c:extract_ss_info()
ofonod[6708]:
drivers/qmimodem/network-registration.c:extract_ss_info() serving
system status 1
ofonod[6708]:
drivers/qmimodem/network-registration.c:extract_ss_info() radio in use
5
ofonod[6708]:
drivers/qmimodem/network-registration.c:extract_ss_info() radio in use
2
ofonod[6708]:
drivers/qmimodem/network-registration.c:extract_ss_info() lac -1
cellid -1 tech -1
ofonod[6708]: src/network.c:ofono_netreg_status_notify() /gobi_0
status 1 tech -1 lac -1 ci -1
ofonod[6708]:
drivers/qmimodem/network-registration.c:qmi_current_operator()
ofonod[6708]: src/network.c:current_operator_callback() 0x5a0c98,
0x5a1f20
ofonod[6708]:
drivers/qmimodem/network-registration.c:qmi_signal_strength()
ofonod[6708]: src/gprs.c:netreg_status_changed() 1
ofonod[6708]: src/gprs.c:gprs_netreg_update() attach: 1,
driver_attached: 1
ofonod[6708]: drivers/qmimodem/network-registration.c:get_rssi_cb()
ofonod[6708]: drivers/qmimodem/network-registration.c:get_rssi_cb()
signal with -56 dBm on 5
ofonod[6708]: drivers/qmimodem/network-registration.c:event_notify()
ofonod[6708]: drivers/qmimodem/network-registration.c:event_notify()
signal with 80%(-55 dBm) on 5
ofonod[6708]: drivers/qmimodem/network-registration.c:event_notify()
ofonod[6708]: drivers/qmimodem/network-registration.c:event_notify()
signal with 80%(-56 dBm) on 5
ofonod[6708]: drivers/qmimodem/voicecall.c:dial_cb() QMI Error 3
ofonod[6708]: src/voicecall.c:dial_handle_result() Dial callback
returned error: Unknown error type
ofonod[6708]: drivers/qmimodem/network-registration.c:event_notify()
ofonod[6708]: drivers/qmimodem/network-registration.c:event_notify()
signal with 80%(-55 dBm) on 5
ofonod[6708]: drivers/qmimodem/network-registration.c:event_notify()
ofonod[6708]: drivers/qmimodem/network-registration.c:event_notify()
signal with 80%(-56 dBm) on 5
ofonod[6708]: drivers/qmimodem/network-registration.c:event_notify()
ofonod[6708]: drivers/qmimodem/network-registration.c:event_notify()
signal with 80%(-55 dBm) on 5



-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 181 bytes --]

Re: Voice calls over qmi was Re: Incoming sms problem on Motorola Droid 4

[Marcel Holtmann]

[-- Attachment #1: Type: text/plain, Size: 662 bytes --]

Hi Joey,

> nice to hear, somebody else is using it.
> 
> The voicecall driver needs to be refactored as Denis already pointed
> out on the mailinglist before it can go upstream.
> 
> The original idea of the voice_generated.c/h files was to create a
> prototype how qmi generated files should look like and then write a
> code generator for it, as the other qmi related projects already do.

I think that code generation is a bad idea. And other projects are drinking some cool aid. So I really prefer not to have generated code for QMI support upstream. It is also a total waste of time. Using QMI is actually really simple.

Regards

Marcel

Re: Voice calls over qmi was Re: Incoming sms problem on Motorola Droid 4

[Alexander Couzens]

[-- Attachment #1: Type: text/plain, Size: 737 bytes --]

Hi Joey,

nice to hear, somebody else is using it.

The voicecall driver needs to be refactored as Denis already pointed
out on the mailinglist before it can go upstream.

The original idea of the voice_generated.c/h files was to create a
prototype how qmi generated files should look like and then write a
code generator for it, as the other qmi related projects already do.

I'ven't yet took a closer look on your patches together with modems and
a gsm network. Hopefully I can do it next week together with the
required refactoring for upstream.

Best,
lynxis
-- 
Alexander Couzens

mail: lynxis(a)fe80.eu
jabber: lynxis(a)fe80.eu
mobile: +4915123277221
gpg: 390D CF78 8BF9 AA50 4F8F  F1E2 C29E 9DA6 A0DF 8604

[-- Attachment #2: attachment.sig --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: Voice calls over qmi was Re: Incoming sms problem on Motorola Droid 4

[Pavel Machek]

[-- Attachment #1: Type: text/plain, Size: 1435 bytes --]

Hi!

