public inbox for linux-xfs@vger.kernel.org
 help / color / mirror / Atom feed
From: Ben Myers <bpm@sgi.com>
To: Christoph Hellwig <hch@infradead.org>
Cc: xfs@oss.sgi.com
Subject: Re: [PATCH 06/11] xfs: replace i_flock with a sleeping bitlock
Date: Fri, 13 Jan 2012 15:49:41 -0600	[thread overview]
Message-ID: <20120113214941.GA16386@sgi.com> (raw)
In-Reply-To: <20111218200131.745039484@bombadil.infradead.org>

On Sun, Dec 18, 2011 at 03:00:09PM -0500, Christoph Hellwig wrote:
> Index: xfs/fs/xfs/xfs_inode.c
> ===================================================================
> --- xfs.orig/fs/xfs/xfs_inode.c	2011-12-18 08:05:01.469974560 -0800
> +++ xfs/fs/xfs/xfs_inode.c	2011-12-18 08:06:23.266640737 -0800

...

> -/*
>   * In-core inode flags.
>   */
> -#define XFS_IRECLAIM		0x0001  /* started reclaiming this inode */
> -#define XFS_ISTALE		0x0002	/* inode has been staled */
> -#define XFS_IRECLAIMABLE	0x0004	/* inode can be reclaimed */
> -#define XFS_INEW		0x0008	/* inode has just been allocated */
> -#define XFS_IFILESTREAM		0x0010	/* inode is in a filestream directory */
> -#define XFS_ITRUNCATED		0x0020	/* truncated down so flush-on-close */
> -#define XFS_IDIRTY_RELEASE	0x0040	/* dirty release already seen */
> +#define XFS_IRECLAIM		(1 << 0) /* started reclaiming this inode */
> +#define XFS_ISTALE		(1 << 1) /* inode has been staled */
> +#define XFS_IRECLAIMABLE	(1 << 2) /* inode can be reclaimed */
> +#define XFS_INEW		(1 << 3) /* inode has just been allocated */
> +#define XFS_IFILESTREAM		(1 << 4) /* inode is in a filestream dir. */
> +#define XFS_ITRUNCATED		(1 << 5) /* truncated down so flush-on-close */
> +#define XFS_IDIRTY_RELEASE	(1 << 6) /* dirty release already seen */
> +#define __XFS_IFLOCK_BIT	7	 /* inode is being flushed right now */
> +#define XFS_IFLOCK		(1 << __XFS_IFLOCK_BIT)

Nice.

> +static inline void xfs_iflock(struct xfs_inode *ip)
> +{
> +	if (!xfs_iflock_nowait(ip))
> +		__xfs_iflock(ip);
> +}

Going after the iflock nowait first seems a little silly, but I'm
guessing that you're trying to avoid touching the zone wait_table in
bit_waitqueue if you can avoid it, along with the rest of the overhead
related to seeting up the wait queue in __xfs_iflock.

> +
> +static inline void xfs_ifunlock(struct xfs_inode *ip)
> +{
> +	xfs_iflags_clear(ip, XFS_IFLOCK);
> +	wake_up_bit(&ip->i_flags, __XFS_IFLOCK_BIT);
> +}

I was going to suggest this:

	spin_lock(&ip->i_flags_lock); 
	ip->i_flags &= ~XFS_IFLOCK;
	wake_up_bit(... 
	spin_unlock(&ip->i_flags_lock);

After some study I believe what you have is ok:

Say this is process A.  If process B got the lock after A cleared IFLOCK
and before A wake_up_bit wakes process C.  C will just go back to sleep
in __xfs_iflock until B calls xfs_ifunlock to wake C again.  

Looks good.

Reviewed-by: Ben Myers <bpm@sgi.com>

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

  reply	other threads:[~2012-01-13 21:49 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-12-18 20:00 [PATCH 00/11] inode shrink and misc updates V2 Christoph Hellwig
2011-12-18 20:00 ` [PATCH 01/11] xfs: remove xfs_itruncate_data Christoph Hellwig
2012-01-03 21:53   ` Ben Myers
2012-01-04  9:27     ` Christoph Hellwig
2011-12-18 20:00 ` [PATCH 02/11] xfs: cleanup xfs_iomap_eof_align_last_fsb Christoph Hellwig
2012-01-04 20:32   ` Ben Myers
2011-12-18 20:00 ` [PATCH 03/11] xfs: remove the unused dm_attrs structure Christoph Hellwig
2012-01-04 21:13   ` Ben Myers
2011-12-18 20:00 ` [PATCH 04/11] xfs: remove the if_ext_max field in struct xfs_ifork Christoph Hellwig
2012-01-06 16:58   ` Ben Myers
2012-01-16 22:45     ` Ben Myers
2012-01-17 15:16       ` Ben Myers
2012-01-17 17:04         ` Mark Tinguely
2011-12-18 20:00 ` [PATCH 05/11] xfs: make i_flags an unsigned long Christoph Hellwig
2011-12-18 20:00 ` [PATCH 06/11] xfs: replace i_flock with a sleeping bitlock Christoph Hellwig
2012-01-13 21:49   ` Ben Myers [this message]
2011-12-18 20:00 ` [PATCH 07/11] xfs: replace i_pin_wait with a bit waitqueue Christoph Hellwig
2012-01-13 22:42   ` Ben Myers
2011-12-18 20:00 ` [PATCH 08/11] xfs: remove the i_size field in struct xfs_inode Christoph Hellwig
2012-01-16 18:32   ` Ben Myers
2012-01-16 19:45     ` Ben Myers
2011-12-18 20:00 ` [PATCH 09/11] xfs: remove the i_new_size " Christoph Hellwig
2011-12-18 22:13   ` Dave Chinner
2012-01-16 22:41   ` Ben Myers
2012-01-17 20:14   ` Ben Myers
2011-12-18 20:00 ` [PATCH 10/11] xfs: always return with the iolock held from xfs_file_aio_write_checks Christoph Hellwig
2012-01-17 20:18   ` Ben Myers
2012-01-20 12:51     ` Jeff Liu
2011-12-18 20:00 ` [PATCH 11/11] xfs: cleanup xfs_file_aio_write Christoph Hellwig
2012-01-17 20:42   ` Ben Myers
  -- strict thread matches above, loose matches on Subject: below --
2011-12-08 15:57 [PATCH 00/11] inode shrink and misc updates Christoph Hellwig
2011-12-08 15:58 ` [PATCH 06/11] xfs: replace i_flock with a sleeping bitlock Christoph Hellwig
2011-12-13 22:19   ` Dave Chinner
2011-12-18 18:11     ` Christoph Hellwig

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=20120113214941.GA16386@sgi.com \
    --to=bpm@sgi.com \
    --cc=hch@infradead.org \
    --cc=xfs@oss.sgi.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox