#include <linux/sched.h>
#include <linux/bio.h>
#include <linux/buffer_head.h>
+#include <linux/blkdev.h>
+#include <linux/random.h>
#include <asm/div64.h>
#include "ctree.h"
#include "extent_map.h"
devices_cur = fs_devices->devices.next;
dev = list_entry(devices_cur, struct btrfs_device,
dev_list);
- printk("uuid cleanup finds %s\n", dev->name);
if (dev->bdev) {
- printk("closing\n");
close_bdev_excl(dev->bdev);
}
list_del(&dev->dev_list);
return 0;
}
-static struct btrfs_device *__find_device(struct list_head *head, u64 devid)
+static struct btrfs_device *__find_device(struct list_head *head, u64 devid,
+ u8 *uuid)
{
struct btrfs_device *dev;
struct list_head *cur;
list_for_each(cur, head) {
dev = list_entry(cur, struct btrfs_device, dev_list);
- if (dev->devid == devid)
+ if (dev->devid == devid &&
+ (!uuid || !memcmp(dev->uuid, uuid, BTRFS_UUID_SIZE))) {
return dev;
+ }
}
return NULL;
}
if (!fs_devices)
return -ENOMEM;
INIT_LIST_HEAD(&fs_devices->devices);
+ INIT_LIST_HEAD(&fs_devices->alloc_list);
list_add(&fs_devices->list, &fs_uuids);
memcpy(fs_devices->fsid, disk_super->fsid, BTRFS_FSID_SIZE);
fs_devices->latest_devid = devid;
fs_devices->num_devices = 0;
device = NULL;
} else {
- device = __find_device(&fs_devices->devices, devid);
+ device = __find_device(&fs_devices->devices, devid,
+ disk_super->dev_item.uuid);
}
if (!device) {
device = kzalloc(sizeof(*device), GFP_NOFS);
return -ENOMEM;
}
device->devid = devid;
+ memcpy(device->uuid, disk_super->dev_item.uuid,
+ BTRFS_UUID_SIZE);
device->barriers = 1;
spin_lock_init(&device->io_lock);
device->name = kstrdup(path, GFP_NOFS);
return -ENOMEM;
}
list_add(&device->dev_list, &fs_devices->devices);
+ list_add(&device->dev_alloc_list, &fs_devices->alloc_list);
fs_devices->num_devices++;
}
}
if (fs_devices->lowest_devid > devid) {
fs_devices->lowest_devid = devid;
- printk("lowest devid now %Lu\n", devid);
}
*fs_devices_ret = fs_devices;
return 0;
device = list_entry(cur, struct btrfs_device, dev_list);
if (device->bdev) {
close_bdev_excl(device->bdev);
- printk("close devices closes %s\n", device->name);
}
device->bdev = NULL;
}
mutex_lock(&uuid_mutex);
- printk("scan one opens %s\n", path);
bdev = open_bdev_excl(path, flags, holder);
if (IS_ERR(bdev)) {
- printk("open failed\n");
ret = PTR_ERR(bdev);
goto error;
}
disk_super = (struct btrfs_super_block *)bh->b_data;
if (strncmp((char *)(&disk_super->magic), BTRFS_MAGIC,
sizeof(disk_super->magic))) {
- printk("no btrfs found on %s\n", path);
ret = -EINVAL;
goto error_brelse;
}
devid = le64_to_cpu(disk_super->dev_item.devid);
transid = btrfs_super_generation(disk_super);
- printk("found device %Lu transid %Lu on %s\n", devid, transid, path);
+ if (disk_super->label[0])
+ printk("device label %s ", disk_super->label);
+ else {
+ /* FIXME, make a readl uuid parser */
+ printk("device fsid %llx-%llx ",
+ *(unsigned long long *)disk_super->fsid,
+ *(unsigned long long *)(disk_super->fsid + 8));
+ }
+ printk("devid %Lu transid %Lu %s\n", devid, transid, path);
ret = device_list_add(path, disk_super, devid, fs_devices_ret);
error_brelse:
* so we make sure to start at an offset of at least 1MB
*/
search_start = max((u64)1024 * 1024, search_start);
+
+ if (root->fs_info->alloc_start + num_bytes <= device->total_bytes)
+ search_start = max(root->fs_info->alloc_start, search_start);
+
key.objectid = device->devid;
key.offset = search_start;
key.type = BTRFS_DEV_EXTENT_KEY;
return ret;
}
+int btrfs_free_dev_extent(struct btrfs_trans_handle *trans,
+ struct btrfs_device *device,
+ u64 start)
+{
+ int ret;
+ struct btrfs_path *path;
+ struct btrfs_root *root = device->dev_root;
