[RFC PATCH v1 2/5] BTRFS hot reloc: add one new block group

From: zwu . kernel
Date: Mon May 20 2013 - 11:11:05 EST

From: Zhi Yong Wu <wuzhy@xxxxxxxxxxxxxxxxxx>

Introduce one new block group BTRFS_BLOCK_GROUP_DATA_NONROT,
which is used to differentiate if the block space is reserved
and allocated from one rotating disk or nonrotating disk.

Signed-off-by: Zhi Yong Wu <wuzhy@xxxxxxxxxxxxxxxxxx>
---
fs/btrfs/ctree.h | 33 ++++++++++++---
fs/btrfs/extent-tree.c | 99 ++++++++++++++++++++++++++++++++++++---------
fs/btrfs/extent_io.c | 59 ++++++++++++++++++++++++++-
fs/btrfs/extent_io.h | 7 ++++
fs/btrfs/file.c | 24 +++++++----
fs/btrfs/free-space-cache.c | 2 +-
fs/btrfs/inode-map.c | 7 ++--
fs/btrfs/inode.c | 94 ++++++++++++++++++++++++++++++++++--------
fs/btrfs/ioctl.c | 17 +++++---
fs/btrfs/relocation.c | 6 ++-
fs/btrfs/super.c | 4 +-
fs/btrfs/volumes.c | 29 ++++++++++++-
12 files changed, 316 insertions(+), 65 deletions(-)

diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 133a6ed..f7a3170 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -963,6 +963,12 @@ struct btrfs_dev_replace_item {
#define BTRFS_BLOCK_GROUP_RAID10 (1ULL << 6)
#define BTRFS_BLOCK_GROUP_RAID5 (1 << 7)
#define BTRFS_BLOCK_GROUP_RAID6 (1 << 8)
+/*
+ * New block groups for use with BTRFS hot relocation feature.
+ * When BTRFS hot relocation is enabled, *_NONROT block group is
+ * forced to nonrotating drives.
+ */
+#define BTRFS_BLOCK_GROUP_DATA_NONROT (1ULL << 9)
#define BTRFS_BLOCK_GROUP_RESERVED BTRFS_AVAIL_ALLOC_BIT_SINGLE

enum btrfs_raid_types {
@@ -978,7 +984,8 @@ enum btrfs_raid_types {

#define BTRFS_BLOCK_GROUP_TYPE_MASK (BTRFS_BLOCK_GROUP_DATA | \
BTRFS_BLOCK_GROUP_SYSTEM | \
- BTRFS_BLOCK_GROUP_METADATA)
+ BTRFS_BLOCK_GROUP_METADATA | \
+ BTRFS_BLOCK_GROUP_DATA_NONROT)

#define BTRFS_BLOCK_GROUP_PROFILE_MASK (BTRFS_BLOCK_GROUP_RAID0 | \
BTRFS_BLOCK_GROUP_RAID1 | \
@@ -1521,6 +1528,7 @@ struct btrfs_fs_info {
struct list_head space_info;

struct btrfs_space_info *data_sinfo;
+ struct btrfs_space_info *nonrot_data_sinfo;

struct reloc_control *reloc_ctl;

@@ -1545,6 +1553,7 @@ struct btrfs_fs_info {
u64 avail_data_alloc_bits;
u64 avail_metadata_alloc_bits;
u64 avail_system_alloc_bits;
+ u64 avail_data_nonrot_alloc_bits;

/* restriper state */
spinlock_t balance_lock;
@@ -1557,6 +1566,7 @@ struct btrfs_fs_info {

unsigned data_chunk_allocations;
unsigned metadata_ratio;
+ unsigned data_nonrot_chunk_allocations;

void *bdev_holder;

@@ -1928,6 +1938,7 @@ struct btrfs_ioctl_defrag_range_args {
#define BTRFS_MOUNT_CHECK_INTEGRITY_INCLUDING_EXTENT_DATA (1 << 21)
#define BTRFS_MOUNT_PANIC_ON_FATAL_ERROR (1 << 22)
#define BTRFS_MOUNT_HOT_TRACK (1 << 23)
+#define BTRFS_MOUNT_HOT_MOVE (1 << 24)

