public inbox for linux-xfs@vger.kernel.org
 help / color / mirror / Atom feed
From: Brian Foster <bfoster@redhat.com>
To: "Darrick J. Wong" <darrick.wong@oracle.com>
Cc: linux-xfs@vger.kernel.org, david@fromorbit.com, hch@lst.de
Subject: Re: [PATCH 5/5] xfs: xfs_defer_capture should absorb remaining transaction reservation
Date: Thu, 1 Oct 2020 13:32:56 -0400	[thread overview]
Message-ID: <20201001173256.GG112884@bfoster> (raw)
In-Reply-To: <160140142459.830233.7194402837807253154.stgit@magnolia>

On Tue, Sep 29, 2020 at 10:43:44AM -0700, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
> 
> When xfs_defer_capture extracts the deferred ops and transaction state
> from a transaction, it should record the transaction reservation type
> from the old transaction so that when we continue the dfops chain, we
> still use the same reservation parameters.
> 
> Doing this means that the log item recovery functions get to determine
> the transaction reservation instead of abusing tr_itruncate in yet
> another part of xfs.
> 
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> ---

Much nicer, and FWIW this is pretty much the approach I was wondering
about wrt to the block reservation in the previous patch..

>  fs/xfs/libxfs/xfs_defer.c |    9 +++++++++
>  fs/xfs/libxfs/xfs_defer.h |    1 +
>  fs/xfs/xfs_log_recover.c  |    4 ++--
>  3 files changed, 12 insertions(+), 2 deletions(-)
> 
> 
> diff --git a/fs/xfs/libxfs/xfs_defer.c b/fs/xfs/libxfs/xfs_defer.c
> index 0cceebb390c4..4caaf5527403 100644
> --- a/fs/xfs/libxfs/xfs_defer.c
> +++ b/fs/xfs/libxfs/xfs_defer.c
> @@ -579,6 +579,15 @@ xfs_defer_ops_capture(
>  	dfc->dfc_blkres = tp->t_blk_res - tp->t_blk_res_used;
>  	tp->t_blk_res = tp->t_blk_res_used;
>  
> +	/*
> +	 * Preserve the transaction reservation type.  The logcount is
> +	 * hardwired to 1 to so that we can make forward progress in recovery
> +	 * no matter how full the log might be, at a cost of more regrants.
> +	 */
> +	dfc->dfc_tres.tr_logres = tp->t_log_res;
> +	dfc->dfc_tres.tr_logcount = 1;
> +	dfc->dfc_tres.tr_logflags = XFS_TRANS_PERM_LOG_RES;

Any real need to allocate these last two fields in every captured chain
when they're basically hardcoded? If not, it might be a bit more
efficient to put an xfs_trans_res on the stack in
xlog_finish_defer_ops() and just save the logres value here.

Brian

> +
>  	return dfc;
>  }
>  
> diff --git a/fs/xfs/libxfs/xfs_defer.h b/fs/xfs/libxfs/xfs_defer.h
> index b1c7b761afd5..c447c79bbe74 100644
> --- a/fs/xfs/libxfs/xfs_defer.h
> +++ b/fs/xfs/libxfs/xfs_defer.h
> @@ -76,6 +76,7 @@ struct xfs_defer_capture {
>  	struct list_head	dfc_dfops;
>  	unsigned int		dfc_tpflags;
>  	unsigned int		dfc_blkres;
> +	struct xfs_trans_res	dfc_tres;
>  };
>  
>  /*
> diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
> index b06c9881a13d..46e750279634 100644
> --- a/fs/xfs/xfs_log_recover.c
> +++ b/fs/xfs/xfs_log_recover.c
> @@ -2442,8 +2442,8 @@ xlog_finish_defer_ops(
>  	int			error = 0;
>  
>  	list_for_each_entry_safe(dfc, next, capture_list, dfc_list) {
> -		error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0,
> -				0, XFS_TRANS_RESERVE, &tp);
> +		error = xfs_trans_alloc(mp, &dfc->dfc_tres, 0, 0,
> +				XFS_TRANS_RESERVE, &tp);
>  		if (error)
>  			return error;
>  
> 


  reply	other threads:[~2020-10-01 17:33 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-09-29 17:43 [PATCH v4 0/5] xfs: fix how we deal with new intents during recovery Darrick J. Wong
2020-09-29 17:43 ` [PATCH 1/5] xfs: remove xfs_defer_reset Darrick J. Wong
2020-10-01 17:30   ` Brian Foster
2020-09-29 17:43 ` [PATCH 2/5] xfs: remove XFS_LI_RECOVERED Darrick J. Wong
2020-10-01 17:30   ` Brian Foster
2020-10-02  7:16   ` Christoph Hellwig
2020-09-29 17:43 ` [PATCH 3/5] xfs: proper replay of deferred ops queued during log recovery Darrick J. Wong
2020-10-01 17:31   ` Brian Foster
2020-10-01 17:41     ` Darrick J. Wong
2020-10-02  7:18   ` Christoph Hellwig
2020-09-29 17:43 ` [PATCH 4/5] xfs: xfs_defer_capture should absorb remaining block reservation Darrick J. Wong
2020-10-01 17:32   ` Brian Foster
2020-10-01 17:46     ` Darrick J. Wong
2020-10-02  4:20   ` [PATCH v4.2 " Darrick J. Wong
2020-10-02  7:22     ` Christoph Hellwig
2020-10-02 10:39     ` Brian Foster
2020-09-29 17:43 ` [PATCH 5/5] xfs: xfs_defer_capture should absorb remaining transaction reservation Darrick J. Wong
2020-10-01 17:32   ` Brian Foster [this message]
2020-10-01 17:52     ` Darrick J. Wong
2020-10-02  4:21   ` [PATCH v4.2 " Darrick J. Wong
2020-10-02  7:24     ` Christoph Hellwig
2020-10-02  7:32       ` Christoph Hellwig
2020-10-02 10:39     ` Brian Foster
  -- strict thread matches above, loose matches on Subject: below --
2020-10-05 18:19 [PATCH v5 0/5] xfs: fix how we deal with new intents during recovery Darrick J. Wong
2020-10-05 18:20 ` [PATCH 5/5] xfs: xfs_defer_capture should absorb remaining transaction reservation Darrick J. Wong
2020-10-06  6:22   ` 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=20201001173256.GG112884@bfoster \
    --to=bfoster@redhat.com \
    --cc=darrick.wong@oracle.com \
    --cc=david@fromorbit.com \
    --cc=hch@lst.de \
    --cc=linux-xfs@vger.kernel.org \
    /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