Apply a PAM session to run/task/sysv/services Finit starts, pam_limits
above all, so a service running as a given user picks up that user's
limits the way a login does.
Add a new `pam` setting for the new block format (only), like the
per-service directories, naming a file in /etc/pam.d:
service weston {
user = "weston"
pam = "weston-autologin"
command = "/usr/bin/weston --continue-without-input"
}
pam_close_session() has to be called by a process still holding the
handle, and the handle does not survive exec(). Hence the keeper: it
holds the handle, drops to the service's credentials, and waits for a
parent-death signal before closing the session. Same shape as
systemd's (sd-pam), for the same reason, and one per fork, so the
script hooks open and close their own.
The keeper closes the descriptors it inherited from Finit and only
those. Closing everything would also take out what pam_open_session()
opened for itself, a keyring fd or a lock file, and leave the modules
to close a session with those pulled out from under them. Closing
nothing, as (sd-pam) does, would leave it holding the write end of the
notify pipe for the service's whole lifetime and starve notify = "s6"
services of their ready signal. So the fds open before pam_start()
are snapshotted and exactly those are closed, while the ones PAM opens
after are marked close-on-exec so the daemon does not inherit them
either.
A refused value, a denied account stack, an uninstalled pam.d file,
and a build without PAM support all keep the service from starting
rather than running it with the stacks skipped: one that quietly loses
pam_limits and its private /tmp, with nothing said. Capabilities a
module like pam_cap.so granted are merged into the IAB Finit applies
instead of being replaced by it, which only helps a service that also
sets capabilities, the other arm being a plain setuid() with nothing
left to restore once permitted is empty.
The test sysroot gains pam_permit.so, pam_deny.so and pam_limits.so,
which ldd cannot see, libpam dlopen()s them, and the test skips when
the host has none to stage. The negative cases pin the exit status
rather than only asserting crashed, which serv reports for any early
exit, so a bad command or an unwritable pidfile cannot pass for a
rejected session.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Stanzas in /lib/finit/system cannot use if:keventd, or any other
built-in service: if: is evaluated at parse time, and the generated
.conf files for built-ins were read after the system directory, so the
referenced service did not exist yet and the stanza was dropped.
The late position was deliberate: it kept bundled services from starting
before udev (5d703fd3). The motivating case was dbus, which has since
moved to a build-time 20-dbus.conf (2e0b1d6f), and the built-ins that
remain want the early slot: watchdogd should start as soon as possible,
keventd is the device manager, and runparts is gated on int/bootstrap so
its position never mattered.
Reading /run/finit/system first also puts the override chain in its
natural order: built-in/generated defaults, then bundled (read-only)
system files in /lib, then the administrator (override) .conf file in
/etc/finit.d, always read last.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
This is a backwards compatible mode for users upgrading and not noticing
that keventd is now build by default.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The line-based format has had the flag since v4.4 (issue #286), where
it prepends -p to the built-in getty, which turns it into login -p and
passes the environment on. The block format was written from the three
documented tty variants and the flags listed in the tty documentation,
and passenv was in neither, so it was left out. Converting a tty line
that used it therefore lost it, with nothing said.
It only reaches the built-in getty. An external getty is handed its
arguments through command, so there is nowhere to put a -p, and the
setting is refused with a warning rather than quietly ignored.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
A block title is a service identity, so two variants of one service
gated differently per platform cannot share a title. They do need to
share the barrier condition downstream services wait for, which until
now was spelled by the identity alone and so could not be shared:
service syslogd:udev {
if = "udevd"
conditions = { "run/udevadm:5/success" }
provides = "pid/syslogd"
command = "-syslogd -F"
}
Any namespace is allowed, since the point is publishing a name that
existing configurations already wait on. A claim on a condition that
is already owned is dropped with a warning naming the owner, and the
service still registers: the overlap is a configuration bug, and an
init system is more useful degraded than refusing to boot. A real
identity outranks a claim, pid/<ident> is how Finit tracks its own
services, so it is not up for grabs.
Claims are dropped before each reload re-reads the .conf files. Doing
it per service as it re-registers is not enough, since services are
read in file order and one re-registering would lose to a claim
another had not dropped yet, flipping the owner on every reload.
initctl cond dump asked who owned a condition only for the pid/
namespace and printed 'static' for usr/, which now hides a provider.
