[dm-devel] dm-table: set non-zero q->limits.discard_granularity

[Mikulas Patocka]

On Thu, 24 Sep 2020, Mike Snitzer wrote:

> On Thu, Sep 24 2020 at 10:34am -0400,
> Mikulas Patocka <mpatocka at redhat.com> wrote:
> 
> > Hi Mike
> > 
> > Could you send this to Linus before v5.9 is released?
> > 
> > Mikulas
> > 
> > 
> > 
> > From: Mikulas Patocka <mpatocka at redhat.com>
> > 
> > If q->limits.discard_granularity is zero, the block core will warn in
> > __blkdev_issue_discard. This patch sets it to a minimum - 512 bytes.
> > 
> > Signed-off-by: Mikulas Patocka <mpatocka at redhat.com>
> > Fixes: b35fd7422c2f ("block: check queue's limits.discard_granularity in __blkdev_issue_discard()")
> > Fixes: 9b15d109a6b2 ("block: improve discard bio alignment in __blkdev_issue_discard()")
> > Fixes: c52abf563049 ("loop: Better discard support for block devices")
> > 
> > ---
> >  drivers/md/dm-table.c |    5 ++++-
> >  1 file changed, 4 insertions(+), 1 deletion(-)
> > 
> > Index: linux-2.6/drivers/md/dm-table.c
> > ===================================================================
> > --- linux-2.6.orig/drivers/md/dm-table.c	2020-08-31 12:08:27.000000000 +0200
> > +++ linux-2.6/drivers/md/dm-table.c	2020-09-24 16:23:55.000000000 +0200
> > @@ -1858,8 +1858,11 @@ void dm_table_set_restrictions(struct dm
> >  		q->limits.discard_granularity = 0;
> >  		q->limits.discard_alignment = 0;
> >  		q->limits.discard_misaligned = 0;
> > -	} else
> > +	} else {
> >  		blk_queue_flag_set(QUEUE_FLAG_DISCARD, q);
> > +		if (q->limits.discard_granularity < 512)
> > +			q->limits.discard_granularity = 512;
> > +	}
> >  
> >  	if (dm_table_supports_secure_erase(t))
> >  		blk_queue_flag_set(QUEUE_FLAG_SECERASE, q);
> 
> This patch is wrong, papers over something else.  Please identify what
> that something else is.
> 
> block core has _always_ cheecked if discard_granularity is non-zero to
> allow discards.
> 
> Why is q->limits.discard_granularity < 512 if discard is supported by
> the DM table and targets?
> 
> That shouldn't ever be.
> 
> Mike

OK

I'm sending another patch that changes the function raid_io_hints. This is 
where the zero came from.

Mikulas