[21/34] xfs: devirtualize ->sf_get_ino and ->sf_put_ino

Message ID	20191101220719.29100-22-hch@lst.de (mailing list archive)
State	Superseded, archived
Headers	show Return-Path: <SRS0=5kOL=YZ=vger.kernel.org=linux-xfs-owner@kernel.org> From: Christoph Hellwig <hch@lst.de> To: linux-xfs@vger.kernel.org Subject: [PATCH 21/34] xfs: devirtualize ->sf_get_ino and ->sf_put_ino Date: Fri, 1 Nov 2019 15:07:06 -0700 Message-Id: <20191101220719.29100-22-hch@lst.de> In-Reply-To: <20191101220719.29100-1-hch@lst.de> References: <20191101220719.29100-1-hch@lst.de> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Sender: linux-xfs-owner@vger.kernel.org Precedence: bulk
Series	[01/34] xfs: move incore structures out of xfs_da_format.h \| expand [01/34] xfs: move incore structures out of xfs_da_format.h [02/34] xfs: refactor btree node scrubbing [03/34] xfs: devirtualize ->node_hdr_from_disk [04/34] xfs: devirtualize ->node_hdr_to_disk [05/34] xfs: add a btree entries pointer to struct xfs_da3_icnode_hdr [06/34] xfs: move the node header size to struct xfs_da_geometry [07/34] xfs: devirtualize ->leaf_hdr_from_disk [08/34] xfs: devirtualize ->leaf_hdr_to_disk [09/34] xfs: add a entries pointer to struct xfs_dir3_icleaf_hdr [10/34] xfs: move the dir2 leaf header size to struct xfs_da_geometry [11/34] xfs: move the max dir2 leaf entries count to struct xfs_da_geometry [12/34] xfs: devirtualize ->free_hdr_from_disk [13/34] xfs: devirtualize ->free_hdr_to_disk [14/34] xfs: make the xfs_dir3_icfree_hdr available to xfs_dir2_node_addname_int [15/34] xfs: add a bests pointer to struct xfs_dir3_icfree_hdr [16/34] xfs: move the dir2 free header size to struct xfs_da_geometry [17/34] xfs: move the max dir2 free bests count to struct xfs_da_geometry [18/34] xfs: devirtualize ->db_to_fdb and ->db_to_fdindex [19/34] xfs: devirtualize ->sf_get_parent_ino and ->sf_put_parent_ino [20/34] xfs: devirtualize ->sf_entsize and ->sf_nextentry [21/34] xfs: devirtualize ->sf_get_ino and ->sf_put_ino [22/34] xfs: devirtualize ->sf_get_ftype and ->sf_put_ftype [23/34] xfs: remove the ->data_dot_entry_p and ->data_dotdot_entry_p methods [24/34] xfs: remove the unused ->data_first_entry_p method [25/34] xfs: remove the ->data_entry_entry_p method [26/34] xfs: devirtualize ->data_entsize [27/34] xfs: devirtualize ->data_entry_tag_p [28/34] xfs: move the dir2 data block fixed offsets to struct xfs_da_geometry [29/34] xfs: cleanup xfs_dir2_data_entsize [30/34] xfs: devirtualize ->data_bestfree_p [31/34] xfs: devirtualize ->data_get_ftype and ->data_put_ftype [32/34] xfs: remove the now unused dir ops infrastructure [33/34] xfs: merge xfs_dir2_data_freescan and xfs_dir2_data_freescan_int [34/34] xfs: always pass a valid hdr to xfs_dir3_leaf_check_int

Message ID

20191101220719.29100-22-hch@lst.de (mailing list archive)

State

Superseded, archived

Headers

From: Christoph Hellwig <hch@lst.de>
To: linux-xfs@vger.kernel.org
Subject: [PATCH 21/34] xfs: devirtualize ->sf_get_ino and ->sf_put_ino
Date: Fri,  1 Nov 2019 15:07:06 -0700
Message-Id: <20191101220719.29100-22-hch@lst.de>
In-Reply-To: <20191101220719.29100-1-hch@lst.de>
References: <20191101220719.29100-1-hch@lst.de>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Sender: linux-xfs-owner@vger.kernel.org
Precedence: bulk

Series

[01/34] xfs: move incore structures out of xfs_da_format.h | expand

Commit Message

Christoph Hellwig Nov. 1, 2019, 10:07 p.m. UTC

Replace the ->sf_get_ino and ->sf_put_ino dir ops methods with directly
called xfs_dir2_sf_get_ino nd xfs_dir2_sf_put_ino helpers that take care
of the difference between the directory format with and without the file
type field.  Also move xfs_dir2_sf_get_parent_ino and
xfs_dir2_sf_put_parent_ino to xfs_dir2_sf.c with the rest of the
low-level short form entry handling.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 fs/xfs/libxfs/xfs_da_format.c  | 94 ----------------------------------
 fs/xfs/libxfs/xfs_dir2.h       |  5 --
 fs/xfs/libxfs/xfs_dir2_block.c |  2 +-
 fs/xfs/libxfs/xfs_dir2_priv.h  |  2 +
 fs/xfs/libxfs/xfs_dir2_sf.c    | 90 +++++++++++++++++++++++++++-----
 fs/xfs/xfs_dir2_readdir.c      |  2 +-
 6 files changed, 81 insertions(+), 114 deletions(-)

Comments

Darrick J. Wong Nov. 4, 2019, 8:33 p.m. UTC | #1

On Fri, Nov 01, 2019 at 03:07:06PM -0700, Christoph Hellwig wrote:
> Replace the ->sf_get_ino and ->sf_put_ino dir ops methods with directly
> called xfs_dir2_sf_get_ino nd xfs_dir2_sf_put_ino helpers that take care

                            ^^^ "and"

> of the difference between the directory format with and without the file
> type field.  Also move xfs_dir2_sf_get_parent_ino and
> xfs_dir2_sf_put_parent_ino to xfs_dir2_sf.c with the rest of the
> low-level short form entry handling.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>

Looks ok, except...

