[PATCH 0/4] xfsprogs: more misc fixes

[PATCH 0/4] xfsprogs: more misc fixes

[Darrick J. Wong]

Hi all,

Here's a bunch more patches for xfsprogs 4.16.

The first patch implements a custom ifork verifier for repair phase 6
so that we can mark directories with invalid parents and later _iget
them so that we can reattach them to whatever parent we find (or
lost+found).

The next couple of patches fix some mountpoint finding bugs so that
users can pass "/home" or "/home/" to programs that take file paths.

The fourth patch enables sparse inodes by default in mkfs.

This probably won't eat your data, and the branch[1] should apply
against for-next.  More patches tomorrow.

--D

[1] https://git.kernel.org/cgit/linux/kernel/git/djwong/xfsprogs-dev.git/log/?h=djwong-devel

[PATCH 1/4] xfs_repair: implement custom ifork verifiers

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

There are a few cases where an early stage of xfs_repair will write an
invalid inode fork buffer to signal to a later stage that it needs to
correct the value.  This happens in phase 4 when we detect an inline
format directory with an invalid .. pointer.  To avoid triggering the
ifork verifiers on this, inject a custom verifier for phase 6 that lets
this pass for now.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 libxfs/libxfs_api_defs.h |    2 +
 repair/phase6.c          |   66 +++++++++++++++++++++++++++++++++++++++++++++-
 2 files changed, 67 insertions(+), 1 deletion(-)


diff --git a/libxfs/libxfs_api_defs.h b/libxfs/libxfs_api_defs.h
index 5d56340..78daca0 100644
--- a/libxfs/libxfs_api_defs.h
+++ b/libxfs/libxfs_api_defs.h
@@ -150,5 +150,7 @@
 #define xfs_rmap_lookup_le_range	libxfs_rmap_lookup_le_range
 #define xfs_refc_block			libxfs_refc_block
 #define xfs_rmap_compare		libxfs_rmap_compare
+#define xfs_dir_get_ops			libxfs_dir_get_ops
+#define xfs_default_ifork_ops		libxfs_default_ifork_ops
 
 #endif /* __LIBXFS_API_DEFS_H__ */
diff --git a/repair/phase6.c b/repair/phase6.c
index aff83bc..e9189af 100644
--- a/repair/phase6.c
+++ b/repair/phase6.c
@@ -39,6 +39,70 @@ static struct xfs_name		xfs_name_dot = {(unsigned char *)".",
 						XFS_DIR3_FT_DIR};
 
 /*
+ * When we're checking directory inodes, we're allowed to set a directory's
+ * dotdot entry to zero to signal that the parent needs to be reconnected
+ * during phase 6.  The ifork verifiers would normally fail that, but we'll
+ * accept this canary so that we can fix the dir.
+ */
+static xfs_failaddr_t
+phase6_verify_dir(
+	struct xfs_inode		*ip)
+{
+	struct xfs_mount		*mp = ip->i_mount;
+	const struct xfs_dir_ops	*dops;
+	struct xfs_ifork		*ifp;
+	struct xfs_dir2_sf_hdr		*sfp;
+	xfs_failaddr_t			fa;
+	xfs_ino_t			old_parent;
+	bool				parent_bypass = false;
+	int				size;
+
+	dops = libxfs_dir_get_ops(mp, NULL);
+
+	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+	sfp = (struct xfs_dir2_sf_hdr *)ifp->if_u1.if_data;
+	size = ifp->if_bytes;
+
+	/* Don't let the NULLFSINO .. entry blow everything up. */
+	if (size > offsetof(struct xfs_dir2_sf_hdr, parent) &&
+	    size >= xfs_dir2_sf_hdr_size(sfp->i8count)) {
+		old_parent = dops->sf_get_parent_ino(sfp);
+		if (old_parent == 0) {
+			dops->sf_put_parent_ino(sfp, mp->m_sb.sb_rootino);
+			parent_bypass = true;
+		}
+	}
+
+	fa = libxfs_default_ifork_ops.verify_dir(ip);
+
+	/* Put it back. */
+	if (parent_bypass)
+		dops->sf_put_parent_ino(sfp, old_parent);
+
+	return fa;
+}
+
+static xfs_failaddr_t
+phase6_verify_attr(
+	struct xfs_inode		*ip)
+{
+	return libxfs_default_ifork_ops.verify_attr(ip);
+}
+
+static xfs_failaddr_t
+phase6_verify_symlink(
+	struct xfs_inode		*ip)
+{
+	return libxfs_default_ifork_ops.verify_symlink(ip);
+}
+
+struct xfs_ifork_ops phase6_default_ifork_ops = {
+	.verify_attr	= phase6_verify_attr,
+	.verify_dir	= phase6_verify_dir,
+	.verify_symlink	= phase6_verify_symlink,
+};
+
+/*
  * Data structures used to keep track of directories where the ".."
  * entries are updated. These must be rebuilt after the initial pass
  */
@@ -2833,7 +2897,7 @@ process_dir_inode(
 
 	ASSERT(!is_inode_refchecked(irec, ino_offset) || dotdot_update);
 
-	error = -libxfs_iget(mp, NULL, ino, 0, &ip, NULL);
+	error = -libxfs_iget(mp, NULL, ino, 0, &ip, &phase6_default_ifork_ops);
 	if (error) {
 		if (!no_modify)
 			do_error(

[PATCH 2/4] libfrog: fs_table_lookup_mount should realpath the argument

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Call realpath on the dir argument so that we're comparing canonical
paths when looking for the mountpoint.  This fixes the problem where
'/home/' doesn't match '/home' even though they refer to the same thing.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 libfrog/paths.c |    5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)


diff --git a/libfrog/paths.c b/libfrog/paths.c
index 62b4eda..19ee1ea 100644
--- a/libfrog/paths.c
+++ b/libfrog/paths.c
@@ -102,16 +102,19 @@ fs_table_lookup_mount(
 	uint		i;
 	dev_t		dev = 0;
 	char		rpath[PATH_MAX];
+	char		dpath[PATH_MAX];
 
 	if (fs_device_number(dir, &dev))
 		return NULL;
+	if (!realpath(dir, dpath))
+		return NULL;
 
 	for (i = 0; i < fs_count; i++) {
 		if (fs_table[i].fs_flags != FS_MOUNT_POINT)
 			continue;
 		if (!realpath(fs_table[i].fs_dir, rpath))
 			continue;
-		if (strcmp(rpath, dir) == 0)
+		if (strcmp(rpath, dpath) == 0)
 			return &fs_table[i];
 	}
 	return NULL;

[PATCH 3/4] xfs_fsr: refactor mountpoint finding to use libfrog paths functions

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Refactor the mount-point finding code in fsr to use the libfrog helpers
instead of open-coding yet another routine.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 fsr/Makefile    |    4 ++-
 fsr/xfs_fsr.c   |   71 +++++--------------------------------------------------
 include/path.h  |    1 +
 libfrog/paths.c |   47 ++++++++++++++++++++++++++++--------
 4 files changed, 47 insertions(+), 76 deletions(-)


diff --git a/fsr/Makefile b/fsr/Makefile
index d3521b2..fc1c33b 100644
--- a/fsr/Makefile
+++ b/fsr/Makefile
@@ -7,7 +7,9 @@ include $(TOPDIR)/include/builddefs
 
 LTCOMMAND = xfs_fsr
 CFILES = xfs_fsr.c
-LLDLIBS = $(LIBHANDLE)
+LLDLIBS = $(LIBHANDLE) $(LIBFROG)
+LTDEPENDENCIES = $(LIBFROG)
+LLDFLAGS = -static-libtool-libs
 
 ifeq ($(HAVE_GETMNTENT),yes)
 LCFLAGS += -DHAVE_GETMNTENT
diff --git a/fsr/xfs_fsr.c b/fsr/xfs_fsr.c
index 2a18ce0..ef6a68f 100644
--- a/fsr/xfs_fsr.c
+++ b/fsr/xfs_fsr.c
@@ -22,6 +22,7 @@
 #include "jdm.h"
 #include "xfs_bmap_btree.h"
 #include "xfs_attr_sf.h"
+#include "path.h"
 
 #include <fcntl.h>
 #include <errno.h>
@@ -167,73 +168,13 @@ aborter(int unused)
 	exit(1);
 }
 
-/*
- * Check if the argument is either the device name or mountpoint of an XFS
- * filesystem.  Note that we do not care about bind mounted regular files
- * here - the code that handles defragmentation of invidual files takes care
- * of that.
- */
-static char *
-find_mountpoint_check(struct stat *sb, struct mntent *t)
-{
-	struct stat ms;
-
-	if (S_ISDIR(sb->st_mode)) {		/* mount point */
-		if (stat(t->mnt_dir, &ms) < 0)
-			return NULL;
-		if (sb->st_ino != ms.st_ino)
-			return NULL;
-		if (sb->st_dev != ms.st_dev)
-			return NULL;
-		if (strcmp(t->mnt_type, MNTTYPE_XFS) != 0)
-			return NULL;
-	} else {				/* device */
-		if (stat(t->mnt_fsname, &ms) < 0)
-			return NULL;
-		if (sb->st_rdev != ms.st_rdev)
-			return NULL;
-		if (strcmp(t->mnt_type, MNTTYPE_XFS) != 0)
-			return NULL;
-		/*
-		 * Make sure the mountpoint given by mtab is accessible
-		 * before using it.
-		 */
-		if (stat(t->mnt_dir, &ms) < 0)
-			return NULL;
-	}
-
-	return t->mnt_dir;
-}
-
-static char *
-find_mountpoint(char *mtab, char *argname, struct stat *sb)
-{
-	struct mntent_cursor cursor;
-	struct mntent *t = NULL;
-	char *mntp = NULL;
-
-	if (platform_mntent_open(&cursor, mtab) != 0){
-		fprintf(stderr, "Error: can't get mntent entries.\n");
-		exit(1);
-	}
-
-	while ((t = platform_mntent_next(&cursor)) != NULL) {
-		mntp = find_mountpoint_check(sb, t);
-		if (mntp == NULL)
-			continue;
-		break;
-	}
-	platform_mntent_close(&cursor);
-	return mntp;
-}
-
 int
 main(int argc, char **argv)
 {
 	struct stat sb;
 	char *argname;
 	int c;
-	char *mntp;
+	struct fs_path	*fsp;
 	char *mtab = NULL;
 
 	setlinebuf(stdout);
@@ -343,9 +284,11 @@ main(int argc, char **argv)
 				sb = sb2;
 			}
 
-			mntp = find_mountpoint(mtab, argname, &sb);
-			if (mntp != NULL) {
-				fsrfs(mntp, 0, 100);
+			fsp = fs_table_lookup_mount(argname);
+			if (!fsp)
+				fsp = fs_table_lookup_blkdev(argname);
+			if (fsp != NULL) {
+				fsrfs(fsp->fs_dir, 0, 100);
 			} else if (S_ISCHR(sb.st_mode)) {
 				fprintf(stderr, _(
 					"%s: char special not supported: %s\n"),
diff --git a/include/path.h b/include/path.h
index 1d3a902..88dc44b 100644
--- a/include/path.h
+++ b/include/path.h
@@ -57,6 +57,7 @@ extern void fs_table_insert_project_path(char *__dir, uint __projid);
 
 extern fs_path_t *fs_table_lookup(const char *__dir, uint __flags);
 extern fs_path_t *fs_table_lookup_mount(const char *__dir);
+extern fs_path_t *fs_table_lookup_blkdev(const char *bdev);
 
 typedef struct fs_cursor {
 	uint		count;		/* total count of mount entries	*/
diff --git a/libfrog/paths.c b/libfrog/paths.c
index 19ee1ea..a80a30b 100644
--- a/libfrog/paths.c
+++ b/libfrog/paths.c
@@ -89,15 +89,10 @@ fs_table_lookup(
 	return NULL;
 }
 
-/*
- * Find the FS table entry describing an actual mount for the given path.
- * Unlike fs_table_lookup(), fs_table_lookup_mount() compares the "dir"
- * argument to actual mount point entries in the table. Accordingly, it
- * will find matches only if the "dir" argument is indeed mounted.
- */
-struct fs_path *
-fs_table_lookup_mount(
-	const char	*dir)
+static struct fs_path *
+__fs_table_lookup_mount(
+	const char	*dir,
+	const char	*blkdev)
 {
 	uint		i;
 	dev_t		dev = 0;
@@ -106,13 +101,17 @@ fs_table_lookup_mount(
 
 	if (fs_device_number(dir, &dev))
 		return NULL;
-	if (!realpath(dir, dpath))
+	if (dir && !realpath(dir, dpath))
+		return NULL;
+	if (blkdev && !realpath(blkdev, dpath))
 		return NULL;
 
 	for (i = 0; i < fs_count; i++) {
 		if (fs_table[i].fs_flags != FS_MOUNT_POINT)
 			continue;
-		if (!realpath(fs_table[i].fs_dir, rpath))
+		if (dir && !realpath(fs_table[i].fs_dir, rpath))
+			continue;
+		if (blkdev && !realpath(fs_table[i].fs_name, rpath))
 			continue;
 		if (strcmp(rpath, dpath) == 0)
 			return &fs_table[i];
@@ -120,6 +119,32 @@ fs_table_lookup_mount(
 	return NULL;
 }
 
+/*
+ * Find the FS table entry describing an actual mount for the given path.
+ * Unlike fs_table_lookup(), fs_table_lookup_mount() compares the "dir"
+ * argument to actual mount point entries in the table. Accordingly, it
+ * will find matches only if the "dir" argument is indeed mounted.
+ */
+struct fs_path *
+fs_table_lookup_mount(
+	const char	*dir)
+{
+	return __fs_table_lookup_mount(dir, NULL);
+}
+
+/*
+ * Find the FS table entry describing an actual mount for the block device.
+ * Unlike fs_table_lookup(), fs_table_lookup_blkdev() compares the "bdev"
+ * argument to actual mount point names in the table. Accordingly, it
+ * will find matches only if the "bdev" argument is indeed mounted.
+ */
+struct fs_path *
+fs_table_lookup_blkdev(
+	const char	*bdev)
+{
+	return __fs_table_lookup_mount(NULL, bdev);
+}
+
 static int
 fs_table_insert(
 	char		*dir,

[PATCH 4/4] mkfs: enable sparse inodes by default

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Enable the sparse inode feature by default.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 mkfs/xfs_mkfs.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)


diff --git a/mkfs/xfs_mkfs.c b/mkfs/xfs_mkfs.c
index 1ca6a2d..78d0ce5 100644
--- a/mkfs/xfs_mkfs.c
+++ b/mkfs/xfs_mkfs.c
@@ -1996,7 +1996,7 @@ _("finobt not supported without CRC support\n"));
 		}
 		cli->sb_feat.finobt = false;
 
-		if (cli->sb_feat.spinodes) {
+		if (cli->sb_feat.spinodes && cli_opt_set(&iopts, I_SPINODES)) {
 			fprintf(stderr,
 _("sparse inodes not supported without CRC support\n"));
 			usage();
@@ -3811,7 +3811,7 @@ main(
 			.crcs_enabled = true,
 			.dirftype = true,
 			.finobt = true,
-			.spinodes = false,
+			.spinodes = true,
 			.rmapbt = false,
 			.reflink = false,
 			.parent_pointers = false,

Re: [PATCH 1/4] xfs_repair: implement custom ifork verifiers

[Eric Sandeen]

On 3/19/18 10:08 PM, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
> 
> There are a few cases where an early stage of xfs_repair will write an
> invalid inode fork buffer to signal to a later stage that it needs to
> correct the value.  This happens in phase 4 when we detect an inline
> format directory with an invalid .. pointer.  To avoid triggering the
> ifork verifiers on this, inject a custom verifier for phase 6 that lets
> this pass for now.
> 
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> ---
>  libxfs/libxfs_api_defs.h |    2 +
>  repair/phase6.c          |   66 +++++++++++++++++++++++++++++++++++++++++++++-
>  2 files changed, 67 insertions(+), 1 deletion(-)
> 
> 
> diff --git a/libxfs/libxfs_api_defs.h b/libxfs/libxfs_api_defs.h
> index 5d56340..78daca0 100644
> --- a/libxfs/libxfs_api_defs.h
> +++ b/libxfs/libxfs_api_defs.h
> @@ -150,5 +150,7 @@
>  #define xfs_rmap_lookup_le_range	libxfs_rmap_lookup_le_range
>  #define xfs_refc_block			libxfs_refc_block
>  #define xfs_rmap_compare		libxfs_rmap_compare
> +#define xfs_dir_get_ops			libxfs_dir_get_ops
> +#define xfs_default_ifork_ops		libxfs_default_ifork_ops
>  
>  #endif /* __LIBXFS_API_DEFS_H__ */
> diff --git a/repair/phase6.c b/repair/phase6.c
> index aff83bc..e9189af 100644
> --- a/repair/phase6.c
> +++ b/repair/phase6.c
> @@ -39,6 +39,70 @@ static struct xfs_name		xfs_name_dot = {(unsigned char *)".",
>  						XFS_DIR3_FT_DIR};
>  
>  /*
> + * When we're checking directory inodes, we're allowed to set a directory's

(a shortform directory only?)

> + * dotdot entry to zero to signal that the parent needs to be reconnected
> + * during phase 6.  The ifork verifiers would normally fail that, but we'll
> + * accept this canary so that we can fix the dir.

hm we actually just replace it temporarily, potato/potahto?

> + */
> +static xfs_failaddr_t
> +phase6_verify_dir(
> +	struct xfs_inode		*ip)
> +{
> +	struct xfs_mount		*mp = ip->i_mount;
> +	const struct xfs_dir_ops	*dops;
> +	struct xfs_ifork		*ifp;
> +	struct xfs_dir2_sf_hdr		*sfp;
> +	xfs_failaddr_t			fa;
> +	xfs_ino_t			old_parent;
> +	bool				parent_bypass = false;
> +	int				size;
> +
> +	dops = libxfs_dir_get_ops(mp, NULL);
> +
> +	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
> +	sfp = (struct xfs_dir2_sf_hdr *)ifp->if_u1.if_data;
> +	size = ifp->if_bytes;
> +
> +	/* Don't let the NULLFSINO .. entry blow everything up. */

NULLFSINO is ((xfs_ino_t)-1) not zero, so is this comment accurate?

Maybe an explicit comment here about this being for shortform dirs?

	/*
	 * If this is a shortform directory, phase4 may have set the parent
	 * inode to zero to indicate that it must be fixed.  Temporarily
	 * set a valid parent so that the directory verifier will pass.
	 */

> +	if (size > offsetof(struct xfs_dir2_sf_hdr, parent) &&
> +	    size >= xfs_dir2_sf_hdr_size(sfp->i8count)) {
> +		old_parent = dops->sf_get_parent_ino(sfp);
> +		if (old_parent == 0) {
> +			dops->sf_put_parent_ino(sfp, mp->m_sb.sb_rootino);
> +			parent_bypass = true;
> +		}
> +	}
> +
> +	fa = libxfs_default_ifork_ops.verify_dir(ip);
> +
> +	/* Put it back. */

	/* Put the special parent == 0 back in place */

> +	if (parent_bypass)
> +		dops->sf_put_parent_ino(sfp, old_parent);
> +
> +	return fa;
> +}
> +
> +static xfs_failaddr_t
> +phase6_verify_attr(
> +	struct xfs_inode		*ip)
> +{
> +	return libxfs_default_ifork_ops.verify_attr(ip);
> +}

Is there a reason for these wrappers vs. just populating the
custom ifork_ops with xfs_attr_shortform_verify and
xfs_symlink_shortform_verify?

> +
> +static xfs_failaddr_t
> +phase6_verify_symlink(
> +	struct xfs_inode		*ip)
> +{
> +	return libxfs_default_ifork_ops.verify_symlink(ip);
> +}
> +
> +struct xfs_ifork_ops phase6_default_ifork_ops = {

Naming a "custom" verifier "default" seems counterintuitive,
is there a reason for the "default" semantics I'm missing?  Not
a huge deal, just makes me go "hmmm...."