> > It looks like the QMI voicecall driver is just a stub with no operations
> actually implemented.
> 
> I hope I'm not stealing their thunder, but Alexander Couzens at sysmocom
> seems to be working on implementing it [0]. Last fall, I had
> incoming/outgoing calls in oFono working on my MSM8930-based Android phone
> with these patches. You will probably want to revert the "NOT_FOR_MERGE"
> commit(s).

Out of interest, what kind of phone how usable it is?

(Bacially everything works on Motorola Droid 4, with exception of
camera, and yes, I have some interesting hacks at places. That is with
current kernel and recent debian / Maemo Leste). 

https://pavelmachek.livejournal.com/141420.html?nojs=1

> I have some rough (outdated?) patches on top of theirs, for DTMF[1], and
> improving the implementation of release_specific()[2]. (I don't know some of
> the technical details of GSM and QMI, but those changes made declining
> incoming calls work the way I expected.)
> 
> -Joey
> 
> [0] https://git.sysmocom.de/ofono/log/?h=lynxis/master

Thanks for pointer.

include/ofono/call-list.h was missing; I provided a symlink to
include/call-list.h.

drivers/qmimodem/lde.c was missing. I provided an empty file. This
can't end well.

									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 181 bytes --]

Re: Voice calls over qmi was Re: Incoming sms problem on Motorola Droid 4

[Harald Welte]

[-- Attachment #1: Type: text/plain, Size: 1442 bytes --]

Hi Denis, Pavel, Marcel,

On Fri, May 11, 2018 at 08:02:01PM -0500, Denis Kenzior wrote:
> This reminds me... Weren't you the one who mentioned that you got voicecalls
> working on QMI devices?  Does that include the circuit switched control
> operations (e.g. dial, etc) as well?  Any chance that code might be
> upstream-able?

let me pull in Alexander "lynxis" Couzens into the loop, he has been doing all
QMI / ofono related developments for us in the Osmocom project.

My knowledge is that we're testing our network-side stack with voice calls
related tests (via ofono/QMI) for quite some months, but lynxis would know
all related details.

As there are no modems that reliably can deliver the actual audio data via USB
(aside some non-supported outdated experimental firmwares), we're only testing
the signaling plane so far.

We're building a custom board that contains four PCM bus slaves (at different clocks!)
in order to pick up the audio/voice data, but this doesn't exist yet.

See http://laforge.gnumonks.org/blog/20170902-cellular_modems-voice/ for more background
on the user/voice plane.

-- 
- Harald Welte <laforge@gnumonks.org>           http://laforge.gnumonks.org/
============================================================================
"Privacy in residential applications is a desirable marketing option."
                                                  (ETSI EN 300 175-7 Ch. A6)

Re: Voice calls over qmi was Re: Incoming sms problem on Motorola Droid 4

[Joey Hewitt]

[-- Attachment #1: Type: text/plain, Size: 978 bytes --]

Hi Denis and Pavel,

 > It looks like the QMI voicecall driver is just a stub with no 
operations actually implemented.

I hope I'm not stealing their thunder, but Alexander Couzens at sysmocom 
seems to be working on implementing it [0]. Last fall, I had 
incoming/outgoing calls in oFono working on my MSM8930-based Android 
phone with these patches. You will probably want to revert the 
"NOT_FOR_MERGE" commit(s).

I have some rough (outdated?) patches on top of theirs, for DTMF[1], and 
improving the implementation of release_specific()[2]. (I don't know 
some of the technical details of GSM and QMI, but those changes made 
declining incoming calls work the way I expected.)

-Joey

[0] https://git.sysmocom.de/ofono/log/?h=lynxis/master
[1] 
https://github.com/scintill/android_external_ofono/commit/306a3a6f19546b7b9b6b8df973d75710ce21bc3c
[2] 
https://github.com/scintill/android_external_ofono/commit/979fce83a1985b8a43f1b8f49ed746710792375f

Re: Voice calls over qmi was Re: Incoming sms problem on Motorola Droid 4

[Denis Kenzior]

[-- Attachment #1: Type: text/plain, Size: 380 bytes --]

Hi Harald,

> ofono supports voice calls over qmi, right? Is it possible that
> mdm6600 does not support them?
> 

This reminds me... Weren't you the one who mentioned that you got 
voicecalls working on QMI devices?  Does that include the circuit 
switched control operations (e.g. dial, etc) as well?  Any chance that 
code might be upstream-able?

