Btrfs-progs: make send/receive compatible with older kernels
[platform/upstream/btrfs-progs.git] / volumes.c
index 3e52d06..8c45851 100644 (file)
--- a/volumes.c
+++ b/volumes.c
@@ -29,6 +29,7 @@
 #include "transaction.h"
 #include "print-tree.h"
 #include "volumes.h"
+#include "math.h"
 
 struct stripe {
        struct btrfs_device *dev;
@@ -52,9 +53,6 @@ static inline int nr_data_stripes(struct map_lookup *map)
 
 #define is_parity_stripe(x) ( ((x) == BTRFS_RAID5_P_STRIPE) || ((x) == BTRFS_RAID6_Q_STRIPE) )
 
-#define map_lookup_size(n) (sizeof(struct map_lookup) + \
-                           (sizeof(struct btrfs_bio_stripe) * (n)))
-
 static LIST_HEAD(fs_uuids);
 
 static struct btrfs_device *__find_device(struct list_head *head, u64 devid,
@@ -116,6 +114,7 @@ static int device_list_add(const char *path,
                        /* we can safely leave the fs_devices entry around */
                        return -ENOMEM;
                }
+               device->fd = -1;
                device->devid = devid;
                memcpy(device->uuid, disk_super->dev_item.uuid,
                       BTRFS_UUID_SIZE);
@@ -125,6 +124,11 @@ static int device_list_add(const char *path,
                        return -ENOMEM;
                }
                device->label = kstrdup(disk_super->label, GFP_NOFS);
+               if (!device->label) {
+                       kfree(device->name);
+                       kfree(device);
+                       return -ENOMEM;
+               }
                device->total_devs = btrfs_super_num_devices(disk_super);
                device->super_bytes_used = btrfs_super_bytes_used(disk_super);
                device->total_bytes =
@@ -161,8 +165,13 @@ int btrfs_close_devices(struct btrfs_fs_devices *fs_devices)
 again:
        list_for_each(cur, &fs_devices->devices) {
                device = list_entry(cur, struct btrfs_device, dev_list);
-               close(device->fd);
-               device->fd = -1;
+               if (device->fd != -1) {
+                       fsync(device->fd);
+                       if (posix_fadvise(device->fd, 0, 0, POSIX_FADV_DONTNEED))
+                               fprintf(stderr, "Warning, could not drop caches\n");
+                       close(device->fd);
+                       device->fd = -1;
+               }
                device->writeable = 0;
        }
 
@@ -186,6 +195,10 @@ int btrfs_open_devices(struct btrfs_fs_devices *fs_devices, int flags)
 
        list_for_each(cur, head) {
                device = list_entry(cur, struct btrfs_device, dev_list);
+               if (!device->name) {
+                       printk("no name for device %llu, skip it now\n", device->devid);
+                       continue;
+               }
 
                fd = open(device->name, flags);
                if (fd < 0) {
@@ -201,7 +214,7 @@ int btrfs_open_devices(struct btrfs_fs_devices *fs_devices, int flags)
                if (device->devid == fs_devices->lowest_devid)
                        fs_devices->lowest_bdev = fd;
                device->fd = fd;
-               if (flags == O_RDWR)
+               if (flags & O_RDWR)
                        device->writeable = 1;
        }
        return 0;
@@ -218,7 +231,6 @@ int btrfs_scan_one_device(int fd, const char *path,
        char *buf;
        int ret;
        u64 devid;
-       char uuidbuf[37];
 
        buf = malloc(4096);
        if (!buf) {
@@ -231,12 +243,11 @@ int btrfs_scan_one_device(int fd, const char *path,
                ret = -EIO;
                goto error_brelse;
        }
-       devid = le64_to_cpu(disk_super->dev_item.devid);
+       devid = btrfs_stack_device_id(&disk_super->dev_item);
        if (btrfs_super_flags(disk_super) & BTRFS_SUPER_FLAG_METADUMP)
                *total_devs = 1;
        else
                *total_devs = btrfs_super_num_devices(disk_super);
-       uuid_unparse(disk_super->fsid, uuidbuf);
 
        ret = device_list_add(path, disk_super, devid, fs_devices_ret);
 
@@ -261,7 +272,7 @@ static int find_free_dev_extent(struct btrfs_trans_handle *trans,
        struct btrfs_dev_extent *dev_extent = NULL;
        u64 hole_size = 0;
        u64 last_byte = 0;
-       u64 search_start = 0;
+       u64 search_start = root->fs_info->alloc_start;
        u64 search_end = device->total_bytes;
        int ret;
        int slot = 0;
@@ -276,10 +287,12 @@ static int find_free_dev_extent(struct btrfs_trans_handle *trans,
        /* we don't want to overwrite the superblock on the drive,
         * so we make sure to start at an offset of at least 1MB
         */
-       search_start = max((u64)1024 * 1024, search_start);
+       search_start = max(BTRFS_BLOCK_RESERVED_1M_FOR_SUPER, search_start);
 
-       if (root->fs_info->alloc_start + num_bytes <= device->total_bytes)
-               search_start = max(root->fs_info->alloc_start, search_start);
+       if (search_start >= search_end) {
+               ret = -ENOSPC;
+               goto error;
+       }
 
        key.objectid = device->devid;
        key.offset = search_start;
@@ -353,7 +366,7 @@ check_pending:
        /* we have to make sure we didn't find an extent that has already
         * been allocated by the map tree or the original allocation
         */
-       btrfs_release_path(root, path);
+       btrfs_release_path(path);
        BUG_ON(*start < search_start);
 
        if (*start + num_bytes > search_end) {
@@ -364,15 +377,15 @@ check_pending:
        return 0;
 
 error:
-       btrfs_release_path(root, path);
+       btrfs_release_path(path);
        return ret;
 }
 
-int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans,
-                          struct btrfs_device *device,
-                          u64 chunk_tree, u64 chunk_objectid,
-                          u64 chunk_offset,
-                          u64 num_bytes, u64 *start)
+static int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans,
+                                 struct btrfs_device *device,
+                                 u64 chunk_tree, u64 chunk_objectid,
+                                 u64 chunk_offset,
+                                 u64 num_bytes, u64 *start)
 {
        int ret;
        struct btrfs_path *path;
@@ -485,7 +498,7 @@ static int find_next_devid(struct btrfs_root *root, struct btrfs_path *path,
        }
        ret = 0;
 error:
-       btrfs_release_path(root, path);
+       btrfs_release_path(path);
        return ret;
 }
 
@@ -623,14 +636,6 @@ int btrfs_add_system_chunk(struct btrfs_trans_handle *trans,
        return 0;
 }
 
-static u64 div_factor(u64 num, int factor)
-{
-       if (factor == 10)
-               return num;
-       num *= factor;
-       return num / 10;
-}
-
 static u64 chunk_bytes_by_type(u64 type, u64 calc_size, int num_stripes,
                               int sub_stripes)
 {
@@ -650,7 +655,95 @@ static u64 chunk_bytes_by_type(u64 type, u64 calc_size, int num_stripes,
 static u32 find_raid56_stripe_len(u32 data_devices, u32 dev_stripe_target)
 {
        /* TODO, add a way to store the preferred stripe size */
-       return 64 * 1024;
+       return BTRFS_STRIPE_LEN;
+}
+
+/*
+ * btrfs_device_avail_bytes - count bytes available for alloc_chunk
+ *
+ * It is not equal to "device->total_bytes - device->bytes_used".
+ * We do not allocate any chunk in 1M at beginning of device, and not
+ * allowed to allocate any chunk before alloc_start if it is specified.
+ * So search holes from max(1M, alloc_start) to device->total_bytes.
+ */
+static int btrfs_device_avail_bytes(struct btrfs_trans_handle *trans,
+                                   struct btrfs_device *device,
+                                   u64 *avail_bytes)
+{
+       struct btrfs_path *path;
+       struct btrfs_root *root = device->dev_root;
+       struct btrfs_key key;
+       struct btrfs_dev_extent *dev_extent = NULL;
+       struct extent_buffer *l;
+       u64 search_start = root->fs_info->alloc_start;
+       u64 search_end = device->total_bytes;
+       u64 extent_end = 0;
+       u64 free_bytes = 0;
+       int ret;
+       int slot = 0;
+
+       search_start = max(BTRFS_BLOCK_RESERVED_1M_FOR_SUPER, search_start);
+
+       path = btrfs_alloc_path();
+       if (!path)
+               return -ENOMEM;
+
+       key.objectid = device->devid;
+       key.offset = root->fs_info->alloc_start;
+       key.type = BTRFS_DEV_EXTENT_KEY;
+
+       path->reada = 2;
+       ret = btrfs_search_slot(trans, root, &key, path, 0, 0);
+       if (ret < 0)
+               goto error;
+       ret = btrfs_previous_item(root, path, 0, key.type);
+       if (ret < 0)
+               goto error;
+
+       while (1) {
+               l = path->nodes[0];
+               slot = path->slots[0];
+               if (slot >= btrfs_header_nritems(l)) {
+                       ret = btrfs_next_leaf(root, path);
+                       if (ret == 0)
+                               continue;
+                       if (ret < 0)
+                               goto error;
+                       break;
+               }
+               btrfs_item_key_to_cpu(l, &key, slot);
+
+               if (key.objectid < device->devid)
+                       goto next;
+               if (key.objectid > device->devid)
+                       break;
+               if (btrfs_key_type(&key) != BTRFS_DEV_EXTENT_KEY)
+                       goto next;
+               if (key.offset > search_end)
+                       break;
+               if (key.offset > search_start)
+                       free_bytes += key.offset - search_start;
+
+               dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent);
+               extent_end = key.offset + btrfs_dev_extent_length(l,
+                                                                 dev_extent);
+               if (extent_end > search_start)
+                       search_start = extent_end;
+               if (search_start > search_end)
+                       break;
+next:
+               path->slots[0]++;
+               cond_resched();
+       }
+
+       if (search_start < search_end)
+               free_bytes += search_end - search_start;
+
+       *avail_bytes = free_bytes;
+       ret = 0;
+error:
+       btrfs_free_path(path);
+       return ret;
 }
 
 int btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
@@ -659,19 +752,19 @@ int btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
 {
        u64 dev_offset;
        struct btrfs_fs_info *info = extent_root->fs_info;
-       struct btrfs_root *chunk_root = extent_root->fs_info->chunk_root;
+       struct btrfs_root *chunk_root = info->chunk_root;
        struct btrfs_stripe *stripes;
        struct btrfs_device *device = NULL;
        struct btrfs_chunk *chunk;
        struct list_head private_devs;
-       struct list_head *dev_list = &extent_root->fs_info->fs_devices->devices;
+       struct list_head *dev_list = &info->fs_devices->devices;
        struct list_head *cur;
        struct map_lookup *map;
        int min_stripe_size = 1 * 1024 * 1024;
        u64 calc_size = 8 * 1024 * 1024;
        u64 min_free;
        u64 max_chunk_size = 4 * calc_size;
-       u64 avail;
+       u64 avail = 0;
        u64 max_avail = 0;
        u64 percent_max;
        int num_stripes = 1;
@@ -680,7 +773,7 @@ int btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
        int looped = 0;
        int ret;
        int index;
-       int stripe_len = 64 * 1024;
+       int stripe_len = BTRFS_STRIPE_LEN;
        struct btrfs_key key;
        u64 offset;
 
@@ -775,7 +868,9 @@ again:
        /* build a private list of devices we will allocate from */
        while(index < num_stripes) {
                device = list_entry(cur, struct btrfs_device, dev_list);
-               avail = device->total_bytes - device->bytes_used;
+               ret = btrfs_device_avail_bytes(trans, device, &avail);
+               if (ret)
+                       return ret;
                cur = cur->next;
                if (avail >= min_free) {
                        list_move_tail(&device->dev_list, &private_devs);
@@ -817,7 +912,7 @@ again:
        if (!chunk)
                return -ENOMEM;
 
-       map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS);
+       map = kmalloc(btrfs_map_lookup_size(num_stripes), GFP_NOFS);
        if (!map) {
                kfree(chunk);
                return -ENOMEM;
@@ -884,9 +979,7 @@ again:
        map->ce.start = key.offset;
        map->ce.size = *num_bytes;
 
-       ret = insert_existing_cache_extent(
-                          &extent_root->fs_info->mapping_tree.cache_tree,
-                          &map->ce);
+       ret = insert_cache_extent(&info->mapping_tree.cache_tree, &map->ce);
        BUG_ON(ret);
 
        if (type & BTRFS_BLOCK_GROUP_SYSTEM) {
@@ -905,11 +998,11 @@ int btrfs_alloc_data_chunk(struct btrfs_trans_handle *trans,
 {
        u64 dev_offset;
        struct btrfs_fs_info *info = extent_root->fs_info;
-       struct btrfs_root *chunk_root = extent_root->fs_info->chunk_root;
+       struct btrfs_root *chunk_root = info->chunk_root;
        struct btrfs_stripe *stripes;
        struct btrfs_device *device = NULL;
        struct btrfs_chunk *chunk;
-       struct list_head *dev_list = &extent_root->fs_info->fs_devices->devices;
+       struct list_head *dev_list = &info->fs_devices->devices;
        struct list_head *cur;
        struct map_lookup *map;
        u64 calc_size = 8 * 1024 * 1024;
@@ -917,7 +1010,7 @@ int btrfs_alloc_data_chunk(struct btrfs_trans_handle *trans,
        int sub_stripes = 0;
        int ret;
        int index;
-       int stripe_len = 64 * 1024;
+       int stripe_len = BTRFS_STRIPE_LEN;
        struct btrfs_key key;
 
        key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
@@ -931,7 +1024,7 @@ int btrfs_alloc_data_chunk(struct btrfs_trans_handle *trans,
        if (!chunk)
                return -ENOMEM;
 
-       map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS);
+       map = kmalloc(btrfs_map_lookup_size(num_stripes), GFP_NOFS);
        if (!map) {
                kfree(chunk);
                return -ENOMEM;
@@ -992,27 +1085,20 @@ int btrfs_alloc_data_chunk(struct btrfs_trans_handle *trans,
        map->ce.start = key.offset;
        map->ce.size = num_bytes;
 
-       ret = insert_existing_cache_extent(
-                          &extent_root->fs_info->mapping_tree.cache_tree,
-                          &map->ce);
+       ret = insert_cache_extent(&info->mapping_tree.cache_tree, &map->ce);
        BUG_ON(ret);
 
        kfree(chunk);
        return ret;
 }
 
-void btrfs_mapping_init(struct btrfs_mapping_tree *tree)
-{
-       cache_tree_init(&tree->cache_tree);
-}
-
 int btrfs_num_copies(struct btrfs_mapping_tree *map_tree, u64 logical, u64 len)
 {
        struct cache_extent *ce;
        struct map_lookup *map;
        int ret;
 
-       ce = find_first_cache_extent(&map_tree->cache_tree, logical);
+       ce = search_cache_extent(&map_tree->cache_tree, logical);
        BUG_ON(!ce);
        BUG_ON(ce->start > logical || ce->start + ce->size < logical);
        map = container_of(ce, struct map_lookup, ce);
@@ -1036,7 +1122,7 @@ int btrfs_next_metadata(struct btrfs_mapping_tree *map_tree, u64 *logical,
        struct cache_extent *ce;
        struct map_lookup *map;
 
-       ce = find_first_cache_extent(&map_tree->cache_tree, *logical);
+       ce = search_cache_extent(&map_tree->cache_tree, *logical);
 
        while (ce) {
                ce = next_cache_extent(ce);
@@ -1067,7 +1153,7 @@ int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree,
        u64 rmap_len;
        int i, j, nr = 0;
 
-       ce = find_first_cache_extent(&map_tree->cache_tree, chunk_start);
+       ce = search_cache_extent(&map_tree->cache_tree, chunk_start);
        BUG_ON(!ce);
        map = container_of(ce, struct map_lookup, ce);
 
@@ -1179,15 +1265,13 @@ int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw,
                stripes_allocated = 1;
        }
 again:
-       ce = find_first_cache_extent(&map_tree->cache_tree, logical);
+       ce = search_cache_extent(&map_tree->cache_tree, logical);
        if (!ce) {
-               if (multi)
-                       kfree(multi);
+               kfree(multi);
                return -ENOENT;
        }
        if (ce->start > logical || ce->start + ce->size < logical) {
-               if (multi)
-                       kfree(multi);
+               kfree(multi);
                return -ENOENT;
        }
 