> +	.verify_attr	= phase6_verify_attr,
> +	.verify_dir	= phase6_verify_dir,
> +	.verify_symlink	= phase6_verify_symlink,
> +};
> +
> +/*
>   * Data structures used to keep track of directories where the ".."
>   * entries are updated. These must be rebuilt after the initial pass
>   */
> @@ -2833,7 +2897,7 @@ process_dir_inode(
>  
>  	ASSERT(!is_inode_refchecked(irec, ino_offset) || dotdot_update);
>  
> -	error = -libxfs_iget(mp, NULL, ino, 0, &ip, NULL);
> +	error = -libxfs_iget(mp, NULL, ino, 0, &ip, &phase6_default_ifork_ops);
>  	if (error) {
>  		if (!no_modify)
>  			do_error(
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

Re: [PATCH 2/4] libfrog: fs_table_lookup_mount should realpath the argument

[Eric Sandeen]

On 3/19/18 10:08 PM, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
> 
> Call realpath on the dir argument so that we're comparing canonical
> paths when looking for the mountpoint.  This fixes the problem where
> '/home/' doesn't match '/home' even though they refer to the same thing.
> 
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>

Yeah, not sure why this wasn't done.  This stuff always seems fiddly
but I can't see anything wrong with doing this, soooo

Reviewed-by: Eric Sandeen <sandeen@redhat.com>

> ---
>  libfrog/paths.c |    5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> 
> diff --git a/libfrog/paths.c b/libfrog/paths.c
> index 62b4eda..19ee1ea 100644
> --- a/libfrog/paths.c
> +++ b/libfrog/paths.c
> @@ -102,16 +102,19 @@ fs_table_lookup_mount(
>  	uint		i;
>  	dev_t		dev = 0;
>  	char		rpath[PATH_MAX];
> +	char		dpath[PATH_MAX];
>  
>  	if (fs_device_number(dir, &dev))
>  		return NULL;
> +	if (!realpath(dir, dpath))
> +		return NULL;
>  
>  	for (i = 0; i < fs_count; i++) {
>  		if (fs_table[i].fs_flags != FS_MOUNT_POINT)
>  			continue;
>  		if (!realpath(fs_table[i].fs_dir, rpath))
>  			continue;
> -		if (strcmp(rpath, dir) == 0)
> +		if (strcmp(rpath, dpath) == 0)
>  			return &fs_table[i];
>  	}
>  	return NULL;
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

Re: [PATCH 4/4] mkfs: enable sparse inodes by default

[Eric Sandeen]

On 3/19/18 10:08 PM, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
> 
> Enable the sparse inode feature by default.
> 
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>

OK - the cli_opt_set stuff feels clunky but I know it's just the pattern
right now...

Reviewed-by: Eric Sandeen <sandeen@redhat.com>

> ---
>  mkfs/xfs_mkfs.c |    4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> 
> diff --git a/mkfs/xfs_mkfs.c b/mkfs/xfs_mkfs.c
> index 1ca6a2d..78d0ce5 100644
> --- a/mkfs/xfs_mkfs.c
> +++ b/mkfs/xfs_mkfs.c
> @@ -1996,7 +1996,7 @@ _("finobt not supported without CRC support\n"));
>  		}
>  		cli->sb_feat.finobt = false;
>  
> -		if (cli->sb_feat.spinodes) {
> +		if (cli->sb_feat.spinodes && cli_opt_set(&iopts, I_SPINODES)) {
>  			fprintf(stderr,
>  _("sparse inodes not supported without CRC support\n"));
>  			usage();
> @@ -3811,7 +3811,7 @@ main(
>  			.crcs_enabled = true,
>  			.dirftype = true,
>  			.finobt = true,
> -			.spinodes = false,
> +			.spinodes = true,
>  			.rmapbt = false,
>  			.reflink = false,
>  			.parent_pointers = false,
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

Re: [PATCH 1/4] xfs_repair: implement custom ifork verifiers

[Darrick J. Wong]

On Tue, Mar 20, 2018 at 02:54:30PM -0500, Eric Sandeen wrote:
> On 3/19/18 10:08 PM, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> > 
> > There are a few cases where an early stage of xfs_repair will write an
> > invalid inode fork buffer to signal to a later stage that it needs to
> > correct the value.  This happens in phase 4 when we detect an inline
> > format directory with an invalid .. pointer.  To avoid triggering the
> > ifork verifiers on this, inject a custom verifier for phase 6 that lets
> > this pass for now.
> > 
> > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> > ---
> >  libxfs/libxfs_api_defs.h |    2 +
> >  repair/phase6.c          |   66 +++++++++++++++++++++++++++++++++++++++++++++-
> >  2 files changed, 67 insertions(+), 1 deletion(-)
> > 
> > 
> > diff --git a/libxfs/libxfs_api_defs.h b/libxfs/libxfs_api_defs.h
> > index 5d56340..78daca0 100644
> > --- a/libxfs/libxfs_api_defs.h
> > +++ b/libxfs/libxfs_api_defs.h
> > @@ -150,5 +150,7 @@
> >  #define xfs_rmap_lookup_le_range	libxfs_rmap_lookup_le_range
> >  #define xfs_refc_block			libxfs_refc_block
> >  #define xfs_rmap_compare		libxfs_rmap_compare
> > +#define xfs_dir_get_ops			libxfs_dir_get_ops
> > +#define xfs_default_ifork_ops		libxfs_default_ifork_ops
> >  
> >  #endif /* __LIBXFS_API_DEFS_H__ */
> > diff --git a/repair/phase6.c b/repair/phase6.c
> > index aff83bc..e9189af 100644
> > --- a/repair/phase6.c
> > +++ b/repair/phase6.c
> > @@ -39,6 +39,70 @@ static struct xfs_name		xfs_name_dot = {(unsigned char *)".",
> >  						XFS_DIR3_FT_DIR};
> >  
> >  /*
> > + * When we're checking directory inodes, we're allowed to set a directory's
> 
> (a shortform directory only?)

I think we do it for any directory, but it's only the shortform dirs
that require this fix.

> > + * dotdot entry to zero to signal that the parent needs to be reconnected
> > + * during phase 6.  The ifork verifiers would normally fail that, but we'll
> > + * accept this canary so that we can fix the dir.
> 
> hm we actually just replace it temporarily, potato/potahto?
> 
> > + */
> > +static xfs_failaddr_t
> > +phase6_verify_dir(
> > +	struct xfs_inode		*ip)
> > +{
> > +	struct xfs_mount		*mp = ip->i_mount;
> > +	const struct xfs_dir_ops	*dops;
> > +	struct xfs_ifork		*ifp;
> > +	struct xfs_dir2_sf_hdr		*sfp;
> > +	xfs_failaddr_t			fa;
> > +	xfs_ino_t			old_parent;
> > +	bool				parent_bypass = false;
> > +	int				size;
> > +
> > +	dops = libxfs_dir_get_ops(mp, NULL);
> > +
> > +	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
> > +	sfp = (struct xfs_dir2_sf_hdr *)ifp->if_u1.if_data;
> > +	size = ifp->if_bytes;
> > +
> > +	/* Don't let the NULLFSINO .. entry blow everything up. */
> 
> NULLFSINO is ((xfs_ino_t)-1) not zero, so is this comment accurate?

Oops. :)

> Maybe an explicit comment here about this being for shortform dirs?
> 
> 	/*
> 	 * If this is a shortform directory, phase4 may have set the parent
> 	 * inode to zero to indicate that it must be fixed.  Temporarily
> 	 * set a valid parent so that the directory verifier will pass.
> 	 */

Much better comment, let's go with that.

> > +	if (size > offsetof(struct xfs_dir2_sf_hdr, parent) &&
> > +	    size >= xfs_dir2_sf_hdr_size(sfp->i8count)) {
> > +		old_parent = dops->sf_get_parent_ino(sfp);
> > +		if (old_parent == 0) {
> > +			dops->sf_put_parent_ino(sfp, mp->m_sb.sb_rootino);
> > +			parent_bypass = true;
> > +		}
> > +	}
> > +
> > +	fa = libxfs_default_ifork_ops.verify_dir(ip);
> > +
> > +	/* Put it back. */
> 
> 	/* Put the special parent == 0 back in place */
> 
> > +	if (parent_bypass)
> > +		dops->sf_put_parent_ino(sfp, old_parent);
> > +
> > +	return fa;
> > +}
> > +
> > +static xfs_failaddr_t
> > +phase6_verify_attr(
> > +	struct xfs_inode		*ip)
> > +{
> > +	return libxfs_default_ifork_ops.verify_attr(ip);
> > +}
> 
> Is there a reason for these wrappers vs. just populating the
> custom ifork_ops with xfs_attr_shortform_verify and
> xfs_symlink_shortform_verify?

gcc whines about non-const expressions.  I tried adding const to
everything that touches an ifork_ops but it still wouldn't compile.

> > +
> > +static xfs_failaddr_t
> > +phase6_verify_symlink(
> > +	struct xfs_inode		*ip)
> > +{
> > +	return libxfs_default_ifork_ops.verify_symlink(ip);
> > +}
> > +
> > +struct xfs_ifork_ops phase6_default_ifork_ops = {
> 
> Naming a "custom" verifier "default" seems counterintuitive,
> is there a reason for the "default" semantics I'm missing?  Not
> a huge deal, just makes me go "hmmm...."

-EBADNAME

phase6_ifork_ops, much better.

--D

> > +	.verify_attr	= phase6_verify_attr,
> > +	.verify_dir	= phase6_verify_dir,
> > +	.verify_symlink	= phase6_verify_symlink,
> > +};
> > +
> > +/*
> >   * Data structures used to keep track of directories where the ".."
> >   * entries are updated. These must be rebuilt after the initial pass
> >   */
> > @@ -2833,7 +2897,7 @@ process_dir_inode(
> >  
> >  	ASSERT(!is_inode_refchecked(irec, ino_offset) || dotdot_update);
> >  
> > -	error = -libxfs_iget(mp, NULL, ino, 0, &ip, NULL);
> > +	error = -libxfs_iget(mp, NULL, ino, 0, &ip, &phase6_default_ifork_ops);
> >  	if (error) {
> >  		if (!no_modify)
> >  			do_error(
> > 
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v2 1/4] xfs_repair: implement custom ifork verifiers

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

There are a few cases where an early stage of xfs_repair will write an
invalid inode fork buffer to signal to a later stage that it needs to
correct the value.  This happens in phase 4 when we detect an inline
format directory with an invalid .. pointer.  To avoid triggering the
ifork verifiers on this, inject a custom verifier for phase 6 that lets
this pass for now.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
v2: fix misleading comments, remove pointless wrappers
---
 libxfs/libxfs_api_defs.h |    2 ++
 repair/phase6.c          |   57 +++++++++++++++++++++++++++++++++++++++++++++-
 2 files changed, 58 insertions(+), 1 deletion(-)

diff --git a/libxfs/libxfs_api_defs.h b/libxfs/libxfs_api_defs.h
index 5d56340..78daca0 100644
--- a/libxfs/libxfs_api_defs.h
+++ b/libxfs/libxfs_api_defs.h
@@ -150,5 +150,7 @@
 #define xfs_rmap_lookup_le_range	libxfs_rmap_lookup_le_range
 #define xfs_refc_block			libxfs_refc_block
 #define xfs_rmap_compare		libxfs_rmap_compare
+#define xfs_dir_get_ops			libxfs_dir_get_ops
+#define xfs_default_ifork_ops		libxfs_default_ifork_ops
 
 #endif /* __LIBXFS_API_DEFS_H__ */
diff --git a/repair/phase6.c b/repair/phase6.c
index aff83bc..ed005e8 100644
--- a/repair/phase6.c
+++ b/repair/phase6.c
@@ -39,6 +39,61 @@ static struct xfs_name		xfs_name_dot = {(unsigned char *)".",
 						XFS_DIR3_FT_DIR};
 
 /*
+ * When we're checking directory inodes, we're allowed to set a directory's
+ * dotdot entry to zero to signal that the parent needs to be reconnected
+ * during phase 6.  If we're handling a shortform directory the ifork
+ * verifiers will fail, so temporarily patch out this canary so that we can
+ * verify the rest of the fork and move on to fixing the dir.
+ */
+static xfs_failaddr_t
+phase6_verify_dir(
+	struct xfs_inode		*ip)
+{
+	struct xfs_mount		*mp = ip->i_mount;
+	const struct xfs_dir_ops	*dops;
+	struct xfs_ifork		*ifp;
+	struct xfs_dir2_sf_hdr		*sfp;
+	xfs_failaddr_t			fa;
+	xfs_ino_t			old_parent;
+	bool				parent_bypass = false;
+	int				size;
+
+	dops = libxfs_dir_get_ops(mp, NULL);
+
+	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+	sfp = (struct xfs_dir2_sf_hdr *)ifp->if_u1.if_data;
+	size = ifp->if_bytes;
+
+	/*
+	 * If this is a shortform directory, phase4 may have set the parent
+	 * inode to zero to indicate that it must be fixed.  Temporarily
+	 * set a valid parent so that the directory verifier will pass.
+	 */
+	if (size > offsetof(struct xfs_dir2_sf_hdr, parent) &&
+	    size >= xfs_dir2_sf_hdr_size(sfp->i8count)) {
+		old_parent = dops->sf_get_parent_ino(sfp);
+		if (old_parent == 0) {
+			dops->sf_put_parent_ino(sfp, mp->m_sb.sb_rootino);
+			parent_bypass = true;
+		}
+	}
+
+	fa = libxfs_default_ifork_ops.verify_dir(ip);
+
+	/* Put it back. */
+	if (parent_bypass)
+		dops->sf_put_parent_ino(sfp, old_parent);
+
+	return fa;
+}
+
+static struct xfs_ifork_ops phase6_ifork_ops = {
+	.verify_attr	= xfs_attr_shortform_verify,
+	.verify_dir	= phase6_verify_dir,
+	.verify_symlink	= xfs_symlink_shortform_verify,
+};
+
+/*
  * Data structures used to keep track of directories where the ".."
  * entries are updated. These must be rebuilt after the initial pass
  */
@@ -2833,7 +2888,7 @@ process_dir_inode(
 
 	ASSERT(!is_inode_refchecked(irec, ino_offset) || dotdot_update);
 
-	error = -libxfs_iget(mp, NULL, ino, 0, &ip, NULL);
+	error = -libxfs_iget(mp, NULL, ino, 0, &ip, &phase6_ifork_ops);
 	if (error) {
 		if (!no_modify)
 			do_error(

Re: [PATCH 3/4] xfs_fsr: refactor mountpoint finding to use libfrog paths functions

[Darrick J. Wong]

On Mon, Mar 19, 2018 at 08:08:38PM -0700, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
> 
> Refactor the mount-point finding code in fsr to use the libfrog helpers
> instead of open-coding yet another routine.
> 
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> ---
>  fsr/Makefile    |    4 ++-
>  fsr/xfs_fsr.c   |   71 +++++--------------------------------------------------
>  include/path.h  |    1 +
>  libfrog/paths.c |   47 ++++++++++++++++++++++++++++--------
>  4 files changed, 47 insertions(+), 76 deletions(-)
> 
> 
> diff --git a/fsr/Makefile b/fsr/Makefile
> index d3521b2..fc1c33b 100644
> --- a/fsr/Makefile
> +++ b/fsr/Makefile
> @@ -7,7 +7,9 @@ include $(TOPDIR)/include/builddefs
>  
>  LTCOMMAND = xfs_fsr
>  CFILES = xfs_fsr.c
> -LLDLIBS = $(LIBHANDLE)
> +LLDLIBS = $(LIBHANDLE) $(LIBFROG)
> +LTDEPENDENCIES = $(LIBFROG)
> +LLDFLAGS = -static-libtool-libs
>  
>  ifeq ($(HAVE_GETMNTENT),yes)
>  LCFLAGS += -DHAVE_GETMNTENT
> diff --git a/fsr/xfs_fsr.c b/fsr/xfs_fsr.c
> index 2a18ce0..ef6a68f 100644
> --- a/fsr/xfs_fsr.c
> +++ b/fsr/xfs_fsr.c
> @@ -22,6 +22,7 @@
>  #include "jdm.h"
>  #include "xfs_bmap_btree.h"
>  #include "xfs_attr_sf.h"
> +#include "path.h"
>  
>  #include <fcntl.h>
>  #include <errno.h>
> @@ -167,73 +168,13 @@ aborter(int unused)
>  	exit(1);
>  }
>  
> -/*
> - * Check if the argument is either the device name or mountpoint of an XFS
> - * filesystem.  Note that we do not care about bind mounted regular files
> - * here - the code that handles defragmentation of invidual files takes care
> - * of that.
> - */
> -static char *
> -find_mountpoint_check(struct stat *sb, struct mntent *t)
> -{
> -	struct stat ms;
> -
> -	if (S_ISDIR(sb->st_mode)) {		/* mount point */
> -		if (stat(t->mnt_dir, &ms) < 0)
> -			return NULL;
> -		if (sb->st_ino != ms.st_ino)
> -			return NULL;
> -		if (sb->st_dev != ms.st_dev)
> -			return NULL;
> -		if (strcmp(t->mnt_type, MNTTYPE_XFS) != 0)
> -			return NULL;
> -	} else {				/* device */
> -		if (stat(t->mnt_fsname, &ms) < 0)
> -			return NULL;
> -		if (sb->st_rdev != ms.st_rdev)
> -			return NULL;
> -		if (strcmp(t->mnt_type, MNTTYPE_XFS) != 0)
> -			return NULL;
> -		/*
> -		 * Make sure the mountpoint given by mtab is accessible
> -		 * before using it.
> -		 */
> -		if (stat(t->mnt_dir, &ms) < 0)
> -			return NULL;
> -	}
> -
> -	return t->mnt_dir;
> -}
> -
> -static char *
> -find_mountpoint(char *mtab, char *argname, struct stat *sb)
> -{
> -	struct mntent_cursor cursor;
> -	struct mntent *t = NULL;
> -	char *mntp = NULL;
> -
> -	if (platform_mntent_open(&cursor, mtab) != 0){
> -		fprintf(stderr, "Error: can't get mntent entries.\n");
> -		exit(1);
> -	}
> -
> -	while ((t = platform_mntent_next(&cursor)) != NULL) {
> -		mntp = find_mountpoint_check(sb, t);
> -		if (mntp == NULL)
> -			continue;
> -		break;
> -	}
> -	platform_mntent_close(&cursor);
> -	return mntp;
> -}
> -
>  int
>  main(int argc, char **argv)
>  {
>  	struct stat sb;
>  	char *argname;
>  	int c;
> -	char *mntp;
> +	struct fs_path	*fsp;
>  	char *mtab = NULL;
>  
>  	setlinebuf(stdout);
> @@ -343,9 +284,11 @@ main(int argc, char **argv)
>  				sb = sb2;
>  			}
>  
> -			mntp = find_mountpoint(mtab, argname, &sb);
> -			if (mntp != NULL) {
> -				fsrfs(mntp, 0, 100);
> +			fsp = fs_table_lookup_mount(argname);

Lovely, we call fs_table_lookup_mount without first calling
fs_table_initialise to initialize the mountpoint table.  Now this
program refuses directory and bdev arguments...

> +			if (!fsp)
> +				fsp = fs_table_lookup_blkdev(argname);
> +			if (fsp != NULL) {
> +				fsrfs(fsp->fs_dir, 0, 100);
>  			} else if (S_ISCHR(sb.st_mode)) {
>  				fprintf(stderr, _(
>  					"%s: char special not supported: %s\n"),
> diff --git a/include/path.h b/include/path.h
> index 1d3a902..88dc44b 100644
> --- a/include/path.h
> +++ b/include/path.h
> @@ -57,6 +57,7 @@ extern void fs_table_insert_project_path(char *__dir, uint __projid);
>  
>  extern fs_path_t *fs_table_lookup(const char *__dir, uint __flags);
>  extern fs_path_t *fs_table_lookup_mount(const char *__dir);
> +extern fs_path_t *fs_table_lookup_blkdev(const char *bdev);
>  
>  typedef struct fs_cursor {
>  	uint		count;		/* total count of mount entries	*/
> diff --git a/libfrog/paths.c b/libfrog/paths.c
> index 19ee1ea..a80a30b 100644
> --- a/libfrog/paths.c
> +++ b/libfrog/paths.c
> @@ -89,15 +89,10 @@ fs_table_lookup(
>  	return NULL;
>  }
>  
> -/*
> - * Find the FS table entry describing an actual mount for the given path.
> - * Unlike fs_table_lookup(), fs_table_lookup_mount() compares the "dir"
> - * argument to actual mount point entries in the table. Accordingly, it
> - * will find matches only if the "dir" argument is indeed mounted.
> - */
> -struct fs_path *
> -fs_table_lookup_mount(
> -	const char	*dir)
> +static struct fs_path *
> +__fs_table_lookup_mount(
> +	const char	*dir,
> +	const char	*blkdev)
>  {
>  	uint		i;
>  	dev_t		dev = 0;
> @@ -106,13 +101,17 @@ fs_table_lookup_mount(
>  
>  	if (fs_device_number(dir, &dev))
>  		return NULL;

...which is just as well, since dir can now be NULL which means that
fs_table_lookup_blkdev will always bail out here.

nak, patch is fubar.