#define btrfs_clear_opt(o, opt) ((o) &= ~BTRFS_MOUNT_##opt)
#define btrfs_set_opt(o, opt) ((o) |= BTRFS_MOUNT_##opt)
@@ -3043,6 +3054,8 @@ int btrfs_pin_extent_for_log_replay(struct btrfs_root *root,
int btrfs_cross_ref_exist(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
u64 objectid, u64 offset, u64 bytenr);
+struct btrfs_block_group_cache *btrfs_lookup_first_block_group(
+ struct btrfs_fs_info *info, u64 bytenr);
struct btrfs_block_group_cache *btrfs_lookup_block_group(
struct btrfs_fs_info *info,
u64 bytenr);
@@ -3093,6 +3106,8 @@ int btrfs_inc_extent_ref(struct btrfs_trans_handle *trans,
struct btrfs_root *root,
u64 bytenr, u64 num_bytes, u64 parent,
u64 root_objectid, u64 owner, u64 offset, int for_cow);
+struct btrfs_block_group_cache *next_block_group(struct btrfs_root *root,
+ struct btrfs_block_group_cache *cache);

int btrfs_write_dirty_block_groups(struct btrfs_trans_handle *trans,
struct btrfs_root *root);
@@ -3122,8 +3137,14 @@ enum btrfs_reserve_flush_enum {
BTRFS_RESERVE_FLUSH_ALL,
};

-int btrfs_check_data_free_space(struct inode *inode, u64 bytes);
-void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes);
+enum {
+ TYPE_ROT, /* rot -> rotating */
+ TYPE_NONROT, /* nonrot -> nonrotating */
+ MAX_RELOC_TYPES,
+};
+
+int btrfs_check_data_free_space(struct inode *inode, u64 bytes, int *flag);
+void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes, int flag);
void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans,
struct btrfs_root *root);
int btrfs_orphan_reserve_metadata(struct btrfs_trans_handle *trans,
@@ -3138,8 +3159,8 @@ void btrfs_subvolume_release_metadata(struct btrfs_root *root,
u64 qgroup_reserved);
int btrfs_delalloc_reserve_metadata(struct inode *inode, u64 num_bytes);
void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes);
-int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes);
-void btrfs_delalloc_release_space(struct inode *inode, u64 num_bytes);
+int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes, int *flag);
+void btrfs_delalloc_release_space(struct inode *inode, u64 num_bytes, int flag);
void btrfs_init_block_rsv(struct btrfs_block_rsv *rsv, unsigned short type);
struct btrfs_block_rsv *btrfs_alloc_block_rsv(struct btrfs_root *root,
unsigned short type);
@@ -3612,7 +3633,7 @@ int btrfs_release_file(struct inode *inode, struct file *file);
int btrfs_dirty_pages(struct btrfs_root *root, struct inode *inode,
struct page **pages, size_t num_pages,
loff_t pos, size_t write_bytes,
- struct extent_state **cached);
+ struct extent_state **cached, int flag);

/* tree-defrag.c */
int btrfs_defrag_leaves(struct btrfs_trans_handle *trans,
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 2305b5c..afc9f77 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -628,7 +628,7 @@ static int cache_block_group(struct btrfs_block_group_cache *cache,
/*
* return the block group that starts at or after bytenr
*/
-static struct btrfs_block_group_cache *
+struct btrfs_block_group_cache *
btrfs_lookup_first_block_group(struct btrfs_fs_info *info, u64 bytenr)
{
struct btrfs_block_group_cache *cache;
@@ -3030,7 +3030,7 @@ fail:

}

-static struct btrfs_block_group_cache *
+struct btrfs_block_group_cache *
next_block_group(struct btrfs_root *root,
struct btrfs_block_group_cache *cache)
{
@@ -3059,6 +3059,7 @@ static int cache_save_setup(struct btrfs_block_group_cache *block_group,
int num_pages = 0;
int retries = 0;
int ret = 0;
+ int flag = TYPE_ROT;

/*
* If this block group is smaller than 100 megs don't bother caching the
@@ -3142,7 +3143,7 @@ again:
num_pages *= 16;
num_pages *= PAGE_CACHE_SIZE;

- ret = btrfs_check_data_free_space(inode, num_pages);
+ ret = btrfs_check_data_free_space(inode, num_pages, &flag);
if (ret)
goto out_put;

