Re: [PATCH] nextfd(2)

From: Alexey Dobriyan <adobriyan@gmail.com>
To: Eric Dumazet <eric.dumazet@gmail.com>
Cc: akpm@linux-foundation.org, viro@zeniv.linux.org.uk,
	torvalds@linux-foundation.org, drepper@gmail.com,
	linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org
Subject: Re: [PATCH] nextfd(2)
Date: Mon, 2 Apr 2012 00:31:54 +0300	[thread overview]
Message-ID: <20120401213154.GB4408@p183.telecom.by> (raw)
In-Reply-To: <1333295005.2325.5204.camel@edumazet-glaptop>

On Sun, Apr 01, 2012 at 05:43:25PM +0200, Eric Dumazet wrote:
> On Sun, 2012-04-01 at 15:57 +0300, Alexey Dobriyan wrote:
> 
> > +
> > +/* Return first opened file descriptor which is >= than the argument. */
> > +SYSCALL_DEFINE1(nextfd, unsigned int, fd)
> > +{
> > +	struct files_struct *files = current->files;
> > +	struct fdtable *fdt;
> > +
> > +	rcu_read_lock();
> > +	fdt = files_fdtable(files);
> > +	while (fd < fdt->max_fds) {
> > +		struct file *file;
> > +
> > +		file = rcu_dereference_check_fdtable(files, fdt->fd[fd]);
> > +		if (file) {
> > +			rcu_read_unlock();
> > +			return fd;
> > +		}
> > +		fd++;
> > +	}
> > +	rcu_read_unlock();
> > +	return -ESRCH;
> > +}
> 
> Interesting idea but what about using fdt->open_fds bitmap to have a
> fast search and less cache pollution ?
> 
> alloc_fd(start, flags) uses find_next_zero_bit(), you could use
> find_next_bit().

Indeed.
I've copied code from /proc/*/fd implementation which does loop.