mount: report unmount failures

Return a failure status from unmount_regular() when filesystems remain
mounted, and use it to report the final shutdown unmount result.

Signed-off-by: Anees Rehman <anees.rehman@atlascopco.com>

.github/workflows/build.yml

@@ -1,12 +1,13 @@
 name: Bob the Builder
 
-# Run on all branches, including all pull requests, except the 'dev'
-# branch since that's where we run Coverity Scan (limited tokens/day)
+# Run on push to release branches (master and maintenance N.x), and
+# on all pull requests.  The 'dev' branch is where we run Coverity
+# Scan (limited tokens/day)
 on:
   push:
     branches:
-      - '**'
-      - '!dev'
+      - master
+      - '[0-9]+.x'
   pull_request:
     types: [opened, synchronize, reopened, labeled]
 
@@ -22,8 +23,6 @@ jobs:
     # Verify we can build on latest Ubuntu with both gcc and clang
     name: ${{ matrix.compiler }}
     runs-on: ubuntu-latest
-    # Skip redundant builds for PRs - prefer PR builds over push builds
-    if: github.event_name != 'push' || github.ref == 'refs/heads/master'
     strategy:
       matrix:
         compiler: [gcc, clang]
@@ -43,7 +42,7 @@ jobs:
           (cd libuev-2.4.1 && ./configure && make -j9 && sudo make install-strip)
           (cd libite-2.6.2 && ./configure && make -j9 && sudo make install-strip)
           sudo ldconfig
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v7
       - name: Static Finit
         run: |
           ./autogen.sh
@@ -81,7 +80,7 @@ jobs:
           make -j1 check || (cat test/test-suite.log; false)
       - name: Upload Test Results
         if: always()
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v7
         with:
           name: finit-test-${{ matrix.compiler }}
           path: test/*.log

.github/workflows/coverity.yml

@@ -19,7 +19,7 @@ jobs:
   coverity:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v7
       - name: Fetch latest Coverity Scan MD5
         id: var
         env:
@@ -29,7 +29,7 @@ jobs:
                --post-data "token=$TOKEN&project=${COVERITY_PROJ}&md5=1" \
                -O coverity-latest.tar.gz.md5
           echo "md5=$(cat coverity-latest.tar.gz.md5)" | tee -a $GITHUB_OUTPUT
-      - uses: actions/cache@v4
+      - uses: actions/cache@v6
         id: cache
         with:
           path: coverity-latest.tar.gz
@@ -86,7 +86,7 @@ jobs:
             --form description="${PROJECT_NAME} $(git rev-parse HEAD)" \
             https://scan.coverity.com/builds?project=${COVERITY_PROJ}
       - name: Upload build.log
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v7
         with:
           name: coverity-build.log
           path: cov-int/build-log.txt

.github/workflows/docs.yml

@@ -4,10 +4,12 @@ on:
   push:
     branches:
       - master
+      - '[0-9]+.x'
     paths:
       - 'doc/**'
       - 'README.md'
       - 'mkdocs.yml'
+      - 'configure.ac'
       - '.github/workflows/docs.yml'
   pull_request:
     types: [opened, synchronize, reopened, labeled]
@@ -15,6 +17,7 @@ on:
       - 'doc/**'
       - 'README.md'
       - 'mkdocs.yml'
+      - 'configure.ac'
       - '.github/workflows/docs.yml'
 
 permissions:
@@ -23,14 +26,23 @@ permissions:
 jobs:
   build:
     runs-on: ubuntu-latest
+    outputs:
+      version: ${{ steps.version.outputs.version }}
     steps:
       - name: Checkout
-        uses: actions/checkout@v4
+        uses: actions/checkout@v7
         with:
           fetch-depth: 0  # Needed for git-revision-date-localized plugin
 
+      - name: Determine docs version
+        id: version
+        run: |
+          version=$(sed -n 's/^AC_INIT(\[Finit\], *\[\([0-9]*\)\..*/\1.x/p' configure.ac)
+          test -n "$version"
+          echo "version=$version" >> $GITHUB_OUTPUT
+
       - name: Setup Python
-        uses: actions/setup-python@v4
+        uses: actions/setup-python@v7
         with:
           python-version: '3.x'
 
@@ -46,25 +58,27 @@ jobs:
         run: mkdocs build --clean
 
       - name: Upload site artifact
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v7
         with:
           name: site
           path: site/
 
   deploy:
-    if: github.ref == 'refs/heads/master' && github.event_name == 'push'
+    if: github.event_name == 'push'
     needs: build
     runs-on: ubuntu-latest
+    env:
+      VERSION: ${{ needs.build.outputs.version }}
 
     steps:
       - name: Download site artifact
-        uses: actions/download-artifact@v4
+        uses: actions/download-artifact@v8
         with:
           name: site
           path: site/
 
       - name: Checkout pages repo
-        uses: actions/checkout@v4
+        uses: actions/checkout@v7
         with:
           repository: finit-project/finit-project.github.io
           path: pages
@@ -79,7 +93,56 @@ jobs:
 
       - name: Sync site to pages repo
         run: |
-          rsync -a --delete --exclude .git site/ pages/
+          rsync -a --delete site/ "pages/$VERSION/"
+
+      - name: Update version index and redirects
+        run: |
+          cd pages/
+
+          # The root holds only version directories and the files
+          # generated below, drop anything else (pre-versioned site)
+          find . -maxdepth 1 ! -name . ! -name .git ! -name CNAME \
+               ! -name '[0-9]*.x' -exec rm -rf {} +
+
+          latest=$(ls -d [0-9]*.x | sort -rV | head -1)
+
+          sep=
+          printf '[' > versions.json
+          for v in $(ls -d [0-9]*.x | sort -rV); do
+              printf '%s{"version": "%s", "title": "%s", "aliases": []}' \
+                     "$sep" "$v" "$v" >> versions.json
+              sep=', '
+          done
+          printf ']\n' >> versions.json
+
+          cat > index.html <<EOF
+          <!DOCTYPE html>
+          <html>
+            <head>
+              <meta charset="utf-8">
+              <meta http-equiv="refresh" content="0; url=$latest/">
+              <link rel="canonical" href="https://finit-project.github.io/$latest/">
+            </head>
+          </html>
+          EOF
+
+          cat > 404.html <<EOF
+          <!DOCTYPE html>
+          <html>
+            <head>
+              <meta charset="utf-8">
+              <title>Finit &mdash; page not found</title>
+              <script>
+                var path = window.location.pathname;
+                if (!/^\/[0-9]+\.x(\/|$)/.test(path))
+                  window.location.replace("/$latest" + path);
+              </script>
+            </head>
+            <body>
+              <p>Page not found.  Try the <a href="/$latest/">latest User Guide</a>.</p>
+            </body>
+          </html>
+          EOF
 