--D

> -	if (!realpath(dir, dpath))
> +	if (dir && !realpath(dir, dpath))
> +		return NULL;
> +	if (blkdev && !realpath(blkdev, dpath))
>  		return NULL;
>  
>  	for (i = 0; i < fs_count; i++) {
>  		if (fs_table[i].fs_flags != FS_MOUNT_POINT)
>  			continue;
> -		if (!realpath(fs_table[i].fs_dir, rpath))
> +		if (dir && !realpath(fs_table[i].fs_dir, rpath))
> +			continue;
> +		if (blkdev && !realpath(fs_table[i].fs_name, rpath))
>  			continue;
>  		if (strcmp(rpath, dpath) == 0)
>  			return &fs_table[i];
> @@ -120,6 +119,32 @@ fs_table_lookup_mount(
>  	return NULL;
>  }
>  
> +/*
> + * Find the FS table entry describing an actual mount for the given path.
> + * Unlike fs_table_lookup(), fs_table_lookup_mount() compares the "dir"
> + * argument to actual mount point entries in the table. Accordingly, it
> + * will find matches only if the "dir" argument is indeed mounted.
> + */
> +struct fs_path *
> +fs_table_lookup_mount(
> +	const char	*dir)
> +{
> +	return __fs_table_lookup_mount(dir, NULL);
> +}
> +
> +/*
> + * Find the FS table entry describing an actual mount for the block device.
> + * Unlike fs_table_lookup(), fs_table_lookup_blkdev() compares the "bdev"
> + * argument to actual mount point names in the table. Accordingly, it
> + * will find matches only if the "bdev" argument is indeed mounted.
> + */
> +struct fs_path *
> +fs_table_lookup_blkdev(
> +	const char	*bdev)
> +{
> +	return __fs_table_lookup_mount(NULL, bdev);
> +}
> +
>  static int
>  fs_table_insert(
>  	char		*dir,
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH v2 3/4] xfs_fsr: refactor mountpoint finding to use libfrog paths functions

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Refactor the mount-point finding code in fsr to use the libfrog helpers
instead of open-coding yet another routine.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
v2: fix obvious brokenness w/ path testing
---
 fsr/Makefile    |    4 ++-
 fsr/xfs_fsr.c   |   73 ++++++-------------------------------------------------
 include/path.h  |    1 +
 libfrog/paths.c |   48 ++++++++++++++++++++++++++----------
 4 files changed, 47 insertions(+), 79 deletions(-)

diff --git a/fsr/Makefile b/fsr/Makefile
index d3521b2..4201b38 100644
--- a/fsr/Makefile
+++ b/fsr/Makefile
@@ -7,7 +7,9 @@ include $(TOPDIR)/include/builddefs
 
 LTCOMMAND = xfs_fsr
 CFILES = xfs_fsr.c
-LLDLIBS = $(LIBHANDLE)
+LLDLIBS = $(LIBHANDLE) $(LIBFROG) $(LIBPTHREAD) $(LIBBLKID)
+LTDEPENDENCIES = $(LIBHANDLE) $(LIBFROG)
+LLDFLAGS = -static-libtool-libs
 
 ifeq ($(HAVE_GETMNTENT),yes)
 LCFLAGS += -DHAVE_GETMNTENT
diff --git a/fsr/xfs_fsr.c b/fsr/xfs_fsr.c
index 2a18ce0..b74a70b 100644
--- a/fsr/xfs_fsr.c
+++ b/fsr/xfs_fsr.c
@@ -22,6 +22,7 @@
 #include "jdm.h"
 #include "xfs_bmap_btree.h"
 #include "xfs_attr_sf.h"
+#include "path.h"
 
 #include <fcntl.h>
 #include <errno.h>
@@ -167,73 +168,13 @@ aborter(int unused)
 	exit(1);
 }
 
-/*
- * Check if the argument is either the device name or mountpoint of an XFS
- * filesystem.  Note that we do not care about bind mounted regular files
- * here - the code that handles defragmentation of invidual files takes care
- * of that.
- */
-static char *
-find_mountpoint_check(struct stat *sb, struct mntent *t)
-{
-	struct stat ms;
-
-	if (S_ISDIR(sb->st_mode)) {		/* mount point */
-		if (stat(t->mnt_dir, &ms) < 0)
-			return NULL;
-		if (sb->st_ino != ms.st_ino)
-			return NULL;
-		if (sb->st_dev != ms.st_dev)
-			return NULL;
-		if (strcmp(t->mnt_type, MNTTYPE_XFS) != 0)
-			return NULL;
-	} else {				/* device */
-		if (stat(t->mnt_fsname, &ms) < 0)
-			return NULL;
-		if (sb->st_rdev != ms.st_rdev)
-			return NULL;
-		if (strcmp(t->mnt_type, MNTTYPE_XFS) != 0)
-			return NULL;
-		/*
-		 * Make sure the mountpoint given by mtab is accessible
-		 * before using it.
-		 */
-		if (stat(t->mnt_dir, &ms) < 0)
-			return NULL;
-	}
-
-	return t->mnt_dir;
-}
-
-static char *
-find_mountpoint(char *mtab, char *argname, struct stat *sb)
-{
-	struct mntent_cursor cursor;
-	struct mntent *t = NULL;
-	char *mntp = NULL;
-
-	if (platform_mntent_open(&cursor, mtab) != 0){
-		fprintf(stderr, "Error: can't get mntent entries.\n");
-		exit(1);
-	}
-
-	while ((t = platform_mntent_next(&cursor)) != NULL) {
-		mntp = find_mountpoint_check(sb, t);
-		if (mntp == NULL)
-			continue;
-		break;
-	}
-	platform_mntent_close(&cursor);
-	return mntp;
-}
-
 int
 main(int argc, char **argv)
 {
 	struct stat sb;
 	char *argname;
 	int c;
-	char *mntp;
+	struct fs_path	*fsp;
 	char *mtab = NULL;
 
 	setlinebuf(stdout);
@@ -322,7 +263,7 @@ main(int argc, char **argv)
 	RealUid = getuid();
 
 	pagesize = getpagesize();
-
+	fs_table_initialise(0, NULL, 0, NULL);
 	if (optind < argc) {
 		for (; optind < argc; optind++) {
 			argname = argv[optind];
@@ -343,9 +284,11 @@ main(int argc, char **argv)
 				sb = sb2;
 			}
 
-			mntp = find_mountpoint(mtab, argname, &sb);
-			if (mntp != NULL) {
-				fsrfs(mntp, 0, 100);
+			fsp = fs_table_lookup_mount(argname);
+			if (!fsp)
+				fsp = fs_table_lookup_blkdev(argname);
+			if (fsp != NULL) {
+				fsrfs(fsp->fs_dir, 0, 100);
 			} else if (S_ISCHR(sb.st_mode)) {
 				fprintf(stderr, _(
 					"%s: char special not supported: %s\n"),
diff --git a/include/path.h b/include/path.h
index 1d3a902..88dc44b 100644
--- a/include/path.h
+++ b/include/path.h
@@ -57,6 +57,7 @@ extern void fs_table_insert_project_path(char *__dir, uint __projid);
 
 extern fs_path_t *fs_table_lookup(const char *__dir, uint __flags);
 extern fs_path_t *fs_table_lookup_mount(const char *__dir);
+extern fs_path_t *fs_table_lookup_blkdev(const char *bdev);
 
 typedef struct fs_cursor {
 	uint		count;		/* total count of mount entries	*/
diff --git a/libfrog/paths.c b/libfrog/paths.c
index 19ee1ea..318b48f 100644
--- a/libfrog/paths.c
+++ b/libfrog/paths.c
@@ -89,30 +89,26 @@ fs_table_lookup(
 	return NULL;
 }
 
-/*
- * Find the FS table entry describing an actual mount for the given path.
- * Unlike fs_table_lookup(), fs_table_lookup_mount() compares the "dir"
- * argument to actual mount point entries in the table. Accordingly, it
- * will find matches only if the "dir" argument is indeed mounted.
- */
-struct fs_path *
-fs_table_lookup_mount(
-	const char	*dir)
+static struct fs_path *
+__fs_table_lookup_mount(
+	const char	*dir,
+	const char	*blkdev)
 {
 	uint		i;
-	dev_t		dev = 0;
 	char		rpath[PATH_MAX];
 	char		dpath[PATH_MAX];
 
-	if (fs_device_number(dir, &dev))
+	if (dir && !realpath(dir, dpath))
 		return NULL;
-	if (!realpath(dir, dpath))
+	if (blkdev && !realpath(blkdev, dpath))
 		return NULL;
 
 	for (i = 0; i < fs_count; i++) {
 		if (fs_table[i].fs_flags != FS_MOUNT_POINT)
 			continue;
-		if (!realpath(fs_table[i].fs_dir, rpath))
+		if (dir && !realpath(fs_table[i].fs_dir, rpath))
+			continue;
+		if (blkdev && !realpath(fs_table[i].fs_name, rpath))
 			continue;
 		if (strcmp(rpath, dpath) == 0)
 			return &fs_table[i];
@@ -120,6 +116,32 @@ fs_table_lookup_mount(
 	return NULL;
 }
 
+/*
+ * Find the FS table entry describing an actual mount for the given path.
+ * Unlike fs_table_lookup(), fs_table_lookup_mount() compares the "dir"
+ * argument to actual mount point entries in the table. Accordingly, it
+ * will find matches only if the "dir" argument is indeed mounted.
+ */
+struct fs_path *
+fs_table_lookup_mount(
+	const char	*dir)
+{
+	return __fs_table_lookup_mount(dir, NULL);
+}
+
+/*
+ * Find the FS table entry describing an actual mount for the block device.
+ * Unlike fs_table_lookup(), fs_table_lookup_blkdev() compares the "bdev"
+ * argument to actual mount point names in the table. Accordingly, it
+ * will find matches only if the "bdev" argument is indeed mounted.
+ */
+struct fs_path *
+fs_table_lookup_blkdev(
+	const char	*bdev)
+{
+	return __fs_table_lookup_mount(NULL, bdev);
+}
+
 static int
 fs_table_insert(
 	char		*dir,

[PATCH 5/4] misc: remove darwin, irix, and freebsd support

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Remove these ports since they're not actively maintained:

IRIX support was partially removed last year; the OS itself hasn't had a
release since 2006.

FreeBSD dropped XFS support in v9.3, which EOLd in January 2017.

Darwin/OSX has never supported XFS.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 include/darwin.h  |  315 -----------------------------------------------------
 include/freebsd.h |  210 -----------------------------------
 libxfs/Makefile   |    2 
 libxfs/darwin.c   |  144 ------------------------
 libxfs/freebsd.c  |  201 ----------------------------------
 libxfs/irix.c     |  111 -------------------
 6 files changed, 1 insertion(+), 982 deletions(-)
 delete mode 100644 include/darwin.h
 delete mode 100644 include/freebsd.h
 delete mode 100644 libxfs/darwin.c
 delete mode 100644 libxfs/freebsd.c
 delete mode 100644 libxfs/irix.c

diff --git a/include/darwin.h b/include/darwin.h
deleted file mode 100644
index 2632e1d..0000000
--- a/include/darwin.h
+++ /dev/null
@@ -1,315 +0,0 @@
-/*
- * Copyright (c) 2004-2006 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public License
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_DARWIN_H__
-#define __XFS_DARWIN_H__
-
-#include <unistd.h>
-#include <sys/types.h>
-#include <string.h>
-#include <uuid/uuid.h>
-#include <libgen.h>
-#include <sys/vm.h>
-#include <sys/stat.h>
-#include <sys/ioctl.h>
-#include <sys/mount.h>
-#include <sys/types.h>
-#include <sys/time.h>
-#include <ftw.h>
-#include <mach/mach_time.h>
-#include <inttypes.h>
-#include <stdio.h>
-#include <sys/mman.h>
-
-#include <machine/endian.h>
-#define __BYTE_ORDER	BYTE_ORDER
-#define __BIG_ENDIAN	BIG_ENDIAN
-#define __LITTLE_ENDIAN	LITTLE_ENDIAN
-
-#include <sys/syscall.h>
-# ifndef SYS_fsctl
-#  define SYS_fsctl	242
-# endif
-
-#ifndef XATTR_LIST_MAX
-#define XATTR_LIST_MAX  65536
-#endif
-
-static __inline__ int xfsctl(const char *path, int fd, int cmd, void *p)
-{
-	return syscall(SYS_fsctl, path, cmd, p, 0);
-}
-
-static __inline__ int platform_test_xfs_fd(int fd)
-{
-	struct statfs buf;
-	if (fstatfs(fd, &buf) < 0)
-		return 0;
-	return strncmp(buf.f_fstypename, "xfs", 4) == 0;
-}
-
-static __inline__ int platform_test_xfs_path(const char *path)
-{
-	struct statfs buf;
-	if (statfs(path, &buf) < 0)
-		return 0;
-	return strncmp(buf.f_fstypename, "xfs", 4) == 0;
-}
-
-static __inline__ int platform_fstatfs(int fd, struct statfs *buf)
-{
-	return fstatfs(fd, buf);
-}
-
-static __inline__ void platform_getoptreset(void)
-{
-	extern int optreset;
-	optreset = 0;
-}
-
-static __inline__ int platform_uuid_compare(uuid_t *uu1, uuid_t *uu2)
-{
-	return uuid_compare(*uu1, *uu2);
-}
-
-static __inline__ void platform_uuid_unparse(uuid_t *uu, char *buffer)
-{
-	uuid_unparse(*uu, buffer);
-}
-
-static __inline__ int platform_uuid_parse(char *buffer, uuid_t *uu)
-{
-	return uuid_parse(buffer, *uu);
-}
-
-static __inline__ int platform_uuid_is_null(uuid_t *uu)
-{
-	return uuid_is_null(*uu);
-}
-
-static __inline__ void platform_uuid_generate(uuid_t *uu)
-{
-	uuid_generate(*uu);
-}
-
-static __inline__ void platform_uuid_clear(uuid_t *uu)
-{
-	uuid_clear(*uu);
-}
-
-static __inline__ void platform_uuid_copy(uuid_t *dst, uuid_t *src)
-{
-	uuid_copy(*dst, *src);
-}
-
-typedef unsigned char		__u8;
-typedef signed char		__s8;
-typedef unsigned short		__u16;
-typedef signed short		__s16;
-typedef unsigned int		__u32;
-typedef signed int		__s32;
-typedef unsigned long long int	__u64;
-typedef signed long long int	__s64;
-
-#define int8_t		int8_t
-#define int16_t		int16_t
-#define int32_t		int32_t
-#define int32_t		int32_t
-#define int64_t		int64_t
-
-typedef off_t		xfs_off_t;
-typedef uint64_t	xfs_ino_t;
-typedef uint32_t	xfs_dev_t;
-typedef int64_t		xfs_daddr_t;
-typedef __u32		xfs_nlink_t;
-
-#define fdatasync	fsync
-#define memalign(a,sz)	valloc(sz)
-
-#define O_LARGEFILE     0
-#ifndef O_DIRECT
-#define O_DIRECT        0
-#endif
-#ifndef O_SYNC
-#define O_SYNC          0
-#endif
-
-#define EFSCORRUPTED	990	/* Filesystem is corrupted */
-#define EFSBADCRC	991	/* Bad CRC detected */
-
-#define HAVE_FID	1
-
-static __inline__ int
-platform_discard_blocks(int fd, uint64_t start, uint64_t len)
-{
-	return 0;
-}
-
-/*
- * POSIX timer replacement.
- * It really just do the minimum we need for xfs_repair.
- * Also, as setitimer can't create multiple timers,
- * the timerid things are useless - we have only one ITIMER_REAL
- * timer.
- */
-#define CLOCK_REALTIME ITIMER_REAL
-#define itimerspec itimerval
-typedef uint64_t timer_t;
-typedef double   timer_c;
-typedef clock_id_t clockid_t;
-
-
-static inline int timer_create (clockid_t __clock_id,
-                         struct sigevent *__restrict __evp,
-                         timer_t *__restrict timer)
-{
-	// set something, to initialize the variable, just in case
-	*timer = 0;
-	return 0;
-}
-
-static inline int timer_settime (timer_t timerid, int flags,
-                          const struct itimerspec *__restrict timerspec,
-                          struct itimerspec *__restrict ovalue)
-{
-	return setitimer(ITIMER_REAL, timerspec, ovalue);
-}
-
-static inline int timer_delete (timer_t timerid)
-{
-	struct itimerspec timespec;
-
-	timespec.it_interval.tv_sec=0;
-	timespec.it_interval.tv_usec=0;
-	timespec.it_value.tv_sec=0;
-	timespec.it_value.tv_usec=0;
-
-	return setitimer(ITIMER_REAL, &timespec, NULL);
-}
-
-static inline int timer_gettime (timer_t timerid, struct itimerspec *value)
-{
-	return getitimer(ITIMER_REAL, value);
-}
-
-/* FSR */
-
-#  include <sys/mount.h>
-#  include <sys/param.h>
-#include <sys/ucred.h>
-#include <errno.h>
-#define		_PATH_MOUNTED   "/etc/mtab"
-
-struct mntent
-{
-	char *mnt_fsname;
-	char *mnt_dir;
-	char *mnt_type;
-	char *mnt_opts;
-	int mnt_freq;
-	int mnt_passno;
-};
-
-static inline void mntinfo2mntent (struct statfs * stats, struct mntent * mnt) {
-	mnt->mnt_fsname = stats->f_mntfromname;
-	mnt->mnt_dir = stats->f_mntonname;
-	mnt->mnt_type = stats->f_fstypename;
-}
-
-
-
-/**
- * Abstraction of mountpoints.
- */
-struct mntent_cursor {
-	FILE *mtabp;
-	struct statfs *stats;
-	int count;
-	int i;
-};
-
-/**
- * OS X uses getmntinfo, which doesn't use a mtab file. So we just ignore it.
- */
-static inline int platform_mntent_open(struct mntent_cursor * cursor, char *mtab)
-{
-	if ((cursor->count = getmntinfo(&cursor->stats, 0)) < 0) {
-		fprintf(stderr, "Error: getmntinfo() failed: %s\n", strerror(errno));
-		return 1;
-	}
-	cursor->i = 0;
-	return 0;
-}
-
-static inline struct mntent * platform_mntent_next(struct mntent_cursor * cursor)
-{
-	struct mntent * t = NULL;
-	if (cursor->i >= cursor->count){
-		return NULL;
-	}
-	mntinfo2mntent(&cursor->stats[cursor->i], t);
-	cursor->i++;
-	return t;
-}
-
-static inline void platform_mntent_close(struct mntent_cursor * cursor)
-{
-	cursor->count = 0;
-	cursor->i = 0;
-}
-
-/* check whether we have to define FS_IOC_FS[GS]ETXATTR ourselves */
-#ifndef HAVE_FSXATTR
-struct fsxattr {
-	__u32		fsx_xflags;	/* xflags field value (get/set) */
-	__u32		fsx_extsize;	/* extsize field value (get/set)*/
-	__u32		fsx_nextents;	/* nextents field value (get)	*/
-	__u32		fsx_projid;	/* project identifier (get/set) */
-	__u32		fsx_cowextsize;	/* cow extsize field value (get/set) */
-	unsigned char	fsx_pad[8];
-};
-
-/*
- * Flags for the fsx_xflags field
- */
-#define FS_XFLAG_REALTIME	0x00000001	/* data in realtime volume */
-#define FS_XFLAG_PREALLOC	0x00000002	/* preallocated file extents */
-#define FS_XFLAG_IMMUTABLE	0x00000008	/* file cannot be modified */
-#define FS_XFLAG_APPEND		0x00000010	/* all writes append */
-#define FS_XFLAG_SYNC		0x00000020	/* all writes synchronous */
-#define FS_XFLAG_NOATIME	0x00000040	/* do not update access time */
-#define FS_XFLAG_NODUMP		0x00000080	/* do not include in backups */
-#define FS_XFLAG_RTINHERIT	0x00000100	/* create with rt bit set */
-#define FS_XFLAG_PROJINHERIT	0x00000200	/* create with parents projid */
-#define FS_XFLAG_NOSYMLINKS	0x00000400	/* disallow symlink creation */
-#define FS_XFLAG_EXTSIZE	0x00000800	/* extent size allocator hint */
-#define FS_XFLAG_EXTSZINHERIT	0x00001000	/* inherit inode extent size */
-#define FS_XFLAG_NODEFRAG	0x00002000	/* do not defragment */
-#define FS_XFLAG_FILESTREAM	0x00004000	/* use filestream allocator */
-#define FS_XFLAG_DAX		0x00008000	/* use DAX for IO */
-#define FS_XFLAG_HASATTR	0x80000000	/* no DIFLAG for this	*/
-
-#define FS_IOC_FSGETXATTR     _IOR ('X', 31, struct fsxattr)
-#define FS_IOC_FSSETXATTR     _IOW ('X', 32, struct fsxattr)
-
-#endif
-
-#ifndef FS_XFLAG_COWEXTSIZE
-#define FS_XFLAG_COWEXTSIZE	0x00010000	/* CoW extent size allocator hint */
-#endif
-
-#endif	/* __XFS_DARWIN_H__ */
diff --git a/include/freebsd.h b/include/freebsd.h
deleted file mode 100644
index f52ed0a..0000000
--- a/include/freebsd.h
+++ /dev/null
@@ -1,210 +0,0 @@
-/*
- * Copyright (c) 2004-2006 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public License
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_FREEBSD_H__
-#define __XFS_FREEBSD_H__
-
-#include <sys/stat.h>
-#include <sys/param.h>
-#include <sys/ioccom.h>
-#include <sys/mount.h>
-#include <ctype.h>
-#include <libgen.h>
-#include <paths.h>
-#include <uuid.h>
-#include <mntent.h>
-
-#include <sys/endian.h>
-#define __BYTE_ORDER	BYTE_ORDER
-#define __BIG_ENDIAN	BIG_ENDIAN
-#define __LITTLE_ENDIAN	LITTLE_ENDIAN
-
-/* FreeBSD file API is 64-bit aware */
-#define fdatasync	fsync
-#define memalign(a,sz)	valloc(sz)
-
-#define EFSCORRUPTED	990	/* Filesystem is corrupted */
-#define EFSBADCRC	991	/* Bad CRC detected */
-
-typedef unsigned char		__u8;
-typedef signed char		__s8;
-typedef unsigned short		__u16;
-typedef signed short		__s16;
-typedef unsigned int		__u32;
-typedef signed int		__s32;
-typedef unsigned long long int	__u64;
-typedef signed long long int	__s64;
-
-typedef off_t		xfs_off_t;
-typedef uint64_t	xfs_ino_t;
-typedef uint32_t	xfs_dev_t;
-typedef int64_t		xfs_daddr_t;
-typedef __u32		xfs_nlink_t;
-
-#define	O_LARGEFILE	0
-
-#define HAVE_FID	1
-
-static __inline__ int xfsctl(const char *path, int fd, int cmd, void *p)
-{
-	return ioctl(fd, cmd, p);
-}
-
-static __inline__ int platform_test_xfs_fd(int fd)
-{
-	struct statfs buf;
-	if (fstatfs(fd, &buf) < 0)
-		return 0;
-	return strncmp(buf.f_fstypename, "xfs", 4) == 0;
-}
-
-static __inline__ int platform_test_xfs_path(const char *path)
-{
-	struct statfs buf;
-	if (statfs(path, &buf) < 0)
-		return 0;
-	return strncmp(buf.f_fstypename, "xfs", 4) == 0;
-}
-
-static __inline__ int platform_fstatfs(int fd, struct statfs *buf)
-{
-	return fstatfs(fd, buf);
-}
-
-static __inline__ void platform_getoptreset(void)
-{
-	extern int optind;
-	optind = 0;
-}
-
-static __inline__ int platform_uuid_compare(uuid_t *uu1, uuid_t *uu2)
-{
-	return uuid_compare(uu1, uu2, NULL);
-}
-
-static __inline__ void platform_uuid_unparse(uuid_t *uu, char *buffer)
-{
-	uint32_t status;
-	char *s;
-	uuid_to_string(uu, &s, &status);
-	if (status == uuid_s_ok)
-		strcpy(buffer, s);
-	else buffer[0] = '\0';
-	free(s);
-}
-
-static __inline__ int platform_uuid_parse(char *buffer, uuid_t *uu)
-{
-	uint32_t status;
-	uuid_from_string(buffer, uu, &status);
-	return (status == uuid_s_ok);
-}
-
-static __inline__ int platform_uuid_is_null(uuid_t *uu)
-{
-	return uuid_is_nil(uu, NULL);
-}
-
-static __inline__ void platform_uuid_generate(uuid_t *uu)
-{
-	uuid_create(uu, NULL);
-}
-
-static __inline__ void platform_uuid_clear(uuid_t *uu)
-{
-	uuid_create_nil(uu, NULL);
-}
-
-static __inline__ void platform_uuid_copy(uuid_t *dst, uuid_t *src)
-{
-	memcpy(dst, src, sizeof(uuid_t));
-}
-
-static __inline__ int
-platform_discard_blocks(int fd, uint64_t start, uint64_t len)
-{
-	return 0;
-}
-
-/**
- * Abstraction of mountpoints.
- */
-struct mntent_cursor {
-	FILE *mtabp;
-};
-
-static inline int platform_mntent_open(struct mntent_cursor * cursor, char *mtab)
-{
-	cursor->mtabp = setmntent(mtab, "r");
-	if (!cursor->mtabp) {
-		fprintf(stderr, "Error: cannot read %s\n", mtab);
-		return 1;
-	}
-	return 0;
-}
-
-static inline  struct mntent * platform_mntent_next(struct mntent_cursor * cursor)
-{
-	return getmntent(cursor->mtabp);
-}
-
-static inline void platform_mntent_close(struct mntent_cursor * cursor)
-{
-	endmntent(cursor->mtabp);
-}
-
-/* check whether we have to define FS_IOC_FS[GS]ETXATTR ourselves */
-#ifndef HAVE_FSXATTR
-struct fsxattr {
-	__u32		fsx_xflags;	/* xflags field value (get/set) */
-	__u32		fsx_extsize;	/* extsize field value (get/set)*/
-	__u32		fsx_nextents;	/* nextents field value (get)	*/
-	__u32		fsx_projid;	/* project identifier (get/set) */
-	__u32		fsx_cowextsize;	/* cow extsize field value (get/set) */
-	unsigned char	fsx_pad[8];
-};
-
-/*
- * Flags for the fsx_xflags field
- */
-#define FS_XFLAG_REALTIME	0x00000001	/* data in realtime volume */
-#define FS_XFLAG_PREALLOC	0x00000002	/* preallocated file extents */
-#define FS_XFLAG_IMMUTABLE	0x00000008	/* file cannot be modified */
-#define FS_XFLAG_APPEND		0x00000010	/* all writes append */
-#define FS_XFLAG_SYNC		0x00000020	/* all writes synchronous */
-#define FS_XFLAG_NOATIME	0x00000040	/* do not update access time */
-#define FS_XFLAG_NODUMP		0x00000080	/* do not include in backups */
-#define FS_XFLAG_RTINHERIT	0x00000100	/* create with rt bit set */
-#define FS_XFLAG_PROJINHERIT	0x00000200	/* create with parents projid */
-#define FS_XFLAG_NOSYMLINKS	0x00000400	/* disallow symlink creation */
-#define FS_XFLAG_EXTSIZE	0x00000800	/* extent size allocator hint */
-#define FS_XFLAG_EXTSZINHERIT	0x00001000	/* inherit inode extent size */
-#define FS_XFLAG_NODEFRAG	0x00002000	/* do not defragment */
-#define FS_XFLAG_FILESTREAM	0x00004000	/* use filestream allocator */
-#define FS_XFLAG_DAX		0x00008000	/* use DAX for IO */
-#define FS_XFLAG_HASATTR	0x80000000	/* no DIFLAG for this	*/
-
-#define FS_IOC_FSGETXATTR     _IOR ('X', 31, struct fsxattr)
-#define FS_IOC_FSSETXATTR     _IOW ('X', 32, struct fsxattr)
-
-#endif
-
-#ifndef FS_XFLAG_COWEXTSIZE
-#define FS_XFLAG_COWEXTSIZE	0x00010000	/* CoW extent size allocator hint */
-#endif
-
-#endif	/* __XFS_FREEBSD_H__ */
diff --git a/libxfs/Makefile b/libxfs/Makefile
index 0470f5f..00df418 100644
--- a/libxfs/Makefile
+++ b/libxfs/Makefile
@@ -98,7 +98,7 @@ CFILES = cache.c \
 	xfs_trans_resv.c
 
