All of lore.kernel.org
 help / color / mirror / Atom feed
From: Mike Snitzer <snitzer@redhat.com>
To: dm-devel@redhat.com
Cc: ejt@redhat.com
Subject: Re: [PATCH 2/3] dm thin metadata: do not write metadata if no changes occurred
Date: Mon, 15 Apr 2019 22:39:03 -0400	[thread overview]
Message-ID: <20190416023903.GA19565@redhat.com> (raw)
In-Reply-To: <20190415211717.18737-3-snitzer@redhat.com>

On Mon, Apr 15 2019 at  5:17pm -0400,
Mike Snitzer <snitzer@redhat.com> wrote:

> Otherwise, just loading a thin-pool and then removing it will cause its
> metadata to be changed (e.g. superblock written) -- even without any
> thin devices being made active or metadata snapshot being manipulated.
> 
> Add 'changed' flag to struct dm_pool_metadata.  Set both td->changed and
> pmd->changed using __set_thin_changed() -- this assures pmd->changed is
> always set when td->changed is.  For methods that don't involve changing
> a td, set pmd->changed directly.
> 
> Rename __delete_device() to __delete_thin() -- improves symmetry with
> __create_thin().
> 
> Update dm_pool_changed_this_transaction() to simply return pmd->changed.
> 
> Fix dm_pool_commit_metadata() to open the next transaction if the return
> from __commit_transaction() is 0.  Not seeing why the early return ever
> made since for a return of 0 given that dm-io's async_io(), as used by
> bufio, always returns 0.

This v1 patch was incomplete.  However this approach's need to establish
pmd->changed everywhere that alters metadata, without doing so in terms
of a thin device, is like whack-a-mole.

But I think this could be simplified to only set a 'pmd->accessed'
(never clear it) and only__commit_transaction() if pmd->accessed --
so only pool messages that change metadata would need to set
pmd->accessed.

  reply	other threads:[~2019-04-16  2:39 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-04-15 21:17 [PATCH 0/3] dm thin metadata: handful of improvements Mike Snitzer
2019-04-15 21:17 ` [PATCH 1/3] dm space map common: zero entire ll_disk Mike Snitzer
2019-04-15 21:17 ` [PATCH 2/3] dm thin metadata: do not write metadata if no changes occurred Mike Snitzer
2019-04-16  2:39   ` Mike Snitzer [this message]
2019-04-17  1:30     ` [PATCH v2 " Mike Snitzer
2019-04-15 21:17 ` [PATCH 3/3] dm thin metadata: check __commit_transaction()'s return Mike Snitzer

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=20190416023903.GA19565@redhat.com \
    --to=snitzer@redhat.com \
    --cc=dm-devel@redhat.com \
    --cc=ejt@redhat.com \
    /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.