Cinterion MV31-W modem support

[Loic Poulain]

On Tue, 3 Aug 2021 at 17:32, Loic Poulain <loic.poulain at linaro.org> wrote:
>
> On Tue, 3 Aug 2021 at 17:13, ROSTANSKI Tomasz
> <tomasz.rostanski at thalesgroup.com> wrote:
> >
> > Hi,
> >
> > The patch works. No QMI is being probed.
> > I think I spotted the reason why not only cinterion module is considered appropriate for the modem.
> > After executing the:
> > mmcli --report-kernel-event="action=add,subsystem=wwan,name=wwan0mbim0"
> >
> > we can see:
> >
> > [2940]: <debug> [base-manager] kernel event reported:
> > [2940]: <debug> [base-manager]   action:    add
> > [2940]: <debug> [base-manager]   subsystem: wwan
> > [2940]: <debug> [base-manager]   name:      wwan0mbim0
> > [2940]: <debug> [base-manager]   uid:       n/a
> > [2940]: <debug> [wwan0mbim0] preloading contents and properties...
> > [2940]: <debug> [wwan0mbim0] sysfs path: /sys/devices/platform/soc at 0/884057000000.pci/pci0002:57/0002:57:00.0/0002:58:00.0/mhi0/wwan/wwan0/wwan0mbim0
> > [2940]: <debug> [wwan0mbim0] port contents loaded:
> > [2940]: <debug> [wwan0mbim0]   bus: pci
> > [2940]: <debug> [wwan0mbim0]   device: /sys/devices/platform/soc at 0/884057000000.pci/pci0002:57/0002:57:00.0
> > [2940]: <debug> [wwan0mbim0]   subsystems: wwan, mhi, pci
> > [2940]: <debug> [wwan0mbim0]   drivers: mhi-pci-generic, pcieport
> > [2940]: <debug> [wwan0mbim0]   vendor: 177d
> > [2940]: <debug> [wwan0mbim0]   product: a22d
> > [2940]: <debug> [wwan0mbim0]   revision: 0002
> > [2940]: <debug> [wwan0mbim0]   ID_VENDOR_ID: 0x177d
> > [2940]: <debug> [wwan0mbim0]   ID_MODEL_ID: 0xa22d
> > [2940]: <debug> [wwan0mbim0]   ID_REVISION: 0x0002
> >
> > The detected vendor id is 0x177d:a22d is a PCI bridge:
> > 0002:57:00.0 PCI bridge: Cavium, Inc. Device a22d (rev 02) (prog-if 00 [Normal decode])
> > While the modem is:
> > 0002:58:00.0 Wireless controller [0d40]: Thomson-CSF/TTM Device 00b3
> > (Thomson-CSF is former name of Thales)
>
> I observe quite the same on my side, looks like the PCI IDs are
> retrieved from the parent instead of the device. It does not happen
> with udev mechanism though. Should be fixed in kernel-generic I
> assume...

Here is a patch:
https://gitlab.freedesktop.org/mobile-broadband/ModemManager/-/merge_requests/599

Regards,
Loic