<snip>

> diff --git a/fs/xfs/libxfs/xfs_dir2_sf.c b/fs/xfs/libxfs/xfs_dir2_sf.c
> index f3a4f0ddfc1a..c33d838b1a5c 100644
> --- a/fs/xfs/libxfs/xfs_dir2_sf.c
> +++ b/fs/xfs/libxfs/xfs_dir2_sf.c
> @@ -63,6 +63,69 @@ xfs_dir2_sf_nextentry(
>  		((char *)sfep + xfs_dir2_sf_entsize(mp, hdr, sfep->namelen));
>  }
>  
> +/*
> + * In short-form directory entries the inode numbers are stored at variable
> + * offset behind the entry name. If the entry stores a filetype value, then it
> + * sits between the name and the inode number.  The actual inode numbers can
> + * come in two formats as well, either 4 bytes or 8 bytes wide.
> + */
> +xfs_ino_t
> +xfs_dir2_sf_get_ino(
> +	struct xfs_mount		*mp,
> +	struct xfs_dir2_sf_hdr		*hdr,
> +	struct xfs_dir2_sf_entry	*sfep)
> +{
> +	uint8_t				*from = sfep->name + sfep->namelen;
> +
> +	if (xfs_sb_version_hasftype(&mp->m_sb))
> +		from++;
> +
> +	if (!hdr->i8count)
> +		return get_unaligned_be32(from);
> +	return get_unaligned_be64(from) & 0x00ffffffffffffffULL;

...maybe we ought to convert this to use XFS_MAXINUMBER instead of
encoding magic numbers?

> +}
> +
> +void
> +xfs_dir2_sf_put_ino(
> +	struct xfs_mount		*mp,
> +	struct xfs_dir2_sf_hdr		*hdr,
> +	struct xfs_dir2_sf_entry	*sfep,
> +	xfs_ino_t			ino)
> +{
> +	uint8_t				*to = sfep->name + sfep->namelen;
> +
> +	ASSERT((ino & 0xff00000000000000ULL) == 0);

Same here...

> +	if (xfs_sb_version_hasftype(&mp->m_sb))
> +		to++;
> +
> +	if (hdr->i8count)
> +		put_unaligned_be64(ino, to);
> +	else
> +		put_unaligned_be32(ino, to);
> +}
> +
> +xfs_ino_t
> +xfs_dir2_sf_get_parent_ino(
> +	struct xfs_dir2_sf_hdr	*hdr)
> +{
> +	if (!hdr->i8count)
> +		return get_unaligned_be32(hdr->parent);
> +	return get_unaligned_be64(hdr->parent) & 0x00ffffffffffffffULL;

And here...

> +}
> +
> +void
> +xfs_dir2_sf_put_parent_ino(
> +	struct xfs_dir2_sf_hdr		*hdr,
> +	xfs_ino_t			ino)
> +{
> +	ASSERT((ino & 0xff00000000000000ULL) == 0);

And here?

