From mboxrd@z Thu Jan  1 00:00:00 1970
From: Jan Schmidt <list.btrfs@jan-o-sch.net>
Subject: [RFC] Passing stream to userspace
Date: Fri, 12 Aug 2011 17:30:32 +0200
Message-ID: <4E454718.6000206@jan-o-sch.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Cc: Chris Mason <chris.mason@oracle.com>,
	Matthew Wilcox <matthew@wil.cx>,
	linux-btrfs <linux-btrfs@vger.kernel.org>,
	linux-fsdevel@vger.kernel.org
To: Al Viro <viro@ZenIV.linux.org.uk>,
	Christoph Hellwig <hch@infradead.org>
Return-path: <linux-btrfs-owner@vger.kernel.org>
List-ID: <linux-btrfs.vger.kernel.org>

Hi there,

I was promised you (viro, hch) have a decided opinion on this topic. The
original mail ("Getting a lot of fs-generated information to user
space") contains a lot of background and is way too long. Point is: I
want to get a bunch of data generated by the kernel (btrfs) to userland
(really).

Matthew Wilcox suggested to use an ioctl changing f_ops.read. Userland
would be like:

> int fd = open("/mnt/btrfs");
> ioctl(fd, BTRFS_IOC_STREAM);
> while (...) {
> 	read(fd, buf, 4096);
> 	...
> }
> close(fd);

To avoid doing buffering myself, I suggested passing a fd to the kernel
where it is expected to dump the generated information, like:

> int fd;
> int pipefd[2];
> struct io_args io_args;
>
> fd = open("/mnt/btrfs");
> pipe(pipefd);
> io_agrs.dest = pipefd[0];
>
> /* thread 1 */
> ioctl(fd, BTRFS_IOC_STREAM, &io_args);
> /* thread 2 */
> while (...) {
> 	read(pipefd[1], buf, 4096);
> 	...
> }

Any opinions on those or different suggestions? Thanks,
-Jan