+ struct btrfs_key key;
+
+ path = btrfs_alloc_path();
+ if (!path)
+ return -ENOMEM;
+
+ key.objectid = device->devid;
+ key.offset = start;
+ key.type = BTRFS_DEV_EXTENT_KEY;
+
+ ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
+ BUG_ON(ret);
+
+ ret = btrfs_del_item(trans, root, path);
+ BUG_ON(ret);
+
+ btrfs_free_path(path);
+ return ret;
+}
+
int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans,
struct btrfs_device *device,
u64 chunk_tree, u64 chunk_objectid,
btrfs_free_path(path);
return ret;
}
+
+int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
+{
+ struct btrfs_trans_handle *trans;
+ struct btrfs_device *device;
+ struct block_device *bdev;
+ struct list_head *cur;
+ struct list_head *devices;
+ u64 total_bytes;
+ int ret = 0;
+
+
+ bdev = open_bdev_excl(device_path, 0, root->fs_info->bdev_holder);
+ if (!bdev) {
+ return -EIO;
+ }
+ mutex_lock(&root->fs_info->fs_mutex);
+ trans = btrfs_start_transaction(root, 1);
+ devices = &root->fs_info->fs_devices->devices;
+ list_for_each(cur, devices) {
+ device = list_entry(cur, struct btrfs_device, dev_list);
+ if (device->bdev == bdev) {
+ ret = -EEXIST;
+ goto out;
+ }
+ }
+
+ device = kzalloc(sizeof(*device), GFP_NOFS);
+ if (!device) {
+ /* we can safely leave the fs_devices entry around */
+ ret = -ENOMEM;
+ goto out_close_bdev;
+ }
+
+ device->barriers = 1;
+ generate_random_uuid(device->uuid);
+ spin_lock_init(&device->io_lock);
+ device->name = kstrdup(device_path, GFP_NOFS);
+ if (!device->name) {
+ kfree(device);
+ goto out_close_bdev;
+ }
+ device->io_width = root->sectorsize;
+ device->io_align = root->sectorsize;
+ device->sector_size = root->sectorsize;
+ device->total_bytes = i_size_read(bdev->bd_inode);
+ device->dev_root = root->fs_info->dev_root;
+ device->bdev = bdev;
+
+ ret = btrfs_add_device(trans, root, device);
+ if (ret)
+ goto out_close_bdev;
+
+ total_bytes = btrfs_super_total_bytes(&root->fs_info->super_copy);
+ btrfs_set_super_total_bytes(&root->fs_info->super_copy,
+ total_bytes + device->total_bytes);
+
+ total_bytes = btrfs_super_num_devices(&root->fs_info->super_copy);
+ btrfs_set_super_num_devices(&root->fs_info->super_copy,
+ total_bytes + 1);
+
+ list_add(&device->dev_list, &root->fs_info->fs_devices->devices);
+ list_add(&device->dev_alloc_list,
+ &root->fs_info->fs_devices->alloc_list);
+ root->fs_info->fs_devices->num_devices++;
+out:
+ btrfs_end_transaction(trans, root);
+ mutex_unlock(&root->fs_info->fs_mutex);
+ return ret;
+
+out_close_bdev:
+ close_bdev_excl(bdev);
+ goto out;
+}
+
int btrfs_update_device(struct btrfs_trans_handle *trans,
struct btrfs_device *device)
{
return ret;
}
+int btrfs_grow_device(struct btrfs_trans_handle *trans,
+ struct btrfs_device *device, u64 new_size)
+{
+ struct btrfs_super_block *super_copy =
+ &device->dev_root->fs_info->super_copy;
+ u64 old_total = btrfs_super_total_bytes(super_copy);
+ u64 diff = new_size - device->total_bytes;
+
+ btrfs_set_super_total_bytes(super_copy, old_total + diff);
+ return btrfs_update_device(trans, device);
+}
+
+static int btrfs_free_chunk(struct btrfs_trans_handle *trans,
+ struct btrfs_root *root,
+ u64 chunk_tree, u64 chunk_objectid,
+ u64 chunk_offset)
+{
+ int ret;
+ struct btrfs_path *path;
+ struct btrfs_key key;
+
+ root = root->fs_info->chunk_root;
+ path = btrfs_alloc_path();
+ if (!path)
+ return -ENOMEM;
+
+ key.objectid = chunk_objectid;
+ key.offset = chunk_offset;
+ key.type = BTRFS_CHUNK_ITEM_KEY;
+
+ ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
+ BUG_ON(ret);
+
+ ret = btrfs_del_item(trans, root, path);
+ BUG_ON(ret);
+
+ btrfs_free_path(path);
+ return 0;
+}
