More read-only filesystems mean less fsck during boot after crashes.
Especially on crappy machines like the Pinebook Poop, I keep /usr
read-only and run with this diff so I still get a relinked kernel.
rc's reorder_libs() already does the same remount dance, but for
multiple directories/filesystems.
My diff works as expected with read-write and read-only /usr as well as
interrupting /usr/libexec/reorder_kernel runs with ^C, i.e. a failed run
will correctly mount /usr ro again if it was ro before the run.
Feedback? OK?
Index: reorder_kernel.sh
===================================================================
RCS file: /cvs/src/libexec/reorder_kernel/reorder_kernel.sh,v
retrieving revision 1.13
diff -u -p -r1.13 reorder_kernel.sh
--- reorder_kernel.sh 7 Nov 2022 15:55:56 -0000 1.13
+++ reorder_kernel.sh 8 Nov 2022 11:02:42 -0000
@@ -27,13 +27,32 @@ LOGFILE=$KERNEL_DIR/$KERNEL/relink.log
PROGNAME=${0##*/}
SHA256=/var/db/kernel.SHA256
-# Silently skip if on a NFS mounted filesystem.
-df -t nonfs $KERNEL_DIR >/dev/null 2>&1
+# Silently skip if on NFS, otherwise remount the filesystem read-write.
+DEV=$(df -t nonfs $KERNEL_DIR 2>/dev/null | awk 'NR == 2 { print $1 }')
+MP=$(mount -t ffs | grep ^$DEV)
+RO=false
+if [[ $MP == *read-only* ]]; then
+ MP=${MP%% *}
+ mount -u -w $MP
+ RO=true
+fi
+
+restore_mount() {
+ if $RO; then
+ # Close the logfile to unbusy $MP by switching back to console.
+ exec 1>/dev/console
+ exec 2>&1
+ mount -u -r $MP
+ fi
+}
# Install trap handlers to inform about success or failure via syslog.
ERRMSG='failed'
-trap 'trap - EXIT; logger -st $PROGNAME "$ERRMSG" >/dev/console 2>&1' ERR
-trap 'logger -t $PROGNAME "kernel relinking done"' EXIT
+trap 'trap - EXIT
+ logger -st $PROGNAME "$ERRMSG" 2>/dev/console
+ restore_mount' ERR
+trap 'logger -t $PROGNAME "kernel relinking done"
+ restore_mount' EXIT
# Create kernel compile dir and redirect stdout/stderr to a logfile.
mkdir -m 700 -p $KERNEL_DIR/$KERNEL
