[PATCH 1/6] blkid: fix ZFS device detection

[PATCH 1/6] blkid: fix ZFS device detection

[Andreas Dilger]

From: Andreas Dilger <andreas.dilger@intel.com>

While the main blkid functionality is in util-linux, there is still
use for blkid on non-Linux platforms.

Fix the ZFS device detection by looking at multiple uberblocks to
see if any are present, rather than looking for the ZFS boot block
which is not always present.

There are slots for up to 128 uberblocks, but the first 4 are not
written to disk on a newly-formatted filesystem, so check several
of them at different offsets within the uberblock array.

Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
---
 lib/blkid/probe.c | 17 +++++++++++++----
 1 file changed, 13 insertions(+), 4 deletions(-)

diff --git a/lib/blkid/probe.c b/lib/blkid/probe.c
index 4b797ab..161f4f1 100644
--- a/lib/blkid/probe.c
+++ b/lib/blkid/probe.c
@@ -1440,10 +1440,19 @@ static struct blkid_magic type_array[] = {
   { "iso9660",	32,	 1,  5, "CD001",		probe_iso9660 },
   { "iso9660",	32,	 9,  5, "CDROM",		probe_iso9660 },
   { "jfs",	32,	 0,  4, "JFS1",			probe_jfs },
-  { "zfs",       8,	 0,  8, "\0\0\x02\xf5\xb0\x07\xb1\x0c", probe_zfs },
-  { "zfs",       8,	 0,  8, "\x0c\xb1\x07\xb0\xf5\x02\0\0", probe_zfs },
-  { "zfs",     264,	 0,  8, "\0\0\x02\xf5\xb0\x07\xb1\x0c", probe_zfs },
-  { "zfs",     264,	 0,  8, "\x0c\xb1\x07\xb0\xf5\x02\0\0", probe_zfs },
+  /* ZFS has 128 root blocks (#4 is the first used), check only 6 of them */
+  { "zfs",     128,	 0,  8, "\0\0\0\0\0\xba\xb1\x0c", probe_zfs },
+  { "zfs",     128,	 0,  8, "\x0c\xb1\xba\0\0\0\0\0", probe_zfs },
+  { "zfs",     132,	 0,  8, "\0\0\0\0\0\xba\xb1\x0c", probe_zfs },
+  { "zfs",     132,	 0,  8, "\x0c\xb1\xba\0\0\0\0\0", probe_zfs },
+  { "zfs",     136,	 0,  8, "\0\0\0\0\0\xba\xb1\x0c", probe_zfs },
+  { "zfs",     136,	 0,  8, "\x0c\xb1\xba\0\0\0\0\0", probe_zfs },
+  { "zfs",     384,	 0,  8, "\0\0\0\0\0\xba\xb1\x0c", probe_zfs },
+  { "zfs",     384,	 0,  8, "\x0c\xb1\xba\0\0\0\0\0", probe_zfs },
+  { "zfs",     388,	 0,  8, "\0\0\0\0\0\xba\xb1\x0c", probe_zfs },
+  { "zfs",     388,	 0,  8, "\x0c\xb1\xba\0\0\0\0\0", probe_zfs },
+  { "zfs",     392,	 0,  8, "\0\0\0\0\0\xba\xb1\x0c", probe_zfs },
+  { "zfs",     392,	 0,  8, "\x0c\xb1\xba\0\0\0\0\0", probe_zfs },
   { "hfsplus",	 1,	 0,  2, "BD",			probe_hfsplus },
   { "hfsplus",	 1,	 0,  2, "H+",			probe_hfsplus },
   { "hfsplus",	 1,	 0,  2, "HX",			probe_hfsplus },
-- 
1.8.0

[PATCH 2/6] debugfs: improve dump_mmp handling

[Andreas Dilger]

If MMP is not enabled on a filesystem (s_mmp_block == 0), print this
clearly rather than "MMP: block number beyond filesystem range".

Add an option to "debugfs dump_mmp" to specify the MMP block number
instead of getting it from the superblock s_mmp_block field.

Signed-off-by: Andreas Dilger <adilger@dilger.ca>
---
 debugfs/debugfs.8.in |  9 +++++++--
 debugfs/debugfs.c    | 25 ++++++++++++++++++++++---
 2 files changed, 29 insertions(+), 5 deletions(-)

diff --git a/debugfs/debugfs.8.in b/debugfs/debugfs.8.in
index bae14db..4928791 100644
--- a/debugfs/debugfs.8.in
+++ b/debugfs/debugfs.8.in
@@ -221,8 +221,13 @@ option is given set the owner, group and permissions information on
 to match
 .IR filespec .
 .TP
-.B dump_mmp
-Display the multiple-mount protection (mmp) field values.
+.BI dump_mmp " [mmp_block]"
+Display the multiple-mount protection (mmp) field values.  If
+.I mmp_block
+is specified then verify and dump the MMP values from the given block
+number, otherwise use the
+.B s_mmp_block
+field in the superblock to locate and use the existing MMP block.
 .TP
 .BI dx_hash " [-h hash_alg] [-s hash_seed] filename"
 Calculate the directory hash of
diff --git a/debugfs/debugfs.c b/debugfs/debugfs.c
index 5423634..260698c 100644
--- a/debugfs/debugfs.c
+++ b/debugfs/debugfs.c
@@ -2351,12 +2351,31 @@ try_again:
 void do_dump_mmp(int argc EXT2FS_ATTR((unused)), char *argv[])
 {
 	struct mmp_struct *mmp_s;
+	unsigned long long mmp_block;
 	time_t t;
 	errcode_t retval = 0;
 
 	if (check_fs_open(argv[0]))
 		return;
 
+	if (argc > 1) {
+		char *end = NULL;
+		mmp_block = strtoull(argv[1], &end, 0);
+		if (end == argv[0] || mmp_block == 0) {
+			fprintf(stderr, "%s: invalid MMP block '%s' given\n",
+				argv[0], argv[1]);
+			return;
+		}
+	} else {
+		mmp_block = current_fs->super->s_mmp_block;
+	}
+
+	if (mmp_block == 0) {
+		fprintf(stderr, "%s: MMP: not active on this filesystem.\n",
+			argv[0]);
+		return;
+	}
+
 	if (current_fs->mmp_buf == NULL) {
 		retval = ext2fs_get_mem(current_fs->blocksize,
 					&current_fs->mmp_buf);
@@ -2368,10 +2387,10 @@ void do_dump_mmp(int argc EXT2FS_ATTR((unused)), char *argv[])
 
 	mmp_s = current_fs->mmp_buf;
 
-	retval = ext2fs_mmp_read(current_fs, current_fs->super->s_mmp_block,
-				 current_fs->mmp_buf);
+	retval = ext2fs_mmp_read(current_fs, mmp_block, current_fs->mmp_buf);
 	if (retval) {
-		com_err(argv[0], retval, "reading MMP block.\n");
+		com_err(argv[0], retval, "reading MMP block %llu.\n",
+			mmp_block);
 		return;
 	}
 
-- 
1.8.0

Re: [PATCH 2/6] debugfs: improve dump_mmp handling

[Theodore Ts'o]

On Sat, Mar 05, 2016 at 05:38:32PM -0700, Andreas Dilger wrote:
> If MMP is not enabled on a filesystem (s_mmp_block == 0), print this
> clearly rather than "MMP: block number beyond filesystem range".
> 
> Add an option to "debugfs dump_mmp" to specify the MMP block number
> instead of getting it from the superblock s_mmp_block field.
> 
> Signed-off-by: Andreas Dilger <adilger@dilger.ca>

Thanks, applied.

					- Ted

[PATCH 3/6] misc: add missing declarations on maint

[Andreas Dilger]

Fix compile warnings for missing declarations on the maint branch.

Signed-off-by: Andreas Dilger <adilger@dilger.ca>
---
 misc/util.c  | 3 +++
 util/subst.c | 4 ++++
 2 files changed, 7 insertions(+)

diff --git a/misc/util.c b/misc/util.c
index 2898830..5a24e7e 100644
--- a/misc/util.c
+++ b/misc/util.c
@@ -21,6 +21,9 @@
 #ifdef HAVE_ERRNO_H
 #include <errno.h>
 #endif
+#if HAVE_UNISTD_H
+#include <unistd.h>
+#endif
 #ifdef HAVE_LINUX_MAJOR_H
 #include <linux/major.h>
 #endif
diff --git a/util/subst.c b/util/subst.c
index 91f6d44..db72923 100644
--- a/util/subst.c
+++ b/util/subst.c
@@ -19,8 +19,12 @@
 #ifdef HAVE_SYS_TIME_H
 #include <sys/time.h>
 #endif
+#ifdef HAVE_SYS_TYPES_H
 #include <sys/types.h>
+#endif
+#ifdef HAVE_SYS_STAT_H
 #include <sys/stat.h>
+#endif
 #include <fcntl.h>
 #include <time.h>
 #include <utime.h>
-- 
1.8.0

Re: [PATCH 3/6] misc: add missing declarations on maint

[Theodore Ts'o]

On Sat, Mar 05, 2016 at 05:38:33PM -0700, Andreas Dilger wrote:
> Fix compile warnings for missing declarations on the maint branch.
> 
> Signed-off-by: Andreas Dilger <adilger@dilger.ca>

Thanks, applied.

						- Ted

[PATCH 4/6] lsattr: treat inode generation as an unsigned int

[Andreas Dilger]

The EXT2_GETVERSION ioctl is defined to take a "long" parameter, but
fgetversion() calls ioctl() with an "int" parameter instead.  This is
handled in the kernel correctly, but the generation is sign-extended
in fgetversion() before return on 64-bit systems and lsattr prints
it as a huge positive number for inode generation above 0x80000000:

        1635574212 -------------e-- /mnt/ost0/O/0/d0/12928
        18446744073045131735 -------------e-- /mnt/ost0/O/0/d0/166240
        782808861 -------------e-- /mnt/ost0/O/0/d0/31744
        18446744072181134840 -------------e-- /mnt/ost0/O/0/d0/135008

Correctly assign the returned generation number as an unsigned value,
and print it with a 10-character field width.  The version is printed
left-aligned for consistency with the old code and to ensure it is
always printed in the first column for use with tools like "cut":

        1635574212 -------------e-- /mnt/ost0/O/0/d0/12928
        3630547415 -------------e-- /mnt/ost0/O/0/d0/166240
        782808861  -------------e-- /mnt/ost0/O/0/d0/31744
        2766550520 -------------e-- /mnt/ost0/O/0/d0/135008

Do not return a random value from the stack as the version on error.
Clean up some style issues and consolidate some duplicate code.

Signed-off-by: Andreas Dilger <adilger@dilger.ca>
---
 lib/e2p/fgetversion.c | 36 +++++++++++++++++++-----------------
 misc/lsattr.c         |  2 +-
 2 files changed, 20 insertions(+), 18 deletions(-)

diff --git a/lib/e2p/fgetversion.c b/lib/e2p/fgetversion.c
index e6cee8b..2ad813b 100644
--- a/lib/e2p/fgetversion.c
+++ b/lib/e2p/fgetversion.c
@@ -37,32 +37,34 @@
 #define OPEN_FLAGS (O_RDONLY|O_NONBLOCK)
 #endif
 
-int fgetversion (const char * name, unsigned long * version)
+int fgetversion(const char *name, unsigned long *version)
 {
+	unsigned int ver = -1;
+	int rc = -1;
 #if HAVE_EXT2_IOCTLS
-#if !APPLE_DARWIN
-	int fd, r, ver, save_errno = 0;
+# if !APPLE_DARWIN
+	int fd, save_errno = 0;
 
-	fd = open (name, OPEN_FLAGS);
+	fd = open(name, OPEN_FLAGS);
 	if (fd == -1)
 		return -1;
-	r = ioctl (fd, EXT2_IOC_GETVERSION, &ver);
-	if (r == -1)
+
+	rc = ioctl(fd, EXT2_IOC_GETVERSION, &ver);
+	if (rc == -1)
 		save_errno = errno;
-	*version = ver;
-	close (fd);
-	if (save_errno)
+	close(fd);
+	if (rc == -1)
 		errno = save_errno;
-	return r;
-#else
-   int ver=-1, err;
-   err = syscall(SYS_fsctl, name, EXT2_IOC_GETVERSION, &ver, 0);
-   *version = ver;
-   return(err);
-#endif
+# else /* APPLE_DARWIN */
+	rc = syscall(SYS_fsctl, name, EXT2_IOC_GETVERSION, &ver, 0);
+# endif /* !APPLE_DARWIN */
 #else /* ! HAVE_EXT2_IOCTLS */
 	extern int errno;
+
 	errno = EOPNOTSUPP;
-	return -1;
 #endif /* ! HAVE_EXT2_IOCTLS */
+	if (rc == 0)
+		*version = ver;
+
+	return rc;
 }
diff --git a/misc/lsattr.c b/misc/lsattr.c
index e5e5969..4c34e2f 100644
--- a/misc/lsattr.c
+++ b/misc/lsattr.c
@@ -92,7 +92,7 @@ static int list_attributes (const char * name)
 				 name);
 			return -1;
 		}
-		printf ("%5lu ", generation);
+		printf ("%-10lu ", generation);
 	}
 	if (pf_options & PFOPT_LONG) {
 		printf("%-28s ", name);
-- 
1.8.0

Re: [PATCH 4/6] lsattr: treat inode generation as an unsigned int

[Theodore Ts'o]

On Sat, Mar 05, 2016 at 05:38:34PM -0700, Andreas Dilger wrote:
> The EXT2_GETVERSION ioctl is defined to take a "long" parameter, but
> fgetversion() calls ioctl() with an "int" parameter instead.  This is
> handled in the kernel correctly, but the generation is sign-extended
> in fgetversion() before return on 64-bit systems and lsattr prints
> it as a huge positive number for inode generation above 0x80000000:
> 
>         1635574212 -------------e-- /mnt/ost0/O/0/d0/12928
>         18446744073045131735 -------------e-- /mnt/ost0/O/0/d0/166240
>         782808861 -------------e-- /mnt/ost0/O/0/d0/31744
>         18446744072181134840 -------------e-- /mnt/ost0/O/0/d0/135008
> 
> Correctly assign the returned generation number as an unsigned value,
> and print it with a 10-character field width.  The version is printed
> left-aligned for consistency with the old code and to ensure it is
> always printed in the first column for use with tools like "cut":
> 
>         1635574212 -------------e-- /mnt/ost0/O/0/d0/12928
>         3630547415 -------------e-- /mnt/ost0/O/0/d0/166240
>         782808861  -------------e-- /mnt/ost0/O/0/d0/31744
>         2766550520 -------------e-- /mnt/ost0/O/0/d0/135008
> 
> Do not return a random value from the stack as the version on error.
> Clean up some style issues and consolidate some duplicate code.
> 
> Signed-off-by: Andreas Dilger <adilger@dilger.ca>

Thanks, applied.

						- Ted

[PATCH 5/6] findsuper: improve output if s_mkfs_time unset

[Andreas Dilger]

If s_mkfs_time is not set in the superblock, print the s_mtime field
instead to identify the different superblocks.  This can happen if the
superblock is corrupted, since s_mkfs_time is not reset by e2fsck.

Signed-off-by: Andreas Dilger <adilger@dilger.ca>
---
 misc/findsuper.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/misc/findsuper.c b/misc/findsuper.c
index eb9130b..b3b76cd 100644
--- a/misc/findsuper.c
+++ b/misc/findsuper.c
@@ -187,7 +187,7 @@ int main(int argc, char *argv[])
 	if (print_jnl_copies)
 		printf(_("[*] probably superblock written in the ext3 "
 			 "journal superblock,\n\tso start/end/grp wrong\n"));
-	printf(_("byte_offset  byte_start     byte_end  fs_blocks blksz  grp  last_mount_time           sb_uuid label\n"));
+	printf(_("byte_offset  byte_start     byte_end  fs_blocks blksz  grp  mkfs/mount_time           sb_uuid label\n"));
 	for (; lseek64(fd, sk, SEEK_SET) != -1 &&
 	       read(fd, &ext2, 512) == 512; sk += skiprate) {
 		static unsigned char last_uuid[16] = "blah";
@@ -230,7 +230,10 @@ int main(int argc, char *argv[])
 			WHY("free_inodes_count > inodes_count (%u > %u)\n",
 			    ext2.s_free_inodes_count, ext2.s_inodes_count);
 
-		tm = ext2.s_mtime;
+		if (ext2.s_mkfs_time != 0)
+			tm = ext2.s_mkfs_time;
+		else
+			tm = ext2.s_mtime;
 		s = ctime(&tm);
 		s[24] = 0;
 		bsize = 1 << (ext2.s_log_block_size + 10);
-- 
1.8.0

Re: [PATCH 5/6] findsuper: improve output if s_mkfs_time unset

[Theodore Ts'o]

On Sat, Mar 05, 2016 at 05:38:35PM -0700, Andreas Dilger wrote:
> If s_mkfs_time is not set in the superblock, print the s_mtime field
> instead to identify the different superblocks.  This can happen if the
> superblock is corrupted, since s_mkfs_time is not reset by e2fsck.
> 
> Signed-off-by: Andreas Dilger <adilger@dilger.ca>

Thanks, applied.

