#include "print-tree.h"
#include "transaction.h"
#include "volumes.h"
+#include "raid56.h"
#include "locking.h"
#include "free-space-cache.h"
#include "math.h"
*actual_bytes = discarded_bytes;
+ if (ret == -EOPNOTSUPP)
+ ret = 0;
return ret;
}
return ret;
}
+static int refs_newer(struct btrfs_delayed_ref_root *delayed_refs, int seq,
+ int count)
+{
+ int val = atomic_read(&delayed_refs->ref_seq);
+
+ if (val < seq || val >= seq + count)
+ return 1;
+ return 0;
+}
+
/*
* this starts processing the delayed reference count updates and
* extent insertions we have queued up so far. count can be
delayed_refs = &trans->transaction->delayed_refs;
INIT_LIST_HEAD(&cluster);
+ if (count == 0) {
+ count = delayed_refs->num_entries * 2;
+ run_most = 1;
+ }
+
+ if (!run_all && !run_most) {
+ int old;
+ int seq = atomic_read(&delayed_refs->ref_seq);
+
+progress:
+ old = atomic_cmpxchg(&delayed_refs->procs_running_refs, 0, 1);
+ if (old) {
+ DEFINE_WAIT(__wait);
+ if (delayed_refs->num_entries < 16348)
+ return 0;
+
+ prepare_to_wait(&delayed_refs->wait, &__wait,
+ TASK_UNINTERRUPTIBLE);
+
+ old = atomic_cmpxchg(&delayed_refs->procs_running_refs, 0, 1);
+ if (old) {
+ schedule();
+ finish_wait(&delayed_refs->wait, &__wait);
+
+ if (!refs_newer(delayed_refs, seq, 256))
+ goto progress;
+ else
+ return 0;
+ } else {
+ finish_wait(&delayed_refs->wait, &__wait);
+ goto again;
+ }
+ }
+
+ } else {
+ atomic_inc(&delayed_refs->procs_running_refs);
+ }
+
again:
loops = 0;
spin_lock(&delayed_refs->lock);
delayed_refs->run_delayed_start = find_middle(&delayed_refs->root);
#endif
- if (count == 0) {
- count = delayed_refs->num_entries * 2;
- run_most = 1;
- }
while (1) {
if (!(run_all || run_most) &&
delayed_refs->num_heads_ready < 64)
if (ret < 0) {
spin_unlock(&delayed_refs->lock);
btrfs_abort_transaction(trans, root, ret);
+ atomic_dec(&delayed_refs->procs_running_refs);
return ret;
}
+ atomic_add(ret, &delayed_refs->ref_seq);
+
count -= min_t(unsigned long, ret, count);
if (count == 0)
goto again;
}
out:
+ atomic_dec(&delayed_refs->procs_running_refs);
+ smp_mb();
+ if (waitqueue_active(&delayed_refs->wait))
+ wake_up(&delayed_refs->wait);
+
spin_unlock(&delayed_refs->lock);
assert_qgroups_uptodate(trans);
return 0;
u64 num_devices = root->fs_info->fs_devices->rw_devices +
root->fs_info->fs_devices->missing_devices;
u64 target;
+ u64 tmp;
/*
* see if restripe for this chunk_type is in progress, if so
}
spin_unlock(&root->fs_info->balance_lock);
+ /* First, mask out the RAID levels which aren't possible */
if (num_devices == 1)
- flags &= ~(BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID0);
+ flags &= ~(BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID0 |
+ BTRFS_BLOCK_GROUP_RAID5);
+ if (num_devices < 3)
+ flags &= ~BTRFS_BLOCK_GROUP_RAID6;
if (num_devices < 4)
flags &= ~BTRFS_BLOCK_GROUP_RAID10;
- if ((flags & BTRFS_BLOCK_GROUP_DUP) &&
- (flags & (BTRFS_BLOCK_GROUP_RAID1 |
- BTRFS_BLOCK_GROUP_RAID10))) {
- flags &= ~BTRFS_BLOCK_GROUP_DUP;
- }
-
- if ((flags & BTRFS_BLOCK_GROUP_RAID1) &&
- (flags & BTRFS_BLOCK_GROUP_RAID10)) {
- flags &= ~BTRFS_BLOCK_GROUP_RAID1;
- }
+ tmp = flags & (BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID0 |
+ BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID5 |
+ BTRFS_BLOCK_GROUP_RAID6 | BTRFS_BLOCK_GROUP_RAID10);
+ flags &= ~tmp;
- if ((flags & BTRFS_BLOCK_GROUP_RAID0) &&
- ((flags & BTRFS_BLOCK_GROUP_RAID1) |
- (flags & BTRFS_BLOCK_GROUP_RAID10) |
- (flags & BTRFS_BLOCK_GROUP_DUP))) {
- flags &= ~BTRFS_BLOCK_GROUP_RAID0;
- }
+ if (tmp & BTRFS_BLOCK_GROUP_RAID6)
+ tmp = BTRFS_BLOCK_GROUP_RAID6;
+ else if (tmp & BTRFS_BLOCK_GROUP_RAID5)
+ tmp = BTRFS_BLOCK_GROUP_RAID5;
+ else if (tmp & BTRFS_BLOCK_GROUP_RAID10)
+ tmp = BTRFS_BLOCK_GROUP_RAID10;
+ else if (tmp & BTRFS_BLOCK_GROUP_RAID1)
+ tmp = BTRFS_BLOCK_GROUP_RAID1;
+ else if (tmp & BTRFS_BLOCK_GROUP_RAID0)
+ tmp = BTRFS_BLOCK_GROUP_RAID0;
- return extended_to_chunk(flags);
+ return extended_to_chunk(flags | tmp);
}
static u64 get_alloc_profile(struct btrfs_root *root, u64 flags)
u64 btrfs_get_alloc_profile(struct btrfs_root *root, int data)
{
u64 flags;
+ u64 ret;
if (data)
flags = BTRFS_BLOCK_GROUP_DATA;
else
flags = BTRFS_BLOCK_GROUP_METADATA;
- return get_alloc_profile(root, flags);
+ ret = get_alloc_profile(root, flags);
+ return ret;
}
/*
{
u64 num_dev;
- if (type & BTRFS_BLOCK_GROUP_RAID10 ||
- type & BTRFS_BLOCK_GROUP_RAID0)
+ if (type & (BTRFS_BLOCK_GROUP_RAID10 |
+ BTRFS_BLOCK_GROUP_RAID0 |
+ BTRFS_BLOCK_GROUP_RAID5 |
+ BTRFS_BLOCK_GROUP_RAID6))
num_dev = root->fs_info->fs_devices->rw_devices;
else if (type & BTRFS_BLOCK_GROUP_RAID1)
num_dev = 2;
/*
* If we have dup, raid1 or raid10 then only half of the free
- * space is actually useable.
+ * space is actually useable. For raid56, the space info used
+ * doesn't include the parity drive, so we don't have to
+ * change the math
*/
if (profile & (BTRFS_BLOCK_GROUP_DUP |
BTRFS_BLOCK_GROUP_RAID1 |
* We make the other tasks wait for the flush only when we can flush
* all things.
*/
- if (ret && flush == BTRFS_RESERVE_FLUSH_ALL) {
+ if (ret && flush != BTRFS_RESERVE_NO_FLUSH) {
flushing = true;
space_info->flush = 1;
}
return ret;
}
-static u64 stripe_align(struct btrfs_root *root, u64 val)
+static u64 stripe_align(struct btrfs_root *root,
+ struct btrfs_block_group_cache *cache,
+ u64 val, u64 num_bytes)
{
- u64 mask = ((u64)root->stripesize - 1);
- u64 ret = (val + mask) & ~mask;
+ u64 mask;
+ u64 ret;
+ mask = ((u64)root->stripesize - 1);
+ ret = (val + mask) & ~mask;
return ret;
}
index = 2;
else if (flags & BTRFS_BLOCK_GROUP_RAID0)
index = 3;
+ else if (flags & BTRFS_BLOCK_GROUP_RAID5)
+ index = 5;
+ else if (flags & BTRFS_BLOCK_GROUP_RAID6)
+ index = 6;
else
- index = 4;
-
+ index = 4; /* BTRFS_BLOCK_GROUP_SINGLE */
return index;
}
int empty_cluster = 2 * 1024 * 1024;
struct btrfs_space_info *space_info;
int loop = 0;
- int index = 0;
+ int index = __get_raid_index(data);
int alloc_type = (data & BTRFS_BLOCK_GROUP_DATA) ?
RESERVE_ALLOC_NO_ACCOUNT : RESERVE_ALLOC;
bool found_uncached_bg = false;
if (!block_group_bits(block_group, data)) {
u64 extra = BTRFS_BLOCK_GROUP_DUP |
BTRFS_BLOCK_GROUP_RAID1 |
+ BTRFS_BLOCK_GROUP_RAID5 |
+ BTRFS_BLOCK_GROUP_RAID6 |
BTRFS_BLOCK_GROUP_RAID10;
/*
* lets look there
*/
if (last_ptr) {
+ unsigned long aligned_cluster;
/*
* the refill lock keeps out other
* people trying to start a new cluster
goto unclustered_alloc;
}
+ aligned_cluster = max_t(unsigned long,
+ empty_cluster + empty_size,
+ block_group->full_stripe_len);
+
/* allocate a cluster in this block group */
ret = btrfs_find_space_cluster(trans, root,
block_group, last_ptr,
search_start, num_bytes,
- empty_cluster + empty_size);
+ aligned_cluster);
if (ret == 0) {
/*
* now pull our allocation out of this
goto loop;
}
checks:
- search_start = stripe_align(root, offset);
+ search_start = stripe_align(root, used_block_group,
+ offset, num_bytes);
/* move on to the next group */
if (search_start + num_bytes >
&wc->flags[level]);
if (ret < 0) {
btrfs_tree_unlock_rw(eb, path->locks[level]);
+ path->locks[level] = 0;
return ret;
}
BUG_ON(wc->refs[level] == 0);
if (wc->refs[level] == 1) {
btrfs_tree_unlock_rw(eb, path->locks[level]);
+ path->locks[level] = 0;
return 1;
}
}
root->fs_info->fs_devices->missing_devices;
stripped = BTRFS_BLOCK_GROUP_RAID0 |
+ BTRFS_BLOCK_GROUP_RAID5 | BTRFS_BLOCK_GROUP_RAID6 |
BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10;
if (num_devices == 1) {
btrfs_release_path(path);
cache->flags = btrfs_block_group_flags(&cache->item);
cache->sectorsize = root->sectorsize;
-
+ cache->full_stripe_len = btrfs_full_stripe_len(root,
+ &root->fs_info->mapping_tree,
+ found_key.objectid);
btrfs_init_free_space_ctl(cache);
/*
if (!(get_alloc_profile(root, space_info->flags) &
(BTRFS_BLOCK_GROUP_RAID10 |
BTRFS_BLOCK_GROUP_RAID1 |
+ BTRFS_BLOCK_GROUP_RAID5 |
+ BTRFS_BLOCK_GROUP_RAID6 |
BTRFS_BLOCK_GROUP_DUP)))
continue;
/*
cache->key.type = BTRFS_BLOCK_GROUP_ITEM_KEY;
cache->sectorsize = root->sectorsize;
cache->fs_info = root->fs_info;
+ cache->full_stripe_len = btrfs_full_stripe_len(root,
+ &root->fs_info->mapping_tree,
+ chunk_offset);
atomic_set(&cache->count, 1);
spin_lock_init(&cache->lock);
}
static struct btrfs_free_space *
-find_free_space(struct btrfs_free_space_ctl *ctl, u64 *offset, u64 *bytes)
+find_free_space(struct btrfs_free_space_ctl *ctl, u64 *offset, u64 *bytes,
+ unsigned long align)
{
struct btrfs_free_space *entry;
struct rb_node *node;
+ u64 ctl_off;
+ u64 tmp;
+ u64 align_off;
int ret;
if (!ctl->free_space_offset.rb_node)
if (entry->bytes < *bytes)
continue;
+ /* make sure the space returned is big enough
+ * to match our requested alignment
+ */
+ if (*bytes >= align) {
+ ctl_off = entry->offset - ctl->start;
+ tmp = ctl_off + align - 1;;
+ do_div(tmp, align);
+ tmp = tmp * align + ctl->start;
+ align_off = tmp - entry->offset;
+ } else {
+ align_off = 0;
+ tmp = entry->offset;
+ }
+
+ if (entry->bytes < *bytes + align_off)
+ continue;
+
if (entry->bitmap) {
- ret = search_bitmap(ctl, entry, offset, bytes);
- if (!ret)
+ ret = search_bitmap(ctl, entry, &tmp, bytes);
+ if (!ret) {
+ *offset = tmp;
return entry;
+ }
continue;
}
- *offset = entry->offset;
- *bytes = entry->bytes;
+ *offset = tmp;
+ *bytes = entry->bytes - align_off;
return entry;
}
{
struct btrfs_free_space_ctl *ctl = block_group->free_space_ctl;
struct btrfs_free_space *info;
- int ret = 0;
+ int ret;
+ bool re_search = false;
spin_lock(&ctl->tree_lock);
again:
+ ret = 0;
if (!bytes)
goto out_lock;
info = tree_search_offset(ctl, offset_to_bitmap(ctl, offset),
1, 0);
if (!info) {
- /* the tree logging code might be calling us before we
- * have fully loaded the free space rbtree for this
- * block group. So it is possible the entry won't
- * be in the rbtree yet at all. The caching code
- * will make sure not to put it in the rbtree if
- * the logging code has pinned it.
