Re: [PATCH 05/14] xfs: per-cpu deferred inode inactivation queues

From: "Darrick J. Wong" <djwong@kernel.org>
To: Dave Chinner <david@fromorbit.com>
Cc: Dave Chinner <dchinner@redhat.com>,
	linux-xfs@vger.kernel.org, hch@infradead.org
Subject: Re: [PATCH 05/14] xfs: per-cpu deferred inode inactivation queues
Date: Mon, 9 Aug 2021 16:36:01 -0700	[thread overview]
Message-ID: <20210809233601.GR3601466@magnolia> (raw)
In-Reply-To: <20210807214900.GB3657114@dread.disaster.area>

On Sun, Aug 08, 2021 at 07:49:00AM +1000, Dave Chinner wrote:
> On Fri, Aug 06, 2021 at 05:21:04PM -0700, Darrick J. Wong wrote:
> > On Wed, Aug 04, 2021 at 07:06:50PM -0700, Darrick J. Wong wrote:
> > > From: Dave Chinner <dchinner@redhat.com>
> > 
> > <megasnip> A couple of minor changes that aren't worth reposting the
> > entire series:
> > 
> > > diff --git a/fs/xfs/xfs_icache.c b/fs/xfs/xfs_icache.c
> > > index b9214733d0c3..fedfa40e3cd6 100644
> > > --- a/fs/xfs/xfs_icache.c
> > > +++ b/fs/xfs/xfs_icache.c
> > 
> > <snip>
> > 
> > > @@ -1767,30 +1801,276 @@ xfs_inode_mark_reclaimable(
> > >  		ASSERT(0);
> > >  	}
> > >  
> > > +	pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino));
> > > +	spin_lock(&pag->pag_ici_lock);
> > > +	spin_lock(&ip->i_flags_lock);
> > > +
> > > +	trace_xfs_inode_set_reclaimable(ip);
> > > +	ip->i_flags &= ~(XFS_NEED_INACTIVE | XFS_INACTIVATING);
> > > +	ip->i_flags |= XFS_IRECLAIMABLE;
> > > +	xfs_perag_set_inode_tag(pag, XFS_INO_TO_AGINO(mp, ip->i_ino),
> > > +			XFS_ICI_RECLAIM_TAG);
> > > +
> > > +	spin_unlock(&ip->i_flags_lock);
> > > +	spin_unlock(&pag->pag_ici_lock);
> > > +	xfs_perag_put(pag);
> > > +}
> > > +
> > > +/*
> > > + * Free all speculative preallocations and possibly even the inode itself.
> > > + * This is the last chance to make changes to an otherwise unreferenced file
> > > + * before incore reclamation happens.
> > > + */
> > > +static void
> > > +xfs_inodegc_inactivate(
> > > +	struct xfs_inode	*ip)
> > > +{
> > > +	struct xfs_mount        *mp = ip->i_mount;
> > > +
> > > +	/*
> > > +	* Inactivation isn't supposed to run when the fs is frozen because
> > > +	* we don't want kernel threads to block on transaction allocation.
> > > +	*/
> > > +	ASSERT(mp->m_super->s_writers.frozen < SB_FREEZE_FS);
> > > +
> > 
> > I solved the problems Dave was complaining about (g/390, x/517) by
> > removing this ASSERT.
> > 
> > > diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
> > > index 19260291ff8b..bd8abb50b33a 100644
> > > --- a/fs/xfs/xfs_trace.h
> > > +++ b/fs/xfs/xfs_trace.h
> > > @@ -157,6 +157,48 @@ DEFINE_PERAG_REF_EVENT(xfs_perag_put);
> > >  DEFINE_PERAG_REF_EVENT(xfs_perag_set_inode_tag);
> > >  DEFINE_PERAG_REF_EVENT(xfs_perag_clear_inode_tag);
> > >  
> > > +#define XFS_STATE_FLAGS \
> > > +	{ (1UL << XFS_STATE_INODEGC_ENABLED),		"inodegc" }
> > 
> > I've also changed the name of this to XFS_OPSTATE_STRINGS because we use
> > _STRINGS everywhere else in this file.
> 
> FWIW, can we define this with the definition of the OPSTATE
> variables in xfs_mount.h? THat makes it much easier to keep up to
> date when we add new flags because it's obvious that there are
> tracing flags that also need to be updated when we add a new state
> flag...

Yeah, I'll move the _STRINGS definition and clean up all these names to
have 'OPSTATE' instead of just 'STATE' too.

--D

> 
> Cheers,
> 
> Dave.
> -- 
> Dave Chinner
> david@fromorbit.com