Linux-Fsdevel Archive on lore.kernel.org
 help / color / Atom feed
From: Christoph Hellwig <hch@infradead.org>
To: Luis Chamberlain <mcgrof@kernel.org>
Cc: rafael@kernel.org, gregkh@linuxfoundation.org,
	viro@zeniv.linux.org.uk, jack@suse.cz, bvanassche@acm.org,
	jeyu@kernel.org, ebiederm@xmission.com, mchehab@kernel.org,
	keescook@chromium.org, linux-fsdevel@vger.kernel.org,
	kernel@tuxforce.de, kexec@lists.infradead.org,
	linux-kernel@vger.kernel.org
Subject: Re: [RFC v2 6/6] fs: add automatic kernel fs freeze / thaw and remove kthread freezing
Date: Tue, 20 Apr 2021 13:59:03 +0100
Message-ID: <20210420125903.GC3604224@infradead.org> (raw)
In-Reply-To: <20210417001026.23858-7-mcgrof@kernel.org>

> This also removes all the superflous freezer calls on all filesystems
> as they are no longer needed as the VFS now performs filesystem
> freezing/thaw if the filesystem has support for it. The filesystem
> therefore is in charge of properly dealing with quiescing of the
> filesystem through its callbacks.

Can you split that out from the main logic change?  Maybe even into one
patch per file system?

> +#ifdef CONFIG_PM_SLEEP
> +static bool super_should_freeze(struct super_block *sb)
> +{
> +	if (!sb->s_root)
> +		return false;
> +	if (!(sb->s_flags & MS_BORN))
> +		return false;

This is already done in the iterate_supers_excl and
iterate_supers_reverse_excl helpers that this helper is always called
through.

> +	/*
> +	 * We don't freeze virtual filesystems, we skip those filesystems with
> +	 * no backing device.
> +	 */
> +	if (sb->s_bdi == &noop_backing_dev_info)
> +		return false;

Why?

> +	/* No need to freeze read-only filesystems */
> +	if (sb_rdonly(sb))
> +		return false;

freeze_super/thaw_super already takes care of read-only file systems,
and IMHO in a better way.

> +	int error = 0;
> +
> +	spin_lock(&sb_lock);
> +	if (!super_should_freeze(sb))
> +		goto out;
> +
> +	pr_info("%s (%s): freezing\n", sb->s_type->name, sb->s_id);
> +
> +	spin_unlock(&sb_lock);

I don't see how super_should_freeze needs sb_lock.  But if it does
the lock should be taken in the function.

> +	atomic_inc(&sb->s_active);

Doesn't this need a atomic_inc_not_zero if we're racing with a delayed
unmount?

> +	error = freeze_locked_super(sb, false);
> +	if (error)
> +		atomic_dec(&sb->s_active);

And this really needs something like deactivate_locked_super.

> +	spin_lock(&sb_lock);
> +	if (error && error != -EBUSY)
> +		pr_notice("%s (%s): Unable to freeze, error=%d",
> +			  sb->s_type->name, sb->s_id, error);
> +
> +out:
> +	spin_unlock(&sb_lock);

Huh, what is the point of sb_lock here?

> +int fs_suspend_freeze(void)
> +{
> +	return iterate_supers_reverse_excl(fs_suspend_freeze_sb, NULL);
> +}

I'd just fold this helper into its only caller.

> +	error = __thaw_super_locked(sb, false);
> +	if (!error)
> +		atomic_dec(&sb->s_active);
> +
> +	spin_lock(&sb_lock);
> +	if (error && error != -EBUSY)
> +		pr_notice("%s (%s): Unable to unfreeze, error=%d",
> +			  sb->s_type->name, sb->s_id, error);
> +
> +out:
> +	spin_unlock(&sb_lock);
> +	return error;
> +}
> +
> +int fs_resume_unfreeze(void)
> +{
> +	return iterate_supers_excl(fs_suspend_thaw_sb, NULL);
> +}

Same comments as on the freeze side.

  reply index

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-04-17  0:10 [RFC v2 0/6] vfs: provide automatic kernel freeze / resume Luis Chamberlain
2021-04-17  0:10 ` [RFC v2 1/6] fs: provide unlocked helper for freeze_super() Luis Chamberlain
2021-04-20 12:03   ` Christoph Hellwig
2021-04-17  0:10 ` [RFC v2 2/6] fs: add frozen sb state helpers Luis Chamberlain
2021-04-17  0:10 ` [RFC v2 3/6] fs: add a helper for thaw_super_locked() which does not unlock Luis Chamberlain
2021-04-17  0:10 ` [RFC v2 4/6] fs: distinguish between user initiated freeze and kernel initiated freeze Luis Chamberlain
2021-04-20 12:46   ` Christoph Hellwig
2021-04-17  0:10 ` [RFC v2 5/6] fs: add iterate_supers_excl() and iterate_supers_reverse_excl() Luis Chamberlain
2021-04-17  0:10 ` [RFC v2 6/6] fs: add automatic kernel fs freeze / thaw and remove kthread freezing Luis Chamberlain
2021-04-20 12:59   ` Christoph Hellwig [this message]
2021-04-20 18:47     ` Luis Chamberlain

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=20210420125903.GC3604224@infradead.org \
    --to=hch@infradead.org \
    --cc=bvanassche@acm.org \
    --cc=ebiederm@xmission.com \
    --cc=gregkh@linuxfoundation.org \
    --cc=jack@suse.cz \
    --cc=jeyu@kernel.org \
    --cc=keescook@chromium.org \
    --cc=kernel@tuxforce.de \
    --cc=kexec@lists.infradead.org \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mcgrof@kernel.org \
    --cc=mchehab@kernel.org \
    --cc=rafael@kernel.org \
    --cc=viro@zeniv.linux.org.uk \
    /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

Linux-Fsdevel Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-fsdevel/0 linux-fsdevel/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-fsdevel linux-fsdevel/ https://lore.kernel.org/linux-fsdevel \
		linux-fsdevel@vger.kernel.org
	public-inbox-index linux-fsdevel

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-fsdevel


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git