Re: [PATCH 5/5] xfs: xfs_defer_capture should absorb remaining transaction reservation

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Thu, Oct 01, 2020 at 01:32:56PM -0400, Brian Foster wrote:
> On Tue, Sep 29, 2020 at 10:43:44AM -0700, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@xxxxxxxxxx>
> > 
> > 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@xxxxxxxxxx>
> > ---
> 
> 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.

Ok, will do.

--D

> 
> 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;
> >  
> > 
> 



[Index of Archives]     [XFS Filesystem Development (older mail)]     [Linux Filesystem Development]     [Linux Audio Users]     [Yosemite Trails]     [Linux Kernel]     [Linux RAID]     [Linux SCSI]


  Powered by Linux