[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-changelog] [xen-unstable] Add ZFS libfsimage support patch



# HG changeset patch
# User Keir Fraser <keir.fraser@xxxxxxxxxx>
# Date 1209656336 -3600
# Node ID 64f790e90d3d4797e298cc5abdfc54cb943687f3
# Parent  8ccaf43782bd43e7d593a0c8e9c7af6f616b86c6
Add ZFS libfsimage support patch

Add support to pygrub and libfsimage to boot ZFS root filesystems.
Boot argument of zfs-bootfs is set to describe ZFS root pool and
boot filesystem object number.  Boot argument bootpath is set to
describe the virtual device root mirror components.

Signed-off-by: Susan Kamm-Worrell <susan.kamm-worrell@xxxxxxx>
---
 tools/libfsimage/Makefile                         |    2 
 tools/libfsimage/common/fsimage.c                 |   29 
 tools/libfsimage/common/fsimage.h                 |    6 
 tools/libfsimage/common/fsimage_grub.c            |    4 
 tools/libfsimage/common/fsimage_grub.h            |    8 
 tools/libfsimage/common/fsimage_priv.h            |    3 
 tools/libfsimage/common/mapfile-GNU               |    3 
 tools/libfsimage/common/mapfile-SunOS             |    3 
 tools/libfsimage/zfs/Makefile                     |   37 
 tools/libfsimage/zfs/fsys_zfs.c                   | 1457 ++++++++++++++++++++++
 tools/libfsimage/zfs/fsys_zfs.h                   |  203 +++
 tools/libfsimage/zfs/mb_info.h                    |  217 +++
 tools/libfsimage/zfs/zfs-include/dmu.h            |  105 +
 tools/libfsimage/zfs/zfs-include/dmu_objset.h     |   35 
 tools/libfsimage/zfs/zfs-include/dnode.h          |   76 +
 tools/libfsimage/zfs/zfs-include/dsl_dataset.h    |   53 
 tools/libfsimage/zfs/zfs-include/dsl_dir.h        |   49 
 tools/libfsimage/zfs/zfs-include/spa.h            |  283 ++++
 tools/libfsimage/zfs/zfs-include/uberblock_impl.h |   49 
 tools/libfsimage/zfs/zfs-include/vdev_impl.h      |   70 +
 tools/libfsimage/zfs/zfs-include/zap_impl.h       |  110 +
 tools/libfsimage/zfs/zfs-include/zap_leaf.h       |  100 +
 tools/libfsimage/zfs/zfs-include/zfs.h            |  112 +
 tools/libfsimage/zfs/zfs-include/zfs_acl.h        |   60 
 tools/libfsimage/zfs/zfs-include/zfs_znode.h      |   68 +
 tools/libfsimage/zfs/zfs-include/zil.h            |   51 
 tools/libfsimage/zfs/zfs-include/zio.h            |   81 +
 tools/libfsimage/zfs/zfs-include/zio_checksum.h   |   42 
 tools/libfsimage/zfs/zfs_fletcher.c               |   93 +
 tools/libfsimage/zfs/zfs_lzjb.c                   |   60 
 tools/libfsimage/zfs/zfs_sha256.c                 |  124 +
 tools/pygrub/src/fsimage/fsimage.c                |   24 
 tools/pygrub/src/pygrub                           |   18 
 33 files changed, 3626 insertions(+), 9 deletions(-)

diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/Makefile
--- a/tools/libfsimage/Makefile Thu May 01 16:37:46 2008 +0100
+++ b/tools/libfsimage/Makefile Thu May 01 16:38:56 2008 +0100
@@ -1,7 +1,7 @@ XEN_ROOT = ../..
 XEN_ROOT = ../..
 include $(XEN_ROOT)/tools/Rules.mk
 
-SUBDIRS-y = common ufs reiserfs iso9660 fat
+SUBDIRS-y = common ufs reiserfs iso9660 fat zfs
 SUBDIRS-y += $(shell env CC="$(CC)" ./check-libext2fs)
 
 .PHONY: all clean install
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/common/fsimage.c
--- a/tools/libfsimage/common/fsimage.c Thu May 01 16:37:46 2008 +0100
+++ b/tools/libfsimage/common/fsimage.c Thu May 01 16:38:56 2008 +0100
@@ -17,7 +17,7 @@
  * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
  * DEALINGS IN THE SOFTWARE.
  *
- * Copyright 2006 Sun Microsystems, Inc.  All rights reserved.
+ * Copyright 2008 Sun Microsystems, Inc.  All rights reserved.
  * Use is subject to license terms.
  */
 
@@ -51,6 +51,7 @@ fsi_t *fsi_open_fsimage(const char *path
        fsi->f_fd = fd;
        fsi->f_off = off;
        fsi->f_data = NULL;
+       fsi->f_bootstring = NULL;
 
        pthread_mutex_lock(&fsi_lock);
        err = find_plugin(fsi, path, options);
@@ -140,3 +141,29 @@ ssize_t fsi_pread_file(fsi_file_t *ffi, 
 
        return (ret);
 }
+
+char *
+fsi_bootstring_alloc(fsi_t *fsi, size_t len)
+{
+       fsi->f_bootstring = malloc(len);
+       if (fsi->f_bootstring == NULL)
+               return (NULL);
+
+       bzero(fsi->f_bootstring, len);
+       return (fsi->f_bootstring);
+}
+
+void
+fsi_bootstring_free(fsi_t *fsi)
+{
+       if (fsi->f_bootstring != NULL) {
+               free(fsi->f_bootstring);
+               fsi->f_bootstring = NULL;
+       }
+}
+
+char *
+fsi_fs_bootstring(fsi_t *fsi)
+{
+       return (fsi->f_bootstring);
+}
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/common/fsimage.h
--- a/tools/libfsimage/common/fsimage.h Thu May 01 16:37:46 2008 +0100
+++ b/tools/libfsimage/common/fsimage.h Thu May 01 16:38:56 2008 +0100
@@ -17,7 +17,7 @@
  * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
  * DEALINGS IN THE SOFTWARE.
  *
- * Copyright 2006 Sun Microsystems, Inc.  All rights reserved.
+ * Copyright 2008 Sun Microsystems, Inc.  All rights reserved.
  * Use is subject to license terms.
  */
 
@@ -45,6 +45,10 @@ ssize_t fsi_read_file(fsi_file_t *, void
 ssize_t fsi_read_file(fsi_file_t *, void *, size_t);
 ssize_t fsi_pread_file(fsi_file_t *, void *, size_t, uint64_t);
 
+char *fsi_bootstring_alloc(fsi_t *, size_t);
+void fsi_bootstring_free(fsi_t *);
+char *fsi_fs_bootstring(fsi_t *);
+
 #ifdef __cplusplus
 };
 #endif
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/common/fsimage_grub.c
--- a/tools/libfsimage/common/fsimage_grub.c    Thu May 01 16:37:46 2008 +0100
+++ b/tools/libfsimage/common/fsimage_grub.c    Thu May 01 16:38:56 2008 +0100
@@ -17,7 +17,7 @@
  * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
  * DEALINGS IN THE SOFTWARE.
  *
- * Copyright 2006 Sun Microsystems, Inc.  All rights reserved.
+ * Copyright 2008 Sun Microsystems, Inc.  All rights reserved.
  * Use is subject to license terms.
  */
 
@@ -286,6 +286,7 @@ fsig_mount(fsi_t *fsi, const char *path,
 
        if (!ops->fpo_mount(ffi, options)) {
                fsip_file_free(ffi);
+               fsi_bootstring_free(fsi);
                free(fsi->f_data);
                fsi->f_data = NULL;
                return (-1);
@@ -299,6 +300,7 @@ static int
 static int
 fsig_umount(fsi_t *fsi)
 {
+       fsi_bootstring_free(fsi);
        free(fsi->f_data);
        return (0);
 }
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/common/fsimage_grub.h
--- a/tools/libfsimage/common/fsimage_grub.h    Thu May 01 16:37:46 2008 +0100
+++ b/tools/libfsimage/common/fsimage_grub.h    Thu May 01 16:38:56 2008 +0100
@@ -17,7 +17,7 @@
  * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
  * DEALINGS IN THE SOFTWARE.
  *
- * Copyright 2006 Sun Microsystems, Inc.  All rights reserved.
+ * Copyright 2008 Sun Microsystems, Inc.  All rights reserved.
  * Use is subject to license terms.
  */
 
@@ -72,6 +72,12 @@ unsigned long fsig_log2(unsigned long);
 #define        ERR_FILELENGTH 1
 #define        ERR_BAD_FILETYPE 1
 #define        ERR_FILE_NOT_FOUND 1
+#define        ERR_BAD_ARGUMENT 1
+#define        ERR_FILESYSTEM_NOT_FOUND 1
+#define        ERR_NO_BOOTPATH 1
+#define        ERR_DEV_VALUES 1
+#define        ERR_WONT_FIT 1
+#define        ERR_READ 1
 
 fsi_plugin_ops_t *fsig_init(fsi_plugin_t *, fsig_plugin_ops_t *);
 
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/common/fsimage_priv.h
--- a/tools/libfsimage/common/fsimage_priv.h    Thu May 01 16:37:46 2008 +0100
+++ b/tools/libfsimage/common/fsimage_priv.h    Thu May 01 16:38:56 2008 +0100
@@ -17,7 +17,7 @@
  * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
  * DEALINGS IN THE SOFTWARE.
  *
- * Copyright 2006 Sun Microsystems, Inc.  All rights reserved.
+ * Copyright 2008 Sun Microsystems, Inc.  All rights reserved.
  * Use is subject to license terms.
  */
 
@@ -46,6 +46,7 @@ struct fsi {
        uint64_t f_off;
        void *f_data;
        fsi_plugin_t *f_plugin;
+       char *f_bootstring;
 };
 
 struct fsi_file {
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/common/mapfile-GNU
--- a/tools/libfsimage/common/mapfile-GNU       Thu May 01 16:37:46 2008 +0100
+++ b/tools/libfsimage/common/mapfile-GNU       Thu May 01 16:38:56 2008 +0100
@@ -8,6 +8,9 @@ VERSION {
                        fsi_close_file;
                        fsi_read_file;
                        fsi_pread_file;
+                       fsi_bootstring_alloc;
+                       fsi_bootstring_free;
+                       fsi_fs_bootstring;
        
                        fsip_fs_set_data;
                        fsip_file_alloc;
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/common/mapfile-SunOS
--- a/tools/libfsimage/common/mapfile-SunOS     Thu May 01 16:37:46 2008 +0100
+++ b/tools/libfsimage/common/mapfile-SunOS     Thu May 01 16:38:56 2008 +0100
@@ -7,6 +7,9 @@ libfsimage.so.1.0 {
                fsi_close_file;
                fsi_read_file;
                fsi_pread_file;
+               fsi_bootstring_alloc;
+               fsi_bootstring_free;
+               fsi_fs_bootstring;
 
                fsip_fs_set_data;
                fsip_file_alloc;
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/zfs/Makefile
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/Makefile     Thu May 01 16:38:56 2008 +0100
@@ -0,0 +1,37 @@
+#
+#  GRUB  --  GRand Unified Bootloader
+#  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+#
+#  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; either version 2 of the License, or
+#  (at your option) any later version.
+#
+#  This program is distributed in the hope that it will 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 to the Free Software
+#  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+#
+
+# 
+#  Copyright 2008 Sun Microsystems, Inc.  All rights reserved.
+#  Use is subject to license terms.
+#
+
+XEN_ROOT = ../../..
+
+LIB_SRCS-y = fsys_zfs.c zfs_lzjb.c zfs_sha256.c zfs_fletcher.c
+
+FS = zfs
+
+.PHONY: all
+all: fs-all
+
+.PHONY: install
+install: fs-install
+
+include $(XEN_ROOT)/tools/libfsimage/Rules.mk
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/zfs/fsys_zfs.c
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/fsys_zfs.c   Thu May 01 16:38:56 2008 +0100
@@ -0,0 +1,1457 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2008 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+/*
+ * All files in the zfs directory are derived from the OpenSolaris
+ * zfs grub files.  All files in the zfs-include directory were
+ * included without changes.
+ */
+
+/*
+ * The zfs plug-in routines for GRUB are:
+ *
+ * zfs_mount() - locates a valid uberblock of the root pool and reads
+ *             in its MOS at the memory address MOS.
+ *
+ * zfs_open() - locates a plain file object by following the MOS
+ *             and places its dnode at the memory address DNODE.
+ *
+ * zfs_read() - read in the data blocks pointed by the DNODE.
+ *
+ * ZFS_SCRATCH is used as a working area.
+ *
+ * (memory addr)   MOS      DNODE      ZFS_SCRATCH
+ *                 |         |          |
+ *         +-------V---------V----------V---------------+
+ *   memory |       | dnode   | dnode    |  scratch      |
+ *         |       | 512B    | 512B     |  area         |
+ *         +--------------------------------------------+
+ */
+
+#include <stdio.h>
+#include <strings.h>
+
+/* From "shared.h" */
+#include "mb_info.h"
+
+/* Boot signature related defines for the findroot command */
+#define        BOOTSIGN_DIR    "/boot/grub/bootsign"
+#define        BOOTSIGN_BACKUP "/etc/bootsign"
+
+/* Maybe redirect memory requests through grub_scratch_mem. */
+#define        RAW_ADDR(x) (x)
+#define        RAW_SEG(x) (x)
+
+/* ZFS will use the top 4 Meg of physical memory (below 4Gig) for sratch */
+#define        ZFS_SCRATCH_SIZE 0x400000
+
+#define        MIN(x, y) ((x) < (y) ? (x) : (y))
+/* End from shared.h */
+
+#include "fsys_zfs.h"
+
+/* cache for a file block of the currently zfs_open()-ed file */
+#define        file_buf zfs_ba->zfs_file_buf
+#define        file_start zfs_ba->zfs_file_start
+#define        file_end zfs_ba->zfs_file_end
+
+/* cache for a dnode block */
+#define        dnode_buf zfs_ba->zfs_dnode_buf
+#define        dnode_mdn zfs_ba->zfs_dnode_mdn
+#define        dnode_start zfs_ba->zfs_dnode_start
+#define        dnode_end zfs_ba->zfs_dnode_end
+
+#define        stackbase zfs_ba->zfs_stackbase
+
+decomp_entry_t decomp_table[ZIO_COMPRESS_FUNCTIONS] =
+{
+       {"noop", 0},
+       {"on", lzjb_decompress},        /* ZIO_COMPRESS_ON */
+       {"off", 0},
+       {"lzjb", lzjb_decompress}       /* ZIO_COMPRESS_LZJB */
+};
+
+/* From disk_io.c */
+/* ZFS root filesystem for booting */
+#define        current_bootpath zfs_ba->zfs_current_bootpath
+#define        current_rootpool zfs_ba->zfs_current_rootpool
+#define        current_bootfs zfs_ba->zfs_current_bootfs
+#define        current_bootfs_obj zfs_ba->zfs_current_bootfs_obj
+#define        is_zfs_mount (*fsig_int1(ffi))
+/* End from disk_io.c */
+
+#define        is_zfs_open zfs_ba->zfs_open
+
+/*
+ * Our own version of bcmp().
+ */
+static int
+zfs_bcmp(const void *s1, const void *s2, size_t n)
+{
+       const unsigned char *ps1 = s1;
+       const unsigned char *ps2 = s2;
+
+       if (s1 != s2 && n != 0) {
+               do {
+                       if (*ps1++ != *ps2++)
+                               return (1);
+               } while (--n != 0);
+       }
+
+       return (0);
+}
+
+/*
+ * Our own version of log2().  Same thing as highbit()-1.
+ */
+static int
+zfs_log2(uint64_t num)
+{
+       int i = 0;
+
+       while (num > 1) {
+               i++;
+               num = num >> 1;
+       }
+
+       return (i);
+}
+
+/* Checksum Functions */
+static void
+zio_checksum_off(const void *buf, uint64_t size, zio_cksum_t *zcp)
+{
+       ZIO_SET_CHECKSUM(zcp, 0, 0, 0, 0);
+}
+
+/* Checksum Table and Values */
+zio_checksum_info_t zio_checksum_table[ZIO_CHECKSUM_FUNCTIONS] = {
+       {{NULL,                 NULL},                  0, 0,   "inherit"},
+       {{NULL,                 NULL},                  0, 0,   "on"},
+       {{zio_checksum_off,     zio_checksum_off},      0, 0,   "off"},
+       {{zio_checksum_SHA256,  zio_checksum_SHA256},   1, 1,   "label"},
+       {{zio_checksum_SHA256,  zio_checksum_SHA256},   1, 1,   "gang_header"},
+       {{fletcher_2_native,    fletcher_2_byteswap},   0, 1,   "zilog"},
+       {{fletcher_2_native,    fletcher_2_byteswap},   0, 0,   "fletcher2"},
+       {{fletcher_4_native,    fletcher_4_byteswap},   1, 0,   "fletcher4"},
+       {{zio_checksum_SHA256,  zio_checksum_SHA256},   1, 0,   "SHA256"}
+};
+
+/*
+ * zio_checksum_verify: Provides support for checksum verification.
+ *
+ * Fletcher2, Fletcher4, and SHA256 are supported.
+ *
+ * Return:
+ *     -1 = Failure
+ *      0 = Success
+ */
+static int
+zio_checksum_verify(blkptr_t *bp, char *data, int size)
+{
+       zio_cksum_t zc = bp->blk_cksum;
+       uint32_t checksum = BP_IS_GANG(bp) ? ZIO_CHECKSUM_GANG_HEADER :
+           BP_GET_CHECKSUM(bp);
+       int byteswap = BP_SHOULD_BYTESWAP(bp);
+       zio_block_tail_t *zbt = (zio_block_tail_t *)(data + size) - 1;
+       zio_checksum_info_t *ci = &zio_checksum_table[checksum];
+       zio_cksum_t actual_cksum, expected_cksum;
+
+       /* byteswap is not supported */
+       if (byteswap)
+               return (-1);
+
+       if (checksum >= ZIO_CHECKSUM_FUNCTIONS || ci->ci_func[0] == NULL)
+               return (-1);
+
+       if (ci->ci_zbt) {
+               if (checksum == ZIO_CHECKSUM_GANG_HEADER) {
+                       /*
+                        * 'gang blocks' is not supported.
+                        */
+                       return (-1);
+               }
+
+               if (zbt->zbt_magic == BSWAP_64(ZBT_MAGIC)) {
+                       /* byte swapping is not supported */
+                       return (-1);
+               } else {
+                       expected_cksum = zbt->zbt_cksum;
+                       zbt->zbt_cksum = zc;
+                       ci->ci_func[0](data, size, &actual_cksum);
+                       zbt->zbt_cksum = expected_cksum;
+               }
+               zc = expected_cksum;
+
+       } else {
+               if (BP_IS_GANG(bp))
+                       return (-1);
+               ci->ci_func[byteswap](data, size, &actual_cksum);
+       }
+
+       if ((actual_cksum.zc_word[0] - zc.zc_word[0]) |
+           (actual_cksum.zc_word[1] - zc.zc_word[1]) |
+           (actual_cksum.zc_word[2] - zc.zc_word[2]) |
+           (actual_cksum.zc_word[3] - zc.zc_word[3]))
+               return (-1);
+
+       return (0);
+}
+
+/*
+ * vdev_label_offset takes "offset" (the offset within a vdev_label) and
+ * returns its physical disk offset (starting from the beginning of the vdev).
+ *
+ * Input:
+ *     psize   : Physical size of this vdev
+ *      l      : Label Number (0-3)
+ *     offset  : The offset with a vdev_label in which we want the physical
+ *               address
+ * Return:
+ *     Success : physical disk offset
+ *     Failure : errnum = ERR_BAD_ARGUMENT, return value is meaningless
+ */
+static uint64_t
+vdev_label_offset(fsi_file_t *ffi, uint64_t psize, int l, uint64_t offset)
+{
+       /* XXX Need to add back label support! */
+       if (l >= VDEV_LABELS/2 || offset > sizeof (vdev_label_t)) {
+               errnum = ERR_BAD_ARGUMENT;
+               return (0);
+       }
+
+       return (offset + l * sizeof (vdev_label_t) + (l < VDEV_LABELS / 2 ?
+           0 : psize - VDEV_LABELS * sizeof (vdev_label_t)));
+
+}
+
+/*
+ * vdev_uberblock_compare takes two uberblock structures and returns an integer
+ * indicating the more recent of the two.
+ *     Return Value = 1 if ub2 is more recent
+ *     Return Value = -1 if ub1 is more recent
+ * The most recent uberblock is determined using its transaction number and
+ * timestamp.  The uberblock with the highest transaction number is
+ * considered "newer".  If the transaction numbers of the two blocks match, the
+ * timestamps are compared to determine the "newer" of the two.
+ */
+static int
+vdev_uberblock_compare(uberblock_t *ub1, uberblock_t *ub2)
+{
+       if (ub1->ub_txg < ub2->ub_txg)
+               return (-1);
+       if (ub1->ub_txg > ub2->ub_txg)
+               return (1);
+
+       if (ub1->ub_timestamp < ub2->ub_timestamp)
+               return (-1);
+       if (ub1->ub_timestamp > ub2->ub_timestamp)
+               return (1);
+
+       return (0);
+}
+
+/*
+ * Three pieces of information are needed to verify an uberblock: the magic
+ * number, the version number, and the checksum.
+ *
+ * Currently Implemented: version number, magic number
+ * Need to Implement: checksum
+ *
+ * Return:
+ *     0 - Success
+ *    -1 - Failure
+ */
+static int
+uberblock_verify(uberblock_phys_t *ub, int offset)
+{
+
+       uberblock_t *uber = &ub->ubp_uberblock;
+       blkptr_t bp;
+
+       BP_ZERO(&bp);
+       BP_SET_CHECKSUM(&bp, ZIO_CHECKSUM_LABEL);
+       BP_SET_BYTEORDER(&bp, ZFS_HOST_BYTEORDER);
+       ZIO_SET_CHECKSUM(&bp.blk_cksum, offset, 0, 0, 0);
+
+       if (zio_checksum_verify(&bp, (char *)ub, UBERBLOCK_SIZE) != 0)
+               return (-1);
+
+       if (uber->ub_magic == UBERBLOCK_MAGIC &&
+           uber->ub_version >= SPA_VERSION_1 &&
+           uber->ub_version <= SPA_VERSION)
+               return (0);
+
+       return (-1);
+}
+
+/*
+ * Find the best uberblock.
+ * Return:
+ *    Success - Pointer to the best uberblock.
+ *    Failure - NULL
+ */
+static uberblock_phys_t *
+find_bestub(fsi_file_t *ffi, uberblock_phys_t *ub_array, int label)
+{
+       uberblock_phys_t *ubbest = NULL;
+       int i, offset;
+
+       for (i = 0; i < (VDEV_UBERBLOCK_RING >> VDEV_UBERBLOCK_SHIFT); i++) {
+               offset = vdev_label_offset(ffi, 0, label,
+                   VDEV_UBERBLOCK_OFFSET(i));
+               if (errnum == ERR_BAD_ARGUMENT)
+                       return (NULL);
+               if (uberblock_verify(&ub_array[i], offset) == 0) {
+                       if (ubbest == NULL) {
+                               ubbest = &ub_array[i];
+                       } else if (vdev_uberblock_compare(
+                           &(ub_array[i].ubp_uberblock),
+                           &(ubbest->ubp_uberblock)) > 0) {
+                               ubbest = &ub_array[i];
+                       }
+               }
+       }
+
+       return (ubbest);
+}
+
+/*
+ * Read in a block and put its uncompressed data in buf.
+ *
+ * Return:
+ *     0 - success
+ *     errnum - failure
+ */
+static int
+zio_read(fsi_file_t *ffi, blkptr_t *bp, void *buf, char *stack)
+{
+       uint64_t offset, sector;
+       int psize, lsize;
+       int i, comp, cksum;
+
+       psize = BP_GET_PSIZE(bp);
+       lsize = BP_GET_LSIZE(bp);
+       comp = BP_GET_COMPRESS(bp);
+       cksum = BP_GET_CHECKSUM(bp);
+
+       if ((unsigned int)comp >= ZIO_COMPRESS_FUNCTIONS ||
+           (comp != ZIO_COMPRESS_OFF &&
+           decomp_table[comp].decomp_func == NULL))
+               return (ERR_FSYS_CORRUPT);
+
+       /* pick a good dva from the block pointer */
+       for (i = 0; i < SPA_DVAS_PER_BP; i++) {
+
+               if (bp->blk_dva[i].dva_word[0] == 0 &&
+                   bp->blk_dva[i].dva_word[1] == 0)
+                       continue;
+
+               /* read in a block */
+               offset = DVA_GET_OFFSET(&bp->blk_dva[i]);
+               sector =  DVA_OFFSET_TO_PHYS_SECTOR(offset);
+
+               if (comp != ZIO_COMPRESS_OFF) {
+
+                       if (devread(ffi, sector, 0, psize, stack) == 0)
+                               continue;
+                       if (zio_checksum_verify(bp, stack, psize) != 0)
+                               continue;
+                       decomp_table[comp].decomp_func(stack, buf, psize,
+                           lsize);
+               } else {
+                       if (devread(ffi, sector, 0, psize, buf) == 0)
+                               continue;
+                       if (zio_checksum_verify(bp, buf, psize) != 0)
+                               continue;
+               }
+               return (0);
+       }
+
+       return (ERR_FSYS_CORRUPT);
+}
+
+/*
+ * Get the block from a block id.
+ * push the block onto the stack.
+ *
+ * Return:
+ *     0 - success
+ *     errnum - failure
+ */
+static int
+dmu_read(fsi_file_t *ffi, dnode_phys_t *dn, uint64_t blkid, void *buf,
+    char *stack)
+{
+       int idx, level;
+       blkptr_t *bp_array = dn->dn_blkptr;
+       int epbs = dn->dn_indblkshift - SPA_BLKPTRSHIFT;
+       blkptr_t *bp, *tmpbuf;
+
+       bp = (blkptr_t *)stack;
+       stack += sizeof (blkptr_t);
+
+       tmpbuf = (blkptr_t *)stack;
+       stack += 1<<dn->dn_indblkshift;
+
+       for (level = dn->dn_nlevels - 1; level >= 0; level--) {
+               idx = (blkid >> (epbs * level)) & ((1<<epbs)-1);
+               *bp = bp_array[idx];
+               if (level == 0)
+                       tmpbuf = buf;
+               if (BP_IS_HOLE(bp)) {
+                       grub_memset(buf, 0,
+                           dn->dn_datablkszsec << SPA_MINBLOCKSHIFT);
+                       break;
+               } else if ((errnum = zio_read(ffi, bp, tmpbuf, stack))) {
+                       return (errnum);
+               }
+               bp_array = tmpbuf;
+       }
+
+       return (0);
+}
+
+/*
+ * mzap_lookup: Looks up property described by "name" and returns the value
+ * in "value".
+ *
+ * Return:
+ *     0 - success
+ *     errnum - failure
+ */
+static int
+mzap_lookup(mzap_phys_t *zapobj, int objsize, char *name,
+       uint64_t *value)
+{
+       int i, chunks;
+       mzap_ent_phys_t *mzap_ent = zapobj->mz_chunk;
+
+       chunks = objsize/MZAP_ENT_LEN - 1;
+       for (i = 0; i < chunks; i++) {
+               if (strcmp(mzap_ent[i].mze_name, name) == 0) {
+                       *value = mzap_ent[i].mze_value;
+                       return (0);
+               }
+       }
+
+       return (ERR_FSYS_CORRUPT);
+}
+
+static uint64_t
+zap_hash(fsi_file_t *ffi, uint64_t salt, const char *name)
+{
+       static uint64_t table[256];
+       const uint8_t *cp;
+       uint8_t c;
+       uint64_t crc = salt;
+
+       if (table[128] == 0) {
+               uint64_t *ct;
+               int i, j;
+               for (i = 0; i < 256; i++) {
+                       for (ct = table + i, *ct = i, j = 8; j > 0; j--)
+                               *ct = (*ct >> 1) ^ (-(*ct & 1) &
+                                   ZFS_CRC64_POLY);
+               }
+       }
+
+       if (crc == 0 || table[128] != ZFS_CRC64_POLY) {
+               errnum = ERR_FSYS_CORRUPT;
+               return (0);
+       }
+
+       for (cp = (const uint8_t *)name; (c = *cp) != '\0'; cp++)
+               crc = (crc >> 8) ^ table[(crc ^ c) & 0xFF];
+
+       /*
+        * Only use 28 bits, since we need 4 bits in the cookie for the
+        * collision differentiator.  We MUST use the high bits, since
+        * those are the onces that we first pay attention to when
+        * chosing the bucket.
+        */
+       crc &= ~((1ULL << (64 - ZAP_HASHBITS)) - 1);
+
+       return (crc);
+}
+
+/*
+ * Only to be used on 8-bit arrays.
+ * array_len is actual len in bytes (not encoded le_value_length).
+ * buf is null-terminated.
+ */
+static int
+zap_leaf_array_equal(zap_leaf_phys_t *l, int blksft, int chunk,
+    int array_len, const char *buf)
+{
+       int bseen = 0;
+
+       while (bseen < array_len) {
+               struct zap_leaf_array *la =
+                   &ZAP_LEAF_CHUNK(l, blksft, chunk).l_array;
+               int toread = MIN(array_len - bseen, ZAP_LEAF_ARRAY_BYTES);
+
+               if (chunk >= ZAP_LEAF_NUMCHUNKS(blksft))
+                       return (0);
+
+               if (zfs_bcmp(la->la_array, buf + bseen, toread) != 0)
+                       break;
+               chunk = la->la_next;
+               bseen += toread;
+       }
+       return (bseen == array_len);
+}
+
+/*
+ * Given a zap_leaf_phys_t, walk thru the zap leaf chunks to get the
+ * value for the property "name".
+ *
+ * Return:
+ *     0 - success
+ *     errnum - failure
+ */
+static int
+zap_leaf_lookup(zap_leaf_phys_t *l, int blksft, uint64_t h,
+    const char *name, uint64_t *value)
+{
+       uint16_t chunk;
+       struct zap_leaf_entry *le;
+
+       /* Verify if this is a valid leaf block */
+       if (l->l_hdr.lh_block_type != ZBT_LEAF)
+               return (ERR_FSYS_CORRUPT);
+       if (l->l_hdr.lh_magic != ZAP_LEAF_MAGIC)
+               return (ERR_FSYS_CORRUPT);
+
+       for (chunk = l->l_hash[LEAF_HASH(blksft, h)];
+           chunk != CHAIN_END; chunk = le->le_next) {
+
+               if (chunk >= ZAP_LEAF_NUMCHUNKS(blksft))
+                       return (ERR_FSYS_CORRUPT);
+
+               le = ZAP_LEAF_ENTRY(l, blksft, chunk);
+
+               /* Verify the chunk entry */
+               if (le->le_type != ZAP_CHUNK_ENTRY)
+                       return (ERR_FSYS_CORRUPT);
+
+               if (le->le_hash != h)
+                       continue;
+
+               if (zap_leaf_array_equal(l, blksft, le->le_name_chunk,
+                   le->le_name_length, name)) {
+
+                       struct zap_leaf_array *la;
+                       uint8_t *ip;
+
+                       if (le->le_int_size != 8 || le->le_value_length != 1)
+                               return (ERR_FSYS_CORRUPT);
+
+                       /* get the uint64_t property value */
+                       la = &ZAP_LEAF_CHUNK(l, blksft,
+                           le->le_value_chunk).l_array;
+                       ip = la->la_array;
+
+                       *value = (uint64_t)ip[0] << 56 | (uint64_t)ip[1] << 48 |
+                           (uint64_t)ip[2] << 40 | (uint64_t)ip[3] << 32 |
+                           (uint64_t)ip[4] << 24 | (uint64_t)ip[5] << 16 |
+                           (uint64_t)ip[6] << 8 | (uint64_t)ip[7];
+
+                       return (0);
+               }
+       }
+
+       return (ERR_FSYS_CORRUPT);
+}
+
+/*
+ * Fat ZAP lookup
+ *
+ * Return:
+ *     0 - success
+ *     errnum - failure
+ */
+static int
+fzap_lookup(fsi_file_t *ffi, dnode_phys_t *zap_dnode, zap_phys_t *zap,
+    char *name, uint64_t *value, char *stack)
+{
+       zap_leaf_phys_t *l;
+       uint64_t hash, idx, blkid;
+       int blksft = zfs_log2(zap_dnode->dn_datablkszsec << DNODE_SHIFT);
+
+       /* Verify if this is a fat zap header block */
+       if (zap->zap_magic != (uint64_t)ZAP_MAGIC)
+               return (ERR_FSYS_CORRUPT);
+
+       hash = zap_hash(ffi, zap->zap_salt, name);
+       if (errnum)
+               return (errnum);
+
+       /* get block id from index */
+       if (zap->zap_ptrtbl.zt_numblks != 0) {
+               /* external pointer tables not supported */
+               return (ERR_FSYS_CORRUPT);
+       }
+       idx = ZAP_HASH_IDX(hash, zap->zap_ptrtbl.zt_shift);
+       blkid = ((uint64_t *)zap)[idx + (1<<(blksft-3-1))];
+
+       /* Get the leaf block */
+       l = (zap_leaf_phys_t *)stack;
+       stack += 1<<blksft;
+       if ((errnum = dmu_read(ffi, zap_dnode, blkid, l, stack)))
+               return (errnum);
+
+       return (zap_leaf_lookup(l, blksft, hash, name, value));
+}
+
+/*
+ * Read in the data of a zap object and find the value for a matching
+ * property name.
+ *
+ * Return:
+ *     0 - success
+ *     errnum - failure
+ */
+static int
+zap_lookup(fsi_file_t *ffi, dnode_phys_t *zap_dnode, char *name,
+    uint64_t *val, char *stack)
+{
+       uint64_t block_type;
+       int size;
+       void *zapbuf;
+
+       /* Read in the first block of the zap object data. */
+       zapbuf = stack;
+       size = zap_dnode->dn_datablkszsec << SPA_MINBLOCKSHIFT;
+       stack += size;
+       if ((errnum = dmu_read(ffi, zap_dnode, 0, zapbuf, stack)))
+               return (errnum);
+
+       block_type = *((uint64_t *)zapbuf);
+
+       if (block_type == ZBT_MICRO) {
+               return (mzap_lookup(zapbuf, size, name, val));
+       } else if (block_type == ZBT_HEADER) {
+               /* this is a fat zap */
+               return (fzap_lookup(ffi, zap_dnode, zapbuf, name,
+                   val, stack));
+       }
+
+       return (ERR_FSYS_CORRUPT);
+}
+
+/*
+ * Get the dnode of an object number from the metadnode of an object set.
+ *
+ * Input
+ *     mdn - metadnode to get the object dnode
+ *     objnum - object number for the object dnode
+ *     buf - data buffer that holds the returning dnode
+ *     stack - scratch area
+ *
+ * Return:
+ *     0 - success
+ *     errnum - failure
+ */
+static int
+dnode_get(fsi_file_t *ffi, dnode_phys_t *mdn, uint64_t objnum,
+    uint8_t type, dnode_phys_t *buf, char *stack)
+{
+       uint64_t blkid, blksz; /* the block id this object dnode is in */
+       int epbs; /* shift of number of dnodes in a block */
+       int idx; /* index within a block */
+       dnode_phys_t *dnbuf;
+       zfs_bootarea_t *zfs_ba = (zfs_bootarea_t *)ffi->ff_fsi->f_data;
+
+       blksz = mdn->dn_datablkszsec << SPA_MINBLOCKSHIFT;
+       epbs = zfs_log2(blksz) - DNODE_SHIFT;
+       blkid = objnum >> epbs;
+       idx = objnum & ((1<<epbs)-1);
+
+       if (dnode_buf != NULL && dnode_mdn == mdn &&
+           objnum >= dnode_start && objnum < dnode_end) {
+               grub_memmove(buf, &dnode_buf[idx], DNODE_SIZE);
+               VERIFY_DN_TYPE(buf, type);
+               return (0);