@@ -3151,7 +3152,8 @@ again:
&alloc_hint);
if (!ret)
dcs = BTRFS_DC_SETUP;
- btrfs_free_reserved_data_space(inode, num_pages);
+
+ btrfs_free_reserved_data_space(inode, num_pages, flag);

out_put:
iput(inode);
@@ -3353,6 +3355,8 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags,
list_add_rcu(&found->list, &info->space_info);
if (flags & BTRFS_BLOCK_GROUP_DATA)
info->data_sinfo = found;
+ else if (flags & BTRFS_BLOCK_GROUP_DATA_NONROT)
+ info->nonrot_data_sinfo = found;
return 0;
}

@@ -3368,6 +3372,8 @@ static void set_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags)
fs_info->avail_metadata_alloc_bits |= extra_flags;
if (flags & BTRFS_BLOCK_GROUP_SYSTEM)
fs_info->avail_system_alloc_bits |= extra_flags;
+ if (flags & BTRFS_BLOCK_GROUP_DATA_NONROT)
+ fs_info->avail_data_nonrot_alloc_bits |= extra_flags;
write_sequnlock(&fs_info->profiles_lock);
}

@@ -3474,18 +3480,27 @@ static u64 get_alloc_profile(struct btrfs_root *root, u64 flags)
flags |= root->fs_info->avail_system_alloc_bits;
else if (flags & BTRFS_BLOCK_GROUP_METADATA)
flags |= root->fs_info->avail_metadata_alloc_bits;
+ else if (flags & BTRFS_BLOCK_GROUP_DATA_NONROT)
+ flags |= root->fs_info->avail_data_nonrot_alloc_bits;
} while (read_seqretry(&root->fs_info->profiles_lock, seq));

return btrfs_reduce_alloc_profile(root, flags);
}

+/*
+ * Turns a chunk_type integer into set of block group flags (a profile).
+ * Hot relocation code adds chunk_type 2 for hot data specific block
+ * group type.
+ */
u64 btrfs_get_alloc_profile(struct btrfs_root *root, int data)
{
u64 flags;
u64 ret;

- if (data)
+ if (data == 1)
flags = BTRFS_BLOCK_GROUP_DATA;
+ else if (data == 2)
+ flags = BTRFS_BLOCK_GROUP_DATA_NONROT;
else if (root == root->fs_info->chunk_root)
flags = BTRFS_BLOCK_GROUP_SYSTEM;
else
@@ -3499,13 +3514,14 @@ u64 btrfs_get_alloc_profile(struct btrfs_root *root, int data)
* This will check the space that the inode allocates from to make sure we have
* enough space for bytes.
*/
-int btrfs_check_data_free_space(struct inode *inode, u64 bytes)
+int btrfs_check_data_free_space(struct inode *inode, u64 bytes, int *flag)
{
struct btrfs_space_info *data_sinfo;
struct btrfs_root *root = BTRFS_I(inode)->root;
struct btrfs_fs_info *fs_info = root->fs_info;
u64 used;
int ret = 0, committed = 0, alloc_chunk = 1;
+ int data, tried = 0;

/* make sure bytes are sectorsize aligned */
bytes = ALIGN(bytes, root->sectorsize);
@@ -3516,7 +3532,15 @@ int btrfs_check_data_free_space(struct inode *inode, u64 bytes)
committed = 1;
}

- data_sinfo = fs_info->data_sinfo;
+ if (*flag == TYPE_NONROT) {
+try_nonrot:
+ data = 2;
+ data_sinfo = fs_info->nonrot_data_sinfo;
+ } else {
+ data = 1;
+ data_sinfo = fs_info->data_sinfo;
+ }
+
if (!data_sinfo)
goto alloc;

@@ -3534,13 +3558,22 @@ again:
* if we don't have enough free bytes in this space then we need
* to alloc a new chunk.
*/
- if (!data_sinfo->full && alloc_chunk) {
+ if (alloc_chunk) {
u64 alloc_target;

+ if (data_sinfo->full) {
+ if (!tried) {
+ tried = 1;
+ spin_unlock(&data_sinfo->lock);
+ goto try_nonrot;
+ } else
+ goto non_alloc;
+ }
+
data_sinfo->force_alloc = CHUNK_ALLOC_FORCE;
spin_unlock(&data_sinfo->lock);
alloc:
- alloc_target = btrfs_get_alloc_profile(root, 1);
+ alloc_target = btrfs_get_alloc_profile(root, data);
trans = btrfs_join_transaction(root);
if (IS_ERR(trans))
return PTR_ERR(trans);
@@ -3557,11 +3590,13 @@ alloc:
}

if (!data_sinfo)
- data_sinfo = fs_info->data_sinfo;
+ data_sinfo = (data == 1) ? fs_info->data_sinfo :
+ fs_info->nonrot_data_sinfo;

goto again;
}

+non_alloc:
/*
* If we have less pinned bytes than we want to allocate then
* don't bother committing the transaction, it won't help us.
@@ -3572,7 +3607,7 @@ alloc:

/* commit the current transaction and try again */
commit_trans:
- if (!committed &&
+ if (!committed && data_sinfo &&
!atomic_read(&root->fs_info->open_ioctl_trans)) {
committed = 1;
trans = btrfs_join_transaction(root);
@@ -3586,6 +3621,10 @@ commit_trans:

return -ENOSPC;
}
+
+ if (tried)
+ *flag = TYPE_NONROT;
+
data_sinfo->bytes_may_use += bytes;
trace_btrfs_space_reservation(root->fs_info, "space_info",
data_sinfo->flags, bytes, 1);
@@ -3597,7 +3636,7 @@ commit_trans:
/*
* Called if we need to clear a data reservation for this inode.
*/
-void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes)
+void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes, int flag)
{
struct btrfs_root *root = BTRFS_I(inode)->root;
struct btrfs_space_info *data_sinfo;
@@ -3605,7 +3644,10 @@ void btrfs_free_reserved_data_space(struct inode *inode, u64 bytes)
/* make sure bytes are sectorsize aligned */
bytes = ALIGN(bytes, root->sectorsize);

- data_sinfo = root->fs_info->data_sinfo;
+ if (flag == TYPE_NONROT)
+ data_sinfo = root->fs_info->nonrot_data_sinfo;
+ else
+ data_sinfo = root->fs_info->data_sinfo;
spin_lock(&data_sinfo->lock);
data_sinfo->bytes_may_use -= bytes;
trace_btrfs_space_reservation(root->fs_info, "space_info",
@@ -3789,6 +3831,13 @@ again:
force_metadata_allocation(fs_info);
}

+ if (flags & BTRFS_BLOCK_GROUP_DATA_NONROT && fs_info->metadata_ratio) {
+ fs_info->data_nonrot_chunk_allocations++;
+ if (!(fs_info->data_nonrot_chunk_allocations %
+ fs_info->metadata_ratio))
+ force_metadata_allocation(fs_info);
+ }
+
/*
* Check if we have enough space in SYSTEM chunk because we may need
* to update devices.
@@ -4495,6 +4544,13 @@ static u64 calc_global_metadata_size(struct btrfs_fs_info *fs_info)
meta_used = sinfo->bytes_used;
spin_unlock(&sinfo->lock);

+ sinfo = __find_space_info(fs_info, BTRFS_BLOCK_GROUP_DATA_NONROT);
+ if (sinfo) {
+ spin_lock(&sinfo->lock);
+ data_used += sinfo->bytes_used;
+ spin_unlock(&sinfo->lock);
+ }
+
num_bytes = (data_used >> fs_info->sb->s_blocksize_bits) *
csum_size * 2;
num_bytes += div64_u64(data_used + meta_used, 50);
@@ -4968,6 +5024,7 @@ void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes)
* btrfs_delalloc_reserve_space - reserve data and metadata space for delalloc
* @inode: inode we're writing to
* @num_bytes: the number of bytes we want to allocate
+ * @flag: indicate if block space is reserved from rotating disk or not
*
* This will do the following things
*
@@ -4979,17 +5036,17 @@ void btrfs_delalloc_release_metadata(struct inode *inode, u64 num_bytes)
*
* This will return 0 for success and -ENOSPC if there is no space left.
*/
-int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes)
+int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes, int *flag)
{
int ret;

- ret = btrfs_check_data_free_space(inode, num_bytes);
+ ret = btrfs_check_data_free_space(inode, num_bytes, flag);
if (ret)
return ret;

ret = btrfs_delalloc_reserve_metadata(inode, num_bytes);
if (ret) {
- btrfs_free_reserved_data_space(inode, num_bytes);
+ btrfs_free_reserved_data_space(inode, num_bytes, *flag);
return ret;
}