+
+int btrfs_del_sys_chunk(struct btrfs_root *root, u64 chunk_objectid, u64
+ chunk_offset)
+{
+ struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
+ struct btrfs_disk_key *disk_key;
+ struct btrfs_chunk *chunk;
+ u8 *ptr;
+ int ret = 0;
+ u32 num_stripes;
+ u32 array_size;
+ u32 len = 0;
+ u32 cur;
+ struct btrfs_key key;
+
+ array_size = btrfs_super_sys_array_size(super_copy);
+
+ ptr = super_copy->sys_chunk_array;
+ cur = 0;
+
+ while (cur < array_size) {
+ disk_key = (struct btrfs_disk_key *)ptr;
+ btrfs_disk_key_to_cpu(&key, disk_key);
+
+ len = sizeof(*disk_key);
+
+ if (key.type == BTRFS_CHUNK_ITEM_KEY) {
+ chunk = (struct btrfs_chunk *)(ptr + len);
+ num_stripes = btrfs_stack_chunk_num_stripes(chunk);
+ len += btrfs_chunk_item_size(num_stripes);
+ } else {
+ ret = -EIO;
+ break;
+ }
+ if (key.objectid == chunk_objectid &&
+ key.offset == chunk_offset) {
+ memmove(ptr, ptr + len, array_size - (cur + len));
+ array_size -= len;
+ btrfs_set_super_sys_array_size(super_copy, array_size);
+ } else {
+ ptr += len;
+ cur += len;
+ }
+ }
+ return ret;
+}
+
+
+int btrfs_relocate_chunk(struct btrfs_root *root,
+ u64 chunk_tree, u64 chunk_objectid,
+ u64 chunk_offset)
+{
+ struct extent_map_tree *em_tree;
+ struct btrfs_root *extent_root;
+ struct btrfs_trans_handle *trans;
+ struct extent_map *em;
+ struct map_lookup *map;
+ int ret;
+ int i;
+
+ root = root->fs_info->chunk_root;
+ extent_root = root->fs_info->extent_root;
+ em_tree = &root->fs_info->mapping_tree.map_tree;
+
+ /* step one, relocate all the extents inside this chunk */
+ ret = btrfs_shrink_extent_tree(extent_root, chunk_offset);
+ BUG_ON(ret);
+
+ trans = btrfs_start_transaction(root, 1);
+ BUG_ON(!trans);
+
+ /*
+ * step two, delete the device extents and the
+ * chunk tree entries
+ */
+ spin_lock(&em_tree->lock);
+ em = lookup_extent_mapping(em_tree, chunk_offset, 1);
+ spin_unlock(&em_tree->lock);
+
+ BUG_ON(em->start > chunk_offset || em->start + em->len < chunk_offset);
+ map = (struct map_lookup *)em->bdev;
+
+ for (i = 0; i < map->num_stripes; i++) {
+ ret = btrfs_free_dev_extent(trans, map->stripes[i].dev,
+ map->stripes[i].physical);
+ BUG_ON(ret);
+ }
+ ret = btrfs_free_chunk(trans, root, chunk_tree, chunk_objectid,
+ chunk_offset);
+
+ BUG_ON(ret);
+
+ if (map->type & BTRFS_BLOCK_GROUP_SYSTEM) {
+ ret = btrfs_del_sys_chunk(root, chunk_objectid, chunk_offset);
+ BUG_ON(ret);
+ goto out;
+ }
+
+
+
+ spin_lock(&em_tree->lock);
+ remove_extent_mapping(em_tree, em);
+ kfree(map);
+ em->bdev = NULL;
+
+ /* once for the tree */
+ free_extent_map(em);
+ spin_unlock(&em_tree->lock);
+
+out:
+ /* once for us */
+ free_extent_map(em);
+
+ btrfs_end_transaction(trans, root);
+ return 0;
+}
+
+static u64 div_factor(u64 num, int factor)
+{
+ if (factor == 10)
+ return num;
+ num *= factor;
+ do_div(num, 10);
+ return num;
+}
+
+
+int btrfs_balance(struct btrfs_root *dev_root)
+{
+ int ret;
+ struct list_head *cur;
+ struct list_head *devices = &dev_root->fs_info->fs_devices->devices;
+ struct btrfs_device *device;
+ u64 old_size;
+ u64 size_to_free;
+ struct btrfs_path *path;
+ struct btrfs_key key;
+ struct btrfs_chunk *chunk;
+ struct btrfs_root *chunk_root = dev_root->fs_info->chunk_root;
+ struct btrfs_trans_handle *trans;
+ struct btrfs_key found_key;
+
+
+ dev_root = dev_root->fs_info->dev_root;
+
+ mutex_lock(&dev_root->fs_info->fs_mutex);
+ /* step one make some room on all the devices */
+ list_for_each(cur, devices) {
+ device = list_entry(cur, struct btrfs_device, dev_list);
+ old_size = device->total_bytes;
+ size_to_free = div_factor(old_size, 1);
+ size_to_free = min(size_to_free, (u64)1 * 1024 * 1024);