+       }
+
+       if (dnode_buf && blksz == 1<<DNODE_BLOCK_SHIFT) {
+               dnbuf = dnode_buf;
+               dnode_mdn = mdn;
+               dnode_start = blkid << epbs;
+               dnode_end = (blkid + 1) << epbs;
+       } else {
+               dnbuf = (dnode_phys_t *)stack;
+               stack += blksz;
+       }
+
+       if ((errnum = dmu_read(ffi, mdn, blkid, (char *)dnbuf, stack)))
+               return (errnum);
+
+       grub_memmove(buf, &dnbuf[idx], DNODE_SIZE);
+       VERIFY_DN_TYPE(buf, type);
+
+       return (0);
+}
+
+/*
+ * Check if this is a special file that resides at the top
+ * dataset of the pool. Currently this is the GRUB menu,
+ * boot signature and boot signature backup.
+ * str starts with '/'.
+ */
+static int
+is_top_dataset_file(char *str)
+{
+       char *tptr;
+
+       if (((tptr = strstr(str, "menu.lst"))) &&
+           (tptr[8] == '\0' || tptr[8] == ' ') &&
+           *(tptr-1) == '/')
+               return (1);
+
+       if (strncmp(str, BOOTSIGN_DIR"/",
+           strlen(BOOTSIGN_DIR) + 1) == 0)
+               return (1);
+
+       if (strcmp(str, BOOTSIGN_BACKUP) == 0)
+               return (1);
+
+       return (0);
+}
+
+/*
+ * Get the file dnode for a given file name where mdn is the meta dnode
+ * for this ZFS object set. When found, place the file dnode in dn.
+ * The 'path' argument will be mangled.
+ *
+ * Return:
+ *     0 - success
+ *     errnum - failure
+ */
+static int
+dnode_get_path(fsi_file_t *ffi, dnode_phys_t *mdn, char *path,
+    dnode_phys_t *dn, char *stack)
+{
+       uint64_t objnum, version;
+       char *cname, ch;
+
+       if ((errnum = dnode_get(ffi, mdn, MASTER_NODE_OBJ, DMU_OT_MASTER_NODE,
+           dn, stack)))
+               return (errnum);
+
+       if ((errnum = zap_lookup(ffi, dn, ZPL_VERSION_STR, &version, stack)))
+               return (errnum);
+       if (version > ZPL_VERSION)
+               return (-1);
+
+       if ((errnum = zap_lookup(ffi, dn, ZFS_ROOT_OBJ, &objnum, stack)))
+               return (errnum);
+
+       if ((errnum = dnode_get(ffi, mdn, objnum, DMU_OT_DIRECTORY_CONTENTS,
+           dn, stack)))
+               return (errnum);
+
+       /* skip leading slashes */
+       while (*path == '/')
+               path++;
+
+       while (*path && !isspace(*path)) {
+
+               /* get the next component name */
+               cname = path;
+               while (*path && !isspace(*path) && *path != '/')
+                       path++;
+               ch = *path;
+               *path = 0;   /* ensure null termination */
+
+               if ((errnum = zap_lookup(ffi, dn, cname, &objnum, stack)))
+                       return (errnum);
+
+               objnum = ZFS_DIRENT_OBJ(objnum);
+               if ((errnum = dnode_get(ffi, mdn, objnum, 0, dn, stack)))
+                       return (errnum);
+
+               *path = ch;
+               while (*path == '/')
+                       path++;
+       }
+
+       /* We found the dnode for this file. Verify if it is a plain file. */
+       VERIFY_DN_TYPE(dn, DMU_OT_PLAIN_FILE_CONTENTS);
+
+       return (0);
+}
+
+/*
+ * Get the default 'bootfs' property value from the rootpool.
+ *
+ * Return:
+ *     0 - success
+ *     errnum -failure
+ */
+static int
+get_default_bootfsobj(fsi_file_t *ffi, dnode_phys_t *mosmdn,
+    uint64_t *obj, char *stack)
+{
+       uint64_t objnum = 0;
+       dnode_phys_t *dn = (dnode_phys_t *)stack;
+       stack += DNODE_SIZE;
+
+       if ((errnum = dnode_get(ffi, mosmdn, DMU_POOL_DIRECTORY_OBJECT,
+           DMU_OT_OBJECT_DIRECTORY, dn, stack)))
+               return (errnum);
+
+       /*
+        * find the object number for 'pool_props', and get the dnode
+        * of the 'pool_props'.
+        */
+       if (zap_lookup(ffi, dn, DMU_POOL_PROPS, &objnum, stack))
+               return (ERR_FILESYSTEM_NOT_FOUND);
+
+       if ((errnum = dnode_get(ffi, mosmdn, objnum, DMU_OT_POOL_PROPS, dn,
+           stack)))
+               return (errnum);
+
+       if (zap_lookup(ffi, dn, ZPOOL_PROP_BOOTFS, &objnum, stack))
+               return (ERR_FILESYSTEM_NOT_FOUND);
+
+       if (!objnum)
+               return (ERR_FILESYSTEM_NOT_FOUND);
+
+
+       *obj = objnum;
+       return (0);
+}
+
+/*
+ * Given a MOS metadnode, get the metadnode of a given filesystem name 
(fsname),
+ * e.g. pool/rootfs, or a given object number (obj), e.g. the object number
+ * of pool/rootfs.
+ *
+ * If no fsname and no obj are given, return the DSL_DIR metadnode.
+ * If fsname is given, return its metadnode and its matching object number.
+ * If only obj is given, return the metadnode for this object number.
+ *
+ * Return:
+ *     0 - success
+ *     errnum - failure
+ */
+static int
+get_objset_mdn(fsi_file_t *ffi, dnode_phys_t *mosmdn, char *fsname,
+    uint64_t *obj, dnode_phys_t *mdn, char *stack)
+{
+       uint64_t objnum, headobj;
+       char *cname, ch;
+       blkptr_t *bp;
+       objset_phys_t *osp;
+
+       if (fsname == NULL && obj) {
+               headobj = *obj;
+               goto skip;
+       }
+
+       if ((errnum = dnode_get(ffi, mosmdn, DMU_POOL_DIRECTORY_OBJECT,
+           DMU_OT_OBJECT_DIRECTORY, mdn, stack)))
+               return (errnum);
+
+       if ((errnum = zap_lookup(ffi, mdn, DMU_POOL_ROOT_DATASET, &objnum,
+           stack)))
+               return (errnum);
+
+       if ((errnum = dnode_get(ffi, mosmdn, objnum, DMU_OT_DSL_DIR, mdn,
+           stack)))
+               return (errnum);
+
+       if (fsname == NULL) {
+               headobj =
+                   ((dsl_dir_phys_t *)DN_BONUS(mdn))->dd_head_dataset_obj;
+               goto skip;
+       }
+
+       /* take out the pool name */
+       while (*fsname && !isspace(*fsname) && *fsname != '/')
+               fsname++;
+
+       while (*fsname && !isspace(*fsname)) {
+               uint64_t childobj;
+
+               while (*fsname == '/')
+                       fsname++;
+
+               cname = fsname;
+               while (*fsname && !isspace(*fsname) && *fsname != '/')
+                       fsname++;
+               ch = *fsname;
+               *fsname = 0;
+
+               childobj =
+                   ((dsl_dir_phys_t *)DN_BONUS(mdn))->dd_child_dir_zapobj;
+               if ((errnum = dnode_get(ffi, mosmdn, childobj,
+                   DMU_OT_DSL_DIR_CHILD_MAP, mdn, stack)))
+                       return (errnum);
+
+               if (zap_lookup(ffi, mdn, cname, &objnum, stack))
+                       return (ERR_FILESYSTEM_NOT_FOUND);
+
+               if ((errnum = dnode_get(ffi, mosmdn, objnum, DMU_OT_DSL_DIR,
+                   mdn, stack)))
+                       return (errnum);
+
+               *fsname = ch;
+       }
+       headobj = ((dsl_dir_phys_t *)DN_BONUS(mdn))->dd_head_dataset_obj;
+       if (obj)
+               *obj = headobj;
+
+skip:
+       if ((errnum = dnode_get(ffi, mosmdn, headobj, DMU_OT_DSL_DATASET, mdn,
+           stack)))
+               return (errnum);
+
+       /* TODO: Add snapshot support here - for fsname=snapshot-name */
+
+       bp = &((dsl_dataset_phys_t *)DN_BONUS(mdn))->ds_bp;
+       osp = (objset_phys_t *)stack;
+       stack += sizeof (objset_phys_t);
+       if ((errnum = zio_read(ffi, bp, osp, stack)))
+               return (errnum);
+
+       grub_memmove((char *)mdn, (char *)&osp->os_meta_dnode, DNODE_SIZE);
+
+       return (0);
+}
+
+/*
+ * For a given XDR packed nvlist, verify the first 4 bytes and move on.
+ *
+ * An XDR packed nvlist is encoded as (comments from nvs_xdr_create) :
+ *
+ *      encoding method/host endian     (4 bytes)
+ *      nvl_version                     (4 bytes)
+ *      nvl_nvflag                      (4 bytes)
+ *     encoded nvpairs:
+ *             encoded size of the nvpair      (4 bytes)
+ *             decoded size of the nvpair      (4 bytes)
+ *             name string size                (4 bytes)
+ *             name string data                (sizeof(NV_ALIGN4(string))
+ *             data type                       (4 bytes)
+ *             # of elements in the nvpair     (4 bytes)
+ *             data
+ *      2 zero's for the last nvpair
+ *             (end of the entire list)        (8 bytes)
+ *
+ * Return:
+ *     0 - success
+ *     1 - failure
+ */
+static int
+nvlist_unpack(char *nvlist, char **out)
+{
+       /* Verify if the 1st and 2nd byte in the nvlist are valid. */
+       if (nvlist[0] != NV_ENCODE_XDR || nvlist[1] != HOST_ENDIAN)
+               return (1);
+
+       nvlist += 4;
+       *out = nvlist;
+       return (0);
+}
+
+static char *
+nvlist_array(char *nvlist, int index)
+{
+       int i, encode_size;
+
+       for (i = 0; i < index; i++) {
+               /* skip the header, nvl_version, and nvl_nvflag */
+               nvlist = nvlist + 4 * 2;
+
+               while ((encode_size = BSWAP_32(*(uint32_t *)nvlist)))
+                       nvlist += encode_size; /* goto the next nvpair */
+
+               nvlist = nvlist + 4 * 2; /* skip the ending 2 zeros - 8 bytes */
+       }
+
+       return (nvlist);
+}
+
+static int
+nvlist_lookup_value(char *nvlist, char *name, void *val, int valtype,
+    int *nelmp)
+{
+       int name_len, type, slen, encode_size;
+       char *nvpair, *nvp_name, *strval = val;
+       uint64_t *intval = val;
+
+       /* skip the header, nvl_version, and nvl_nvflag */
+       nvlist = nvlist + 4 * 2;
+
+       /*
+        * Loop thru the nvpair list
+        * The XDR representation of an integer is in big-endian byte order.
+        */
+       while ((encode_size = BSWAP_32(*(uint32_t *)nvlist)))  {
+
+               nvpair = nvlist + 4 * 2; /* skip the encode/decode size */
+
+               name_len = BSWAP_32(*(uint32_t *)nvpair);
+               nvpair += 4;
+
+               nvp_name = nvpair;
+               nvpair = nvpair + ((name_len + 3) & ~3); /* align */
+
+               type = BSWAP_32(*(uint32_t *)nvpair);
+               nvpair += 4;
+
+               if (((strncmp(nvp_name, name, name_len) == 0) &&
+                   type == valtype)) {
+                       int nelm;
+
+                       if (((nelm = BSWAP_32(*(uint32_t *)nvpair)) < 1))
+                               return (1);
+                       nvpair += 4;
+
+                       switch (valtype) {
+                       case DATA_TYPE_STRING:
+                               slen = BSWAP_32(*(uint32_t *)nvpair);
+                               nvpair += 4;
+                               grub_memmove(strval, nvpair, slen);
+                               strval[slen] = '\0';
+                               return (0);
+
+                       case DATA_TYPE_UINT64:
+                               *intval = BSWAP_64(*(uint64_t *)nvpair);
+                               return (0);
+
+                       case DATA_TYPE_NVLIST:
+                               *(void **)val = (void *)nvpair;
+                               return (0);
+
+                       case DATA_TYPE_NVLIST_ARRAY:
+                               *(void **)val = (void *)nvpair;
+                               if (nelmp)
+                                       *nelmp = nelm;
+                               return (0);
+                       }
+               }
+
+               nvlist += encode_size; /* goto the next nvpair */
+       }
+
+       return (1);
+}
+
+/*
+ * Check if this vdev is online and is in a good state.
+ */
+static int
+vdev_validate(char *nv)
+{
+       uint64_t ival;
+
+       if (nvlist_lookup_value(nv, ZPOOL_CONFIG_OFFLINE, &ival,
+           DATA_TYPE_UINT64, NULL) == 0 ||
+           nvlist_lookup_value(nv, ZPOOL_CONFIG_FAULTED, &ival,
+           DATA_TYPE_UINT64, NULL) == 0 ||
+           nvlist_lookup_value(nv, ZPOOL_CONFIG_DEGRADED, &ival,
+           DATA_TYPE_UINT64, NULL) == 0 ||
+           nvlist_lookup_value(nv, ZPOOL_CONFIG_REMOVED, &ival,
+           DATA_TYPE_UINT64, NULL) == 0)
+               return (ERR_DEV_VALUES);
+
+       return (0);
+}
+
+/*
+ * Get a list of valid vdev pathname from the boot device.
+ * The caller should already allocate MAXNAMELEN memory for bootpath.
+ */
+static int
+vdev_get_bootpath(char *nv, char *bootpath)
+{
+       char type[16];
+
+       bootpath[0] = '\0';
+       if (nvlist_lookup_value(nv, ZPOOL_CONFIG_TYPE, &type, DATA_TYPE_STRING,
+           NULL))
+               return (ERR_FSYS_CORRUPT);
+
+       if (strcmp(type, VDEV_TYPE_DISK) == 0) {
+               if (vdev_validate(nv) != 0 ||
+                   nvlist_lookup_value(nv, ZPOOL_CONFIG_PHYS_PATH, bootpath,
+                   DATA_TYPE_STRING, NULL) != 0)
+                       return (ERR_NO_BOOTPATH);
+
+       } else if (strcmp(type, VDEV_TYPE_MIRROR) == 0) {
+               int nelm, i;
+               char *child;
+
+               if (nvlist_lookup_value(nv, ZPOOL_CONFIG_CHILDREN, &child,
+                   DATA_TYPE_NVLIST_ARRAY, &nelm))
+                       return (ERR_FSYS_CORRUPT);
+
+               for (i = 0; i < nelm; i++) {
+                       char tmp_path[MAXNAMELEN];
+                       char *child_i;
+
+                       child_i = nvlist_array(child, i);
+                       if (vdev_validate(child_i) != 0)
+                               continue;
+
+                       if (nvlist_lookup_value(child_i, ZPOOL_CONFIG_PHYS_PATH,
+                           tmp_path, DATA_TYPE_STRING, NULL) != 0)
+                               return (ERR_NO_BOOTPATH);
+
+                       if ((strlen(bootpath) + strlen(tmp_path)) > MAXNAMELEN)
+                               return (ERR_WONT_FIT);
+
+                       if (strlen(bootpath) == 0)
+                               sprintf(bootpath, "%s", tmp_path);
+                       else
+                               sprintf(bootpath, "%s %s", bootpath, tmp_path);
+               }
+       }
+
+       return (strlen(bootpath) > 0 ? 0 : ERR_NO_BOOTPATH);
+}
+
+/*
+ * Check the disk label information and retrieve needed vdev name-value pairs.
+ *
+ * Return:
+ *     0 - success
+ *     ERR_* - failure
+ */
+static int
+check_pool_label(fsi_file_t *ffi, int label, char *stack)
+{
+       vdev_phys_t *vdev;
+       uint64_t sector, pool_state, txg = 0;
+       char *nvlist, *nv;
+       zfs_bootarea_t *zfs_ba = (zfs_bootarea_t *)ffi->ff_fsi->f_data;
+
+       sector = (label * sizeof (vdev_label_t) + VDEV_SKIP_SIZE +
+           VDEV_BOOT_HEADER_SIZE) >> SPA_MINBLOCKSHIFT;
+
+       /* Read in the vdev name-value pair list (112K). */
+       if (devread(ffi, sector, 0, VDEV_PHYS_SIZE, stack) == 0)
+               return (ERR_READ);
+
+       vdev = (vdev_phys_t *)stack;
+
+       if (nvlist_unpack(vdev->vp_nvlist, &nvlist))
+               return (ERR_FSYS_CORRUPT);
+
+       if (nvlist_lookup_value(nvlist, ZPOOL_CONFIG_POOL_STATE, &pool_state,
+           DATA_TYPE_UINT64, NULL))
+               return (ERR_FSYS_CORRUPT);
+
+       if (pool_state == POOL_STATE_DESTROYED)
+               return (ERR_FILESYSTEM_NOT_FOUND);
+
+       if (nvlist_lookup_value(nvlist, ZPOOL_CONFIG_POOL_NAME,
+           current_rootpool, DATA_TYPE_STRING, NULL))
+               return (ERR_FSYS_CORRUPT);
+
+       if (nvlist_lookup_value(nvlist, ZPOOL_CONFIG_POOL_TXG, &txg,
+           DATA_TYPE_UINT64, NULL))
+               return (ERR_FSYS_CORRUPT);
+
+       /* not an active device */
+       if (txg == 0)
+               return (ERR_NO_BOOTPATH);
+
+       if (nvlist_lookup_value(nvlist, ZPOOL_CONFIG_VDEV_TREE, &nv,
+           DATA_TYPE_NVLIST, NULL))
+               return (ERR_FSYS_CORRUPT);
+
+       if (vdev_get_bootpath(nv, current_bootpath))
+               return (ERR_NO_BOOTPATH);
+
+       return (0);
+}
+
+/*
+ * zfs_mount() locates a valid uberblock of the root pool and read in its MOS
+ * to the memory address MOS.
+ *
+ * Return:
+ *     1 - success
+ *     0 - failure
+ */
+int
+zfs_mount(fsi_file_t *ffi, const char *options)
+{
+       char *stack;
+       int label = 0;
+       uberblock_phys_t *ub_array, *ubbest = NULL;
+       objset_phys_t *osp;
+       zfs_bootarea_t *zfs_ba;
+
+       /* if zfs is already mounted, don't do it again */
+       if (is_zfs_mount == 1)
+               return (1);
+
+       /* get much bigger data block for zfs */
+       if (((zfs_ba = malloc(sizeof (zfs_bootarea_t))) == NULL)) {
+               return (1);
+       }
+       bzero(zfs_ba, sizeof (zfs_bootarea_t));
+
+       /* replace small data area in fsi with big one */
+       free(ffi->ff_fsi->f_data);
+       ffi->ff_fsi->f_data = (void *)zfs_ba;
+
+       /* If an boot filesystem is passed in, set it to current_bootfs */
+       if (options != NULL) {
+               if (strlen(options) < MAXNAMELEN) {
+                       strcpy(current_bootfs, options);
+               }
+       }
+
+       stackbase = ZFS_SCRATCH;
+       stack = stackbase;
+       ub_array = (uberblock_phys_t *)stack;
+       stack += VDEV_UBERBLOCK_RING;
+
+       osp = (objset_phys_t *)stack;
+       stack += sizeof (objset_phys_t);
+
+       /* XXX add back labels support? */
+       for (label = 0; ubbest == NULL && label < (VDEV_LABELS/2); label++) {
+               uint64_t sector = (label * sizeof (vdev_label_t) +
+                   VDEV_SKIP_SIZE + VDEV_BOOT_HEADER_SIZE +
+                   VDEV_PHYS_SIZE) >> SPA_MINBLOCKSHIFT;
+
+
+               /* Read in the uberblock ring (128K). */
+               if (devread(ffi, sector, 0, VDEV_UBERBLOCK_RING,
+                   (char *)ub_array) == 0)
+                       continue;
+
+               if ((ubbest = find_bestub(ffi, ub_array, label)) != NULL &&
+                   zio_read(ffi, &ubbest->ubp_uberblock.ub_rootbp, osp, stack)
+                   == 0) {
+
+                       VERIFY_OS_TYPE(osp, DMU_OST_META);
+
+                       /* Got the MOS. Save it at the memory addr MOS. */
+                       grub_memmove(MOS, &osp->os_meta_dnode, DNODE_SIZE);
+
+                       if (check_pool_label(ffi, label, stack))
+                               return (0);
+
+                       /*
+                        * Copy fsi->f_data to ffi->ff_data since
+                        * fsig_mount copies from ff_data to f_data
+                        * overwriting fsi->f_data.
+                        */
+                       bcopy(zfs_ba, fsig_file_buf(ffi), FSYS_BUFLEN);
+
+                       is_zfs_mount = 1;
+                       return (1);
+               }
+       }
+
+       return (0);
+}
+
+/*
+ * zfs_open() locates a file in the rootpool by following the
+ * MOS and places the dnode of the file in the memory address DNODE.
+ *
+ * Return:
+ *     1 - success
+ *     0 - failure
+ */
+int
+zfs_open(fsi_file_t *ffi, char *filename)
+{
+       char *stack;
+       dnode_phys_t *mdn;
+       char *bootstring;
+       zfs_bootarea_t *zfs_ba = (zfs_bootarea_t *)ffi->ff_fsi->f_data;
+
+       file_buf = NULL;
+       stackbase = ZFS_SCRATCH;
+       stack = stackbase;
+
+       mdn = (dnode_phys_t *)stack;
+       stack += sizeof (dnode_phys_t);
+
+       dnode_mdn = NULL;
+       dnode_buf = (dnode_phys_t *)stack;
+       stack += 1<<DNODE_BLOCK_SHIFT;
+
+       /*
+        * menu.lst is placed at the root pool filesystem level,
+        * do not goto 'current_bootfs'.
+        */
+       if (is_top_dataset_file(filename)) {
+               if ((errnum = get_objset_mdn(ffi, MOS, NULL, NULL, mdn, stack)))
+                       return (0);
+
+               current_bootfs_obj = 0;
+       } else {
+               if (current_bootfs[0] == '\0') {
+                       /* Get the default root filesystem object number */
+                       if ((errnum = get_default_bootfsobj(ffi, MOS,
+                           &current_bootfs_obj, stack)))
+                               return (0);
+                       if ((errnum = get_objset_mdn(ffi, MOS, NULL,
+                           &current_bootfs_obj, mdn, stack)))
+                               return (0);
+               } else {
+                       if ((errnum = get_objset_mdn(ffi, MOS,
+                           current_bootfs, &current_bootfs_obj, mdn, stack)))
+                               return (0);
+               }
+
+               /*
+                * Put zfs rootpool and boot obj number into bootstring.
+                */
+               if (is_zfs_open == 0) {
+                       char temp[25];          /* needs to hold long long */
+                       int alloc_size;
+                       char zfs_bootstr[] = "zfs-bootfs=";
+                       char zfs_bootpath[] = ",bootpath='";
+
+                       sprintf(temp, "%llu", (unsigned long long)
+                           current_bootfs_obj);
+                       alloc_size = strlen(zfs_bootstr) +
+                           strlen(current_rootpool) +
+                           strlen(temp) + strlen(zfs_bootpath) +
+                           strlen(current_bootpath) + 3;
+                       bootstring = fsi_bootstring_alloc(ffi->ff_fsi,
+                           alloc_size);
+                       if (bootstring != NULL) {
+                               strcpy(bootstring, zfs_bootstr);
+                               strcat(bootstring, current_rootpool);
+                               strcat(bootstring, "/");
+                               strcat(bootstring, temp);
+                               strcat(bootstring, zfs_bootpath);
+                               strcat(bootstring, current_bootpath);
+                               strcat(bootstring, "'");
+                               is_zfs_open = 1;
+                       }
+               }
+       }
+
+       if (dnode_get_path(ffi, mdn, filename, DNODE, stack)) {
+               errnum = ERR_FILE_NOT_FOUND;
+               return (0);
+       }
+
+       /* get the file size and set the file position to 0 */
+       filemax = ((znode_phys_t *)DN_BONUS(DNODE))->zp_size;
+       filepos = 0;
+
+       dnode_buf = NULL;
+       return (1);
+}
+
+/*
+ * zfs_read reads in the data blocks pointed by the DNODE.
+ *
+ * Return:
+ *     len - the length successfully read in to the buffer
+ *     0   - failure
+ */
+int
+zfs_read(fsi_file_t *ffi, char *buf, int len)
+{
+       char *stack;
+       int blksz, length, movesize;
+       zfs_bootarea_t *zfs_ba = (zfs_bootarea_t *)ffi->ff_fsi->f_data;
