Quectel modem ceases working

Quectel modem ceases working

[Timo Ketola]

Hi,

We have successfully used ofono with Quectel EG25-G modem in our iMX6
Arm system. But now things break due to BuildRoot upgrade which upgrades
ofono from 2.2 to 2.18. I have pinpointed the problem to commit
281b49763818, "gobi: ensure Set Data Format changes were applied".
ofonod gives an error:

> Setting Data Format had no effect

list-modems shows:

> [ /gobi_0 ]
>     Online = 0
>     Powered = 0
>     Lockdown = 0
>     Emergency = 0
>     SystemPath = /sys/devices/soc0/soc/2100000.aips-bus/2184200.usb/ci_hdrc.1/usb2/2-1
>     Interfaces = 
>     Features = 
>     Type = hardware
>     Capabilities = dbus.Array([dbus.String('lte')], signature=dbus.Signature('s'), variant_level=1)

If I ignore the error by hacking plugins/gobi.c:

>  		ofono_error("Setting Data Format had no effect");
> +		goto done;
>  	}

list-modems shows expected info again:

> [ /quectelqmi_0 ]
>     Online = 0
>     Powered = 1
>     Lockdown = 0
>     Emergency = 0
>     Manufacturer = QUALCOMM INCORPORATED
>     Model = QUECTEL Mobile Broadband Module
>     Revision = EG25GGBR07A08M2G
> ...

I have no idea if there is other harmful side effects.

How this should be fixed properly? Any ideas?

-- 
Freedom To Create Together.

+------------------------------------------------+
| Timo Ketola            | Exertus Oy            |
| Specialist,            | Kampusranta 9C        |
|       FW & Electronics | 60320 Seinäjoki       |
| +358 40 5311 244       | Finland               |
| timo.ketola@exertus.fi | http://www.exertus.fi |
+------------------------+-----------------------+
| /"\                                            |
| \ / ASCII RIBBON CAMPAIGN AGAINST HTML MAIL    |
|  X  http://www.asciiribbon.org/                |
| / \                                            |
+------------------------------------------------+

Re: Quectel modem ceases working

[Denis Kenzior]

Hi Timo,

 > We have successfully used ofono with Quectel EG25-G modem in our iMX6
> Arm system. But now things break due to BuildRoot upgrade which upgrades
> ofono from 2.2 to 2.18. I have pinpointed the problem to commit
> 281b49763818, "gobi: ensure Set Data Format changes were applied".
> ofonod gives an error:
> 

Hard to say.  There has been quite a bit of changes between those versions and I 
don't have access to that particular hardware.

Can you provide a full debug log?  ofonod -n -d '*gobi*'

What kernel are you running?  Is the modem a USB device or directly connected to 
the SoC?  Assuming it is USB and the device gets detected at /dev/cdc-wdm0, what 
does the output of:
qmicli -d /dev/cdc-wdm0 --wda-get-data-format
look like?

Regards,
-Denis

Re: Quectel modem ceases working

[Timo Ketola]

Hi Denis,

Thanks for looking at this!

My further testing unsurprisingly shows that my hack does not really
work. list-modems looks better and the IF gets IP address but actual
traffic does not work.

On 9.9.2025 18.23, Denis Kenzior wrote:
> Can you provide a full debug log?  ofonod -n -d '*gobi*'

Sure. Working 2.12:

