Re: [patch 2/3] kcmp: Add KCMP_EPOLL_TFD mode to compare epoll target files

From: Cyrill Gorcunov <gorcunov@gmail.com>
To: Andrei Vagin <avagin@virtuozzo.com>
Cc: linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
	linux-api@vger.kernel.org, viro@zeniv.linux.org.uk,
	akpm@linuxfoundation.org, xemul@virtuozzo.com,
	mtk.manpages@gmail.com, kir@openvz.org,
	Andrey Vagin <avagin@openvz.org>, Jason Baron <jbaron@akamai.com>,
	Andy Lutomirski <luto@amacapital.net>
Subject: Re: [patch 2/3] kcmp: Add KCMP_EPOLL_TFD mode to compare epoll target files
Date: Mon, 13 Mar 2017 21:37:58 +0300	[thread overview]
Message-ID: <20170313183758.GE6743@uranus.lan> (raw)
In-Reply-To: <20170313173723.GA2855@outlook.office365.com>

On Mon, Mar 13, 2017 at 10:37:24AM -0700, Andrei Vagin wrote:
> 
> Here is one question inline.
> 
> Acked-by: Andrei Vagin <avagin@virtuozzo.com>

Thanks!

> > +
> > +	files = get_files_struct(task2);
> > +	if (!files)
> > +		return -EBADF;
> > +
> > +	spin_lock(&files->file_lock);
> > +	filp_epoll = fcheck_files(files, slot.efd);
> > +	if (filp_epoll)
> > +		get_file(filp_epoll);
> > +	else
> > +		filp_tgt = ERR_PTR(-EBADF);
> > +	spin_unlock(&files->file_lock);
> > +	put_files_struct(files);
> 
> Why can we not use fget here ^^^^ ?

Because it's not @current's file table but foreigner process.
So we make sure there is no expand_fdtable in progress and
we test for file descriptor which is being freed.