On Wed, Jan 30, 2019 at 07:35:39AM -0600, Eric W. Biederman wrote:

I suspect that as long as userspace supports /etc/fstab and we in turn
support /proc/mounts there is going to be a lot of pressure to keep
the majority of options so they  can be encoded in a string separated by
commas.

Well, we're doing many crazy things in userspace ;-) For example we do
not distinguish between VFS flags (MS_*), FS specific mount options
and userspace mount options (loop=) in our config files. 

You already need to parse fstab/command line before you can use the 
strings for mount(2) syscall. It's already not straightforward, see
all the code in libmount...

/proc/mounts is only for backward compatibility, /proc/self/mountinfo
is better way, because it allows to see VFS and FS, etc.

IMHO it would be better to not care about way how we use mount options 
strings now (in userspace) when you think about the new API design.

    Karel

-- 
 Karel Zak  [off-list ref]
 http://karelzak.blogspot.com