Re: [PATCH] fix writing to the filesystem after unmount

[Al Viro <viro@zeniv.linux.org.uk>]

On Wed, Sep 06, 2023 at 05:03:34PM +0200, Mikulas Patocka wrote:

> > IOW, you'd also hang on any umount of a bind-mount. IOW, every
> > single container making use of this filesystems via bind-mounts would
> > hang on umount and shutdown.
> 
> bind-mount doesn't modify "s->s_writers.frozen", so the patch does nothing 
> in this case. I tried unmounting bind-mounts and there was no deadlock.

You are making *any* mount destruction hang if the sucker is frozen.
Which includes the things like exit(2) of the last process within
a namespace, etc.

And it does include the things like mount --bind /usr/bin/gcc /tmp/cc; umount /tmp/cc
if /usr happened to be frozen at the moment.

This is really not an option.

> BTW. what do you think that unmount of a frozen filesystem should properly 
> do? Fail with -EBUSY? Or, unfreeze the filesystem and unmount it? Or 
> something else?

It's not just umount(2).  It's exit(2).  And close(2).  And AF_UNIX garbage
collector taking out an undeliverable SCM_RIGHTS datagram that happens to
contain a reference to the last opened file on lazy-umounted fs, etc.