Re: [PATCH] xfs: get rid of xfs_ag_resv_rmapbt_alloc

[Long Li <leo.lilong@huawei.com>]

On Tue, Jul 02, 2024 at 10:14:46PM -0700, Darrick J. Wong wrote:
> On Tue, Jul 02, 2024 at 09:48:51PM +0800, Long Li wrote:
> > The pag in xfs_ag_resv_rmapbt_alloc() is already held when the struct
> > xfs_btree_cur is initialized in xfs_rmapbt_init_cursor(), so there is no
> > need to get pag again.
> > 
> > On the other hand, in xfs_rmapbt_free_block(), the similar function
> > xfs_ag_resv_rmapbt_free() was removed in commit 92a005448f6f ("xfs: get
> > rid of unnecessary xfs_perag_{get,put} pairs"), xfs_ag_resv_rmapbt_alloc()
> > was left because scrub used it, but now scrub has removed it. Therefore,
> > we could get rid of xfs_ag_resv_rmapbt_alloc() just like the rmap free
> > block, make the code cleaner.
> > 
> > Signed-off-by: Long Li <leo.lilong@huawei.com>
> > ---
> >  fs/xfs/libxfs/xfs_ag_resv.h    | 19 -------------------
> >  fs/xfs/libxfs/xfs_rmap_btree.c |  8 +++++++-
> >  2 files changed, 7 insertions(+), 20 deletions(-)
> > 
> > diff --git a/fs/xfs/libxfs/xfs_ag_resv.h b/fs/xfs/libxfs/xfs_ag_resv.h
> > index ff20ed93de77..f247eeff7358 100644
> > --- a/fs/xfs/libxfs/xfs_ag_resv.h
> > +++ b/fs/xfs/libxfs/xfs_ag_resv.h
> > @@ -33,23 +33,4 @@ xfs_perag_resv(
> >  	}
> >  }
> >  
> > -/*
> > - * RMAPBT reservation accounting wrappers. Since rmapbt blocks are sourced from
> > - * the AGFL, they are allocated one at a time and the reservation updates don't
> > - * require a transaction.
> > - */
> > -static inline void
> > -xfs_ag_resv_rmapbt_alloc(
> > -	struct xfs_mount	*mp,
> > -	xfs_agnumber_t		agno)
> > -{
> > -	struct xfs_alloc_arg	args = { NULL };
> > -	struct xfs_perag	*pag;
> > -
> > -	args.len = 1;
> > -	pag = xfs_perag_get(mp, agno);
> > -	xfs_ag_resv_alloc_extent(pag, XFS_AG_RESV_RMAPBT, &args);
> > -	xfs_perag_put(pag);
> > -}
> > -
> >  #endif	/* __XFS_AG_RESV_H__ */
> > diff --git a/fs/xfs/libxfs/xfs_rmap_btree.c b/fs/xfs/libxfs/xfs_rmap_btree.c
> > index 9e759efa81cc..aa1d29814b74 100644
> > --- a/fs/xfs/libxfs/xfs_rmap_btree.c
> > +++ b/fs/xfs/libxfs/xfs_rmap_btree.c
> > @@ -88,6 +88,7 @@ xfs_rmapbt_alloc_block(
> >  	struct xfs_buf		*agbp = cur->bc_ag.agbp;
> >  	struct xfs_agf		*agf = agbp->b_addr;
> >  	struct xfs_perag	*pag = cur->bc_ag.pag;
> > +	struct xfs_alloc_arg    args = { NULL };
> 
> You could make this even more compact with
> 
> 	struct xfs_alloc_arg	args = { .len = 1 };

It's ok for me, I will send a new version. thanks!

> 
> Otherwise this looks ok to me as a cleanup.
> 
> --D
> 
> >  	int			error;
> >  	xfs_agblock_t		bno;
> >  
> > @@ -107,7 +108,12 @@ xfs_rmapbt_alloc_block(
> >  	be32_add_cpu(&agf->agf_rmap_blocks, 1);
> >  	xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_RMAP_BLOCKS);
> >  
> > -	xfs_ag_resv_rmapbt_alloc(cur->bc_mp, pag->pag_agno);
> > +	/*
> > +	 * Since rmapbt blocks are sourced from the AGFL, they are allocated one
> > +	 * at a time and the reservation updates don't require a transaction.
> > +	 */
> > +	args.len = 1;
> > +	xfs_ag_resv_alloc_extent(pag, XFS_AG_RESV_RMAPBT, &args);
> >  
> >  	*stat = 1;
> >  	return 0;
> > -- 
> > 2.39.2
> > 
> >