+
+       if (file_buf == NULL) {
+               file_buf = stackbase;
+               stackbase += SPA_MAXBLOCKSIZE;
+               file_start = file_end = 0;
+       }
+       stack = stackbase;
+
+       /*
+        * If offset is in memory, move it into the buffer provided and return.
+        */
+       if (filepos >= file_start && filepos+len <= file_end) {
+               grub_memmove(buf, file_buf + filepos - file_start, len);
+               filepos += len;
+               return (len);
+       }
+
+       blksz = DNODE->dn_datablkszsec << SPA_MINBLOCKSHIFT;
+
+       /*
+        * Entire Dnode is too big to fit into the space available.  We
+        * will need to read it in chunks.  This could be optimized to
+        * read in as large a chunk as there is space available, but for
+        * now, this only reads in one data block at a time.
+        */
+       length = len;
+       while (length) {
+               /*
+                * Find requested blkid and the offset within that block.
+                */
+               uint64_t blkid = filepos / blksz;
+
+               if ((errnum = dmu_read(ffi, DNODE, blkid, file_buf, stack)))
+                       return (0);
+
+               file_start = blkid * blksz;
+               file_end = file_start + blksz;
+
+               movesize = MIN(length, file_end - filepos);
+
+               grub_memmove(buf, file_buf + filepos - file_start,
+                   movesize);
+               buf += movesize;
+               length -= movesize;
+               filepos += movesize;
+       }
+
+       return (len);
+}
+
+/*
+ * No-Op
+ */
+int
+zfs_embed(int *start_sector, int needed_sectors)
+{
+       return (1);
+}
+
+fsi_plugin_ops_t *
+fsi_init_plugin(int version, fsi_plugin_t *fp, const char **name)
+{
+       static fsig_plugin_ops_t ops = {
+               FSIMAGE_PLUGIN_VERSION,
+               .fpo_mount = zfs_mount,
+               .fpo_dir = zfs_open,
+               .fpo_read = zfs_read
+       };
+
+       *name = "zfs";
+       return (fsig_init(fp, &ops));
+}
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/zfs/fsys_zfs.h
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/fsys_zfs.h   Thu May 01 16:38:56 2008 +0100
@@ -0,0 +1,203 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2008 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+#ifndef _FSYS_ZFS_H
+#define        _FSYS_ZFS_H
+
+#include <fsimage_grub.h>
+#include <fsimage_priv.h>
+
+#include "zfs-include/zfs.h"
+#include "zfs-include/dmu.h"
+#include "zfs-include/spa.h"
+#include "zfs-include/zio.h"
+#include "zfs-include/zio_checksum.h"
+#include "zfs-include/vdev_impl.h"
+#include "zfs-include/zap_impl.h"
+#include "zfs-include/zap_leaf.h"
+#include "zfs-include/uberblock_impl.h"
+#include "zfs-include/dnode.h"
+#include "zfs-include/dsl_dir.h"
+#include "zfs-include/zfs_acl.h"
+#include "zfs-include/zfs_znode.h"
+#include "zfs-include/dsl_dataset.h"
+#include "zfs-include/zil.h"
+#include "zfs-include/dmu_objset.h"
+
+/*
+ * Global Memory addresses to store MOS and DNODE data
+ */
+#define        MOS             ((dnode_phys_t *)(((zfs_bootarea_t *) \
+                           (ffi->ff_fsi->f_data))->zfs_data))
+#define        DNODE           (MOS+1) /* move sizeof(dnode_phys_t) bytes */
+#define        ZFS_SCRATCH     ((char *)(DNODE+1))
+
+#define        MAXNAMELEN      256
+
+typedef struct zfs_bootarea {
+       char zfs_current_bootpath[MAXNAMELEN];
+       char zfs_current_rootpool[MAXNAMELEN];
+       char zfs_current_bootfs[MAXNAMELEN];
+       uint64_t zfs_current_bootfs_obj;
+       int zfs_open;
+
+       /* cache for a file block of the currently zfs_open()-ed file */
+       void *zfs_file_buf;
+       uint64_t zfs_file_start;
+       uint64_t zfs_file_end;
+
+       /* cache for a dnode block */
+       dnode_phys_t *zfs_dnode_buf;
+       dnode_phys_t *zfs_dnode_mdn;
+       uint64_t zfs_dnode_start;
+       uint64_t zfs_dnode_end;
+
+       char *zfs_stackbase;
+       char zfs_data[0x400000];
+} zfs_bootarea_t;
+
+/*
+ * Verify dnode type.
+ * Can only be used in functions returning non-0 for failure.
+ */
+#define        VERIFY_DN_TYPE(dnp, type) \
+       if (type && (dnp)->dn_type != type) { \
+               return (ERR_FSYS_CORRUPT); \
+       }
+
+/*
+ * Verify object set type.
+ * Can only be used in functions returning 0 for failure.
+ */
+#define        VERIFY_OS_TYPE(osp, type) \
+       if (type && (osp)->os_type != type) { \
+               errnum = ERR_FSYS_CORRUPT; \
+               return (0); \
+       }
+
+#define        ZPOOL_PROP_BOOTFS               "bootfs"
+
+/* General macros */
+#define        BSWAP_8(x)      ((x) & 0xff)
+#define        BSWAP_16(x)     ((BSWAP_8(x) << 8) | BSWAP_8((x) >> 8))
+#define        BSWAP_32(x)     ((BSWAP_16(x) << 16) | BSWAP_16((x) >> 16))
+#define        BSWAP_64(x)     ((BSWAP_32(x) << 32) | BSWAP_32((x) >> 32))
+#define        P2ROUNDUP(x, align)     (-(-(x) & -(align)))
+
+/*
+ * XXX Match these macro up with real zfs once we have nvlist support so that 
we
+ * can support large sector disks.
+ */
+#define        UBERBLOCK_SIZE          (1ULL << UBERBLOCK_SHIFT)
+#undef offsetof
+#define        offsetof(t, m)   (size_t)(&(((t *)0)->m))
+#define        VDEV_UBERBLOCK_SHIFT    UBERBLOCK_SHIFT
+#define        VDEV_UBERBLOCK_OFFSET(n) \
+offsetof(vdev_label_t, vl_uberblock[(n) << VDEV_UBERBLOCK_SHIFT])
+
+typedef struct uberblock uberblock_t;
+
+/* XXX Uberblock_phys_t is no longer in the kernel zfs */
+typedef struct uberblock_phys {
+       uberblock_t     ubp_uberblock;
+       char            ubp_pad[UBERBLOCK_SIZE - sizeof (uberblock_t) -
+                               sizeof (zio_block_tail_t)];
+       zio_block_tail_t ubp_zbt;
+} uberblock_phys_t;
+
+/*
+ * Macros to get fields in a bp or DVA.
+ */
+#define        P2PHASE(x, align)               ((x) & ((align) - 1))
+#define        DVA_OFFSET_TO_PHYS_SECTOR(offset) \
+       ((offset + VDEV_LABEL_START_SIZE) >> SPA_MINBLOCKSHIFT)
+
+/*
+ * For nvlist manipulation. (from nvpair.h)
+ */
+#define        NV_ENCODE_NATIVE        0
+#define        NV_ENCODE_XDR           1
+#define        HOST_ENDIAN             1       /* for x86 machine */
+#define        DATA_TYPE_UINT64        8
+#define        DATA_TYPE_STRING        9
+#define        DATA_TYPE_NVLIST        19
+#define        DATA_TYPE_NVLIST_ARRAY  20
+
+/*
+ * Decompression Entry - lzjb
+ */
+#ifndef        NBBY
+#define        NBBY    8
+#endif
+
+typedef int zfs_decomp_func_t(void *s_start, void *d_start, size_t s_len,
+                       size_t d_len);
+typedef struct decomp_entry {
+       char *name;
+       zfs_decomp_func_t *decomp_func;
+} decomp_entry_t;
+
+/*
+ * FAT ZAP data structures
+ */
+#define        ZFS_CRC64_POLY 0xC96C5795D7870F42ULL /* ECMA-182, reflected 
form */
+#define        ZAP_HASH_IDX(hash, n)   (((n) == 0) ? 0 : ((hash) >> (64 - 
(n))))
+#define        CHAIN_END       0xffff  /* end of the chunk chain */
+
+/*
+ * The amount of space within the chunk available for the array is:
+ * chunk size - space for type (1) - space for next pointer (2)
+ */
+#define        ZAP_LEAF_ARRAY_BYTES (ZAP_LEAF_CHUNKSIZE - 3)
+
+#define        ZAP_LEAF_HASH_SHIFT(bs) (bs - 5)
+#define        ZAP_LEAF_HASH_NUMENTRIES(bs) (1 << ZAP_LEAF_HASH_SHIFT(bs))
+#define        LEAF_HASH(bs, h) \
+       ((ZAP_LEAF_HASH_NUMENTRIES(bs)-1) & \
+       ((h) >> (64 - ZAP_LEAF_HASH_SHIFT(bs)-l->l_hdr.lh_prefix_len)))
+
+/*
+ * The amount of space available for chunks is:
+ * block size shift - hash entry size (2) * number of hash
+ * entries - header space (2*chunksize)
+ */
+#define        ZAP_LEAF_NUMCHUNKS(bs) \
+       (((1<<bs) - 2*ZAP_LEAF_HASH_NUMENTRIES(bs)) / \
+       ZAP_LEAF_CHUNKSIZE - 2)
+
+/*
+ * The chunks start immediately after the hash table.  The end of the
+ * hash table is at l_hash + HASH_NUMENTRIES, which we simply cast to a
+ * chunk_t.
+ */
+#define        ZAP_LEAF_CHUNK(l, bs, idx) \
+       ((zap_leaf_chunk_t *)(l->l_hash + ZAP_LEAF_HASH_NUMENTRIES(bs)))[idx]
+#define        ZAP_LEAF_ENTRY(l, bs, idx) (&ZAP_LEAF_CHUNK(l, bs, idx).l_entry)
+
+extern void fletcher_2_native(const void *, uint64_t, zio_cksum_t *);
+extern void fletcher_2_byteswap(const void *, uint64_t, zio_cksum_t *);
+extern void fletcher_4_native(const void *, uint64_t, zio_cksum_t *);
+extern void fletcher_4_byteswap(const void *, uint64_t, zio_cksum_t *);
+extern void zio_checksum_SHA256(const void *, uint64_t, zio_cksum_t *);
+extern int lzjb_decompress(void *, void *, size_t, size_t);
+
+#endif /* !_FSYS_ZFS_H */
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/zfs/mb_info.h
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/mb_info.h    Thu May 01 16:38:56 2008 +0100
@@ -0,0 +1,217 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 2000,2003  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+
+/*
+ *  The structure type "mod_list" is used by the "multiboot_info" structure.
+ */
+
+struct mod_list
+{
+  /* the memory used goes from bytes 'mod_start' to 'mod_end-1' inclusive */
+  unsigned long mod_start;
+  unsigned long mod_end;
+  
+  /* Module command line */
+  unsigned long cmdline;
+  
+  /* padding to take it to 16 bytes (must be zero) */
+  unsigned long pad;
+};
+
+
+/*
+ *  INT-15, AX=E820 style "AddressRangeDescriptor"
+ *  ...with a "size" parameter on the front which is the structure size - 4,
+ *  pointing to the next one, up until the full buffer length of the memory
+ *  map has been reached.
+ */
+
+struct AddrRangeDesc
+{
+  unsigned long size;
+  unsigned long long BaseAddr;
+  unsigned long long Length;
+  unsigned long Type;
+  
+  /* unspecified optional padding... */
+} __attribute__ ((packed));
+
+/* usable memory "Type", all others are reserved.  */
+#define MB_ARD_MEMORY          1
+
+
+/* Drive Info structure.  */
+struct drive_info
+{
+  /* The size of this structure.  */
+  unsigned long size;
+
+  /* The BIOS drive number.  */
+  unsigned char drive_number;
+
+  /* The access mode (see below).  */
+  unsigned char drive_mode;
+
+  /* The BIOS geometry.  */
+  unsigned short drive_cylinders;
+  unsigned char drive_heads;
+  unsigned char drive_sectors;
+
+  /* The array of I/O ports used for the drive.  */
+  unsigned short drive_ports[0];
+};
+
+/* Drive Mode.  */
+#define MB_DI_CHS_MODE         0
+#define MB_DI_LBA_MODE         1
+
+
+/* APM BIOS info.  */
+struct apm_info
+{
+  unsigned short version;
+  unsigned short cseg;
+  unsigned long offset;
+  unsigned short cseg_16;
+  unsigned short dseg_16;
+  unsigned short cseg_len;
+  unsigned short cseg_16_len;
+  unsigned short dseg_16_len;
+};
+
+
+/*
+ *  MultiBoot Info description
+ *
+ *  This is the struct passed to the boot image.  This is done by placing
+ *  its address in the EAX register.
+ */
+
+struct multiboot_info
+{
+  /* MultiBoot info version number */
+  unsigned long flags;
+  
+  /* Available memory from BIOS */
+  unsigned long mem_lower;
+  unsigned long mem_upper;
+  
+  /* "root" partition */
+  unsigned long boot_device;
+  
+  /* Kernel command line */
+  unsigned long cmdline;
+  
+  /* Boot-Module list */
+  unsigned long mods_count;
+  unsigned long mods_addr;
+  
+  union
+  {
+    struct
+    {
+      /* (a.out) Kernel symbol table info */
+      unsigned long tabsize;
+      unsigned long strsize;
+      unsigned long addr;
+      unsigned long pad;
+    }
+    a;
+    
+    struct
+    {
+      /* (ELF) Kernel section header table */
+      unsigned long num;
+      unsigned long size;
+      unsigned long addr;
+      unsigned long shndx;
+    }
+    e;
+  }
+  syms;
+  
+  /* Memory Mapping buffer */
+  unsigned long mmap_length;
+  unsigned long mmap_addr;
+  
+  /* Drive Info buffer */
+  unsigned long drives_length;
+  unsigned long drives_addr;
+  
+  /* ROM configuration table */
+  unsigned long config_table;
+  
+  /* Boot Loader Name */
+  unsigned long boot_loader_name;
+
+  /* APM table */
+  unsigned long apm_table;
+
+  /* Video */
+  unsigned long vbe_control_info;
+  unsigned long vbe_mode_info;
+  unsigned short vbe_mode;
+  unsigned short vbe_interface_seg;
+  unsigned short vbe_interface_off;
+  unsigned short vbe_interface_len;
+};
+
+/*
+ *  Flags to be set in the 'flags' parameter above
+ */
+
+/* is there basic lower/upper memory information? */
+#define MB_INFO_MEMORY                 0x00000001
+/* is there a boot device set? */
+#define MB_INFO_BOOTDEV                        0x00000002
+/* is the command-line defined? */
+#define MB_INFO_CMDLINE                        0x00000004
+/* are there modules to do something with? */
+#define MB_INFO_MODS                   0x00000008
+
+/* These next two are mutually exclusive */
+
+/* is there a symbol table loaded? */
+#define MB_INFO_AOUT_SYMS              0x00000010
+/* is there an ELF section header table? */
+#define MB_INFO_ELF_SHDR               0x00000020
+
+/* is there a full memory map? */
+#define MB_INFO_MEM_MAP                        0x00000040
+
+/* Is there drive info?  */
+#define MB_INFO_DRIVE_INFO             0x00000080
+
+/* Is there a config table?  */
+#define MB_INFO_CONFIG_TABLE           0x00000100
+
+/* Is there a boot loader name?  */
+#define MB_INFO_BOOT_LOADER_NAME       0x00000200
+
+/* Is there a APM table?  */
+#define MB_INFO_APM_TABLE              0x00000400
+
+/* Is there video information?  */
+#define MB_INFO_VIDEO_INFO             0x00000800
+
+/*
+ *  The following value must be present in the EAX register.
+ */
+
+#define MULTIBOOT_VALID                        0x2BADB002
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/zfs/zfs-include/dmu.h
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/zfs-include/dmu.h    Thu May 01 16:38:56 2008 +0100
@@ -0,0 +1,105 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef        _SYS_DMU_H
+#define        _SYS_DMU_H
+
+/*
+ * This file describes the interface that the DMU provides for its
+ * consumers.
+ *
+ * The DMU also interacts with the SPA.  That interface is described in
+ * dmu_spa.h.
+ */
+typedef enum dmu_object_type {
+       DMU_OT_NONE,
+       /* general: */
+       DMU_OT_OBJECT_DIRECTORY,        /* ZAP */
+       DMU_OT_OBJECT_ARRAY,            /* UINT64 */
+       DMU_OT_PACKED_NVLIST,           /* UINT8 (XDR by nvlist_pack/unpack) */
+       DMU_OT_PACKED_NVLIST_SIZE,      /* UINT64 */
+       DMU_OT_BPLIST,                  /* UINT64 */
+       DMU_OT_BPLIST_HDR,              /* UINT64 */
+       /* spa: */
+       DMU_OT_SPACE_MAP_HEADER,        /* UINT64 */
+       DMU_OT_SPACE_MAP,               /* UINT64 */
+       /* zil: */
+       DMU_OT_INTENT_LOG,              /* UINT64 */
+       /* dmu: */
+       DMU_OT_DNODE,                   /* DNODE */
+       DMU_OT_OBJSET,                  /* OBJSET */
+       /* dsl: */
+       DMU_OT_DSL_DIR,                 /* UINT64 */
+       DMU_OT_DSL_DIR_CHILD_MAP,       /* ZAP */
+       DMU_OT_DSL_DS_SNAP_MAP,         /* ZAP */
+       DMU_OT_DSL_PROPS,               /* ZAP */
+       DMU_OT_DSL_DATASET,             /* UINT64 */
+       /* zpl: */
+       DMU_OT_ZNODE,                   /* ZNODE */
+       DMU_OT_ACL,                     /* ACL */
+       DMU_OT_PLAIN_FILE_CONTENTS,     /* UINT8 */
+       DMU_OT_DIRECTORY_CONTENTS,      /* ZAP */
+       DMU_OT_MASTER_NODE,             /* ZAP */
+       DMU_OT_UNLINKED_SET,            /* ZAP */
+       /* zvol: */
+       DMU_OT_ZVOL,                    /* UINT8 */
+       DMU_OT_ZVOL_PROP,               /* ZAP */
+       /* other; for testing only! */
+       DMU_OT_PLAIN_OTHER,             /* UINT8 */
+       DMU_OT_UINT64_OTHER,            /* UINT64 */
+       DMU_OT_ZAP_OTHER,               /* ZAP */
+       /* new object types: */
+       DMU_OT_ERROR_LOG,               /* ZAP */
+       DMU_OT_SPA_HISTORY,             /* UINT8 */
+       DMU_OT_SPA_HISTORY_OFFSETS,     /* spa_his_phys_t */
+       DMU_OT_POOL_PROPS,              /* ZAP */
+
+       DMU_OT_NUMTYPES
+} dmu_object_type_t;
+
+typedef enum dmu_objset_type {
+       DMU_OST_NONE,
+       DMU_OST_META,
+       DMU_OST_ZFS,
+       DMU_OST_ZVOL,
+       DMU_OST_OTHER,                  /* For testing only! */
+       DMU_OST_ANY,                    /* Be careful! */
+       DMU_OST_NUMTYPES
+} dmu_objset_type_t;
+
+/*
+ * The names of zap entries in the DIRECTORY_OBJECT of the MOS.
+ */
+#define        DMU_POOL_DIRECTORY_OBJECT       1
+#define        DMU_POOL_CONFIG                 "config"
+#define        DMU_POOL_ROOT_DATASET           "root_dataset"
+#define        DMU_POOL_SYNC_BPLIST            "sync_bplist"
+#define        DMU_POOL_ERRLOG_SCRUB           "errlog_scrub"
+#define        DMU_POOL_ERRLOG_LAST            "errlog_last"
+#define        DMU_POOL_SPARES                 "spares"
+#define        DMU_POOL_DEFLATE                "deflate"
+#define        DMU_POOL_HISTORY                "history"
+#define        DMU_POOL_PROPS                  "pool_props"
+#define        DMU_POOL_L2CACHE                "l2cache"
+
+#endif /* _SYS_DMU_H */
diff -r 8ccaf43782bd -r 64f790e90d3d 
tools/libfsimage/zfs/zfs-include/dmu_objset.h
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/zfs-include/dmu_objset.h     Thu May 01 16:38:56 
2008 +0100
@@ -0,0 +1,35 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef        _SYS_DMU_OBJSET_H
+#define        _SYS_DMU_OBJSET_H
+
+typedef struct objset_phys {
+       dnode_phys_t os_meta_dnode;
+       zil_header_t os_zil_header;
+       uint64_t os_type;
+       char os_pad[1024 - sizeof (dnode_phys_t) - sizeof (zil_header_t) -
+           sizeof (uint64_t)];
+} objset_phys_t;
+
+#endif /* _SYS_DMU_OBJSET_H */
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/zfs/zfs-include/dnode.h
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/zfs-include/dnode.h  Thu May 01 16:38:56 2008 +0100
@@ -0,0 +1,76 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef        _SYS_DNODE_H
+#define        _SYS_DNODE_H
+
+/*
+ * Fixed constants.
+ */
+#define        DNODE_SHIFT             9       /* 512 bytes */
+#define        DN_MIN_INDBLKSHIFT      10      /* 1k */
+#define        DN_MAX_INDBLKSHIFT      14      /* 16k */
+#define        DNODE_BLOCK_SHIFT       14      /* 16k */
+#define        DNODE_CORE_SIZE         64      /* 64 bytes for dnode sans 
blkptrs */
+#define        DN_MAX_OBJECT_SHIFT     48      /* 256 trillion (zfs_fid_t 
limit) */
+#define        DN_MAX_OFFSET_SHIFT     64      /* 2^64 bytes in a dnode */
+
+/*
+ * Derived constants.
+ */
+#define        DNODE_SIZE      (1 << DNODE_SHIFT)
+#define        DN_MAX_NBLKPTR  ((DNODE_SIZE - DNODE_CORE_SIZE) >> 
SPA_BLKPTRSHIFT)
+#define        DN_MAX_BONUSLEN (DNODE_SIZE - DNODE_CORE_SIZE - (1 << 
SPA_BLKPTRSHIFT))
+#define        DN_MAX_OBJECT   (1ULL << DN_MAX_OBJECT_SHIFT)
+
+#define        DNODES_PER_BLOCK_SHIFT  (DNODE_BLOCK_SHIFT - DNODE_SHIFT)
+#define        DNODES_PER_BLOCK        (1ULL << DNODES_PER_BLOCK_SHIFT)
+#define        DNODES_PER_LEVEL_SHIFT  (DN_MAX_INDBLKSHIFT - SPA_BLKPTRSHIFT)
+
+#define        DN_BONUS(dnp)   ((void*)((dnp)->dn_bonus + \
+       (((dnp)->dn_nblkptr - 1) * sizeof (blkptr_t))))
+
+typedef struct dnode_phys {
+       uint8_t dn_type;                /* dmu_object_type_t */
+       uint8_t dn_indblkshift;         /* ln2(indirect block size) */
+       uint8_t dn_nlevels;             /* 1=dn_blkptr->data blocks */
+       uint8_t dn_nblkptr;             /* length of dn_blkptr */
+       uint8_t dn_bonustype;           /* type of data in bonus buffer */
+       uint8_t dn_checksum;            /* ZIO_CHECKSUM type */
+       uint8_t dn_compress;            /* ZIO_COMPRESS type */
+       uint8_t dn_flags;               /* DNODE_FLAG_* */
+       uint16_t dn_datablkszsec;       /* data block size in 512b sectors */
+       uint16_t dn_bonuslen;           /* length of dn_bonus */
+       uint8_t dn_pad2[4];
+
+       /* accounting is protected by dn_dirty_mtx */
+       uint64_t dn_maxblkid;           /* largest allocated block ID */
+       uint64_t dn_used;               /* bytes (or sectors) of disk space */
+
+       uint64_t dn_pad3[4];
+
+       blkptr_t dn_blkptr[1];
+       uint8_t dn_bonus[DN_MAX_BONUSLEN];
+} dnode_phys_t;
+
+#endif /* _SYS_DNODE_H */
diff -r 8ccaf43782bd -r 64f790e90d3d 
tools/libfsimage/zfs/zfs-include/dsl_dataset.h
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/zfs-include/dsl_dataset.h    Thu May 01 16:38:56 
2008 +0100
@@ -0,0 +1,53 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef        _SYS_DSL_DATASET_H
+#define        _SYS_DSL_DATASET_H
+
+typedef struct dsl_dataset_phys {
+       uint64_t ds_dir_obj;
+       uint64_t ds_prev_snap_obj;
+       uint64_t ds_prev_snap_txg;
+       uint64_t ds_next_snap_obj;
+       uint64_t ds_snapnames_zapobj;   /* zap obj of snaps; ==0 for snaps */
+       uint64_t ds_num_children;       /* clone/snap children; ==0 for head */
+       uint64_t ds_creation_time;      /* seconds since 1970 */
+       uint64_t ds_creation_txg;
+       uint64_t ds_deadlist_obj;
+       uint64_t ds_used_bytes;
+       uint64_t ds_compressed_bytes;
+       uint64_t ds_uncompressed_bytes;
+       uint64_t ds_unique_bytes;       /* only relevant to snapshots */
+       /*
+        * The ds_fsid_guid is a 56-bit ID that can change to avoid
+        * collisions.  The ds_guid is a 64-bit ID that will never
+        * change, so there is a small probability that it will collide.
+        */
+       uint64_t ds_fsid_guid;
+       uint64_t ds_guid;
+       uint64_t ds_flags;
+       blkptr_t ds_bp;
+       uint64_t ds_pad[8]; /* pad out to 320 bytes for good measure */
+} dsl_dataset_phys_t;
+
+#endif /* _SYS_DSL_DATASET_H */
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/zfs/zfs-include/dsl_dir.h
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/zfs-include/dsl_dir.h        Thu May 01 16:38:56 
2008 +0100
@@ -0,0 +1,49 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef        _SYS_DSL_DIR_H
+#define        _SYS_DSL_DIR_H
+
+typedef struct dsl_dir_phys {
+       uint64_t dd_creation_time; /* not actually used */
+       uint64_t dd_head_dataset_obj;
+       uint64_t dd_parent_obj;
+       uint64_t dd_clone_parent_obj;
+       uint64_t dd_child_dir_zapobj;
+       /*
+        * how much space our children are accounting for; for leaf
+        * datasets, == physical space used by fs + snaps
+        */
+       uint64_t dd_used_bytes;
+       uint64_t dd_compressed_bytes;
+       uint64_t dd_uncompressed_bytes;
+       /* Administrative quota setting */
+       uint64_t dd_quota;
+       /* Administrative reservation setting */
+       uint64_t dd_reserved;
+       uint64_t dd_props_zapobj;
+       uint64_t dd_deleg_zapobj;       /* dataset permissions */
+       uint64_t dd_pad[20]; /* pad out to 256 bytes for good measure */
+} dsl_dir_phys_t;
+
+#endif /* _SYS_DSL_DIR_H */
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/zfs/zfs-include/spa.h
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/zfs-include/spa.h    Thu May 01 16:38:56 2008 +0100
@@ -0,0 +1,283 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef _SYS_SPA_H
+#define        _SYS_SPA_H
+
+/*
+ * General-purpose 32-bit and 64-bit bitfield encodings.
+ */
+#define        BF32_DECODE(x, low, len)        P2PHASE((x) >> (low), 1U << 
(len))
+#define        BF64_DECODE(x, low, len)        P2PHASE((x) >> (low), 1ULL << 
(len))
+#define        BF32_ENCODE(x, low, len)        (P2PHASE((x), 1U << (len)) << 
(low))
+#define        BF64_ENCODE(x, low, len)        (P2PHASE((x), 1ULL << (len)) << 
(low))
+
+#define        BF32_GET(x, low, len)           BF32_DECODE(x, low, len)
+#define        BF64_GET(x, low, len)           BF64_DECODE(x, low, len)
+
+#define        BF32_SET(x, low, len, val)      \
+       ((x) ^= BF32_ENCODE((x >> low) ^ (val), low, len))
+#define        BF64_SET(x, low, len, val)      \
+       ((x) ^= BF64_ENCODE((x >> low) ^ (val), low, len))
+
+#define        BF32_GET_SB(x, low, len, shift, bias)   \
+       ((BF32_GET(x, low, len) + (bias)) << (shift))
+#define        BF64_GET_SB(x, low, len, shift, bias)   \
+       ((BF64_GET(x, low, len) + (bias)) << (shift))
+
+#define        BF32_SET_SB(x, low, len, shift, bias, val)      \
+       BF32_SET(x, low, len, ((val) >> (shift)) - (bias))
+#define        BF64_SET_SB(x, low, len, shift, bias, val)      \
+       BF64_SET(x, low, len, ((val) >> (shift)) - (bias))
+
+/*
+ * We currently support nine block sizes, from 512 bytes to 128K.
+ * We could go higher, but the benefits are near-zero and the cost
+ * of COWing a giant block to modify one byte would become excessive.
+ */
+#define        SPA_MINBLOCKSHIFT       9
+#define        SPA_MAXBLOCKSHIFT       17
+#define        SPA_MINBLOCKSIZE        (1ULL << SPA_MINBLOCKSHIFT)
+#define        SPA_MAXBLOCKSIZE        (1ULL << SPA_MAXBLOCKSHIFT)
+
+#define        SPA_BLOCKSIZES          (SPA_MAXBLOCKSHIFT - SPA_MINBLOCKSHIFT 
+ 1)
+
+/*
+ * The DVA size encodings for LSIZE and PSIZE support blocks up to 32MB.
+ * The ASIZE encoding should be at least 64 times larger (6 more bits)
+ * to support up to 4-way RAID-Z mirror mode with worst-case gang block
+ * overhead, three DVAs per bp, plus one more bit in case we do anything
+ * else that expands the ASIZE.
+ */
+#define        SPA_LSIZEBITS           16      /* LSIZE up to 32M (2^16 * 512) 
*/
+#define        SPA_PSIZEBITS           16      /* PSIZE up to 32M (2^16 * 512) 
*/
+#define        SPA_ASIZEBITS           24      /* ASIZE up to 64 times larger  
*/
+
+/*
+ * All SPA data is represented by 128-bit data virtual addresses (DVAs).
+ * The members of the dva_t should be considered opaque outside the SPA.
+ */
+typedef struct dva {
+       uint64_t        dva_word[2];
+} dva_t;
+
+/*
+ * Each block has a 256-bit checksum -- strong enough for cryptographic hashes.
+ */
+typedef struct zio_cksum {
+       uint64_t        zc_word[4];
+} zio_cksum_t;
+
+/*
+ * Each block is described by its DVAs, time of birth, checksum, etc.
+ * The word-by-word, bit-by-bit layout of the blkptr is as follows:
+ *
+ *     64      56      48      40      32      24      16      8       0
+ *     +-------+-------+-------+-------+-------+-------+-------+-------+
+ * 0   |               vdev1           | GRID  |         ASIZE         |
+ *     +-------+-------+-------+-------+-------+-------+-------+-------+
+ * 1   |G|                      offset1                                |
+ *     +-------+-------+-------+-------+-------+-------+-------+-------+
+ * 2   |               vdev2           | GRID  |         ASIZE         |
+ *     +-------+-------+-------+-------+-------+-------+-------+-------+
+ * 3   |G|                      offset2                                |
+ *     +-------+-------+-------+-------+-------+-------+-------+-------+
+ * 4   |               vdev3           | GRID  |         ASIZE         |
+ *     +-------+-------+-------+-------+-------+-------+-------+-------+
+ * 5   |G|                      offset3                                |
+ *     +-------+-------+-------+-------+-------+-------+-------+-------+
+ * 6   |E| lvl | type  | cksum | comp  |     PSIZE     |     LSIZE     |
+ *     +-------+-------+-------+-------+-------+-------+-------+-------+
+ * 7   |                       padding                                 |
+ *     +-------+-------+-------+-------+-------+-------+-------+-------+
+ * 8   |                       padding                                 |
+ *     +-------+-------+-------+-------+-------+-------+-------+-------+
+ * 9   |                       padding                                 |
+ *     +-------+-------+-------+-------+-------+-------+-------+-------+
+ * a   |                       birth txg                               |
+ *     +-------+-------+-------+-------+-------+-------+-------+-------+
+ * b   |                       fill count                              |
+ *     +-------+-------+-------+-------+-------+-------+-------+-------+
+ * c   |                       checksum[0]                             |
+ *     +-------+-------+-------+-------+-------+-------+-------+-------+
+ * d   |                       checksum[1]                             |
+ *     +-------+-------+-------+-------+-------+-------+-------+-------+
+ * e   |                       checksum[2]                             |
+ *     +-------+-------+-------+-------+-------+-------+-------+-------+
+ * f   |                       checksum[3]                             |
+ *     +-------+-------+-------+-------+-------+-------+-------+-------+
+ *
+ * Legend:
+ *
+ * vdev                virtual device ID
+ * offset      offset into virtual device
+ * LSIZE       logical size
+ * PSIZE       physical size (after compression)
+ * ASIZE       allocated size (including RAID-Z parity and gang block headers)
+ * GRID                RAID-Z layout information (reserved for future use)
+ * cksum       checksum function
+ * comp                compression function
+ * G           gang block indicator
+ * E           endianness
+ * type                DMU object type
+ * lvl         level of indirection
+ * birth txg   transaction group in which the block was born
+ * fill count  number of non-zero blocks under this bp
+ * checksum[4] 256-bit checksum of the data this bp describes
+ */
+typedef struct blkptr {
+       dva_t           blk_dva[3];     /* 128-bit Data Virtual Address */
+       uint64_t        blk_prop;       /* size, compression, type, etc */
+       uint64_t        blk_pad[3];     /* Extra space for the future   */
+       uint64_t        blk_birth;      /* transaction group at birth   */
+       uint64_t        blk_fill;       /* fill count                   */
+       zio_cksum_t     blk_cksum;      /* 256-bit checksum             */
+} blkptr_t;
+
+#define        SPA_BLKPTRSHIFT 7               /* blkptr_t is 128 bytes        
*/
+#define        SPA_DVAS_PER_BP 3               /* Number of DVAs in a bp       
*/
+
+/*
+ * Macros to get and set fields in a bp or DVA.
+ */
+#define        DVA_GET_ASIZE(dva)      \
+       BF64_GET_SB((dva)->dva_word[0], 0, 24, SPA_MINBLOCKSHIFT, 0)
+#define        DVA_SET_ASIZE(dva, x)   \
+       BF64_SET_SB((dva)->dva_word[0], 0, 24, SPA_MINBLOCKSHIFT, 0, x)
+
+#define        DVA_GET_GRID(dva)       BF64_GET((dva)->dva_word[0], 24, 8)
+#define        DVA_SET_GRID(dva, x)    BF64_SET((dva)->dva_word[0], 24, 8, x)
+
+#define        DVA_GET_VDEV(dva)       BF64_GET((dva)->dva_word[0], 32, 32)
+#define        DVA_SET_VDEV(dva, x)    BF64_SET((dva)->dva_word[0], 32, 32, x)
+
+#define        DVA_GET_OFFSET(dva)     \
+       BF64_GET_SB((dva)->dva_word[1], 0, 63, SPA_MINBLOCKSHIFT, 0)
+#define        DVA_SET_OFFSET(dva, x)  \
+       BF64_SET_SB((dva)->dva_word[1], 0, 63, SPA_MINBLOCKSHIFT, 0, x)
+
+#define        DVA_GET_GANG(dva)       BF64_GET((dva)->dva_word[1], 63, 1)
+#define        DVA_SET_GANG(dva, x)    BF64_SET((dva)->dva_word[1], 63, 1, x)
+
+#define        BP_GET_LSIZE(bp)        \
+       (BP_IS_HOLE(bp) ? 0 : \
+       BF64_GET_SB((bp)->blk_prop, 0, 16, SPA_MINBLOCKSHIFT, 1))
+#define        BP_SET_LSIZE(bp, x)     \
+       BF64_SET_SB((bp)->blk_prop, 0, 16, SPA_MINBLOCKSHIFT, 1, x)
+
+#define        BP_GET_PSIZE(bp)        \
+       BF64_GET_SB((bp)->blk_prop, 16, 16, SPA_MINBLOCKSHIFT, 1)
+#define        BP_SET_PSIZE(bp, x)     \
+       BF64_SET_SB((bp)->blk_prop, 16, 16, SPA_MINBLOCKSHIFT, 1, x)
+
+#define        BP_GET_COMPRESS(bp)     BF64_GET((bp)->blk_prop, 32, 8)
+#define        BP_SET_COMPRESS(bp, x)  BF64_SET((bp)->blk_prop, 32, 8, x)
+
+#define        BP_GET_CHECKSUM(bp)     BF64_GET((bp)->blk_prop, 40, 8)
+#define        BP_SET_CHECKSUM(bp, x)  BF64_SET((bp)->blk_prop, 40, 8, x)
+
+#define        BP_GET_TYPE(bp)         BF64_GET((bp)->blk_prop, 48, 8)
+#define        BP_SET_TYPE(bp, x)      BF64_SET((bp)->blk_prop, 48, 8, x)
+
+#define        BP_GET_LEVEL(bp)        BF64_GET((bp)->blk_prop, 56, 5)
+#define        BP_SET_LEVEL(bp, x)     BF64_SET((bp)->blk_prop, 56, 5, x)
+
+#define        BP_GET_BYTEORDER(bp)    (0 - BF64_GET((bp)->blk_prop, 63, 1))
+#define        BP_SET_BYTEORDER(bp, x) BF64_SET((bp)->blk_prop, 63, 1, x)
+
+#define        BP_GET_ASIZE(bp)        \
+       (DVA_GET_ASIZE(&(bp)->blk_dva[0]) + DVA_GET_ASIZE(&(bp)->blk_dva[1]) + \
+               DVA_GET_ASIZE(&(bp)->blk_dva[2]))
+
+#define        BP_GET_UCSIZE(bp) \
+       ((BP_GET_LEVEL(bp) > 0 || dmu_ot[BP_GET_TYPE(bp)].ot_metadata) ? \
+       BP_GET_PSIZE(bp) : BP_GET_LSIZE(bp));
+
+#define        BP_GET_NDVAS(bp)        \
+       (!!DVA_GET_ASIZE(&(bp)->blk_dva[0]) + \
+       !!DVA_GET_ASIZE(&(bp)->blk_dva[1]) + \
+       !!DVA_GET_ASIZE(&(bp)->blk_dva[2]))
+
+#define        BP_COUNT_GANG(bp)       \
+       (DVA_GET_GANG(&(bp)->blk_dva[0]) + \
+       DVA_GET_GANG(&(bp)->blk_dva[1]) + \
+       DVA_GET_GANG(&(bp)->blk_dva[2]))
+
+#define        DVA_EQUAL(dva1, dva2)   \
+       ((dva1)->dva_word[1] == (dva2)->dva_word[1] && \
+       (dva1)->dva_word[0] == (dva2)->dva_word[0])
+
+#define        ZIO_CHECKSUM_EQUAL(zc1, zc2) \
+       (0 == (((zc1).zc_word[0] - (zc2).zc_word[0]) | \
+       ((zc1).zc_word[1] - (zc2).zc_word[1]) | \
+       ((zc1).zc_word[2] - (zc2).zc_word[2]) | \
+       ((zc1).zc_word[3] - (zc2).zc_word[3])))
+
+
+#define        DVA_IS_VALID(dva)       (DVA_GET_ASIZE(dva) != 0)
+
+#define        ZIO_SET_CHECKSUM(zcp, w0, w1, w2, w3)   \
+{                                              \
+       (zcp)->zc_word[0] = w0;                 \
+       (zcp)->zc_word[1] = w1;                 \
+       (zcp)->zc_word[2] = w2;                 \
+       (zcp)->zc_word[3] = w3;                 \
+}
+
+#define        BP_IDENTITY(bp)         (&(bp)->blk_dva[0])
+#define        BP_IS_GANG(bp)          DVA_GET_GANG(BP_IDENTITY(bp))
+#define        BP_IS_HOLE(bp)          ((bp)->blk_birth == 0)
+#define        BP_IS_OLDER(bp, txg)    (!BP_IS_HOLE(bp) && (bp)->blk_birth < 
(txg))
+
+#define        BP_ZERO(bp)                             \
+{                                              \
+       (bp)->blk_dva[0].dva_word[0] = 0;       \
+       (bp)->blk_dva[0].dva_word[1] = 0;       \
+       (bp)->blk_dva[1].dva_word[0] = 0;       \
+       (bp)->blk_dva[1].dva_word[1] = 0;       \
+       (bp)->blk_dva[2].dva_word[0] = 0;       \
+       (bp)->blk_dva[2].dva_word[1] = 0;       \
+       (bp)->blk_prop = 0;                     \
+       (bp)->blk_pad[0] = 0;                   \
+       (bp)->blk_pad[1] = 0;                   \
+       (bp)->blk_pad[2] = 0;                   \
+       (bp)->blk_birth = 0;                    \
+       (bp)->blk_fill = 0;                     \
+       ZIO_SET_CHECKSUM(&(bp)->blk_cksum, 0, 0, 0, 0); \
+}
+
+/*
+ * Note: the byteorder is either 0 or -1, both of which are palindromes.
+ * This simplifies the endianness handling a bit.
+ */
+#ifdef _BIG_ENDIAN
+#define        ZFS_HOST_BYTEORDER      (0ULL)
+#else
+#define        ZFS_HOST_BYTEORDER      (-1ULL)
+#endif
+
+#define        BP_SHOULD_BYTESWAP(bp)  (BP_GET_BYTEORDER(bp) != 
ZFS_HOST_BYTEORDER)
+
+#define        BP_SPRINTF_LEN  320
+
+#endif /* _SYS_SPA_H */
diff -r 8ccaf43782bd -r 64f790e90d3d 
tools/libfsimage/zfs/zfs-include/uberblock_impl.h
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/zfs-include/uberblock_impl.h Thu May 01 16:38:56 
2008 +0100
@@ -0,0 +1,49 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef _SYS_UBERBLOCK_IMPL_H
+#define        _SYS_UBERBLOCK_IMPL_H
+
+/*
+ * The uberblock version is incremented whenever an incompatible on-disk
+ * format change is made to the SPA, DMU, or ZAP.
+ *
+ * Note: the first two fields should never be moved.  When a storage pool
+ * is opened, the uberblock must be read off the disk before the version
+ * can be checked.  If the ub_version field is moved, we may not detect
+ * version mismatch.  If the ub_magic field is moved, applications that
+ * expect the magic number in the first word won't work.
+ */
+#define        UBERBLOCK_MAGIC         0x00bab10c              /* oo-ba-bloc!  
*/
+#define        UBERBLOCK_SHIFT         10                      /* up to 1K     
*/
+
+struct uberblock {
+       uint64_t        ub_magic;       /* UBERBLOCK_MAGIC              */
+       uint64_t        ub_version;     /* ZFS_VERSION                  */
+       uint64_t        ub_txg;         /* txg of last sync             */
+       uint64_t        ub_guid_sum;    /* sum of all vdev guids        */
+       uint64_t        ub_timestamp;   /* UTC time of last sync        */
+       blkptr_t        ub_rootbp;      /* MOS objset_phys_t            */
+};
+
+#endif /* _SYS_UBERBLOCK_IMPL_H */
diff -r 8ccaf43782bd -r 64f790e90d3d 
tools/libfsimage/zfs/zfs-include/vdev_impl.h
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/zfs-include/vdev_impl.h      Thu May 01 16:38:56 
2008 +0100
@@ -0,0 +1,70 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef _SYS_VDEV_IMPL_H
+#define        _SYS_VDEV_IMPL_H
+
+#define        VDEV_SKIP_SIZE          (8 << 10)
+#define        VDEV_BOOT_HEADER_SIZE   (8 << 10)
+#define        VDEV_PHYS_SIZE          (112 << 10)
+#define        VDEV_UBERBLOCK_RING     (128 << 10)
+
+/* ZFS boot block */
+#define        VDEV_BOOT_MAGIC         0x2f5b007b10cULL
+#define        VDEV_BOOT_VERSION       1               /* version number       
*/
+
+typedef struct vdev_boot_header {
+       uint64_t        vb_magic;               /* VDEV_BOOT_MAGIC      */
+       uint64_t        vb_version;             /* VDEV_BOOT_VERSION    */
+       uint64_t        vb_offset;              /* start offset (bytes) */
+       uint64_t        vb_size;                /* size (bytes)         */
+       char            vb_pad[VDEV_BOOT_HEADER_SIZE - 4 * sizeof (uint64_t)];
+} vdev_boot_header_t;
+
+typedef struct vdev_phys {
+       char            vp_nvlist[VDEV_PHYS_SIZE - sizeof (zio_block_tail_t)];
+       zio_block_tail_t vp_zbt;
+} vdev_phys_t;
+
+typedef struct vdev_label {
+       char            vl_pad[VDEV_SKIP_SIZE];                 /*   8K */
+       vdev_boot_header_t vl_boot_header;                      /*   8K */
+       vdev_phys_t     vl_vdev_phys;                           /* 112K */
+       char            vl_uberblock[VDEV_UBERBLOCK_RING];      /* 128K */
+} vdev_label_t;                                                        /* 256K 
total */
+
+/*
+ * Size and offset of embedded boot loader region on each label.
+ * The total size of the first two labels plus the boot area is 4MB.
+ */
+#define        VDEV_BOOT_OFFSET        (2 * sizeof (vdev_label_t))
+#define        VDEV_BOOT_SIZE          (7ULL << 19)                    /* 3.5M 
*/
+
+/*
+ * Size of label regions at the start and end of each leaf device.
+ */
+#define        VDEV_LABEL_START_SIZE   (2 * sizeof (vdev_label_t) + 
VDEV_BOOT_SIZE)
+#define        VDEV_LABEL_END_SIZE     (2 * sizeof (vdev_label_t))
+#define        VDEV_LABELS             4
+
+#endif /* _SYS_VDEV_IMPL_H */
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/zfs/zfs-include/zap_impl.h
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/zfs-include/zap_impl.h       Thu May 01 16:38:56 
2008 +0100
@@ -0,0 +1,110 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef        _SYS_ZAP_IMPL_H
+#define        _SYS_ZAP_IMPL_H
+
+#define        ZAP_MAGIC 0x2F52AB2ABULL
+
+#define        ZAP_HASHBITS            28
+#define        MZAP_ENT_LEN            64
+#define        MZAP_NAME_LEN           (MZAP_ENT_LEN - 8 - 4 - 2)
+#define        MZAP_MAX_BLKSHIFT       SPA_MAXBLOCKSHIFT
+#define        MZAP_MAX_BLKSZ          (1 << MZAP_MAX_BLKSHIFT)
+
+typedef struct mzap_ent_phys {
+       uint64_t mze_value;
+       uint32_t mze_cd;
+       uint16_t mze_pad;       /* in case we want to chain them someday */
+       char mze_name[MZAP_NAME_LEN];
+} mzap_ent_phys_t;
+
+typedef struct mzap_phys {
+       uint64_t mz_block_type; /* ZBT_MICRO */
+       uint64_t mz_salt;
+       uint64_t mz_pad[6];
+       mzap_ent_phys_t mz_chunk[1];
+       /* actually variable size depending on block size */
+} mzap_phys_t;
+
+/*
+ * The (fat) zap is stored in one object. It is an array of
+ * 1<<FZAP_BLOCK_SHIFT byte blocks. The layout looks like one of:
+ *
+ * ptrtbl fits in first block:
+ *     [zap_phys_t zap_ptrtbl_shift < 6] [zap_leaf_t] ...
+ *
+ * ptrtbl too big for first block:
+ *     [zap_phys_t zap_ptrtbl_shift >= 6] [zap_leaf_t] [ptrtbl] ...
+ *
+ */
+
+#define        ZBT_LEAF                ((1ULL << 63) + 0)
+#define        ZBT_HEADER              ((1ULL << 63) + 1)
+#define        ZBT_MICRO               ((1ULL << 63) + 3)
+/* any other values are ptrtbl blocks */
+
+/*
+ * the embedded pointer table takes up half a block:
+ * block size / entry size (2^3) / 2
+ */
+#define        ZAP_EMBEDDED_PTRTBL_SHIFT(zap) (FZAP_BLOCK_SHIFT(zap) - 3 - 1)
+
+/*
+ * The embedded pointer table starts half-way through the block.  Since
+ * the pointer table itself is half the block, it starts at (64-bit)
+ * word number (1<<ZAP_EMBEDDED_PTRTBL_SHIFT(zap)).
+ */
+#define        ZAP_EMBEDDED_PTRTBL_ENT(zap, idx) \
+       ((uint64_t *)(zap)->zap_f.zap_phys) \
+       [(idx) + (1<<ZAP_EMBEDDED_PTRTBL_SHIFT(zap))]
+
+/*
+ * TAKE NOTE:
+ * If zap_phys_t is modified, zap_byteswap() must be modified.
+ */
+typedef struct zap_phys {
+       uint64_t zap_block_type;        /* ZBT_HEADER */
+       uint64_t zap_magic;             /* ZAP_MAGIC */
+
+       struct zap_table_phys {
+               uint64_t zt_blk;        /* starting block number */
+               uint64_t zt_numblks;    /* number of blocks */
+               uint64_t zt_shift;      /* bits to index it */
+               uint64_t zt_nextblk;    /* next (larger) copy start block */
+               uint64_t zt_blks_copied; /* number source blocks copied */
+       } zap_ptrtbl;
+
+       uint64_t zap_freeblk;           /* the next free block */
+       uint64_t zap_num_leafs;         /* number of leafs */
+       uint64_t zap_num_entries;       /* number of entries */
+       uint64_t zap_salt;              /* salt to stir into hash function */
+       /*
+        * This structure is followed by padding, and then the embedded
+        * pointer table.  The embedded pointer table takes up second
+        * half of the block.  It is accessed using the
+        * ZAP_EMBEDDED_PTRTBL_ENT() macro.
+        */
+} zap_phys_t;
+
+#endif /* _SYS_ZAP_IMPL_H */
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/zfs/zfs-include/zap_leaf.h
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/zfs-include/zap_leaf.h       Thu May 01 16:38:56 
2008 +0100
@@ -0,0 +1,100 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef        _SYS_ZAP_LEAF_H
+#define        _SYS_ZAP_LEAF_H
+
+#define        ZAP_LEAF_MAGIC 0x2AB1EAF
+
+/* chunk size = 24 bytes */
+#define        ZAP_LEAF_CHUNKSIZE 24
+
+/*
+ * The amount of space within the chunk available for the array is:
+ * chunk size - space for type (1) - space for next pointer (2)
+ */
+#define        ZAP_LEAF_ARRAY_BYTES (ZAP_LEAF_CHUNKSIZE - 3)
+
+typedef enum zap_chunk_type {
+       ZAP_CHUNK_FREE = 253,
+       ZAP_CHUNK_ENTRY = 252,
+       ZAP_CHUNK_ARRAY = 251,
+       ZAP_CHUNK_TYPE_MAX = 250
+} zap_chunk_type_t;
+
+/*
+ * TAKE NOTE:
+ * If zap_leaf_phys_t is modified, zap_leaf_byteswap() must be modified.
+ */
+typedef struct zap_leaf_phys {
+       struct zap_leaf_header {
+               uint64_t lh_block_type;         /* ZBT_LEAF */
+               uint64_t lh_pad1;
+               uint64_t lh_prefix;             /* hash prefix of this leaf */
+               uint32_t lh_magic;              /* ZAP_LEAF_MAGIC */
+               uint16_t lh_nfree;              /* number free chunks */
+               uint16_t lh_nentries;           /* number of entries */
+               uint16_t lh_prefix_len;         /* num bits used to id this */
+
+/* above is accessable to zap, below is zap_leaf private */
+
+               uint16_t lh_freelist;           /* chunk head of free list */
+               uint8_t lh_pad2[12];
+       } l_hdr; /* 2 24-byte chunks */
+
+       /*
+        * The header is followed by a hash table with
+        * ZAP_LEAF_HASH_NUMENTRIES(zap) entries.  The hash table is
+        * followed by an array of ZAP_LEAF_NUMCHUNKS(zap)
+        * zap_leaf_chunk structures.  These structures are accessed
+        * with the ZAP_LEAF_CHUNK() macro.
+        */
+
+       uint16_t l_hash[1];
+} zap_leaf_phys_t;
+
+typedef union zap_leaf_chunk {
+       struct zap_leaf_entry {
+               uint8_t le_type;                /* always ZAP_CHUNK_ENTRY */
+               uint8_t le_int_size;            /* size of ints */
+               uint16_t le_next;               /* next entry in hash chain */
+               uint16_t le_name_chunk;         /* first chunk of the name */
+               uint16_t le_name_length;        /* bytes in name, incl null */
+               uint16_t le_value_chunk;        /* first chunk of the value */
+               uint16_t le_value_length;       /* value length in ints */
+               uint32_t le_cd;                 /* collision differentiator */
+               uint64_t le_hash;               /* hash value of the name */
+       } l_entry;
+       struct zap_leaf_array {
+               uint8_t la_type;                /* always ZAP_CHUNK_ARRAY */
+               uint8_t la_array[ZAP_LEAF_ARRAY_BYTES];
+               uint16_t la_next;               /* next blk or CHAIN_END */
+       } l_array;
+       struct zap_leaf_free {
+               uint8_t lf_type;                /* always ZAP_CHUNK_FREE */
+               uint8_t lf_pad[ZAP_LEAF_ARRAY_BYTES];
+               uint16_t lf_next;       /* next in free list, or CHAIN_END */
+       } l_free;
+} zap_leaf_chunk_t;
+
+#endif /* _SYS_ZAP_LEAF_H */
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/zfs/zfs-include/zfs.h
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/zfs-include/zfs.h    Thu May 01 16:38:56 2008 +0100
@@ -0,0 +1,112 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2008 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef        _SYS_FS_ZFS_H
+#define        _SYS_FS_ZFS_H
+
+
+/*
+ * On-disk version number.
+ */
+#define        SPA_VERSION_1                   1ULL
+#define        SPA_VERSION_2                   2ULL
+#define        SPA_VERSION_3                   3ULL
+#define        SPA_VERSION_4                   4ULL
+#define        SPA_VERSION_5                   5ULL
+#define        SPA_VERSION_6                   6ULL
+#define        SPA_VERSION_7                   7ULL
+#define        SPA_VERSION_8                   8ULL
+#define        SPA_VERSION_9                   9ULL
+#define        SPA_VERSION_10                  10ULL
+#define        SPA_VERSION                     SPA_VERSION_10
+
+/*
+ * The following are configuration names used in the nvlist describing a pool's
+ * configuration.
+ */
+#define        ZPOOL_CONFIG_VERSION            "version"
+#define        ZPOOL_CONFIG_POOL_NAME          "name"
+#define        ZPOOL_CONFIG_POOL_STATE         "state"
+#define        ZPOOL_CONFIG_POOL_TXG           "txg"
+#define        ZPOOL_CONFIG_POOL_GUID          "pool_guid"
+#define        ZPOOL_CONFIG_CREATE_TXG         "create_txg"
+#define        ZPOOL_CONFIG_TOP_GUID           "top_guid"
+#define        ZPOOL_CONFIG_VDEV_TREE          "vdev_tree"
+#define        ZPOOL_CONFIG_TYPE               "type"
+#define        ZPOOL_CONFIG_CHILDREN           "children"
+#define        ZPOOL_CONFIG_ID                 "id"
+#define        ZPOOL_CONFIG_GUID               "guid"
+#define        ZPOOL_CONFIG_PATH               "path"
+#define        ZPOOL_CONFIG_DEVID              "devid"
+#define        ZPOOL_CONFIG_METASLAB_ARRAY     "metaslab_array"
+#define        ZPOOL_CONFIG_METASLAB_SHIFT     "metaslab_shift"
+#define        ZPOOL_CONFIG_ASHIFT             "ashift"
+#define        ZPOOL_CONFIG_ASIZE              "asize"
+#define        ZPOOL_CONFIG_DTL                "DTL"
+#define        ZPOOL_CONFIG_STATS              "stats"
+#define        ZPOOL_CONFIG_WHOLE_DISK         "whole_disk"
+#define        ZPOOL_CONFIG_ERRCOUNT           "error_count"
+#define        ZPOOL_CONFIG_NOT_PRESENT        "not_present"
+#define        ZPOOL_CONFIG_SPARES             "spares"
+#define        ZPOOL_CONFIG_IS_SPARE           "is_spare"
+#define        ZPOOL_CONFIG_NPARITY            "nparity"
+#define        ZPOOL_CONFIG_PHYS_PATH          "phys_path"
+#define        ZPOOL_CONFIG_L2CACHE            "l2cache"
+/*
+ * The persistent vdev state is stored as separate values rather than a single
+ * 'vdev_state' entry.  This is because a device can be in multiple states, 
such
+ * as offline and degraded.
+ */
+#define        ZPOOL_CONFIG_OFFLINE            "offline"
+#define        ZPOOL_CONFIG_FAULTED            "faulted"
+#define        ZPOOL_CONFIG_DEGRADED           "degraded"
+#define        ZPOOL_CONFIG_REMOVED            "removed"
+
+#define        VDEV_TYPE_ROOT                  "root"
+#define        VDEV_TYPE_MIRROR                "mirror"
+#define        VDEV_TYPE_REPLACING             "replacing"
+#define        VDEV_TYPE_RAIDZ                 "raidz"
+#define        VDEV_TYPE_DISK                  "disk"
+#define        VDEV_TYPE_FILE                  "file"
+#define        VDEV_TYPE_MISSING               "missing"
+#define        VDEV_TYPE_SPARE                 "spare"
+#define        VDEV_TYPE_L2CACHE               "l2cache"
+
+/*
+ * pool state.  The following states are written to disk as part of the normal
+ * SPA lifecycle: ACTIVE, EXPORTED, DESTROYED, SPARE, L2CACHE.  The remaining
+ * states are software abstractions used at various levels to communicate pool
+ * state.
+ */
+typedef enum pool_state {
+       POOL_STATE_ACTIVE = 0,          /* In active use                */
+       POOL_STATE_EXPORTED,            /* Explicitly exported          */
+       POOL_STATE_DESTROYED,           /* Explicitly destroyed         */
+       POOL_STATE_SPARE,               /* Reserved for hot spare use   */
+       POOL_STATE_L2CACHE,             /* Level 2 ARC device           */
+       POOL_STATE_UNINITIALIZED,       /* Internal spa_t state         */
+       POOL_STATE_UNAVAIL,             /* Internal libzfs state        */
+       POOL_STATE_POTENTIALLY_ACTIVE   /* Internal libzfs state        */
+} pool_state_t;
+
+#endif /* _SYS_FS_ZFS_H */
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/zfs/zfs-include/zfs_acl.h
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/zfs-include/zfs_acl.h        Thu May 01 16:38:56 
2008 +0100
@@ -0,0 +1,60 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef        _SYS_FS_ZFS_ACL_H
+#define        _SYS_FS_ZFS_ACL_H
+
+#ifndef _UID_T
+#define        _UID_T
+typedef        unsigned int uid_t;                     /* UID type */
+#endif /* _UID_T */
+
+typedef struct zfs_oldace {
+       uint32_t        z_fuid;         /* "who" */
+       uint32_t        z_access_mask;  /* access mask */
+       uint16_t        z_flags;        /* flags, i.e inheritance */
+       uint16_t        z_type;         /* type of entry allow/deny */
+} zfs_oldace_t;
+
+#define        ACE_SLOT_CNT    6
+
+typedef struct zfs_znode_acl_v0 {
+       uint64_t        z_acl_extern_obj;         /* ext acl pieces */
+       uint32_t        z_acl_count;              /* Number of ACEs */
+       uint16_t        z_acl_version;            /* acl version */
+       uint16_t        z_acl_pad;                /* pad */
+       zfs_oldace_t    z_ace_data[ACE_SLOT_CNT]; /* 6 standard ACEs */
+} zfs_znode_acl_v0_t;
+
+#define        ZFS_ACE_SPACE   (sizeof (zfs_oldace_t) * ACE_SLOT_CNT)
+
+typedef struct zfs_znode_acl {
+       uint64_t        z_acl_extern_obj;         /* ext acl pieces */
+       uint32_t        z_acl_size;               /* Number of bytes in ACL */
+       uint16_t        z_acl_version;            /* acl version */
+       uint16_t        z_acl_count;              /* ace count */
+       uint8_t         z_ace_data[ZFS_ACE_SPACE]; /* space for embedded ACEs */
+} zfs_znode_acl_t;
+
+
+#endif /* _SYS_FS_ZFS_ACL_H */
diff -r 8ccaf43782bd -r 64f790e90d3d 
tools/libfsimage/zfs/zfs-include/zfs_znode.h
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/zfs-include/zfs_znode.h      Thu May 01 16:38:56 
2008 +0100
@@ -0,0 +1,68 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef        _SYS_FS_ZFS_ZNODE_H
+#define        _SYS_FS_ZFS_ZNODE_H
+
+#define        MASTER_NODE_OBJ 1
+#define        ZFS_ROOT_OBJ            "ROOT"
+#define        ZPL_VERSION_STR         "VERSION"
+
+#define        ZPL_VERSION             3ULL
+
+#define        ZFS_DIRENT_OBJ(de) BF64_GET(de, 0, 48)
+
+/*
+ * This is the persistent portion of the znode.  It is stored
+ * in the "bonus buffer" of the file.  Short symbolic links
+ * are also stored in the bonus buffer.
+ */
+typedef struct znode_phys {
+       uint64_t zp_atime[2];           /*  0 - last file access time */
+       uint64_t zp_mtime[2];           /* 16 - last file modification time */
+       uint64_t zp_ctime[2];           /* 32 - last file change time */
+       uint64_t zp_crtime[2];          /* 48 - creation time */
+       uint64_t zp_gen;                /* 64 - generation (txg of creation) */
+       uint64_t zp_mode;               /* 72 - file mode bits */
+       uint64_t zp_size;               /* 80 - size of file */
+       uint64_t zp_parent;             /* 88 - directory parent (`..') */
+       uint64_t zp_links;              /* 96 - number of links to file */
+       uint64_t zp_xattr;              /* 104 - DMU object for xattrs */
+       uint64_t zp_rdev;               /* 112 - dev_t for VBLK & VCHR files */
+       uint64_t zp_flags;              /* 120 - persistent flags */
+       uint64_t zp_uid;                /* 128 - file owner */
+       uint64_t zp_gid;                /* 136 - owning group */
+       uint64_t zp_pad[4];             /* 144 - future */
+       zfs_znode_acl_t zp_acl;         /* 176 - 263 ACL */
+       /*
+        * Data may pad out any remaining bytes in the znode buffer, eg:
+        *
+        * |<---------------------- dnode_phys (512) ------------------------>|
+        * |<-- dnode (192) --->|<----------- "bonus" buffer (320) ---------->|
+        *                      |<---- znode (264) ---->|<---- data (56) ---->|
+        *
+        * At present, we only use this space to store symbolic links.
+        */
+} znode_phys_t;
+
+#endif /* _SYS_FS_ZFS_ZNODE_H */
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/zfs/zfs-include/zil.h
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/zfs-include/zil.h    Thu May 01 16:38:56 2008 +0100
@@ -0,0 +1,51 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef        _SYS_ZIL_H
+#define        _SYS_ZIL_H
+
+/*
+ * Intent log format:
+ *
+ * Each objset has its own intent log.  The log header (zil_header_t)
+ * for objset N's intent log is kept in the Nth object of the SPA's
+ * intent_log objset.  The log header points to a chain of log blocks,
+ * each of which contains log records (i.e., transactions) followed by
+ * a log block trailer (zil_trailer_t).  The format of a log record
+ * depends on the record (or transaction) type, but all records begin
+ * with a common structure that defines the type, length, and txg.
+ */
+
+/*
+ * Intent log header - this on disk structure holds fields to manage
+ * the log.  All fields are 64 bit to easily handle cross architectures.
+ */
+typedef struct zil_header {
+       uint64_t zh_claim_txg;  /* txg in which log blocks were claimed */
+       uint64_t zh_replay_seq; /* highest replayed sequence number */
+       blkptr_t zh_log;        /* log chain */
+       uint64_t zh_claim_seq;  /* highest claimed sequence number */
+       uint64_t zh_pad[5];
+} zil_header_t;
+
+#endif /* _SYS_ZIL_H */
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/zfs/zfs-include/zio.h
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/zfs-include/zio.h    Thu May 01 16:38:56 2008 +0100
@@ -0,0 +1,81 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef _ZIO_H
+#define        _ZIO_H
+
+#define        ZBT_MAGIC       0x210da7ab10c7a11ULL    /* zio data bloc tail */
+
+typedef struct zio_block_tail {
+       uint64_t        zbt_magic;      /* for validation, endianness   */
+       zio_cksum_t     zbt_cksum;      /* 256-bit checksum             */
+} zio_block_tail_t;
+
+/*
+ * Gang block headers are self-checksumming and contain an array
+ * of block pointers.
+ */
+#define        SPA_GANGBLOCKSIZE       SPA_MINBLOCKSIZE
+#define        SPA_GBH_NBLKPTRS        ((SPA_GANGBLOCKSIZE - \
+       sizeof (zio_block_tail_t)) / sizeof (blkptr_t))
+#define        SPA_GBH_FILLER          ((SPA_GANGBLOCKSIZE - \
+       sizeof (zio_block_tail_t) - \
+       (SPA_GBH_NBLKPTRS * sizeof (blkptr_t))) /\
+       sizeof (uint64_t))
+
+#define        ZIO_GET_IOSIZE(zio)     \
+       (BP_IS_GANG((zio)->io_bp) ? \
+       SPA_GANGBLOCKSIZE : BP_GET_PSIZE((zio)->io_bp))
+
+typedef struct zio_gbh {
+       blkptr_t                zg_blkptr[SPA_GBH_NBLKPTRS];
+       uint64_t                zg_filler[SPA_GBH_FILLER];
+       zio_block_tail_t        zg_tail;
+} zio_gbh_phys_t;
+
+enum zio_checksum {
+       ZIO_CHECKSUM_INHERIT = 0,
+       ZIO_CHECKSUM_ON,
+       ZIO_CHECKSUM_OFF,
+       ZIO_CHECKSUM_LABEL,
+       ZIO_CHECKSUM_GANG_HEADER,
+       ZIO_CHECKSUM_ZILOG,
+       ZIO_CHECKSUM_FLETCHER_2,
+       ZIO_CHECKSUM_FLETCHER_4,
+       ZIO_CHECKSUM_SHA256,
+       ZIO_CHECKSUM_FUNCTIONS
+};
+
+#define        ZIO_CHECKSUM_ON_VALUE   ZIO_CHECKSUM_FLETCHER_2
+#define        ZIO_CHECKSUM_DEFAULT    ZIO_CHECKSUM_ON
+
+enum zio_compress {
+       ZIO_COMPRESS_INHERIT = 0,
+       ZIO_COMPRESS_ON,
+       ZIO_COMPRESS_OFF,
+       ZIO_COMPRESS_LZJB,
+       ZIO_COMPRESS_EMPTY,
+       ZIO_COMPRESS_FUNCTIONS
+};
+
+#endif /* _ZIO_H */
diff -r 8ccaf43782bd -r 64f790e90d3d 
tools/libfsimage/zfs/zfs-include/zio_checksum.h
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/zfs-include/zio_checksum.h   Thu May 01 16:38:56 
2008 +0100
@@ -0,0 +1,42 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#ifndef _SYS_ZIO_CHECKSUM_H
+#define        _SYS_ZIO_CHECKSUM_H
+
+/*
+ * Signature for checksum functions.
+ */
+typedef void zio_checksum_t(const void *data, uint64_t size, zio_cksum_t *zcp);
+
+/*
+ * Information about each checksum function.
+ */
+typedef struct zio_checksum_info {
+       zio_checksum_t  *ci_func[2]; /* checksum function for each byteorder */
+       int             ci_correctable; /* number of correctable bits   */
+       int             ci_zbt;         /* uses zio block tail? */
+       char            *ci_name;       /* descriptive name */
+} zio_checksum_info_t;
+
+#endif /* _SYS_ZIO_CHECKSUM_H */
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/zfs/zfs_fletcher.c
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/zfs_fletcher.c       Thu May 01 16:38:56 2008 +0100
@@ -0,0 +1,93 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#include "fsys_zfs.h"
+
+
+void
+fletcher_2_native(const void *buf, uint64_t size, zio_cksum_t *zcp)
+{
+       const uint64_t *ip = buf;
+       const uint64_t *ipend = ip + (size / sizeof (uint64_t));
+       uint64_t a0, b0, a1, b1;
+
+       for (a0 = b0 = a1 = b1 = 0; ip < ipend; ip += 2) {
+               a0 += ip[0];
+               a1 += ip[1];
+               b0 += a0;
+               b1 += a1;
+       }
+
+       ZIO_SET_CHECKSUM(zcp, a0, a1, b0, b1);
+}
+
+void
+fletcher_2_byteswap(const void *buf, uint64_t size, zio_cksum_t *zcp)
+{
+       const uint64_t *ip = buf;
+       const uint64_t *ipend = ip + (size / sizeof (uint64_t));
+       uint64_t a0, b0, a1, b1;
+
+       for (a0 = b0 = a1 = b1 = 0; ip < ipend; ip += 2) {
+               a0 += BSWAP_64(ip[0]);
+               a1 += BSWAP_64(ip[1]);
+               b0 += a0;
+               b1 += a1;
+       }
+
+       ZIO_SET_CHECKSUM(zcp, a0, a1, b0, b1);
+}
+
+void
+fletcher_4_native(const void *buf, uint64_t size, zio_cksum_t *zcp)
+{
+       const uint32_t *ip = buf;
+       const uint32_t *ipend = ip + (size / sizeof (uint32_t));
+       uint64_t a, b, c, d;
+
+       for (a = b = c = d = 0; ip < ipend; ip++) {
+               a += ip[0];
+               b += a;
+               c += b;
+               d += c;
+       }
+
+       ZIO_SET_CHECKSUM(zcp, a, b, c, d);
+}
+
+void
+fletcher_4_byteswap(const void *buf, uint64_t size, zio_cksum_t *zcp)
+{
+       const uint32_t *ip = buf;
+       const uint32_t *ipend = ip + (size / sizeof (uint32_t));
+       uint64_t a, b, c, d;
+
+       for (a = b = c = d = 0; ip < ipend; ip++) {
+               a += BSWAP_32(ip[0]);
+               b += a;
+               c += b;
+               d += c;
+       }
+
+       ZIO_SET_CHECKSUM(zcp, a, b, c, d);
+}
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/zfs/zfs_lzjb.c
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/zfs_lzjb.c   Thu May 01 16:38:56 2008 +0100
@@ -0,0 +1,60 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#include "fsys_zfs.h"
+
+#define        MATCH_BITS      6
+#define        MATCH_MIN       3
+#define        OFFSET_MASK     ((1 << (16 - MATCH_BITS)) - 1)
+
+
+/*ARGSUSED*/
+int
+lzjb_decompress(void *s_start, void *d_start, size_t s_len, size_t d_len)
+{
+       unsigned char *src = s_start;
+       unsigned char *dst = d_start;
+       unsigned char *d_end = (unsigned char *)d_start + d_len;
+       unsigned char *cpy;
+       unsigned char copymap = '\0';
+       int copymask = 1 << (NBBY - 1);
+
+       while (dst < d_end) {
+               if ((copymask <<= 1) == (1 << NBBY)) {
+                       copymask = 1;
+                       copymap = *src++;
+               }
+               if (copymap & (unsigned char)copymask) {
+                       int mlen = (src[0] >> (NBBY - MATCH_BITS)) + MATCH_MIN;
+                       int offset = ((src[0] << NBBY) | src[1]) & OFFSET_MASK;
+                       src += 2;
+                       if ((cpy = dst - offset) < (unsigned char *)d_start)
+                               return (-1);
+                       while (--mlen >= 0 && dst < d_end)
+                               *dst++ = *cpy++;
+               } else {
+                       *dst++ = *src++;
+               }
+       }
+       return (0);
+}
diff -r 8ccaf43782bd -r 64f790e90d3d tools/libfsimage/zfs/zfs_sha256.c
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libfsimage/zfs/zfs_sha256.c Thu May 01 16:38:56 2008 +0100
@@ -0,0 +1,124 @@
+/*
+ *  GRUB  --  GRand Unified Bootloader
+ *  Copyright (C) 1999,2000,2001,2002,2003,2004  Free Software Foundation, Inc.
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will 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 to the Free Software
+ *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ */
+/*
+ * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
+ * Use is subject to license terms.
+ */
+
+#include "fsys_zfs.h"
+
+/*
+ * SHA-256 checksum, as specified in FIPS 180-2, available at:
+ * http://csrc.nist.gov/cryptval
+ *
+ * This is a very compact implementation of SHA-256.
+ * It is designed to be simple and portable, not to be fast.
+ */
+
+/*
+ * The literal definitions according to FIPS180-2 would be:
+ *
+ *     Ch(x, y, z)     (((x) & (y)) ^ ((~(x)) & (z)))
+ *     Maj(x, y, z)    (((x) & (y)) | ((x) & (z)) | ((y) & (z)))
+ *
+ * We use logical equivalents which require one less op.
+ */
+#define        Ch(x, y, z)     ((z) ^ ((x) & ((y) ^ (z))))
+#define        Maj(x, y, z)    (((x) & (y)) ^ ((z) & ((x) ^ (y))))
+#define        Rot32(x, s)     (((x) >> s) | ((x) << (32 - s)))
+#define        SIGMA0(x)       (Rot32(x, 2) ^ Rot32(x, 13) ^ Rot32(x, 22))
+#define        SIGMA1(x)       (Rot32(x, 6) ^ Rot32(x, 11) ^ Rot32(x, 25))
+#define        sigma0(x)       (Rot32(x, 7) ^ Rot32(x, 18) ^ ((x) >> 3))
+#define        sigma1(x)       (Rot32(x, 17) ^ Rot32(x, 19) ^ ((x) >> 10))
+
+static const uint32_t SHA256_K[64] = {
+       0x428a2f98, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5,
+       0x3956c25b, 0x59f111f1, 0x923f82a4, 0xab1c5ed5,
+       0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3,
+       0x72be5d74, 0x80deb1fe, 0x9bdc06a7, 0xc19bf174,
+       0xe49b69c1, 0xefbe4786, 0x0fc19dc6, 0x240ca1cc,
+       0x2de92c6f, 0x4a7484aa, 0x5cb0a9dc, 0x76f988da,
+       0x983e5152, 0xa831c66d, 0xb00327c8, 0xbf597fc7,
+       0xc6e00bf3, 0xd5a79147, 0x06ca6351, 0x14292967,
+       0x27b70a85, 0x2e1b2138, 0x4d2c6dfc, 0x53380d13,
+       0x650a7354, 0x766a0abb, 0x81c2c92e, 0x92722c85,
+       0xa2bfe8a1, 0xa81a664b, 0xc24b8b70, 0xc76c51a3,
+       0xd192e819, 0xd6990624, 0xf40e3585, 0x106aa070,
+       0x19a4c116, 0x1e376c08, 0x2748774c, 0x34b0bcb5,
+       0x391c0cb3, 0x4ed8aa4a, 0x5b9cca4f, 0x682e6ff3,
+       0x748f82ee, 0x78a5636f, 0x84c87814, 0x8cc70208,
+       0x90befffa, 0xa4506ceb, 0xbef9a3f7, 0xc67178f2
+};
+
+static void
+SHA256Transform(uint32_t *H, const uint8_t *cp)
+{
+       uint32_t a, b, c, d, e, f, g, h, t, T1, T2, W[64];
+
+       for (t = 0; t < 16; t++, cp += 4)
+               W[t] = (cp[0] << 24) | (cp[1] << 16) | (cp[2] << 8) | cp[3];
+
+       for (t = 16; t < 64; t++)
+               W[t] = sigma1(W[t - 2]) + W[t - 7] +
+                   sigma0(W[t - 15]) + W[t - 16];
+
+       a = H[0]; b = H[1]; c = H[2]; d = H[3];
+       e = H[4]; f = H[5]; g = H[6]; h = H[7];
+
+       for (t = 0; t < 64; t++) {
+               T1 = h + SIGMA1(e) + Ch(e, f, g) + SHA256_K[t] + W[t];
+               T2 = SIGMA0(a) + Maj(a, b, c);
+               h = g; g = f; f = e; e = d + T1;
+               d = c; c = b; b = a; a = T1 + T2;
+       }
+
+       H[0] += a; H[1] += b; H[2] += c; H[3] += d;
+       H[4] += e; H[5] += f; H[6] += g; H[7] += h;
+}
+
+void
+zio_checksum_SHA256(const void *buf, uint64_t size, zio_cksum_t *zcp)
+{
+       uint32_t H[8] = { 0x6a09e667, 0xbb67ae85, 0x3c6ef372, 0xa54ff53a,
+           0x510e527f, 0x9b05688c, 0x1f83d9ab, 0x5be0cd19 };
+       uint8_t pad[128];
+       int padsize = size & 63;
+       int i;
+
+       for (i = 0; i < size - padsize; i += 64)
+               SHA256Transform(H, (uint8_t *)buf + i);
+
+       for (i = 0; i < padsize; i++)
+               pad[i] = ((uint8_t *)buf)[i];
+
+       for (pad[padsize++] = 0x80; (padsize & 63) != 56; padsize++)
+               pad[padsize] = 0;
+
+       for (i = 0; i < 8; i++)
+               pad[padsize++] = (size << 3) >> (56 - 8 * i);
+
+       for (i = 0; i < padsize; i += 64)
+               SHA256Transform(H, pad + i);
+
+       ZIO_SET_CHECKSUM(zcp,
+           (uint64_t)H[0] << 32 | H[1],
+           (uint64_t)H[2] << 32 | H[3],
+           (uint64_t)H[4] << 32 | H[5],
+           (uint64_t)H[6] << 32 | H[7]);
+}
diff -r 8ccaf43782bd -r 64f790e90d3d tools/pygrub/src/fsimage/fsimage.c
--- a/tools/pygrub/src/fsimage/fsimage.c        Thu May 01 16:37:46 2008 +0100
+++ b/tools/pygrub/src/fsimage/fsimage.c        Thu May 01 16:38:56 2008 +0100
@@ -17,7 +17,7 @@
  * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
  * DEALINGS IN THE SOFTWARE.
  *
- * Copyright 2006 Sun Microsystems, Inc.  All rights reserved.
+ * Copyright 2008 Sun Microsystems, Inc.  All rights reserved.
  * Use is subject to license terms.
  */
 
@@ -281,6 +281,22 @@ fsimage_open(PyObject *o, PyObject *args
        return (PyObject *)fs;
 }
 
+static PyObject *
+fsimage_getbootstring(PyObject *o, PyObject *args)
+{
+       PyObject *fs;
+       char    *bootstring;
+       fsi_t   *fsi;
+
+       if (!PyArg_ParseTuple(args, "O", &fs))
+               return (NULL);
+
+       fsi = ((fsimage_fs_t *)fs)->fs;
+       bootstring = fsi_fs_bootstring(fsi);
+
+       return Py_BuildValue("s", bootstring);
+}
+
 PyDoc_STRVAR(fsimage_open__doc__,
     "open(name, [offset=off]) - Open the given file as a filesystem image.\n"
     "\n"
@@ -288,9 +304,15 @@ PyDoc_STRVAR(fsimage_open__doc__,
     "offset - offset of file system within file image.\n"
     "options - mount options string.\n");
 
+PyDoc_STRVAR(fsimage_getbootstring__doc__,
+    "getbootstring(fs) - Return the boot string needed for this file system "
+    "or NULL if none is needed.\n");
+
 static struct PyMethodDef fsimage_module_methods[] = {
        { "open", (PyCFunction)fsimage_open,
            METH_VARARGS|METH_KEYWORDS, fsimage_open__doc__ },
+       { "getbootstring", (PyCFunction)fsimage_getbootstring,
+           METH_VARARGS, fsimage_getbootstring__doc__ },
        { NULL, NULL, 0, NULL }
 };
 
