[dm-devel] block: add a bi_error field to struct bio
Christoph Hellwig
hch at lst.de
Wed Jun 10 08:11:38 UTC 2015
On Thu, Jun 04, 2015 at 11:31:07AM -0400, Mike Snitzer wrote:
> This patch _really_ concerns me because just in DM alone I found you
> took liberties that you shouldn't have and created a regression. First
> issue is a real bug (your proposed dm-io.c:dmio_complete change missed
> that dm-io uses error_bits and not traditional error code like expected)
Point taken. I already wanted to complain about the mess due to the bio
error abuse with it's own values in DM in the first posting, guess I
need to add that to the second one. I don't think overloading common
interfaces with your private error codes is a good idea, but let's
leave that for a separate discussion.
> the other issue being you added extra branching that isn't needed and
> made review more tedious (dm.c:clone_endio).
I think the code is better than what it was before, but it's still
a bit of a mess. What do you think of the patch below which I'd
like to add before the big bi_error patch as a preparatory one?
> For DM, please add Signed-off-by: Mike Snitzer <snitzer at redhat.com> once
> you've folded in this patch, thanks!
FYI, that wasn't a foldable patch but updated hunks of the old one. Not
really a problem, but a little confusing.
>From f095cbeba5135afa6cf102718319f0d0c1e7b422 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch at lst.de>
Date: Wed, 10 Jun 2015 10:04:45 +0200
Subject: dm: use a single error code variable in clone_endio
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
clone_endio currently uses two variables for tracking error state, with
values getting bounceѕ forth and back between the two, which makes the
code hard to read.
Signed-off-by: Christoph Hellwig <hch at lst.de>
---
drivers/md/dm.c | 22 ++++++++++------------
1 file changed, 10 insertions(+), 12 deletions(-)
diff --git a/drivers/md/dm.c b/drivers/md/dm.c
index 2161ed9..8467976 100644
--- a/drivers/md/dm.c
+++ b/drivers/md/dm.c
@@ -956,7 +956,6 @@ static void disable_write_same(struct mapped_device *md)
static void clone_endio(struct bio *bio, int error)
{
- int r = error;
struct dm_target_io *tio = container_of(bio, struct dm_target_io, clone);
struct dm_io *io = tio->io;
struct mapped_device *md = tio->io->md;
@@ -966,23 +965,22 @@ static void clone_endio(struct bio *bio, int error)
error = -EIO;
if (endio) {
- r = endio(tio->ti, bio, error);
- if (r < 0 || r == DM_ENDIO_REQUEUE)
- /*
- * error and requeue request are handled
- * in dec_pending().
- */
- error = r;
- else if (r == DM_ENDIO_INCOMPLETE)
+ error = endio(tio->ti, bio, error);
+ if (error == DM_ENDIO_INCOMPLETE) {
/* The target will handle the io */
return;
- else if (r) {
- DMWARN("unimplemented target endio return value: %d", r);
+ }
+
+ if (error > 0 && error != DM_ENDIO_REQUEUE) {
+ DMWARN("unimplemented target endio return value: %d",
+ error);
BUG();
}
+
+ /* Error and requeue request are handled in dec_pending(). */
}
- if (unlikely(r == -EREMOTEIO && (bio->bi_rw & REQ_WRITE_SAME) &&
+ if (unlikely(error == -EREMOTEIO && (bio->bi_rw & REQ_WRITE_SAME) &&
!bdev_get_queue(bio->bi_bdev)->limits.max_write_same_sectors))
disable_write_same(md);
--
1.9.1
More information about the dm-devel
mailing list