+ /*
+ * If we found a partial bit of our free space in a
+ * bitmap but then couldn't find the other part this may
+ * be a problem, so WARN about it.
*/
+ WARN_ON(re_search);
goto out_lock;
}
}
+ re_search = false;
if (!info->bitmap) {
unlink_free_space(ctl, info);
if (offset == info->offset) {
}
ret = remove_from_bitmap(ctl, info, &offset, &bytes);
- if (ret == -EAGAIN)
+ if (ret == -EAGAIN) {
+ re_search = true;
goto again;
+ }
BUG_ON(ret); /* logic error */
out_lock:
spin_unlock(&ctl->tree_lock);
struct btrfs_free_space *entry = NULL;
u64 bytes_search = bytes + empty_size;
u64 ret = 0;
+ u64 align_gap = 0;
+ u64 align_gap_len = 0;
spin_lock(&ctl->tree_lock);
- entry = find_free_space(ctl, &offset, &bytes_search);
+ entry = find_free_space(ctl, &offset, &bytes_search,
+ block_group->full_stripe_len);
if (!entry)
goto out;
if (!entry->bytes)
free_bitmap(ctl, entry);
} else {
+
unlink_free_space(ctl, entry);
- entry->offset += bytes;
- entry->bytes -= bytes;
+ align_gap_len = offset - entry->offset;
+ align_gap = entry->offset;
+
+ entry->offset = offset + bytes;
+ WARN_ON(entry->bytes < bytes + align_gap_len);
+
+ entry->bytes -= bytes + align_gap_len;
if (!entry->bytes)
kmem_cache_free(btrfs_free_space_cachep, entry);
else
out:
spin_unlock(&ctl->tree_lock);
+ if (align_gap_len)
+ __btrfs_add_free_space(ctl, align_gap, align_gap_len);
return ret;
}
#include <linux/slab.h>
#include <linux/ratelimit.h>
#include <linux/mount.h>
+#include <linux/blkdev.h>
#include "compat.h"
#include "ctree.h"
#include "disk-io.h"
[S_IFLNK >> S_SHIFT] = BTRFS_FT_SYMLINK,
};
- static int btrfs_setsize(struct inode *inode, loff_t newsize);
+ static int btrfs_setsize(struct inode *inode, struct iattr *attr);
static int btrfs_truncate(struct inode *inode);
static int btrfs_finish_ordered_io(struct btrfs_ordered_extent *ordered_extent);
static noinline int cow_file_range(struct inode *inode,
* extent_io.c merge_bio_hook, this must check the chunk tree to make sure
* we don't create bios that span stripes or chunks
*/
-int btrfs_merge_bio_hook(struct page *page, unsigned long offset,
+int btrfs_merge_bio_hook(int rw, struct page *page, unsigned long offset,
size_t size, struct bio *bio,
unsigned long bio_flags)
{
length = bio->bi_size;
map_length = length;
- ret = btrfs_map_block(root->fs_info, READ, logical,
+ ret = btrfs_map_block(root->fs_info, rw, logical,
&map_length, NULL, 0);
/* Will always return 0 with map_multi == NULL */
BUG_ON(ret < 0);
continue;
}
nr_truncate++;
+
+ /* 1 for the orphan item deletion. */
+ trans = btrfs_start_transaction(root, 1);
+ if (IS_ERR(trans)) {
+ ret = PTR_ERR(trans);
+ goto out;
+ }
+ ret = btrfs_orphan_add(trans, inode);
+ btrfs_end_transaction(trans, root);
+ if (ret)
+ goto out;
+
ret = btrfs_truncate(inode);
} else {
nr_unlink++;
block_end - cur_offset, 0);
if (IS_ERR(em)) {
err = PTR_ERR(em);
+ em = NULL;
break;
}
last_byte = min(extent_map_end(em), block_end);
return err;
}
- static int btrfs_setsize(struct inode *inode, loff_t newsize)
+ static int btrfs_setsize(struct inode *inode, struct iattr *attr)
{
struct btrfs_root *root = BTRFS_I(inode)->root;
struct btrfs_trans_handle *trans;
loff_t oldsize = i_size_read(inode);
+ loff_t newsize = attr->ia_size;
+ int mask = attr->ia_valid;
int ret;
if (newsize == oldsize)
return 0;
+ /*
+ * The regular truncate() case without ATTR_CTIME and ATTR_MTIME is a
+ * special case where we need to update the times despite not having
+ * these flags set. For all other operations the VFS set these flags
+ * explicitly if it wants a timestamp update.
+ */
+ if (newsize != oldsize && (!(mask & (ATTR_CTIME | ATTR_MTIME))))
+ inode->i_ctime = inode->i_mtime = current_fs_time(inode->i_sb);
+
if (newsize > oldsize) {
truncate_pagecache(inode, oldsize, newsize);
ret = btrfs_cont_expand(inode, oldsize, newsize);
set_bit(BTRFS_INODE_ORDERED_DATA_CLOSE,
&BTRFS_I(inode)->runtime_flags);
+ /*
+ * 1 for the orphan item we're going to add
+ * 1 for the orphan item deletion.
+ */
+ trans = btrfs_start_transaction(root, 2);
+ if (IS_ERR(trans))
+ return PTR_ERR(trans);
+
+ /*
+ * We need to do this in case we fail at _any_ point during the
+ * actual truncate. Once we do the truncate_setsize we could
+ * invalidate pages which forces any outstanding ordered io to
+ * be instantly completed which will give us extents that need
+ * to be truncated. If we fail to get an orphan inode down we
+ * could have left over extents that were never meant to live,
+ * so we need to garuntee from this point on that everything
+ * will be consistent.
+ */
+ ret = btrfs_orphan_add(trans, inode);
+ btrfs_end_transaction(trans, root);
+ if (ret)
+ return ret;
+
/* we don't support swapfiles, so vmtruncate shouldn't fail */
truncate_setsize(inode, newsize);
ret = btrfs_truncate(inode);
+ if (ret && inode->i_nlink)
+ btrfs_orphan_del(NULL, inode);
}
return ret;
return err;
if (S_ISREG(inode->i_mode) && (attr->ia_valid & ATTR_SIZE)) {
- err = btrfs_setsize(inode, attr->ia_size);
+ err = btrfs_setsize(inode, attr);
if (err)
return err;
}
return em;
if (em) {
/*
- * if our em maps to a hole, there might
- * actually be delalloc bytes behind it
+ * if our em maps to
+ * - a hole or
+ * - a pre-alloc extent,
+ * there might actually be delalloc bytes behind it.
*/
- if (em->block_start != EXTENT_MAP_HOLE)
+ if (em->block_start != EXTENT_MAP_HOLE &&
+ !test_bit(EXTENT_FLAG_PREALLOC, &em->flags))
return em;
else
hole_em = em;
*/
em->block_start = hole_em->block_start;
em->block_len = hole_len;
+ if (test_bit(EXTENT_FLAG_PREALLOC, &hole_em->flags))
+ set_bit(EXTENT_FLAG_PREALLOC, &em->flags);
} else {
em->start = range_start;
em->len = found;
int async_submit = 0;
map_length = orig_bio->bi_size;
- ret = btrfs_map_block(root->fs_info, READ, start_sector << 9,
+ ret = btrfs_map_block(root->fs_info, rw, start_sector << 9,
&map_length, NULL, 0);
if (ret) {
bio_put(orig_bio);
return -EIO;
}
-
if (map_length >= orig_bio->bi_size) {
bio = orig_bio;
goto submit;
}
- async_submit = 1;
+ /* async crcs make it difficult to collect full stripe writes. */
+ if (btrfs_get_alloc_profile(root, 1) &
+ (BTRFS_BLOCK_GROUP_RAID5 | BTRFS_BLOCK_GROUP_RAID6))
+ async_submit = 0;
+ else
+ async_submit = 1;
+
bio = btrfs_dio_bio_alloc(orig_bio->bi_bdev, start_sector, GFP_NOFS);
if (!bio)
return -ENOMEM;
bio->bi_end_io = btrfs_end_dio_bio;
map_length = orig_bio->bi_size;
- ret = btrfs_map_block(root->fs_info, READ,
+ ret = btrfs_map_block(root->fs_info, rw,
start_sector << 9,
&map_length, NULL, 0);
if (ret) {
{
struct file *file = iocb->ki_filp;
struct inode *inode = file->f_mapping->host;
+ ssize_t ret;
if (check_direct_IO(BTRFS_I(inode)->root, rw, iocb, iov,
offset, nr_segs))
return 0;
- return __blockdev_direct_IO(rw, iocb, inode,
+ ret = __blockdev_direct_IO(rw, iocb, inode,
BTRFS_I(inode)->root->fs_info->fs_devices->latest_bdev,
iov, offset, nr_segs, btrfs_get_blocks_direct, NULL,
btrfs_submit_direct, 0);
+ return ret;
}
#define BTRFS_FIEMAP_FLAGS (FIEMAP_FLAG_SYNC)
/*
* 1 for the truncate slack space
- * 1 for the orphan item we're going to add
- * 1 for the orphan item deletion
* 1 for updating the inode.
*/
- trans = btrfs_start_transaction(root, 4);
+ trans = btrfs_start_transaction(root, 2);
if (IS_ERR(trans)) {
err = PTR_ERR(trans);
goto out;
min_size);
BUG_ON(ret);
- ret = btrfs_orphan_add(trans, inode);
- if (ret) {
- btrfs_end_transaction(trans, root);
- goto out;
- }
-
/*
* setattr is responsible for setting the ordered_data_close flag,
* but that is only tested during the last file release. That
ret = btrfs_orphan_del(trans, inode);
if (ret)
err = ret;
- } else if (ret && inode->i_nlink > 0) {
- /*
- * Failed to do the truncate, remove us from the in memory
- * orphan list.
- */
- ret = btrfs_orphan_del(NULL, inode);
}
if (trans) {
*/
int btrfs_start_delalloc_inodes(struct btrfs_root *root, int delay_iput)
{
- struct list_head *head = &root->fs_info->delalloc_inodes;
struct btrfs_inode *binode;
struct inode *inode;
struct btrfs_delalloc_work *work, *next;
struct list_head works;
+ struct list_head splice;
int ret = 0;
if (root->fs_info->sb->s_flags & MS_RDONLY)
return -EROFS;
INIT_LIST_HEAD(&works);
-
+ INIT_LIST_HEAD(&splice);
+ again:
spin_lock(&root->fs_info->delalloc_lock);
- while (!list_empty(head)) {
- binode = list_entry(head->next, struct btrfs_inode,
+ list_splice_init(&root->fs_info->delalloc_inodes, &splice);
+ while (!list_empty(&splice)) {
+ binode = list_entry(splice.next, struct btrfs_inode,
delalloc_inodes);
+
+ list_del_init(&binode->delalloc_inodes);
+
inode = igrab(&binode->vfs_inode);
if (!inode)
- list_del_init(&binode->delalloc_inodes);
+ continue;
+
+ list_add_tail(&binode->delalloc_inodes,
+ &root->fs_info->delalloc_inodes);
spin_unlock(&root->fs_info->delalloc_lock);
- if (inode) {
- work = btrfs_alloc_delalloc_work(inode, 0, delay_iput);
- if (!work) {
- ret = -ENOMEM;
- goto out;
- }
- list_add_tail(&work->list, &works);
- btrfs_queue_worker(&root->fs_info->flush_workers,
- &work->work);
+
+ work = btrfs_alloc_delalloc_work(inode, 0, delay_iput);
+ if (unlikely(!work)) {
+ ret = -ENOMEM;
+ goto out;
}
+ list_add_tail(&work->list, &works);
+ btrfs_queue_worker(&root->fs_info->flush_workers,
+ &work->work);
+
cond_resched();
spin_lock(&root->fs_info->delalloc_lock);
}
spin_unlock(&root->fs_info->delalloc_lock);
+ list_for_each_entry_safe(work, next, &works, list) {
+ list_del_init(&work->list);
+ btrfs_wait_and_free_delalloc_work(work);
+ }
+
+ spin_lock(&root->fs_info->delalloc_lock);
+ if (!list_empty(&root->fs_info->delalloc_inodes)) {
+ spin_unlock(&root->fs_info->delalloc_lock);
+ goto again;
+ }
+ spin_unlock(&root->fs_info->delalloc_lock);
+
/* the filemap_flush will queue IO into the worker threads, but
* we have to make sure the IO is actually started and that
* ordered extents get created before we return
atomic_read(&root->fs_info->async_delalloc_pages) == 0));
}
atomic_dec(&root->fs_info->async_submit_draining);
+ return 0;
out:
list_for_each_entry_safe(work, next, &works, list) {
list_del_init(&work->list);
btrfs_wait_and_free_delalloc_work(work);
}
+
+ if (!list_empty_careful(&splice)) {
+ spin_lock(&root->fs_info->delalloc_lock);
+ list_splice_tail(&splice, &root->fs_info->delalloc_inodes);
+ spin_unlock(&root->fs_info->delalloc_lock);
+ }
return ret;
}
spin_lock_init(&cur_trans->commit_lock);
spin_lock_init(&cur_trans->delayed_refs.lock);
+ atomic_set(&cur_trans->delayed_refs.procs_running_refs, 0);
+ atomic_set(&cur_trans->delayed_refs.ref_seq, 0);
+ init_waitqueue_head(&cur_trans->delayed_refs.wait);
INIT_LIST_HEAD(&cur_trans->pending_snapshots);
list_add_tail(&cur_trans->list, &fs_info->trans_list);
if (!list_empty(&trans->new_bgs))
btrfs_create_pending_block_groups(trans, root);
- while (count < 2) {
+ while (count < 1) {
unsigned long cur = trans->delayed_ref_updates;
trans->delayed_ref_updates = 0;
if (cur &&
}
count++;
}
+
btrfs_trans_release_metadata(trans, root);
trans->block_rsv = NULL;
struct extent_state *cached_state = NULL;
u64 start = 0;
u64 end;
+ struct blk_plug plug;
+ blk_start_plug(&plug);
while (!find_first_extent_bit(dirty_pages, start, &start, &end,
mark, &cached_state)) {
convert_extent_bit(dirty_pages, start, end, EXTENT_NEED_WAIT,
}
if (err)
werr = err;
+ blk_finish_plug(&plug);
return werr;
}
goto cleanup_transaction;
}
- if (cur_trans->aborted) {
+ /* Stop the commit early if ->aborted is set */
+ if (unlikely(ACCESS_ONCE(cur_trans->aborted))) {
ret = cur_trans->aborted;
goto cleanup_transaction;
}
wait_event(cur_trans->writer_wait,
atomic_read(&cur_trans->num_writers) == 1);
+ /* ->aborted might be set after the previous check, so check it */
+ if (unlikely(ACCESS_ONCE(cur_trans->aborted))) {
+ ret = cur_trans->aborted;
+ goto cleanup_transaction;
+ }
/*
* the reloc mutex makes sure that we stop
* the balancing code from coming in and moving
goto cleanup_transaction;
}
+ /*
+ * The tasks which save the space cache and inode cache may also
+ * update ->aborted, check it.
+ */
+ if (unlikely(ACCESS_ONCE(cur_trans->aborted))) {
+ ret = cur_trans->aborted;
+ mutex_unlock(&root->fs_info->tree_log_mutex);
+ mutex_unlock(&root->fs_info->reloc_mutex);
+ goto cleanup_transaction;
+ }
+
btrfs_prepare_extent_commit(trans, root);
cur_trans = root->fs_info->running_transaction;
#include <linux/capability.h>
#include <linux/ratelimit.h>
#include <linux/kthread.h>
+#include <linux/raid/pq.h>
+#include <asm/div64.h>
#include "compat.h"
#include "ctree.h"
#include "extent_map.h"
#include "transaction.h"
#include "print-tree.h"
#include "volumes.h"
+#include "raid56.h"
#include "async-thread.h"
#include "check-integrity.h"
#include "rcu-string.h"
goto out;
}
+ if ((all_avail & BTRFS_BLOCK_GROUP_RAID5) &&
+ root->fs_info->fs_devices->rw_devices <= 2) {
+ printk(KERN_ERR "btrfs: unable to go below two "
+ "devices on raid5\n");
+ ret = -EINVAL;
+ goto out;
+ }
+ if ((all_avail & BTRFS_BLOCK_GROUP_RAID6) &&
+ root->fs_info->fs_devices->rw_devices <= 3) {
+ printk(KERN_ERR "btrfs: unable to go below three "
+ "devices on raid6\n");
+ ret = -EINVAL;
+ goto out;
+ }
+
if (strcmp(device_path, "missing") == 0) {
struct list_head *devices;
struct btrfs_device *tmp;
}
} else {
ret = btrfs_get_bdev_and_sb(device_path,
- FMODE_READ | FMODE_EXCL,
+ FMODE_WRITE | FMODE_EXCL,
root->fs_info->bdev_holder, 0,
&bdev, &bh);
if (ret)
ret = 0;
/* Notify udev that device has changed */
- btrfs_kobject_uevent(bdev, KOBJ_CHANGE);
+ if (bdev)
+ btrfs_kobject_uevent(bdev, KOBJ_CHANGE);
error_brelse:
brelse(bh);
cache = btrfs_lookup_block_group(fs_info, chunk_offset);
chunk_used = btrfs_block_group_used(&cache->item);
- user_thresh = div_factor_fine(cache->key.offset, bargs->usage);
+ if (bargs->usage == 0)
+ user_thresh = 0;
+ else if (bargs->usage > 100)
+ user_thresh = cache->key.offset;
+ else
+ user_thresh = div_factor_fine(cache->key.offset,
+ bargs->usage);
+
if (chunk_used < user_thresh)
ret = 0;
return 0;
if (btrfs_chunk_type(leaf, chunk) & (BTRFS_BLOCK_GROUP_DUP |
- BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10))
- factor = 2;
- else
- factor = 1;
- factor = num_stripes / factor;
+ BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10)) {
+ factor = num_stripes / 2;
+ } else if (btrfs_chunk_type(leaf, chunk) & BTRFS_BLOCK_GROUP_RAID5) {
+ factor = num_stripes - 1;
+ } else if (btrfs_chunk_type(leaf, chunk) & BTRFS_BLOCK_GROUP_RAID6) {
+ factor = num_stripes - 2;
+ } else {
+ factor = num_stripes;
+ }
for (i = 0; i < num_stripes; i++) {
stripe = btrfs_stripe_nr(chunk, i);
unset_balance_control(fs_info);
ret = del_balance_item(fs_info->tree_root);
BUG_ON(ret);
+
+ atomic_set(&fs_info->mutually_exclusive_operation_running, 0);
}
void update_ioctl_balance_args(struct btrfs_fs_info *fs_info, int lock,
int mixed = 0;
int ret;
u64 num_devices;
+ int cancel = 0;
if (btrfs_fs_closing(fs_info) ||
atomic_read(&fs_info->balance_pause_req) ||
allowed |= (BTRFS_BLOCK_GROUP_RAID0 | BTRFS_BLOCK_GROUP_RAID1);
else
allowed |= (BTRFS_BLOCK_GROUP_RAID0 | BTRFS_BLOCK_GROUP_RAID1 |
- BTRFS_BLOCK_GROUP_RAID10);
+ BTRFS_BLOCK_GROUP_RAID10 |
+ BTRFS_BLOCK_GROUP_RAID5 |
+ BTRFS_BLOCK_GROUP_RAID6);
if ((bctl->data.flags & BTRFS_BALANCE_ARGS_CONVERT) &&
(!alloc_profile_is_valid(bctl->data.target, 1) ||
/* allow to reduce meta or sys integrity only if force set */
allowed = BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1 |
- BTRFS_BLOCK_GROUP_RAID10;
+ BTRFS_BLOCK_GROUP_RAID10 |
+ BTRFS_BLOCK_GROUP_RAID5 |
+ BTRFS_BLOCK_GROUP_RAID6;
+
if (((bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT) &&
(fs_info->avail_system_alloc_bits & allowed) &&
!(bctl->sys.target & allowed)) ||
}
if ((ret && ret != -ECANCELED && ret != -ENOSPC) ||
- balance_need_close(fs_info)) {
- __cancel_balance(fs_info);
- }
+ balance_need_close(fs_info))
+ cancel = 1;
if (bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT) {
fs_info->num_tolerated_disk_barrier_failures =
btrfs_calc_num_tolerated_disk_barrier_failures(fs_info);
}
+ if (cancel)
+ __cancel_balance(fs_info);
+
wake_up(&fs_info->balance_wait_q);
return ret;
out:
if (bctl->flags & BTRFS_BALANCE_RESUME)
__cancel_balance(fs_info);
- else
+ else {
kfree(bctl);
+ atomic_set(&fs_info->mutually_exclusive_operation_running, 0);
+ }
return ret;
}
ret = btrfs_balance(fs_info->balance_ctl, NULL);
}
- atomic_set(&fs_info->mutually_exclusive_operation_running, 0);
mutex_unlock(&fs_info->balance_mutex);
mutex_unlock(&fs_info->volume_mutex);
return 0;
}
- WARN_ON(atomic_xchg(&fs_info->mutually_exclusive_operation_running, 1));
tsk = kthread_run(balance_kthread, fs_info, "btrfs-balance");
if (IS_ERR(tsk))
return PTR_ERR(tsk);
btrfs_balance_sys(leaf, item, &disk_bargs);
btrfs_disk_balance_args_to_cpu(&bctl->sys, &disk_bargs);