diff -r 8ccaf43782bd -r 64f790e90d3d tools/pygrub/src/pygrub
--- a/tools/pygrub/src/pygrub   Thu May 01 16:37:46 2008 +0100
+++ b/tools/pygrub/src/pygrub   Thu May 01 16:38:56 2008 +0100
@@ -646,7 +646,13 @@ if __name__ == "__main__":
         print "  args: %s" % chosencfg["args"]
         sys.exit(0)
 
-    fs = fsimage.open(file, get_fs_offset(file))
+    # if boot filesystem is set then pass to fsimage.open
+    bootfsargs = '"%s"' % incfg["args"]
+    bootfsgroup = re.findall('zfs-bootfs=(.*?)[\s\,\"]', bootfsargs)
+    if bootfsgroup:
+        fs = fsimage.open(file, get_fs_offset(file), bootfsgroup[0])
+    else:
+        fs = fsimage.open(file, get_fs_offset(file))
 
     chosencfg = sniff_solaris(fs, incfg)
 
@@ -672,7 +678,15 @@ if __name__ == "__main__":
     if bootcfg["ramdisk"]:
         sxp += "(ramdisk %s)" % bootcfg["ramdisk"]
     if chosencfg["args"]:
-        sxp += "(args \"%s\")" % chosencfg["args"]
+        zfsinfo = fsimage.getbootstring(fs)
+        if zfsinfo is None:
+            sxp += "(args \"%s\")" % chosencfg["args"]
+        else:
+            e = re.compile("zfs-bootfs=[\w\-\.\:@/]+" )
+            (chosencfg["args"],count) = e.subn(zfsinfo, chosencfg["args"])
+            if count == 0:
+               chosencfg["args"] += " -B %s" % zfsinfo
+            sxp += "(args \"%s\")" % (chosencfg["args"])
 
     sys.stdout.flush()
     os.write(fd, sxp)

_______________________________________________
Xen-changelog mailing list
Xen-changelog@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-changelog


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.