@@ -1287,7 +1371,7 @@ again:
                                BTRFS_BLOCK_GROUP_RAID6)) {
 
                if (raid_map) {
-                       int i, rot;
+                       int rot;
                        u64 tmp;
                        u64 raid56_full_stripe_start;
                        u64 full_stripe_len = nr_data_stripes(map) * map->stripe_len;
@@ -1387,75 +1471,21 @@ struct btrfs_device *btrfs_find_device(struct btrfs_root *root, u64 devid,
        return NULL;
 }
 
-struct btrfs_device *btrfs_find_device_by_devid(struct btrfs_root *root,
-                                               u64 devid, int instance)
+struct btrfs_device *
+btrfs_find_device_by_devid(struct btrfs_fs_devices *fs_devices,
+                          u64 devid, int instance)
 {
-       struct list_head *head = &root->fs_info->fs_devices->devices;
+       struct list_head *head = &fs_devices->devices;
        struct btrfs_device *dev;
-       struct list_head *cur;
        int num_found = 0;
 
-       list_for_each(cur, head) {
-               dev = list_entry(cur, struct btrfs_device, dev_list);
+       list_for_each_entry(dev, head, dev_list) {
                if (dev->devid == devid && num_found++ == instance)
                        return dev;
        }
        return NULL;
 }
 
-int btrfs_bootstrap_super_map(struct btrfs_mapping_tree *map_tree,
-                             struct btrfs_fs_devices *fs_devices)
-{
-       struct map_lookup *map;
-       u64 logical = BTRFS_SUPER_INFO_OFFSET;
-       u64 length = BTRFS_SUPER_INFO_SIZE;
-       int num_stripes = 0;
-       int sub_stripes = 0;
-       int ret;
-       int i;
-       struct list_head *cur;
-
-       list_for_each(cur, &fs_devices->devices) {
-               num_stripes++;
-       }
-       map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS);
-       if (!map)
-               return -ENOMEM;
-
-       map->ce.start = logical;
-       map->ce.size = length;
-       map->num_stripes = num_stripes;
-       map->sub_stripes = sub_stripes;
-       map->io_width = length;
-       map->io_align = length;
-       map->sector_size = length;
-       map->stripe_len = length;
-       map->type = BTRFS_BLOCK_GROUP_RAID1;
-
-       i = 0;
-       list_for_each(cur, &fs_devices->devices) {
-               struct btrfs_device *device = list_entry(cur,
-                                                        struct btrfs_device,
-                                                        dev_list);
-               map->stripes[i].physical = logical;
-               map->stripes[i].dev = device;
-               i++;
-       }
-       ret = insert_existing_cache_extent(&map_tree->cache_tree, &map->ce);
-       if (ret == -EEXIST) {
-               struct cache_extent *old;
-               struct map_lookup *old_map;
-               old = find_cache_extent(&map_tree->cache_tree, logical, length);
-               old_map = container_of(old, struct map_lookup, ce);
-               remove_cache_extent(&map_tree->cache_tree, old);
-               kfree(old_map);
-               ret = insert_existing_cache_extent(&map_tree->cache_tree,
-                                                  &map->ce);
-       }
-       BUG_ON(ret);
-       return 0;
-}
-
 int btrfs_chunk_readonly(struct btrfs_root *root, u64 chunk_offset)
 {
        struct cache_extent *ce;
@@ -1464,8 +1494,15 @@ int btrfs_chunk_readonly(struct btrfs_root *root, u64 chunk_offset)
        int readonly = 0;
        int i;
 
-       ce = find_first_cache_extent(&map_tree->cache_tree, chunk_offset);
-       BUG_ON(!ce);
+       /*
+        * During chunk recovering, we may fail to find block group's
+        * corresponding chunk, we will rebuild it later
+        */
+       ce = search_cache_extent(&map_tree->cache_tree, chunk_offset);
+       if (!root->fs_info->is_chunk_recover)
+               BUG_ON(!ce);
+       else
+               return 0;
 
        map = container_of(ce, struct map_lookup, ce);
        for (i = 0; i < map->num_stripes; i++) {
@@ -1506,7 +1543,7 @@ static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key,
        logical = key->offset;
        length = btrfs_chunk_length(leaf, chunk);
 