       - name: Commit and push
         run: |

.github/workflows/release.yml

@@ -14,7 +14,7 @@ jobs:
     if: startsWith(github.ref, 'refs/tags/')
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v7
       - name: Installing dependencies ...
         run: |
           sudo apt-get -y update
@@ -31,12 +31,17 @@ jobs:
         run: |
           ver=${GITHUB_REF#refs/tags/}
           echo "ver=${ver}" >> $GITHUB_OUTPUT
-          if echo $ver | grep -qE '^[0-9]+\.[0-9]+(\.[0-9]+)?(-alpha|-beta|-rc)[0-9]*$'; then
+          stable='[0-9]+\.[0-9]+(\.[0-9]+)?'
+          if echo $ver | grep -qE "^${stable}(-alpha|-beta|-rc)[0-9]*$"; then
             echo "pre=true"  >> $GITHUB_OUTPUT
           else
             echo "pre=false" >> $GITHUB_OUTPUT
           fi
-          if echo $ver | grep -qE '^[0-9.]+\.[0-9.]+(\.[0-9]+)?$'; then
+          # Mark as latest release only if this is the highest stable
+          # tag, so a 4.x maintenance release doesn't outrank 5.x
+          newest=$(git ls-remote --tags --refs origin | awk -F/ '{print $NF}' \
+                       | grep -E "^${stable}$" | sort -V | tail -1)
+          if [ "$ver" = "$newest" ]; then
             echo "latest=true"  >> $GITHUB_OUTPUT
           else
             echo "latest=false" >> $GITHUB_OUTPUT
@@ -64,7 +69,7 @@ jobs:
           mv ../*.tar.* artifacts/
       - name: Upload Test Results
         if: always()
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v7
         with:
           name: finit-release-test-logs
           path: finit-*/_build/sub/test/*.log

.github/workflows/weekly.yml

@@ -14,22 +14,31 @@ jobs:
   schedule:
     runs-on: ubuntu-latest
     outputs:
-      run: ${{ steps.schedule.outputs.run }}
+      branches: ${{ steps.schedule.outputs.branches }}
     steps:
-      - uses: actions/checkout@v4
-      - name: Check for any commits since last week
+      - uses: actions/checkout@v7
+        with:
+          fetch-depth: 0
+      - name: Check for branches with no commits since last week
         id: schedule
         run: |
-          commits="$(git log --format=%H --since='1 week ago')"
-          if [ -z "$commits" ]; then
-              echo "run=true"  >> $GITHUB_OUTPUT
-          else
-              echo "run=false" >> $GITHUB_OUTPUT
-          fi
+          branches=
+          maint=$(git branch -r --format='%(refname:lstrip=3)' | grep -E '^[0-9]+\.x$')
+          for branch in master $maint; do
+              commits="$(git log --format=%H --since='1 week ago' origin/$branch)"
+              if [ -z "$commits" ]; then
+                  branches="$branches\"$branch\","
+              fi
+          done
+          echo "branches=[${branches%,}]" >> $GITHUB_OUTPUT
   distcheck:
     needs: schedule
-    if: ${{ needs.schedule.outputs.run == 'true' }}
+    if: ${{ needs.schedule.outputs.branches != '[]' }}
     runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        branch: ${{ fromJSON(needs.schedule.outputs.branches) }}
+      fail-fast: false
     steps:
       - name: Install dependencies
         run: |
@@ -42,7 +51,9 @@ jobs:
           (cd libuev-2.4.1 && ./configure && make -j9 && sudo make install-strip)
           (cd libite-2.6.2 && ./configure && make -j9 && sudo make install-strip)
           sudo ldconfig
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v7
+        with:
+          ref: ${{ matrix.branch }}
       - name: Base run
         run: |
           ./autogen.sh
@@ -61,7 +72,7 @@ jobs:
           make -j1 distcheck || (cat $dir/test/test-suite.log; false)
       - name: Upload Test Results
         if: always()
-        uses: actions/upload-artifact@v4
+        uses: actions/upload-artifact@v7
         with:
-          name: finit-weekly-test-logs
+          name: finit-weekly-test-logs-${{ matrix.branch }}
           path: finit-*/_build/sub/test/*.log

configure.ac

@@ -1,4 +1,4 @@
-AC_INIT([Finit], [4.17-rc1], [https://github.com/finit-project/finit/issues],
+AC_INIT([Finit], [4.18-pre], [https://github.com/finit-project/finit/issues],
 	[finit], [https://http://finit-project.github.io/])
 AC_CONFIG_AUX_DIR(aux)
 AM_INIT_AUTOMAKE([1.11 foreign subdir-objects])

doc/ChangeLog.md

@@ -3,7 +3,53 @@ Change Log
 
 All relevant changes are documented in this file.
 
-[4.17][] - 2026-04-xx
+[Unreleased]
+------------
+
+### Changes
+
+- Restart log now spells out the signal name and flags core dumps,
+  e.g. `killed by SIGKILL` or `killed by SIGSEGV, core dumped`, in
+  place of the bare numeric `by signal: N`.  Gives operators a much
+  stronger breadcrumb when a daemon dies unexpectedly
+- Service `user:` and `group:` names, including supplementary groups,
+  can now be up to 63 characters, up from 15.  Longer names were
+  silently truncated before the lookup, so the service never started.
+  By Anees Rehman
+- Shutdown now stops after the first successful remount of `/` read
+  only.  The plain `mount -n -o remount,ro /`, which reads
+  `/etc/fstab`, is only used when both device-agnostic forms fail,
+  the same order as Debian's `umountroot`.  By Anees Rehman
+
+### Fixes
+
+- Remove stale daemon-owned (`pid:!`) pidfiles after unclean exits.
+  When a daemon dies via SIGKILL/OOM/segfault, or exits early during
+  startup, its pidfile lingers and can prevent the next instance from
+  starting -- `dbus-daemon` for example refuses to start when its
+  pidfile already exists, so Finit's restart loop gives up.  Finit now
+  drops the file when it still names the just-reaped PID and that PID
+  is no longer alive (the liveness check guards against PID reuse).
+  This is a controlled exception to the long-standing rule that Finit
+  does not touch service-owned pidfiles
+- Fix misspelled `SIGUNKOWN` returned by `sig_name()` for unknown
+  signal numbers, now spelled correctly as `SIGUNKNOWN`.  Surfaced by
+  the new restart log above
+- Fix mode of existing `d`/`D` directories in tmpfiles, the chmod branch
+  never ran because the directory creation swallowed `EEXIST`.  Mode and
+  owner of `d`/`D`, `e`, and `f`/`F` entries, and of the directories
+  `mksubsys()` creates for the dbus plugin, are now set through an open fd
+  and only when they differ, so an immutable directory like `/var/empty`
+  no longer warns on every boot.  By Sam Brkopac
+- Fix crash at config load for a `stop:` or `reload:` script written
+  with a timeout, e.g. `stop:5,/bin/true`, which took PID 1 down
+- Fix a lost reload for a service paused mid-reload by a condition in
+  flux.  A second reload arriving meanwhile re-parsed the unchanged .conf
+  and cleared the pending mark, so the service resumed without ever being
+  reloaded.  Seen on Infix, where sshd kept its old listen addresses after
+  a change that landed as two reloads in a row
+
+[4.17][] - 2026-04-28
 ---------------------
 
 ### Changes

doc/config/services.md

@@ -31,6 +31,14 @@ basename of the binary to guess the PID file to watch for the PID:
 `/var/run/ntpd.pid`.  If Finit guesses wrong, you have to submit the
 full `pid:!/path/to/file.pid`.
 
+With `pid:!/path`, the file belongs to the service: Finit reads it
+but does not create or remove it.  The one exception is *stale*
+cleanup — if the service dies without removing its own pidfile
+(SIGKILL, OOM, segfault), and the file still names the just-reaped
+PID, Finit removes it before the next retry.  This prevents daemons
+that refuse to start on an existing pidfile (e.g. `dbus-daemon`)
+from getting stuck in a crash-restart loop.
+
 **Example:**
 
 In the case of `ospfd` (below), we omit the `-d` flag (daemonize) to

mkdocs.yml

@@ -1,9 +1,9 @@
-site_name: ""
+site_name: User's Guide
 site_description: Fast Init for Linux Systems
-site_url: http://finit-project.github.io
+site_url: https://finit-project.github.io
 repo_url: https://github.com/finit-project/finit
 repo_name: Finit Project
-copyright: Copyright © 2008-2025 Joachim Wiberg
+copyright: Copyright © 2008-2026 Joachim Wiberg
 docs_dir: doc/
 edit_uri: edit/master/doc/
 extra_css:
@@ -13,42 +13,6 @@ exclude_docs: |
   TODO.md
   ChangeLog.md
 
-theme:
-  logo: img/logo.png
-  name: material
-  features:
-    - toc.follow
-    - toc.integrate
-    - navigation.path
-    - navigation.instant
-    - navigation.instant.progress
-    - navigation.tracking
-    - navigation.tabs
-#    - navigation.sections
-#    - navigation.expand
-#    - navigation.top
-    - navigation.footer
-    - search.highlight
-    - search.share
-#    - content.action.edit
-    - content.code.copy
-    - content.code.annotate
-  palette:
-    - media: "(prefers-color-scheme: light)"
-      scheme: default
-      primary: orange
-      accent: orange
-      toggle:
-        icon: material/weather-night
-        name: Switch to dark mode
-    - media: "(prefers-color-scheme: dark)"
-      scheme: slate
-      primary: black
-      accent: black
-      toggle:
-        icon: material/weather-sunny
-        name: Switch to light mode
-
 nav:
   - Getting Started:
     - Introduction: index.md
@@ -94,14 +58,62 @@ nav:
     - Distributions: distro.md
     - Requirements: requirements.md
 
+theme:
+  logo: img/logo.png
+  name: material
+  features:
+    - toc.follow
+#    - toc.integrate
+    - navigation.path
+    - navigation.instant
+    - navigation.instant.progress
+    - navigation.tracking
+    - navigation.indexes
+#    - navigation.tabs
+#    - navigation.sections
+#    - navigation.expand
+#    - navigation.top
+#    - navigation.footer
+    - search.highlight
+    - search.share
+#    - content.action.edit
+    - content.code.copy
+    - content.code.annotate
+    - content.footnote.tooltips
+  palette:
+    - media: "(prefers-color-scheme: light)"
+      scheme: default
+      primary: orange
+      accent: orange
+      toggle:
+        icon: material/weather-night
+        name: Switch to dark mode
+    - media: "(prefers-color-scheme: dark)"
+      scheme: slate
+      primary: black
+      accent: black
+      toggle:
+        icon: material/weather-sunny
+        name: Switch to light mode
+
 markdown_extensions:
   - admonition
   - attr_list
+  - footnotes
+  - pymdownx.blocks.caption
+  - pymdownx.critic
+  - pymdownx.caret
+  - pymdownx.keys
+  - pymdownx.mark
+  - pymdownx.tilde
   - pymdownx.details
   - pymdownx.superfences
   - pymdownx.highlight:
       anchor_linenums: true
   - pymdownx.inlinehilite
+  - pymdownx.emoji:
+      emoji_index: !!python/name:material.extensions.emoji.twemoji
+      emoji_generator: !!python/name:material.extensions.emoji.to_svg
   - pymdownx.snippets
   - pymdownx.tabbed:
       alternate_style: true
@@ -112,10 +124,29 @@ markdown_extensions:
 plugins:
   - search
   - callouts
-  - glightbox
+  - glightbox:
+      touchNavigation: true
+      loop: false
+      effect: zoom
+      slide_effect: slide
+      width: 100%
+      height: auto
+      zoomable: true
+      draggable: true
+      skip_classes:
+        - custom-skip-class-name
+      auto_themed: true
+      auto_caption: false
+      caption_position: bottom
+      background: black
+      shadow: false
+      manual: false
 