+ WARN_ON(atomic_xchg(&fs_info->mutually_exclusive_operation_running, 1));
+
mutex_lock(&fs_info->volume_mutex);
mutex_lock(&fs_info->balance_mutex);
}
struct btrfs_raid_attr btrfs_raid_array[BTRFS_NR_RAID_TYPES] = {
+ /*
+ * sub_stripes info for map,
+ * dev_stripes -- stripes per dev, 2 for DUP, 1 other wise
+ * devs_max -- max devices per stripe, 0 for unlimited
+ * devs_min -- min devices per stripe
+ * devs_increment -- ndevs must be a multiple of this
+ * ncopies -- how many copies of the data we have
+ */
{ 2, 1, 0, 4, 2, 2 /* raid10 */ },
{ 1, 1, 2, 2, 2, 2 /* raid1 */ },
{ 1, 2, 1, 1, 1, 2 /* dup */ },
{ 1, 1, 0, 2, 1, 1 /* raid0 */ },
- { 1, 1, 0, 1, 1, 1 /* single */ },
+ { 1, 1, 1, 1, 1, 1 /* single */ },
+ { 1, 1, 0, 2, 1, 2 /* raid5 */ },
+ { 1, 1, 0, 3, 1, 3 /* raid6 */ },
};
+static u32 find_raid56_stripe_len(u32 data_devices, u32 dev_stripe_target)
+{
+ /* TODO allow them to set a preferred stripe size */
+ return 64 * 1024;
+}
+
+static void check_raid56_incompat_flag(struct btrfs_fs_info *info, u64 type)
+{
+ u64 features;
+
+ if (!(type & (BTRFS_BLOCK_GROUP_RAID5 | BTRFS_BLOCK_GROUP_RAID6)))
+ return;
+
+ features = btrfs_super_incompat_flags(info->super_copy);
+ if (features & BTRFS_FEATURE_INCOMPAT_RAID56)
+ return;
+
+ features |= BTRFS_FEATURE_INCOMPAT_RAID56;
+ btrfs_set_super_incompat_flags(info->super_copy, features);
+ printk(KERN_INFO "btrfs: setting RAID5/6 feature flag\n");
+}
+
static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
struct btrfs_root *extent_root,
struct map_lookup **map_ret,
struct btrfs_device_info *devices_info = NULL;
u64 total_avail;
int num_stripes; /* total number of stripes to allocate */
+ int data_stripes; /* number of stripes that count for
+ block group size */
int sub_stripes; /* sub_stripes info for map */
int dev_stripes; /* stripes per dev */
int devs_max; /* max devs to use */
u64 max_chunk_size;
u64 stripe_size;
u64 num_bytes;
+ u64 raid_stripe_len = BTRFS_STRIPE_LEN;
int ndevs;
int i;
int j;
stripe_size = devices_info[ndevs-1].max_avail;
num_stripes = ndevs * dev_stripes;
+ /*
+ * this will have to be fixed for RAID1 and RAID10 over
+ * more drives
+ */
+ data_stripes = num_stripes / ncopies;
+
if (stripe_size * ndevs > max_chunk_size * ncopies) {
stripe_size = max_chunk_size * ncopies;
do_div(stripe_size, ndevs);
}
-
+ if (type & BTRFS_BLOCK_GROUP_RAID5) {
+ raid_stripe_len = find_raid56_stripe_len(ndevs - 1,
+ btrfs_super_stripesize(info->super_copy));
+ data_stripes = num_stripes - 1;
+ }
+ if (type & BTRFS_BLOCK_GROUP_RAID6) {
+ raid_stripe_len = find_raid56_stripe_len(ndevs - 2,
+ btrfs_super_stripesize(info->super_copy));
+ data_stripes = num_stripes - 2;
+ }
do_div(stripe_size, dev_stripes);
/* align to BTRFS_STRIPE_LEN */
- do_div(stripe_size, BTRFS_STRIPE_LEN);
- stripe_size *= BTRFS_STRIPE_LEN;
+ do_div(stripe_size, raid_stripe_len);
+ stripe_size *= raid_stripe_len;
map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS);
if (!map) {
}
}
map->sector_size = extent_root->sectorsize;
- map->stripe_len = BTRFS_STRIPE_LEN;
- map->io_align = BTRFS_STRIPE_LEN;
- map->io_width = BTRFS_STRIPE_LEN;
+ map->stripe_len = raid_stripe_len;
+ map->io_align = raid_stripe_len;
+ map->io_width = raid_stripe_len;
map->type = type;
map->sub_stripes = sub_stripes;
*map_ret = map;
- num_bytes = stripe_size * (num_stripes / ncopies);
+ num_bytes = stripe_size * data_stripes;
*stripe_size_out = stripe_size;
*num_bytes_out = num_bytes;
}
}
+ check_raid56_incompat_flag(extent_root->fs_info, type);
+
kfree(devices_info);
return 0;
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;
free_extent_map(em);
return ret;
}
+unsigned long btrfs_full_stripe_len(struct btrfs_root *root,
+ struct btrfs_mapping_tree *map_tree,
+ u64 logical)
+{
+ struct extent_map *em;
+ struct map_lookup *map;
+ struct extent_map_tree *em_tree = &map_tree->map_tree;
+ unsigned long len = root->sectorsize;
+
+ read_lock(&em_tree->lock);
+ em = lookup_extent_mapping(em_tree, logical, len);
+ read_unlock(&em_tree->lock);
+ 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_RAID5 |
+ BTRFS_BLOCK_GROUP_RAID6)) {
+ len = map->stripe_len * nr_data_stripes(map);
+ }
+ free_extent_map(em);
+ return len;
+}
+
+int btrfs_is_parity_mirror(struct btrfs_mapping_tree *map_tree,
+ u64 logical, u64 len, int mirror_num)
+{
+ struct extent_map *em;
+ struct map_lookup *map;
+ struct extent_map_tree *em_tree = &map_tree->map_tree;
+ int ret = 0;
+
+ read_lock(&em_tree->lock);
+ em = lookup_extent_mapping(em_tree, logical, len);
+ read_unlock(&em_tree->lock);
+ 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_RAID5 |
+ BTRFS_BLOCK_GROUP_RAID6))
+ ret = 1;
+ free_extent_map(em);
+ return ret;
+}
+
static int find_live_mirror(struct btrfs_fs_info *fs_info,