> # /usr/sbin/ofonod-2.12 -nd '*gobi*'
> ofonod-2.12[588]: oFono version 2.12
> src/main.c:main() Using configuration directory: /etc/ofono
> ofonod-2.12[588]: plugins/gobi.c:gobi_probe() 0x19bd18
> ofonod-2.12[588]: plugins/gobi.c:gobi_probe() net: wwan0[qmi_wwan](5) usb
> ofonod-2.12[588]: plugins/gobi.c:gobi_enable() 0x19bd18
> ofonod-2.12[588]: plugins/gobi.c:init_powered_down_cb() error: 0
> ofonod-2.12[588]: plugins/gobi.c:discover_cb() 
> ofonod-2.12[588]: plugins/gobi.c:request_service_cb() 
> ofonod-2.12[588]: plugins/gobi.c:request_service_cb() Requesting: 3
> ofonod-2.12[588]: plugins/gobi.c:request_service_cb() 
> ofonod-2.12[588]: plugins/gobi.c:request_service_cb() Requesting: 1
> ofonod-2.12[588]: plugins/gobi.c:request_service_cb() 
> ofonod-2.12[588]: plugins/gobi.c:request_service_cb() Requesting: 5
> ofonod-2.12[588]: plugins/gobi.c:request_service_cb() 
> ofonod-2.12[588]: plugins/gobi.c:request_service_cb() Requesting: 9
> ofonod-2.12[588]: plugins/gobi.c:request_service_cb() 
> ofonod-2.12[588]: plugins/gobi.c:request_service_cb() Requesting: 11
> ofonod-2.12[588]: plugins/gobi.c:request_service_cb() 
> ofonod-2.12[588]: plugins/gobi.c:request_service_cb() Requesting: 1
> ofonod-2.12[588]: plugins/gobi.c:request_service_cb() 
> ofonod-2.12[588]: plugins/gobi.c:request_service_cb() Requesting: 1
> ofonod-2.12[588]: plugins/gobi.c:request_service_cb() 
> ofonod-2.12[588]: plugins/gobi.c:request_service_cb() All services requested, pr
> oceeding to create WDA
> ofonod-2.12[588]: plugins/gobi.c:create_wda_cb() 
> ofonod-2.12[588]: plugins/gobi.c:get_data_format_cb() 
> ofonod-2.12[588]: plugins/gobi.c:setup_qmi_wwan() raw_ip: Y, want: Y
> ofonod-2.12[588]: plugins/gobi.c:get_caps_cb() 
> ofonod-2.12[588]: plugins/gobi.c:get_caps_cb() service capabilities 4
> ofonod-2.12[588]: plugins/gobi.c:get_caps_cb() sim supported 2
> ofonod-2.12[588]: plugins/gobi.c:get_caps_cb() radio = 4
> ofonod-2.12[588]: plugins/gobi.c:get_caps_cb() radio = 5
> ofonod-2.12[588]: plugins/gobi.c:get_caps_cb() radio = 8
> ofonod-2.12[588]: plugins/gobi.c:get_oper_mode_cb() 
> ofonod-2.12[588]: plugins/gobi.c:gobi_pre_sim() 0x19bd18
> ofonod-2.12[588]: Requested file structure differs from SIM: 6fb7
> ofonod-2.12[588]: plugins/gobi.c:gobi_post_sim() 0x19bd18
> ofonod-2.12[588]: plugins/gobi.c:gobi_set_online() 0x19bd18 online using_mux: no
> ofonod-2.12[588]: plugins/gobi.c:powered_up_cb() error: 0
> ofonod-2.12[588]: plugins/gobi.c:set_online_cb() 
> ofonod-2.12[588]: plugins/gobi.c:gobi_post_online() 0x19bd18

Hacked 2.18:

> # /usr/sbin/ofonod-2.18 -nd '*gobi*'
> ofonod-2.18[601]: oFono version 2.18
> src/main.c:main() Using configuration directory: /etc/ofono
> ofonod-2.18[601]: plugins/gobi.c:gobi_probe() 0x19ff88
> ofonod-2.18[601]: plugins/gobi.c:gobi_probe() net: wwan0[qmi_wwan](5) usb
> ofonod-2.18[601]: plugins/gobi.c:gobi_enable() 0x19ff88
> ofonod-2.18[601]: plugins/gobi.c:init_powered_down_cb() error: 0
> ofonod-2.18[601]: plugins/gobi.c:init_powered_down_cb() Setting QMI_WWAN to 802.
> 3 mode
> ofonod-2.18[601]: plugins/gobi.c:discover_cb() 
> ofonod-2.18[601]: plugins/gobi.c:create_wda_cb() 
> ofonod-2.18[601]: plugins/gobi.c:set_data_format_cb() 
> ofonod-2.18[601]: Setting Data Format had no effect
> ofonod-2.18[601]: plugins/gobi.c:set_data_format_cb() Set Data Format succeeded,
>  try to set MTU...
> ofonod-2.18[601]: plugins/gobi.c:enable_set_mtu_cb() error: 0
> ofonod-2.18[601]: plugins/gobi.c:start_service_requests() 
> ofonod-2.18[601]: plugins/gobi.c:request_service_cb() 
> ofonod-2.18[601]: plugins/gobi.c:request_service_cb() Requesting: 3
> ofonod-2.18[601]: plugins/gobi.c:request_service_cb() 
> ofonod-2.18[601]: plugins/gobi.c:request_service_cb() Requesting: 1
> ofonod-2.18[601]: plugins/gobi.c:request_service_cb() 
> ofonod-2.18[601]: plugins/gobi.c:request_service_cb() Requesting: 5
> ofonod-2.18[601]: plugins/gobi.c:request_service_cb() 
> ofonod-2.18[601]: plugins/gobi.c:request_service_cb() Requesting: 9
> ofonod-2.18[601]: plugins/gobi.c:request_service_cb() 
> ofonod-2.18[601]: plugins/gobi.c:request_service_cb() Requesting: 11
> ofonod-2.18[601]: plugins/gobi.c:request_service_cb() 
> ofonod-2.18[601]: plugins/gobi.c:request_service_cb() Requesting: 1
> ofonod-2.18[601]: plugins/gobi.c:request_service_cb() 
> ofonod-2.18[601]: plugins/gobi.c:request_service_cb() Requesting: 1
> ofonod-2.18[601]: plugins/gobi.c:request_service_cb() 
> ofonod-2.18[601]: plugins/gobi.c:request_service_cb() All services requested, qu
> ery DMS Capabilities
> ofonod-2.18[601]: plugins/gobi.c:get_caps_cb() 
> ofonod-2.18[601]: plugins/gobi.c:get_caps_cb() service capabilities 4
> ofonod-2.18[601]: plugins/gobi.c:get_caps_cb() sim supported 2
> ofonod-2.18[601]: plugins/gobi.c:get_caps_cb() radio = 4
> ofonod-2.18[601]: plugins/gobi.c:get_caps_cb() radio = 5
> ofonod-2.18[601]: plugins/gobi.c:get_caps_cb() radio = 8
> ofonod-2.18[601]: plugins/gobi.c:get_oper_mode_cb() 
> ofonod-2.18[601]: plugins/gobi.c:gobi_pre_sim() 0x19ff88
> ofonod-2.18[601]: Requested file structure differs from SIM: 6fb7
> ofonod-2.18[601]: plugins/gobi.c:gobi_post_sim() 0x19ff88
> ofonod-2.18[601]: plugins/gobi.c:gobi_set_online() 0x19ff88 online using_mux: no
> ofonod-2.18[601]: plugins/gobi.c:powered_up_cb() error: 0
> ofonod-2.18[601]: plugins/gobi.c:set_online_cb() 
> ofonod-2.18[601]: plugins/gobi.c:gobi_post_online() 0x19ff88

Pristine 2.18:

> # ./ofonod-2.18 -nd '*gobi*'
> ofonod-2.18[622]: oFono version 2.18
> src/main.c:main() Using configuration directory: /etc/ofono
> ofonod-2.18[622]: plugins/gobi.c:gobi_probe() 0x19ff88
> ofonod-2.18[622]: plugins/gobi.c:gobi_probe() net: wwan0[qmi_wwan](5) usb
> ofonod-2.18[622]: plugins/gobi.c:gobi_enable() 0x19ff88
> ofonod-2.18[622]: plugins/gobi.c:init_powered_down_cb() error: 0
> ofonod-2.18[622]: plugins/gobi.c:init_powered_down_cb() Setting QMI_WWAN to 802.
> 3 mode
> ofonod-2.18[622]: plugins/gobi.c:discover_cb() 
> ofonod-2.18[622]: plugins/gobi.c:create_wda_cb() 
> ofonod-2.18[622]: plugins/gobi.c:set_data_format_cb() 
> ofonod-2.18[622]: Setting Data Format had no effect
> ofonod-2.18[622]: plugins/gobi.c:shutdown_device() 0x19ff88
> ofonod-2.18[622]: plugins/gobi.c:shutdown_cb() 

> What kernel are you running?

5.4.78 with RT patch.

> Is the modem a USB device or directly connected to the SoC?

Quectel EG25-G MiniPCIe module with USB interface.

> Assuming it is USB and the device gets detected at /dev/cdc-wdm0,
> what does the output of: qmicli -d /dev/cdc-wdm0 --wda-get-data-
> format look like?