-       ce = find_first_cache_extent(&map_tree->cache_tree, logical);
+       ce = search_cache_extent(&map_tree->cache_tree, logical);
 
        /* already mapped? */
        if (ce && ce->start <= logical && ce->start + ce->size > logical) {
@@ -1514,7 +1551,7 @@ static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key,
        }
 
        num_stripes = btrfs_chunk_num_stripes(leaf, chunk);
-       map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS);
+       map = kmalloc(btrfs_map_lookup_size(num_stripes), GFP_NOFS);
        if (!map)
                return -ENOMEM;
 
@@ -1544,7 +1581,7 @@ static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key,
                }
 
        }
-       ret = insert_existing_cache_extent(&map_tree->cache_tree, &map->ce);
+       ret = insert_cache_extent(&map_tree->cache_tree, &map->ce);
        BUG_ON(ret);
 
        return 0;
@@ -1628,10 +1665,10 @@ static int read_one_dev(struct btrfs_root *root,
        if (!device) {
                printk("warning devid %llu not found already\n",
                        (unsigned long long)devid);
-               device = kmalloc(sizeof(*device), GFP_NOFS);
+               device = kzalloc(sizeof(*device), GFP_NOFS);
                if (!device)
                        return -ENOMEM;
-               device->total_ios = 0;
+               device->fd = -1;
                list_add(&device->dev_list,
                         &root->fs_info->fs_devices->devices);
        }
@@ -1641,15 +1678,6 @@ static int read_one_dev(struct btrfs_root *root,
        return ret;
 }
 
-int btrfs_read_super_device(struct btrfs_root *root, struct extent_buffer *buf)
-{
-       struct btrfs_dev_item *dev_item;
-
-       dev_item = (struct btrfs_dev_item *)offsetof(struct btrfs_super_block,
-                                                    dev_item);
-       return read_one_dev(root, buf, dev_item);
-}
-
 int btrfs_read_sys_array(struct btrfs_root *root)
 {
        struct btrfs_super_block *super_copy = root->fs_info->super_copy;
@@ -1658,11 +1686,9 @@ int btrfs_read_sys_array(struct btrfs_root *root)
        struct btrfs_chunk *chunk;
        struct btrfs_key key;
        u32 num_stripes;
-       u32 array_size;
        u32 len = 0;
        u8 *ptr;
-       unsigned long sb_ptr;
-       u32 cur;
+       u8 *array_end;
        int ret = 0;
 
        sb = btrfs_find_create_tree_block(root, BTRFS_SUPER_INFO_OFFSET,
@@ -1671,7 +1697,8 @@ int btrfs_read_sys_array(struct btrfs_root *root)
                return -ENOMEM;
        btrfs_set_buffer_uptodate(sb);
        write_extent_buffer(sb, super_copy, 0, sizeof(*super_copy));
-       array_size = btrfs_super_sys_array_size(super_copy);
+       array_end = ((u8 *)super_copy->sys_chunk_array) +
+                   btrfs_super_sys_array_size(super_copy);
 
        /*
         * we do this loop twice, once for the device items and
@@ -1679,20 +1706,16 @@ int btrfs_read_sys_array(struct btrfs_root *root)
         * structs filled in for every chunk
         */
        ptr = super_copy->sys_chunk_array;
-       sb_ptr = offsetof(struct btrfs_super_block, sys_chunk_array);
-       cur = 0;
 
-       while (cur < array_size) {
+       while (ptr < array_end) {
                disk_key = (struct btrfs_disk_key *)ptr;
                btrfs_disk_key_to_cpu(&key, disk_key);
 
                len = sizeof(*disk_key);
                ptr += len;
-               sb_ptr += len;
-               cur += len;
 
                if (key.type == BTRFS_CHUNK_ITEM_KEY) {
-                       chunk = (struct btrfs_chunk *)sb_ptr;
+                       chunk = (struct btrfs_chunk *)(ptr - (u8 *)super_copy);
                        ret = read_one_chunk(root, &key, sb, chunk);
                        if (ret)
                                break;
@@ -1702,8 +1725,6 @@ int btrfs_read_sys_array(struct btrfs_root *root)
                        BUG();
                }
                ptr += len;
-               sb_ptr += len;
-               cur += len;
        }
        free_extent_buffer(sb);
        return ret;
@@ -1724,15 +1745,18 @@ int btrfs_read_chunk_tree(struct btrfs_root *root)
        if (!path)
                return -ENOMEM;
 