 extra:
+  generator: false
   homepage: https://finit-project.github.io/
+  version:
+    provider: mike
   social:
     - icon: fontawesome/brands/github
       link: https://github.com/finit-project/finit

src/helpers.c

@@ -379,7 +379,6 @@ void print(int rc, const char *fmt, ...)
 		/* Pending state: show description with spinner, no final status yet */
 		if (!buf[0])
 			return;
-		delline();
 		if (progress_style == PROGRESS_CLASSIC)
 			cprintf("\r%s ", pad(buf, sizeof(buf), ".", sizeof(buf)));
 		else
@@ -391,15 +390,21 @@ void print(int rc, const char *fmt, ...)
 	 * Final status.  Emit description + status in a single cprintf() so
 	 * both reach the console in one write(), preventing kernel messages or
 	 * command output from splitting the description from its [ OK ]/[FAIL].
+	 *
+	 * status() returns a pointer to a static buffer, so we must copy the
+	 * result for rc before calling status(3) — the second call overwrites
+	 * the same buffer, corrupting the first result.
 	 */
 	last_desc[0] = 0;
 
 	if (buf[0]) {
-		delline();
+		char st[64];
+
+		strlcpy(st, status(rc), sizeof(st));
 		if (progress_style == PROGRESS_CLASSIC)
-			cprintf("\r%s %s\n", pad(buf, sizeof(buf), ".", sizeof(buf)), status(rc));
+			cprintf("\r%s %s\n", pad(buf, sizeof(buf), ".", sizeof(buf)), st);
 		else
-			cprintf("\r\e[K%s%s\r%s\n", status(3), buf, status(rc));
+			cprintf("\r\e[K%s%s\r%s\n", status(3), buf, st);
 	} else {
 		if (progress_style == PROGRESS_CLASSIC)
 			cprintf("%s\n", status(rc));

src/mount.c

@@ -121,15 +121,19 @@ void unmount_tmpfs(void)
 	}
 }
 
-void unmount_regular(void)
+int unmount_regular(void)
 {
 	const struct mntent *mnt;
 	FILE *fp = NULL;
+	int rc = 0;
 
 	while ((mnt = iterator("/proc/mounts", &fp))) {
-		if (!unmount(mnt->mnt_dir))
+		rc = unmount(mnt->mnt_dir);
+		if (!rc)
 			iterator_end(&fp);  /* Restart iteration */
 	}
+
+	return !!rc;
 }
 
 /**

src/pid.c

@@ -201,7 +201,10 @@ int pid_file_set(svc_t *svc, char *file, int not)
  * pid:!foo          --> !/run/foo.pid
  * pid:!/run/foo.pid --> !/run/foo.pid
  *
- * Note, nothing is created or removed by Finit in this latter form.
+ * Nothing is created or removed by Finit in this latter form, with one
+ * exception: a verifiably stale pidfile (still names the just-reaped
+ * PID, and that PID is no longer alive) is removed so the next instance
+ * can start.  See service_clean_pidfile() in src/service.c.
  */
 int pid_file_parse(svc_t *svc, char *arg)
 {

src/service.c

@@ -1120,6 +1120,35 @@ static void service_notify_stop(svc_t *svc)
 	}
 }
 
+/*
+ * Drop a daemon-owned (pid:!) pidfile if it still names the just-reaped
+ * PID and that PID is gone.  The liveness check guards against reuse.
+ */
+static void service_clean_pidfile(svc_t *svc, pid_t reaped)
+{
+	pid_t pid;
+	char *fn;
+
+	if (reaped <= 1)
+		return;
+
+	fn = pid_file(svc);
+	if (!fn)
+		return;
+
+	pid = pid_file_read(fn);
+	if (pid != reaped || pid_alive(pid))
+		return;
+
+	if (remove(fn) && errno != ENOENT) {
+		logit(LOG_CRIT, "Failed removing stale service %s pidfile %s",
+		      svc_ident(svc, NULL, 0), fn);
+		return;
+	}
+
+	dbg("Removed stale service %s pidfile %s", svc_ident(svc, NULL, 0), fn);
+}
+
 /*
  * Clean up any lingering state from dead/killed services
  */
@@ -1137,6 +1166,8 @@ static void service_cleanup(svc_t *svc)
 		if (remove(fn) && errno != ENOENT)
 			logit(LOG_CRIT, "Failed removing service %s pidfile %s",
 			      svc_ident(svc, NULL, 0), fn);
+	} else if (svc->pidfile[0] == '!') {
+		service_clean_pidfile(svc, svc->pid);
 	}
 
 	/*
@@ -1623,7 +1654,8 @@ static void parse_script(svc_t *svc, char *type, char *script, int *tmo, char *b
 			     script, errstr);
 			goto err;
 		}
-		*tmo = (int)(sec * 1000);
+		if (tmo)
+			*tmo = (int)(sec * 1000);
 	} else {
 		path = script;
 		if (tmo)
@@ -2305,11 +2337,14 @@ int service_register(int type, char *cfg, struct rlimit rlimit[], char *file)
 	if (cgroup)
 		parse_cgroup(svc, cgroup);
 
-	/* New, recently modified or unchanged ... used on reload. */
+	/*
+	 * New or modified since the last reload.  The mark is cleared when
+	 * the change has been applied, on start or reload, so one that is
+	 * still set here is a change that has not been applied yet, e.g.
+	 * the service is paused waiting for a condition.  Leave it.
+	 */
 	if ((file && conf_changed(file)) || conf_changed(svc_getenv(svc)) || svc->args_dirty)
 		svc_mark_dirty(svc);
-	else
-		svc_mark_clean(svc);
 
 	svc_enable(svc);
 
