All of lore.kernel.org
 help / color / mirror / Atom feed
From: Mike Snitzer <snitzer@redhat.com>
To: Christoph Hellwig <hch@lst.de>
Cc: Jens Axboe <axboe@fb.com>, dm-devel@redhat.com
Subject: Re: block, dm: don't copy bios for request clones
Date: Fri, 24 Apr 2015 20:15:24 -0400	[thread overview]
Message-ID: <20150425001524.GB18601@redhat.com> (raw)
In-Reply-To: <1429904020-16363-1-git-send-email-hch@lst.de>

On Fri, Apr 24 2015 at  3:33pm -0400,
Christoph Hellwig <hch@lst.de> wrote:

> Currently dm-multipath has to clone the bios for every request sent
> to the lower devices, which wastes cpu cycles and ties down memory.
> 
> This patch instead adds a new REQ_CLONE flag that instructs req_bio_endio
> to not complete bios attached to a request, which we set on clone
> requests similar to bios in a flush sequence.  With this change I/O
> errors on a path failure only get propagated to dm-multipath, which
> can then either resubmit the I/O or complete the bios on the original
> request.
> 
> I've done some basic testing of this on a Linux target with ALUA support,
> and it survives path failures during I/O nicely.

Thanks for working on this (it slipped in priority on my TODO list).
Will be great to get this reviewed/tested/staged for 4.2.

But your patch needs rebasing against latest upstream code.  I'd imagine
this means you also haven't tested with "use_blk_mq" enabled (from
commit 17e149b8f).

I'll do a much more careful review once you rebase but I noticed one
thing worth mentioning now:

> diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c
> index 6554d91..4ac0a47 100644
> --- a/drivers/md/dm-table.c
> +++ b/drivers/md/dm-table.c
> @@ -941,23 +941,30 @@ bool dm_table_mq_request_based(struct dm_table *t)
>  
>  static int dm_table_alloc_md_mempools(struct dm_table *t)
>  {
> -	unsigned type = dm_table_get_type(t);
> +	int type = dm_table_get_type(t);
>  	unsigned per_bio_data_size = 0;
> -	struct dm_target *tgt;
>  	unsigned i;
>  
> -	if (unlikely(type == DM_TYPE_NONE)) {
> +	switch (type) {

You're missing a case statement here for DM_TYPE_BIO_BASED:

> +		for (i = 0; i < t->num_targets; i++) {
> +			struct dm_target *tgt = t->targets + i;
> +
> +			per_bio_data_size = max(per_bio_data_size,
> +						tgt->per_bio_data_size);
> +		}
> +
> +		t->mempools = dm_alloc_bio_mempools(t->integrity_supported,
> +				per_bio_data_size);
> +		break;
> +	case DM_TYPE_REQUEST_BASED:
> +	case DM_TYPE_MQ_REQUEST_BASED:
> +		t->mempools = dm_alloc_rq_mempools(type);
> +		break;
> +	default:
>  		DMWARN("no table type is set, can't allocate mempools");
>  		return -EINVAL;
>  	}
>  
> -	if (type == DM_TYPE_BIO_BASED)
> -		for (i = 0; i < t->num_targets; i++) {
> -			tgt = t->targets + i;
> -			per_bio_data_size = max(per_bio_data_size, tgt->per_bio_data_size);
> -		}
> -
> -	t->mempools = dm_alloc_md_mempools(type, t->integrity_supported, per_bio_data_size);
>  	if (!t->mempools)
>  		return -ENOMEM;
>  

      reply	other threads:[~2015-04-25  0:15 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-04-24 19:33 [PATCH] block, dm: don't copy bios for request clones Christoph Hellwig
2015-04-25  0:15 ` Mike Snitzer [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20150425001524.GB18601@redhat.com \
    --to=snitzer@redhat.com \
    --cc=axboe@fb.com \
    --cc=dm-devel@redhat.com \
    --cc=hch@lst.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.