Re: [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting business

From	NeilBrown <neilb@suse.com>
Newsgroups	linux.kernel
Subject	Re: [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting business
Date	2017-03-06 04:10 +0100
Message-ID	<thRc6-1Dq-9@gated-at.bofh.it> (permalink)
References	<thEHU-ZX-9@gated-at.bofh.it>
Organization	linux.* mail to news gateway

Show all headers | View raw

[Multipart message — attachments visible in raw view] - view raw

On Sun, Mar 05 2017, 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.
>
> 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.

Speaking of rabbit holes... I thought to wonder how IO error propagate
up from NFS.
It doesn't use SetPageError or mapping_set_error() for files (except in
one case that looks a bit odd).
It has an "nfs_open_context" and store the latest error in ctx->error.

So when you get around to documenting how this is supposed to work, it
would be worth while describing the required observable behaviour, and
note that while filesystems can use mapping_set_error() to achieve this,
they don't have to.

I notice that
  drivers/staging/lustre/lustre/llite/rw.c
  fs/afs/write.c
  fs/btrfs/extent_io.c
  fs/cifs/file.c
  fs/jffs2/file.c
  fs/jfs/jfs_metapage.c
  fs/ntfs/aops.c

(and possible others) all have SetPageError() calls that seem to be
in response to a write error to a file, but don't appear to have
matching mapping_set_error() calls.  Did you look at these?  Did I miss
something?

Thanks,
NeilBrown

>
> 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(-)
>
> -- 
> 2.9.3

Thread

[PATCH 0/3] mm/fs: get PG_error out of the writeback reporting business Jeff Layton <jlayton@redhat.com> - 2017-03-05 14:50 +0100
  [PATCH 3/3] mm: set mapping error when launder_pages fails Jeff Layton <jlayton@redhat.com> - 2017-03-05 14:50 +0100
  [PATCH 1/3] nilfs2: set the mapping error when calling SetPageError on writeback Jeff Layton <jlayton@redhat.com> - 2017-03-05 15:00 +0100
    Re: [PATCH 1/3] nilfs2: set the mapping error when calling  SetPageError on writeback Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> - 2017-03-07 17:50 +0100
  Re: [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting  business Jeff Layton <jlayton@redhat.com> - 2017-03-05 15:50 +0100
    Re: [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting  business Ross Zwisler <ross.zwisler@linux.intel.com> - 2017-03-07 03:00 +0100
      Re: [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting  business Jan Kara <jack@suse.cz> - 2017-03-07 11:40 +0100
        Re: [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting  business Jeff Layton <jlayton@redhat.com> - 2017-03-07 15:40 +0100
        Re: [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting  business Ross Zwisler <ross.zwisler@linux.intel.com> - 2017-03-07 17:10 +0100
          Re: [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting  business Jan Kara <jack@suse.cz> - 2017-03-07 17:50 +0100
        Re: [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting  business Theodore Ts'o <tytso@mit.edu> - 2017-03-09 04:00 +0100
          Re: [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting  business Jan Kara <jack@suse.cz> - 2017-03-09 10:30 +0100
            Re: [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting  business Jeff Layton <jlayton@redhat.com> - 2017-03-09 12:00 +0100
              Re: [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting  business Jan Kara <jack@suse.cz> - 2017-03-09 12:10 +0100
                Re: [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting  business Jeff Layton <jlayton@redhat.com> - 2017-03-09 13:50 +0100
                Re: [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting  business Brian Foster <bfoster@redhat.com> - 2017-03-09 14:30 +0100
                Re: [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting  business Theodore Ts'o <tytso@mit.edu> - 2017-03-09 16:10 +0100
  [PATCH 2/3] mm: don't TestClearPageError in __filemap_fdatawait_range Jeff Layton <jlayton@redhat.com> - 2017-03-05 16:30 +0100
  Re: [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting business NeilBrown <neilb@suse.com> - 2017-03-06 04:10 +0100
    Re: [PATCH 0/3] mm/fs: get PG_error out of the writeback reporting  business Jeff Layton <jlayton@redhat.com> - 2017-03-06 12:50 +0100

csiph-web