From mboxrd@z Thu Jan  1 00:00:00 1970
From: Miklos Szeredi <miklos@szeredi.hu>
Subject: Re: [fuse-devel] [BUG] fuse: request prepared before protocol version
 is known
Date: Thu, 8 Jan 2015 15:14:18 +0100
Message-ID: <CAJfpegtqBEZLdhKjXe3VhZe5mp1LvsQRTOjHw3RcnWeqi+U6NA@mail.gmail.com>
References: <20141223175803.GA9789@zzz>
Mime-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Cc: fuse-devel <fuse-devel@lists.sourceforge.net>,
	Linux-Fsdevel <linux-fsdevel@vger.kernel.org>,
	"mszeredi@suse.cz" <mszeredi@suse.cz>,
	Kernel Mailing List <linux-kernel@vger.kernel.org>
To: Eric Biggers <ebiggers3@gmail.com>
Return-path: <linux-kernel-owner@vger.kernel.org>
In-Reply-To: <20141223175803.GA9789@zzz>
Sender: linux-kernel-owner@vger.kernel.org
List-Id: linux-fsdevel.vger.kernel.org

On Tue, Dec 23, 2014 at 6:58 PM, Eric Biggers <ebiggers3@gmail.com> wrote:
> I am using Linux 3.19-rc1 and I encountered a bug where stat() would
> sporatically fail with EIO when called on a file in a FUSE filesystem very
> shortly after that filesystem has been mounted.  This is a regression caused by
> commit 7078187a: "fuse: introduce fuse_simple_request() helper".
>
> The underlying issue is that the arguments to the FUSE request are being
> initialized before the protocol version (fc->minor) is known.  This was not a
> problem with the old code because fuse_get_req() will wait until the channel has
> been initialized.
>
> The only potential solutions I have in mind are adding an explicit waits in all
> the needed places, or reverting the commit.

Thanks for the report.  See for-linus branch of

 git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse.git for-linus

for the fix.

Thanks,
Miklos