Regards,
-Denis

Re: Voice calls over qmi was Re: Incoming sms problem on Motorola Droid 4

[Denis Kenzior]

[-- Attachment #1: Type: text/plain, Size: 234 bytes --]

Hi Pavel,

> ofono supports voice calls over qmi, right? Is it possible that
> mdm6600 does not support them?

It looks like the QMI voicecall driver is just a stub with no operations 
actually implemented.

Regards,
-Denis

Voice calls over qmi was Re: Incoming sms problem on Motorola Droid 4

[Pavel Machek]

[-- Attachment #1: Type: text/plain, Size: 3294 bytes --]

Hi!

> > More explanation needed it seems:
> > 
> > I can already do USB with AT commands, no multiplexer needed. We'd
> > actually prefer to go over serial (with multiplexer), because USB
> > needs to be powered down in idle, and you still need communication
> > ovre serial to know that wakeup of USB is needed.
> > 
> > Yes, it is complex :-(.
> 
> I missed the part that the weird serial multiplexer is wired independently. Then do everything via that serial multiplexer and never look back. However if that is as broken, then you will be really out of luck. Making this work reliable will be really hard.
> 
> While we have seen QMI protocol issues as well, but they have been less bad than the bolted on AT commands. Manufactures always get AT commands wrong since they let humans test it and not a real telephony stack like oFono.
> 

Yes, I understand why AT commands are bad. IIRC messages should work
over qmi... but voice calls do not:

user(a)devuan:/my/ofono$ sudo python test/dial-number 604343103
Using modem /gobi_0
Traceback (most recent call last):
  File "test/dial-number", line 40, in <module>
      path = vcm.Dial(number, hide_callerid)
        File "/usr/lib/python2.7/dist-packages/dbus/proxies.py", line
	70, in __call__
	    return self._proxy_method(*args, **keywords)
 File "/usr/lib/python2.7/dist-packages/dbus/proxies.py",
 line 145, in __call__
 **keywords)
 File
 "/usr/lib/python2.7/dist-packages/dbus/connection.py",
 line 651, in call_blocking message, timeout) dbus.exceptions.DBusException:
 org.ofono.Error.NotImplemented: Implementation not provided
user(a)devuan:/my/ofono$
			

Communication with modem is definitely happening:

ofonod[7881]: drivers/qmimodem/network-registration.c:event_notify()
ofonod[7881]: drivers/qmimodem/network-registration.c:event_notify()
signal with 80%(-55 dBm) on 5
ofonod[7881]: src/network.c:ofono_netreg_strength_notify() strength 80
ofonod[7881]: drivers/qmimodem/network-registration.c:event_notify()
ofonod[7881]: drivers/qmimodem/network-registration.c:event_notify()
signal with 100%(-54 dBm) on 5

ofono supports voice calls over qmi, right? Is it possible that
mdm6600 does not support them?

qmicli has some voice options, but not option to do a call, either...

user(a)devuan:/my/ofono$ qmicli -d /dev/cdc-wdm0 --voice-get-config
error: couldn't open the QmiDevice: Cannot open device file '/dev/cdc-wdm0': Permission denied
user(a)devuan:/my/ofono$ sudo qmicli -d /dev/cdc-wdm0 --voice-get-config
[/dev/cdc-wdm0] Successfully retrieved Voice configuration:
Current TTY mode: 'off'Current Preferred Voice SO:
	NAM ID: '0'
	EVRC capability: 'enabled'
	Home Page Voice SO: 'evrc'
	Home Origination Voice SO: 'evrc'
	Roaming Origination Voice SO: 'evrc'
AMR Status:
	GSM: 'enabled'
	WCDMA: 'not-supported, gsm-hr-amr' (0x0005)
Current Voice Privacy Preference: 'enhanced'
user(a)devuan:/my/ofono$ sudo qmicli -d /dev/cdc-wdm0 --voice-get-supported-messages
error: couldn't get supported VOICE messages: QMI protocol error (71): 'InvalidQmiCommand'
user(a)devuan:/my/ofono$ 


Thanks,
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 181 bytes --]