It asks first and falls back to what the namespace implies.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
A daemon known under more than one name had to be declared once per
name, each stanza carrying nowarn so the ones that were not installed
were skipped quietly. That leaned on the line-based format having no
titles. With a title as the service identity the repetition is no
longer available, so the candidates move inside the block:
command = { "/lib/systemd/systemd-udevd", "-udevd" }
Finit starts the first one whose binary resolves. A candidate that is
not installed is expected here, so nothing is logged for the ones that
lose, and only the winner is looked up again by service_register().
The leading '-' keeps its meaning and is read from the candidate that
wins, or from the last one when none resolve.
libconfuse accepts a bare string for a list option, so the common
`command = "prog args"` is unchanged, and whichp() already skips any
arguments to the command, so the candidate goes to it as written.
tty blocks take the same list. Their command is the getty to run, and
it has the same reason to name alternatives. Both block types now
share svc_command(), which also puts the leading '-' rule back in one
place: tty_translate() had its own copy. A tty needs no command at
all, it may name a device or notty instead, so svc_command() returns
NULL for an empty section and each caller decides whether that is an
error.
system/10-hotplug.conf.in returns to one udevd block, which is what it
meant all along.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The block title is the service identity, and libconfuse merges two
sections that share one, without a word. Two blocks titled the same
in one file therefore loaded as a single service holding a mix of both
declarations, with scalars taken from the last block and lists reset
by it.
system/10-hotplug.conf.in is written this way: two udevd blocks, one
per candidate binary, the way the line-based format spelled a
fallback. Only the second survived the merge, so a system that has
/lib/systemd/systemd-udevd but no udevd got no udevd service at all,
and the whole `if = "udevd"` chain behind it went with it.
CFGF_NO_TITLE_DUPES turns the merge into a parse error naming the file
and the title, and the file is then rejected as a whole. The same
title in another file is untouched, that is how an administrator
overrides a system .conf.
Format detection had to stop agreeing with it. is_new_format() probes
by parsing, so a duplicate title made it answer "not block format" and
conf_parse_file() handed the file to the legacy parser, whose errors
buried the real message. The probe now clears the flag on its own
copy of the option array, keeping the verdict syntactic.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The reload-signal translation calls str2sig() and compares against
SIGHUP, but conf.c never included signal.h. glibc pulls it in
transitively, so the omission went unnoticed until a cross-compile
against uClibc-ng in Buildroot failed to build.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The runparts directory and the dbus pidfile and daemon paths are
embedded in double-quoted values of the generated block files. A
literal quote in either ends the value early and libconfuse rejects
the whole file, and a backslash is read as an escape sequence,
silently mangling the path. The legacy one-liners had no quoting, so
neither failure existed before the block conversion.
conf_escape() doubles backslashes and escapes quotes; verified by
round-tripping hostile paths through cfg_parse_buf().
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Finit's own generated services -- watchdogd, keventd, runparts, and
the dbus plugin -- still went through conf_save_service() as legacy
one-liners, so `initctl show keventd` taught the old format on a
system otherwise converted to the new one.
conf_save_service() now takes the block title and a printf-style body
and writes the file itself:
# Generated by finit:conf_save_service()
service keventd {
description = "Finit kernel event daemon"
runlevel = "S12345789"
notify = "none"
cgroup init {}
command = "/libexec/finit/keventd"
}
vfprintf() into the file also removes the fixed-size staging buffers
in the callers, where a long dbus pidfile path could truncate inside
a quoted string and take the whole generated file with it.
Semantics preserved: watch-only pid:! maps to pidfile without
pidfile-create, the watchdog keeps its watchdog:finit identity, and
log:console becomes log { file = "/dev/console" }.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Aaron Andersen points out in the #492 discussion that the *Directory
settings carry more contract than create-and-chown: per-directory
modes, specific ownership rules, and cleanup toggles. Without them
config-dir was chowned to the service user, which systemd never does,
an existing directory with drifted ownership was left wrong, and the
runtime directory could not survive a restart.
Now matching systemd.exec(5), and where the man page is vague, the
code in setup_exec_directory():
- each directory takes a matching -mode key, octal with the leading
zero, default 0755. The mode of the named directory is locked
down again on every start, also when it already exists
- config-dir is created but never chowned
- the contents of an existing directory are left alone as long as
the owner is right; on drift everything under it is chowned back
- runtime-dir-preserve = no | restart | yes maps
RuntimeDirectoryPreserve=. A service still qualified to run when
the runtime directory would be removed is restarting, not
stopping, which is what svc_enabled() answers
The dir mechanics move to mksubsysd(), taking resolved ids, with
mksubsys() reduced to a name-resolving wrapper for the dbus plugin.