 CFILES += $(PKG_PLATFORM).c
-PCFILES = darwin.c freebsd.c irix.c linux.c
+PCFILES = linux.c
 LSRCFILES = $(shell echo $(PCFILES) | sed -e "s/$(PKG_PLATFORM).c//g")
 LSRCFILES += gen_crc32table.c
 
diff --git a/libxfs/darwin.c b/libxfs/darwin.c
deleted file mode 100644
index 16d2c35..0000000
--- a/libxfs/darwin.c
+++ /dev/null
@@ -1,144 +0,0 @@
-/*
- * Copyright (c) 2003,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-
-#include <sys/disk.h>
-#include <sys/stat.h>
-#include <sys/mount.h>
-#include <sys/ioctl.h>
-#include <sys/sysctl.h>
-#include "libxfs.h"
-
-int platform_has_uuid = 1;
-extern char *progname;
-
-int
-platform_check_ismounted(char *name, char *block, struct stat *s, int verbose)
-{
-	return 0;
-}
-
-int
-platform_check_iswritable(char *name, char *block, struct stat *s)
-{
-	int	fd, writable;
-
-	if ((fd = open(block, O_RDONLY, 0)) < 0) {
-		fprintf(stderr, _("%s: "
-			"error opening the device special file \"%s\": %s\n"),
-			progname, block, strerror(errno));
-		exit(1);
-	}
-
-	if (ioctl(fd, DKIOCISWRITABLE, &writable) < 0) {
-		fprintf(stderr, _("%s: can't tell if \"%s\" is writable: %s\n"),
-			progname, block, strerror(errno));
-		exit(1);
-	}
-	close(fd);
-	return writable == 0;
-}
-
-int
-platform_set_blocksize(int fd, char *path, dev_t device, int blocksize, int fatal)
-{
-	return fatal;
-}
-
-void
-platform_flush_device(int fd, dev_t device)
-{
-	ioctl(fd, DKIOCSYNCHRONIZECACHE, NULL);
-}
-
-void
-platform_findsizes(char *path, int fd, long long *sz, int *bsz)
-{
-	uint64_t	size;
-	struct stat	st;
-
-	if (fstat(fd, &st) < 0) {
-		fprintf(stderr,
-			_("%s: cannot stat the device file \"%s\": %s\n"),
-			progname, path, strerror(errno));
-		exit(1);
-	}
-	if ((st.st_mode & S_IFMT) == S_IFREG) {
-		*sz = (long long)(st.st_size >> 9);
-		*bsz = BBSIZE;
-		return;
-	}
-	if (ioctl(fd, DKIOCGETBLOCKCOUNT, &size) < 0) {
-		fprintf(stderr, _("%s: can't determine device size: %s\n"),
-			progname, strerror(errno));
-		exit(1);
-	}
-	*sz = (long long)size;
-	*bsz = BBSIZE;
-}
-
-char *
-platform_findrawpath(char *path)
-{
-	return path;
-}
-
-char *
-platform_findblockpath(char *path)
-{
-	return path;
-}
-
-int
-platform_direct_blockdev(void)
-{
-	return 0;
-}
-
-int
-platform_align_blockdev(void)
-{
-	return sizeof(void *);
-}
-
-int
-platform_nproc(void)
-{
-	int		ncpu;
-	size_t		len = sizeof(ncpu);
-	static int	mib[2] = {CTL_HW, HW_NCPU};
-
-	if (sysctl(mib, 2, &ncpu, &len, NULL, 0) < 0)
-		ncpu = 1;
-
-	return ncpu;
-}
-
-unsigned long
-platform_physmem(void)
-{
-	unsigned long	physmem;
-	size_t		len = sizeof(physmem);
-	static int	mib[2] = {CTL_HW, HW_PHYSMEM};
-
-	if (sysctl(mib, 2, &physmem, &len, NULL, 0) < 0) {
-		fprintf(stderr, _("%s: can't determine memory size\n"),
-			progname);
-		exit(1);
-	}
-	return physmem >> 10;
-}
diff --git a/libxfs/freebsd.c b/libxfs/freebsd.c
deleted file mode 100644
index 5b9ef29..0000000
--- a/libxfs/freebsd.c
+++ /dev/null
@@ -1,201 +0,0 @@
-/*
- * Copyright (c) 2003,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-
-#include "libxfs.h"
-#include <sys/stat.h>
-#include <sys/disk.h>
-#include <sys/mount.h>
-#include <sys/ioctl.h>
-#include <sys/sysctl.h>
-
-int platform_has_uuid = 1;
-extern char *progname;
-
-int
-platform_check_ismounted(char *name, char *block, struct stat *s, int verbose)
-{
-	struct stat	st;
-        int cnt, i;
-        struct statfs *fsinfo;
-
-	if (!s) {
-		if (stat(block, &st) < 0)
-			return 0;
-		s = &st;
-	}
-
-	/* Remember, FreeBSD can now mount char devices! -- adrian */
-	if (((st.st_mode & S_IFMT) != S_IFBLK) &&
-	    ((st.st_mode & S_IFMT) != S_IFCHR))
-		return 0;
-
-	if ((cnt = getmntinfo(&fsinfo, MNT_NOWAIT)) == 0) {
-		fprintf(stderr,
-		    _("%s: %s possibly contains a mounted filesystem\n"),
-		    progname, name);
-		return 1;
-	}
-
-        for (i = 0; i < cnt; i++) {
-                if (strcmp (name, fsinfo[i].f_mntfromname) != 0)
-			continue;
-
-		if (verbose)
-			fprintf(stderr,
-			    _("%s: %s contains a mounted filesystem\n"),
-			    progname, name);
-		break;
-	}
-
-        return i < cnt;
-}
-
-int
-platform_check_iswritable(char *name, char *block, struct stat *s)
-{
-        int cnt, i;
-        struct statfs *fsinfo;
-
-        if ((cnt = getmntinfo(&fsinfo, MNT_NOWAIT)) == 0) {
-		fprintf(stderr, _("%s: %s contains a possibly writable, "
-				"mounted filesystem\n"), progname, name);
-			return 1;
-	}
-
-        for (i = 0; i < cnt; i++) {
-                if (strcmp (name, fsinfo[i].f_mntfromname) != 0)
-			continue;
-
-		if (fsinfo[i].f_flags &= MNT_RDONLY)
-			break;
-	}
-
-        if (i == cnt) {
-		fprintf(stderr, _("%s: %s contains a mounted and writable "
-				"filesystem\n"), progname, name);
-		return 1;
-	}
-	return 0;
-}
-
-int
-platform_set_blocksize(int fd, char *path, dev_t device, int blocksize, int fatal)
-{
-	return fatal;
-}
-
-void
-platform_flush_device(int fd, dev_t device)
-{
-	return;
-}
-
-void
-platform_findsizes(char *path, int fd, long long *sz, int *bsz)
-{
-	struct stat	st;
-	int64_t		size;
-	uint		ssize;
-
-	if (fstat(fd, &st) < 0) {
-		fprintf(stderr, _("%s: "
-			"cannot stat the device file \"%s\": %s\n"),
-			progname, path, strerror(errno));
-		exit(1);
-	}
-
-	if ((st.st_mode & S_IFMT) == S_IFREG) {
-		*sz = (long long)(st.st_size >> 9);
-		*bsz = 512;
-		return;
-	}
-
-	if ((st.st_mode & S_IFMT) != S_IFCHR) {
-		fprintf(stderr, _("%s: Not a device or file: \"%s\"\n"),
-			progname, path);
-		exit(1);
-	}
-
-	if (ioctl(fd, DIOCGMEDIASIZE, &size) != 0) {
-		fprintf(stderr, _("%s: DIOCGMEDIASIZE failed on \"%s\": %s\n"),
-			progname, path, strerror(errno));
-		exit(1);
-	}
-
-	if (ioctl(fd, DIOCGSECTORSIZE, &ssize) != 0) {
-		fprintf(stderr, _("%s: "
-			"DIOCGSECTORSIZE failed on \"%s\": %s\n"),
-			progname, path, strerror(errno));
-		exit(1);
-	}
-
-	*sz = (long long) (size / ssize);
-	*bsz = (int)ssize;
-}
-
-char *
-platform_findrawpath(char *path)
-{
-	return path;
-}
-
-char *
-platform_findblockpath(char *path)
-{
-	return path;
-}
-
-int
-platform_direct_blockdev(void)
-{
-	return 0;
-}
-
-int
-platform_align_blockdev(void)
-{
-	return sizeof(void *);
-}
-
-int
-platform_nproc(void)
-{
-	int		ncpu;
-	size_t		len = sizeof(ncpu);
-	static int	mib[2] = {CTL_HW, HW_NCPU};
-
-	if (sysctl(mib, 2, &ncpu, &len, NULL, 0) < 0)
-		ncpu = 1;
-
-	return ncpu;
-}
-
-unsigned long
-platform_physmem(void)
-{
-	unsigned long	physmem;
-	size_t		len = sizeof(physmem);
-	static int	mib[2] = {CTL_HW, HW_PHYSMEM};
-
-	if (sysctl(mib, 2, &physmem, &len, NULL, 0) < 0) {
-		fprintf(stderr, _("%s: can't determine memory size\n"),
-			progname);
-		exit(1);
-	}
-	return physmem >> 10;
-}
diff --git a/libxfs/irix.c b/libxfs/irix.c
deleted file mode 100644
index 0f14aec..0000000
--- a/libxfs/irix.c
+++ /dev/null
@@ -1,111 +0,0 @@
-/*
- * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-
-#include "libxfs.h"
-#include <diskinfo.h>
-#include <sys/sysmp.h>
-
-int platform_has_uuid = 0;
-extern char *progname;
-extern int64_t findsize(char *);
-
-int
-platform_check_ismounted(char *name, char *block, struct stat *s, int verbose)
-{
-	return 0;
-}
-
-int
-platform_check_iswritable(char *name, char *block, struct stat *s)
-{
-	return 1;
-}
-
-int
-platform_set_blocksize(int fd, char *path, dev_t device, int blocksize, int fatal)
-{
-	return fatal;
-}
-
-void
-platform_flush_device(int fd, dev_t device)
-{
-	return;
-}
-
-void
-platform_findsizes(char *path, int fd, long long *sz, int *bsz)
-{
-	struct stat		st;
-
-	if (fstat(fd, &st) < 0) {
-		fprintf(stderr,
-			_("%s: cannot stat the device file \"%s\": %s\n"),
-			progname, path, strerror(errno));
-		exit(1);
-	}
-	if ((st.st_mode & S_IFMT) == S_IFREG) {
-		*sz = (long long)(st.st_size >> 9);
-	} else {
-		*sz = findsize(path);
-	}
-	*bsz = BBSIZE;
-}
-
-char *
-platform_findrawpath(char *path)
-{
-	return findrawpath(path);
-}
-
-char *
-platform_findblockpath(char *path)
-{
-	return findblockpath(path);
-}
-
-int
-platform_direct_blockdev(void)
-{
-	return 0;
-}
-
-int
-platform_align_blockdev(void)
-{
-	return sizeof(void *);
-}
-
-int
-platform_nproc(void)
-{
-	return sysmp(MP_NPROCS);
-}
-
-unsigned long
-platform_physmem(void)
-{
-	struct rminfo ri;
-
-	if (sysmp(MP_SAGET, MPSA_RMINFO, &ri, sizeof(ri)) < 0)
-		fprintf(stderr, _("%s: can't determine memory size\n"),
-			progname);
-		exit(1);
-	}
-	return (ri.physmem >> 10) * getpagesize();	/* kilobytes */
-}

[PATCH 6/4] libfrog: absorb platform specific code

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Move the linux support code to libfrog, which should remove the final
dependency of libfrog on libxfs.  libfrog is the runtime support library
anyway.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 libfrog/Makefile |    4 +
 libfrog/linux.c  |  265 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
 libxfs/Makefile  |    3 -
 libxfs/linux.c   |  265 ------------------------------------------------------
 4 files changed, 269 insertions(+), 268 deletions(-)
 create mode 100644 libfrog/linux.c
 delete mode 100644 libxfs/linux.c

diff --git a/libfrog/Makefile b/libfrog/Makefile
index 230b08f..e3065e6 100644
--- a/libfrog/Makefile
+++ b/libfrog/Makefile
@@ -22,6 +22,10 @@ topology.c \
 util.c \
 workqueue.c
 
