For a lot of people having the 1 second timeout results in not finding
the sim at all. Let's rather wait up to 4 seconds for the sim to appear,
it was reported that 3 is enough in a report where 1 wasn't.
Related: issue 2072
[ci:skip-build]: already built successfully in CI
Like alcatel-idol3, these devices have a tunning partition. However, the
modem calls it modem_tng instead of modem_tunning. This patch adds the
alternative name for the partition.
An upstream pull request has already been submitted (see patch).
1. Upgrades the library to version 6.0.0004
2. Adds runtime overrides for detected device and board,
to facilitate running on devices where autodetect fails.
Ideally tqftpserv should be started before rmtfs. This is because rmtfs
is currently responsible for starting the modem, and it's best for
performance if all the services needed by the modem are already up.
Also change "want" -> "use" for consistency with rmtfs. The difference
between "want" and "use" is that "want" just triggers starting of the
other service, while "use" also ensures that the other service starts
successfully.
Also include z3ntu's patch fixing a buffer overread issue which sadly
has not been merged for several months now.
rmtfs should be ideally started and stopped before NetworkManager
because otherwise the modem network interface might be removed before
NetworkManager can set it down. This has caused kernel issues at some
point (although it seems to be fine again at the moment).
Recent libqmi update in Alpine [1] splits off binary utilities into a
subpackage. This breaks the selection script which depends on qmicli.
Add the new qmi-utils as a dependency to make sure the script can
continue to work.
[1] - a1e61165f4
Unfortunately it looks like some modem firmwares (or SIM cards?) need
more time until the SIM card shows up in the modem. This means that the
SIM card might not get configured, making it look like it was not
detected at all. On the other hand, having the previous long wait loop
affects _all devices_ when there is really no SIM inserted.
The proper solution would be to make it possible to configure SIM cards
once they show up (probably in ModemManager). Until then, try to make it
possible to workaround this by introducing a configurable delay to wait
for the SIM card. It is set to minimal "1 seconds" by default but could
be increased by the user if needed to give the modem enough time to detect
the SIM card.
This reverts commit 683b543a3b ("msm-modem-uim-selection: Don't wait
for SIM (MR 2470)" but with a new configuration mechanism and a default
of 1 second only.
[ci:skip-build] already built successfully in CI
Unfortunately, msm-modem-uim-selection causes boot delays on APQ*
SoCs (without most modem functionality) at the moment. The modem will
never show up so it keeps waiting and waiting.
To allow using the same rootfs on devices that have MSM* and APQ*
variants (e.g. Samsung MSM8916 tablets), add a quick check at the
beginning of the service to cancel early when running on an APQ* SoC.
This is easy to check because the SoC is actually exposed in sysfs.
The diag-router service is required to be running to stop ATH10k WLAN
firmware from crashing constantly on MSM8998 SoC devices running
mainline currently.
This is no longer used by anyone and also should not be used anymore.
The new WWAN RPMSG CTRL for the WWAN subsystem is now the preferred
way to talk to the modem on MSM8916/MSM8939/MSM8974 and other similar SoCs.
[ci:skip-build]: already built successfully in CI
The delay to wait for the SIM card to show up appears to be unneeded on
our modems, remove it so that we don't delay boot by 10 seconds if your
device doesn't have a SIM card installed.
[ci:skip-build] already built successfully in CI
uim-selection has "after *" in it's dependencies to make sure the
theoretical up to ~1min delay can't be observed. Unfortunaely this
causes incorrect dependencies ordering if one of other services depends
on ofono or modemmanager. Drop "after *" so modem can always be started
properly.
On some devices, primarily with two sim slots, modem doesn't choose
the sim automatically. Add an init script that would detect in which
slot the sim is present and configure the modem to use it.
At the moment, oFono/MM are usually stopped after rmtfs on shutdown.
This does not make sense, since the modem managed by oFono/MM will
disappear once rmtfs is stopped. Add a "before ofono modemmanager"
to make sure that OpenRC will stop oFono/MM before rmtfs on shutdown.
This may or may not fix some occasional hangs during shutdown,
but it definitely makes it possible for oFono/MM to shutdown cleanly.
When I upgraded tinyalsa to the new 2.0.0 major release (with breaking
changes) I failed to test q6voiced properly, even though I explicitly
had to rebuild it against the new tinyalsa version (see MR 1941).
