Skip to contentSkip to conversation

xfs: minor cleanups for userspace sanity

PATCH Superseded Dave Chinner 3 participants 8 messages Aug 30, 2013 – Sep 3, 2013 XFS filesystem

lore
Share & tools
lore
#1 lore raw

PATCH 1/2 xfs: dtype changed xfs_dir2_sfe_put_ino to xfs_dir3_sfe_put_ino

From: Dave Chinner <[email protected]>

So fix up the export in xfs_dir2.h that is needed by userspace.

---
 fs/xfs/xfs_dir2.h | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
fs/xfs/xfs_dir2.h +3 -2
index 768ddad..9910401 100644
@@ -70,8 +70,9 @@ extern void xfs_dir2_sf_put_parent_ino(struct xfs_dir2_sf_hdr *sfp,
 		xfs_ino_t ino);
 extern xfs_ino_t xfs_dir3_sfe_get_ino(struct xfs_mount *mp,
 		struct xfs_dir2_sf_hdr *sfp, struct xfs_dir2_sf_entry *sfep);
-extern void xfs_dir2_sfe_put_ino( struct xfs_dir2_sf_hdr *,
-		struct xfs_dir2_sf_entry *sfep, xfs_ino_t ino);
+extern void xfs_dir3_sfe_put_ino(struct xfs_mount *mp,
+		struct xfs_dir2_sf_hdr *hdr, struct xfs_dir2_sf_entry *sfep,
+		xfs_ino_t ino);
 
 extern int xfs_dir2_isblock(struct xfs_trans *tp, struct xfs_inode *dp, int *r);
 extern int xfs_dir2_isleaf(struct xfs_trans *tp, struct xfs_inode *dp, int *r);
Signature · 6 lines
1.8.3.2

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Reviewed-by on v1 1/2
#2 lore raw

Re: [PATCH 1/2] xfs: dtype changed xfs_dir2_sfe_put_ino to xfs_dir3_sfe_put_ino

Quoted 15 lines · Dave Chinner

From: Dave Chinner<[email protected]>

So fix up the export in xfs_dir2.h that is needed by userspace.

---
  fs/xfs/xfs_dir2.h | 5 +++--
  1 file changed, 3 insertions(+), 2 deletions(-)
fs/xfs/xfs_dir2.h +0 -0
index 768ddad..9910401 100644
@@ -70,8 +70,9 @@ extern void xfs_dir2_sf_put_parent_ino(struct xfs_dir2_sf_hdr *sfp,
  		xfs_ino_t ino);

fs/xfs/xfs_dir2.h extern void xfs_dir2_sf_put_parent_ino(struct xfs_dir2_sf_hdr *sfp, · line 78

Quoted 10 lines
fs/xfs/xfs_dir2.h +3 -2
  extern xfs_ino_t xfs_dir3_sfe_get_ino(struct xfs_mount *mp,
  		struct xfs_dir2_sf_hdr *sfp, struct xfs_dir2_sf_entry *sfep);
-extern void xfs_dir2_sfe_put_ino( struct xfs_dir2_sf_hdr *,
-		struct xfs_dir2_sf_entry *sfep, xfs_ino_t ino);
+extern void xfs_dir3_sfe_put_ino(struct xfs_mount *mp,
+		struct xfs_dir2_sf_hdr *hdr, struct xfs_dir2_sf_entry *sfep,
+		xfs_ino_t ino);
 
  extern int xfs_dir2_isblock(struct xfs_trans *tp, struct xfs_inode *dp, int *r);
  extern int xfs_dir2_isleaf(struct xfs_trans *tp, struct xfs_inode *dp, int *r);

Okay xfs_dir2_sfe_put_ino() does not exist. xfs_dir3_sfe_put_ino() is remotely used in xfs_repair but not in kernel.

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

#3 lore raw

PATCH v2 1/1 xfs: dtype changed xfs_dir2_sfe_put_ino to xfs_dir3_sfe_put_ino

In reply to Dave Chinner
Quoted 5 lines · Dave Chinner

From: Dave Chinner <[email protected]>

So fix up the export in xfs_dir2.h that is needed by userspace.

Well, this patch now causes a compilation failure. Why? xfs_dir3_sfe_put_ino() got marked as static because it didn't have a matching export.

Perhaps we should be a little less hasty in merging auto-generated patches from random build-farms? The patch was wrong - fixing the prototype like in v1 of this patch was the correct thing to do....

So, the updated patch below adds the revert.

Cheers,

Dave.

PS: Ben, when you merge a patch from a reply in an email thread, can you please pull the commit subject line from the patch, not the title of the email thread? One of the recent commits got it wrong. i.e. commit 3780437

"XFS: Assertion failed: first <= last && last < BBTOB(bp->b_length), file:"

should have had the title

"xfs: fix calculation of the number of node entries in a dir3 node"

Signature · 23 lines
Dave Chinner
[email protected]

xfs: dtype changed xfs_dir2_sfe_put_ino to xfs_dir3_sfe_put_ino

From: Dave Chinner <[email protected]>

So fix up the export in xfs_dir2.h that is needed by userspace.

<sigh>

Now xfs_dir3_sfe_put_ino has been made static. Revert 98f7462 ("xfs:
xfs_dir3_sfe_put_ino can be static") to being non static so that the
code shared with userspace is identical again.

Signed-off-by: Dave Chinner <[email protected]>
---
V2: revert 98f7462 so that xfs_dir3_sfe_put_ino is exported again to
match what userspace needs.

 fs/xfs/xfs_dir2.h    | 5 +++--
 fs/xfs/xfs_dir2_sf.c | 2 +-
 2 files changed, 4 insertions(+), 3 deletions(-)
fs/xfs/xfs_dir2.h +3 -2
index 768ddad..9910401 100644
@@ -70,8 +70,9 @@ extern void xfs_dir2_sf_put_parent_ino(struct xfs_dir2_sf_hdr *sfp,
 		xfs_ino_t ino);
 extern xfs_ino_t xfs_dir3_sfe_get_ino(struct xfs_mount *mp,
 		struct xfs_dir2_sf_hdr *sfp, struct xfs_dir2_sf_entry *sfep);
-extern void xfs_dir2_sfe_put_ino( struct xfs_dir2_sf_hdr *,
-		struct xfs_dir2_sf_entry *sfep, xfs_ino_t ino);
+extern void xfs_dir3_sfe_put_ino(struct xfs_mount *mp,
+		struct xfs_dir2_sf_hdr *hdr, struct xfs_dir2_sf_entry *sfep,
+		xfs_ino_t ino);
 
 extern int xfs_dir2_isblock(struct xfs_trans *tp, struct xfs_inode *dp, int *r);
 extern int xfs_dir2_isleaf(struct xfs_trans *tp, struct xfs_inode *dp, int *r);
fs/xfs/xfs_dir2_sf.c +1 -1
index 579b56b..bb6e284 100644
@@ -129,7 +129,7 @@ xfs_dir3_sfe_get_ino(
 	return xfs_dir2_sf_get_ino(hdr, xfs_dir3_sfe_inop(mp, sfep));
 }
 
-static void
+void
 xfs_dir3_sfe_put_ino(
 	struct xfs_mount	*mp,
 	struct xfs_dir2_sf_hdr	*hdr,

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

Reviewed-by on v2 1/1
#4 lore raw

Re: [PATCH V2] xfs: dtype changed xfs_dir2_sfe_put_ino to xfs_dir3_sfe_put_ino

Hi Dave,

Quoted 26 lines · Dave Chinner
Quoted 5 lines · Dave Chinner

From: Dave Chinner <[email protected]>

So fix up the export in xfs_dir2.h that is needed by userspace.

Well, this patch now causes a compilation failure. Why? xfs_dir3_sfe_put_ino() got marked as static because it didn't have a matching export.