+CFILES += $(PKG_PLATFORM).c
+PCFILES = linux.c
+LSRCFILES = $(shell echo $(PCFILES) | sed -e "s/$(PKG_PLATFORM).c//g")
+
 ifeq ($(HAVE_GETMNTENT),yes)
 LCFLAGS += -DHAVE_GETMNTENT
 endif
diff --git a/libfrog/linux.c b/libfrog/linux.c
new file mode 100644
index 0000000..0bace3e
--- /dev/null
+++ b/libfrog/linux.c
@@ -0,0 +1,265 @@
+/*
+ * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include <mntent.h>
+#include <sys/stat.h>
+#include <sys/mount.h>
+#include <sys/ioctl.h>
+#include <sys/sysinfo.h>
+
+#include "libxfs_priv.h"
+#include "xfs_fs.h"
+
+int platform_has_uuid = 1;
+extern char *progname;
+static int max_block_alignment;
+
+#ifndef BLKGETSIZE64
+# define BLKGETSIZE64	_IOR(0x12,114,size_t)
+#endif
+#ifndef BLKBSZSET
+# define BLKBSZSET	_IOW(0x12,113,size_t)
+#endif
+#ifndef BLKSSZGET
+# define BLKSSZGET	_IO(0x12,104)
+#endif
+
+#ifndef RAMDISK_MAJOR
+#define RAMDISK_MAJOR	1	/* ramdisk major number */
+#endif
+
+#define PROC_MOUNTED	"/proc/mounts"
+
+/*
+ * Check if the filesystem is mounted.  Be verbose if asked, and
+ * optionally restrict check to /writable/ mounts (i.e. RO is OK)
+ */
+#define	CHECK_MOUNT_VERBOSE	0x1
+#define	CHECK_MOUNT_WRITABLE	0x2
+
+static int
+platform_check_mount(char *name, char *block, struct stat *s, int flags)
+{
+	FILE		*f;
+	struct stat	st, mst;
+	struct mntent	*mnt;
+	char		mounts[MAXPATHLEN];
+
+	if (!s) {
+		/* If either fails we are not mounted */
+		if (stat(block, &st) < 0)
+			return 0;
+		if ((st.st_mode & S_IFMT) != S_IFBLK)
+			return 0;
+		s = &st;
+	}
+
+	strcpy(mounts, (!access(PROC_MOUNTED, R_OK)) ? PROC_MOUNTED : MOUNTED);
+	if ((f = setmntent(mounts, "r")) == NULL) {
+		/* Unexpected failure, warn unconditionally */
+		fprintf(stderr,
+		    _("%s: %s possibly contains a mounted filesystem\n"),
+		    progname, name);
+		return 1;
+	}
+	while ((mnt = getmntent(f)) != NULL) {
+		if (stat(mnt->mnt_dir, &mst) < 0)
+			continue;
+		if (mst.st_dev != s->st_rdev)
+			continue;
+		/* Found our device, is RO OK? */
+		if ((flags & CHECK_MOUNT_WRITABLE) && hasmntopt(mnt, MNTOPT_RO))
+			continue;
+		else
+			break;
+	}
+	endmntent(f);
+
+	/* No mounts contained the condition we were looking for */
+	if (mnt == NULL)
+		return 0;
+
+	if (flags & CHECK_MOUNT_VERBOSE) {
+		if (flags & CHECK_MOUNT_WRITABLE) {
+			fprintf(stderr,
+_("%s: %s contains a mounted and writable filesystem\n"),
+				progname, name);
+		} else {
+			fprintf(stderr,
+_("%s: %s contains a mounted filesystem\n"),
+				progname, name);
+		}
+	}
+	return 1;
+}
+
+int
+platform_check_ismounted(char *name, char *block, struct stat *s, int verbose)
+{
+	int flags;
+
+	flags = verbose ? CHECK_MOUNT_VERBOSE : 0;
+	return platform_check_mount(name, block, s, flags);
+}
+
+int
+platform_check_iswritable(char *name, char *block, struct stat *s)
+{
+	int flags;
+
+	/* Writable checks are always verbose */
+	flags = CHECK_MOUNT_WRITABLE | CHECK_MOUNT_VERBOSE;
+	return platform_check_mount(name, block, s, flags);
+}
+
+int
+platform_set_blocksize(int fd, char *path, dev_t device, int blocksize, int fatal)
+{
+	int error = 0;
+
+	if (major(device) != RAMDISK_MAJOR) {
+		if ((error = ioctl(fd, BLKBSZSET, &blocksize)) < 0) {
+			fprintf(stderr, _("%s: %s - cannot set blocksize "
+					"%d on block device %s: %s\n"),
+				progname, fatal ? "error": "warning",
+				blocksize, path, strerror(errno));
+		}
+	}
+	return error;
+}
+
+void
+platform_flush_device(int fd, dev_t device)
+{
+	struct stat	st;
+	if (major(device) == RAMDISK_MAJOR)
+		return;
+
+	if (fstat(fd, &st) < 0)
+		return;
+
+	if (S_ISREG(st.st_mode))
+		fsync(fd);
+	else
+		ioctl(fd, BLKFLSBUF, 0);
+}
+
+void
+platform_findsizes(char *path, int fd, long long *sz, int *bsz)
+{
+	struct stat	st;
+	uint64_t	size;
+	int		error;
+
+	if (fstat(fd, &st) < 0) {
+		fprintf(stderr, _("%s: "
+			"cannot stat the device file \"%s\": %s\n"),
+			progname, path, strerror(errno));
+		exit(1);
+	}
+
+	if ((st.st_mode & S_IFMT) == S_IFREG) {
+		struct dioattr	da;
+
+		*sz = (long long)(st.st_size >> 9);
+
+		if (ioctl(fd, XFS_IOC_DIOINFO, &da) < 0) {
+			/*
+			 * fall back to BBSIZE; mkfs might fail if there's a
+			 * size mismatch between the image & the host fs...
+			 */
+			*bsz = BBSIZE;
+		} else
+			*bsz = da.d_miniosz;
+
+		if (*bsz > max_block_alignment)
+			max_block_alignment = *bsz;
+		return;
+	}
+
+	error = ioctl(fd, BLKGETSIZE64, &size);
+	if (error >= 0) {
+		/* BLKGETSIZE64 returns size in bytes not 512-byte blocks */
+		*sz = (long long)(size >> 9);
+	} else {
+		/* If BLKGETSIZE64 fails, try BLKGETSIZE */
+		unsigned long tmpsize;
+
+		error = ioctl(fd, BLKGETSIZE, &tmpsize);
+		if (error < 0) {
+			fprintf(stderr, _("%s: can't determine device size\n"),
+				progname);
+			exit(1);
+		}
+		*sz = (long long)tmpsize;
+	}
+
+	if (ioctl(fd, BLKSSZGET, bsz) < 0) {
+		fprintf(stderr, _("%s: warning - cannot get sector size "
+				"from block device %s: %s\n"),
+			progname, path, strerror(errno));
+		*bsz = BBSIZE;
+	}
+	if (*bsz > max_block_alignment)
+		max_block_alignment = *bsz;
+}
+
+char *
+platform_findrawpath(char *path)
+{
+	return path;
+}
+
+char *
+platform_findblockpath(char *path)
+{
+	return path;
+}
+
+int
+platform_direct_blockdev(void)
+{
+	return 1;
+}
+
+int
+platform_align_blockdev(void)
+{
+	if (!max_block_alignment)
+		return getpagesize();
+	return max_block_alignment;
+}
+
+int
+platform_nproc(void)
+{
+	return sysconf(_SC_NPROCESSORS_ONLN);
+}
+
+unsigned long
+platform_physmem(void)
+{
+	struct sysinfo  si;
+
+	if (sysinfo(&si) < 0) {
+		fprintf(stderr, _("%s: can't determine memory size\n"),
+			progname);
+		exit(1);
+	}
+	return (si.totalram >> 10) * si.mem_unit;	/* kilobytes */
+}
diff --git a/libxfs/Makefile b/libxfs/Makefile
index 00df418..7cde18d 100644
--- a/libxfs/Makefile
+++ b/libxfs/Makefile
@@ -97,9 +97,6 @@ CFILES = cache.c \
 	xfs_symlink_remote.c \
 	xfs_trans_resv.c
 
-CFILES += $(PKG_PLATFORM).c
-PCFILES = linux.c
-LSRCFILES = $(shell echo $(PCFILES) | sed -e "s/$(PKG_PLATFORM).c//g")
 LSRCFILES += gen_crc32table.c
 
 #
diff --git a/libxfs/linux.c b/libxfs/linux.c
deleted file mode 100644
index 0bace3e..0000000
--- a/libxfs/linux.c
+++ /dev/null
@@ -1,265 +0,0 @@
-/*
- * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-
-#include <mntent.h>
-#include <sys/stat.h>
-#include <sys/mount.h>
-#include <sys/ioctl.h>
-#include <sys/sysinfo.h>
-
-#include "libxfs_priv.h"
-#include "xfs_fs.h"
-
-int platform_has_uuid = 1;
-extern char *progname;
-static int max_block_alignment;
-
-#ifndef BLKGETSIZE64
-# define BLKGETSIZE64	_IOR(0x12,114,size_t)
-#endif
-#ifndef BLKBSZSET
-# define BLKBSZSET	_IOW(0x12,113,size_t)
-#endif
-#ifndef BLKSSZGET
-# define BLKSSZGET	_IO(0x12,104)
-#endif
-
-#ifndef RAMDISK_MAJOR
-#define RAMDISK_MAJOR	1	/* ramdisk major number */
-#endif
-
-#define PROC_MOUNTED	"/proc/mounts"
-
-/*
- * Check if the filesystem is mounted.  Be verbose if asked, and
- * optionally restrict check to /writable/ mounts (i.e. RO is OK)
- */
-#define	CHECK_MOUNT_VERBOSE	0x1
-#define	CHECK_MOUNT_WRITABLE	0x2
-
-static int
-platform_check_mount(char *name, char *block, struct stat *s, int flags)
-{
-	FILE		*f;
-	struct stat	st, mst;
-	struct mntent	*mnt;
-	char		mounts[MAXPATHLEN];
-
-	if (!s) {
-		/* If either fails we are not mounted */
-		if (stat(block, &st) < 0)
-			return 0;
-		if ((st.st_mode & S_IFMT) != S_IFBLK)
-			return 0;
-		s = &st;
-	}
-
-	strcpy(mounts, (!access(PROC_MOUNTED, R_OK)) ? PROC_MOUNTED : MOUNTED);
-	if ((f = setmntent(mounts, "r")) == NULL) {
-		/* Unexpected failure, warn unconditionally */
-		fprintf(stderr,
-		    _("%s: %s possibly contains a mounted filesystem\n"),
-		    progname, name);
-		return 1;
-	}
-	while ((mnt = getmntent(f)) != NULL) {
-		if (stat(mnt->mnt_dir, &mst) < 0)
-			continue;
-		if (mst.st_dev != s->st_rdev)
-			continue;
-		/* Found our device, is RO OK? */
-		if ((flags & CHECK_MOUNT_WRITABLE) && hasmntopt(mnt, MNTOPT_RO))
-			continue;
-		else
-			break;
-	}
-	endmntent(f);
-
-	/* No mounts contained the condition we were looking for */
-	if (mnt == NULL)
-		return 0;
-
-	if (flags & CHECK_MOUNT_VERBOSE) {
-		if (flags & CHECK_MOUNT_WRITABLE) {
-			fprintf(stderr,
-_("%s: %s contains a mounted and writable filesystem\n"),
-				progname, name);
-		} else {
-			fprintf(stderr,
-_("%s: %s contains a mounted filesystem\n"),
-				progname, name);
-		}
-	}
-	return 1;
-}
-
-int
-platform_check_ismounted(char *name, char *block, struct stat *s, int verbose)
-{
-	int flags;
-
-	flags = verbose ? CHECK_MOUNT_VERBOSE : 0;
-	return platform_check_mount(name, block, s, flags);
-}
-
-int
-platform_check_iswritable(char *name, char *block, struct stat *s)
-{
-	int flags;
-
-	/* Writable checks are always verbose */
-	flags = CHECK_MOUNT_WRITABLE | CHECK_MOUNT_VERBOSE;
-	return platform_check_mount(name, block, s, flags);
-}
-
-int
-platform_set_blocksize(int fd, char *path, dev_t device, int blocksize, int fatal)
-{
-	int error = 0;
-
-	if (major(device) != RAMDISK_MAJOR) {
-		if ((error = ioctl(fd, BLKBSZSET, &blocksize)) < 0) {
-			fprintf(stderr, _("%s: %s - cannot set blocksize "
-					"%d on block device %s: %s\n"),
-				progname, fatal ? "error": "warning",
-				blocksize, path, strerror(errno));
-		}
-	}
-	return error;
-}
-
-void
-platform_flush_device(int fd, dev_t device)
-{
-	struct stat	st;
-	if (major(device) == RAMDISK_MAJOR)
-		return;
-
-	if (fstat(fd, &st) < 0)
-		return;
-
-	if (S_ISREG(st.st_mode))
-		fsync(fd);
-	else
-		ioctl(fd, BLKFLSBUF, 0);
-}
-
-void
-platform_findsizes(char *path, int fd, long long *sz, int *bsz)
-{
-	struct stat	st;
-	uint64_t	size;
-	int		error;
-
-	if (fstat(fd, &st) < 0) {
-		fprintf(stderr, _("%s: "
-			"cannot stat the device file \"%s\": %s\n"),
-			progname, path, strerror(errno));
-		exit(1);
-	}
-
-	if ((st.st_mode & S_IFMT) == S_IFREG) {
-		struct dioattr	da;
-
-		*sz = (long long)(st.st_size >> 9);
-
-		if (ioctl(fd, XFS_IOC_DIOINFO, &da) < 0) {
-			/*
-			 * fall back to BBSIZE; mkfs might fail if there's a
-			 * size mismatch between the image & the host fs...
-			 */
-			*bsz = BBSIZE;
-		} else
-			*bsz = da.d_miniosz;
-
-		if (*bsz > max_block_alignment)
-			max_block_alignment = *bsz;
-		return;
-	}
-
-	error = ioctl(fd, BLKGETSIZE64, &size);
-	if (error >= 0) {
-		/* BLKGETSIZE64 returns size in bytes not 512-byte blocks */
-		*sz = (long long)(size >> 9);
-	} else {
-		/* If BLKGETSIZE64 fails, try BLKGETSIZE */
-		unsigned long tmpsize;
-
-		error = ioctl(fd, BLKGETSIZE, &tmpsize);
-		if (error < 0) {
-			fprintf(stderr, _("%s: can't determine device size\n"),
-				progname);
-			exit(1);
-		}
-		*sz = (long long)tmpsize;
-	}
-
-	if (ioctl(fd, BLKSSZGET, bsz) < 0) {
-		fprintf(stderr, _("%s: warning - cannot get sector size "
-				"from block device %s: %s\n"),
-			progname, path, strerror(errno));
-		*bsz = BBSIZE;
-	}
-	if (*bsz > max_block_alignment)
-		max_block_alignment = *bsz;
-}
-
-char *
-platform_findrawpath(char *path)
-{
-	return path;
-}
-
-char *
-platform_findblockpath(char *path)
-{
-	return path;
-}
-
-int
-platform_direct_blockdev(void)
-{
-	return 1;
-}
-
-int
-platform_align_blockdev(void)
-{
-	if (!max_block_alignment)
-		return getpagesize();
-	return max_block_alignment;
-}
-
-int
-platform_nproc(void)
-{
-	return sysconf(_SC_NPROCESSORS_ONLN);
-}
-
-unsigned long
-platform_physmem(void)
-{
-	struct sysinfo  si;
-
-	if (sysinfo(&si) < 0) {
-		fprintf(stderr, _("%s: can't determine memory size\n"),
-			progname);
-		exit(1);
-	}
-	return (si.totalram >> 10) * si.mem_unit;	/* kilobytes */
-}

Re: [PATCH v2 1/4] xfs_repair: implement custom ifork verifiers

[Eric Sandeen]

On 3/20/18 4:47 PM, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
> 
> There are a few cases where an early stage of xfs_repair will write an
> invalid inode fork buffer to signal to a later stage that it needs to
> correct the value.  This happens in phase 4 when we detect an inline
> format directory with an invalid .. pointer.  To avoid triggering the
> ifork verifiers on this, inject a custom verifier for phase 6 that lets
> this pass for now.
> 
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>

Looks good, thanks.

Reviewed-by: Eric Sandeen <sandeen@redhat.com>

> ---
> v2: fix misleading comments, remove pointless wrappers
> ---
>  libxfs/libxfs_api_defs.h |    2 ++
>  repair/phase6.c          |   57 +++++++++++++++++++++++++++++++++++++++++++++-
>  2 files changed, 58 insertions(+), 1 deletion(-)
> 
> diff --git a/libxfs/libxfs_api_defs.h b/libxfs/libxfs_api_defs.h
> index 5d56340..78daca0 100644
> --- a/libxfs/libxfs_api_defs.h
> +++ b/libxfs/libxfs_api_defs.h
> @@ -150,5 +150,7 @@
>  #define xfs_rmap_lookup_le_range	libxfs_rmap_lookup_le_range
>  #define xfs_refc_block			libxfs_refc_block
>  #define xfs_rmap_compare		libxfs_rmap_compare
> +#define xfs_dir_get_ops			libxfs_dir_get_ops
> +#define xfs_default_ifork_ops		libxfs_default_ifork_ops
>  
>  #endif /* __LIBXFS_API_DEFS_H__ */
> diff --git a/repair/phase6.c b/repair/phase6.c
> index aff83bc..ed005e8 100644
> --- a/repair/phase6.c
> +++ b/repair/phase6.c
> @@ -39,6 +39,61 @@ static struct xfs_name		xfs_name_dot = {(unsigned char *)".",
>  						XFS_DIR3_FT_DIR};
>  
>  /*
> + * When we're checking directory inodes, we're allowed to set a directory's
> + * dotdot entry to zero to signal that the parent needs to be reconnected
> + * during phase 6.  If we're handling a shortform directory the ifork
> + * verifiers will fail, so temporarily patch out this canary so that we can
> + * verify the rest of the fork and move on to fixing the dir.
> + */
> +static xfs_failaddr_t
> +phase6_verify_dir(
> +	struct xfs_inode		*ip)
> +{
> +	struct xfs_mount		*mp = ip->i_mount;
> +	const struct xfs_dir_ops	*dops;
> +	struct xfs_ifork		*ifp;
> +	struct xfs_dir2_sf_hdr		*sfp;
> +	xfs_failaddr_t			fa;
> +	xfs_ino_t			old_parent;
> +	bool				parent_bypass = false;
> +	int				size;
> +
> +	dops = libxfs_dir_get_ops(mp, NULL);
> +
> +	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
> +	sfp = (struct xfs_dir2_sf_hdr *)ifp->if_u1.if_data;
> +	size = ifp->if_bytes;
> +
> +	/*
> +	 * If this is a shortform directory, phase4 may have set the parent
> +	 * inode to zero to indicate that it must be fixed.  Temporarily
> +	 * set a valid parent so that the directory verifier will pass.
> +	 */
> +	if (size > offsetof(struct xfs_dir2_sf_hdr, parent) &&
> +	    size >= xfs_dir2_sf_hdr_size(sfp->i8count)) {
> +		old_parent = dops->sf_get_parent_ino(sfp);
> +		if (old_parent == 0) {
> +			dops->sf_put_parent_ino(sfp, mp->m_sb.sb_rootino);
> +			parent_bypass = true;
> +		}
> +	}
> +
> +	fa = libxfs_default_ifork_ops.verify_dir(ip);
> +
> +	/* Put it back. */
> +	if (parent_bypass)
> +		dops->sf_put_parent_ino(sfp, old_parent);
> +
> +	return fa;
> +}
> +
> +static struct xfs_ifork_ops phase6_ifork_ops = {
> +	.verify_attr	= xfs_attr_shortform_verify,
> +	.verify_dir	= phase6_verify_dir,
> +	.verify_symlink	= xfs_symlink_shortform_verify,
> +};
> +
> +/*
>   * Data structures used to keep track of directories where the ".."
>   * entries are updated. These must be rebuilt after the initial pass
>   */
> @@ -2833,7 +2888,7 @@ process_dir_inode(
>  
>  	ASSERT(!is_inode_refchecked(irec, ino_offset) || dotdot_update);
>  
> -	error = -libxfs_iget(mp, NULL, ino, 0, &ip, NULL);
> +	error = -libxfs_iget(mp, NULL, ino, 0, &ip, &phase6_ifork_ops);
>  	if (error) {
>  		if (!no_modify)
>  			do_error(
> 

Re: [PATCH v2 3/4] xfs_fsr: refactor mountpoint finding to use libfrog paths functions

[Eric Sandeen]

On 3/20/18 10:19 PM, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
> 
> Refactor the mount-point finding code in fsr to use the libfrog helpers
> instead of open-coding yet another routine.
> 
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>

Seems fine, I'll test further but:

Reviewed-by: Eric Sandeen <sandeen@redhat.com>

> ---
> v2: fix obvious brokenness w/ path testing
> ---
>  fsr/Makefile    |    4 ++-
>  fsr/xfs_fsr.c   |   73 ++++++-------------------------------------------------
>  include/path.h  |    1 +
>  libfrog/paths.c |   48 ++++++++++++++++++++++++++----------
>  4 files changed, 47 insertions(+), 79 deletions(-)
> 
> diff --git a/fsr/Makefile b/fsr/Makefile
> index d3521b2..4201b38 100644
> --- a/fsr/Makefile
> +++ b/fsr/Makefile
> @@ -7,7 +7,9 @@ include $(TOPDIR)/include/builddefs
>  
>  LTCOMMAND = xfs_fsr
>  CFILES = xfs_fsr.c
> -LLDLIBS = $(LIBHANDLE)
> +LLDLIBS = $(LIBHANDLE) $(LIBFROG) $(LIBPTHREAD) $(LIBBLKID)
> +LTDEPENDENCIES = $(LIBHANDLE) $(LIBFROG)
> +LLDFLAGS = -static-libtool-libs
>  
>  ifeq ($(HAVE_GETMNTENT),yes)
>  LCFLAGS += -DHAVE_GETMNTENT
> diff --git a/fsr/xfs_fsr.c b/fsr/xfs_fsr.c
> index 2a18ce0..b74a70b 100644
> --- a/fsr/xfs_fsr.c
> +++ b/fsr/xfs_fsr.c
> @@ -22,6 +22,7 @@
>  #include "jdm.h"
>  #include "xfs_bmap_btree.h"
>  #include "xfs_attr_sf.h"
> +#include "path.h"
>  
>  #include <fcntl.h>
>  #include <errno.h>
> @@ -167,73 +168,13 @@ aborter(int unused)
>  	exit(1);
>  }
>  
> -/*
> - * Check if the argument is either the device name or mountpoint of an XFS
> - * filesystem.  Note that we do not care about bind mounted regular files
> - * here - the code that handles defragmentation of invidual files takes care
> - * of that.
> - */
> -static char *
> -find_mountpoint_check(struct stat *sb, struct mntent *t)
> -{
> -	struct stat ms;
> -
> -	if (S_ISDIR(sb->st_mode)) {		/* mount point */
> -		if (stat(t->mnt_dir, &ms) < 0)
> -			return NULL;
> -		if (sb->st_ino != ms.st_ino)
> -			return NULL;
> -		if (sb->st_dev != ms.st_dev)
> -			return NULL;
> -		if (strcmp(t->mnt_type, MNTTYPE_XFS) != 0)
> -			return NULL;
> -	} else {				/* device */
> -		if (stat(t->mnt_fsname, &ms) < 0)
> -			return NULL;
> -		if (sb->st_rdev != ms.st_rdev)
> -			return NULL;
> -		if (strcmp(t->mnt_type, MNTTYPE_XFS) != 0)
> -			return NULL;
> -		/*
> -		 * Make sure the mountpoint given by mtab is accessible
> -		 * before using it.
> -		 */
> -		if (stat(t->mnt_dir, &ms) < 0)
> -			return NULL;
> -	}
> -
> -	return t->mnt_dir;
> -}
> -
> -static char *
> -find_mountpoint(char *mtab, char *argname, struct stat *sb)
> -{
> -	struct mntent_cursor cursor;
> -	struct mntent *t = NULL;
> -	char *mntp = NULL;
> -
> -	if (platform_mntent_open(&cursor, mtab) != 0){
> -		fprintf(stderr, "Error: can't get mntent entries.\n");
> -		exit(1);
> -	}
> -
> -	while ((t = platform_mntent_next(&cursor)) != NULL) {
> -		mntp = find_mountpoint_check(sb, t);
> -		if (mntp == NULL)
> -			continue;
> -		break;
> -	}
> -	platform_mntent_close(&cursor);
> -	return mntp;
> -}
> -
>  int
>  main(int argc, char **argv)
>  {
>  	struct stat sb;
>  	char *argname;
>  	int c;
> -	char *mntp;
> +	struct fs_path	*fsp;
>  	char *mtab = NULL;
>  
>  	setlinebuf(stdout);
> @@ -322,7 +263,7 @@ main(int argc, char **argv)
>  	RealUid = getuid();
>  
>  	pagesize = getpagesize();
> -
> +	fs_table_initialise(0, NULL, 0, NULL);
>  	if (optind < argc) {
>  		for (; optind < argc; optind++) {
>  			argname = argv[optind];
> @@ -343,9 +284,11 @@ main(int argc, char **argv)
>  				sb = sb2;
>  			}
>  
> -			mntp = find_mountpoint(mtab, argname, &sb);
> -			if (mntp != NULL) {
> -				fsrfs(mntp, 0, 100);
> +			fsp = fs_table_lookup_mount(argname);
> +			if (!fsp)
> +				fsp = fs_table_lookup_blkdev(argname);
> +			if (fsp != NULL) {
> +				fsrfs(fsp->fs_dir, 0, 100);
>  			} else if (S_ISCHR(sb.st_mode)) {
>  				fprintf(stderr, _(
>  					"%s: char special not supported: %s\n"),
> diff --git a/include/path.h b/include/path.h
> index 1d3a902..88dc44b 100644
> --- a/include/path.h
> +++ b/include/path.h
> @@ -57,6 +57,7 @@ extern void fs_table_insert_project_path(char *__dir, uint __projid);
>  
>  extern fs_path_t *fs_table_lookup(const char *__dir, uint __flags);
>  extern fs_path_t *fs_table_lookup_mount(const char *__dir);
> +extern fs_path_t *fs_table_lookup_blkdev(const char *bdev);
>  
>  typedef struct fs_cursor {
>  	uint		count;		/* total count of mount entries	*/
> diff --git a/libfrog/paths.c b/libfrog/paths.c
> index 19ee1ea..318b48f 100644
> --- a/libfrog/paths.c
> +++ b/libfrog/paths.c
> @@ -89,30 +89,26 @@ fs_table_lookup(
>  	return NULL;
>  }
>  
> -/*
> - * Find the FS table entry describing an actual mount for the given path.
> - * Unlike fs_table_lookup(), fs_table_lookup_mount() compares the "dir"
> - * argument to actual mount point entries in the table. Accordingly, it
> - * will find matches only if the "dir" argument is indeed mounted.
> - */
> -struct fs_path *
> -fs_table_lookup_mount(
> -	const char	*dir)
> +static struct fs_path *
> +__fs_table_lookup_mount(
> +	const char	*dir,
> +	const char	*blkdev)
>  {
>  	uint		i;
> -	dev_t		dev = 0;
>  	char		rpath[PATH_MAX];
>  	char		dpath[PATH_MAX];
>  
> -	if (fs_device_number(dir, &dev))
> +	if (dir && !realpath(dir, dpath))
>  		return NULL;
> -	if (!realpath(dir, dpath))
> +	if (blkdev && !realpath(blkdev, dpath))
>  		return NULL;
>  
>  	for (i = 0; i < fs_count; i++) {
>  		if (fs_table[i].fs_flags != FS_MOUNT_POINT)
>  			continue;
> -		if (!realpath(fs_table[i].fs_dir, rpath))
> +		if (dir && !realpath(fs_table[i].fs_dir, rpath))
> +			continue;
> +		if (blkdev && !realpath(fs_table[i].fs_name, rpath))
>  			continue;
>  		if (strcmp(rpath, dpath) == 0)
>  			return &fs_table[i];
> @@ -120,6 +116,32 @@ fs_table_lookup_mount(
>  	return NULL;
>  }
>  
> +/*
> + * Find the FS table entry describing an actual mount for the given path.
> + * Unlike fs_table_lookup(), fs_table_lookup_mount() compares the "dir"
> + * argument to actual mount point entries in the table. Accordingly, it
> + * will find matches only if the "dir" argument is indeed mounted.
> + */
> +struct fs_path *
> +fs_table_lookup_mount(
> +	const char	*dir)
> +{
> +	return __fs_table_lookup_mount(dir, NULL);
> +}
> +
> +/*
> + * Find the FS table entry describing an actual mount for the block device.
> + * Unlike fs_table_lookup(), fs_table_lookup_blkdev() compares the "bdev"
> + * argument to actual mount point names in the table. Accordingly, it
> + * will find matches only if the "bdev" argument is indeed mounted.
> + */
> +struct fs_path *
> +fs_table_lookup_blkdev(
> +	const char	*bdev)
> +{
> +	return __fs_table_lookup_mount(NULL, bdev);
> +}
> +
>  static int
>  fs_table_insert(
>  	char		*dir,
> 

Re: [PATCH 5/4] misc: remove darwin, irix, and freebsd support

[Eric Sandeen]

On 3/20/18 10:19 PM, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
> 
> Remove these ports since they're not actively maintained:
> 
> IRIX support was partially removed last year; the OS itself hasn't had a
> release since 2006.
> 
> FreeBSD dropped XFS support in v9.3, which EOLd in January 2017.
> 
> Darwin/OSX has never supported XFS.
> 
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>

Probably should clean up more than just whole files, see bits & pieces
sprinkled throughout:

include/builddefs.in 
include/platform_defs.h.in
include/gnukfreebsd.h (?)
include/xfs.h 
libhandle/Makefile
quota/Makefile
io/Makefile
*/Makefile
m4/package_libcdev.m4 (AC_HAVE_GETMNTINFO)

...maybe even a wholesale removal of all references to PKG_PLATFORM at this point?

I suppose we can start with this and clean up more as we go.

> ---
>  include/darwin.h  |  315 -----------------------------------------------------
>  include/freebsd.h |  210 -----------------------------------
>  libxfs/Makefile   |    2 
>  libxfs/darwin.c   |  144 ------------------------
>  libxfs/freebsd.c  |  201 ----------------------------------
>  libxfs/irix.c     |  111 -------------------

Re: [PATCH 5/4] misc: remove darwin, irix, and freebsd support

[Eric Sandeen]

On 3/21/18 1:59 PM, Eric Sandeen wrote:
> On 3/20/18 10:19 PM, Darrick J. Wong wrote:
>> From: Darrick J. Wong <darrick.wong@oracle.com>
>>
>> Remove these ports since they're not actively maintained:
>>
>> IRIX support was partially removed last year; the OS itself hasn't had a
>> release since 2006.
>>
>> FreeBSD dropped XFS support in v9.3, which EOLd in January 2017.
>>
>> Darwin/OSX has never supported XFS.
>>
>> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> 
> Probably should clean up more than just whole files, see bits & pieces
> sprinkled throughout:
> 
> include/builddefs.in 
> include/platform_defs.h.in
> include/gnukfreebsd.h (?)
> include/xfs.h 
> libhandle/Makefile
> quota/Makefile
> io/Makefile
> */Makefile
> m4/package_libcdev.m4 (AC_HAVE_GETMNTINFO)
> 
> ...maybe even a wholesale removal of all references to PKG_PLATFORM at this point?
> 
> I suppose we can start with this and clean up more as we go.

./quota/darwin.c
./quota/freebsd.c
./quota/irix.c


...

[PATCH 5.5/4] misc: remove darwin, irix, and freebsd support

[Eric Sandeen]

Remove even more unsupported platform bits.

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
---

diff --git a/Makefile b/Makefile
index 7ddfa31..4ff2cac 100644
--- a/Makefile
+++ b/Makefile
@@ -47,13 +47,9 @@ HDR_SUBDIRS = include libxfs
 LIBFROG_SUBDIR = libfrog
 DLIB_SUBDIRS = libxlog libxcmd libhandle
 LIB_SUBDIRS = libxfs $(DLIB_SUBDIRS)
-TOOL_SUBDIRS = copy db estimate fsck growfs io logprint mkfs quota \
+TOOL_SUBDIRS = copy db estimate fsck fsr growfs io logprint mkfs quota \
 		mdrestore repair rtcp m4 man doc debian spaceman scrub
 
-ifneq ("$(PKG_PLATFORM)","darwin")
-TOOL_SUBDIRS += fsr
-endif
-
 ifneq ("$(XGETTEXT)","")
 TOOL_SUBDIRS += po
 endif
diff --git a/include/Makefile b/include/Makefile
index 6077507..09fe42c 100644
--- a/include/Makefile
+++ b/include/Makefile
@@ -47,14 +47,12 @@ LIBHFILES = libxfs.h \
 
 HFILES = handle.h \
 	jdm.h \
-	$(PKG_PLATFORM).h \
+	linux.h \
 	xfs.h \
 	xqm.h \
 	xfs_arch.h
 
-PHFILES = darwin.h freebsd.h linux.h gnukfreebsd.h
-LSRCFILES = $(shell echo $(PHFILES) | sed -e "s/$(PKG_PLATFORM).h//g")
-LSRCFILES += platform_defs.h.in builddefs.in buildmacros buildrules install-sh
+LSRCFILES = platform_defs.h.in builddefs.in buildmacros buildrules install-sh
 LSRCFILES += $(DKHFILES) $(LIBHFILES)
 LDIRT = disk
 LDIRDIRT = xfs
diff --git a/include/builddefs.in b/include/builddefs.in
index df76b2c..7f1f58c 100644
--- a/include/builddefs.in
+++ b/include/builddefs.in
@@ -48,7 +48,6 @@ PKG_USER	= @pkg_user@
 PKG_GROUP	= @pkg_group@
 PKG_RELEASE	= @pkg_release@
 PKG_VERSION	= @pkg_version@
-PKG_PLATFORM	= @pkg_platform@
 PKG_DISTRIBUTION= @pkg_distribution@
 
 prefix		= @prefix@
@@ -135,25 +134,12 @@ CROND_DIR = @crond_dir@
 GCCFLAGS = -funsigned-char -fno-strict-aliasing -Wall
 #	   -Wbitwise -Wno-transparent-union -Wno-old-initializer -Wno-decl
 
-ifeq ($(PKG_PLATFORM),linux)
 PCFLAGS = -D_GNU_SOURCE $(GCCFLAGS)
 ifeq ($(HAVE_UMODE_T),yes)
 PCFLAGS += -DHAVE_UMODE_T
 endif
 DEPENDFLAGS = -D__linux__
-endif
-ifeq ($(PKG_PLATFORM),gnukfreebsd)
-PCFLAGS = -D_GNU_SOURCE $(GCCFLAGS)
-endif
-ifeq ($(PKG_PLATFORM),darwin)
-PCFLAGS = $(GCCFLAGS)
-DEPENDFLAGS = -D__APPLE__ -D_DARWIN_FEATURE_64_BIT_INODE
-endif
-ifeq ($(PKG_PLATFORM),freebsd)
-PLDLIBS = -L/usr/local/lib -lintl
-PCFLAGS = -I/usr/local/include $(GCCFLAGS)
-DEPENDFLAGS = -D__FreeBSD__
-endif
+
 ifeq ($(HAVE_FLS),yes)
 LCFLAGS+= -DHAVE_FLS
 endif
diff --git a/include/gnukfreebsd.h b/include/gnukfreebsd.h
deleted file mode 100644
index 1db3f4f..0000000
--- a/include/gnukfreebsd.h
+++ /dev/null
@@ -1,156 +0,0 @@
-/*
- * Copyright (c) 2004-2006 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public License
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-#ifndef __XFS_KFREEBSD_H__
-#define __XFS_KFREEBSD_H__
-
-#include <uuid/uuid.h>
-#include <sys/vfs.h>
-#include <sys/ioctl.h>
-#include <sys/sysmacros.h>
-#include <malloc.h>
-#include <getopt.h>
-#include <endian.h>
-#include <sys/stat.h>
-#include <sys/param.h>
-#include <sys/mount.h>
-#include <ctype.h>
-#include <libgen.h>
-#include <paths.h>
-#include <mntent.h>
-
-#define EFSCORRUPTED	990	/* Filesystem is corrupted */
-#define EFSBADCRC	991	/* Bad CRC detected */
-
-typedef unsigned char		__u8;
-typedef signed char		__s8;
-typedef unsigned short		__u16;
-typedef signed short		__s16;
-typedef unsigned int		__u32;
-typedef signed int		__s32;
-typedef unsigned long long int	__u64;
-typedef signed long long int	__s64;
-
-typedef off_t		xfs_off_t;
-typedef uint64_t	xfs_ino_t;
-typedef uint32_t	xfs_dev_t;
-typedef int64_t		xfs_daddr_t;
-typedef __u32		xfs_nlink_t;
-
-#define HAVE_FID	1
-
-static __inline__ int xfsctl(const char *path, int fd, int cmd, void *p)
-{
-	return ioctl(fd, cmd, p);
-}
-
-static __inline__ int platform_test_xfs_fd(int fd)
-{
-	struct statfs buf;
-	if (fstatfs(fd, &buf) < 0)
-		return 0;
-	return strncmp(buf.f_fstypename, "xfs", 4) == 0;
-}
-
-static __inline__ int platform_test_xfs_path(const char *path)
-{
-	struct statfs buf;
-	if (statfs(path, &buf) < 0)
-		return 0;
-	return strncmp(buf.f_fstypename, "xfs", 4) == 0;
-}
-
-static __inline__ int platform_fstatfs(int fd, struct statfs *buf)
-{
-	return fstatfs(fd, buf);
-}
-
-static __inline__ void platform_getoptreset(void)
-{
-	extern int optind;
-	optind = 0;
-}
-
-static __inline__ int platform_uuid_compare(uuid_t *uu1, uuid_t *uu2)
-{
-	return uuid_compare(*uu1, *uu2);
-}
-
-static __inline__ void platform_uuid_unparse(uuid_t *uu, char *buffer)
-{
-	uuid_unparse(*uu, buffer);
-}
-
-static __inline__ int platform_uuid_parse(char *buffer, uuid_t *uu)
-{
-	return uuid_parse(buffer, *uu);
-}
-
-static __inline__ int platform_uuid_is_null(uuid_t *uu)
-{
-	return uuid_is_null(*uu);
-}
-
-static __inline__ void platform_uuid_generate(uuid_t *uu)
-{
-	uuid_generate(*uu);
-}
-
-static __inline__ void platform_uuid_clear(uuid_t *uu)
-{
-	uuid_clear(*uu);
-}
-
-static __inline__ void platform_uuid_copy(uuid_t *dst, uuid_t *src)
-{
-	uuid_copy(*dst, *src);
-}
-
-static __inline__ int
-platform_discard_blocks(int fd, uint64_t start, uint64_t len)
-{
-	return 0;
-}
-
-/**
- * Abstraction of mountpoints.
- */
-struct mntent_cursor {
-	FILE *mtabp;
-};
-
-static inline int platform_mntent_open(struct mntent_cursor * cursor, char *mtab)
-{
-	cursor->mtabp = setmntent(mtab, "r");
-	if (!cursor->mtabp) {
-		fprintf(stderr, "Error: cannot read %s\n", mtab);
-		return 1;
-	}
-	return 0;
-}
-
-static inline struct mntent * platform_mntent_next(struct mntent_cursor * cursor)
-{
-	return getmntent(cursor->mtabp);
-}
-
-static inline void platform_mntent_close(struct mntent_cursor * cursor)
-{
-	endmntent(cursor->mtabp);
-}
-
-#endif	/* __XFS_KFREEBSD_H__ */
diff --git a/include/xfs.h b/include/xfs.h
index a40ca0c..1cbe3ad 100644
--- a/include/xfs.h
+++ b/include/xfs.h
@@ -35,12 +35,6 @@
 
 #if defined(__linux__)
 #include <xfs/linux.h>
-#elif defined(__FreeBSD__)
-#include <xfs/freebsd.h>
-#elif defined(__FreeBSD_kernel__)
-#include <xfs/gnukfreebsd.h>
-#elif defined(__APPLE__)
-#include <xfs/darwin.h>
 #else
 # error unknown platform... have fun porting!
 #endif
diff --git a/io/Makefile b/io/Makefile
index 8055d4b..a251b28 100644
--- a/io/Makefile
+++ b/io/Makefile
@@ -9,10 +9,10 @@ LTCOMMAND = xfs_io
 LSRCFILES = xfs_bmap.sh xfs_freeze.sh xfs_mkfile.sh
 HFILES = init.h io.h
 CFILES = init.c \
-	attr.c bmap.c cowextsize.c encrypt.c file.c freeze.c fsync.c \
-	getrusage.c imap.c link.c mmap.c open.c parent.c pread.c prealloc.c \
-	pwrite.c reflink.c scrub.c seek.c shutdown.c stat.c sync.c truncate.c \
-	utimes.c
+	attr.c bmap.c cowextsize.c encrypt.c file.c freeze.c fsmap.c fsync.c \
+	getrusage.c imap.c inject.c link.c mmap.c open.c parent.c pread.c \
+	prealloc.c pwrite.c reflink.c resblks.c scrub.c seek.c shutdown.c \
+	stat.c sync.c truncate.c utimes.c
 
 LLDLIBS = $(LIBXCMD) $(LIBHANDLE) $(LIBFROG) $(LIBPTHREAD)
 LTDEPENDENCIES = $(LIBXCMD) $(LIBHANDLE) $(LIBFROG)
@@ -53,13 +53,6 @@ else
 LSRCFILES += fiemap.c
 endif
 
-ifeq ($(PKG_PLATFORM),irix)
-LSRCFILES += inject.c resblks.c
-else
-CFILES += inject.c resblks.c
-LCFLAGS += -DHAVE_INJECT -DHAVE_RESBLKS
-endif
-
 ifeq ($(HAVE_COPY_FILE_RANGE),yes)
 CFILES += copy_file_range.c
 LCFLAGS += -DHAVE_COPY_FILE_RANGE
@@ -114,13 +107,6 @@ LLDLIBS += $(LIBDEVMAPPER)
 LCFLAGS += -DHAVE_DEVMAPPER
 endif
 