In tinyalsa 2.0.0, calling pcm_open() does not automatically call
pcm_prepare() anymore, which is responsible for actually starting any
audio streams. This means that the old version of q6voiced is almost
entirely broken - only by pure luck it's sometimes possible to get
voice call audio working if you play other audio simultaenously while
a voice call is active. This does not work for the microphone though.
The new q6voiced version now calls pcm_prepare() explicitly to fix this.
On some downstream kernels it seems like we need to explicitly keep
/dev/subsys_modem open (without writing anything), otherwise the modem
will be stopped (or never started). Weird.
The -s switch to automatically start/stop the modem remoteproc only
works on mainline, on downstream it fails with "Failed to get rprocfd".
Let's abuse /usr/lib/preload/libqipcrtr4msmipc.so to check if we are
(probably) running on a downstream installation, and omit the -s
argument in this case. The modem remoteproc needs to be started
differently on downstream.
qrtr-ns is not needed (or working) on downstream, also it is no longer
needed for mainline starting with Linux 5.9. Convert the "need" dependency
to a "use" so qrtr-ns is only started if it is really needed.
In most cases, the qrtr-ns is not needed anymore now, but we still
need the libqrtr library. Separate that out into a separate package
so it can be installed independently.
So far we have only created /dev/modem based on the DATA5_CNTL
channel available from the modem. The DATAX_CNTL channels allow
sending QMI messages to the modem. This is sufficient for oFono
to work.
Unfortunately, ModemManager currently does not support starting
calls through the QMI interface. Instead, it uses serial AT commands
that can be alternatively used on all USB-based Qualcomm modems.
It turns out that we can also send serial AT commands through
the RPMSG interface: the DATAX channels (without _CNTL) all respond
to serial AT commands. We set it up at /dev/modem-at, configure
ModemManager accordingly and then we are able to start calls. Yay!
Newer Qualcomm SoCs (e.g. SDM845. MSM8953) seem to use a different
method to talk to the modem. There are no longer separate rpmsg/SMD
channels for QMI messags, instead the modem is also available through QRTR.
On these newer SoCs, installing msm-modem-mainline is pointless,
because the created /dev/modem device will not actually allow
communicating with the modem. However, you still need the part from
the base package (msm-modem), which installs+enables rmtfs.
To avoid confusion, rename the msm-modem-mainline subpackage to
msm-modem-rpmsg. Install only "msm-modem" on SDM845 because that
likely does not allow communication with the modem through RPMSG
anyway.
It is possible for msm-modem to get installed before
rmtfs-openrc, which causes the following error in
post-install:
* rc-update: service `rmtfs' does not exist
ERROR: msm-modem-1-r0.post-install: script exited with error 1
Add a dependency to ensure rmtfs-openrc gets installed first
linux-postmarketos-qcom-msm8916 now supports voice call audio.
However, at the moment the audio stream needs to be explicitly started
by opening some PCM audio devices when the call is started
(even though no audio data is written to the PCM device).
Eventually this should go away by implementing codec2codec support
for q6voice in the kernel. In that case, the stream would be started
once the audio routes are set up using ALSA UCM. However, this is a bit
complicated actually so for now it's nice to have a workaround:
q6voiced listens on dbus for signals from oFono, and opens/closes the
PCM device when a call is initiated/ended in oFono. This essentially
makes voice call audio work out of the box (provided that the audio
routing, e.g. Earpiece and a microphone is set up appropriately).
Co-Authored-By: Nikita Travkin <nikitos.tr@gmail.com>
Combine modem-qcom-msm-{mainline,downstream}-common to a single APKBUILD
with mainline and downstream subpackages.
Enable rmtfs service here instead of directly in the package.
[ci:skip-vercheck]: for qrtr version fix
- use supervise-daemon to restart when necessary
- run normally, without wrapping with sh and logger
- add "need qrtr-ns" to make sure it gets started first
- stop enabling service directly in rmtfs package
(moved to msm-modem)
- Fix version
- Rename init script to qrtr-ns ("nameservice") to clarify which
service is actually started
- Use supervise-daemon
- Configure and start daemon in one command (no need for qrtr-cfg)
- Remove "before qcom_rmtfs": it should depend on qrtr-ns,
not the other way around
- Do not activate qrtr daemon by default - this should happen automatically
if other service files have "need qrtr-ns"