						- Ted

[PATCH 6/6] debugfs: update to handle newer inode fields

[Andreas Dilger]

Remove the printing of fragment fields that are HURD only, and
instead print the i_checksum_lo and i_checksum_hi fields.

Allow setting newer fields in extra inode space with modify_inode.

Add inode_includes() macro to simplify checking for fields in the
large inode, rather than hard-coding field offsets in the code.

Fix corresponding tests for newly added Checksum field.

Signed-off-by: Andreas Dilger <adilger@dilger.ca>
---
 debugfs/debugfs.c            | 197 +++++++++++++++++++++++++++----------------
 lib/ext2fs/ext2_fs.h         |  22 ++---
 tests/d_special_files/expect |  10 +--
 tests/f_badcluster/expect    |  14 +--
 4 files changed, 140 insertions(+), 103 deletions(-)

diff --git a/debugfs/debugfs.c b/debugfs/debugfs.c
index 260698c..332fd5a 100644
--- a/debugfs/debugfs.c
+++ b/debugfs/debugfs.c
@@ -527,6 +527,10 @@ static void dump_xattr_string(FILE *out, const char *str, int len)
 			fprintf(out, "%02x ", (unsigned char)str[i]);
 }
 
+#define inode_includes(size, field)					\
+	(size >= (sizeof(((struct ext2_inode_large *)0)->field) +	\
+		  offsetof(struct ext2_inode_large, field)))
+
 static void internal_dump_inode_extra(FILE *out,
 				      const char *prefix EXT2FS_ATTR((unused)),
 				      ext2_ino_t inode_num EXT2FS_ATTR((unused)),
@@ -725,15 +729,12 @@ void internal_dump_inode(FILE *out, const char *prefix,
 			 int do_dump_blocks)
 {
 	const char *i_type;
-	char frag, fsize;
 	int os = current_fs->super->s_creator_os;
-	struct ext2_inode_large *large_inode;
-	int is_large_inode = 0;
+	struct ext2_inode_large *large_inode = (struct ext2_inode_large *)inode;
+	unsigned inode_size = EXT2_GOOD_OLD_INODE_SIZE;
 
 	if (EXT2_INODE_SIZE(current_fs->super) > EXT2_GOOD_OLD_INODE_SIZE)
-		is_large_inode = 1;
-	large_inode = (struct ext2_inode_large *) inode;
-
+		inode_size += large_inode->i_extra_isize;
 	if (LINUX_S_ISDIR(inode->i_mode)) i_type = "directory";
 	else if (LINUX_S_ISREG(inode->i_mode)) i_type = "regular";
 	else if (LINUX_S_ISLNK(inode->i_mode)) i_type = "symlink";
@@ -745,14 +746,14 @@ void internal_dump_inode(FILE *out, const char *prefix,
 	fprintf(out, "%sInode: %u   Type: %s    ", prefix, inode_num, i_type);
 	fprintf(out, "%sMode:  %04o   Flags: 0x%x\n",
 		prefix, inode->i_mode & 0777, inode->i_flags);
-	if (is_large_inode && large_inode->i_extra_isize >= 24) {
+
+	if (inode_includes(inode_size, i_version_hi))
 		fprintf(out, "%sGeneration: %u    Version: 0x%08x:%08x\n",
 			prefix, inode->i_generation, large_inode->i_version_hi,
 			inode->osd1.linux1.l_i_version);
-	} else {
+	else
 		fprintf(out, "%sGeneration: %u    Version: 0x%08x\n", prefix,
 			inode->i_generation, inode->osd1.linux1.l_i_version);
-	}
 	fprintf(out, "%sUser: %5d   Group: %5d   Size: ",
 		prefix, inode_uid(*inode), inode_gid(*inode));
 	if (LINUX_S_ISREG(inode->i_mode))
@@ -763,7 +764,8 @@ void internal_dump_inode(FILE *out, const char *prefix,
 		fprintf(out,
 			"%sFile ACL: %d    Directory ACL: %d Translator: %d\n",
 			prefix,
-			inode->i_file_acl, LINUX_S_ISDIR(inode->i_mode) ? inode->i_dir_acl : 0,
+			inode->i_file_acl, LINUX_S_ISDIR(inode->i_mode) ?
+							inode->i_dir_acl : 0,
 			inode->osd1.hurd1.h_i_translator);
 	else
 		fprintf(out, "%sFile ACL: %llu    Directory ACL: %d\n",
@@ -780,46 +782,53 @@ void internal_dump_inode(FILE *out, const char *prefix,
 	else
 		fprintf(out, "%sLinks: %d   Blockcount: %u\n",
 			prefix, inode->i_links_count, inode->i_blocks);
-	switch (os) {
-	    case EXT2_OS_HURD:
-		frag = inode->osd2.hurd2.h_i_frag;
-		fsize = inode->osd2.hurd2.h_i_fsize;
-		break;
-	    default:
-		frag = fsize = 0;
-	}
-	fprintf(out, "%sFragment:  Address: %d    Number: %d    Size: %d\n",
-		prefix, inode->i_faddr, frag, fsize);
-	if (is_large_inode && large_inode->i_extra_isize >= 24) {
-		fprintf(out, "%s ctime: 0x%08x:%08x -- %s", prefix,
+	if (os == EXT2_OS_HURD)
+		fprintf(out,
+			"%sFragment:  Address: %d    Number: %d    Size: %d\n",
+			prefix, inode->i_faddr, inode->osd2.hurd2.h_i_frag,
+			inode->osd2.hurd2.h_i_fsize);
+	else
+		fprintf(out,
+			"%sFragment Address: %d    Checksum: 0x%04x%04x\n",
+			prefix, inode->i_faddr,
+			inode_includes(inode_size, i_checksum_hi) ?
+				large_inode->i_checksum_hi : 0,
+			inode->osd2.linux2.l_i_checksum_lo);
+	if (inode_includes(inode_size, i_ctime_extra))
+		fprintf(out, "%sctime: 0x%08x:%08x -- %s", prefix,
 			inode->i_ctime, large_inode->i_ctime_extra,
 			time_to_string(inode->i_ctime));
-		fprintf(out, "%s atime: 0x%08x:%08x -- %s", prefix,
+	else
+		fprintf(out, "%sctime: 0x%08x -- %s", prefix, inode->i_ctime,
+			time_to_string(inode->i_ctime));
+	if (inode_includes(inode_size, i_atime_extra))
+		fprintf(out, "%satime: 0x%08x:%08x -- %s", prefix,
 			inode->i_atime, large_inode->i_atime_extra,
 			time_to_string(inode->i_atime));
-		fprintf(out, "%s mtime: 0x%08x:%08x -- %s", prefix,
+	else
+		fprintf(out, "%satime: 0x%08x -- %s", prefix, inode->i_atime,
+			time_to_string(inode->i_atime));
+	if (inode_includes(inode_size, i_mtime_extra))
+		fprintf(out, "%smtime: 0x%08x:%08x -- %s", prefix,
 			inode->i_mtime, large_inode->i_mtime_extra,
 			time_to_string(inode->i_mtime));
+	else
+		fprintf(out, "%smtime: 0x%08x -- %s", prefix, inode->i_mtime,
+			time_to_string(inode->i_mtime));
+	if (inode_includes(inode_size, i_crtime_extra))
 		fprintf(out, "%scrtime: 0x%08x:%08x -- %s", prefix,
 			large_inode->i_crtime, large_inode->i_crtime_extra,
 			time_to_string(large_inode->i_crtime));
-	} else {
-		fprintf(out, "%sctime: 0x%08x -- %s", prefix, inode->i_ctime,
-			time_to_string(inode->i_ctime));
-		fprintf(out, "%satime: 0x%08x -- %s", prefix, inode->i_atime,
-			time_to_string(inode->i_atime));
-		fprintf(out, "%smtime: 0x%08x -- %s", prefix, inode->i_mtime,
-			time_to_string(inode->i_mtime));
-	}
 	if (inode->i_dtime)
-	  fprintf(out, "%sdtime: 0x%08x -- %s", prefix, inode->i_dtime,
-		  time_to_string(inode->i_dtime));
-	if (EXT2_INODE_SIZE(current_fs->super) > EXT2_GOOD_OLD_INODE_SIZE)
+		fprintf(out, "%sdtime: 0x%08x -- %s", prefix, inode->i_dtime,
+			time_to_string(inode->i_dtime));
+	if (inode_size > EXT2_GOOD_OLD_INODE_SIZE)
 		internal_dump_inode_extra(out, prefix, inode_num,
-					  (struct ext2_inode_large *) inode);
-	if (LINUX_S_ISLNK(inode->i_mode) && ext2fs_inode_data_blocks(current_fs,inode) == 0)
+					  (struct ext2_inode_large *)inode);
+	if (LINUX_S_ISLNK(inode->i_mode) &&
+	    ext2fs_inode_data_blocks(current_fs, inode) == 0)
 		fprintf(out, "%sFast_link_dest: %.*s\n", prefix,
-			(int) inode->i_size, (char *)inode->i_block);
+			(int)inode->i_size, (char *)inode->i_block);
 	else if (LINUX_S_ISBLK(inode->i_mode) || LINUX_S_ISCHR(inode->i_mode)) {
 		int major, minor;
 		const char *devnote;
@@ -1202,15 +1211,15 @@ static void modify_u32(char *com, const char *prompt,
 		*val = v;
 }
 
-
 void do_modify_inode(int argc, char *argv[])
 {
-	struct ext2_inode inode;
+	struct ext2_inode_large inode;
+	struct ext2_inode *inodep;
 	ext2_ino_t	inode_num;
-	int 		i;
-	unsigned char	*frag, *fsize;
+	int		i;
+	int		inode_size;
 	char		buf[80];
-	int 		os;
+	int		os;
 	const char	*hex_format = "0x%x";
 	const char	*octal_format = "0%o";
 	const char	*decimal_format = "%d";
@@ -1220,53 +1229,61 @@ void do_modify_inode(int argc, char *argv[])
 		return;
 
 	os = current_fs->super->s_creator_os;
+	inode_size = EXT2_INODE_SIZE(current_fs->super);
 
-	if (debugfs_read_inode(inode_num, &inode, argv[1]))
+	inodep = (struct ext2_inode *)&inode;
+	if (debugfs_read_inode_full(inode_num, inodep, argv[1], inode_size))
 		return;
 
 	modify_u16(argv[0], "Mode", octal_format, &inode.i_mode);
 	modify_u16(argv[0], "User ID", decimal_format, &inode.i_uid);
 	modify_u16(argv[0], "Group ID", decimal_format, &inode.i_gid);
 	modify_u32(argv[0], "Size", unsignedlong_format, &inode.i_size);
-	modify_u32(argv[0], "Creation time", decimal_format, &inode.i_ctime);
-	modify_u32(argv[0], "Modification time", decimal_format, &inode.i_mtime);
+	modify_u32(argv[0], "High 32bits of size", decimal_format,
+		   &inode.i_size_high);
+	modify_u32(argv[0], "Change time", decimal_format, &inode.i_ctime);
+	modify_u32(argv[0], "Modification time",decimal_format, &inode.i_mtime);
 	modify_u32(argv[0], "Access time", decimal_format, &inode.i_atime);
 	modify_u32(argv[0], "Deletion time", decimal_format, &inode.i_dtime);
 	modify_u16(argv[0], "Link count", decimal_format, &inode.i_links_count);
+	modify_u32(argv[0], "Block count",unsignedlong_format, &inode.i_blocks);
 	if (os == EXT2_OS_LINUX)
-		modify_u16(argv[0], "Block count high", unsignedlong_format,
+		modify_u16(argv[0], "High 16bits of lock count",
+			   unsignedlong_format,
 			   &inode.osd2.linux2.l_i_blocks_hi);
-	modify_u32(argv[0], "Block count", unsignedlong_format, &inode.i_blocks);
 	modify_u32(argv[0], "File flags", hex_format, &inode.i_flags);
 	modify_u32(argv[0], "Generation", hex_format, &inode.i_generation);
-#if 0
-	modify_u32(argv[0], "Reserved1", decimal_format, &inode.i_reserved1);
-#endif
-	modify_u32(argv[0], "File acl", decimal_format, &inode.i_file_acl);
-	if (LINUX_S_ISDIR(inode.i_mode))
-		modify_u32(argv[0], "Directory acl", decimal_format, &inode.i_dir_acl);
-	else
-		modify_u32(argv[0], "High 32bits of size", decimal_format, &inode.i_size_high);
+	modify_u32(argv[0], "File ACL", decimal_format, &inode.i_file_acl);
 
-	if (os == EXT2_OS_HURD)
+	if (os == EXT2_OS_HURD) {
 		modify_u32(argv[0], "Translator Block",
 			    decimal_format, &inode.osd1.hurd1.h_i_translator);
-
-	modify_u32(argv[0], "Fragment address", decimal_format, &inode.i_faddr);
-	switch (os) {
-	    case EXT2_OS_HURD:
-		frag = &inode.osd2.hurd2.h_i_frag;
-		fsize = &inode.osd2.hurd2.h_i_fsize;
-		break;
-	    default:
-		frag = fsize = 0;
-	}
-	if (frag)
-		modify_u8(argv[0], "Fragment number", decimal_format, frag);
-	if (fsize)
-		modify_u8(argv[0], "Fragment size", decimal_format, fsize);
-
-	for (i=0;  i < EXT2_NDIR_BLOCKS; i++) {
+		modify_u32(argv[0], "Fragment address", decimal_format,
+			   &inode.i_faddr);
+		modify_u8(argv[0], "Fragment number", decimal_format,
+			  &inode.osd2.hurd2.h_i_frag);
+		modify_u8(argv[0], "Fragment size", decimal_format,
+			  &inode.osd2.hurd2.h_i_fsize);
+	} else {
+		modify_u32(argv[0], "Fragment address", decimal_format,
+			   &inode.i_faddr);
+		modify_u32(argv[0], "Change Version", decimal_format,
+			   &inode.osd1.linux1.l_i_version);
+		modify_u16(argv[0], "High 16bits of file ACL",
+			   unsignedlong_format,
+			   &inode.osd2.linux2.l_i_file_acl_high);
+		modify_u16(argv[0], "High 16bits of user ID",
+			   unsignedlong_format,
+			   &inode.osd2.linux2.l_i_uid_high);
+		modify_u16(argv[0], "High 16bits of group ID",
+			   unsignedlong_format,
+			   &inode.osd2.linux2.l_i_gid_high);
+		modify_u16(argv[0], "Low 16bits of checksum",
+			   unsignedlong_format,
+			   &inode.osd2.linux2.l_i_checksum_lo);
+	}
+
+	for (i = 0; i < EXT2_NDIR_BLOCKS; i++) {
 		sprintf(buf, "Direct Block #%d", i);
 		modify_u32(argv[0], buf, decimal_format, &inode.i_block[i]);
 	}
@@ -1276,7 +1293,39 @@ void do_modify_inode(int argc, char *argv[])
 		    &inode.i_block[EXT2_DIND_BLOCK]);
 	modify_u32(argv[0], "Triple Indirect Block", decimal_format,
 		    &inode.i_block[EXT2_TIND_BLOCK]);
-	if (debugfs_write_inode(inode_num, &inode, argv[1]))
+
+	if (inode_includes(inode_size, i_extra_isize)) {
+		modify_u16(argv[0], "Extra inode size", decimal_format,
+			   &inode.i_extra_isize);
+		i = inode.i_extra_isize;
+	} else {
+		i = 0;
+	}
+
+	if (inode_includes(inode_size, i_checksum_hi))
+		modify_u16(argv[0], "High 16bits of checksum", decimal_format,
+			   &inode.i_checksum_hi);
+	if (inode_includes(inode_size, i_ctime_extra))
+		modify_u32(argv[0], "Extra change time",
+			   decimal_format, &inode.i_ctime_extra);
+	if (inode_includes(inode_size, i_mtime_extra))
+		modify_u32(argv[0], "Extra modification time",
+			   decimal_format, &inode.i_mtime_extra);
+	if (inode_includes(inode_size, i_atime_extra))
+		modify_u32(argv[0], "Extra access time",
+			   decimal_format, &inode.i_atime_extra);
+	if (inode_includes(inode_size, i_crtime))
+		modify_u32(argv[0], "Creation time",
+			   decimal_format, &inode.i_crtime);
+	if (inode_includes(inode_size, i_crtime_extra))
+		modify_u32(argv[0], "Extra creation time",
+			   decimal_format, &inode.i_crtime_extra);
+	if (inode_includes(inode_size, i_version_hi))
+		modify_u32(argv[0], "High 32bits change version",
+			   decimal_format, &inode.i_version_hi);
+
+	if (debugfs_write_inode_full(inode_num, inodep, argv[1],
+				     EXT2_INODE_SIZE(current_fs->super)))
 		return;
 }
 #endif /* READ_ONLY */
diff --git a/lib/ext2fs/ext2_fs.h b/lib/ext2fs/ext2_fs.h
index 6c3620c..42024fa 100644
--- a/lib/ext2fs/ext2_fs.h
+++ b/lib/ext2fs/ext2_fs.h
@@ -461,26 +461,14 @@ struct ext2_inode_large {
 
