Re: [PATCH 1/4] locks: new locks_mandatory_area calling convention

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

On Thu, Dec 03, 2015 at 12:59:49PM +0100, Christoph Hellwig wrote:
> Pass a loff_t end for the last byte instead of the 32-bit count
> parameter to allow full file clones even on 32-bit architectures.
> While we're at it also drop the pointless inode argument and simplify
> the read/write selection.

locks_mandatory_area() contains this:
        if (filp && !(filp->f_flags & O_NONBLOCK))
                sleep = true;
which is a strong hint that filp might be NULL.  And indeed it might -
        error = locks_verify_truncate(inode, NULL, length);
in vfs_truncate() and
        host_err = locks_verify_truncate(inode, NULL, iap->ia_size);
in nfsd_get_write_access().  Both are broken by that commit.

Where the hell would truncate(2) get struct file, anyway?  IOW, the inode
argument is _not_ pointless; re-added.