@@ -5000,6 +5057,7 @@ int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes)
* btrfs_delalloc_release_space - release data and metadata space for delalloc
* @inode: inode we're releasing space for
* @num_bytes: the number of bytes we want to free up
+ * @flag: indicate if block space is freed from rotating disk or not
*
* This must be matched with a call to btrfs_delalloc_reserve_space. This is
* called in the case that we don't need the metadata AND data reservations
@@ -5009,10 +5067,10 @@ int btrfs_delalloc_reserve_space(struct inode *inode, u64 num_bytes)
* decrement ->delalloc_bytes and remove it from the fs_info delalloc_inodes
* list if there are no delalloc bytes left.
*/
-void btrfs_delalloc_release_space(struct inode *inode, u64 num_bytes)
+void btrfs_delalloc_release_space(struct inode *inode, u64 num_bytes, int flag)
{
btrfs_delalloc_release_metadata(inode, num_bytes);
- btrfs_free_reserved_data_space(inode, num_bytes);
+ btrfs_free_reserved_data_space(inode, num_bytes, flag);
}

static int update_block_group(struct btrfs_root *root,
@@ -5888,7 +5946,8 @@ static noinline int find_free_extent(struct btrfs_trans_handle *trans,
struct btrfs_space_info *space_info;
int loop = 0;
int index = __get_raid_index(flags);
- int alloc_type = (flags & BTRFS_BLOCK_GROUP_DATA) ?
+ int alloc_type = ((flags & BTRFS_BLOCK_GROUP_DATA)
+ || (flags & BTRFS_BLOCK_GROUP_DATA_NONROT)) ?
RESERVE_ALLOC_NO_ACCOUNT : RESERVE_ALLOC;
bool found_uncached_bg = false;
bool failed_cluster_refill = false;
@@ -8360,6 +8419,8 @@ static void clear_avail_alloc_bits(struct btrfs_fs_info *fs_info, u64 flags)
fs_info->avail_metadata_alloc_bits &= ~extra_flags;
if (flags & BTRFS_BLOCK_GROUP_SYSTEM)
fs_info->avail_system_alloc_bits &= ~extra_flags;
+ if (flags & BTRFS_BLOCK_GROUP_DATA_NONROT)
+ fs_info->avail_data_nonrot_alloc_bits &= ~extra_flags;
write_sequnlock(&fs_info->profiles_lock);
}

diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index 32d67a8..2b1f132 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -1216,6 +1216,34 @@ int clear_extent_uptodate(struct extent_io_tree *tree, u64 start, u64 end,
cached_state, mask);
}

+void set_extent_hot(struct inode *inode, u64 start, u64 end,
+ struct extent_state **cached_state,
+ int type, int flag)
+{
+ int set_bits = 0, clear_bits = 0;
+
+ if (flag) {
+ set_bits = EXTENT_DELALLOC | EXTENT_UPTODATE;
+ clear_bits = EXTENT_DIRTY | EXTENT_DELALLOC |
+ EXTENT_DO_ACCOUNTING;
+ }
+
+ if (type == TYPE_NONROT) {
+ set_bits |= EXTENT_HOT;
+ clear_bits |= EXTENT_COLD;
+ } else {
+ set_bits |= EXTENT_COLD;
+ clear_bits |= EXTENT_HOT;
+ }
+
+ clear_extent_bit(&BTRFS_I(inode)->io_tree,
+ start, end, clear_bits,
+ 0, 0, cached_state, GFP_NOFS);
+ set_extent_bit(&BTRFS_I(inode)->io_tree, start,
+ end, set_bits, NULL,
+ cached_state, GFP_NOFS);
+}
+
/*
* either insert or lock state struct between start and end use mask to tell
* us if waiting is desired.
@@ -1417,9 +1445,11 @@ static noinline u64 find_delalloc_range(struct extent_io_tree *tree,
{
struct rb_node *node;
struct extent_state *state;
+ struct btrfs_root *root;
u64 cur_start = *start;
u64 found = 0;
u64 total_bytes = 0;
+ int flag = EXTENT_DELALLOC;

spin_lock(&tree->lock);

@@ -1434,13 +1464,27 @@ static noinline u64 find_delalloc_range(struct extent_io_tree *tree,
goto out;
}