+ if (device->total_bytes - device->bytes_used > size_to_free)
+ continue;
+
+ ret = btrfs_shrink_device(device, old_size - size_to_free);
+ BUG_ON(ret);
+
+ trans = btrfs_start_transaction(dev_root, 1);
+ BUG_ON(!trans);
+
+ ret = btrfs_grow_device(trans, device, old_size);
+ BUG_ON(ret);
+
+ btrfs_end_transaction(trans, dev_root);
+ }
+
+ /* step two, relocate all the chunks */
+ path = btrfs_alloc_path();
+ BUG_ON(!path);
+
+ key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
+ key.offset = (u64)-1;
+ key.type = BTRFS_CHUNK_ITEM_KEY;
+
+ while(1) {
+ ret = btrfs_search_slot(NULL, chunk_root, &key, path, 0, 0);
+ if (ret < 0)
+ goto error;
+
+ /*
+ * this shouldn't happen, it means the last relocate
+ * failed
+ */
+ if (ret == 0)
+ break;
+
+ ret = btrfs_previous_item(chunk_root, path, 0,
+ BTRFS_CHUNK_ITEM_KEY);
+ if (ret) {
+ break;
+ }
+ btrfs_item_key_to_cpu(path->nodes[0], &found_key,
+ path->slots[0]);
+ if (found_key.objectid != key.objectid)
+ break;
+ chunk = btrfs_item_ptr(path->nodes[0],
+ path->slots[0],
+ struct btrfs_chunk);
+ key.offset = found_key.offset;
+ /* chunk zero is special */
+ if (key.offset == 0)
+ break;
+
+ ret = btrfs_relocate_chunk(chunk_root,
+ chunk_root->root_key.objectid,
+ found_key.objectid,
+ found_key.offset);
+ BUG_ON(ret);
+ btrfs_release_path(chunk_root, path);
+ }
+ ret = 0;
+error:
+ btrfs_free_path(path);
+ mutex_unlock(&dev_root->fs_info->fs_mutex);
+ return ret;
+}
+
+/*
+ * shrinking a device means finding all of the device extents past
+ * the new size, and then following the back refs to the chunks.
+ * The chunk relocation code actually frees the device extent
+ */
+int btrfs_shrink_device(struct btrfs_device *device, u64 new_size)
+{
+ struct btrfs_trans_handle *trans;
+ struct btrfs_root *root = device->dev_root;
+ struct btrfs_dev_extent *dev_extent = NULL;
+ struct btrfs_path *path;
+ u64 length;
+ u64 chunk_tree;
+ u64 chunk_objectid;
+ u64 chunk_offset;
+ int ret;
+ int slot;
+ struct extent_buffer *l;
+ struct btrfs_key key;
+ struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
+ u64 old_total = btrfs_super_total_bytes(super_copy);
+ u64 diff = device->total_bytes - new_size;
+
+
+ path = btrfs_alloc_path();
+ if (!path)
+ return -ENOMEM;
+
+ trans = btrfs_start_transaction(root, 1);
+ if (!trans) {
+ ret = -ENOMEM;
+ goto done;
+ }
+
+ path->reada = 2;
+
+ device->total_bytes = new_size;
+ ret = btrfs_update_device(trans, device);
+ if (ret) {
+ btrfs_end_transaction(trans, root);
+ goto done;
+ }
+ WARN_ON(diff > old_total);
+ btrfs_set_super_total_bytes(super_copy, old_total - diff);
+ btrfs_end_transaction(trans, root);
+
+ key.objectid = device->devid;
+ key.offset = (u64)-1;
+ key.type = BTRFS_DEV_EXTENT_KEY;
+
+ while (1) {
+ ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+ if (ret < 0)
+ goto done;
+
+ ret = btrfs_previous_item(root, path, 0, key.type);
+ if (ret < 0)
+ goto done;
+ if (ret) {
+ ret = 0;
+ goto done;
+ }
+
+ l = path->nodes[0];
+ slot = path->slots[0];
+ btrfs_item_key_to_cpu(l, &key, path->slots[0]);
+
+ if (key.objectid != device->devid)
+ goto done;
+
+ dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent);
+ length = btrfs_dev_extent_length(l, dev_extent);
+
+ if (key.offset + length <= new_size)
+ goto done;
+
+ chunk_tree = btrfs_dev_extent_chunk_tree(l, dev_extent);
+ chunk_objectid = btrfs_dev_extent_chunk_objectid(l, dev_extent);
+ chunk_offset = btrfs_dev_extent_chunk_offset(l, dev_extent);
+ btrfs_release_path(root, path);
+
+ ret = btrfs_relocate_chunk(root, chunk_tree, chunk_objectid,
+ chunk_offset);
+ if (ret)
+ goto done;
+ }
+
+done:
+ btrfs_free_path(path);
+ return ret;
+}
+
int btrfs_add_system_chunk(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
struct btrfs_key *key,
return 0;
}
+static u64 chunk_bytes_by_type(u64 type, u64 calc_size, int num_stripes,
+ int sub_stripes)
+{
+ if (type & (BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_DUP))
+ return calc_size;
+ else if (type & BTRFS_BLOCK_GROUP_RAID10)
+ return calc_size * (num_stripes / sub_stripes);
+ else
+ return calc_size * num_stripes;
+}
+
+
int btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
struct btrfs_root *extent_root, u64 *start,
u64 *num_bytes, u64 type)
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_path *path;
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;
struct list_head *cur;
struct extent_map_tree *em_tree;
struct map_lookup *map;
struct extent_map *em;
+ int min_stripe_size = 1 * 1024 * 1024;
u64 physical;
u64 calc_size = 1024 * 1024 * 1024;
- u64 min_free = calc_size;
+ u64 max_chunk_size = calc_size;
+ u64 min_free;
u64 avail;
u64 max_avail = 0;
+ u64 percent_max;
int num_stripes = 1;
+ int min_stripes = 1;
int sub_stripes = 0;
int looped = 0;
int ret;
int stripe_len = 64 * 1024;
struct btrfs_key key;
+ if ((type & BTRFS_BLOCK_GROUP_RAID1) &&
+ (type & BTRFS_BLOCK_GROUP_DUP)) {
+ WARN_ON(1);
+ type &= ~BTRFS_BLOCK_GROUP_DUP;
+ }
+ dev_list = &extent_root->fs_info->fs_devices->alloc_list;
if (list_empty(dev_list))
return -ENOSPC;
- if (type & (BTRFS_BLOCK_GROUP_RAID0))
+ if (type & (BTRFS_BLOCK_GROUP_RAID0)) {
num_stripes = btrfs_super_num_devices(&info->super_copy);
- if (type & (BTRFS_BLOCK_GROUP_DUP))
+ min_stripes = 2;
+ }
+ if (type & (BTRFS_BLOCK_GROUP_DUP)) {
num_stripes = 2;
+ min_stripes = 2;
+ }
if (type & (BTRFS_BLOCK_GROUP_RAID1)) {
num_stripes = min_t(u64, 2,
btrfs_super_num_devices(&info->super_copy));
+ if (num_stripes < 2)
+ return -ENOSPC;
+ min_stripes = 2;
}
if (type & (BTRFS_BLOCK_GROUP_RAID10)) {
num_stripes = btrfs_super_num_devices(&info->super_copy);
return -ENOSPC;
num_stripes &= ~(u32)1;
sub_stripes = 2;
+ min_stripes = 4;
+ }
+
+ if (type & BTRFS_BLOCK_GROUP_DATA) {
+ max_chunk_size = 10 * calc_size;
+ min_stripe_size = 64 * 1024 * 1024;
+ } else if (type & BTRFS_BLOCK_GROUP_METADATA) {
+ max_chunk_size = 4 * calc_size;
+ min_stripe_size = 32 * 1024 * 1024;
+ } else if (type & BTRFS_BLOCK_GROUP_SYSTEM) {
+ calc_size = 8 * 1024 * 1024;
+ max_chunk_size = calc_size * 2;
+ min_stripe_size = 1 * 1024 * 1024;
}
+
+ path = btrfs_alloc_path();
+ if (!path)
+ return -ENOMEM;
+
+ /* we don't want a chunk larger than 10% of the FS */
+ percent_max = div_factor(btrfs_super_total_bytes(&info->super_copy), 1);
+ max_chunk_size = min(percent_max, max_chunk_size);
+
again:
+ if (calc_size * num_stripes > max_chunk_size) {
+ calc_size = max_chunk_size;
+ do_div(calc_size, num_stripes);
+ do_div(calc_size, stripe_len);
+ calc_size *= stripe_len;
+ }
+ /* we don't want tiny stripes */
+ calc_size = max_t(u64, min_stripe_size, calc_size);
+
+ do_div(calc_size, stripe_len);
+ calc_size *= stripe_len;
+
INIT_LIST_HEAD(&private_devs);
cur = dev_list->next;
index = 0;
if (type & BTRFS_BLOCK_GROUP_DUP)
min_free = calc_size * 2;
+ else
+ min_free = calc_size;
+
+ /* we add 1MB because we never use the first 1MB of the device */
+ min_free += 1024 * 1024;
/* build a private list of devices we will allocate from */
while(index < num_stripes) {
- device = list_entry(cur, struct btrfs_device, dev_list);
+ device = list_entry(cur, struct btrfs_device, dev_alloc_list);
avail = device->total_bytes - device->bytes_used;
cur = cur->next;
- if (avail > max_avail)
- max_avail = avail;
+
if (avail >= min_free) {
- list_move_tail(&device->dev_list, &private_devs);
- index++;
- if (type & BTRFS_BLOCK_GROUP_DUP)
+ u64 ignored_start = 0;
+ ret = find_free_dev_extent(trans, device, path,
+ min_free,
+ &ignored_start);
+ if (ret == 0) {
+ list_move_tail(&device->dev_alloc_list,
+ &private_devs);
index++;
- }
+ if (type & BTRFS_BLOCK_GROUP_DUP)
+ index++;
+ }
+ } else if (avail > max_avail)
+ max_avail = avail;
if (cur == dev_list)
break;
}
if (index < num_stripes) {
list_splice(&private_devs, dev_list);
+ if (index >= min_stripes) {
+ num_stripes = index;
+ if (type & (BTRFS_BLOCK_GROUP_RAID10)) {
+ num_stripes /= sub_stripes;
+ num_stripes *= sub_stripes;
+ }
+ looped = 1;
+ goto again;
+ }
if (!looped && max_avail > 0) {
looped = 1;
calc_size = max_avail;
goto again;
}
+ btrfs_free_path(path);
return -ENOSPC;
}
-
key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
key.type = BTRFS_CHUNK_ITEM_KEY;
ret = find_next_chunk(chunk_root, BTRFS_FIRST_CHUNK_TREE_OBJECTID,
&key.offset);
- if (ret)
+ if (ret) {
+ btrfs_free_path(path);
return ret;
+ }
chunk = kmalloc(btrfs_chunk_item_size(num_stripes), GFP_NOFS);
- if (!chunk)
+ if (!chunk) {
+ btrfs_free_path(path);
return -ENOMEM;
+ }
map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS);
if (!map) {
kfree(chunk);
+ btrfs_free_path(path);
return -ENOMEM;
}
+ btrfs_free_path(path);
+ path = NULL;
stripes = &chunk->stripe;
-
- if (type & (BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_DUP))
- *num_bytes = calc_size;
- else if (type & BTRFS_BLOCK_GROUP_RAID10)
- *num_bytes = calc_size * num_stripes / sub_stripes;
- else
- *num_bytes = calc_size * num_stripes;
+ *num_bytes = chunk_bytes_by_type(type, calc_size,
+ num_stripes, sub_stripes);
index = 0;
-printk("new chunk type %Lu start %Lu size %Lu\n", type, key.offset, *num_bytes);
while(index < num_stripes) {
struct btrfs_stripe *stripe;
BUG_ON(list_empty(&private_devs));
cur = private_devs.next;
- device = list_entry(cur, struct btrfs_device, dev_list);
+ device = list_entry(cur, struct btrfs_device, dev_alloc_list);
/* loop over this device again if we're doing a dup group */
if (!(type & BTRFS_BLOCK_GROUP_DUP) ||
(index == num_stripes - 1))
- list_move_tail(&device->dev_list, dev_list);
+ list_move_tail(&device->dev_alloc_list, dev_list);
ret = btrfs_alloc_dev_extent(trans, device,
info->chunk_root->root_key.objectid,
BTRFS_FIRST_CHUNK_TREE_OBJECTID, key.offset,
calc_size, &dev_offset);
BUG_ON(ret);
-printk("alloc chunk start %Lu size %Lu from dev %Lu type %Lu\n", key.offset, calc_size, device->devid, type);
device->bytes_used += calc_size;
ret = btrfs_update_device(trans, device);
BUG_ON(ret);
em->len = *num_bytes;
em->block_start = 0;
+ if (type & BTRFS_BLOCK_GROUP_SYSTEM) {
+ ret = btrfs_add_system_chunk(trans, chunk_root, &key,
+ chunk, btrfs_chunk_item_size(num_stripes));
+ BUG_ON(ret);
+ }
kfree(chunk);
em_tree = &extent_root->fs_info->mapping_tree.map_tree;
return ret;
}
-int btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw,
- u64 logical, u64 *length,
- struct btrfs_multi_bio **multi_ret, int mirror_num)
+static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw,
+ u64 logical, u64 *length,
+ struct btrfs_multi_bio **multi_ret,
+ int mirror_num, struct page *unplug_page)
{
struct extent_map *em;
struct map_lookup *map;
int stripes_required = 1;
int stripe_index;
int i;
+ int num_stripes;
+ int max_errors = 0;
struct btrfs_multi_bio *multi = NULL;
if (multi_ret && !(rw & (1 << BIO_RW))) {
GFP_NOFS);
if (!multi)
return -ENOMEM;
+
+ atomic_set(&multi->error, 0);
}
spin_lock(&em_tree->lock);
em = lookup_extent_mapping(em_tree, logical, *length);
spin_unlock(&em_tree->lock);
+
+ if (!em && unplug_page)
+ return 0;
+
if (!em) {
printk("unable to find logical %Lu\n", logical);
+ BUG();
}
- BUG_ON(!em);
BUG_ON(em->start > logical || em->start + em->len < logical);
map = (struct map_lookup *)em->bdev;
if (map->type & (BTRFS_BLOCK_GROUP_RAID1 |
BTRFS_BLOCK_GROUP_DUP)) {
stripes_required = map->num_stripes;
+ max_errors = 1;
} else if (map->type & BTRFS_BLOCK_GROUP_RAID10) {
stripes_required = map->sub_stripes;
+ max_errors = 1;
}
}
if (multi_ret && rw == WRITE &&
} else {
*length = em->len - offset;
}
- if (!multi_ret)
+
+ if (!multi_ret && !unplug_page)
goto out;
- multi->num_stripes = 1;
+ num_stripes = 1;
stripe_index = 0;
if (map->type & BTRFS_BLOCK_GROUP_RAID1) {
- if (rw & (1 << BIO_RW))
- multi->num_stripes = map->num_stripes;
- else if (mirror_num) {
+ if (unplug_page || (rw & (1 << BIO_RW)))
+ num_stripes = map->num_stripes;
+ else if (mirror_num)
stripe_index = mirror_num - 1;
- } else {
- int i;
- u64 least = (u64)-1;
- struct btrfs_device *cur;
-
- for (i = 0; i < map->num_stripes; i++) {
- cur = map->stripes[i].dev;
- spin_lock(&cur->io_lock);
- if (cur->total_ios < least) {
- least = cur->total_ios;
- stripe_index = i;
- }
- spin_unlock(&cur->io_lock);
- }
- }
+ else
+ stripe_index = current->pid % map->num_stripes;
+
} else if (map->type & BTRFS_BLOCK_GROUP_DUP) {
if (rw & (1 << BIO_RW))
- multi->num_stripes = map->num_stripes;
+ num_stripes = map->num_stripes;
else if (mirror_num)
stripe_index = mirror_num - 1;
+
} else if (map->type & BTRFS_BLOCK_GROUP_RAID10) {
int factor = map->num_stripes / map->sub_stripes;
- int orig_stripe_nr = stripe_nr;
stripe_index = do_div(stripe_nr, factor);
stripe_index *= map->sub_stripes;
- if (rw & (1 << BIO_RW))
- multi->num_stripes = map->sub_stripes;
+ if (unplug_page || (rw & (1 << BIO_RW)))
+ num_stripes = map->sub_stripes;
else if (mirror_num)
stripe_index += mirror_num - 1;
else
- stripe_index += orig_stripe_nr % map->sub_stripes;
+ stripe_index += current->pid % map->sub_stripes;
} else {
/*
* after this do_div call, stripe_nr is the number of stripes
}
BUG_ON(stripe_index >= map->num_stripes);
- for (i = 0; i < multi->num_stripes; i++) {
- multi->stripes[i].physical =
- map->stripes[stripe_index].physical + stripe_offset +
- stripe_nr * map->stripe_len;
- multi->stripes[i].dev = map->stripes[stripe_index].dev;
+ for (i = 0; i < num_stripes; i++) {
+ if (unplug_page) {
+ struct btrfs_device *device;
+ struct backing_dev_info *bdi;
+
+ device = map->stripes[stripe_index].dev;
+ bdi = blk_get_backing_dev_info(device->bdev);
+ if (bdi->unplug_io_fn) {
+ bdi->unplug_io_fn(bdi, unplug_page);
+ }
+ } else {
+ multi->stripes[i].physical =
+ map->stripes[stripe_index].physical +
+ stripe_offset + stripe_nr * map->stripe_len;
+ multi->stripes[i].dev = map->stripes[stripe_index].dev;
+ }
stripe_index++;
}
- *multi_ret = multi;
+ if (multi_ret) {
+ *multi_ret = multi;
+ multi->num_stripes = num_stripes;
+ multi->max_errors = max_errors;
+ }
out:
free_extent_map(em);
return 0;
}
+int btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw,
+ u64 logical, u64 *length,
+ struct btrfs_multi_bio **multi_ret, int mirror_num)
+{
+ return __btrfs_map_block(map_tree, rw, logical, length, multi_ret,
+ mirror_num, NULL);
+}
+
+int btrfs_unplug_page(struct btrfs_mapping_tree *map_tree,
+ u64 logical, struct page *page)
+{
+ u64 length = PAGE_CACHE_SIZE;
+ return __btrfs_map_block(map_tree, READ, logical, &length,
+ NULL, 0, page);
+}
+
+
#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,23)
static void end_bio_multi_stripe(struct bio *bio, int err)
#else
return 1;
#endif
if (err)
- multi->error = err;
+ atomic_inc(&multi->error);
if (atomic_dec_and_test(&multi->stripes_pending)) {
bio->bi_private = multi->private;
bio->bi_end_io = multi->end_io;
- if (!err && multi->error)
- err = multi->error;
+ /* only send an error to the higher layers if it is
+ * beyond the tolerance of the multi-bio
+ */
+ if (atomic_read(&multi->error) > multi->max_errors)
+ err = -EIO;
+ else
+ err = 0;
kfree(multi);
#if LINUX_VERSION_CODE <= KERNEL_VERSION(2,6,23)
u64 logical = bio->bi_sector << 9;
u64 length = 0;
u64 map_length;
- struct bio_vec *bvec;
struct btrfs_multi_bio *multi = NULL;
- int i;
int ret;
int dev_nr = 0;
int total_devs = 1;
- bio_for_each_segment(bvec, bio, i) {
- length += bvec->bv_len;
- }
-
+ length = bio->bi_size;
map_tree = &root->fs_info->mapping_tree;
map_length = length;
}
bio->bi_sector = multi->stripes[dev_nr].physical >> 9;
dev = multi->stripes[dev_nr].dev;
+
bio->bi_bdev = dev->bdev;
spin_lock(&dev->io_lock);
dev->total_ios++;
return 0;
}
-struct btrfs_device *btrfs_find_device(struct btrfs_root *root, u64 devid)
+struct btrfs_device *btrfs_find_device(struct btrfs_root *root, u64 devid,
+ u8 *uuid)
{
struct list_head *head = &root->fs_info->fs_devices->devices;
- return __find_device(head, devid);
+ return __find_device(head, devid, uuid);
}
static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key,
u64 logical;
u64 length;
u64 devid;
+ u8 uuid[BTRFS_UUID_SIZE];
int num_stripes;
int ret;
int i;
map->stripes[i].physical =
btrfs_stripe_offset_nr(leaf, chunk, i);
devid = btrfs_stripe_devid_nr(leaf, chunk, i);
- map->stripes[i].dev = btrfs_find_device(root, devid);
+ read_extent_buffer(leaf, uuid, (unsigned long)
+ btrfs_stripe_dev_uuid_nr(chunk, i),
+ BTRFS_UUID_SIZE);
+ map->stripes[i].dev = btrfs_find_device(root, devid, uuid);
if (!map->stripes[i].dev) {
kfree(map);
free_extent_map(em);
struct btrfs_device *device;
u64 devid;
int ret;
+ u8 dev_uuid[BTRFS_UUID_SIZE];
+
devid = btrfs_device_id(leaf, dev_item);
- device = btrfs_find_device(root, devid);
+ read_extent_buffer(leaf, dev_uuid,
+ (unsigned long)btrfs_device_uuid(dev_item),
+ BTRFS_UUID_SIZE);
+ device = btrfs_find_device(root, devid, dev_uuid);
if (!device) {
printk("warning devid %Lu not found already\n", devid);
device = kzalloc(sizeof(*device), GFP_NOFS);
return -ENOMEM;
list_add(&device->dev_list,
&root->fs_info->fs_devices->devices);
+ list_add(&device->dev_alloc_list,
+ &root->fs_info->fs_devices->alloc_list);
device->barriers = 1;
spin_lock_init(&device->io_lock);
}
struct extent_buffer *sb = root->fs_info->sb_buffer;
struct btrfs_disk_key *disk_key;
struct btrfs_chunk *chunk;
- struct btrfs_key key;
+ u8 *ptr;
+ unsigned long sb_ptr;
+ int ret = 0;
u32 num_stripes;
u32 array_size;
u32 len = 0;
- u8 *ptr;
- unsigned long sb_ptr;
u32 cur;
- int ret;
+ struct btrfs_key key;
array_size = btrfs_super_sys_array_size(super_copy);
- /*
- * we do this loop twice, once for the device items and
- * once for all of the chunks. This way there are device
- * structs filled in for every chunk
- */
ptr = super_copy->sys_chunk_array;
sb_ptr = offsetof(struct btrfs_super_block, sys_chunk_array);
cur = 0;
if (key.type == BTRFS_CHUNK_ITEM_KEY) {
chunk = (struct btrfs_chunk *)sb_ptr;
ret = read_one_chunk(root, &key, sb, chunk);
- BUG_ON(ret);
+ if (ret)
+ break;
num_stripes = btrfs_chunk_num_stripes(sb, chunk);
len = btrfs_chunk_item_size(num_stripes);
} else {
- BUG();
+ ret = -EIO;
+ break;
}
ptr += len;
sb_ptr += len;
cur += len;
}
- return 0;
+ return ret;
}
int btrfs_read_chunk_tree(struct btrfs_root *root)
projects / deliverable / linux.git / blobdiff