[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20250210190541.GG1264@sol.localdomain>
Date: Mon, 10 Feb 2025 11:05:41 -0800
From: Eric Biggers <ebiggers@...nel.org>
To: Sami Tolvanen <samitolvanen@...gle.com>
Cc: Mikulas Patocka <mpatocka@...hat.com>,
Akilesh Kailash <akailash@...gle.com>, kernel-team@...roid.com,
Alasdair Kergon <agk@...hat.com>, Mike Snitzer <snitzer@...nel.org>,
Milan Broz <gmazyland@...il.com>, dm-devel@...ts.linux.dev,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH] dm-verity: do forward error correction on metadata I/O
errors
On Mon, Feb 10, 2025 at 08:57:55AM -0800, Sami Tolvanen wrote:
> Hi Mikulas,
>
> On Mon, Feb 10, 2025 at 7:04 AM Mikulas Patocka <mpatocka@...hat.com> wrote:
> >
> > Do forward error correction if metadata I/O fails.
> >
> > Signed-off-by: Mikulas Patocka <mpatocka@...hat.com>
> >
> > ---
> > drivers/md/dm-verity-target.c | 19 +++++++++++++++++--
> > 1 file changed, 17 insertions(+), 2 deletions(-)
> >
> > Index: linux-2.6/drivers/md/dm-verity-target.c
> > ===================================================================
> > --- linux-2.6.orig/drivers/md/dm-verity-target.c 2025-02-04 13:52:45.000000000 +0100
> > +++ linux-2.6/drivers/md/dm-verity-target.c 2025-02-10 15:55:42.000000000 +0100
> > @@ -324,8 +324,22 @@ static int verity_verify_level(struct dm
> > &buf, bio->bi_ioprio);
> > }
> >
> > - if (IS_ERR(data))
> > - return PTR_ERR(data);
> > + if (IS_ERR(data)) {
> > + r = PTR_ERR(data);
> > + data = dm_bufio_new(v->bufio, hash_block, &buf);
> > + if (IS_ERR(data))
> > + return r;
> > + if (verity_fec_decode(v, io, DM_VERITY_BLOCK_TYPE_METADATA,
> > + hash_block, data) == 0) {
> > + aux = dm_bufio_get_aux_data(buf);
> > + aux->hash_verified = 1;
> > + goto release_ok;
> > + } else {
> > + dm_bufio_release(buf);
> > + dm_bufio_forget(v->bufio, hash_block);
> > + return r;
> > + }
> > + }
>
> Don't we still have to check for io->in_bh before trying to correct the error?
>
> Overall, it would be nice not to duplicate code here. Should the
> metadata error correction / handling be moved to a separate function
> similar to verity_handle_data_hash_mismatch?
>
It's also incorrect to do this when skip_verified=true, since in that case
want_digest (which verity_fec_decode() uses) has not been initialized yet.
- Eric
Powered by blists - more mailing lists