@@ -2405,7 +2440,10 @@ void service_monitor(pid_t lost, int status)
 	if (svc_is_forking(svc)) {
 		/* Likely start script exiting */
 		if (svc_is_starting(svc)) {
-			svc->pid = 0;	/* Expect no more activity from this one */
+			/* Daemon died before clearing 'starting'; drop any stale pidfile. */
+			service_clean_pidfile(svc, lost);
+			svc->oldpid = lost;	/* So service_retry() logs the real PID */
+			svc->pid = 0;		/* Expect no more activity from this one */
 			goto cont;
 		}
 
@@ -2794,13 +2832,18 @@ static void service_retry(svc_t *svc)
 	timeout = ((*restart_cnt) <= (svc->restart_max / 2)) ? 2000 : 5000;
 	/* If a longer timeout was specified in the conf, use that instead. */
 	svc->restart_tmo = max(svc->restart_tmo, timeout);
-	logit(LOG_CONSOLE|LOG_WARNING, "Service %s[%d] died (%s%d), restarting (retry in %d msec) (attempt: %d/%d)",
-	      svc_ident(svc, NULL, 0), svc->oldpid,
-	      WIFEXITED(svc->status) ? "with exit status: " : "by signal: ",
-	      WIFEXITED(svc->status) ? WEXITSTATUS(svc->status) : WTERMSIG(svc->status),
-	      svc->restart_tmo,
-	      *restart_cnt,
-	      svc->restart_max);
+	if (WIFEXITED(svc->status))
+		logit(LOG_CONSOLE|LOG_WARNING,
+		      "Service %s[%d] died (exit status: %d), restarting (retry in %d msec) (attempt: %d/%d)",
+		      svc_ident(svc, NULL, 0), svc->oldpid, WEXITSTATUS(svc->status),
+		      svc->restart_tmo, *restart_cnt, svc->restart_max);
+	else
+		logit(LOG_CONSOLE|LOG_WARNING,
+		      "Service %s[%d] died (killed by %s%s), restarting (retry in %d msec) (attempt: %d/%d)",
+		      svc_ident(svc, NULL, 0), svc->oldpid,
+		      sig_name(WTERMSIG(svc->status)),
+		      WCOREDUMP(svc->status) ? ", core dumped" : "",
+		      svc->restart_tmo, *restart_cnt, svc->restart_max);
 
 	svc_unblock(svc);
 	service_step(svc);

src/sig.c

@@ -170,7 +170,7 @@ static struct sigmap {
 
 void mdadm_wait(void);
 void unmount_tmpfs(void);
-void unmount_regular(void);
+int unmount_regular(void);
 
 static void fs_swapoff(void)
 {
@@ -302,6 +302,7 @@ void do_shutdown(shutop_t op)
 	struct sched_param sched_param = { .sched_priority = 99 };
 	int in_cont = in_container();
 	int signo = SIGTERM;
+	int rc;
 
 	if (SHUTDOWN_DEBUG) cprintf("do_shutdown: entered, op=%d\n", op);
 
@@ -372,12 +373,13 @@ void do_shutdown(shutop_t op)
 	}
 
 	/* Unmount any tmpfs before unmounting swap ... */
-	print(0, "Unmounting filesystems ...");
+	print(-1, "Unmounting filesystems ...");
 	unmount_tmpfs();
 	fs_swapoff();
 
 	/* ... unmount remaining regular file systems. */
-	unmount_regular();
+	rc = unmount_regular();
+	print(rc, NULL);
 
 	/*
 	 * We sit on / so we must remount it ro, try all the things!
@@ -386,9 +388,9 @@ void do_shutdown(shutop_t op)
 	 *  - https://bugs.launchpad.net/ubuntu/+source/util-linux/+bug/29187
 	 */
 	sync();
-	run("mount -n -o remount,ro -t dummytype dummydev /", NULL);
-	run("mount -n -o remount,ro dummydev /", NULL);
-	run("mount -n -o remount,ro /", "mount");
+	if (run("mount -n -o remount,ro -t dummytype dummydev /", NULL) &&
+	    run("mount -n -o remount,ro dummydev /", NULL))
+		run("mount -n -o remount,ro /", "mount");
 
 	/* Call mdadm to mark any RAID array(s) as clean before halting. */
 	mdadm_wait();
@@ -630,7 +632,7 @@ const char *sig_name(int signo)
 			return signames[i].name;
 	}
 
-	return "SIGUNKOWN";
+	return "SIGUNKNOWN";
 }
 
 /*

src/svc.h

@@ -100,7 +100,7 @@ typedef enum {
 #define MAX_IDENT_LEN    (MAX_ARG_LEN + MAX_ID_LEN + 1)
 #define MAX_STR_LEN      64
 #define MAX_COND_LEN     (MAX_ARG_LEN * 3)
-#define MAX_USER_LEN     16
+#define MAX_USER_LEN     MAX_ARG_LEN
 #define MAX_NUM_SUPGROUPS 4
 #define MAX_NUM_FDS      64	     /* Max number of I/O plugins */
 #define MAX_NUM_SVC_ARGS 64

src/tmpfiles.c

@@ -24,6 +24,7 @@
 #include "config.h"		/* Generated by configure script */
 
 #include <dirent.h>
+#include <fcntl.h>
 #include <ftw.h>
 #include <getopt.h>
 #include <glob.h>
@@ -586,67 +587,74 @@ static void tmpfiles(char *line)
 				if (gid < 0)
 					gid = 0;
 
-				rc = makedir(path, mode ?: 0755);
-				if (rc && errno == EEXIST)
-					rc = chmod(path, mode ?: 0755);
-				if (chown(path, uid, gid))
-					warn("Failed chown(%s, %d, %d)", path, uid, gid);
+				if (!mode)
+					mode = 0755;
+				rc = makedir(path, mode);
+				if (!rc && dirperm(path, mode, uid, gid))
+					warn("Failed setting mode/owner on %s", path);
 			}
 			umask(omask);
 			break;
 		}
-		case 'e':
+		case 'e': {
+			int uid, gid;
+
 			if (glob(path, GLOB_NOESCAPE, NULL, &gl))
 				break;
 
