On Mon, Jan 28 2019 at 11:00am -0500,
Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> On Mon, Jan 28, 2019 at 10:31:41AM -0500, Mike Snitzer wrote:
> > On Mon, Jan 28 2019 at 7:50am -0500,
> > gregkh@xxxxxxxxxxxxxxxxxxx <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> >
> > >
> > > The patch below does not apply to the 4.20-stable tree.
> > > If someone wants it applied there, or to any other stable or longterm
> > > tree, then please email the backport, including the original git commit
> > > id to <stable@xxxxxxxxxxxxxxx>.
> > >
> > > thanks,
> > >
> > > greg k-h
> > >
> > > ------------------ original commit in Linus's tree ------------------
> > >
> > > From a1e1cb72d96491277ede8d257ce6b48a381dd336 Mon Sep 17 00:00:00 2001
> > > From: Mike Snitzer <snitzer@xxxxxxxxxx>
> > > Date: Thu, 17 Jan 2019 10:48:01 -0500
> > > Subject: [PATCH] dm: fix redundant IO accounting for bios that need splitting
> > >
> > > The risk of redundant IO accounting was not taken into consideration
> > > when commit 18a25da84354 ("dm: ensure bio submission follows a
> > > depth-first tree walk") introduced IO splitting in terms of recursion
> > > via generic_make_request().
> > >
> > > Fix this by subtracting the split bio's payload from the IO stats that
> > > were already accounted for by start_io_acct() upon dm_make_request()
> > > entry. This repeat oscillation of the IO accounting, up then down,
> > > isn't ideal but refactoring DM core's IO splitting to pre-split bios
> > > _before_ they are accounted turned out to be an excessive amount of
> > > change that will need a full development cycle to refine and verify.
> > >
> > > Before this fix:
> > >
> > > /dev/mapper/stripe_dev is a 4-way stripe using a 32k chunksize, so
> > > bios are split on 32k boundaries.
> > >
> > > # fio --name=16M --filename=/dev/mapper/stripe_dev --rw=write --bs=64k --size=16M \
> > > --iodepth=1 --ioengine=libaio --direct=1 --refill_buffers
> > >
> > > with debugging added:
> > > [103898.310264] device-mapper: core: start_io_acct: dm-2 WRITE bio->bi_iter.bi_sector=0 len=128
> > > [103898.318704] device-mapper: core: __split_and_process_bio: recursing for following split bio:
> > > [103898.329136] device-mapper: core: start_io_acct: dm-2 WRITE bio->bi_iter.bi_sector=64 len=64
> > > ...
> > >
> > > 16M written yet 136M (278528 * 512b) accounted:
> > > # cat /sys/block/dm-2/stat | awk '{ print $7 }'
> > > 278528
> > >
> > > After this fix:
> > >
> > > 16M written and 16M (32768 * 512b) accounted:
> > > # cat /sys/block/dm-2/stat | awk '{ print $7 }'
> > > 32768
> > >
> > > Fixes: 18a25da84354 ("dm: ensure bio submission follows a depth-first tree walk")
> > > Cc: stable@xxxxxxxxxxxxxxx # 4.16+
> > > Reported-by: Bryan Gurney <bgurney@xxxxxxxxxx>
> > > Reviewed-by: Ming Lei <ming.lei@xxxxxxxxxx>
> > > Signed-off-by: Mike Snitzer <snitzer@xxxxxxxxxx>
> > >
> > > diff --git a/drivers/md/dm.c b/drivers/md/dm.c
> > > index fcb97b0a5743..fbadda68e23b 100644
> > > --- a/drivers/md/dm.c
> > > +++ b/drivers/md/dm.c
> > > @@ -1584,6 +1584,9 @@ static void init_clone_info(struct clone_info *ci, struct mapped_device *md,
> > > ci->sector = bio->bi_iter.bi_sector;
> > > }
> > >
> > > +#define __dm_part_stat_sub(part, field, subnd) \
> > > + (part_stat_get(part, field) -= (subnd))
> > > +
> > > /*
> > > * Entry point to split a bio into clones and submit them to the targets.
> > > */
> > > @@ -1638,6 +1641,19 @@ static blk_qc_t __split_and_process_bio(struct mapped_device *md,
> > > struct bio *b = bio_split(bio, bio_sectors(bio) - ci.sector_count,
> > > GFP_NOIO, &md->queue->bio_split);
> > > ci.io->orig_bio = b;
> > > +
> > > + /*
> > > + * Adjust IO stats for each split, otherwise upon queue
> > > + * reentry there will be redundant IO accounting.
> > > + * NOTE: this is a stop-gap fix, a proper fix involves
> > > + * significant refactoring of DM core's bio splitting
> > > + * (by eliminating DM's splitting and just using bio_split)
> > > + */
> > > + part_stat_lock();
> > > + __dm_part_stat_sub(&dm_disk(md)->part0,
> > > + sectors[op_stat_group(bio_op(bio))], ci.sector_count);
> > > + part_stat_unlock();
> > > +
> > > bio_chain(b, bio);
> > > ret = generic_make_request(bio);
> > > break;
> > >
> >
> > Seems to apply fine.. not sure what the problem is on your end:
> >
> > $ git checkout stable/linux-4.20.y
> > Previous HEAD position was 8fe28cb58bcb... Linux 4.20
> > HEAD is now at 9f1a389a0b5b... Linux 4.20.5
> >
> > $ git show a1e1cb72d96491277ede8d257ce6b48a381dd336 | patch -p1 --dry
> > patching file drivers/md/dm.c
> > Hunk #1 succeeded at 1578 (offset -6 lines).
> > Hunk #2 succeeded at 1626 (offset -15 lines).
> >
> > $ git cherry-pick a1e1cb72d96491277ede8d257ce6b48a381dd336
> > [detached HEAD 3d6015ea633a] dm: fix redundant IO accounting for bios that need splitting
> > Date: Thu Jan 17 10:48:01 2019 -0500
> > 1 file changed, 16 insertions(+)
>
> Try building it, it blows up into tiny pieces :)
>
> I guess I need a different script that says, "the patch applied, but
> broke the build", but it is so rare it's almost not worth it...
Ah, gotcha. Because of part_stat_get() it implicitly depends on commit
1226b8dd0e913 but that shouldn't go to stable.
stable@ would need to factor out __part_stat_sub(), like
__part_stat_add(), and part_stat_sub() updated to use __part_stat_sub().
As is, existing part_stat_sub() is broken on all kernels (and with no
callers nobody cares).
Now that I've woken the dragon (Jens) and told him I papered over block
core's broekn part_stat_sub() in DM.. I'll do whatever Jens wants me to
do ;)
Mike
p.s. since this bug has existed for 1.5 years maybe nobody cares that
DM's io stats are completely bogus and we can just ignore it? Yeah,
that is a cop-out...
