imsd/README.md
Jorijn van der Graaf 142adf057c
engine: register fresh when the refresh is refused or nothing answers
After bring-up the daemon recovered a dead protected flow by
reconnecting through the existing SA and refreshing, on a 5 s → 5 min
backoff, forever. That was the whole recovery: a refresh the network
refused, or a P-CSCF that stopped answering, kept it in that loop on the
same Call-ID and SA until someone restarted the service. O2 UK,
2026-09-24: the 30-minute refresh answered 503, 21 refusals over 70 min;
after a manual restart the new registration's flow died 8 s in and no
connect over its SA succeeded for 2 h 39 min; a second restart
registered within two seconds (13:27:55 → 13:27:57 in the journal; ims
ledger F16).

Do in place the rest of what a restart does: when the refresh keeps
being refused, or reconnect cycles keep ending without an answer unless
the modem reports no packet service, register fresh. The fresh
registration is the expensive step, so a failed one waits before the
next: on a schedule (21 min, growing to 60) after a refusal or a
throttle, on a backoff after silence, and a Retry-After from the network
is honoured up to an hour. What the network set — a Retry-After, the
schedule after a refused or throttled attempt — is kept across a restart
in /run/imsd.holds, for the address it was set on. A refused initial
registration at bring-up — a final from 300 up to the unprotected
REGISTER (its 401 is the challenge), any final but 200 to the protected
one, a second 401 (rejected credentials) included — now waits on the
same schedule instead of exiting for a systemd restart every two minutes.

The policy is a pure reducer in Imsd:Engine (RegRecovery), shaped like
the call machine: the daemon supplies the clock and the network, the
reducer holds the deadlines and names the next step, and tests/Engine
pin it. tests/Engine now exits 0 or 1 instead of its failure count: with
more than 255 checks a count of 256 would exit 0, and crafter-build
reads 77 as a skipped suite. The bring-up's deferral is logged in the
policy's form: `bring-up deferred: <why> — attempt N, next fresh
registration in M min` becomes `bring-up deferred: <why>; next: fresh
registration in M min`.

A bring-up writes the verdict that the old registration is gone as well:
a start that goes on to register fresh — its resume failed, or there was
nothing to resume over — records it before the attempt, and when the
attempt then fails without an answer the daemon exits with the file in
place. The next start on the same address and boot does not resume,
where 0.3.9 retried the resume: it registers fresh at once. On KPN, with
the old registration in fact still live — a start during an outage,
staged on the dev phone — that attempt was throttled, and the 21-min
schedule began. A bring-up does not retry its resume through an outage
the way recovery reconnects.

Tested on the dev phone (KPN) in builds of this series; not every path
has run there.
2026-10-06 14:01:33 +02:00

21 KiB
Raw Blame History

imsd

Userspace IMS/VoLTE daemon for mainline Linux phones.

Modem firmware on typical mainline-Linux phones will bring up the IMS PDN but can't be used for actual call audio. imsd implements the whole IMS client in userspace instead: SIP registration with USIM AKA authentication over kernel IPsec (ESP), call signalling, and the AMR-WB RTP media plane wired into PipeWire so a stock mobile dialer — Plasma Dialer, or GNOME Calls under Phosh and GNOME Mobile — can place and receive real VoLTE calls on a commercial network.

Status

Call-capable, both directions. Outgoing VoLTE calls place, answer, carry full-duplex AMR-WB audio, and tear down; incoming calls ring, answer via Accept, and carry media the same way — driven end-to-end from a stock dialer over the frozen D-Bus ABI. Current state:

The engine core (imsd-core) is a GLib-free, pure-std static library whose every decision — SIP framing, SDP selection, the AKAv1-MD5 digest, the ip xfrm command sequence, and the whole call FSM — is unit-tested on a dev box with no bus, modem, or phone; several suites pin output byte-for-byte against recorded network traffic. GLib/GDBus, sockets, and subprocess control live only in the imsd daemon shell.