+ root = BTRFS_I(tree->mapping->host)->root;
while (1) {
state = rb_entry(node, struct extent_state, rb_node);
if (found && (state->start != cur_start ||
(state->state & EXTENT_BOUNDARY))) {
goto out;
}
- if (!(state->state & EXTENT_DELALLOC)) {
+ if (btrfs_test_opt(root, HOT_MOVE)) {
+ if (!(state->state & EXTENT_DELALLOC) ||
+ (!(state->state & EXTENT_HOT) &&
+ !(state->state & EXTENT_COLD))) {
+ if (!found)
+ *end = state->end;
+ goto out;
+ } else {
+ if (!found)
+ flag = (state->state & EXTENT_HOT) ?
+ EXTENT_HOT : EXTENT_COLD;
+ }
+ }
+ if (!(state->state & flag)) {
if (!found)
*end = state->end;
goto out;
@@ -1627,7 +1671,13 @@ again:
lock_extent_bits(tree, delalloc_start, delalloc_end, 0, &cached_state);

/* then test to make sure it is all still delalloc */
- ret = test_range_bit(tree, delalloc_start, delalloc_end,
+ if (btrfs_test_opt(BTRFS_I(inode)->root, HOT_MOVE)) {
+ ret = test_range_bit(tree, delalloc_start, delalloc_end,
+ EXTENT_DELALLOC | EXTENT_HOT, 1, cached_state);
+ ret |= test_range_bit(tree, delalloc_start, delalloc_end,
+ EXTENT_DELALLOC | EXTENT_COLD, 1, cached_state);
+ } else
+ ret = test_range_bit(tree, delalloc_start, delalloc_end,
EXTENT_DELALLOC, 1, cached_state);
if (!ret) {
unlock_extent_cached(tree, delalloc_start, delalloc_end,
@@ -1665,6 +1715,11 @@ int extent_clear_unlock_delalloc(struct inode *inode,
if (op & EXTENT_CLEAR_DELALLOC)
clear_bits |= EXTENT_DELALLOC;

+ if (op & EXTENT_CLEAR_HOT)
+ clear_bits |= EXTENT_HOT;
+ if (op & EXTENT_CLEAR_COLD)
+ clear_bits |= EXTENT_COLD;
+
clear_extent_bit(tree, start, end, clear_bits, 1, 0, NULL, GFP_NOFS);
if (!(op & (EXTENT_CLEAR_UNLOCK_PAGE | EXTENT_CLEAR_DIRTY |
EXTENT_SET_WRITEBACK | EXTENT_END_WRITEBACK |
diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h
index a2c03a1..a3bfc9d 100644
--- a/fs/btrfs/extent_io.h
+++ b/fs/btrfs/extent_io.h
@@ -19,6 +19,8 @@
#define EXTENT_FIRST_DELALLOC (1 << 12)
#define EXTENT_NEED_WAIT (1 << 13)
#define EXTENT_DAMAGED (1 << 14)
+#define EXTENT_HOT (1 << 15)
+#define EXTENT_COLD (1 << 16)
#define EXTENT_IOBITS (EXTENT_LOCKED | EXTENT_WRITEBACK)
#define EXTENT_CTLBITS (EXTENT_DO_ACCOUNTING | EXTENT_FIRST_DELALLOC)

@@ -51,6 +53,8 @@
#define EXTENT_END_WRITEBACK 0x20
#define EXTENT_SET_PRIVATE2 0x40
#define EXTENT_CLEAR_ACCOUNTING 0x80
+#define EXTENT_CLEAR_HOT 0x100
+#define EXTENT_CLEAR_COLD 0x200

/*
* page->private values. Every page that is controlled by the extent
@@ -237,6 +241,9 @@ int set_extent_delalloc(struct extent_io_tree *tree, u64 start, u64 end,
struct extent_state **cached_state, gfp_t mask);
int set_extent_defrag(struct extent_io_tree *tree, u64 start, u64 end,
struct extent_state **cached_state, gfp_t mask);
+void set_extent_hot(struct inode *inode, u64 start, u64 end,
+ struct extent_state **cached_state,
+ int type, int flag);
int find_first_extent_bit(struct extent_io_tree *tree, u64 start,
u64 *start_ret, u64 *end_ret, unsigned long bits,
struct extent_state **cached_state);
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index 4205ba7..4cbf236 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -41,6 +41,7 @@
#include "locking.h"
#include "compat.h"
#include "volumes.h"
+#include "hot_relocate.h"

static struct kmem_cache *btrfs_inode_defrag_cachep;
/*
@@ -500,7 +501,7 @@ static void btrfs_drop_pages(struct page **pages, size_t num_pages)
int btrfs_dirty_pages(struct btrfs_root *root, struct inode *inode,
struct page **pages, size_t num_pages,
loff_t pos, size_t write_bytes,
- struct extent_state **cached)
+ struct extent_state **cached, int flag)
{
int err = 0;
int i;
@@ -514,6 +515,11 @@ int btrfs_dirty_pages(struct btrfs_root *root, struct inode *inode,
num_bytes = ALIGN(write_bytes + pos - start_pos, root->sectorsize);