It fails if ofono is running but otherwise:

> # qmicli -d /dev/cdc-wdm0 --wda-get-data-format
> [/dev/cdc-wdm0] Successfully got data format
>                    QoS flow header: no
>                Link layer protocol: 'raw-ip'
>   Uplink data aggregation protocol: 'disabled'
> Downlink data aggregation protocol: 'disabled'
>                      NDP signature: '0'
> Downlink data aggregation max datagrams: '0'
> Downlink data aggregation max size: '0'

--
Timo

Re: Quectel modem ceases working

[Denis Kenzior]

Hi Timo,

On 9/10/25 5:59 AM, Timo Ketola wrote:
> Hi Denis,
> 
> Thanks for looking at this!
> 
> My further testing unsurprisingly shows that my hack does not really
> work. list-modems looks better and the IF gets IP address but actual
> traffic does not work.

Yeah, looks like the raw_ip setting isn't being reset back to what is expected.

> 
>> What kernel are you running?
> 
> 5.4.78 with RT patch.

Ugh.  I really have no desire to support kernels pre-5.15 when the passthrough 
capability was introduced.  I also have no hardware that is not 802.3 mode 
capable.  You can try the following patch though and see if things work again.

Regards,
-Denis

diff --git a/plugins/gobi.c b/plugins/gobi.c
index 52ead377de7f..1b77e5e32ed6 100644
--- a/plugins/gobi.c
+++ b/plugins/gobi.c
@@ -665,10 +665,21 @@ static void set_data_format_cb(struct qmi_result *result, 
void *user_data)
  	DBG("");

  	if (!qmi_result_set_error(result, NULL)) {
+		struct qmi_wda_data_format actual;
+
  		if (!compare_data_format(data, result))
  			goto done;

  		ofono_error("Setting Data Format had no effect");
+
+		if (!qmi_wda_parse_data_format(result, &actual) &&
+				data->data_format == WDA_DATA_FORMAT_802_3 &&
+				actual.ll_protocol ==
+				QMI_WDA_DATA_LINK_PROTOCOL_RAW_IP) {
+			ofono_warn("Modem seems to only work in raw-ip mode");
+			data->data_format = WDA_DATA_FORMAT_RAW_IP;
+			goto done;
+		}
  	}

  	if (data->data_format == WDA_DATA_FORMAT_802_3)

Re: Quectel modem ceases working

[adam]

Hi Dennis

Is this the same/similar issue I reported in Dec 24, where the pinephone modem 
stopped working between 2.12 and 2.13?

Thanks

Adam

On Wednesday, 10 September 2025 16:31:05 British Summer Time Denis Kenzior 
wrote:
> Hi Timo,
> 
> On 9/10/25 5:59 AM, Timo Ketola wrote:
> > Hi Denis,
> > 
> > Thanks for looking at this!
> > 
> > My further testing unsurprisingly shows that my hack does not really
> > work. list-modems looks better and the IF gets IP address but actual
> > traffic does not work.
> 
> Yeah, looks like the raw_ip setting isn't being reset back to what is
> expected.
> >> What kernel are you running?
> > 
> > 5.4.78 with RT patch.
> 
> Ugh.  I really have no desire to support kernels pre-5.15 when the
> passthrough capability was introduced.  I also have no hardware that is not
> 802.3 mode capable.  You can try the following patch though and see if
> things work again.
> 
> Regards,
> -Denis
> 
> diff --git a/plugins/gobi.c b/plugins/gobi.c
> index 52ead377de7f..1b77e5e32ed6 100644
> --- a/plugins/gobi.c
> +++ b/plugins/gobi.c
> @@ -665,10 +665,21 @@ static void set_data_format_cb(struct qmi_result
> *result, void *user_data)
>   	DBG("");
> 
>   	if (!qmi_result_set_error(result, NULL)) {
> +		struct qmi_wda_data_format actual;
> +
>   		if (!compare_data_format(data, result))
>   			goto done;
> 
>   		ofono_error("Setting Data Format had no effect");
> +
> +		if (!qmi_wda_parse_data_format(result, &actual) &&
> +				data->data_format == 
WDA_DATA_FORMAT_802_3 &&
> +				actual.ll_protocol ==
> +				QMI_WDA_DATA_LINK_PROTOCOL_RAW_IP) 
{
> +			ofono_warn("Modem seems to only work in raw-
ip mode");
> +			data->data_format = WDA_DATA_FORMAT_RAW_IP;
> +			goto done;
> +		}
>   	}
> 
>   	if (data->data_format == WDA_DATA_FORMAT_802_3)

Re: Quectel modem ceases working

[Denis Kenzior]

Hi Adam,

 >
> Is this the same/similar issue I reported in Dec 24, where the pinephone modem
> stopped working between 2.12 and 2.13?

