Expected single user behavior to NOT run the rcS script when runlevel S is given to sysvinit. From the busybox code - when going into single user, the logic is conditional to not even parse the inittab and jump into the single user shell. busybox/init/init.c /* Check if we are supposed to be in single user mode */ if (argv[1] && (strcmp(argv[1], "single") == 0 || strcmp(argv[1], "-s") == 0 || LONE_CHAR(argv[1], '1')) ) { /* ??? shouldn't we set RUNLEVEL="b" here? */ /* Start a shell on console */ new_init_action(RESPAWN, bb_default_login_shell, ""); } else { /* Not in single user mode - see what inittab says */ /* NOTE that if CONFIG_FEATURE_USE_INITTAB is NOT defined, * then parse_inittab() simply adds in some default * actions (i.e., INIT_SCRIPT and a pair * of "askfirst" shells) */ parse_inittab(); } However, the sysvinit uses the runlevels and the sysinit action runs before a runlevel has been selected - so, the rcS always runs. I believe this diff gets what should happen, but their might be a context that I don't know about for not doing this. diff --git a/package/sysvinit/inittab b/package/sysvinit/inittab index 27eb4a6..fad91ea 100644 --- a/package/sysvinit/inittab +++ b/package/sysvinit/inittab @@ -10,7 +10,7 @@ si2::sysinit:/bin/mkdir -p /dev/pts si3::sysinit:/bin/mkdir -p /dev/shm si4::sysinit:/bin/mount -a si5::sysinit:/bin/hostname -F /etc/hostname -si6::sysinit:/etc/init.d/rcS +rcS:12345:wait:/etc/init.d/rcS # S0:1:respawn:/sbin/getty -L ttyS0 115200 vt100 # GENERIC_SERIAL
Fixed on the next branch, thanks: commit eda809cd052dea73bca02385c7a10a4d7795c912 Author: Charles Hardin <ckhardin@exablox.com> Date: Sat May 21 13:11:10 2016 -0700 sysvinit: update the inittab to support "single" from the kernel Closes #8911 When the kernel passes single in the command line, this translates into an init -s option that is suppose to drop into a shell after the sysinit and before the runlevel. So, in busybox this is hardcoded - but, in sysvinit using the sysinit action for the rcS means that it will always be executed even when trying to get into single user mode for repair. This change should run rcS in all of the expected runlevels 1-5 and should achieve the desired result compared to the busybox and still allow a single user shell to get started as expected before running rcS. Signed-off-by: Charles Hardin <ckhardin@exablox.com> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>