Architecture

C++26 modules, built with Crafter Build. Five build products:

  • imsd-core (static library) — the engine, pure import std C++ with no GLib and no I/O side effects. Everything here is unit-testable on any dev box without a modem, a bus, or a phone; several suites pin output byte-for-byte against recorded network traffic.
  • imsd (executable) — the daemon shell: GLib main loop, GDBus service, sockets, subprocess control (qmicli/mmcli/ip), and the registration sequence + run loop that pumps the core. All GLib and I/O live here.
  • imsd-media (executable) — the RTP/AMR-WB data plane, spawned once per call. Pure import std + POSIX, no GLib; links -ldl and dlopen's the codecs. A separate process for crash isolation and the far-end-hangup exit-code contract.
  • imsd-dialerd (executable) — the Plasma Mobile integration: a session daemon owning org.kde.telephony.{CallUtils,DeviceUtils,UssdUtils}, bridging them to imsd on the system bus. GDBus translation only, no core. It is a separate process because those are per-user session-bus names that the root system daemon cannot own; built with -- --product=dialerd, autostarted from an .desktop file in place of plasma-dialer's modem-daemon (whose autostart must be disabled).
  • imsd-ofonod (executable) — the GNOME Calls integration, for Phosh and GNOME Mobile: a system daemon owning org.ofono and presenting imsd as an oFono modem, which is the seam GNOME Calls' bundled ofono provider plugin drives. GDBus translation only, no core; built with -- --product=ofonod. See GNOME Calls (Phosh, GNOME Mobile).

Desktop integration

imsd itself is desktop-agnostic: everything a dialer needs is on the system bus as net.catcrafts.IMS1. Each desktop gets a small translation daemon speaking whatever backend protocol its dialer already consumes, so the dialer, the shell and the call-history/contacts stack all stay unmodified.

Desktop Daemon Owns Consumed by
Plasma Mobile imsd-dialerd org.kde.telephony.* (session bus) plasma-dialer + kde-telephony-daemon, replacing modem-daemon
Phosh / GNOME Mobile imsd-ofonod org.ofono (system bus) GNOME Calls' ofono provider plugin

Phosh's own call UI needs nothing extra: it consumes org.gnome.Calls, which GNOME Calls exports once it has a working origin.

Terminating-network dependency (incoming)

A UE only receives a VoLTE call as a SIP INVITE if the network's Terminating Access Domain Selection (T-ADS) routes it to the PS/IMS domain rather than paging the modem over CS. That requires the IMS registration to advertise voice capability — the MMTEL ICSI (+g.3gpp.icsi-ref=...mmtel) and +sip.instance media-feature tags in the REGISTER Contact, which Imsd:Messages emits. T-ADS may still select CS for reasons outside the client's control (operator policy, the modem's own radio-capability signalling, a competing registration). When it does, no INVITE reaches imsd; that is a network/modem-integration matter, not a daemon defect.

Supported hardware & carriers

imsd talks to the modem through ModemManager plus qmicli over QRTR (USIM AKA via a UIM logical channel), so in practice it currently needs a Qualcomm modem on a mainline kernel with working WWAN/MM integration, kernel ESP (ip xfrm), and PipeWire for audio.

Everything below has been verified by community members to be fully workking:

Device OS Carrier
The Fairphone (Gen. 6) postmarketOS, Linux 7.1.2 KPN NL
The Fairphone (Gen. 6) postmarketOS, Linux 7.1.2 Telekom Deutschland GER
The Fairphone (Gen. 6) postmarketOS, Linux 7.1.2 Phonero
The Fairphone (Gen. 6) postmarketOS, Linux 7.1.2 Telia Norge

Build

Crafter Build + gio-2.0 headers (glib2 on Arch, glib-dev on Alpine).

crafter-build                       # bin/imsd-<target>-<march>/imsd (the daemon)
crafter-build -- --product=media    # bin/imsd-media-.../imsd-media (the media leg)
crafter-build -- --product=dialerd  # bin/imsd-dialerd-.../imsd-dialerd (Plasma backend)
crafter-build -- --product=ofonod   # bin/imsd-ofonod-.../imsd-ofonod (GNOME Calls backend)
crafter-build test                  # unit tests (Util, Aka, Ipsec, Messages, Engine, Sip, Sdp)

Building with make (distro packaging)

For environments without crafter-build, the Makefile mirrors the same clang module pipeline: GNU make, clang++ with libc++ and the libc++ std module sources (std.cppm — package llvm-runtimes on Alpine, libc++ on Arch), lld, and gio-2.0 headers.

make            # build/make/{imsd,imsd-media,imsd-dialerd,imsd-ofonod}
make check      # the same 7 unit-test suites
make install    # DESTDIR/PREFIX staged install incl. the packaging/ files

project.cpp stays the canonical build description; the Makefile tracks it.