-       /* first we search for all of the device items, and then we
-        * read in all of the chunk items.  This way we can create chunk
-        * mappings that reference all of the devices that are afound
+       /*
+        * Read all device items, and then all the chunk items. All
+        * device items are found before any chunk item (their object id
+        * is smaller than the lowest possible object id for a chunk
+        * item - BTRFS_FIRST_CHUNK_TREE_OBJECTID).
         */
        key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
        key.offset = 0;
        key.type = 0;
-again:
        ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+       if (ret < 0)
+               goto error;
        while(1) {
                leaf = path->nodes[0];
                slot = path->slots[0];
@@ -1745,16 +1769,12 @@ again:
                        break;
                }
                btrfs_item_key_to_cpu(leaf, &found_key, slot);
-               if (key.objectid == BTRFS_DEV_ITEMS_OBJECTID) {
-                       if (found_key.objectid != BTRFS_DEV_ITEMS_OBJECTID)
-                               break;
-                       if (found_key.type == BTRFS_DEV_ITEM_KEY) {
-                               struct btrfs_dev_item *dev_item;
-                               dev_item = btrfs_item_ptr(leaf, slot,
+               if (found_key.type == BTRFS_DEV_ITEM_KEY) {
+                       struct btrfs_dev_item *dev_item;
+                       dev_item = btrfs_item_ptr(leaf, slot,
                                                  struct btrfs_dev_item);
-                               ret = read_one_dev(root, leaf, dev_item);
-                               BUG_ON(ret);
-                       }
+                       ret = read_one_dev(root, leaf, dev_item);
+                       BUG_ON(ret);
                } else if (found_key.type == BTRFS_CHUNK_ITEM_KEY) {
                        struct btrfs_chunk *chunk;
                        chunk = btrfs_item_ptr(leaf, slot, struct btrfs_chunk);
@@ -1763,11 +1783,6 @@ again:
                }
                path->slots[0]++;
        }
-       if (key.objectid == BTRFS_DEV_ITEMS_OBJECTID) {
-               key.objectid = 0;
-               btrfs_release_path(root, path);
-               goto again;
-       }
 
