From: Jeff Layton <jlayton@redhat.com>
To: viro@zeniv.linux.org.uk, konishi.ryusuke@lab.ntt.co.jp
Cc: linux-fsdevel@vger.kernel.org, linux-mm@kvack.org,
linux-kernel@vger.kernel.org, linux-nilfs@vger.kernel.org,
NeilBrown <neilb@suse.com>,
Ross Zwisler <ross.zwisler@linux.intel.com>
Subject: Re: [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting business
Date: Sun, 05 Mar 2017 09:40:54 -0500 [thread overview]
Message-ID: <1488724854.2925.6.camel@redhat.com> (raw)
In-Reply-To: <20170305133535.6516-1-jlayton@redhat.com>
On Sun, 2017-03-05 at 08:35 -0500, Jeff Layton wrote:
> I recently did some work to wire up -ENOSPC handling in ceph, and found
> I could get back -EIO errors in some cases when I should have instead
> gotten -ENOSPC. The problem was that the ceph writeback code would set
> PG_error on a writeback error, and that error would clobber the mapping
> error.
>
I should also note that relying on PG_error to report writeback errors
is inherently unreliable as well. If someone calls sync() before your
fsync gets in there, then you'll likely lose it anyway.
filemap_fdatawait_keep_errors will preserve the error in the mapping,
but not the individual PG_error flags, so I think we do want to ensure
that the mapping error is set when there is a writeback error and not
rely on PG_error bit for that.
> While I fixed that problem by simply not setting that bit on errors,
> that led me down a rabbit hole of looking at how PG_error is being
> handled in the kernel.
>
> This patch series is a few fixes for things that I 100% noticed by
> inspection. I don't have a great way to test these since they involve
> error handling. I can certainly doctor up a kernel to inject errors
> in this code and test by hand however if these look plausible up front.
>
> Jeff Layton (3):
> nilfs2: set the mapping error when calling SetPageError on writeback
> mm: don't TestClearPageError in __filemap_fdatawait_range
> mm: set mapping error when launder_pages fails
>
> fs/nilfs2/segment.c | 1 +
> mm/filemap.c | 19 ++++---------------
> mm/truncate.c | 6 +++++-
> 3 files changed, 10 insertions(+), 16 deletions(-)
>
(cc'ing Ross...)
Just when I thought that only NILFS2 needed a little work here, I see
another spot...
I think that we should also need to fix dax_writeback_mapping_range to
set a mapping error on writeback as well. It looks like that's not
happening today. Something like the patch below (obviously untested).
I'll also plan to follow up with a patch to vfs.txt to outline how
writeback errors should be handled by filesystems, assuming that this
patchset isn't completely off base.
-------------------8<-----------------------
[PATCH] dax: set error in mapping when writeback fails
In order to get proper error codes from fsync, we must set an error in
the mapping range when writeback fails.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
---
fs/dax.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/fs/dax.c b/fs/dax.c
index c45598b912e1..9005d90deeda 100644
--- a/fs/dax.c
+++ b/fs/dax.c
@@ -888,8 +888,10 @@ int dax_writeback_mapping_range(struct address_space *mapping,
ret = dax_writeback_one(bdev, mapping, indices[i],
pvec.pages[i]);
- if (ret < 0)
+ if (ret < 0) {
+ mapping_set_error(mapping, ret);
return ret;
+ }
}
}
return 0;
--
2.9.3
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
next prev parent reply other threads:[~2017-03-05 14:40 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-03-05 13:35 [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting business Jeff Layton
2017-03-05 13:35 ` [PATCH 1/3] nilfs2: set the mapping error when calling SetPageError on writeback Jeff Layton
2017-03-07 13:46 ` Ryusuke Konishi
2017-03-05 13:35 ` [PATCH 2/3] mm: don't TestClearPageError in __filemap_fdatawait_range Jeff Layton
2017-03-05 13:35 ` [PATCH 3/3] mm: set mapping error when launder_pages fails Jeff Layton
2017-03-05 14:40 ` Jeff Layton [this message]
2017-03-06 23:08 ` [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting business Ross Zwisler
2017-03-07 10:26 ` Jan Kara
2017-03-07 14:03 ` Jeff Layton
2017-03-07 15:59 ` Ross Zwisler
2017-03-07 16:17 ` Jan Kara
2017-03-09 2:57 ` Theodore Ts'o
2017-03-09 9:04 ` Jan Kara
2017-03-09 10:47 ` Jeff Layton
2017-03-09 11:02 ` Jan Kara
2017-03-09 12:43 ` Jeff Layton
2017-03-09 13:22 ` Brian Foster
2017-03-09 14:21 ` Theodore Ts'o
2017-03-15 5:07 ` [RFC PATCH] mm: retry writepages() on ENOMEM when doing an data integrity writeback Theodore Ts'o
2017-03-15 11:59 ` Jan Kara
2017-03-15 14:09 ` Theodore Ts'o
2017-03-15 13:03 ` Michal Hocko
2017-03-16 10:18 ` Tetsuo Handa
2017-03-06 3:06 ` [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting business NeilBrown
2017-03-06 11:43 ` Jeff Layton
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1488724854.2925.6.camel@redhat.com \
--to=jlayton@redhat.com \
--cc=konishi.ryusuke@lab.ntt.co.jp \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-nilfs@vger.kernel.org \
--cc=neilb@suse.com \
--cc=ross.zwisler@linux.intel.com \
--cc=viro@zeniv.linux.org.uk \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).