+			uid = parse_uid(user);
+			gid = parse_gid(group);
+			if (gid < 0)
+				gid = 0;
+
 			for (size_t i = 0; i < gl.gl_pathc; i++) {
 				char *p = gl.gl_pathv[i];
-				int uid, gid;
 
-				/* e only adjusts existing directories */
-				if (!fisdir(p))
-					continue;
-
-				uid = parse_uid(user);
-				gid = parse_gid(group);
-				if (gid < 0)
-					gid = 0;
-
-				if (mode)
-					chmod(p, mode);
-				if (uid >= 0 && chown(p, uid, gid))
-					warn("Failed chown(%s, %d, %d)", p, uid, gid);
+				/* skip non-directories */
+				if (dirperm(p, mode, uid, gid) && errno != ENOTDIR)
+					warn("Failed setting mode/owner on %s", p);
 			}
 			break;
+		}
 		case 'f':
-		case 'F':
-			mkparent(path, 0755);
+		case 'F': {
+			int flags = O_WRONLY | O_CREAT | O_NOFOLLOW | O_CLOEXEC;
+			int fd, uid, gid;
+
 			if (type[1] == '+' || type[0] == 'F') {
-				/* f+/F will create or truncate the file */
-				fp = fopen(path, "w+");
+				flags |= O_TRUNC;	/* create or truncate */
 			} else {
-				/* f will create the file if it doesn't exist */
-				if (strc)
-					fp = fopen(path, "w");
+				if (!strc)
+					break;		/* create only if missing */
+				flags |= O_EXCL;
 			}
 
-			if (fp) {
-				int uid, gid;
-
-				write_arg(fp, arg);
-				rc = fclose(fp);
-
-				/* Apply mode and ownership */
-				if (mode)
-					chmod(path, mode);
-				uid = parse_uid(user);
-				gid = parse_gid(group);
-				if (gid < 0)
-					gid = 0;
-				if (uid >= 0 && chown(path, uid, gid))
-					warn("Failed chown(%s, %d, %d)", path, uid, gid);
+			mkparent(path, 0755);
+			fd = open(path, flags, 0666);
+			if (fd < 0) {
+				if (errno != EEXIST)
+					warn("Failed creating %s", path);
+				break;
 			}
+
+			fp = fdopen(fd, "w");
+			if (!fp) {
+				close(fd);
+				break;
+			}
+
+			write_arg(fp, arg);
+
+			uid = parse_uid(user);
+			gid = parse_gid(group);
+			if (gid < 0)
+				gid = 0;
+			if (fdperm(fd, mode, uid, gid))
+				warn("Failed setting mode/owner on %s", path);
+
+			rc = fclose(fp);
 			break;
+		}
 		case 'l': /* Finit extension, like 'L' but only if target exists */
 			if (!arg) {
 				paste(buf, sizeof(buf), "/usr/share/factory", path);

src/util.c

@@ -25,6 +25,7 @@
 
 #include <ctype.h>		/* isprint() */
 #include <errno.h>
+#include <fcntl.h>
 #include <grp.h>
 #include <pwd.h>
 #ifdef HAVE_MNTENT_H
@@ -280,6 +281,43 @@ int getcgroup(char *buf, size_t len)
 	return 0;
 }
 
+/*
+ * Set mode and owner on an open fd, only what differs is touched, the
+ * file may be immutable.  mode 0 skips the chmod, uid -1 the chown.
+ */
+int fdperm(int fd, mode_t mode, uid_t uid, gid_t gid)
+{
+	struct stat st;
+	int rc;
+
+	rc = fstat(fd, &st);
+	if (!rc && mode && (st.st_mode & 07777) != mode)
+		rc = fchmod(fd, mode);
+	if (!rc && uid != (uid_t)-1 && (st.st_uid != uid || st.st_gid != gid))
+		rc = fchown(fd, uid, gid);
+
+	return rc;
+}
+
+/*
+ * Same for a directory that may already exist.  Opened with
+ * O_NOFOLLOW | O_DIRECTORY, so the change lands on the directory
+ * itself and not on whatever a link at @path points to.
+ */
+int dirperm(const char *path, mode_t mode, uid_t uid, gid_t gid)
+{
+	int fd, rc;
+
+	fd = open(path, O_RDONLY | O_DIRECTORY | O_NOFOLLOW | O_CLOEXEC);
+	if (fd < 0)
+		return -1;
+
+	rc = fdperm(fd, mode, uid, gid);
+	close(fd);
+
+	return rc;
+}
+
 int mksubsys(const char *dir, mode_t mode, char *user, char *group)
 {
 	mode_t omask;
@@ -295,10 +333,8 @@ int mksubsys(const char *dir, mode_t mode, char *user, char *group)
 			gid = 0;
 
 		rc = makedir(dir, mode);
-		if (rc && errno == EEXIST)
-			rc = chmod(dir, mode);
-		if (chown(dir, uid, gid))
-			err(1, "Failed chown(%s, %d, %d)", dir, uid, gid);
+		if (!rc && dirperm(dir, mode, uid, gid))
+			warn("Failed setting mode/owner on %s", dir);
 	}
 
 	umask(omask);

src/util.h

@@ -68,6 +68,8 @@ int   getgroup     (const char *group);
 int   getcuser     (char *buf, size_t len);
 int   getcgroup    (char *buf, size_t len);
 
