[lxc-devel] [RFC PATCH 0/2] Loop device psuedo filesystem

[Seth Forshee]

On Wed, May 28, 2014 at 04:47:24PM -0700, H. Peter Anvin wrote:
> On 05/27/2014 02:58 PM, Seth Forshee wrote:
> > 
> > The patches implement a psuedo filesystem for loop devices, which will
> > allow use of loop devices in containters using standard utilities. Under
> > normal use a loopfs mount will initially contain a single device node
> > for loop-control which can be used to request and release loop devices.
> > Any devices allocated via this node will automatically appear in that
> > loopfs mount (and in devtmpfs) but not in any other loopfs mounts.
> > CAP_SYS_ADMIN in the userns of the process which performed the mount is
> > allowed to perform privileged loop ioctls on these devices.
> > 
> > Alternately loopfs can be mounted with the hostmount option, intended
> > for mounting /dev/loop in the host. This is the default mount for any
> > devices not created via loop-control in a loopfs mount (e.g. devices
> > created during driver init, devices created via /dev/loop-control, etc).
> > This is only available to system-wide CAP_SYS_ADMIN.
> > 
> 
> May I instead strongly advocate a slightly different solution: leave
> legacy loop devices where they are, with the current semantics, and let
> them be.  Make the loopfs loop devices completely independent.  Consider
> this equivalent of Unix98 ptys versus legacy BSD ptys.
> 
> Then, hopefully, use of the legacy ones will disappear over time.
> Enabling the new ones in losetup and friends is simple enough; this is
> not like ptys where the old scheme was hard-coded into a hundred
> different applications.

I'm not really sure what you're thinking should be changed about the
loop driver. Sure, I can think of a few things I'd change, but nothing
intractable.

If it's the semantics, I'm not really changing those in any significant
way. Today losetup opens /dev/loop-control and asks for a free device,
and it receives either an existing, unused device or a new device which
appears at /dev/loopN. All that changes here is that it would need to
try /dev/loop/loop-control as well, and devices would appear at
/dev/loop/N (which is a convention losetup already understands, it just
needs to look there in some cases where it doesn't currently).

Or perhaps you're suggesting a more radical change to the semantics?

Thanks,
Seth