Mount the rootfs read-only on both images

The rootfs mounted read-write, so a slot ran with its own files open to
change, and the factory-slot mounts rode along on the release image.
Both images now carry the read-only-rootfs feature: the ro root line and
the rcS default, the volatile links made at rootfs time, a writable copy
of /var/lib at boot, a build failure for a post-install that needs the
machine, and the removal of shadow, base-passwd, update-rc.d and
update-alternatives.

What must last or change at run time is handled file by file:

- forgefirm-users renders the four account files from the record into
  /run/forgefirm/accounts and bind-mounts each copy over its /etc file
  (useradd and the rest are gone with shadow); a render writes through
  the mount, and the image's own files apply until the first render.
- forgefirm-banner bind-mounts a copy of /etc/issue and writes the
  address block through it.
- sshd keeps its host keys under /data/forgefirm/ssh, so the fingerprint
  survives updates; both sshd configs carry the same HostKey lines.
- forgefirm-logging passes logrotate a state file under /var/run
  (logrotate refuses to run without one).
- forgefirm-persist points the boot timestamp and the random seed at
  /data/forgefirm.

The dev image appends the /factory slot mounts, without nofail (busybox
mount hands it to the kernel, which rejects it). The rootfs command
entries lose their semicolons: on scarthgap the value is the task's
vardeps, split on whitespace, so "name;" left the function body out of
the signature and a changed body did not remake the rootfs; with the
bodies tracked, the dev image's DATETIME string needs a vardepsexclude.
release.sh gains the read-only gate (root ro, no /factory line,
ROOTFS_READ_ONLY=yes, host keys on /data). image.health checks the
mounts, the account binds, the banner bind, the host keys and the
dev-only /factory mounts.

Proven on the bench reference (dev image 20260909140901): / ro, /data
rw, /var/lib a tmpfs copy, the four account files and /etc/issue bound
from tmpfs, the host keys in /data/forgefirm/ssh, no "Read-only file
system" line in any log; forgectrl.auth and commission.account-login (a
temporary account rendered, logged in over HTTPS and removed again),
kernel.latch-locked-idle and motion.liveness-probe PASS; logrotate runs
with the volatile state. forgetest unit tests 335 OK; both images build
clean, and debugfs on the built rootfs shows every setting above.

meta-forgefirm/recipes-forgefirm/images/forgefirm-image-dev.bb

@@ -43,6 +43,10 @@ IMAGE_ROOTFS_EXTRA_SPACE = "262144"
 # Dev builds identify by build timestamp (matches the artifact name),
 # tagged so a bench machine is never mistaken for a release.
 FORGEFIRM_VERSION_STRING = "${DATETIME} (dev)"
+# The rootfs functions that write the string are in the do_rootfs
+# signature; DATETIME must not be, or the basehash changes at every parse
+# and bitbake refuses the build as non-deterministic.
+FORGEFIRM_VERSION_STRING[vardepsexclude] += "DATETIME"
 
 # /etc/forgefirm-dev marks a dev image on the rootfs; its content is the
 # version string. The sshd init script starts sshd on a dev image without
@@ -52,4 +56,20 @@ write_forgefirm_dev_marker() {
     echo "${FORGEFIRM_VERSION_STRING}" > ${IMAGE_ROOTFS}${sysconfdir}/forgefirm-dev
 }
 write_forgefirm_dev_marker[vardepsexclude] += "DATETIME"
-ROOTFS_POSTPROCESS_COMMAND += "write_forgefirm_dev_marker;"
+ROOTFS_POSTPROCESS_COMMAND += "write_forgefirm_dev_marker "
+
+# The two factory rootfs slots, read-only under /factory/img1 and
+# /factory/img2: a bench convenience for reading a factory image in place
+# (the ffboot inventory reuses the mounts). The release image mounts no
+# factory slot; it reads one through a temporary read-only mount when it
+# needs to (ffboot -l). No nofail: busybox mount hands it to the kernel
+# as a filesystem parameter, which the kernel rejects (the mount fails
+# with EINVAL); mount -a goes on past a slot that does not mount anyway.
+add_factory_slot_mounts() {
+    install -d ${IMAGE_ROOTFS}/factory/img1 ${IMAGE_ROOTFS}/factory/img2
+    printf '%s\n' \
+        '/dev/mmcblk2p1       /factory/img1        auto       ro,noatime            0  0' \
+        '/dev/mmcblk2p2       /factory/img2        auto       ro,noatime            0  0' \
+        >> ${IMAGE_ROOTFS}${sysconfdir}/fstab
+}
+ROOTFS_POSTPROCESS_COMMAND += "add_factory_slot_mounts "

meta-forgefirm/recipes-forgefirm/images/forgefirm-image.bb

@@ -45,17 +45,35 @@ IMAGE_INSTALL:remove = "python3 ${FORGEFIRM_RELEASE_TRIM}"
 # VIRTUAL-RUNTIME_base-utils-syslog (conf/distro/forgefirm.conf).
 IMAGE_INSTALL:append = " grblhal-glowforge forgectrl gfhome gfcloud v4l-utils fwup ffboot slotmigrate forgefirm-logging"
 
-# forgefirm-users: replays the operator account record
+# forgefirm-users: renders the operator account record
 # (/data/forgefirm/users, written by forgectrl) into the system account
 # files at boot, before sshd, and on reload; also installs the warning an
 # interactive root shell prints. forgefirm-banner: keeps the control
 # panel addresses in the serial-console banner (/etc/issue).
+# forgefirm-persist: the boot timestamp and the random seed on /data.
 # avahi-daemon: mDNS, so the panel answers at https://forgefirm.local/
 # and shows up in service browsers. The daemon is installed by name (the
 # zeroconf distro feature stays off: it would bring libnss-mdns); the
 # build options and the configuration are in conf/distro/forgefirm.conf
 # and recipes-connectivity/avahi.
-IMAGE_INSTALL:append = " forgefirm-users forgefirm-banner avahi-daemon"
+IMAGE_INSTALL:append = " forgefirm-users forgefirm-banner forgefirm-persist avahi-daemon"
+
+# The rootfs mounts read-only on both images; /data (p3) is the writable
+# partition. read-only-rootfs is poky's feature for it: the root line of
+# /etc/fstab (the BSP's, already ro) and ROOTFS_READ_ONLY in
+# /etc/default/rcS, the volatile links made at rootfs time
+# (populate-volatile.sh: /etc/resolv.conf, /tmp), a writable copy of
+# /var/lib at boot (read-only-rootfs-hook.sh), a build failure for a
+# package whose post-install must run on the machine, and the removal of
+# the packages a read-only rootfs cannot use (shadow, base-passwd,
+# update-rc.d, update-alternatives; the account files stay). What must
+# last or change at run time is handled file by file: the account files
+# and /etc/issue (forgefirm-users, forgefirm-banner), the sshd host keys
+# (recipes-connectivity/openssh), the timestamp and the random seed
+# (forgefirm-persist). The facts are on the docs site,
+# technical/forgefirm/image-and-bsp; scripts/release.sh checks the built
+# rootfs for this state.
+IMAGE_FEATURES += "read-only-rootfs"
 
 # Root policy. root has no password and logs in at the serial console
 # only: that is the recovery path when the network, the panel or an
@@ -105,7 +123,9 @@ write_forgefirm_version() {
     echo "ForgeFIRM ${FORGEFIRM_VERSION_STRING}" > ${IMAGE_ROOTFS}${sysconfdir}/motd
 }
 write_forgefirm_version[vardepsexclude] += "DATETIME"
-ROOTFS_POSTPROCESS_COMMAND += "write_forgefirm_version;"
+# No semicolon after a function name here or below (the vardeps rule in
+# classes/forgefirm-image-manifest.bbclass).
+ROOTFS_POSTPROCESS_COMMAND += "write_forgefirm_version "
 
 # The license texts ride with the software. The license class writes
 # the image's license manifest (every installed package with its
@@ -132,4 +152,4 @@ pack_licenses() {
         > "${IMAGE_ROOTFS}${datadir}/forgefirm/licenses.tar.gz"
     rm -rf "$d"
 }
-ROOTFS_POSTPROCESS_COMMAND += "pack_licenses;"
+ROOTFS_POSTPROCESS_COMMAND += "pack_licenses "