+int   fdperm       (int fd, mode_t mode, uid_t uid, gid_t gid);
+int   dirperm      (const char *path, mode_t mode, uid_t uid, gid_t gid);
 int   mksubsys     (const char *dir, mode_t mode, char *user, char *group);
 
 int   fnread       (char *buf, size_t len, char *fmt, ...) __attribute__ ((format (printf, 3, 4)));

test/Makefile.am

@@ -30,9 +30,10 @@ EXTRA_DIST		+= setup-sysroot.sh
 EXTRA_DIST		+= add-remove-dynamic-service.sh
 EXTRA_DIST		+= add-remove-dynamic-service-sub-config.sh
 EXTRA_DIST		+= bootstrap-crash.sh
-EXTRA_DIST		+= cond-start-task.sh
+EXTRA_DIST		+= script-timeout.sh
 EXTRA_DIST		+= crashing.sh
 EXTRA_DIST		+= dep-chain-reload.sh
+EXTRA_DIST		+= reload-while-paused.sh
 EXTRA_DIST		+= depserv.sh
 EXTRA_DIST		+= devmon.sh
 EXTRA_DIST		+= failing-sysv.sh
@@ -41,6 +42,7 @@ EXTRA_DIST		+= global-envs.sh
 EXTRA_DIST		+= initctl-status-subset.sh
 EXTRA_DIST		+= notify.sh
 EXTRA_DIST		+= pidfile.sh
+EXTRA_DIST		+= stale-pidfile.sh
 EXTRA_DIST		+= pre-post-serv.sh
 EXTRA_DIST		+= pre-fail.sh
 EXTRA_DIST		+= process-depends.sh
@@ -74,8 +76,10 @@ TESTS			+= add-remove-dynamic-service.sh
 TESTS			+= add-remove-dynamic-service-sub-config.sh
 TESTS			+= bootstrap-crash.sh
 TESTS			+= cond-start-task.sh
+TESTS			+= script-timeout.sh
 TESTS			+= crashing.sh
 TESTS			+= dep-chain-reload.sh
+TESTS			+= reload-while-paused.sh
 TESTS			+= depserv.sh
 TESTS			+= devmon.sh
 TESTS			+= failing-sysv.sh
@@ -84,6 +88,7 @@ TESTS			+= global-envs.sh
 TESTS			+= initctl-status-subset.sh
 TESTS			+= notify.sh
 TESTS			+= pidfile.sh
+TESTS			+= stale-pidfile.sh
 TESTS			+= pre-post-serv.sh
 TESTS			+= pre-fail.sh
 TESTS			+= process-depends.sh

test/lib/setup.sh

@@ -255,8 +255,12 @@ wdstart()
 
 wdkill()
 {
+    [ -n "${wdpid:-}" ] || return 0
+
     say "Stopping test watchdog, pid $wdpid"
-    kill -KILL $wdpid
+    # Reap the sleep first, killing its subshell orphans it to PID 1.
+    pkill -KILL -P "$wdpid" 2>/dev/null || true
+    kill -KILL "$wdpid" 2>/dev/null || true
 }
 
 teardown()
@@ -305,6 +309,18 @@ trap teardown EXIT
 SYSROOT="${SYSROOT:-$(pwd)/${TEST_DIR}/sysroot}"
 export SYSROOT
 
+# 'make check' refreshes the sysroot through the setup-chroot rule, but
+# running a test script by hand does not, and then the test silently
+# exercises whatever finit was installed last.  Compare and refuse.
+top_builddir="${top_builddir:-$TEST_DIR/..}"
+sysroot_finit="$SYSROOT/sbin/finit"
+built_finit="$top_builddir/src/finit"
+[ -x "$top_builddir/src/.libs/finit" ] && built_finit="$top_builddir/src/.libs/finit"
+
+if [ -x "$built_finit" ] && [ -e "$sysroot_finit" ] && ! cmp -s "$built_finit" "$sysroot_finit"; then
+	fail "Stale $sysroot_finit, run 'make -C test setup-chroot' or use 'make check'"
+fi
+
 TEST_TIMEOUT=300
 
 # shellcheck source=/dev/null

test/reload-while-paused.sh

@@ -0,0 +1,76 @@
+#!/bin/sh
+# A reload must not be lost on a service paused by a condition
+#
+# svc_s writes its PID file one second after a SIGHUP, so a service
+# with <pid/svc_s> is paused for that second whenever svc_s reloads.
+# svc_x counts the SIGHUPs it gets.  Both .conf files are touched and
+# two reloads are requested back to back, so the second one arrives
+# while svc_x is paused with its own reload still pending.  svc_x must
+# get exactly one SIGHUP and keep its PID.
+
+set -eu
+
+TEST_DIR=$(dirname "$0")
+
+test_teardown()
+{
+    say "Running test teardown."
+    run "rm -f $FINIT_RCSD/svc_s.conf $FINIT_RCSD/svc_x.conf /tmp/hup.sh /tmp/hup.log"
+}
+
+pidof()
+{
+    texec initctl -j status "$1" | jq .pid
+}
+
+test_setup()
+{
+    run "cat > /tmp/hup.sh" <<EOF
+#!/bin/sh
+trap 'echo HUP >> /tmp/hup.log' HUP
+trap 'exit 0' TERM
+while true; do
+    sleep 1
+done
+EOF
+    run "cat > $FINIT_RCSD/svc_s.conf" <<EOF
+service log:stdout notify:pid pid:!/run/service.pid name:svc_s service.sh -- Slow to reassert
+EOF
+    run "cat > $FINIT_RCSD/svc_x.conf" <<EOF
+service log:stdout <pid/svc_s> name:svc_x /bin/sh /tmp/hup.sh -- Paused while svc_s reloads
+EOF
+}
+
+# shellcheck source=/dev/null
+. "$TEST_DIR/lib/setup.sh"
+
+sep "Configuration"
+run "cat $FINIT_RCSD/svc_s.conf"
+run "cat $FINIT_RCSD/svc_x.conf"
+
+say "Reload Finit to start both services"
+run "initctl reload"
+retry 'assert_status "svc_x" "running"' 10 1
+
+pid_x=$(pidof svc_x)
+say "svc_x PID before: $pid_x"
+
+sep "Touch both, reload twice"
+run "rm -f /tmp/hup.log"
+run "initctl touch svc_s.conf"
+run "initctl touch svc_x.conf"
+run "initctl reload"
+run "initctl reload"
+
+say "Wait for services to settle"
+retry 'assert_status "svc_x" "running"' 15 1
+run "initctl status"
+
+# shellcheck disable=SC2016
+retry 'assert "svc_x got its SIGHUP" "$(texec sh -c "cat /tmp/hup.log 2>/dev/null | wc -l")" -eq 1' 10 1
+
+new_pid_x=$(pidof svc_x)
+# shellcheck disable=SC2086
+assert "svc_x was not restarted" $new_pid_x -eq $pid_x
+
+return 0