Not the same but likely related.  There's quite a bit of variance in how 
Qualcomm derived firmware behaves in this area.

Regards,
-Denis

Re: Quectel modem ceases working

[Adam Pigg]

[-- Attachment #1: Type: text/plain, Size: 446 bytes --]

On Wednesday, 10 September 2025 20:34:08 British Summer Time Denis Kenzior 
wrote:
> Hi Adam,
> 
> > Is this the same/similar issue I reported in Dec 24, where the pinephone
> > modem stopped working between 2.12 and 2.13?
> 
> Not the same but likely related.  There's quite a bit of variance in how
> Qualcomm derived firmware behaves in this area.
> 
In any case, it would be great to get both issues fixed on the EG25 :)
> Regards,
> -Denis


[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 525 bytes --]

Re: Quectel modem ceases working

[Timo Ketola]

Hi Denis,

On 10.9.2025 18.31, Denis Kenzior wrote:
> Ugh.  I really have no desire to support kernels pre-5.15 when the 
> passthrough capability was introduced...

Are you saying that if we had 5.15 kernel or later, latest ofono should
have been working out of the box?

> You can try the following patch though and see if things work again.

Thanks a lot! I'll try it tomorrow and report back.

--

Timo

Re: Quectel modem ceases working

[Denis Kenzior]

Hi Timo,

On 9/10/25 12:49 PM, Timo Ketola wrote:
> Hi Denis,
> 
> On 10.9.2025 18.31, Denis Kenzior wrote:
>> Ugh.  I really have no desire to support kernels pre-5.15 when the
>> passthrough capability was introduced...
> 
> Are you saying that if we had 5.15 kernel or later, latest ofono should
> have been working out of the box?

I'm not saying that.  It would likely still be broken since the firmware doesn't 
support 802.3 mode.

Long term I want to drop support for the strange single 802.3 / raw IP interface 
setup that is being used here and only support multi-context setups.  qmiwwan's 
behavior in this area was quite different prior to 5.15 and I have no desire to 
support that legacy.

Regards,
-Denis

Re: Quectel modem ceases working

[Timo Ketola]

Hi Denis,

On 10.9.2025 18.31, Denis Kenzior wrote:
> You can try the following patch though and see if things work again.

Yes, it works! Thanks a lot!

Regarding longer term upgrade path, I'm completely on FM-broadcast-band
(finnish idiom). Is it Qualcomm/Quectel FW or qmiwwan Linux driver that
is the ultimate culprit here? Or is this modem completely deteriorating
out of support?

--

Timo

Re: Quectel modem ceases working

[Denis Kenzior]

Hi Timo,

On 9/11/25 7:45 AM, Timo Ketola wrote:
> Hi Denis,
> 
> On 10.9.2025 18.31, Denis Kenzior wrote:
>> You can try the following patch though and see if things work again.
> 
> Yes, it works! Thanks a lot!

Nice, thanks for testing.  I cleaned up the commit description and pushed it out.

> 
> Regarding longer term upgrade path, I'm completely on FM-broadcast-band
> (finnish idiom). Is it Qualcomm/Quectel FW or qmiwwan Linux driver that
> is the ultimate culprit here? Or is this modem completely deteriorating
> out of support?

My comment is mostly about the kernel version.  It is impractical for me to test 
changes on old kernels or to support legacy driver behavior.

Regards,
-Denis

Re: Quectel modem ceases working

[adam]

On Thursday, 11 September 2025 15:43:55 British Summer Time Denis Kenzior 
wrote:
> Hi Timo,
> 
> On 9/11/25 7:45 AM, Timo Ketola wrote:
> > Hi Denis,
> > 
> > On 10.9.2025 18.31, Denis Kenzior wrote:
> >> You can try the following patch though and see if things work again.
> > 
> > Yes, it works! Thanks a lot!
> 
> Nice, thanks for testing.  I cleaned up the commit description and pushed it
> out.

Is this likely to fix the issue I have Dennis?  I will of course try testing it 
when I can.

> > Regarding longer term upgrade path, I'm completely on FM-broadcast-band
> > (finnish idiom). Is it Qualcomm/Quectel FW or qmiwwan Linux driver that
> > is the ultimate culprit here? Or is this modem completely deteriorating
> > out of support?
> 
> My comment is mostly about the kernel version.  It is impractical for me to
> test changes on old kernels or to support legacy driver behavior.
> 
> Regards,
> -Denis