Message-ID: <1492000232.2937.2.camel@redhat.com>
Subject: Re: [PATCH v2 04/17] ext2: don't test/clear AS_EIO flag
From: Jeff Layton <jlayton@redhat.com>
To: Jan Kara <jack@suse.cz>
Cc: linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
        akpm@linux-foundation.org, tytso@mit.edu, willy@infradead.org,
        neilb@suse.com, viro@zeniv.linux.org.uk
Date: Wed, 12 Apr 2017 08:30:32 -0400
In-Reply-To: <20170412122921.GD1289@quack2.suse.cz>
References: <20170412120614.6111-1-jlayton@redhat.com>
         <20170412120614.6111-5-jlayton@redhat.com>
         <20170412122921.GD1289@quack2.suse.cz>
Content-Type: text/plain; charset="UTF-8"
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1662
Lines: 47

On Wed, 2017-04-12 at 14:29 +0200, Jan Kara wrote:
> On Wed 12-04-17 08:06:01, Jeff Layton wrote:
> > ext2 does a test+clear of AS_EIO flag. With the new error reporting
> > infrastructure, we don't need to clear anything. Sample the file's
> > current error code, and after writeback just report whether any
> > errors have been seen since then.
> > 
> > Signed-off-by: Jeff Layton <jlayton@redhat.com>
> > ---
> >  fs/ext2/file.c | 7 ++++++-
> >  1 file changed, 6 insertions(+), 1 deletion(-)
> > 
> > diff --git a/fs/ext2/file.c b/fs/ext2/file.c
> > index b21891a6bfca..0ca77d337c94 100644
> > --- a/fs/ext2/file.c
> > +++ b/fs/ext2/file.c
> > @@ -177,7 +177,12 @@ int ext2_fsync(struct file *file, loff_t start, loff_t end, int datasync)
> >  	struct address_space *mapping = sb->s_bdev->bd_inode->i_mapping;
> >  
> >  	ret = generic_file_fsync(file, start, end, datasync);
> > -	if (ret == -EIO || test_and_clear_bit(AS_EIO, &mapping->flags)) {
> > +	if (ret == -EIO)
> > +		ret = -EIO;
> > +	else
> > +		ret = filemap_check_errors(mapping);
> > +
> 
> IMO making __generic_file_fsync() perform the filemap_check_errors() after
> sync_mapping_buffers() is better and deals with all filesystem using
> generic_file_fsync(). Then we can just remove the AS_EIO check here.
> 
> 								Honza
> 

Thanks, I was just looking at this patch and thinking that it wasn't
quite right. I'll take your approach on the next set.

> > +	if (ret) {
> >  		/* We don't really know where the IO error happened... */
> >  		ext2_error(sb, __func__,
> >  			   "detected IO error when writing metadata buffers");
> > -- 
> > 2.9.3
> > 

-- 
Jeff Layton <jlayton@redhat.com>