test/script-timeout.sh

@@ -0,0 +1,36 @@
+#!/bin/sh
+# A timeout on a stop: or reload: script must not take PID 1 with it.
+# Those two hooks passed a NULL timeout pointer to parse_script(),
+# which wrote through it whenever the script was prefixed with a
+# valid number.  Only a valid number reached the store, so
+# 'stop:abc,/bin/true' was harmless while 'stop:5,/bin/true' was not.
+set -eu
+
+TEST_DIR=$(dirname "$0")
+
+test_teardown()
+{
+    say "Running test teardown."
+    run "rm -f $FINIT_CONF"
+}
+
+# shellcheck source=/dev/null
+. "$TEST_DIR/lib/setup.sh"
+
+# shellcheck disable=SC2154
+assert_alive()
+{
+	assert "Finit survived $1" "$(kill -0 "$finit_pid" 2>/dev/null && echo yes)" = "yes"
+}
+
+for hook in stop reload post pre; do
+	say "Timeout on a $hook: script"
+	run "echo 'service $hook:5,/bin/true service.sh -- Timeout test' > $FINIT_CONF"
+	run "initctl reload" || true
+	assert_alive "$hook:5,/bin/true"
+done
+
+say 'The service still runs afterwards'
+run "echo 'service stop:5,/bin/true service.sh -- Timeout test' > $FINIT_CONF"
+run "initctl reload"
+retry 'assert_num_children 1 service.sh'

test/src/serv.c

@@ -34,6 +34,7 @@ volatile sig_atomic_t reloading = 1;
 volatile sig_atomic_t running   = 1;
 static char *ident = PROGNM;
 static char fn[80];
+static int  exclusive = 0;	/* -x: refuse to start if pidfile exists (dbus-style) */
 
 static void verify_env(char *arg)
 {
@@ -122,7 +123,7 @@ static void writefn(char *fn, int val)
 
 static int checkfn(char *fn)
 {
-	return !access(fn, R_OK);
+	return !access(fn, F_OK);
 }
 
 static void mine(char *fn)
@@ -133,12 +134,18 @@ static void mine(char *fn)
 
 static void pidfile(char *pidfn)
 {
+	static int once = 0;
+
 	if (!pidfn) {
 		if (fn[0] == 0)
 			snprintf(fn, sizeof(fn), "%s%s.pid", _PATH_VARRUN, ident);
 		pidfn = fn;
 	}
 
+	if (exclusive && !once && checkfn(pidfn))
+		errx(EX_SOFTWARE, "PID file %s exists, refusing to start", pidfn);
+	once = 1;
+
 	if (!checkfn(pidfn)) {
 		pid_t pid;
 
@@ -148,7 +155,7 @@ static void pidfile(char *pidfn)
 		atexit(cleanup);
 	} else {
 		inf("Touching PID file %s", pidfn);
-		utimensat(0, fn, NULL, 0);
+		utimensat(AT_FDCWD, pidfn, NULL, 0);
 	}
 }
 
@@ -182,6 +189,7 @@ static int usage(int rc)
 		" -p       Create PID file despite running in foreground\n"
 		" -P FILE  Create PID file using FILE\n"
 		" -r SVC   Call initctl to restart service SVC (self)\n"
+		" -x       Refuse to start if PID file already exists (dbus-style)\n"
 		"\n"
 		"By default this program daemonizes itself to the background, and,\n"
 		"when it's done setting up its signal handler(s), creates a PID file\n"
@@ -212,7 +220,7 @@ int main(int argc, char *argv[])
 	char cmd[80];
 	int c;
 
-	while ((c = getopt(argc, argv, "cCe:E:f:F:hi:nN:pP:r:")) != EOF) {
+	while ((c = getopt(argc, argv, "cCe:E:f:F:hi:nN:pP:r:x")) != EOF) {
 		switch (c) {
 		case 'c':
 			do_crash = 1;
@@ -249,12 +257,17 @@ int main(int argc, char *argv[])
 			break;
 		case 'P':
 			pidfn = optarg;
+			if ((size_t)snprintf(fn, sizeof(fn), "%s", optarg) >= sizeof(fn))
+				errx(EX_USAGE, "-P path too long (max %zu)", sizeof(fn) - 1);
 			do_pidfile++;
 			break;
 		case 'r':
 			snprintf(cmd, sizeof(cmd), "initctl restart %s", optarg);
 			do_restart = 1;
 			break;
+		case 'x':
+			exclusive = 1;
+			break;
 		default:
 			return usage(1);
 		}

test/stale-pidfile.sh

@@ -0,0 +1,38 @@
+#!/bin/sh
+# Regression: Finit must remove a daemon-owned (pid:!) stale pidfile
+# left behind by an unclean exit (SIGKILL), so the next instance can
+# start.  Simulates the dbus-daemon pattern: 'serv -x' refuses to
+# start if its pidfile already exists.
+
+set -eu
+
+TEST_DIR=$(dirname "$0")
+PIDFN="/run/serv.pid"
+
+test_teardown()
+{
+    say "Running test teardown."
+    run "rm -f $FINIT_CONF $PIDFN"
+}
+
+# shellcheck source=/dev/null
+. "$TEST_DIR/lib/setup.sh"
+
+say "Add service stanza '$FINIT_CONF' with pid:!$PIDFN"
+run "echo 'service pid:!$PIDFN serv -np -P $PIDFN -x' > $FINIT_CONF"
+run "initctl reload"
+
+retry "assert_num_children 1 serv"
+assert_is_pidfile "serv" "$PIDFN"
+
+PID=$(texec cat "$PIDFN")
+say "SIGKILL serv ($PID) -- leaves stale pidfile"
+run "kill -9 $PID"
+
+# Without the fix, 'serv -x' refuses to start on each retry until
+# Finit hits restart_max and marks the service crashed.  With the fix
+# Finit removes the stale pidfile and the next instance comes up.
+retry "assert_pidiff serv $PID"
+retry "assert_num_children 1 serv"
+
+sep