Cross-compiling for a phone (aarch64 Alpine/postmarketOS)

packaging/make-sysroot.sh                  # one-time: Alpine aarch64 sysroot from the CDN
crafter-build -- --target=aarch64-alpine-linux-musl \
    --sysroot=$HOME/.cache/imsd/sysroot-aarch64-alpine \
    --march=armv8.6-a+fp16fml+aes+sha3+sm4 --mtune=cortex-a720
# same flags after `crafter-build test --target=aarch64-alpine-linux-musl`
# run the suite under qemu-aarch64 against the sysroot

Run (development)

./bin/imsd-*/imsd --session   # own net.catcrafts.IMS1 on the session bus
busctl --user call net.catcrafts.IMS1 /net/catcrafts/IMS1 net.catcrafts.IMS1 GetStatus

On a phone the daemon runs as root on the system bus (policy in packaging/net.catcrafts.IMS1.conf — it grants the postmarketOS default account user access; adjust it alongside AUDIO_USER if your session account is named differently), started by packaging/imsd.service after packaging/ims-pdn-up.sh has brought up the ims PDN through ModemManager. The bring-up never lets the ims PDN become the phone's IPv6 default route: the PDN's router advertisement is not accepted as a default router (accept_ra_defrtr=0 on the netdev; its prefix and MTU still apply) and the P-CSCF gets a host route instead, so on IPv4-only wifi the phone's other IPv6 traffic (apk, browsers) is not sent into the IMS walled garden, where it would hang. Media peers get a host route per call; anything else on the IMS network (its DNS servers, a traceroute target) needs its own (ip -6 route add <addr> dev qmapmux0.0). The PDN MTU ModemManager reports is set on the link. It spawns /usr/libexec/imsd-media per call. Registration resumes an in-kernel IPsec SA when one is present (avoiding the network's fresh-SA throttle) and keeps itself alive with a re-REGISTER refresh at half the granted lifetime. packaging/APKBUILD builds the apk from source via the Makefile; packaging/APKBUILD.binary + make-bin-tarball.sh wrap a cross-compiled build into an apk instead.

GNOME Calls (Phosh, GNOME Mobile)

imsd-ofonod presents imsd on the system bus as an oFono modem — a Manager at /, one Modem + VoiceCallManager at /imsd, and a VoiceCall object per live call — which is what GNOME Calls' bundled ofono provider plugin drives. Calls, Phosh and GNOME Shell run unmodified.

org.ofono is the seam because it is the only one a third party can implement. org.gnome.Calls.Call is what Calls exports so a shell can observe and control a call that already exists (Accept, Hangup, SendDtmf, Silence — there is no Dial); Phosh consumes it from Calls. Calls' own backends are libpeas provider plugins (mm, ofono, sip, dummy) linked against private headers that no distro installs.

Setup:

apk add imsd-ofono                    # enables imsd-ofonod via a systemd preset
systemctl enable --now imsd-ofonod    # hand installs: system service, org.ofono is a system name
# point Calls at the ofono provider (the package ships this as a gschema
# override; this is the per-user equivalent)
gsettings set org.gnome.Calls autoload-plugins "['ofono']"

Drop mm from autoload-plugins rather than adding ofono alongside it: the ModemManager provider would offer the modem's own CS voice path as a second origin, and on a phone whose voice path is imsd that origin carries no audio.

Do not run this alongside a real ofonod — there is one org.ofono name and one owner; the unit declares Conflicts=ofono.service and the daemon exits if it loses the name. Running imsd-dialerd at the same time is fine (different bus, different consumer).

Verified on a Fairphone 6 (postmarketOS, dbus-broker) with GNOME Calls 50.0: outgoing and incoming calls with audio both ways, answered and hung up through org.gnome.Calls.Call — the surface Phosh drives. One dbus-broker detail is load-bearing and lives in the policy file: dbus-broker checks a broadcast against the sender's send policy for every name the receiver owns, so root (imsd) must be allowed send_destination="org.ofono" or imsd's call signals never reach the daemon.

Known gaps, all inherited rather than introduced:

  • USSD is refused with org.ofono.Error.NotSupported; the IMS stack has no USSD path. It is deliberately not advertised in the modem's Interfaces.
  • No hold, multiparty, transfer or call waiting — imsd is a single-call stack, so those methods return org.ofono.Error.NotImplemented.
  • Per-call caller-ID withholding is refused rather than silently ignored: imsd cannot honour a CLIR request, and failing is safer than dialling with the caller ID exposed.
  • Phosh's emergency-call screen lists nothing. Not for want of the data: imsd-ofonod publishes org.ofono.VoiceCallManager.EmergencyNumbers (112, 911, plus EMERGENCY_NUMBERS), which is oFono's standard place for it. Calls' ofono provider simply never reads it — it makes no GetProperties call on the VoiceCallManager, and returns NULL for the origin's emergency-numbers, which is what org.gnome.Calls.EmergencyCalls.GetEmergencyContacts is built from. (The ModemManager provider does read it, via mm_sim_dup_emergency_numbers().) An upstream fix would close this. Dialling an emergency number from the normal dialpad is unaffected — that reaches imsd's own classifier.
  • DTMF is forwarded to imsd, where SendDtmf is still a stub (RFC 4733 is not implemented yet) — the same limitation Plasma has.
  • Restart GNOME Calls after restarting imsd-ofonod itself. When the modem's VoiceCallManager interface goes away and comes back, Calls 50.0 adds a new origin but leaves the old origin's CallAdded handler connected: after one such cycle every call is added twice, after a few the stale handler runs on freed memory and Calls dies with SIGBUS in g_hash_table_lookup (measured on the FP6). Calls-side; an upstream fix belongs in plugins/provider/ofono/calls-ofono-provider.c. To keep imsd's own restarts and registration drops from triggering it, imsd-ofonod lists the interface permanently: Online tracks registration, and Dial fails with "imsd is not registered" / "imsd is not running" instead. Only org.ofono vanishing — this daemon restarting — still cycles the origin.

Development, without a phone or root:

./bin/imsd-*/imsd --session &          # imsd's ABI on the session bus
./bin/imsd-ofonod-*/imsd-ofonod --session
busctl --user call org.ofono / org.ofono.Manager GetModems
# or keep org.ofono on the system bus and imsd on the session bus:
IMSD_BUS=session imsd-ofonod

Configuration

Everything is environment variables. The packaged unit reads /run/imsd.env (written by ims-pdn-up.sh: the connected ims netdev) and then /etc/imsd.env (your configuration, wins on conflict). Minimum viable /etc/imsd.env:

# REQUIRED until P-CSCF discovery from the PDN's PCO is implemented: your
# carrier's P-CSCF address. Find it in a stock-firmware capture or your
# carrier's IMS documentation.
PCSCF=2001:db8::105

imsd

Variable Default Meaning
PCSCF (none — required) P-CSCF address; without it registration fails with a clear error
PCSCF_PORT 5060 unprotected SIP port for the initial REGISTER
DEV qmapmux0.0 ims-PDN netdev; the packaged service passes the real one via /run/imsd.env
LOCAL (auto) UE address; default = the global IPv6 on DEV
USER_AGENT imsd/<version> REGISTER User-Agent. Some networks fingerprint UAs; setting your device's stock build string reproduces the stock modem's registration exactly. Empty omits the header
STATE_FILE /var/lib/imsd/imsreg.state (root) / $XDG_STATE_HOME/imsd/imsreg.state persisted registration context for warm resume
HOLDS_FILE /run/imsd.holds (root) / $XDG_RUNTIME_DIR/imsd.holds (/tmp/imsd.holds without a runtime dir) the waits the network imposed (a Retry-After, the refusal schedule) and where that schedule stands, kept across a restart of imsd for the address they were set on — not across a reboot; /dev/null switches the file off (only a regular file is ever removed; anything else is emptied)
RESUME auto 0 forces a fresh registration (ignores a warm SA)
REFRESH_INTERVAL (auto) keepalive re-REGISTER period in s; default = half the granted expiry, clamped to [120, 1800]
EALG aes-cbc offered ESP cipher (aes-cbc, des-ede3-cbc, null). From the 401's Security-Server list imsd selects the highest-preference ipsec-3gpp mechanism that matches this offer (RFC 3329 §2.3.1) and takes the SPIs/ports from the entry carrying them (some cores list every mechanism they support and put the SPIs on the one they applied); when no entry matches, the P-CSCF's SPI-carrying entry is installed as listed (hmac-md5-96, an absent ealg = null) and the journal says so. The line, the selection and the SA parameters are logged at every fresh registration
SIP_TRANSPORT auto transport of the protected leg (the second REGISTER and everything after it): tcp, udp, or auto = TCP, falling back to a fresh registration over UDP when the P-CSCF leaves two TCP connects unanswered — but only on a phone that has never registered over TCP: once the state file (/var/lib/imsd/imsreg.state) records a TCP registration, a silent connect is treated as an outage, not a policy (delete the file or set udp to override, e.g. after a SIM change). The challenge (first REGISTER) is always UDP. Over UDP a REGISTER or INVITE larger than ~1.2 KB leaves as IPv6 fragments at the ims PDN's 1280-byte MTU (the journal says so once per flow); there are no SIP retransmission timers over UDP yet, so a lost datagram costs a timeout. The SAs, the listener and the firewall rule cover both transports
EMERGENCY_NUMBERS (empty) comma-separated additions to the builtin 112/911 emergency numbers (e.g. a private test core's short code). SIM EF_ECC is not read yet
RTP_PORT 50004 local RTP port advertised in SDP
PRECOND 0 1 offers SDP QoS preconditions
CODECS (empty — defaults) comma-separated codec preference list over AMR-WB, AMR (or AMR-NB), PCMA, PCMU: restricts and orders both the codecs offered on an outgoing call and those accepted from an inbound offer (default: offer AMR-WB + AMR, accept all four in that order). A bench knob — a network whose gateway transcodes every caller up to AMR-WB otherwise never lets the narrowband path run
DUMP_SIP 0 1 writes raw REGISTER-200/SUBSCRIBE-200 dumps, the last refused REGISTER or SUBSCRIBE final (imsd-register-fail.raw, imsd-subscribe-fail.raw), the last inbound INVITE (imsd-invite-in.raw) and the last outgoing INVITE (imsd-invite-out.raw), mode 0600, for debugging — they contain your IMSI/MSISDN and addresses. A refused REGISTER, and a refused SUBSCRIBE final the run loop reads, are logged with the status line and any Retry-After / Warning / Reason / Server headers
DUMP_DIR, OUT_DIR state dir where dumps / per-call media stats land
IMSD_MEDIA (auto) path to imsd-media (default: next to imsd, else /usr/libexec/imsd-media)

