I guess so, though it is very hard to see how the nodes
could go wrong.

Andrew Hewus Fresh <andrew@afresh1.com> wrote:

> While adjusting style in rootdisk_nodes, I noticed that we don't
> verify that rrootdisk is in sync, just the main one.  While this is
> really only protecting against _rootdev changing, not an accident, but
> it's just one extra stat.
> 
> Comments, OK?
> 
> Index: rc
> ===================================================================
> RCS file: /cvs/src/etc/rc,v
> diff -u -p -r1.596 rc
> --- rc	21 Sep 2026 02:52:29 -0000	1.596
> +++ rc	21 Sep 2026 02:54:38 -0000
> @@ -316,16 +316,17 @@ rootdisk_nodes() {
>  	_rootdev=$(mount | grep ' / ' | grep -v ':' | cut -d ' ' -f 1)
>  	[[ -n $_rootdev ]] || return
>  
> +	_rrootdev=$(echo "$_rootdev" | sed -e 's,dev/,dev/r,')
>  	_maj=$(stat -f %Hr $_rootdev)
> +	_rmaj=$(stat -f %Hr $_rrootdev)
>  	_omin=$(stat -f %Lr $_rootdev)
>  	_min=$((_omin / 16 * 64))
> -	[[ -b /dev/rootdisk ]] &&
> +	[[ -b /dev/rootdisk ]] && [[ -b /dev/rrootdisk ]] &&
>  	    [[ "$(stat -qf %Hr,%Lr /dev/rootdisk)" = "$_maj,$_min" ]] &&
> +	    [[ "$(stat -qf %Hr,%Lr /dev/rrootdisk)" = "$_rmaj,$_min" ]] &&
>  	    return
>  
>  	# does not exist or device changed, redo
> -	_rrootdev=$(echo "$_rootdev" | sed -e 's,dev/,dev/r,')
> -	_rmaj=$(stat -f %Hr $_rrootdev)
>  	rm -f /dev/rrootdisk /dev/rootdisk
>  	mknod -m 600 /dev/rrootdisk c $_rmaj $_min
>  	mknod -m 600 /dev/rootdisk b $_maj $_min
>