 #define i_dir_acl	i_size_high
 
-#if defined(__KERNEL__) || defined(__linux__)
-#define i_reserved1	osd1.linux1.l_i_reserved1
-#define i_frag		osd2.linux2.l_i_frag
-#define i_fsize		osd2.linux2.l_i_fsize
-#define i_uid_low	i_uid
-#define i_gid_low	i_gid
-#define i_uid_high	osd2.linux2.l_i_uid_high
-#define i_gid_high	osd2.linux2.l_i_gid_high
-#else
 #if defined(__GNU__)
-
-#define i_translator	osd1.hurd1.h_i_translator
-#define i_frag		osd2.hurd2.h_i_frag;
-#define i_fsize		osd2.hurd2.h_i_fsize;
 #define i_uid_high	osd2.hurd2.h_i_uid_high
 #define i_gid_high	osd2.hurd2.h_i_gid_high
-#define i_author	osd2.hurd2.h_i_author
-
-#endif  /* __GNU__ */
-#endif	/* defined(__KERNEL__) || defined(__linux__) */
+#else
+#define i_uid_high	osd2.linux2.l_i_uid_high
+#define i_gid_high	osd2.linux2.l_i_gid_high
+#define i_checksum_lo	osd2.linux2.l_i_checksum_lo
+#endif	/* __GNU__ */
 