-# On linux we get fsmap from the system or define it ourselves
-# so include this based on platform type.  If this reverts to only
-# the autoconf check w/o local definition, change to testing HAVE_GETFSMAP
-ifeq ($(PKG_PLATFORM),linux)
-CFILES += fsmap.c
-endif
-
 ifeq ($(HAVE_STATFS_FLAGS),yes)
 LCFLAGS += -DHAVE_STATFS_FLAGS
 endif
diff --git a/io/io.h b/io/io.h
index 5cf7c30..4516554 100644
--- a/io/io.h
+++ b/io/io.h
@@ -116,6 +116,7 @@ extern void		pread_init(void);
 extern void		prealloc_init(void);
 extern void		pwrite_init(void);
 extern void		quit_init(void);
+extern void		resblks_init(void);
 extern void		seek_init(void);
 extern void		shutdown_init(void);
 extern void		stat_init(void);
@@ -129,12 +130,6 @@ extern void		fadvise_init(void);
 #define fadvise_init()	do { } while (0)
 #endif
 
-#ifdef HAVE_RESBLKS
-extern void		resblks_init(void);
-#else
-#define resblks_init()	do { } while (0)
-#endif
-
 #ifdef HAVE_SENDFILE
 extern void		sendfile_init(void);
 #else
diff --git a/libhandle/Makefile b/libhandle/Makefile
index fe1a2af..cc4ad1d 100644
--- a/libhandle/Makefile
+++ b/libhandle/Makefile
@@ -10,11 +10,7 @@ LT_CURRENT = 1
 LT_REVISION = 3
 LT_AGE = 0
 
-ifeq ($(PKG_PLATFORM),darwin)
-LTLDFLAGS += -Wl,libhandle.sym
-else
 LTLDFLAGS += -Wl,--version-script,libhandle.sym
-endif
 
 CFILES = handle.c jdm.c
 LSRCFILES = libhandle.sym
diff --git a/libxfs/Makefile b/libxfs/Makefile
index 00df418..48f66d4 100644
--- a/libxfs/Makefile
+++ b/libxfs/Makefile
@@ -58,6 +58,7 @@ CFILES = cache.c \
 	defer_item.c \
 	init.c \
 	kmem.c \
+	linux.c \
 	logitem.c \
 	rdwr.c \
 	trans.c \
@@ -97,9 +98,6 @@ CFILES = cache.c \
 	xfs_symlink_remote.c \
 	xfs_trans_resv.c
 
-CFILES += $(PKG_PLATFORM).c
-PCFILES = linux.c
-LSRCFILES = $(shell echo $(PCFILES) | sed -e "s/$(PKG_PLATFORM).c//g")
 LSRCFILES += gen_crc32table.c
 
 #
diff --git a/m4/package_globals.m4 b/m4/package_globals.m4
index e469671..892e3bd 100644
--- a/m4/package_globals.m4
+++ b/m4/package_globals.m4
@@ -40,8 +40,4 @@ AC_DEFUN([AC_PACKAGE_GLOBALS],
     pkg_distribution=`uname -s`
     test -z "$DISTRIBUTION" || pkg_distribution="$DISTRIBUTION"
     AC_SUBST(pkg_distribution)
-
-    pkg_platform=`uname -s | tr 'A-Z' 'a-z' | tr -d / | sed -e 's/irix64/irix/'`
-    test -z "$PLATFORM" || pkg_platform="$PLATFORM"
-    AC_SUBST(pkg_platform)
   ])
diff --git a/quota/Makefile b/quota/Makefile
index 120af2e..62cc8a0 100644
--- a/quota/Makefile
+++ b/quota/Makefile
@@ -8,11 +8,7 @@ include $(TOPDIR)/include/builddefs
 LTCOMMAND = xfs_quota
 HFILES = init.h quota.h
 CFILES = init.c util.c \
-	edit.c free.c path.c project.c quot.c quota.c report.c state.c
-
-CFILES += $(PKG_PLATFORM).c
-PCFILES = darwin.c freebsd.c irix.c linux.c
-LSRCFILES = $(shell echo $(PCFILES) | sed -e "s/$(PKG_PLATFORM).c//g")
+	edit.c free.c linux.c path.c project.c quot.c quota.c report.c state.c
 
 LLDLIBS = $(LIBXCMD) $(LIBFROG)
 LTDEPENDENCIES = $(LIBXCMD) $(LIBFROG)
diff --git a/quota/darwin.c b/quota/darwin.c
deleted file mode 100644
index fb81f7f..0000000
--- a/quota/darwin.c
+++ /dev/null
@@ -1,33 +0,0 @@
-/*
- * Copyright (c) 2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-
-#include "quota.h"
-#include <sys/quota.h>
-
-int
-xfsquotactl(
-	int		command,
-	const char	*device,
-	uint		type,
-	uint		id,
-	void		*addr)
-{
-	/* return quotactl(device, QCMD(command, type), id, addr); */
-	errno = -ENOSYS;
-	return -1;
-}
diff --git a/quota/freebsd.c b/quota/freebsd.c
deleted file mode 100644
index 7770b3c..0000000
--- a/quota/freebsd.c
+++ /dev/null
@@ -1,31 +0,0 @@
-/*
- * Copyright (c) 2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-
-#include "quota.h"
-
-int
-xfsquotactl(
-	int		command,
-	const char	*device,
-	uint		type,
-	uint		id,
-	void		*addr)
-{
-	errno = -ENOSYS;
-	return -1;
-}
diff --git a/quota/irix.c b/quota/irix.c
deleted file mode 100644
index bbbcd18..0000000
--- a/quota/irix.c
+++ /dev/null
@@ -1,66 +0,0 @@
-/*
- * Copyright (c) 2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License as
- * published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it would be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with this program; if not, write the Free Software Foundation,
- * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
- */
-
-#include "quota.h"
-#include <sys/quota.h>
-
-static int
-xcommand_to_qcommand(
-	uint		command,
-	uint		type)
-{
-	switch (command) {
-	case XFS_QUOTAON:
-		return Q_XQUOTAON;
-	case XFS_QUOTAOFF:
-		return Q_XQUOTAOFF;
-	case XFS_GETQUOTA:
-		if (type == XFS_GROUP_QUOTA)
-			return Q_XGETGQUOTA;
-		if (type == XFS_PROJ_QUOTA)
-			return Q_XGETPQUOTA;
-		return Q_XGETQUOTA;
-	case XFS_SETQLIM:
-		if (type == XFS_GROUP_QUOTA)
-			return Q_XSETGQLIM;
-		if (type == XFS_PROJ_QUOTA)
-			return Q_XSETPQLIM;
-		return Q_XSETQLIM;
-	case XFS_GETQSTAT:
-		return Q_XGETQSTAT;
-	case XFS_QUOTARM:
-		return Q_XQUOTARM;
-	case XFS_QSYNC:
-		return Q_SYNC;
-	}
-	return 0;
-}
-
-int
-xfsquotactl(
-	int		command,
-	const char	*device,
-	uint		type,
-	uint		id,
-	void		*addr)
-{
-	int		qcommand;
-
-	qcommand = xcommand_to_qcommand(command, type);
-	return quotactl(qcommand, (char *)device, id, addr);
-}
diff --git a/scrub/Makefile b/scrub/Makefile
index 0632794..94c6313 100644
--- a/scrub/Makefile
+++ b/scrub/Makefile
@@ -8,9 +8,9 @@ include $(TOPDIR)/include/builddefs
 # On linux we get fsmap from the system or define it ourselves
 # so include this based on platform type.  If this reverts to only
 # the autoconf check w/o local definition, change to testing HAVE_GETFSMAP
-SCRUB_PREREQS=$(PKG_PLATFORM)$(HAVE_OPENAT)$(HAVE_FSTATAT)
+SCRUB_PREREQS=$(HAVE_OPENAT)$(HAVE_FSTATAT)
 
-ifeq ($(SCRUB_PREREQS),linuxyesyes)
+ifeq ($(SCRUB_PREREQS),yesyes)
 LTCOMMAND = xfs_scrub
 INSTALL_SCRUB = install-scrub
 XFS_SCRUB_ALL_PROG = xfs_scrub_all
diff --git a/spaceman/Makefile b/spaceman/Makefile
index 8b31030..a137efc 100644
--- a/spaceman/Makefile
+++ b/spaceman/Makefile
@@ -7,7 +7,7 @@ include $(TOPDIR)/include/builddefs
 
 LTCOMMAND = xfs_spaceman
 HFILES = init.h space.h
-CFILES = init.c file.c prealloc.c trim.c
+CFILES = freesp.c init.c file.c prealloc.c trim.c
 
 LLDLIBS = $(LIBXCMD) $(LIBFROG)
 LTDEPENDENCIES = $(LIBXCMD) $(LIBFROG)
@@ -21,13 +21,6 @@ ifeq ($(ENABLE_EDITLINE),yes)
 LLDLIBS += $(LIBEDITLINE) $(LIBTERMCAP)
 endif
 
-# On linux we get fsmap from the system or define it ourselves
-# so include this based on platform type.  If this reverts to only
-# the autoconf check w/o local definition, change to testing HAVE_GETFSMAP
-ifeq ($(PKG_PLATFORM),linux)
-CFILES += freesp.c
-endif
-
 default: depend $(LTCOMMAND)
 
 include $(BUILDRULES)

Re: [PATCH 6/4] libfrog: absorb platform specific code

[Eric Sandeen]

On 3/20/18 10:20 PM, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
> 
> Move the linux support code to libfrog, which should remove the final
> dependency of libfrog on libxfs.  libfrog is the runtime support library
> anyway.
> 
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>

sorry, my patch 5.5/4 probably kinda blew this up.

We can fold it in, or whatever ...

> ---
>  libfrog/Makefile |    4 +
>  libfrog/linux.c  |  265 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
>  libxfs/Makefile  |    3 -
>  libxfs/linux.c   |  265 ------------------------------------------------------
>  4 files changed, 269 insertions(+), 268 deletions(-)
>  create mode 100644 libfrog/linux.c
>  delete mode 100644 libxfs/linux.c
> 
> diff --git a/libfrog/Makefile b/libfrog/Makefile
> index 230b08f..e3065e6 100644
> --- a/libfrog/Makefile
> +++ b/libfrog/Makefile
> @@ -22,6 +22,10 @@ topology.c \
>  util.c \
>  workqueue.c
>  
> +CFILES += $(PKG_PLATFORM).c
> +PCFILES = linux.c
> +LSRCFILES = $(shell echo $(PCFILES) | sed -e "s/$(PKG_PLATFORM).c//g")

Ok so this doesn't really do anything at this point :)
(LSRCFILES will always be empty)

Just add linux.c to the main CFILES list and delete this

The rest makes sense, I think, unless quota/linux.c should get the same
treatment.  Probably not.

-Eric

Re: [PATCH 5.5/4] misc: remove darwin, irix, and freebsd support

[Darrick J. Wong]

On Wed, Mar 21, 2018 at 02:42:28PM -0500, Eric Sandeen wrote:
> Remove even more unsupported platform bits.
> 
> Signed-off-by: Eric Sandeen <sandeen@redhat.com>

Looks fine to me, though we should wait a few days to see if anyone
complains...

Looking at osx again, there /was/ a readonly fuse driver.  It claims to
support OSX 10.7 (EOLd 3.5 years ago) and bundles its own copy of
xfsprogs 3.1.4 (imported 6 years ago)... so I think it's probably ok to
cut it loose.

Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>

--D

> ---
> 
> diff --git a/Makefile b/Makefile
> index 7ddfa31..4ff2cac 100644
> --- a/Makefile
> +++ b/Makefile
> @@ -47,13 +47,9 @@ HDR_SUBDIRS = include libxfs
>  LIBFROG_SUBDIR = libfrog
>  DLIB_SUBDIRS = libxlog libxcmd libhandle
>  LIB_SUBDIRS = libxfs $(DLIB_SUBDIRS)
> -TOOL_SUBDIRS = copy db estimate fsck growfs io logprint mkfs quota \
> +TOOL_SUBDIRS = copy db estimate fsck fsr growfs io logprint mkfs quota \
>  		mdrestore repair rtcp m4 man doc debian spaceman scrub
>  
> -ifneq ("$(PKG_PLATFORM)","darwin")
> -TOOL_SUBDIRS += fsr
> -endif
> -
>  ifneq ("$(XGETTEXT)","")
>  TOOL_SUBDIRS += po
>  endif
> diff --git a/include/Makefile b/include/Makefile
> index 6077507..09fe42c 100644
> --- a/include/Makefile
> +++ b/include/Makefile
> @@ -47,14 +47,12 @@ LIBHFILES = libxfs.h \
>  
>  HFILES = handle.h \
>  	jdm.h \
> -	$(PKG_PLATFORM).h \
> +	linux.h \
>  	xfs.h \
>  	xqm.h \
>  	xfs_arch.h
>  
> -PHFILES = darwin.h freebsd.h linux.h gnukfreebsd.h
> -LSRCFILES = $(shell echo $(PHFILES) | sed -e "s/$(PKG_PLATFORM).h//g")
> -LSRCFILES += platform_defs.h.in builddefs.in buildmacros buildrules install-sh
> +LSRCFILES = platform_defs.h.in builddefs.in buildmacros buildrules install-sh
>  LSRCFILES += $(DKHFILES) $(LIBHFILES)
>  LDIRT = disk
>  LDIRDIRT = xfs
> diff --git a/include/builddefs.in b/include/builddefs.in
> index df76b2c..7f1f58c 100644
> --- a/include/builddefs.in
> +++ b/include/builddefs.in
> @@ -48,7 +48,6 @@ PKG_USER	= @pkg_user@
>  PKG_GROUP	= @pkg_group@
>  PKG_RELEASE	= @pkg_release@
>  PKG_VERSION	= @pkg_version@
> -PKG_PLATFORM	= @pkg_platform@
>  PKG_DISTRIBUTION= @pkg_distribution@
>  
>  prefix		= @prefix@
> @@ -135,25 +134,12 @@ CROND_DIR = @crond_dir@
>  GCCFLAGS = -funsigned-char -fno-strict-aliasing -Wall
>  #	   -Wbitwise -Wno-transparent-union -Wno-old-initializer -Wno-decl
>  
> -ifeq ($(PKG_PLATFORM),linux)
>  PCFLAGS = -D_GNU_SOURCE $(GCCFLAGS)
>  ifeq ($(HAVE_UMODE_T),yes)
>  PCFLAGS += -DHAVE_UMODE_T
>  endif
>  DEPENDFLAGS = -D__linux__
> -endif
> -ifeq ($(PKG_PLATFORM),gnukfreebsd)
> -PCFLAGS = -D_GNU_SOURCE $(GCCFLAGS)
> -endif
> -ifeq ($(PKG_PLATFORM),darwin)
> -PCFLAGS = $(GCCFLAGS)
> -DEPENDFLAGS = -D__APPLE__ -D_DARWIN_FEATURE_64_BIT_INODE
> -endif
> -ifeq ($(PKG_PLATFORM),freebsd)
> -PLDLIBS = -L/usr/local/lib -lintl
> -PCFLAGS = -I/usr/local/include $(GCCFLAGS)
> -DEPENDFLAGS = -D__FreeBSD__
> -endif
> +
>  ifeq ($(HAVE_FLS),yes)
>  LCFLAGS+= -DHAVE_FLS
>  endif
> diff --git a/include/gnukfreebsd.h b/include/gnukfreebsd.h
> deleted file mode 100644
> index 1db3f4f..0000000
> --- a/include/gnukfreebsd.h
> +++ /dev/null
> @@ -1,156 +0,0 @@
> -/*
> - * Copyright (c) 2004-2006 Silicon Graphics, Inc.
> - * All Rights Reserved.
> - *
> - * This program is free software; you can redistribute it and/or
> - * modify it under the terms of the GNU Lesser General Public License
> - * as published by the Free Software Foundation.
> - *
> - * This program is distributed in the hope that it would be useful,
> - * but WITHOUT ANY WARRANTY; without even the implied warranty of
> - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> - * GNU Lesser General Public License for more details.
> - *
> - * You should have received a copy of the GNU Lesser General Public License
> - * along with this program; if not, write the Free Software Foundation,
> - * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
> - */
> -#ifndef __XFS_KFREEBSD_H__
> -#define __XFS_KFREEBSD_H__
> -
> -#include <uuid/uuid.h>
> -#include <sys/vfs.h>
> -#include <sys/ioctl.h>
> -#include <sys/sysmacros.h>
> -#include <malloc.h>
> -#include <getopt.h>
> -#include <endian.h>
> -#include <sys/stat.h>
> -#include <sys/param.h>
> -#include <sys/mount.h>
> -#include <ctype.h>
> -#include <libgen.h>
> -#include <paths.h>
> -#include <mntent.h>
> -
> -#define EFSCORRUPTED	990	/* Filesystem is corrupted */
> -#define EFSBADCRC	991	/* Bad CRC detected */
> -
> -typedef unsigned char		__u8;
> -typedef signed char		__s8;
> -typedef unsigned short		__u16;
> -typedef signed short		__s16;
> -typedef unsigned int		__u32;
> -typedef signed int		__s32;
> -typedef unsigned long long int	__u64;
> -typedef signed long long int	__s64;
> -
> -typedef off_t		xfs_off_t;
> -typedef uint64_t	xfs_ino_t;
> -typedef uint32_t	xfs_dev_t;
> -typedef int64_t		xfs_daddr_t;
> -typedef __u32		xfs_nlink_t;
> -
> -#define HAVE_FID	1
> -
> -static __inline__ int xfsctl(const char *path, int fd, int cmd, void *p)
> -{
> -	return ioctl(fd, cmd, p);
> -}
> -
> -static __inline__ int platform_test_xfs_fd(int fd)
> -{
> -	struct statfs buf;
> -	if (fstatfs(fd, &buf) < 0)
> -		return 0;
> -	return strncmp(buf.f_fstypename, "xfs", 4) == 0;
> -}
> -
> -static __inline__ int platform_test_xfs_path(const char *path)
> -{
> -	struct statfs buf;
> -	if (statfs(path, &buf) < 0)
> -		return 0;
> -	return strncmp(buf.f_fstypename, "xfs", 4) == 0;
> -}
> -
> -static __inline__ int platform_fstatfs(int fd, struct statfs *buf)
> -{
> -	return fstatfs(fd, buf);
> -}
> -
> -static __inline__ void platform_getoptreset(void)
> -{
> -	extern int optind;
> -	optind = 0;
> -}
> -
> -static __inline__ int platform_uuid_compare(uuid_t *uu1, uuid_t *uu2)
> -{
> -	return uuid_compare(*uu1, *uu2);
> -}
> -
> -static __inline__ void platform_uuid_unparse(uuid_t *uu, char *buffer)
> -{
> -	uuid_unparse(*uu, buffer);
> -}
> -
> -static __inline__ int platform_uuid_parse(char *buffer, uuid_t *uu)
> -{
> -	return uuid_parse(buffer, *uu);
> -}
> -
> -static __inline__ int platform_uuid_is_null(uuid_t *uu)
> -{
> -	return uuid_is_null(*uu);
> -}
> -
> -static __inline__ void platform_uuid_generate(uuid_t *uu)
> -{
> -	uuid_generate(*uu);
> -}
> -
> -static __inline__ void platform_uuid_clear(uuid_t *uu)
> -{
> -	uuid_clear(*uu);
> -}
> -
> -static __inline__ void platform_uuid_copy(uuid_t *dst, uuid_t *src)
> -{
> -	uuid_copy(*dst, *src);
> -}
> -
> -static __inline__ int
> -platform_discard_blocks(int fd, uint64_t start, uint64_t len)
> -{
> -	return 0;
> -}
> -
> -/**
> - * Abstraction of mountpoints.
> - */
> -struct mntent_cursor {
> -	FILE *mtabp;
> -};
> -
> -static inline int platform_mntent_open(struct mntent_cursor * cursor, char *mtab)
> -{
> -	cursor->mtabp = setmntent(mtab, "r");
> -	if (!cursor->mtabp) {
> -		fprintf(stderr, "Error: cannot read %s\n", mtab);
> -		return 1;
> -	}
> -	return 0;
> -}
> -
> -static inline struct mntent * platform_mntent_next(struct mntent_cursor * cursor)
> -{
> -	return getmntent(cursor->mtabp);
> -}
> -
> -static inline void platform_mntent_close(struct mntent_cursor * cursor)
> -{
> -	endmntent(cursor->mtabp);
> -}
> -
> -#endif	/* __XFS_KFREEBSD_H__ */
> diff --git a/include/xfs.h b/include/xfs.h
> index a40ca0c..1cbe3ad 100644
> --- a/include/xfs.h
> +++ b/include/xfs.h
> @@ -35,12 +35,6 @@
>  
>  #if defined(__linux__)
>  #include <xfs/linux.h>
> -#elif defined(__FreeBSD__)
> -#include <xfs/freebsd.h>
> -#elif defined(__FreeBSD_kernel__)
> -#include <xfs/gnukfreebsd.h>
> -#elif defined(__APPLE__)
> -#include <xfs/darwin.h>
>  #else
>  # error unknown platform... have fun porting!
>  #endif
> diff --git a/io/Makefile b/io/Makefile
> index 8055d4b..a251b28 100644
> --- a/io/Makefile
> +++ b/io/Makefile
> @@ -9,10 +9,10 @@ LTCOMMAND = xfs_io
>  LSRCFILES = xfs_bmap.sh xfs_freeze.sh xfs_mkfile.sh
>  HFILES = init.h io.h
>  CFILES = init.c \
> -	attr.c bmap.c cowextsize.c encrypt.c file.c freeze.c fsync.c \
> -	getrusage.c imap.c link.c mmap.c open.c parent.c pread.c prealloc.c \
> -	pwrite.c reflink.c scrub.c seek.c shutdown.c stat.c sync.c truncate.c \
> -	utimes.c
> +	attr.c bmap.c cowextsize.c encrypt.c file.c freeze.c fsmap.c fsync.c \
> +	getrusage.c imap.c inject.c link.c mmap.c open.c parent.c pread.c \
> +	prealloc.c pwrite.c reflink.c resblks.c scrub.c seek.c shutdown.c \
> +	stat.c sync.c truncate.c utimes.c
>  
>  LLDLIBS = $(LIBXCMD) $(LIBHANDLE) $(LIBFROG) $(LIBPTHREAD)
>  LTDEPENDENCIES = $(LIBXCMD) $(LIBHANDLE) $(LIBFROG)
> @@ -53,13 +53,6 @@ else
>  LSRCFILES += fiemap.c
>  endif
>  
> -ifeq ($(PKG_PLATFORM),irix)
> -LSRCFILES += inject.c resblks.c
> -else
> -CFILES += inject.c resblks.c
> -LCFLAGS += -DHAVE_INJECT -DHAVE_RESBLKS
> -endif
> -
>  ifeq ($(HAVE_COPY_FILE_RANGE),yes)
>  CFILES += copy_file_range.c
>  LCFLAGS += -DHAVE_COPY_FILE_RANGE
> @@ -114,13 +107,6 @@ LLDLIBS += $(LIBDEVMAPPER)
>  LCFLAGS += -DHAVE_DEVMAPPER
>  endif
>  
> -# On linux we get fsmap from the system or define it ourselves
> -# so include this based on platform type.  If this reverts to only
> -# the autoconf check w/o local definition, change to testing HAVE_GETFSMAP
> -ifeq ($(PKG_PLATFORM),linux)
> -CFILES += fsmap.c
> -endif
> -
>  ifeq ($(HAVE_STATFS_FLAGS),yes)
>  LCFLAGS += -DHAVE_STATFS_FLAGS
>  endif
> diff --git a/io/io.h b/io/io.h
> index 5cf7c30..4516554 100644
> --- a/io/io.h
> +++ b/io/io.h
> @@ -116,6 +116,7 @@ extern void		pread_init(void);
>  extern void		prealloc_init(void);
>  extern void		pwrite_init(void);
>  extern void		quit_init(void);
> +extern void		resblks_init(void);
>  extern void		seek_init(void);
>  extern void		shutdown_init(void);
>  extern void		stat_init(void);
> @@ -129,12 +130,6 @@ extern void		fadvise_init(void);
>  #define fadvise_init()	do { } while (0)
>  #endif
>  
> -#ifdef HAVE_RESBLKS
> -extern void		resblks_init(void);
> -#else
> -#define resblks_init()	do { } while (0)
> -#endif
> -
>  #ifdef HAVE_SENDFILE
>  extern void		sendfile_init(void);
>  #else
> diff --git a/libhandle/Makefile b/libhandle/Makefile
> index fe1a2af..cc4ad1d 100644
> --- a/libhandle/Makefile
> +++ b/libhandle/Makefile
> @@ -10,11 +10,7 @@ LT_CURRENT = 1
>  LT_REVISION = 3
>  LT_AGE = 0
>  
> -ifeq ($(PKG_PLATFORM),darwin)
> -LTLDFLAGS += -Wl,libhandle.sym
> -else
>  LTLDFLAGS += -Wl,--version-script,libhandle.sym
> -endif
>  
>  CFILES = handle.c jdm.c
>  LSRCFILES = libhandle.sym
> diff --git a/libxfs/Makefile b/libxfs/Makefile
> index 00df418..48f66d4 100644
> --- a/libxfs/Makefile
> +++ b/libxfs/Makefile
> @@ -58,6 +58,7 @@ CFILES = cache.c \
>  	defer_item.c \
>  	init.c \
>  	kmem.c \
> +	linux.c \
>  	logitem.c \
>  	rdwr.c \
>  	trans.c \
> @@ -97,9 +98,6 @@ CFILES = cache.c \
>  	xfs_symlink_remote.c \
>  	xfs_trans_resv.c
>  
> -CFILES += $(PKG_PLATFORM).c
> -PCFILES = linux.c
> -LSRCFILES = $(shell echo $(PCFILES) | sed -e "s/$(PKG_PLATFORM).c//g")
>  LSRCFILES += gen_crc32table.c
>  
>  #
> diff --git a/m4/package_globals.m4 b/m4/package_globals.m4
> index e469671..892e3bd 100644
> --- a/m4/package_globals.m4
> +++ b/m4/package_globals.m4
> @@ -40,8 +40,4 @@ AC_DEFUN([AC_PACKAGE_GLOBALS],
>      pkg_distribution=`uname -s`
>      test -z "$DISTRIBUTION" || pkg_distribution="$DISTRIBUTION"
>      AC_SUBST(pkg_distribution)
> -
> -    pkg_platform=`uname -s | tr 'A-Z' 'a-z' | tr -d / | sed -e 's/irix64/irix/'`
> -    test -z "$PLATFORM" || pkg_platform="$PLATFORM"
> -    AC_SUBST(pkg_platform)
>    ])
> diff --git a/quota/Makefile b/quota/Makefile
> index 120af2e..62cc8a0 100644
> --- a/quota/Makefile
> +++ b/quota/Makefile
> @@ -8,11 +8,7 @@ include $(TOPDIR)/include/builddefs
>  LTCOMMAND = xfs_quota
>  HFILES = init.h quota.h
>  CFILES = init.c util.c \
> -	edit.c free.c path.c project.c quot.c quota.c report.c state.c
> -
> -CFILES += $(PKG_PLATFORM).c
> -PCFILES = darwin.c freebsd.c irix.c linux.c
> -LSRCFILES = $(shell echo $(PCFILES) | sed -e "s/$(PKG_PLATFORM).c//g")
> +	edit.c free.c linux.c path.c project.c quot.c quota.c report.c state.c
>  
>  LLDLIBS = $(LIBXCMD) $(LIBFROG)
>  LTDEPENDENCIES = $(LIBXCMD) $(LIBFROG)
> diff --git a/quota/darwin.c b/quota/darwin.c
> deleted file mode 100644
> index fb81f7f..0000000
> --- a/quota/darwin.c
> +++ /dev/null
> @@ -1,33 +0,0 @@
> -/*
> - * Copyright (c) 2005 Silicon Graphics, Inc.
> - * All Rights Reserved.
> - *
> - * This program is free software; you can redistribute it and/or
> - * modify it under the terms of the GNU General Public License as
> - * published by the Free Software Foundation.
> - *
> - * This program is distributed in the hope that it would be useful,
> - * but WITHOUT ANY WARRANTY; without even the implied warranty of
> - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> - * GNU General Public License for more details.
> - *
> - * You should have received a copy of the GNU General Public License
> - * along with this program; if not, write the Free Software Foundation,
> - * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
> - */
> -
> -#include "quota.h"
> -#include <sys/quota.h>
> -
> -int
> -xfsquotactl(
> -	int		command,
> -	const char	*device,
> -	uint		type,
> -	uint		id,
> -	void		*addr)
> -{
> -	/* return quotactl(device, QCMD(command, type), id, addr); */
> -	errno = -ENOSYS;
> -	return -1;
> -}
> diff --git a/quota/freebsd.c b/quota/freebsd.c
> deleted file mode 100644
> index 7770b3c..0000000
> --- a/quota/freebsd.c
> +++ /dev/null
> @@ -1,31 +0,0 @@
> -/*
> - * Copyright (c) 2005 Silicon Graphics, Inc.
> - * All Rights Reserved.
> - *
> - * This program is free software; you can redistribute it and/or
> - * modify it under the terms of the GNU General Public License as
> - * published by the Free Software Foundation.
> - *
> - * This program is distributed in the hope that it would be useful,
> - * but WITHOUT ANY WARRANTY; without even the implied warranty of
> - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> - * GNU General Public License for more details.
> - *
> - * You should have received a copy of the GNU General Public License
> - * along with this program; if not, write the Free Software Foundation,
> - * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
> - */
> -
> -#include "quota.h"
> -
> -int
> -xfsquotactl(
> -	int		command,
> -	const char	*device,
> -	uint		type,
> -	uint		id,
> -	void		*addr)
> -{
> -	errno = -ENOSYS;
> -	return -1;
> -}
> diff --git a/quota/irix.c b/quota/irix.c
> deleted file mode 100644
> index bbbcd18..0000000
> --- a/quota/irix.c
> +++ /dev/null
> @@ -1,66 +0,0 @@
> -/*
> - * Copyright (c) 2005 Silicon Graphics, Inc.
> - * All Rights Reserved.
> - *
> - * This program is free software; you can redistribute it and/or
> - * modify it under the terms of the GNU General Public License as
> - * published by the Free Software Foundation.
> - *
> - * This program is distributed in the hope that it would be useful,
> - * but WITHOUT ANY WARRANTY; without even the implied warranty of
> - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> - * GNU General Public License for more details.
> - *
> - * You should have received a copy of the GNU General Public License
> - * along with this program; if not, write the Free Software Foundation,
> - * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
> - */
> -
> -#include "quota.h"
> -#include <sys/quota.h>
> -
> -static int
> -xcommand_to_qcommand(
> -	uint		command,
> -	uint		type)
> -{
> -	switch (command) {
> -	case XFS_QUOTAON:
> -		return Q_XQUOTAON;
> -	case XFS_QUOTAOFF:
> -		return Q_XQUOTAOFF;
> -	case XFS_GETQUOTA:
> -		if (type == XFS_GROUP_QUOTA)
> -			return Q_XGETGQUOTA;
> -		if (type == XFS_PROJ_QUOTA)
> -			return Q_XGETPQUOTA;
> -		return Q_XGETQUOTA;
> -	case XFS_SETQLIM:
> -		if (type == XFS_GROUP_QUOTA)
> -			return Q_XSETGQLIM;
> -		if (type == XFS_PROJ_QUOTA)
> -			return Q_XSETPQLIM;
> -		return Q_XSETQLIM;
> -	case XFS_GETQSTAT:
> -		return Q_XGETQSTAT;
> -	case XFS_QUOTARM:
> -		return Q_XQUOTARM;
> -	case XFS_QSYNC:
> -		return Q_SYNC;
> -	}
> -	return 0;
> -}
> -
> -int
> -xfsquotactl(
> -	int		command,
> -	const char	*device,
> -	uint		type,
> -	uint		id,
> -	void		*addr)
> -{
> -	int		qcommand;
> -
> -	qcommand = xcommand_to_qcommand(command, type);
> -	return quotactl(qcommand, (char *)device, id, addr);
> -}
> diff --git a/scrub/Makefile b/scrub/Makefile
> index 0632794..94c6313 100644
> --- a/scrub/Makefile
> +++ b/scrub/Makefile
> @@ -8,9 +8,9 @@ include $(TOPDIR)/include/builddefs
>  # On linux we get fsmap from the system or define it ourselves
>  # so include this based on platform type.  If this reverts to only
>  # the autoconf check w/o local definition, change to testing HAVE_GETFSMAP
> -SCRUB_PREREQS=$(PKG_PLATFORM)$(HAVE_OPENAT)$(HAVE_FSTATAT)
> +SCRUB_PREREQS=$(HAVE_OPENAT)$(HAVE_FSTATAT)
>  
> -ifeq ($(SCRUB_PREREQS),linuxyesyes)
> +ifeq ($(SCRUB_PREREQS),yesyes)
>  LTCOMMAND = xfs_scrub
>  INSTALL_SCRUB = install-scrub
>  XFS_SCRUB_ALL_PROG = xfs_scrub_all
> diff --git a/spaceman/Makefile b/spaceman/Makefile
> index 8b31030..a137efc 100644
> --- a/spaceman/Makefile
> +++ b/spaceman/Makefile
> @@ -7,7 +7,7 @@ include $(TOPDIR)/include/builddefs
>  
>  LTCOMMAND = xfs_spaceman
>  HFILES = init.h space.h
> -CFILES = init.c file.c prealloc.c trim.c
> +CFILES = freesp.c init.c file.c prealloc.c trim.c
>  
>  LLDLIBS = $(LIBXCMD) $(LIBFROG)
>  LTDEPENDENCIES = $(LIBXCMD) $(LIBFROG)
> @@ -21,13 +21,6 @@ ifeq ($(ENABLE_EDITLINE),yes)
>  LLDLIBS += $(LIBEDITLINE) $(LIBTERMCAP)
>  endif
>  
> -# On linux we get fsmap from the system or define it ourselves
> -# so include this based on platform type.  If this reverts to only
> -# the autoconf check w/o local definition, change to testing HAVE_GETFSMAP
> -ifeq ($(PKG_PLATFORM),linux)
> -CFILES += freesp.c
> -endif
> -
>  default: depend $(LTCOMMAND)
>  
>  include $(BUILDRULES)
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 5/4] misc: remove darwin, irix, and freebsd support

[Dave Chinner]

On Wed, Mar 21, 2018 at 01:59:26PM -0500, Eric Sandeen wrote:
> On 3/20/18 10:19 PM, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> > 
> > Remove these ports since they're not actively maintained:
> > 
> > IRIX support was partially removed last year; the OS itself hasn't had a
> > release since 2006.
> > 
> > FreeBSD dropped XFS support in v9.3, which EOLd in January 2017.
> > 
> > Darwin/OSX has never supported XFS.
> > 
> > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> 
> Probably should clean up more than just whole files, see bits & pieces
> sprinkled throughout:
> 
> include/builddefs.in 
> include/platform_defs.h.in
> include/gnukfreebsd.h (?)

So, this makes me wonder. IIRC, that's the header for the debian
distro on a freebsd kernel. That's a thing, and we package xfsprogs
in debian. Do they build xfsprogs for that platform, and if they do
do we care if we break it?

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

Re: [PATCH 5/4] misc: remove darwin, irix, and freebsd support

[Darrick J. Wong]

On Thu, Mar 22, 2018 at 08:26:17AM +1100, Dave Chinner wrote:
> On Wed, Mar 21, 2018 at 01:59:26PM -0500, Eric Sandeen wrote:
> > On 3/20/18 10:19 PM, Darrick J. Wong wrote:
> > > From: Darrick J. Wong <darrick.wong@oracle.com>
> > > 
> > > Remove these ports since they're not actively maintained:
> > > 
> > > IRIX support was partially removed last year; the OS itself hasn't had a
> > > release since 2006.
> > > 
> > > FreeBSD dropped XFS support in v9.3, which EOLd in January 2017.
> > > 
> > > Darwin/OSX has never supported XFS.
> > > 
> > > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> > 
> > Probably should clean up more than just whole files, see bits & pieces
> > sprinkled throughout:
> > 
> > include/builddefs.in 
> > include/platform_defs.h.in
> > include/gnukfreebsd.h (?)
> 
> So, this makes me wonder. IIRC, that's the header for the debian
> distro on a freebsd kernel. That's a thing, and we package xfsprogs
> in debian. Do they build xfsprogs for that platform, and if they do
> do we care if we break it?

>From the looks of it[1] they (Debian) don't package xfsprogs for
kfreebsd (not even the wheezy release that supposedly had an xfs driver
in the kernle) so I doubt anyone would care.

--D

[1] https://packages.debian.org/search?keywords=xfsprogs

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

Re: [PATCH 5/4] misc: remove darwin, irix, and freebsd support

[Dave Chinner]

On Wed, Mar 21, 2018 at 02:31:38PM -0700, Darrick J. Wong wrote:
> On Thu, Mar 22, 2018 at 08:26:17AM +1100, Dave Chinner wrote:
> > On Wed, Mar 21, 2018 at 01:59:26PM -0500, Eric Sandeen wrote:
> > > On 3/20/18 10:19 PM, Darrick J. Wong wrote:
> > > > From: Darrick J. Wong <darrick.wong@oracle.com>
> > > > 
> > > > Remove these ports since they're not actively maintained:
> > > > 
> > > > IRIX support was partially removed last year; the OS itself hasn't had a
> > > > release since 2006.
> > > > 
> > > > FreeBSD dropped XFS support in v9.3, which EOLd in January 2017.
> > > > 
> > > > Darwin/OSX has never supported XFS.
> > > > 
> > > > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> > > 
> > > Probably should clean up more than just whole files, see bits & pieces
> > > sprinkled throughout:
> > > 
> > > include/builddefs.in 
> > > include/platform_defs.h.in
> > > include/gnukfreebsd.h (?)
> > 
> > So, this makes me wonder. IIRC, that's the header for the debian
> > distro on a freebsd kernel. That's a thing, and we package xfsprogs
> > in debian. Do they build xfsprogs for that platform, and if they do
> > do we care if we break it?
> 
> From the looks of it[1] they (Debian) don't package xfsprogs for
> kfreebsd (not even the wheezy release that supposedly had an xfs driver
> in the kernle) so I doubt anyone would care.

No worries, just making sure we cover all bases, because it seems
not every knows what "gnu/freebsd" is :P

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

[PATCH 1.5/4] xfs_repair: use custom ifork verifier in mv_orphanage

[Eric Sandeen]

Now that we have a custom verifier which can ignore parent
inode numbers, use it in mv_orphanage() as well; orphan inodes
may have invalid parents, and we're about to reconnect
them anyway, so override that test when we get them off disk.

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
---

diff --git a/repair/phase6.c b/repair/phase6.c
index ed005e8..498a3b5 100644
--- a/repair/phase6.c
+++ b/repair/phase6.c
@@ -1138,7 +1138,8 @@ mv_orphanage(
 		xname.len = snprintf((char *)fname, sizeof(fname), "%llu.%d",
 					(unsigned long long)ino, ++incr);
 
-	err = -libxfs_iget(mp, NULL, ino, 0, &ino_p, &xfs_default_ifork_ops);
+	/* Orphans may not have a proper parent, so use custom ops here */
+	err = -libxfs_iget(mp, NULL, ino, 0, &ino_p, &phase6_ifork_ops);
 	if (err)
 		do_error(_("%d - couldn't iget disconnected inode\n"), err);
 

Re: [PATCH 1.5/4] xfs_repair: use custom ifork verifier in mv_orphanage

[Darrick J. Wong]

On Thu, Mar 22, 2018 at 02:35:34PM -0500, Eric Sandeen wrote:
> Now that we have a custom verifier which can ignore parent
> inode numbers, use it in mv_orphanage() as well; orphan inodes
> may have invalid parents, and we're about to reconnect
> them anyway, so override that test when we get them off disk.
> 
> Signed-off-by: Eric Sandeen <sandeen@redhat.com>

Looks ok,
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>

--D

> ---
> 
> diff --git a/repair/phase6.c b/repair/phase6.c
> index ed005e8..498a3b5 100644
> --- a/repair/phase6.c
> +++ b/repair/phase6.c
> @@ -1138,7 +1138,8 @@ mv_orphanage(
>  		xname.len = snprintf((char *)fname, sizeof(fname), "%llu.%d",
>  					(unsigned long long)ino, ++incr);
>  
> -	err = -libxfs_iget(mp, NULL, ino, 0, &ino_p, &xfs_default_ifork_ops);
> +	/* Orphans may not have a proper parent, so use custom ops here */
> +	err = -libxfs_iget(mp, NULL, ino, 0, &ino_p, &phase6_ifork_ops);
>  	if (err)
>  		do_error(_("%d - couldn't iget disconnected inode\n"), err);
>  
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH 7/4] xfs_spaceman: remove incorrect linux/fs.h include

[Darrick J. Wong]

From: Darrick J. Wong <darrick.wong@oracle.com>

Remove the direct linux/fs.h include from spaceman because all xfs
utilities should include xfs.h so that we can wrap missing kernel header
declarations properly.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
 spaceman/trim.c |    1 -
 1 file changed, 1 deletion(-)

diff --git a/spaceman/trim.c b/spaceman/trim.c
index 6df9e6a..c1efb52 100644
--- a/spaceman/trim.c
+++ b/spaceman/trim.c
@@ -17,7 +17,6 @@
  */
 
 #include "libxfs.h"
-#include <linux/fs.h>
 #include "command.h"
 #include "init.h"
 #include "path.h"