The child resolves uid/gid once for both directory setup and
privilege drop.
The symlink form, RuntimeDirectory=foo:bar, is not adopted.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
A service that drops privileges cannot create its own PID file in
/run, root owns it. Finit can create the file with pidfile-create,
but the daemon still cannot touch it to confirm a SIGHUP.
Five new settings, block format only: runtime-dir, state-dir,
cache-dir, logs-dir, and config-dir. The value is a directory name,
resolved under /run, /var/lib, /var/cache, /var/log, and /etc,
respectively. The directory is created before the service starts,
mode 0755 owned by user/group, and the full path is exported to the
process as RUNTIME_DIRECTORY, STATE_DIRECTORY, CACHE_DIRECTORY,
LOGS_DIRECTORY, and CONFIGURATION_DIRECTORY. Mode and ownership are
asserted at creation only, a daemon may tighten them afterwards.
The runtime directory is removed when the unit stops, after any
exec-stop-post script, like systemd with RuntimeDirectoryPreserve=no.
A completed run/task counts as stopped unless remain-after-exit keeps
it up. The other four persist across restarts.
These are the first settings with no legacy token: they are validated
by service_set_dir() and stored on the svc that service_register()
now returns. systemd accepts a list of directories per setting; this
is a single name for now, widening later is compatible since
libconfuse accepts a bare value for a list option.
The test sysroot gains libnss_files.so.2, which ldd cannot see, glibc
dlopen()s it. Without it getpwnam() fails inside the chroot, so
user/group settings never resolved and directory ownership could not
be tested.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The block format spells conditions as bare strings everywhere else, so
requiring `if = "<usr/foo>"` left one sigil behind, carried over from
the line-based `if:` token. A namespace separator already tells the two
apart: a value with a '/' is a condition, anything else is a service
name.
svc_ifthen() picks its mode from the start of the statement and applies
it to the whole, so a statement naming both kinds cannot be evaluated.
That is now an error, as are the old angle brackets, and either one
skips the block:
/etc/finit.conf: mixed: if: cannot mix a service name with a
condition in 'anchor,usr/enable-me', a statement must be all of
one kind, skipping
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
parse_cgroup() takes two arguments that are not cgroupfs files: the
leaf directory to place the service in, and whether to hand the subtree
over to it. The block format could express neither. 'name' happened
to work, because a free-form key is emitted as name:VALUE and that is
what the parser looks for, but 'delegate' came out as delegate:true and
was filed as a cgroup setting, so it silently did nothing.
Declare both, and emit delegate as the bare flag the parser expects.
Neither means anything on a top-level group definition, so say so there
rather than emitting something that would be written to cgroupfs.
service podman {
cgroup containers { name = "podman" delegate = true }
...
}
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Both keys prompt the question they should be answering.
'remain' decides whether a finished run or task keeps existing: without
it the entry is pruned, so the work re-runs on every runlevel entry,
initctl cannot see it, and its post script never fires. With it the
entry stays, is not re-run, and gets a teardown when stopped or when it
leaves its runlevels. That is systemd's RemainAfterExit, and 'remain'
is that name with the informative half cut off.
'manual' says how a service is started but not that it is about
starting at all.
remain -> remain-after-exit
manual -> manual-start
Both keep their old spelling as an alias, which they qualify for twice
over, as abbreviations of the canonical name and as the legacy
spellings.
While here, give sec_getbool() the alias argument its string and list
counterparts already take.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
A condition list could be led by '!', which is not a condition and not
a negation. It is a flag on the block, and it means two unrelated
things depending on which block it sits in: a service or sysv does not
handle SIGHUP and must be restarted to reload, while a run or task
must not hold up bootstrap. Writing '<!>' with no condition at all is
legal, which gives away that it was never an operator.
Give each meaning its own key, valid only where it applies:
service foo { reload-signal = "none" } # restart to reload
task bar { required = false } # do not hold up bootstrap
Using either on a block type it does not apply to warns, as does a '!'
left in a conditions list. Both still translate to that same '!',
which is all a legacy line can carry, so reload-signal takes SIGHUP or
none for now; str2sig() already accepts any case and an optional SIG
prefix.
This also clears the way for the conditions list to grow real
operators, '+' and '-' for asserted and deasserted, without '!'
sitting among them meaning something else entirely.
The '~' prefix stays. It belongs to the list: it marks a dependency
whose reload should propagate here.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The initial implementation was done using a naive translator of the
legacy one-liner format key by key, so it inherited encodings that the
block format exists to remove: a timeout packed into a script path, a
small comma-and-colon language inside the log string, sigils standing in
for booleans, and a log key (services) carrying three (!) types.
Settled naming against systemd, OpenRC, FreeBSD rc.subr, s6 and SMF.
Match systemd's semantics, not its naming.
pid -> pidfile, plus pidfile-create for the rare case
where Finit writes the file rather than the daemon
environment -> envfile, since it names a file to source, and the
top-level environment {} block sets variables
pre/post/... -> exec-start-pre, exec-start-ready, exec-stop,
exec-stop-post, exec-reload, exec-cleanup, each
with its own -timeout instead of "SEC,script"
halt, kill -> stop-signal, stop-timeout
restart -> restart for the policy, restart-max for the count
log -> a block with file, priority and identity, where
/dev/null and /dev/console are spelled as paths
group -> group and extra-groups, no longer positional
nowarn -> a leading - on command, as on envfile
List-valued keys take plural names. Aliases are desc, cond, mod,
caps, env, halt and kill; an alias may abbreviate the canonical name
or preserve a legacy spelling, nothing else.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Three private ones had grown: fnread() in util.c, flen() behind
pid_cmdline()/pid_cgroup() in cgutil.c, and conf_read_template() in
conf.c. Two of them were also wrong in ways the others were not.
fnread() formatted the path into a char[256] and stat()ed it before
opening, so a longer path was silently truncated and then read from
whichever file the truncation happened to name, and the size could
change between the look and the read. flen() existed because neither
of those approaches works on procfs at all, where stat() reports zero
and the only way to learn the size is to read to EOF.
Add fslurp() to util.[ch], which every tool already links. It opens
first and sizes the fd it holds, treats st_size as a hint, and reads
until EOF, so procfs and regular files take the same path. Paths are
formatted by libite's vfopenf(), which allocates to fit. Callers that
need the byte count, /proc/PID/cmdline embeds NUL, ask for it.
fnread() keeps its signature and becomes a bounded copy out of the
result, so its one caller is unaffected.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
A template in the block format registered garbage. conf_parse_file()
routed every file with an '@' in its name straight to the legacy
parser, which read the block line by line: the section header became a
service whose command was the section title, and each key = value line
below it became an environment variable.
service serv:%i { ... } -> service 'serv:eth0' with argument '{'
Substitute %i over the whole file before parsing instead, so format
detection and both parsers see finished text. A bare name@.conf is
still skipped, it is the template rather than an instance of one.
The legacy parser no longer opens the file or substitutes per line, it
is handed the instantiated buffer, so the template convention now has
one implementation instead of two. conf_is_template() applies
basenm(), a directory with an '@' in its name is not a template.
libconfuse cannot name a buffer it parses before 3.4, so a typo in a
template would be reported against "[buf]". Parse through fmemopen()
with the file name preset until the floor moves.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Relocate process-wide global variables from legacy parser that ended up
there because it used to be conf.c, but which is now now frozen at the
4.x feature set. Each variable is moved to their respective "owner".
Give cgroup_current[] and cgroup_settings_current[] named bounds. Their
extern declarations were unsized, so sizeof() on them stopped compiling
once the definitions moved to another translation unit.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The one-liner format has grown crowded and very wide, and every new
service option makes it worse.
Add a second, block-based format, parsed with libconfuse:
service sshd {
description = "OpenSSH daemon"
runlevel = "2345"
command = "/usr/sbin/sshd -D $SSHD_OPTS"
}
Both formats keep the .conf extension and are detected per file by
content. Try-parse strictly with libconfuse; on a parse error,
re-parse leniently to tell a block file with a typo from a one-liner
file. Only a one-liner file reaches the legacy parser, a typo is
reported with its file and line.
Each block is translated to the canonical one-liner and registered
through the existing entry points, so the two formats cannot drift.
The one-liner parser is frozen at the 4.x feature set, new options
land only in the block schema. libconfuse 3.3 or later is required,
CFGF_KEYSTRVAL does not exist before it.
Covers service, task, run, sysv and tty blocks, the static directives,
and the cgroup, rlimit, set and log blocks. Templating and the
documentation rewrite are still to come.
The regression test covers translation of a service block to the
one-liner, a block-format /etc/finit.conf booting with set {} applied
at bootstrap, both formats side by side, and rejection of a typo at
block and at root level.
A rejected file must not fall through to the legacy parser, which
registers a bogus unstartable service per line. assert_num_children
cannot see that, the bogus service has no children either, so the
check is assert_num_services.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The one-liner parser is about to be joined by a second, block-based
format. Give it a name that says which of the two it implements,
before any content changes make the diff hard to follow.
No functional change.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Conditions in Finit are dependencies: if A is asserted, service B is
allowed to run. When A goes through FLUX (e.g., upstream reloads),
dependents are PAUSED and then simply resumed when the condition is
reasserted -- this is the correct behavior for barrier-style deps
like <pid/syslogd>.
However, some setups have tightly coupled services where dependents
must be reloaded/restarted when an upstream service reloads, not just
resumed. E.g., the FRR routing stack on Infix OS:
netd <pid/mgmtd> ← zebra <!pid/netd> ← {staticd,ripd} <!pid/zebra>
When netd reloads (SIGHUP), zebra and its dependents must be restarted
to pick up the new configuration.
The new '~' condition prefix marks a dependency as flux-sensitive:
service <!~pid/netd> name:zebra ...
When the upstream condition goes FLUX and returns to ON, the dependent
is reloaded (SIGHUP) or restarted (noreload '!') instead of merely
resumed. Transitivity follows naturally through the condition chain.
Closes#416Closes#476
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
This commit activates the use of clone3() for service_fork(), to allow
Linux to create the new process directly in the correct cgroup instead
of later moving it there -- much cheaper and less error prone.
To facilitate this a few new helper cgroup functions have been added and
two new configuration directives introduced: delegate and name:leafname.
The delegate option is for running, e.g., container runtimes that want
to create their own cgroup v2 structur, and the name:leafname allows a
user to change the name of the subgroup under user/system/init.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
These variables really belong in conf.c. Relocate and move external
decls. from finit.h -> conf.h to simplify linking of other programs.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
This refactor should have been done years ago when we first introduced
the big state machine. The old functions service_runvel() and the oddly
named service_reload_dynamic() are actually state machine control fns,
so let relocate them.
Also in this commit, remove the global variable 'sm' and rename a few of
the critical functions to better names.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
When building with global CFLAGS=-Wunused-parameter the output from the
build looks terrible. This commit explicitly declares unused variables
to prevent triggering -Wunused-parameter
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Follow-up to 465bc17, which addressed unintended restart of siblings.
This patch fixes a problem where template instantiated services are not
properly reloaded/restarted when marked as "dirty" with `initctl touch`.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Consider the case where container@.conf is an available template. When
creating a container@foo.conf it will share the same base .conf as an
existing container@bar.conf, but we do not expect to restart bar just
because foo is instantiated.
Up until this change, all template siblings were considered "dirty" if a
new one was created or updated. Skipping realpath() for all files that
have a '@' works around the problem.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The responsibility of the kernel event daemon is to relay kernel events
to Finit. At shutdown and reboot it is too late for more events and the
daemon should just shut down with other services.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
There exist two possible basename functions, a xpg compliant one in libgen.h
and a GLIBC exclusive one declared in string.h, that was previously also declared by musl libc.
Both implementations are expecting different parameter types (`const char *` for GLIBC and `char *` for xpg)
With the removal of the basename function from string.h in musl libc, we could only rely on the xpg implementation.
Unfortunately, the xpg implementation of basename does modify the contents of whatever you put in it,
even though that there really is no need for it.
This is an issue in some cases, where we might want to get the basename of a read-only variable, e.g. a `const char *`,
as trying to modify something read-only is undefined behavior.
So in order to keep things consistent for us, we implement our own version of basename called `basenm`,
that does not modify the passed argument.
2023-11-13T04:02:23 [WRN]: conf_change_act():failed registering /etc/finit.conf/ event mask 00000004: Not a directory
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
This fixes an issue when finding a global environment variable with
spaces in the variable name:
set COLORTERM=yes
Literally, 'set COLORTERM' was the name of the variable.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
This change expands the readiness notification system in Finit with the
native 'pid' style, which will remain the default readiness in Finit 4.x
For systems that want to transition to Finit 5.x early, a global option
to set 'readiness none' in /etc/finit.conf, has been added. This change
the service default notification mode to 'notify:none', which can also
be set by Finit 4.x ('readiness pid') for select services.
Fixes#386.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Joachim Wiberg