 #define inode_uid(inode)	((inode).i_uid | (inode).osd2.linux2.l_i_uid_high << 16)
 #define inode_gid(inode)	((inode).i_gid | (inode).osd2.linux2.l_i_gid_high << 16)
diff --git a/tests/d_special_files/expect b/tests/d_special_files/expect
index 2b2dbfa..cfdc505 100644
--- a/tests/d_special_files/expect
+++ b/tests/d_special_files/expect
@@ -7,7 +7,7 @@ Generation: 0    Version: 0x00000000
 User:     0   Group:     0   Size: 3
 File ACL: 0    Directory ACL: 0
 Links: 1   Blockcount: 0
-Fragment:  Address: 0    Number: 0    Size: 0
+Fragment Address: 0    Checksum: 0x00000000
 ctime: 0x50f560e0 -- Tue Jan 15 14:00:00 2013
 atime: 0x50f560e0 -- Tue Jan 15 14:00:00 2013
 mtime: 0x50f560e0 -- Tue Jan 15 14:00:00 2013
@@ -19,7 +19,7 @@ Generation: 0    Version: 0x00000000
 User:     0   Group:     0   Size: 80
 File ACL: 0    Directory ACL: 0
 Links: 1   Blockcount: 2
-Fragment:  Address: 0    Number: 0    Size: 0
+Fragment Address: 0    Checksum: 0x00000000
 ctime: 0x50f560e0 -- Tue Jan 15 14:00:00 2013
 atime: 0x50f560e0 -- Tue Jan 15 14:00:00 2013
 mtime: 0x50f560e0 -- Tue Jan 15 14:00:00 2013
@@ -44,7 +44,7 @@ Generation: 0    Version: 0x00000000
 User:     0   Group:     0   Size: 0
 File ACL: 0    Directory ACL: 0
 Links: 1   Blockcount: 0
-Fragment:  Address: 0    Number: 0    Size: 0
+Fragment Address: 0    Checksum: 0x00000000
 ctime: 0x50f560e0 -- Tue Jan 15 14:00:00 2013
 atime: 0x50f560e0 -- Tue Jan 15 14:00:00 2013
 mtime: 0x50f560e0 -- Tue Jan 15 14:00:00 2013
@@ -57,7 +57,7 @@ Generation: 0    Version: 0x00000000
 User:     0   Group:     0   Size: 0
 File ACL: 0    Directory ACL: 0
 Links: 1   Blockcount: 0
-Fragment:  Address: 0    Number: 0    Size: 0
+Fragment Address: 0    Checksum: 0x00000000
 ctime: 0x50f560e0 -- Tue Jan 15 14:00:00 2013
 atime: 0x50f560e0 -- Tue Jan 15 14:00:00 2013
 mtime: 0x50f560e0 -- Tue Jan 15 14:00:00 2013
@@ -69,7 +69,7 @@ Generation: 0    Version: 0x00000000
 User:     0   Group:     0   Size: 0
 File ACL: 0    Directory ACL: 0
 Links: 1   Blockcount: 0
-Fragment:  Address: 0    Number: 0    Size: 0
+Fragment Address: 0    Checksum: 0x00000000
 ctime: 0x50f560e0 -- Tue Jan 15 14:00:00 2013
 atime: 0x50f560e0 -- Tue Jan 15 14:00:00 2013
 mtime: 0x50f560e0 -- Tue Jan 15 14:00:00 2013
diff --git a/tests/f_badcluster/expect b/tests/f_badcluster/expect
index 6cfffb7..c10a0c7 100644
--- a/tests/f_badcluster/expect
+++ b/tests/f_badcluster/expect
@@ -115,7 +115,7 @@ Generation: 1117152157    Version: 0x00000001
 User:     0   Group:     0   Size: 3072
 File ACL: 0    Directory ACL: 0
 Links: 1   Blockcount: 32
-Fragment:  Address: 0    Number: 0    Size: 0
+Fragment Address: 0    Checksum: 0x00000000
 ctime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
 atime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
 mtime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
@@ -127,7 +127,7 @@ Generation: 1117152158    Version: 0x00000001
 User:     0   Group:     0   Size: 3072
 File ACL: 0    Directory ACL: 0
 Links: 1   Blockcount: 32
-Fragment:  Address: 0    Number: 0    Size: 0
+Fragment Address: 0    Checksum: 0x00000000
 ctime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
 atime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
 mtime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
@@ -139,7 +139,7 @@ Generation: 1117152159    Version: 0x00000001
 User:     0   Group:     0   Size: 3072
 File ACL: 0    Directory ACL: 0
 Links: 1   Blockcount: 32
-Fragment:  Address: 0    Number: 0    Size: 0
+Fragment Address: 0    Checksum: 0x00000000
 ctime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
 atime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
 mtime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
@@ -151,7 +151,7 @@ Generation: 1117152160    Version: 0x00000001
 User:     0   Group:     0   Size: 3072
 File ACL: 0    Directory ACL: 0
 Links: 1   Blockcount: 32
-Fragment:  Address: 0    Number: 0    Size: 0
+Fragment Address: 0    Checksum: 0x00000000
 ctime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
 atime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
 mtime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
@@ -165,7 +165,7 @@ Generation: 1117152161    Version: 0x00000001
 User:     0   Group:     0   Size: 6144
 File ACL: 0    Directory ACL: 0
 Links: 1   Blockcount: 32
-Fragment:  Address: 0    Number: 0    Size: 0
+Fragment Address: 0    Checksum: 0x00000000
 ctime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
 atime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
 mtime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
@@ -177,7 +177,7 @@ Generation: 1117152162    Version: 0x00000001
 User:     0   Group:     0   Size: 3072
 File ACL: 0    Directory ACL: 0
 Links: 1   Blockcount: 32
-Fragment:  Address: 0    Number: 0    Size: 0
+Fragment Address: 0    Checksum: 0x00000000
 ctime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
 atime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
 mtime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
@@ -189,7 +189,7 @@ Generation: 1117152163    Version: 0x00000001
 User:     0   Group:     0   Size: 3072
 File ACL: 0    Directory ACL: 0
 Links: 1   Blockcount: 32
-Fragment:  Address: 0    Number: 0    Size: 0
+Fragment Address: 0    Checksum: 0x00000000
 ctime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
 atime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
 mtime: 0x539ff5b2 -- Tue Jun 17 08:00:50 2014
-- 
1.8.0

Re: [PATCH 6/6] debugfs: update to handle newer inode fields

[Theodore Ts'o]

On Sat, Mar 05, 2016 at 05:38:36PM -0700, Andreas Dilger wrote:
> Remove the printing of fragment fields that are HURD only, and
> instead print the i_checksum_lo and i_checksum_hi fields.
> 
> Allow setting newer fields in extra inode space with modify_inode.
> 
> Add inode_includes() macro to simplify checking for fields in the
> large inode, rather than hard-coding field offsets in the code.
> 
> Fix corresponding tests for newly added Checksum field.
> 
> Signed-off-by: Andreas Dilger <adilger@dilger.ca>

This patch changes way too much at one time, and won't apply given
other changes that have been made by Li Xi and myself.

I wouldn't bother adding stuff to modify_inode; that's an old
interface that is extremely awkward to use, and is there because there
might be a few people who might like it.  These days, set_inode_field
is the new interface that is much more flexible and easier to use.

Li Xi has added inode_includes into lib/ext2fs/ext2_fs.h.  I will
probably rename it to ext2fs_inode_includes() just to avoid namespace
collision problems.

Cheers,

						- Ted

Re: [PATCH 1/6] blkid: fix ZFS device detection

[Theodore Ts'o]

On Sat, Mar 05, 2016 at 05:38:31PM -0700, Andreas Dilger wrote:
> From: Andreas Dilger <andreas.dilger@intel.com>
> 
> While the main blkid functionality is in util-linux, there is still
> use for blkid on non-Linux platforms.
> 
> Fix the ZFS device detection by looking at multiple uberblocks to
> see if any are present, rather than looking for the ZFS boot block
> which is not always present.
> 
> There are slots for up to 128 uberblocks, but the first 4 are not
> written to disk on a newly-formatted filesystem, so check several
> of them at different offsets within the uberblock array.
> 
> Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>

Thanks, applied.

					- Ted