struct map_lookup *map, int first, int num,
int optimal, int dev_replace_is_ongoing)
return optimal;
}
+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_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;
+ }
+ }
+ }
+}
+
static int __btrfs_map_block(struct btrfs_fs_info *fs_info, int rw,
u64 logical, u64 *length,
struct btrfs_bio **bbio_ret,
- int mirror_num)
+ int mirror_num, u64 **raid_map_ret)
{
struct extent_map *em;
struct map_lookup *map;
u64 stripe_nr;
u64 stripe_nr_orig;
u64 stripe_nr_end;
+ u64 stripe_len;
+ u64 *raid_map = NULL;
int stripe_index;
int i;
int ret = 0;
int num_alloc_stripes;
int patch_the_first_stripe_for_dev_replace = 0;
u64 physical_to_patch_in_first_stripe = 0;
+ u64 raid56_full_stripe_start = (u64)-1;
read_lock(&em_tree->lock);
em = lookup_extent_mapping(em_tree, logical, *length);
map = (struct map_lookup *)em->bdev;
offset = logical - em->start;
+ if (mirror_num > map->num_stripes)
+ mirror_num = 0;
+
+ stripe_len = map->stripe_len;
stripe_nr = offset;
/*
* stripe_nr counts the total number of stripes we have to stride
* to get to this block
*/
- do_div(stripe_nr, map->stripe_len);
+ do_div(stripe_nr, stripe_len);
- stripe_offset = stripe_nr * map->stripe_len;
+ stripe_offset = stripe_nr * stripe_len;
BUG_ON(offset < stripe_offset);
/* stripe_offset is the offset of this block in its stripe*/
stripe_offset = offset - stripe_offset;
- if (rw & REQ_DISCARD)
+ /* if we're here for raid56, we need to know the stripe aligned start */
+ if (map->type & (BTRFS_BLOCK_GROUP_RAID5 | BTRFS_BLOCK_GROUP_RAID6)) {
+ unsigned long full_stripe_len = stripe_len * nr_data_stripes(map);
+ raid56_full_stripe_start = offset;
+
+ /* allow a write of a full stripe, but make sure we don't
+ * allow straddling of stripes
+ */
+ do_div(raid56_full_stripe_start, full_stripe_len);
+ raid56_full_stripe_start *= full_stripe_len;
+ }
+
+ if (rw & REQ_DISCARD) {
+ /* we don't discard raid56 yet */
+ if (map->type &
+ (BTRFS_BLOCK_GROUP_RAID5 | BTRFS_BLOCK_GROUP_RAID6)) {
+ ret = -EOPNOTSUPP;
+ goto out;
+ }
*length = min_t(u64, em->len - offset, *length);
- else if (map->type & BTRFS_BLOCK_GROUP_PROFILE_MASK) {
- /* we limit the length of each bio to what fits in a stripe */
- *length = min_t(u64, em->len - offset,
- map->stripe_len - stripe_offset);
+ } else if (map->type & BTRFS_BLOCK_GROUP_PROFILE_MASK) {
+ u64 max_len;
+ /* For writes to RAID[56], allow a full stripeset across all disks.
+ For other RAID types and for RAID[56] reads, just allow a single
+ stripe (on a single disk). */
+ if (map->type & (BTRFS_BLOCK_GROUP_RAID5 | BTRFS_BLOCK_GROUP_RAID6) &&
+ (rw & REQ_WRITE)) {
+ max_len = stripe_len * nr_data_stripes(map) -
+ (offset - raid56_full_stripe_start);
+ } else {
+ /* we limit the length of each bio to what fits in a stripe */
+ max_len = stripe_len - stripe_offset;
+ }
+ *length = min_t(u64, em->len - offset, max_len);
} else {
*length = em->len - offset;
}
+ /* This is for when we're called from btrfs_merge_bio_hook() and all
+ it cares about is the length */
if (!bbio_ret)
goto out;
u64 physical_of_found = 0;
ret = __btrfs_map_block(fs_info, REQ_GET_READ_MIRRORS,
- logical, &tmp_length, &tmp_bbio, 0);
+ logical, &tmp_length, &tmp_bbio, 0, NULL);
if (ret) {
WARN_ON(tmp_bbio != NULL);
goto out;
do_div(stripe_nr_end, map->stripe_len);
stripe_end_offset = stripe_nr_end * map->stripe_len -
(offset + *length);
+
if (map->type & BTRFS_BLOCK_GROUP_RAID0) {
if (rw & REQ_DISCARD)
num_stripes = min_t(u64, map->num_stripes,
dev_replace_is_ongoing);
mirror_num = stripe_index - old_stripe_index + 1;
}
+
+ } else if (map->type & (BTRFS_BLOCK_GROUP_RAID5 |
+ BTRFS_BLOCK_GROUP_RAID6)) {
+ u64 tmp;
+
+ if (bbio_ret && ((rw & REQ_WRITE) || mirror_num > 1)
+ && raid_map_ret) {
+ int i, rot;
+
+ /* push stripe_nr back to the start of the full stripe */
+ stripe_nr = raid56_full_stripe_start;
+ do_div(stripe_nr, stripe_len);
+
+ stripe_index = do_div(stripe_nr, nr_data_stripes(map));
+
+ /* RAID[56] write or recovery. Return all stripes */
+ num_stripes = map->num_stripes;
+ max_errors = nr_parity_stripes(map);
+
+ raid_map = kmalloc(sizeof(u64) * num_stripes,
+ GFP_NOFS);
+ if (!raid_map) {
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ /* Work out the disk rotation on this stripe-set */
+ tmp = stripe_nr;
+ rot = do_div(tmp, 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) % num_stripes] =
+ em->start + (tmp + i) * map->stripe_len;
+
+ raid_map[(i+rot) % map->num_stripes] = RAID5_P_STRIPE;
+ if (map->type & BTRFS_BLOCK_GROUP_RAID6)
+ raid_map[(i+rot+1) % num_stripes] =
+ RAID6_Q_STRIPE;
+
+ *length = map->stripe_len;
+ stripe_index = 0;
+ stripe_offset = 0;
+ } else {
+ /*
+ * Mirror #0 or #1 means the original data block.
+ * Mirror #2 is RAID5 parity block.
+ * Mirror #3 is RAID6 Q block.
+ */
+ stripe_index = do_div(stripe_nr, nr_data_stripes(map));
+ if (mirror_num > 1)
+ stripe_index = nr_data_stripes(map) +
+ mirror_num - 2;
+
+ /* We distribute the parity blocks across stripes */
+ tmp = stripe_nr + stripe_index;
+ stripe_index = do_div(tmp, map->num_stripes);
+ }
} else {
/*
* after this do_div call, stripe_nr is the number of stripes
if (rw & (REQ_WRITE | REQ_GET_READ_MIRRORS)) {
if (map->type & (BTRFS_BLOCK_GROUP_RAID1 |
BTRFS_BLOCK_GROUP_RAID10 |
+ BTRFS_BLOCK_GROUP_RAID5 |
BTRFS_BLOCK_GROUP_DUP)) {
max_errors = 1;
+ } else if (map->type & BTRFS_BLOCK_GROUP_RAID6) {
+ max_errors = 2;
}
}
bbio->stripes[0].physical = physical_to_patch_in_first_stripe;
bbio->mirror_num = map->num_stripes + 1;
}
+ if (raid_map) {
+ sort_parity_stripes(bbio, raid_map);
+ *raid_map_ret = raid_map;
+ }
out:
if (dev_replace_is_ongoing)
btrfs_dev_replace_unlock(dev_replace);
struct btrfs_bio **bbio_ret, int mirror_num)
{
return __btrfs_map_block(fs_info, rw, logical, length, bbio_ret,
- mirror_num);
+ mirror_num, NULL);
}
int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree,
u64 bytenr;
u64 length;
u64 stripe_nr;
+ u64 rmap_len;
int i, j, nr = 0;
read_lock(&em_tree->lock);
map = (struct map_lookup *)em->bdev;
length = em->len;
+ rmap_len = map->stripe_len;
+
if (map->type & BTRFS_BLOCK_GROUP_RAID10)
do_div(length, map->num_stripes / map->sub_stripes);
else if (map->type & BTRFS_BLOCK_GROUP_RAID0)
do_div(length, map->num_stripes);
+ else if (map->type & (BTRFS_BLOCK_GROUP_RAID5 |
+ BTRFS_BLOCK_GROUP_RAID6)) {
+ do_div(length, nr_data_stripes(map));
+ rmap_len = map->stripe_len * nr_data_stripes(map);
+ }
buf = kzalloc(sizeof(u64) * map->num_stripes, GFP_NOFS);
BUG_ON(!buf); /* -ENOMEM */
do_div(stripe_nr, map->sub_stripes);
} else if (map->type & BTRFS_BLOCK_GROUP_RAID0) {
stripe_nr = stripe_nr * map->num_stripes + i;
- }
- bytenr = chunk_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 = chunk_start + stripe_nr * rmap_len;
WARN_ON(nr >= map->num_stripes);
for (j = 0; j < nr; j++) {
if (buf[j] == bytenr)
*logical = buf;
*naddrs = nr;
- *stripe_len = map->stripe_len;
+ *stripe_len = rmap_len;
free_extent_map(em);
return 0;
bio->bi_bdev = (struct block_device *)
(unsigned long)bbio->mirror_num;
/* only send an error to the higher layers if it is
- * beyond the tolerance of the multi-bio
+ * beyond the tolerance of the btrfs bio
*/
if (atomic_read(&bbio->error) > bbio->max_errors) {
err = -EIO;
* This will add one bio to the pending list for a device and make sure
* the work struct is scheduled.
*/
-static noinline void schedule_bio(struct btrfs_root *root,
+noinline void btrfs_schedule_bio(struct btrfs_root *root,
struct btrfs_device *device,
int rw, struct bio *bio)
{
int should_queue = 1;
struct btrfs_pending_bios *pending_bios;
+ if (device->missing || !device->bdev) {
+ bio_endio(bio, -EIO);
+ return;
+ }
+
/* don't bother with additional async steps for reads, right now */
if (!(rw & REQ_WRITE)) {
bio_get(bio);
#endif
bio->bi_bdev = dev->bdev;
if (async)
- schedule_bio(root, dev, rw, bio);
+ btrfs_schedule_bio(root, dev, rw, bio);
else
btrfsic_submit_bio(rw, bio);
}
u64 logical = (u64)bio->bi_sector << 9;
u64 length = 0;
u64 map_length;
+ u64 *raid_map = NULL;
int ret;
int dev_nr = 0;
int total_devs = 1;
length = bio->bi_size;
map_length = length;
- ret = btrfs_map_block(root->fs_info, rw, logical, &map_length, &bbio,
- mirror_num);
- if (ret)
+ ret = __btrfs_map_block(root->fs_info, rw, logical, &map_length, &bbio,
+ mirror_num, &raid_map);
+ if (ret) /* -ENOMEM */
return ret;
total_devs = bbio->num_stripes;
+ bbio->orig_bio = first_bio;
+ bbio->private = first_bio->bi_private;
+ bbio->end_io = first_bio->bi_end_io;
+ atomic_set(&bbio->stripes_pending, bbio->num_stripes);
+
+ if (raid_map) {
+ /* In this case, map_length has been set to the length of
+ a single stripe; not the whole write */
+ if (rw & WRITE) {
+ return raid56_parity_write(root, bio, bbio,
+ raid_map, map_length);
+ } else {
+ return raid56_parity_recover(root, bio, bbio,
+ raid_map, map_length,
+ mirror_num);
+ }
+ }
+
if (map_length < length) {
printk(KERN_CRIT "btrfs: mapping failed logical %llu bio len %llu "
"len %llu\n", (unsigned long long)logical,
BUG();
}
- bbio->orig_bio = first_bio;
- bbio->private = first_bio->bi_private;
- bbio->end_io = first_bio->bi_end_io;
- atomic_set(&bbio->stripes_pending, bbio->num_stripes);
-
while (dev_nr < total_devs) {
dev = bbio->stripes[dev_nr].dev;
if (!dev || !dev->bdev || (rw & WRITE && !dev->writeable)) {