ci: harden the fusermount3 repair (#10485)

* ci: move the fusermount3 repair into a composite action

Three copies of the same block were already drifting apart, and the
target comes from PATH: only ever add setuid root to a root-owned,
non-symlink binary under the system bin paths, and say why otherwise.

* test: say that the process exited in the wait errors

"process exit status 1 before ... accepted connections" is missing its
verb. Also mark the SIGTERM return discarded - it fails with
os.ErrProcessDone exactly when the select below already handles it.

* ci: prefer the distro fusermount3 over escalating a shadow copy

The shadowing /usr/local/bin/fusermount3 is not root-owned either, so
setting its setuid bit would have handed root to a binary the runner
user owns - the repair now symlinks the distro one earlier in PATH and
touches nothing, keeping the in-place chmod for a root-owned binary with
no distro alternative. A setuid bit only grants root when root owns the
file, so accept an existing one only then.

* ci: run the FUSE workflows when the shared action changes

Their paths filters listed each workflow file but not the composite
action all three now call.
This commit is contained in:
Chris Lu
2026-07-29 14:02:33 -07:00
committed by GitHub
parent c4798979d8
commit a4692005e9
5 changed files with 87 additions and 40 deletions
@@ -0,0 +1,66 @@
name: Fix fusermount3 setuid
description: >
Make sure the fusermount3 an unprivileged mount will find can actually mount.
Some runner images carry a second, source-built fusermount3 in /usr/local/bin
that shadows the distro one in PATH; it is neither setuid nor root-owned, so
every mount fails with "mount failed: Operation not permitted". Both the Go
mount and sw-fuse resolve the helper through PATH.
Run it after the step that apt-installs fuse3.
runs:
using: composite
steps:
- name: Point PATH at a fusermount3 that can mount
shell: bash
run: |
set -euo pipefail
# setuid only grants root when root owns the file, and exec follows
# symlinks, so judge the target.
can_mount() {
[ -u "$1" ] && [ "$(stat -Lc %u "$1")" = 0 ]
}
bin=$(command -v fusermount3 || true)
if [ -z "$bin" ]; then
echo "no fusermount3 in PATH" >&2
exit 1
fi
if can_mount "$bin"; then
ls -l "$bin"
exit 0
fi
echo "$bin cannot mount unprivileged:"
ls -l "$bin"
# The distro fusermount3 is setuid root and is the one meant to be used.
# Reach it through a symlink earlier in PATH: exec resolves the link, so
# the target keeps its setuid bit, and nothing on the image is modified.
for distro in /usr/bin/fusermount3 /bin/fusermount3; do
if [ "$distro" != "$bin" ] && can_mount "$distro"; then
mkdir -p "$RUNNER_TEMP/fuse-bin"
ln -sf "$distro" "$RUNNER_TEMP/fuse-bin/fusermount3"
echo "$RUNNER_TEMP/fuse-bin" >> "$GITHUB_PATH"
echo "using $distro instead"
ls -l "$distro"
exit 0
fi
done
# No usable distro binary, so setting the bit is the only way out - but
# the target came from PATH: do that only for a root-owned system binary,
# never for anything else that happens to sit there.
case "$bin" in
/bin/*|/sbin/*|/usr/bin/*|/usr/sbin/*|/usr/local/bin/*|/usr/local/sbin/*) ;;
*) echo "refusing to setuid $bin: outside the system bin paths" >&2; exit 1 ;;
esac
if [ -L "$bin" ] || [ ! -f "$bin" ] || [ ! -x "$bin" ]; then
echo "refusing to setuid $bin: not a regular executable file" >&2
exit 1
fi
if [ "$(stat -c %u "$bin")" != 0 ] || [ "$(stat -c %g "$bin")" != 0 ]; then
echo "refusing to setuid $bin: not owned by root:root" >&2
exit 1
fi
sudo chmod u+s "$bin"
ls -l "$bin"