Perhaps we should be a little less hasty in merging auto-generated patches from random build-farms? The patch was wrong - fixing the prototype like in v1 of this patch was the correct thing to do....

So, the updated patch below adds the revert.

Cheers,

Dave.

PS: Ben, when you merge a patch from a reply in an email thread, can you please pull the commit subject line from the patch, not the title of the email thread? One of the recent commits got it wrong. i.e. commit 3780437

Quoted 5 lines

"XFS: Assertion failed: first <= last && last < BBTOB(bp->b_length), file:"

should have had the title

"xfs: fix calculation of the number of node entries in a dir3 node"

I'll keep an eye out, thanks.

Quoted 17 lines
Signature · 16 lines
Dave Chinner
[email protected]

xfs: dtype changed xfs_dir2_sfe_put_ino to xfs_dir3_sfe_put_ino

From: Dave Chinner <[email protected]>

So fix up the export in xfs_dir2.h that is needed by userspace.

<sigh>

Now xfs_dir3_sfe_put_ino has been made static. Revert 98f7462 ("xfs:
xfs_dir3_sfe_put_ino can be static") to being non static so that the
code shared with userspace is identical again.

Signed-off-by: Dave Chinner <[email protected]>

Ah.. sorry I missed that it's used in your userspace series. Could you repost your userspace series soon?

Thanks,
Ben

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs

#5 lore raw

PATCH 2/2 xfs: XFS_MOUNT_QUOTA_ALL needed by userspace

In reply to Dave Chinner

From: Dave Chinner <[email protected]>

So move it to a header file shared with userspace.

---
 fs/xfs/xfs_log_format.h | 6 ++++++
 fs/xfs/xfs_quota.h      | 7 -------
 2 files changed, 6 insertions(+), 7 deletions(-)
fs/xfs/xfs_log_format.h +6 -0
index 5b7ba98..ca7e28a 100644
@@ -825,6 +825,12 @@ typedef struct xfs_qoff_logformat {
 #define XFS_ALL_QUOTA_CHKD	\
 		(XFS_UQUOTA_CHKD | XFS_GQUOTA_CHKD | XFS_PQUOTA_CHKD)
 
+#define XFS_MOUNT_QUOTA_ALL	(XFS_UQUOTA_ACCT|XFS_UQUOTA_ENFD|\
+				 XFS_UQUOTA_CHKD|XFS_GQUOTA_ACCT|\
+				 XFS_GQUOTA_ENFD|XFS_GQUOTA_CHKD|\
+				 XFS_PQUOTA_ACCT|XFS_PQUOTA_ENFD|\
+				 XFS_PQUOTA_CHKD)
+
 /*
  * Inode create log item structure
  *
fs/xfs/xfs_quota.h +0 -7
index 66522da..e7d84d2 100644
@@ -48,13 +48,6 @@ struct xfs_trans;
 	 (XFS_IS_PQUOTA_ON(mp) && \
 		(mp->m_sb.sb_qflags & XFS_PQUOTA_CHKD) == 0))
 
-#define XFS_MOUNT_QUOTA_ALL	(XFS_UQUOTA_ACCT|XFS_UQUOTA_ENFD|\
-				 XFS_UQUOTA_CHKD|XFS_GQUOTA_ACCT|\
-				 XFS_GQUOTA_ENFD|XFS_GQUOTA_CHKD|\
-				 XFS_PQUOTA_ACCT|XFS_PQUOTA_ENFD|\
-				 XFS_PQUOTA_CHKD)
-
-
 /*
  * The structure kept inside the xfs_trans_t keep track of dquot changes
  * within a transaction and apply them later.
Signature · 6 lines
1.8.3.2

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Reviewed-by on v1 2/2
#6 lore raw

Re: [PATCH 2/2] xfs: XFS_MOUNT_QUOTA_ALL needed by userspace

Quoted 6 lines · Dave Chinner

From: Dave Chinner<[email protected]>

So move it to a header file shared with userspace.

---

xfs_quota.h is kernel only. xfs_format.h is the correct header.

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs