#include "transaction.h"
#include "print-tree.h"
#include "volumes.h"
+#include "math.h"
struct stripe {
struct btrfs_device *dev;
u64 physical;
};
-struct map_lookup {
- struct cache_extent ce;
- u64 type;
- int io_align;
- int io_width;
- int stripe_len;
- int sector_size;
- int num_stripes;
- int sub_stripes;
- struct btrfs_bio_stripe stripes[];
-};
+static inline int nr_parity_stripes(struct map_lookup *map)
+{
+ if (map->type & BTRFS_BLOCK_GROUP_RAID5)
+ return 1;
+ else if (map->type & BTRFS_BLOCK_GROUP_RAID6)
+ return 2;
+ else
+ return 0;
+}
-#define map_lookup_size(n) (sizeof(struct map_lookup) + \
- (sizeof(struct btrfs_bio_stripe) * (n)))
+static inline int nr_data_stripes(struct map_lookup *map)
+{
+ return map->num_stripes - nr_parity_stripes(map);
+}
+
+#define is_parity_stripe(x) ( ((x) == BTRFS_RAID5_P_STRIPE) || ((x) == BTRFS_RAID6_Q_STRIPE) )
static LIST_HEAD(fs_uuids);
/* 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);
btrfs_stack_device_bytes_used(&disk_super->dev_item);
list_add(&device->dev_list, &fs_devices->devices);
device->fs_devices = fs_devices;
- }
+ } else if (!device->name || strcmp(device->name, path)) {
+ char *name = strdup(path);
+ if (!name)
+ return -ENOMEM;
+ kfree(device->name);
+ device->name = name;
+ }
+
if (found_transid > fs_devices->latest_trans) {
fs_devices->latest_devid = devid;
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;
}
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) {
goto fail;
}
+ if (posix_fadvise(fd, 0, 0, POSIX_FADV_DONTNEED))
+ fprintf(stderr, "Warning, could not drop caches\n");
+
if (device->devid == fs_devices->latest_devid)
fs_devices->latest_bdev = fd;
if (device->devid == fs_devices->lowest_devid)
struct btrfs_key *key,
struct btrfs_chunk *chunk, int item_size)
{
- struct btrfs_super_block *super_copy = &root->fs_info->super_copy;
+ struct btrfs_super_block *super_copy = root->fs_info->super_copy;
struct btrfs_disk_key disk_key;
u32 array_size;
u8 *ptr;
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)
{
return calc_size;
else if (type & BTRFS_BLOCK_GROUP_RAID10)
return calc_size * (num_stripes / sub_stripes);
+ else if (type & BTRFS_BLOCK_GROUP_RAID5)
+ return calc_size * (num_stripes - 1);
+ else if (type & BTRFS_BLOCK_GROUP_RAID6)
+ return calc_size * (num_stripes - 2);
else
return calc_size * 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;
+}
+
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_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;
int index;
int stripe_len = 64 * 1024;
struct btrfs_key key;
+ u64 offset;
if (list_empty(dev_list)) {
return -ENOSPC;
}
if (type & (BTRFS_BLOCK_GROUP_RAID0 | BTRFS_BLOCK_GROUP_RAID1 |
+ BTRFS_BLOCK_GROUP_RAID5 | BTRFS_BLOCK_GROUP_RAID6 |
BTRFS_BLOCK_GROUP_RAID10 |
BTRFS_BLOCK_GROUP_DUP)) {
if (type & BTRFS_BLOCK_GROUP_SYSTEM) {
}
if (type & BTRFS_BLOCK_GROUP_RAID1) {
num_stripes = min_t(u64, 2,
- btrfs_super_num_devices(&info->super_copy));
+ btrfs_super_num_devices(info->super_copy));
if (num_stripes < 2)
return -ENOSPC;
min_stripes = 2;
min_stripes = 2;
}
if (type & (BTRFS_BLOCK_GROUP_RAID0)) {
- num_stripes = btrfs_super_num_devices(&info->super_copy);
+ num_stripes = btrfs_super_num_devices(info->super_copy);
min_stripes = 2;
}
if (type & (BTRFS_BLOCK_GROUP_RAID10)) {
- num_stripes = btrfs_super_num_devices(&info->super_copy);
+ num_stripes = btrfs_super_num_devices(info->super_copy);
if (num_stripes < 4)
return -ENOSPC;
num_stripes &= ~(u32)1;
sub_stripes = 2;
min_stripes = 4;
}
+ if (type & (BTRFS_BLOCK_GROUP_RAID5)) {
+ num_stripes = btrfs_super_num_devices(info->super_copy);
+ if (num_stripes < 2)
+ return -ENOSPC;
+ min_stripes = 2;
+ stripe_len = find_raid56_stripe_len(num_stripes - 1,
+ btrfs_super_stripesize(info->super_copy));
+ }
+ if (type & (BTRFS_BLOCK_GROUP_RAID6)) {
+ num_stripes = btrfs_super_num_devices(info->super_copy);
+ if (num_stripes < 3)
+ return -ENOSPC;
+ min_stripes = 3;
+ stripe_len = find_raid56_stripe_len(num_stripes - 2,
+ btrfs_super_stripesize(info->super_copy));
+ }
/* we don't want a chunk larger than 10% of the FS */
- percent_max = div_factor(btrfs_super_total_bytes(&info->super_copy), 1);
+ percent_max = div_factor(btrfs_super_total_bytes(info->super_copy), 1);
max_chunk_size = min(percent_max, max_chunk_size);
again:
}
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);
+ &offset);
if (ret)
return ret;
+ key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
+ key.type = BTRFS_CHUNK_ITEM_KEY;
+ key.offset = offset;
chunk = kmalloc(btrfs_chunk_item_size(num_stripes), GFP_NOFS);
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;
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) {
{
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;
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;
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);
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);
ret = map->num_stripes;
else if (map->type & BTRFS_BLOCK_GROUP_RAID10)
ret = map->sub_stripes;
+ else if (map->type & BTRFS_BLOCK_GROUP_RAID5)
+ ret = 2;
+ else if (map->type & BTRFS_BLOCK_GROUP_RAID6)
+ ret = 3;
else
ret = 1;
return ret;
}
+int btrfs_next_metadata(struct btrfs_mapping_tree *map_tree, u64 *logical,
+ u64 *size)
+{
+ struct cache_extent *ce;
+ struct map_lookup *map;
+
+ ce = search_cache_extent(&map_tree->cache_tree, *logical);
+
+ while (ce) {
+ ce = next_cache_extent(ce);
+ if (!ce)
+ return -ENOENT;
+
+ map = container_of(ce, struct map_lookup, ce);
+ if (map->type & BTRFS_BLOCK_GROUP_METADATA) {
+ *logical = ce->start;
+ *size = ce->size;
+ return 0;
+ }
+ }
+
+ return -ENOENT;
+}
+
int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree,
u64 chunk_start, u64 physical, u64 devid,
u64 **logical, int *naddrs, int *stripe_len)
u64 bytenr;
u64 length;
u64 stripe_nr;
+ 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);
length = ce->size;
+ rmap_len = map->stripe_len;
if (map->type & BTRFS_BLOCK_GROUP_RAID10)
length = ce->size / (map->num_stripes / map->sub_stripes);
else if (map->type & BTRFS_BLOCK_GROUP_RAID0)
length = ce->size / map->num_stripes;
+ else if (map->type & (BTRFS_BLOCK_GROUP_RAID5 |
+ BTRFS_BLOCK_GROUP_RAID6)) {
+ length = ce->size / nr_data_stripes(map);
+ rmap_len = map->stripe_len * nr_data_stripes(map);
+ }
buf = kzalloc(sizeof(u64) * map->num_stripes, GFP_NOFS);
map->sub_stripes;
} else if (map->type & BTRFS_BLOCK_GROUP_RAID0) {
stripe_nr = stripe_nr * map->num_stripes + i;
- }
- bytenr = ce->start + stripe_nr * map->stripe_len;
+ } /* else if RAID[56], multiply by nr_data_stripes().
+ * Alternatively, just use rmap_len below instead of
+ * map->stripe_len */
+
+ bytenr = ce->start + stripe_nr * rmap_len;
for (j = 0; j < nr; j++) {
if (buf[j] == bytenr)
break;
*logical = buf;
*naddrs = nr;
- *stripe_len = map->stripe_len;
+ *stripe_len = rmap_len;
return 0;
}
+static inline int parity_smaller(u64 a, u64 b)
+{
+ return a > b;
+}
+
+/* Bubble-sort the stripe set to put the parity/syndrome stripes last */
+static void sort_parity_stripes(struct btrfs_multi_bio *bbio, u64 *raid_map)
+{
+ struct btrfs_bio_stripe s;
+ int i;
+ u64 l;
+ int again = 1;
+
+ while (again) {
+ again = 0;
+ for (i = 0; i < bbio->num_stripes - 1; i++) {
+ if (parity_smaller(raid_map[i], raid_map[i+1])) {
+ s = bbio->stripes[i];
+ l = raid_map[i];
+ bbio->stripes[i] = bbio->stripes[i+1];
+ raid_map[i] = raid_map[i+1];
+ bbio->stripes[i+1] = s;
+ raid_map[i+1] = l;
+ again = 1;
+ }
+ }
+ }
+}
+
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 btrfs_multi_bio **multi_ret, int mirror_num,
+ u64 **raid_map_ret)
+{
+ return __btrfs_map_block(map_tree, rw, logical, length, NULL,
+ multi_ret, mirror_num, raid_map_ret);
+}
+
+int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw,
+ u64 logical, u64 *length, u64 *type,
+ struct btrfs_multi_bio **multi_ret, int mirror_num,
+ u64 **raid_map_ret)
{
struct cache_extent *ce;
struct map_lookup *map;
u64 offset;
u64 stripe_offset;
u64 stripe_nr;
+ u64 *raid_map = NULL;
int stripes_allocated = 8;
int stripes_required = 1;
int stripe_index;
stripes_allocated = 1;
}
again:
+ ce = search_cache_extent(&map_tree->cache_tree, logical);
+ if (!ce) {
+ if (multi)
+ kfree(multi);
+ return -ENOENT;
+ }
+ if (ce->start > logical || ce->start + ce->size < logical) {
+ if (multi)
+ kfree(multi);
+ return -ENOENT;
+ }
+
if (multi_ret) {
multi = kzalloc(btrfs_multi_bio_size(stripes_allocated),
GFP_NOFS);
if (!multi)
return -ENOMEM;
}
-
- ce = find_first_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);
offset = logical - ce->start;
stripes_required = map->sub_stripes;
}
}
+ if (map->type & (BTRFS_BLOCK_GROUP_RAID5 | BTRFS_BLOCK_GROUP_RAID6)
+ && multi_ret && ((rw & WRITE) || mirror_num > 1) && raid_map_ret) {
+ /* RAID[56] write or recovery. Return all stripes */
+ stripes_required = map->num_stripes;
+
+ /* Only allocate the map if we've already got a large enough multi_ret */
+ if (stripes_allocated >= stripes_required) {
+ raid_map = kmalloc(sizeof(u64) * map->num_stripes, GFP_NOFS);
+ if (!raid_map) {
+ kfree(multi);
+ return -ENOMEM;
+ }
+ }
+ }
+
/* if our multi bio struct is too small, back off and try again */
- if (multi_ret && rw == WRITE &&
- stripes_allocated < stripes_required) {
- stripes_allocated = map->num_stripes;
+ if (multi_ret && stripes_allocated < stripes_required) {
+ stripes_allocated = stripes_required;
kfree(multi);
+ multi = NULL;
goto again;
}
stripe_nr = offset;
stripe_offset = offset - stripe_offset;
if (map->type & (BTRFS_BLOCK_GROUP_RAID0 | BTRFS_BLOCK_GROUP_RAID1 |
+ BTRFS_BLOCK_GROUP_RAID5 | BTRFS_BLOCK_GROUP_RAID6 |
BTRFS_BLOCK_GROUP_RAID10 |
BTRFS_BLOCK_GROUP_DUP)) {
/* we limit the length of each bio to what fits in a stripe */
multi->num_stripes = map->sub_stripes;
else if (mirror_num)
stripe_index += mirror_num - 1;
- else
- stripe_index = stripe_nr % map->sub_stripes;
stripe_nr = stripe_nr / factor;
} else if (map->type & BTRFS_BLOCK_GROUP_DUP) {
multi->num_stripes = map->num_stripes;
else if (mirror_num)
stripe_index = mirror_num - 1;
+ } else if (map->type & (BTRFS_BLOCK_GROUP_RAID5 |
+ BTRFS_BLOCK_GROUP_RAID6)) {
+
+ if (raid_map) {
+ int i, rot;
+ u64 tmp;
+ u64 raid56_full_stripe_start;
+ u64 full_stripe_len = nr_data_stripes(map) * map->stripe_len;
+
+ /*
+ * align the start of our data stripe in the logical
+ * address space
+ */
+ raid56_full_stripe_start = offset / full_stripe_len;
+ raid56_full_stripe_start *= full_stripe_len;
+
+ /* get the data stripe number */
+ stripe_nr = raid56_full_stripe_start / map->stripe_len;
+ stripe_nr = stripe_nr / nr_data_stripes(map);
+
+ /* Work out the disk rotation on this stripe-set */
+ rot = stripe_nr % map->num_stripes;
+
+ /* Fill in the logical address of each stripe */
+ tmp = stripe_nr * nr_data_stripes(map);
+
+ for (i = 0; i < nr_data_stripes(map); i++)
+ raid_map[(i+rot) % map->num_stripes] =
+ ce->start + (tmp + i) * map->stripe_len;
+
+ raid_map[(i+rot) % map->num_stripes] = BTRFS_RAID5_P_STRIPE;
+ if (map->type & BTRFS_BLOCK_GROUP_RAID6)
+ raid_map[(i+rot+1) % map->num_stripes] = BTRFS_RAID6_Q_STRIPE;
+
+ *length = map->stripe_len;
+ stripe_index = 0;
+ stripe_offset = 0;
+ multi->num_stripes = map->num_stripes;
+ } else {
+ stripe_index = stripe_nr % nr_data_stripes(map);
+ stripe_nr = stripe_nr / nr_data_stripes(map);
+
+ /*
+ * Mirror #0 or #1 means the original data block.
+ * Mirror #2 is RAID5 parity block.
+ * Mirror #3 is RAID6 Q block.
+ */
+ if (mirror_num > 1)
+ stripe_index = nr_data_stripes(map) + mirror_num - 2;
+
+ /* We distribute the parity blocks across stripes */
+ stripe_index = (stripe_nr + stripe_index) % map->num_stripes;
+ }
} else {
/*
* after this do_div call, stripe_nr is the number of stripes
stripe_index++;
}
*multi_ret = multi;
+
+ if (type)
+ *type = map->type;
+
+ if (raid_map) {
+ sort_parity_stripes(multi, raid_map);
+ *raid_map_ret = raid_map;
+ }
out:
return 0;
}
return NULL;
}
+struct btrfs_device *
+btrfs_find_device_by_devid(struct btrfs_fs_devices *fs_devices,
+ u64 devid, int instance)
+{
+ struct list_head *head = &fs_devices->devices;
+ struct btrfs_device *dev;
+ int num_found = 0;
+
+ 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)
{
list_for_each(cur, &fs_devices->devices) {
num_stripes++;
}
- map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS);
+ map = kmalloc(btrfs_map_lookup_size(num_stripes), GFP_NOFS);
if (!map)
return -ENOMEM;
map->stripes[i].dev = device;
i++;
}
- ret = insert_existing_cache_extent(&map_tree->cache_tree, &map->ce);
+ ret = insert_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 = lookup_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,
+ ret = insert_cache_extent(&map_tree->cache_tree,
&map->ce);
}
BUG_ON(ret);
int readonly = 0;
int i;
- ce = find_first_cache_extent(&map_tree->cache_tree, chunk_offset);
+ ce = search_cache_extent(&map_tree->cache_tree, chunk_offset);
BUG_ON(!ce);
map = container_of(ce, struct map_lookup, ce);
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) {
}
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;
}
}
- 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;
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);
}
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;
+ struct btrfs_super_block *super_copy = root->fs_info->super_copy;
struct extent_buffer *sb;
struct btrfs_disk_key *disk_key;
struct btrfs_chunk *chunk;
u8 *ptr;
unsigned long sb_ptr;
u32 cur;
- int ret;
+ int ret = 0;
sb = btrfs_find_create_tree_block(root, BTRFS_SUPER_INFO_OFFSET,
BTRFS_SUPER_INFO_SIZE);
if (!sb)
return -ENOMEM;
btrfs_set_buffer_uptodate(sb);
- write_extent_buffer(sb, super_copy, 0, BTRFS_SUPER_INFO_SIZE);
+ write_extent_buffer(sb, super_copy, 0, sizeof(*super_copy));
array_size = btrfs_super_sys_array_size(super_copy);
/*
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 {
cur += len;
}
free_extent_buffer(sb);
- return 0;
+ return ret;
}
int btrfs_read_chunk_tree(struct btrfs_root *root)
goto again;
}
- btrfs_free_path(path);
ret = 0;
error:
+ btrfs_free_path(path);
return ret;
}
{
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[multi->num_stripes], *p_eb = NULL, *q_eb = NULL;
+ int i;
+ int j;
+ int ret;
+ int alloc_size = eb->len;
+
+ 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[multi->num_stripes];
+ 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);
+ } 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]);
+ }
+ return 0;
+}