--D

> +	if (hdr->i8count)
> +		put_unaligned_be64(ino, hdr->parent);
> +	else
> +		put_unaligned_be32(ino, hdr->parent);
> +}
> +
>  /*
>   * Given a block directory (dp/block), calculate its size as a shortform (sf)
>   * directory and a header for the sf directory, if it will fit it the
> @@ -240,7 +303,7 @@ xfs_dir2_block_to_sf(
>  				(xfs_dir2_data_aoff_t)
>  				((char *)dep - (char *)hdr));
>  			memcpy(sfep->name, dep->name, dep->namelen);
> -			dp->d_ops->sf_put_ino(sfp, sfep,
> +			xfs_dir2_sf_put_ino(mp, sfp, sfep,
>  					      be64_to_cpu(dep->inumber));
>  			dp->d_ops->sf_put_ftype(sfep,
>  					dp->d_ops->data_get_ftype(dep));
> @@ -413,7 +476,7 @@ xfs_dir2_sf_addname_easy(
>  	sfep->namelen = args->namelen;
>  	xfs_dir2_sf_put_offset(sfep, offset);
>  	memcpy(sfep->name, args->name, sfep->namelen);
> -	dp->d_ops->sf_put_ino(sfp, sfep, args->inumber);
> +	xfs_dir2_sf_put_ino(mp, sfp, sfep, args->inumber);
>  	dp->d_ops->sf_put_ftype(sfep, args->filetype);
>  
>  	/*
> @@ -503,7 +566,7 @@ xfs_dir2_sf_addname_hard(
>  	sfep->namelen = args->namelen;
>  	xfs_dir2_sf_put_offset(sfep, offset);
>  	memcpy(sfep->name, args->name, sfep->namelen);
> -	dp->d_ops->sf_put_ino(sfp, sfep, args->inumber);
> +	xfs_dir2_sf_put_ino(mp, sfp, sfep, args->inumber);
>  	dp->d_ops->sf_put_ftype(sfep, args->filetype);
>  	sfp->count++;
>  	if (args->inumber > XFS_DIR2_MAX_SHORT_INUM && !objchange)
> @@ -620,7 +683,7 @@ xfs_dir2_sf_check(
>  	     i < sfp->count;
>  	     i++, sfep = xfs_dir2_sf_nextentry(dp->i_mount, sfp, sfep)) {
>  		ASSERT(xfs_dir2_sf_get_offset(sfep) >= offset);
> -		ino = dp->d_ops->sf_get_ino(sfp, sfep);
> +		ino = xfs_dir2_sf_get_ino(dp->i_mount, sfp, sfep);
>  		i8count += ino > XFS_DIR2_MAX_SHORT_INUM;
>  		offset =
>  			xfs_dir2_sf_get_offset(sfep) +
> @@ -712,7 +775,7 @@ xfs_dir2_sf_verify(
>  			return __this_address;
>  
>  		/* Check the inode number. */
> -		ino = dops->sf_get_ino(sfp, sfep);
> +		ino = xfs_dir2_sf_get_ino(mp, sfp, sfep);
>  		i8count += ino > XFS_DIR2_MAX_SHORT_INUM;
>  		error = xfs_dir_ino_validate(mp, ino);
>  		if (error)
> @@ -861,7 +924,7 @@ xfs_dir2_sf_lookup(
>  								sfep->namelen);
>  		if (cmp != XFS_CMP_DIFFERENT && cmp != args->cmpresult) {
>  			args->cmpresult = cmp;
> -			args->inumber = dp->d_ops->sf_get_ino(sfp, sfep);
> +			args->inumber = xfs_dir2_sf_get_ino(mp, sfp, sfep);
>  			args->filetype = dp->d_ops->sf_get_ftype(sfep);
>  			if (cmp == XFS_CMP_EXACT)
>  				return -EEXIST;
> @@ -920,7 +983,7 @@ xfs_dir2_sf_removename(
>  	     i++, sfep = xfs_dir2_sf_nextentry(mp, sfp, sfep)) {
>  		if (xfs_da_compname(args, sfep->name, sfep->namelen) ==
>  								XFS_CMP_EXACT) {
> -			ASSERT(dp->d_ops->sf_get_ino(sfp, sfep) ==
> +			ASSERT(xfs_dir2_sf_get_ino(mp, sfp, sfep) ==
>  			       args->inumber);
>  			break;
>  		}
> @@ -1041,9 +1104,10 @@ xfs_dir2_sf_replace(
>  		     i++, sfep = xfs_dir2_sf_nextentry(mp, sfp, sfep)) {
>  			if (xfs_da_compname(args, sfep->name, sfep->namelen) ==
>  								XFS_CMP_EXACT) {
> -				ino = dp->d_ops->sf_get_ino(sfp, sfep);
> +				ino = xfs_dir2_sf_get_ino(mp, sfp, sfep);
>  				ASSERT(args->inumber != ino);
> -				dp->d_ops->sf_put_ino(sfp, sfep, args->inumber);
> +				xfs_dir2_sf_put_ino(mp, sfp, sfep,
> +						args->inumber);
>  				dp->d_ops->sf_put_ftype(sfep, args->filetype);
>  				break;
>  			}
> @@ -1148,8 +1212,8 @@ xfs_dir2_sf_toino4(
>  		sfep->namelen = oldsfep->namelen;
>  		memcpy(sfep->offset, oldsfep->offset, sizeof(sfep->offset));
>  		memcpy(sfep->name, oldsfep->name, sfep->namelen);
> -		dp->d_ops->sf_put_ino(sfp, sfep,
> -				      dp->d_ops->sf_get_ino(oldsfp, oldsfep));
> +		xfs_dir2_sf_put_ino(mp, sfp, sfep,
> +				xfs_dir2_sf_get_ino(mp, oldsfp, oldsfep));
>  		dp->d_ops->sf_put_ftype(sfep, dp->d_ops->sf_get_ftype(oldsfep));
>  	}
>  	/*
> @@ -1220,8 +1284,8 @@ xfs_dir2_sf_toino8(
>  		sfep->namelen = oldsfep->namelen;
>  		memcpy(sfep->offset, oldsfep->offset, sizeof(sfep->offset));
>  		memcpy(sfep->name, oldsfep->name, sfep->namelen);
> -		dp->d_ops->sf_put_ino(sfp, sfep,
> -				      dp->d_ops->sf_get_ino(oldsfp, oldsfep));
> +		xfs_dir2_sf_put_ino(mp, sfp, sfep,
> +				xfs_dir2_sf_get_ino(mp, oldsfp, oldsfep));
>  		dp->d_ops->sf_put_ftype(sfep, dp->d_ops->sf_get_ftype(oldsfep));
>  	}
>  	/*
> diff --git a/fs/xfs/xfs_dir2_readdir.c b/fs/xfs/xfs_dir2_readdir.c
> index 7d150e914d00..9d318f091a73 100644
> --- a/fs/xfs/xfs_dir2_readdir.c
> +++ b/fs/xfs/xfs_dir2_readdir.c
> @@ -114,7 +114,7 @@ xfs_dir2_sf_getdents(
>  			continue;
>  		}
>  
> -		ino = dp->d_ops->sf_get_ino(sfp, sfep);
> +		ino = xfs_dir2_sf_get_ino(mp, sfp, sfep);
>  		filetype = dp->d_ops->sf_get_ftype(sfep);
>  		ctx->pos = off & 0x7fffffff;
>  		if (!xfs_dir2_namecheck(sfep->name, sfep->namelen)) {
> -- 
> 2.20.1
>

Darrick J. Wong Nov. 7, 2019, 1:05 a.m. UTC | #2

On Mon, Nov 04, 2019 at 12:33:34PM -0800, Darrick J. Wong wrote:
> On Fri, Nov 01, 2019 at 03:07:06PM -0700, Christoph Hellwig wrote:
> > Replace the ->sf_get_ino and ->sf_put_ino dir ops methods with directly
> > called xfs_dir2_sf_get_ino nd xfs_dir2_sf_put_ino helpers that take care
> 
>                             ^^^ "and"
> 
> > of the difference between the directory format with and without the file
> > type field.  Also move xfs_dir2_sf_get_parent_ino and
> > xfs_dir2_sf_put_parent_ino to xfs_dir2_sf.c with the rest of the
> > low-level short form entry handling.
> > 
> > Signed-off-by: Christoph Hellwig <hch@lst.de>
> 
> Looks ok, except...
> 
> <snip>
> 
> > diff --git a/fs/xfs/libxfs/xfs_dir2_sf.c b/fs/xfs/libxfs/xfs_dir2_sf.c
> > index f3a4f0ddfc1a..c33d838b1a5c 100644
> > --- a/fs/xfs/libxfs/xfs_dir2_sf.c
> > +++ b/fs/xfs/libxfs/xfs_dir2_sf.c
> > @@ -63,6 +63,69 @@ xfs_dir2_sf_nextentry(
> >  		((char *)sfep + xfs_dir2_sf_entsize(mp, hdr, sfep->namelen));
> >  }
> >  
> > +/*
> > + * In short-form directory entries the inode numbers are stored at variable
> > + * offset behind the entry name. If the entry stores a filetype value, then it
> > + * sits between the name and the inode number.  The actual inode numbers can
> > + * come in two formats as well, either 4 bytes or 8 bytes wide.
> > + */
> > +xfs_ino_t
> > +xfs_dir2_sf_get_ino(
> > +	struct xfs_mount		*mp,
> > +	struct xfs_dir2_sf_hdr		*hdr,
> > +	struct xfs_dir2_sf_entry	*sfep)
> > +{
> > +	uint8_t				*from = sfep->name + sfep->namelen;
> > +
> > +	if (xfs_sb_version_hasftype(&mp->m_sb))
> > +		from++;
> > +
> > +	if (!hdr->i8count)
> > +		return get_unaligned_be32(from);
> > +	return get_unaligned_be64(from) & 0x00ffffffffffffffULL;
> 
> ...maybe we ought to convert this to use XFS_MAXINUMBER instead of
> encoding magic numbers?
> 
> > +}
> > +
> > +void
> > +xfs_dir2_sf_put_ino(

Also, I think these helpers can be static now...

--D

> > +	struct xfs_mount		*mp,
> > +	struct xfs_dir2_sf_hdr		*hdr,
> > +	struct xfs_dir2_sf_entry	*sfep,
> > +	xfs_ino_t			ino)
> > +{
> > +	uint8_t				*to = sfep->name + sfep->namelen;
> > +
> > +	ASSERT((ino & 0xff00000000000000ULL) == 0);
> 
> Same here...
> 
> > +	if (xfs_sb_version_hasftype(&mp->m_sb))
> > +		to++;
> > +
> > +	if (hdr->i8count)
> > +		put_unaligned_be64(ino, to);
> > +	else
> > +		put_unaligned_be32(ino, to);
> > +}
> > +
> > +xfs_ino_t
> > +xfs_dir2_sf_get_parent_ino(
> > +	struct xfs_dir2_sf_hdr	*hdr)
> > +{
> > +	if (!hdr->i8count)
> > +		return get_unaligned_be32(hdr->parent);
> > +	return get_unaligned_be64(hdr->parent) & 0x00ffffffffffffffULL;
> 
> And here...
> 
> > +}
> > +
> > +void
> > +xfs_dir2_sf_put_parent_ino(
> > +	struct xfs_dir2_sf_hdr		*hdr,
> > +	xfs_ino_t			ino)
> > +{
> > +	ASSERT((ino & 0xff00000000000000ULL) == 0);
> 
> And here?
> 
> --D
> 
> > +	if (hdr->i8count)
> > +		put_unaligned_be64(ino, hdr->parent);
> > +	else
> > +		put_unaligned_be32(ino, hdr->parent);
> > +}
> > +
> >  /*
> >   * Given a block directory (dp/block), calculate its size as a shortform (sf)
> >   * directory and a header for the sf directory, if it will fit it the
> > @@ -240,7 +303,7 @@ xfs_dir2_block_to_sf(
> >  				(xfs_dir2_data_aoff_t)
> >  				((char *)dep - (char *)hdr));
> >  			memcpy(sfep->name, dep->name, dep->namelen);
> > -			dp->d_ops->sf_put_ino(sfp, sfep,
> > +			xfs_dir2_sf_put_ino(mp, sfp, sfep,
> >  					      be64_to_cpu(dep->inumber));
> >  			dp->d_ops->sf_put_ftype(sfep,
> >  					dp->d_ops->data_get_ftype(dep));
> > @@ -413,7 +476,7 @@ xfs_dir2_sf_addname_easy(
> >  	sfep->namelen = args->namelen;
> >  	xfs_dir2_sf_put_offset(sfep, offset);
> >  	memcpy(sfep->name, args->name, sfep->namelen);
> > -	dp->d_ops->sf_put_ino(sfp, sfep, args->inumber);
> > +	xfs_dir2_sf_put_ino(mp, sfp, sfep, args->inumber);
> >  	dp->d_ops->sf_put_ftype(sfep, args->filetype);
> >  
> >  	/*
> > @@ -503,7 +566,7 @@ xfs_dir2_sf_addname_hard(
> >  	sfep->namelen = args->namelen;
> >  	xfs_dir2_sf_put_offset(sfep, offset);
> >  	memcpy(sfep->name, args->name, sfep->namelen);
> > -	dp->d_ops->sf_put_ino(sfp, sfep, args->inumber);
> > +	xfs_dir2_sf_put_ino(mp, sfp, sfep, args->inumber);
> >  	dp->d_ops->sf_put_ftype(sfep, args->filetype);
> >  	sfp->count++;
> >  	if (args->inumber > XFS_DIR2_MAX_SHORT_INUM && !objchange)
> > @@ -620,7 +683,7 @@ xfs_dir2_sf_check(
> >  	     i < sfp->count;
> >  	     i++, sfep = xfs_dir2_sf_nextentry(dp->i_mount, sfp, sfep)) {
> >  		ASSERT(xfs_dir2_sf_get_offset(sfep) >= offset);
> > -		ino = dp->d_ops->sf_get_ino(sfp, sfep);
> > +		ino = xfs_dir2_sf_get_ino(dp->i_mount, sfp, sfep);
> >  		i8count += ino > XFS_DIR2_MAX_SHORT_INUM;
> >  		offset =
> >  			xfs_dir2_sf_get_offset(sfep) +
> > @@ -712,7 +775,7 @@ xfs_dir2_sf_verify(
> >  			return __this_address;
> >  
> >  		/* Check the inode number. */
> > -		ino = dops->sf_get_ino(sfp, sfep);
> > +		ino = xfs_dir2_sf_get_ino(mp, sfp, sfep);
> >  		i8count += ino > XFS_DIR2_MAX_SHORT_INUM;
> >  		error = xfs_dir_ino_validate(mp, ino);
> >  		if (error)
> > @@ -861,7 +924,7 @@ xfs_dir2_sf_lookup(
> >  								sfep->namelen);
> >  		if (cmp != XFS_CMP_DIFFERENT && cmp != args->cmpresult) {
> >  			args->cmpresult = cmp;
> > -			args->inumber = dp->d_ops->sf_get_ino(sfp, sfep);
> > +			args->inumber = xfs_dir2_sf_get_ino(mp, sfp, sfep);
> >  			args->filetype = dp->d_ops->sf_get_ftype(sfep);
> >  			if (cmp == XFS_CMP_EXACT)
> >  				return -EEXIST;
> > @@ -920,7 +983,7 @@ xfs_dir2_sf_removename(
> >  	     i++, sfep = xfs_dir2_sf_nextentry(mp, sfp, sfep)) {
> >  		if (xfs_da_compname(args, sfep->name, sfep->namelen) ==
> >  								XFS_CMP_EXACT) {
> > -			ASSERT(dp->d_ops->sf_get_ino(sfp, sfep) ==
> > +			ASSERT(xfs_dir2_sf_get_ino(mp, sfp, sfep) ==
> >  			       args->inumber);
> >  			break;
> >  		}
> > @@ -1041,9 +1104,10 @@ xfs_dir2_sf_replace(
> >  		     i++, sfep = xfs_dir2_sf_nextentry(mp, sfp, sfep)) {
> >  			if (xfs_da_compname(args, sfep->name, sfep->namelen) ==
> >  								XFS_CMP_EXACT) {
> > -				ino = dp->d_ops->sf_get_ino(sfp, sfep);
> > +				ino = xfs_dir2_sf_get_ino(mp, sfp, sfep);
> >  				ASSERT(args->inumber != ino);
> > -				dp->d_ops->sf_put_ino(sfp, sfep, args->inumber);
> > +				xfs_dir2_sf_put_ino(mp, sfp, sfep,
> > +						args->inumber);
> >  				dp->d_ops->sf_put_ftype(sfep, args->filetype);
> >  				break;
> >  			}
> > @@ -1148,8 +1212,8 @@ xfs_dir2_sf_toino4(
> >  		sfep->namelen = oldsfep->namelen;
> >  		memcpy(sfep->offset, oldsfep->offset, sizeof(sfep->offset));
> >  		memcpy(sfep->name, oldsfep->name, sfep->namelen);
> > -		dp->d_ops->sf_put_ino(sfp, sfep,
> > -				      dp->d_ops->sf_get_ino(oldsfp, oldsfep));
> > +		xfs_dir2_sf_put_ino(mp, sfp, sfep,
> > +				xfs_dir2_sf_get_ino(mp, oldsfp, oldsfep));
> >  		dp->d_ops->sf_put_ftype(sfep, dp->d_ops->sf_get_ftype(oldsfep));
> >  	}
> >  	/*
> > @@ -1220,8 +1284,8 @@ xfs_dir2_sf_toino8(
> >  		sfep->namelen = oldsfep->namelen;
> >  		memcpy(sfep->offset, oldsfep->offset, sizeof(sfep->offset));
> >  		memcpy(sfep->name, oldsfep->name, sfep->namelen);
> > -		dp->d_ops->sf_put_ino(sfp, sfep,
> > -				      dp->d_ops->sf_get_ino(oldsfp, oldsfep));
> > +		xfs_dir2_sf_put_ino(mp, sfp, sfep,
> > +				xfs_dir2_sf_get_ino(mp, oldsfp, oldsfep));
> >  		dp->d_ops->sf_put_ftype(sfep, dp->d_ops->sf_get_ftype(oldsfep));
> >  	}
> >  	/*
> > diff --git a/fs/xfs/xfs_dir2_readdir.c b/fs/xfs/xfs_dir2_readdir.c
> > index 7d150e914d00..9d318f091a73 100644
> > --- a/fs/xfs/xfs_dir2_readdir.c
> > +++ b/fs/xfs/xfs_dir2_readdir.c
> > @@ -114,7 +114,7 @@ xfs_dir2_sf_getdents(
> >  			continue;
> >  		}
> >  
> > -		ino = dp->d_ops->sf_get_ino(sfp, sfep);
> > +		ino = xfs_dir2_sf_get_ino(mp, sfp, sfep);
> >  		filetype = dp->d_ops->sf_get_ftype(sfep);
> >  		ctx->pos = off & 0x7fffffff;
> >  		if (!xfs_dir2_namecheck(sfep->name, sfep->namelen)) {
> > -- 
> > 2.20.1
> >

Christoph Hellwig Nov. 7, 2019, 8:29 a.m. UTC | #3

On Wed, Nov 06, 2019 at 05:05:55PM -0800, Darrick J. Wong wrote:
> > > +}
> > > +
> > > +void
> > > +xfs_dir2_sf_put_ino(
> 
> Also, I think these helpers can be static now...

xfs_dir2_sf_put_ino can and should be marked static.  xfs_dir2_sf_get_ino
is used elsewhere, though.

diff --git a/fs/xfs/libxfs/xfs_da_format.c b/fs/xfs/libxfs/xfs_da_format.c
index f80495a38a76..f427f141d001 100644
--- a/fs/xfs/libxfs/xfs_da_format.c
+++ b/fs/xfs/libxfs/xfs_da_format.c
@@ -58,94 +58,6 @@  xfs_dir3_sfe_put_ftype(
 	sfep->name[sfep->namelen] = ftype;
 }
 
-/*
- * Inode numbers in short-form directories can come in two versions,
- * either 4 bytes or 8 bytes wide.  These helpers deal with the
- * two forms transparently by looking at the headers i8count field.
- *
- * For 64-bit inode number the most significant byte must be zero.
- */
-static xfs_ino_t
-xfs_dir2_sf_get_ino(
-	struct xfs_dir2_sf_hdr	*hdr,
-	uint8_t			*from)
-{
-	if (hdr->i8count)
-		return get_unaligned_be64(from) & 0x00ffffffffffffffULL;
-	else
-		return get_unaligned_be32(from);
-}
-
-static void
-xfs_dir2_sf_put_ino(
-	struct xfs_dir2_sf_hdr	*hdr,
-	uint8_t			*to,
-	xfs_ino_t		ino)
-{
-	ASSERT((ino & 0xff00000000000000ULL) == 0);
-
-	if (hdr->i8count)
-		put_unaligned_be64(ino, to);
-	else
-		put_unaligned_be32(ino, to);
-}
-
-xfs_ino_t
-xfs_dir2_sf_get_parent_ino(
-	struct xfs_dir2_sf_hdr	*hdr)
-{
-	return xfs_dir2_sf_get_ino(hdr, hdr->parent);
-}
-
-void
-xfs_dir2_sf_put_parent_ino(
-	struct xfs_dir2_sf_hdr	*hdr,
-	xfs_ino_t		ino)
-{
-	xfs_dir2_sf_put_ino(hdr, hdr->parent, ino);
-}
-
-/*
- * In short-form directory entries the inode numbers are stored at variable
- * offset behind the entry name. If the entry stores a filetype value, then it
- * sits between the name and the inode number. Hence the inode numbers may only
- * be accessed through the helpers below.
- */
-static xfs_ino_t
-xfs_dir2_sfe_get_ino(
-	struct xfs_dir2_sf_hdr	*hdr,
-	struct xfs_dir2_sf_entry *sfep)
-{
-	return xfs_dir2_sf_get_ino(hdr, &sfep->name[sfep->namelen]);
-}
-
-static void
-xfs_dir2_sfe_put_ino(
-	struct xfs_dir2_sf_hdr	*hdr,
-	struct xfs_dir2_sf_entry *sfep,
-	xfs_ino_t		ino)
-{
-	xfs_dir2_sf_put_ino(hdr, &sfep->name[sfep->namelen], ino);
-}
-
-static xfs_ino_t
-xfs_dir3_sfe_get_ino(
-	struct xfs_dir2_sf_hdr	*hdr,
-	struct xfs_dir2_sf_entry *sfep)
-{
-	return xfs_dir2_sf_get_ino(hdr, &sfep->name[sfep->namelen + 1]);
-}
-
-static void
-xfs_dir3_sfe_put_ino(
-	struct xfs_dir2_sf_hdr	*hdr,
-	struct xfs_dir2_sf_entry *sfep,
-	xfs_ino_t		ino)
-{
-	xfs_dir2_sf_put_ino(hdr, &sfep->name[sfep->namelen + 1], ino);
-}
-
-
 /*
  * Directory data block operations
  */
@@ -361,8 +273,6 @@  xfs_dir3_data_unused_p(struct xfs_dir2_data_hdr *hdr)
 static const struct xfs_dir_ops xfs_dir2_ops = {
 	.sf_get_ftype = xfs_dir2_sfe_get_ftype,
 	.sf_put_ftype = xfs_dir2_sfe_put_ftype,
-	.sf_get_ino = xfs_dir2_sfe_get_ino,
-	.sf_put_ino = xfs_dir2_sfe_put_ino,
 
 	.data_entsize = xfs_dir2_data_entsize,
 	.data_get_ftype = xfs_dir2_data_get_ftype,
@@ -388,8 +298,6 @@  static const struct xfs_dir_ops xfs_dir2_ops = {
 static const struct xfs_dir_ops xfs_dir2_ftype_ops = {
 	.sf_get_ftype = xfs_dir3_sfe_get_ftype,
 	.sf_put_ftype = xfs_dir3_sfe_put_ftype,
-	.sf_get_ino = xfs_dir3_sfe_get_ino,
-	.sf_put_ino = xfs_dir3_sfe_put_ino,
 
 	.data_entsize = xfs_dir3_data_entsize,
 	.data_get_ftype = xfs_dir3_data_get_ftype,
@@ -415,8 +323,6 @@  static const struct xfs_dir_ops xfs_dir2_ftype_ops = {
 static const struct xfs_dir_ops xfs_dir3_ops = {
 	.sf_get_ftype = xfs_dir3_sfe_get_ftype,
 	.sf_put_ftype = xfs_dir3_sfe_put_ftype,
-	.sf_get_ino = xfs_dir3_sfe_get_ino,
-	.sf_put_ino = xfs_dir3_sfe_put_ino,
 
 	.data_entsize = xfs_dir3_data_entsize,
 	.data_get_ftype = xfs_dir3_data_get_ftype,
diff --git a/fs/xfs/libxfs/xfs_dir2.h b/fs/xfs/libxfs/xfs_dir2.h
index c99f0ad6607a..049d844d6a18 100644
--- a/fs/xfs/libxfs/xfs_dir2.h
+++ b/fs/xfs/libxfs/xfs_dir2.h
@@ -35,11 +35,6 @@  struct xfs_dir_ops {
 	uint8_t (*sf_get_ftype)(struct xfs_dir2_sf_entry *sfep);
 	void	(*sf_put_ftype)(struct xfs_dir2_sf_entry *sfep,
 				uint8_t ftype);
-	xfs_ino_t (*sf_get_ino)(struct xfs_dir2_sf_hdr *hdr,
-				struct xfs_dir2_sf_entry *sfep);
-	void	(*sf_put_ino)(struct xfs_dir2_sf_hdr *hdr,
-			      struct xfs_dir2_sf_entry *sfep,
-			      xfs_ino_t ino);
 
 	int	(*data_entsize)(int len);
 	uint8_t (*data_get_ftype)(struct xfs_dir2_data_entry *dep);
diff --git a/fs/xfs/libxfs/xfs_dir2_block.c b/fs/xfs/libxfs/xfs_dir2_block.c
index 88b3d1884c74..0f3024386a5c 100644
--- a/fs/xfs/libxfs/xfs_dir2_block.c
+++ b/fs/xfs/libxfs/xfs_dir2_block.c
@@ -1214,7 +1214,7 @@  xfs_dir2_sf_to_block(
 		 * Copy a real entry.
 		 */
 		dep = (xfs_dir2_data_entry_t *)((char *)hdr + newoffset);
-		dep->inumber = cpu_to_be64(dp->d_ops->sf_get_ino(sfp, sfep));
+		dep->inumber = cpu_to_be64(xfs_dir2_sf_get_ino(mp, sfp, sfep));
 		dep->namelen = sfep->namelen;
 		dp->d_ops->data_put_ftype(dep, dp->d_ops->sf_get_ftype(sfep));
 		memcpy(dep->name, sfep->name, dep->namelen);
diff --git a/fs/xfs/libxfs/xfs_dir2_priv.h b/fs/xfs/libxfs/xfs_dir2_priv.h
index a96b9ba7909b..57c0f7aee7a4 100644
--- a/fs/xfs/libxfs/xfs_dir2_priv.h
+++ b/fs/xfs/libxfs/xfs_dir2_priv.h
@@ -135,6 +135,8 @@  extern int xfs_dir2_free_read(struct xfs_trans *tp, struct xfs_inode *dp,
 		xfs_dablk_t fbno, struct xfs_buf **bpp);
 
 /* xfs_dir2_sf.c */
+xfs_ino_t xfs_dir2_sf_get_ino(struct xfs_mount *mp, struct xfs_dir2_sf_hdr *hdr,
+		struct xfs_dir2_sf_entry *sfep);
 xfs_ino_t xfs_dir2_sf_get_parent_ino(struct xfs_dir2_sf_hdr *hdr);
 void xfs_dir2_sf_put_parent_ino(struct xfs_dir2_sf_hdr *hdr, xfs_ino_t ino);
 struct xfs_dir2_sf_entry *xfs_dir2_sf_nextentry(struct xfs_mount *mp,
diff --git a/fs/xfs/libxfs/xfs_dir2_sf.c b/fs/xfs/libxfs/xfs_dir2_sf.c
index f3a4f0ddfc1a..c33d838b1a5c 100644
--- a/fs/xfs/libxfs/xfs_dir2_sf.c
+++ b/fs/xfs/libxfs/xfs_dir2_sf.c
@@ -63,6 +63,69 @@  xfs_dir2_sf_nextentry(
 		((char *)sfep + xfs_dir2_sf_entsize(mp, hdr, sfep->namelen));
 }
 
+/*
+ * In short-form directory entries the inode numbers are stored at variable
+ * offset behind the entry name. If the entry stores a filetype value, then it
+ * sits between the name and the inode number.  The actual inode numbers can
+ * come in two formats as well, either 4 bytes or 8 bytes wide.
+ */
+xfs_ino_t
+xfs_dir2_sf_get_ino(
+	struct xfs_mount		*mp,
+	struct xfs_dir2_sf_hdr		*hdr,
+	struct xfs_dir2_sf_entry	*sfep)
+{
+	uint8_t				*from = sfep->name + sfep->namelen;
+
+	if (xfs_sb_version_hasftype(&mp->m_sb))
+		from++;
+
+	if (!hdr->i8count)
+		return get_unaligned_be32(from);
+	return get_unaligned_be64(from) & 0x00ffffffffffffffULL;
+}
+
+void
+xfs_dir2_sf_put_ino(
+	struct xfs_mount		*mp,
+	struct xfs_dir2_sf_hdr		*hdr,
+	struct xfs_dir2_sf_entry	*sfep,
+	xfs_ino_t			ino)
+{
+	uint8_t				*to = sfep->name + sfep->namelen;
+
+	ASSERT((ino & 0xff00000000000000ULL) == 0);
+
+	if (xfs_sb_version_hasftype(&mp->m_sb))
+		to++;
+
+	if (hdr->i8count)
+		put_unaligned_be64(ino, to);
+	else
+		put_unaligned_be32(ino, to);
+}
+
+xfs_ino_t
+xfs_dir2_sf_get_parent_ino(
+	struct xfs_dir2_sf_hdr	*hdr)
+{
+	if (!hdr->i8count)
+		return get_unaligned_be32(hdr->parent);
+	return get_unaligned_be64(hdr->parent) & 0x00ffffffffffffffULL;
+}
+
+void
+xfs_dir2_sf_put_parent_ino(
+	struct xfs_dir2_sf_hdr		*hdr,
+	xfs_ino_t			ino)
+{
+	ASSERT((ino & 0xff00000000000000ULL) == 0);
+	if (hdr->i8count)
+		put_unaligned_be64(ino, hdr->parent);
+	else
+		put_unaligned_be32(ino, hdr->parent);
+}
+
 /*
  * Given a block directory (dp/block), calculate its size as a shortform (sf)
  * directory and a header for the sf directory, if it will fit it the
@@ -240,7 +303,7 @@  xfs_dir2_block_to_sf(
 				(xfs_dir2_data_aoff_t)
 				((char *)dep - (char *)hdr));
 			memcpy(sfep->name, dep->name, dep->namelen);
-			dp->d_ops->sf_put_ino(sfp, sfep,
+			xfs_dir2_sf_put_ino(mp, sfp, sfep,
 					      be64_to_cpu(dep->inumber));
 			dp->d_ops->sf_put_ftype(sfep,
 					dp->d_ops->data_get_ftype(dep));
@@ -413,7 +476,7 @@  xfs_dir2_sf_addname_easy(
 	sfep->namelen = args->namelen;
 	xfs_dir2_sf_put_offset(sfep, offset);
 	memcpy(sfep->name, args->name, sfep->namelen);
-	dp->d_ops->sf_put_ino(sfp, sfep, args->inumber);
+	xfs_dir2_sf_put_ino(mp, sfp, sfep, args->inumber);
 	dp->d_ops->sf_put_ftype(sfep, args->filetype);
 
 	/*
@@ -503,7 +566,7 @@  xfs_dir2_sf_addname_hard(
 	sfep->namelen = args->namelen;
 	xfs_dir2_sf_put_offset(sfep, offset);
 	memcpy(sfep->name, args->name, sfep->namelen);
-	dp->d_ops->sf_put_ino(sfp, sfep, args->inumber);
+	xfs_dir2_sf_put_ino(mp, sfp, sfep, args->inumber);
 	dp->d_ops->sf_put_ftype(sfep, args->filetype);
 	sfp->count++;
 	if (args->inumber > XFS_DIR2_MAX_SHORT_INUM && !objchange)
@@ -620,7 +683,7 @@  xfs_dir2_sf_check(
 	     i < sfp->count;
 	     i++, sfep = xfs_dir2_sf_nextentry(dp->i_mount, sfp, sfep)) {
 		ASSERT(xfs_dir2_sf_get_offset(sfep) >= offset);
-		ino = dp->d_ops->sf_get_ino(sfp, sfep);
+		ino = xfs_dir2_sf_get_ino(dp->i_mount, sfp, sfep);
 		i8count += ino > XFS_DIR2_MAX_SHORT_INUM;
 		offset =
 			xfs_dir2_sf_get_offset(sfep) +
@@ -712,7 +775,7 @@  xfs_dir2_sf_verify(
 			return __this_address;
 
 		/* Check the inode number. */
-		ino = dops->sf_get_ino(sfp, sfep);
+		ino = xfs_dir2_sf_get_ino(mp, sfp, sfep);
 		i8count += ino > XFS_DIR2_MAX_SHORT_INUM;
 		error = xfs_dir_ino_validate(mp, ino);
 		if (error)
@@ -861,7 +924,7 @@  xfs_dir2_sf_lookup(
 								sfep->namelen);
 		if (cmp != XFS_CMP_DIFFERENT && cmp != args->cmpresult) {
 			args->cmpresult = cmp;
-			args->inumber = dp->d_ops->sf_get_ino(sfp, sfep);
+			args->inumber = xfs_dir2_sf_get_ino(mp, sfp, sfep);
 			args->filetype = dp->d_ops->sf_get_ftype(sfep);
 			if (cmp == XFS_CMP_EXACT)
 				return -EEXIST;
@@ -920,7 +983,7 @@  xfs_dir2_sf_removename(
 	     i++, sfep = xfs_dir2_sf_nextentry(mp, sfp, sfep)) {
 		if (xfs_da_compname(args, sfep->name, sfep->namelen) ==
 								XFS_CMP_EXACT) {
-			ASSERT(dp->d_ops->sf_get_ino(sfp, sfep) ==
+			ASSERT(xfs_dir2_sf_get_ino(mp, sfp, sfep) ==
 			       args->inumber);
 			break;
 		}
@@ -1041,9 +1104,10 @@  xfs_dir2_sf_replace(
 		     i++, sfep = xfs_dir2_sf_nextentry(mp, sfp, sfep)) {
 			if (xfs_da_compname(args, sfep->name, sfep->namelen) ==
 								XFS_CMP_EXACT) {
-				ino = dp->d_ops->sf_get_ino(sfp, sfep);
+				ino = xfs_dir2_sf_get_ino(mp, sfp, sfep);
 				ASSERT(args->inumber != ino);
-				dp->d_ops->sf_put_ino(sfp, sfep, args->inumber);
+				xfs_dir2_sf_put_ino(mp, sfp, sfep,
+						args->inumber);
 				dp->d_ops->sf_put_ftype(sfep, args->filetype);
 				break;
 			}
@@ -1148,8 +1212,8 @@  xfs_dir2_sf_toino4(
 		sfep->namelen = oldsfep->namelen;
 		memcpy(sfep->offset, oldsfep->offset, sizeof(sfep->offset));
 		memcpy(sfep->name, oldsfep->name, sfep->namelen);
-		dp->d_ops->sf_put_ino(sfp, sfep,
-				      dp->d_ops->sf_get_ino(oldsfp, oldsfep));
+		xfs_dir2_sf_put_ino(mp, sfp, sfep,
+				xfs_dir2_sf_get_ino(mp, oldsfp, oldsfep));
 		dp->d_ops->sf_put_ftype(sfep, dp->d_ops->sf_get_ftype(oldsfep));
 	}
 	/*
@@ -1220,8 +1284,8 @@  xfs_dir2_sf_toino8(
 		sfep->namelen = oldsfep->namelen;
 		memcpy(sfep->offset, oldsfep->offset, sizeof(sfep->offset));
 		memcpy(sfep->name, oldsfep->name, sfep->namelen);
-		dp->d_ops->sf_put_ino(sfp, sfep,
-				      dp->d_ops->sf_get_ino(oldsfp, oldsfep));
+		xfs_dir2_sf_put_ino(mp, sfp, sfep,
+				xfs_dir2_sf_get_ino(mp, oldsfp, oldsfep));
 		dp->d_ops->sf_put_ftype(sfep, dp->d_ops->sf_get_ftype(oldsfep));
 	}
 	/*
diff --git a/fs/xfs/xfs_dir2_readdir.c b/fs/xfs/xfs_dir2_readdir.c
index 7d150e914d00..9d318f091a73 100644
--- a/fs/xfs/xfs_dir2_readdir.c
+++ b/fs/xfs/xfs_dir2_readdir.c
@@ -114,7 +114,7 @@  xfs_dir2_sf_getdents(
 			continue;
 		}
 
-		ino = dp->d_ops->sf_get_ino(sfp, sfep);
+		ino = xfs_dir2_sf_get_ino(mp, sfp, sfep);
 		filetype = dp->d_ops->sf_get_ftype(sfep);
 		ctx->pos = off & 0x7fffffff;
 		if (!xfs_dir2_namecheck(sfep->name, sfep->namelen)) {

[21/34] xfs: devirtualize ->sf_get_ino and ->sf_put_ino

Commit Message

Comments

Patch