The protected client/server ports are fixed at 45061/45062 (imsd::util::kPortUc/kPortUs) — changing them in source also requires updating the firewall rules ims-pdn-up.sh installs.

When the registration stops working (the refresh is refused, or the protected flow dies and nothing answers the reconnect) imsd does in place what a restart does: it reconnects through the existing SA and refreshes; after two refused refreshes or two unanswered reconnect cycles (such a cycle is not counted while the modem reports no packet service) it registers fresh (new Call-ID, new challenge, new SA pair). Reconnects back off 5 s → 5 min; a refused or throttled fresh attempt waits on the throttle schedule (21 min, growing to 60), a silent one 30 s → 10 min, and a network Retry-After is honoured, up to an hour (the schedule's longest wait): a longer one is cut to it and the refusal heard again then. While a fresh registration is deferred, the existing SA is still retried on the reconnect backoff, unless the old registration is known gone: a 403 once a fresh registration is due, or an attempt that got past the SA setup. A Dial that reaches imsd while it is unregistered brings the next recovery step forward (the reconnect, or the fresh attempt once it is due), never a wait the network imposed (a Retry-After, or the schedule after a refused or throttled fresh attempt); imsd-ofonod refuses a Dial then, and imsd-dialerd lists no modem. A bring-up the network deferred waits the same schedule out with nothing to reconnect over (the resume just found the stored SAs unusable), then tries again. The network's holds outlive a restart of imsd: they are kept in /run/imsd.holds for the address they were set on, with the schedule's place (a restart does not start it over at 21 min), and waited out at the next start, before the resume (a refresh's Retry-After) or before the fresh registration (a fresh attempt's schedule or Retry-After); a reboot or another address starts clean. The journal says which (…; next: fresh registration in N s, registering fresh (recovery attempt N), fresh registration deferred/refused/failed: …, bring-up deferred: …).

imsd-media (per-call, set through imsd's environment)

Variable Default Meaning
MIC, PLAY 1, 1 uplink mic capture / downlink playout via PipeWire
AUDIO_USER user desktop user whose PipeWire session carries call audio (postmarketOS default account)
GAIN, PLAY_GAIN 10, 1.0 uplink / downlink gain
AMR_MODE 2 AMR-WB encoder mode (0–8)
DTX 0 encoder discontinuous transmission
MEDIA_TIMEOUT 6.0 seconds of downlink silence before exit 3 (far-end-hangup signal)
RTP_DUMP 0 1 captures raw downlink RTP next to the stats file

D-Bus ABI

net.catcrafts.IMS1 at /net/catcrafts/IMS1. The ABI is frozen — dialer integrations depend on it:

Member Signature
Dial(number) s → s (callUni)
HangUp(callUni) s
Accept(callUni) s
SendDtmf(callUni, tones) ss
GetCalls() → aa{sv}
GetStatus() → a{sv}
CallAdded s a{sv}
CallStateChanged s s s (callUni, state, reason)
CallDeleted s
RegistrationChanged b

Both imsd-dialerd and imsd-ofonod are pure translations of this ABI; a new desktop needs a new translation daemon, not changes here.

License

GPL-3.0-only — see LICENSE.

Copyright (C) 2026 Catcrafts® catcrafts.net