        ret = 0;
 error:
@@ -1779,3 +1794,158 @@ struct list_head *btrfs_scanned_uuids(void)
 {
        return &fs_uuids;
 }
+
+static int rmw_eb(struct btrfs_fs_info *info,
+                 struct extent_buffer *eb, struct extent_buffer *orig_eb)
+{
+       int ret;
+       unsigned long orig_off = 0;
+       unsigned long dest_off = 0;
+       unsigned long copy_len = eb->len;
+
+       ret = read_whole_eb(info, eb, 0);
+       if (ret)
+               return ret;
+
+       if (eb->start + eb->len <= orig_eb->start ||
+           eb->start >= orig_eb->start + orig_eb->len)
+               return 0;
+       /*
+        * | ----- orig_eb ------- |
+        *         | ----- stripe -------  |
+        *         | ----- orig_eb ------- |
+        *              | ----- orig_eb ------- |
+        */
+       if (eb->start > orig_eb->start)
+               orig_off = eb->start - orig_eb->start;
+       if (orig_eb->start > eb->start)
+               dest_off = orig_eb->start - eb->start;
+
+       if (copy_len > orig_eb->len - orig_off)
+               copy_len = orig_eb->len - orig_off;
+       if (copy_len > eb->len - dest_off)
+               copy_len = eb->len - dest_off;
+
+       memcpy(eb->data + dest_off, orig_eb->data + orig_off, copy_len);
+       return 0;
+}
+
+static void split_eb_for_raid56(struct btrfs_fs_info *info,
+                               struct extent_buffer *orig_eb,
+                              struct extent_buffer **ebs,
+                              u64 stripe_len, u64 *raid_map,
+                              int num_stripes)
+{
+       struct extent_buffer *eb;
+       u64 start = orig_eb->start;
+       u64 this_eb_start;
+       int i;
+       int ret;
+
+       for (i = 0; i < num_stripes; i++) {
+               if (raid_map[i] >= BTRFS_RAID5_P_STRIPE)
+                       break;
+
+               eb = malloc(sizeof(struct extent_buffer) + stripe_len);
+               if (!eb)
+                       BUG();
+               memset(eb, 0, sizeof(struct extent_buffer) + stripe_len);
+
+               eb->start = raid_map[i];
+               eb->len = stripe_len;
+               eb->refs = 1;
+               eb->flags = 0;
+               eb->fd = -1;
+               eb->dev_bytenr = (u64)-1;
+
+               this_eb_start = raid_map[i];
+
+               if (start > this_eb_start ||
+                   start + orig_eb->len < this_eb_start + stripe_len) {
+                       ret = rmw_eb(info, eb, orig_eb);
+                       BUG_ON(ret);
+               } else {
+                       memcpy(eb->data, orig_eb->data + eb->start - start, stripe_len);
+               }
+               ebs[i] = eb;
+       }
+}
+
+int write_raid56_with_parity(struct btrfs_fs_info *info,
+                            struct extent_buffer *eb,
+                            struct btrfs_multi_bio *multi,
+                            u64 stripe_len, u64 *raid_map)
+{
+       struct extent_buffer **ebs, *p_eb = NULL, *q_eb = NULL;
+       int i;
+       int j;
+       int ret;
+       int alloc_size = eb->len;
+
+       ebs = kmalloc(sizeof(*ebs) * multi->num_stripes, GFP_NOFS);
+       BUG_ON(!ebs);
+
+       if (stripe_len > alloc_size)
+               alloc_size = stripe_len;
+
+       split_eb_for_raid56(info, eb, ebs, stripe_len, raid_map,
+                           multi->num_stripes);
+
+       for (i = 0; i < multi->num_stripes; i++) {
+               struct extent_buffer *new_eb;
+               if (raid_map[i] < BTRFS_RAID5_P_STRIPE) {
+                       ebs[i]->dev_bytenr = multi->stripes[i].physical;
+                       ebs[i]->fd = multi->stripes[i].dev->fd;
+                       multi->stripes[i].dev->total_ios++;
+                       BUG_ON(ebs[i]->start != raid_map[i]);
+                       continue;
+               }
+               new_eb = kmalloc(sizeof(*eb) + alloc_size, GFP_NOFS);
+               BUG_ON(!new_eb);
+               new_eb->dev_bytenr = multi->stripes[i].physical;
+               new_eb->fd = multi->stripes[i].dev->fd;
+               multi->stripes[i].dev->total_ios++;
+               new_eb->len = stripe_len;
+
+               if (raid_map[i] == BTRFS_RAID5_P_STRIPE)
+                       p_eb = new_eb;
+               else if (raid_map[i] == BTRFS_RAID6_Q_STRIPE)
+                       q_eb = new_eb;
+       }
+       if (q_eb) {
+               void **pointers;
+
+               pointers = kmalloc(sizeof(*pointers) * multi->num_stripes,
+                                  GFP_NOFS);
+               BUG_ON(!pointers);
+
+               ebs[multi->num_stripes - 2] = p_eb;
+               ebs[multi->num_stripes - 1] = q_eb;
+
+               for (i = 0; i < multi->num_stripes; i++)
+                       pointers[i] = ebs[i]->data;
+
+               raid6_gen_syndrome(multi->num_stripes, stripe_len, pointers);
+               kfree(pointers);
+       } else {
+               ebs[multi->num_stripes - 1] = p_eb;
+               memcpy(p_eb->data, ebs[0]->data, stripe_len);
+               for (j = 1; j < multi->num_stripes - 1; j++) {
+                       for (i = 0; i < stripe_len; i += sizeof(unsigned long)) {
+                               *(unsigned long *)(p_eb->data + i) ^=
+                                       *(unsigned long *)(ebs[j]->data + i);
+                       }
+               }
+       }
+
+       for (i = 0; i < multi->num_stripes; i++) {
+               ret = write_extent_to_disk(ebs[i]);
+               BUG_ON(ret);
+               if (ebs[i] != eb)
+                       kfree(ebs[i]);
+       }
+
+       kfree(ebs);
+
+       return 0;
+}