end_of_last_block = start_pos + num_bytes - 1;
+
+ if (btrfs_test_opt(root, HOT_MOVE))
+ set_extent_hot(inode, start_pos, end_of_last_block,
+ cached, flag, 0);
+
err = btrfs_set_extent_delalloc(inode, start_pos, end_of_last_block,
cached);
if (err)
@@ -1350,6 +1356,7 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
size_t dirty_pages;
size_t copied;
+ int flag = TYPE_ROT;

WARN_ON(num_pages > nrptrs);

@@ -1363,7 +1370,7 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
}

ret = btrfs_delalloc_reserve_space(inode,
- num_pages << PAGE_CACHE_SHIFT);
+ num_pages << PAGE_CACHE_SHIFT, &flag);
if (ret)
break;

@@ -1377,7 +1384,7 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
force_page_uptodate);
if (ret) {
btrfs_delalloc_release_space(inode,
- num_pages << PAGE_CACHE_SHIFT);
+ num_pages << PAGE_CACHE_SHIFT, flag);
break;
}

@@ -1416,16 +1423,16 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file,
}
btrfs_delalloc_release_space(inode,
(num_pages - dirty_pages) <<
- PAGE_CACHE_SHIFT);
+ PAGE_CACHE_SHIFT, flag);
}

if (copied > 0) {
ret = btrfs_dirty_pages(root, inode, pages,
dirty_pages, pos, copied,
- NULL);
+ NULL, flag);
if (ret) {
btrfs_delalloc_release_space(inode,
- dirty_pages << PAGE_CACHE_SHIFT);
+ dirty_pages << PAGE_CACHE_SHIFT, flag);
btrfs_drop_pages(pages, num_pages);
break;
}
@@ -2150,6 +2157,7 @@ static long btrfs_fallocate(struct file *file, int mode,
u64 locked_end;
struct extent_map *em;
int blocksize = BTRFS_I(inode)->root->sectorsize;
+ int flag = TYPE_ROT;
int ret;

alloc_start = round_down(offset, blocksize);
@@ -2166,7 +2174,7 @@ static long btrfs_fallocate(struct file *file, int mode,
* Make sure we have enough space before we do the
* allocation.
*/
- ret = btrfs_check_data_free_space(inode, alloc_end - alloc_start);
+ ret = btrfs_check_data_free_space(inode, alloc_end - alloc_start, &flag);
if (ret)
return ret;
if (root->fs_info->quota_enabled) {
@@ -2281,7 +2289,7 @@ out:
btrfs_qgroup_free(root, alloc_end - alloc_start);
out_reserve_fail:
/* Let go of our reservation. */
- btrfs_free_reserved_data_space(inode, alloc_end - alloc_start);
+ btrfs_free_reserved_data_space(inode, alloc_end - alloc_start, flag);
return ret;
}

diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c
index ecca6c7..58a1cc3 100644
--- a/fs/btrfs/free-space-cache.c
+++ b/fs/btrfs/free-space-cache.c
@@ -1007,7 +1007,7 @@ static int __btrfs_write_out_cache(struct btrfs_root *root, struct inode *inode,
io_ctl_zero_remaining_pages(&io_ctl);

ret = btrfs_dirty_pages(root, inode, io_ctl.pages, io_ctl.num_pages,
- 0, i_size_read(inode), &cached_state);
+ 0, i_size_read(inode), &cached_state, TYPE_ROT);
io_ctl_drop_pages(&io_ctl);
unlock_extent_cached(&BTRFS_I(inode)->io_tree, 0,
i_size_read(inode) - 1, &cached_state, GFP_NOFS);
diff --git a/fs/btrfs/inode-map.c b/fs/btrfs/inode-map.c
index d26f67a..ef0c79d 100644
--- a/fs/btrfs/inode-map.c
+++ b/fs/btrfs/inode-map.c
@@ -403,6 +403,7 @@ int btrfs_save_ino_cache(struct btrfs_root *root,
u64 alloc_hint = 0;
int ret;
int prealloc;
+ int flag = TYPE_ROT;
bool retry = false;

/* only fs tree and subvol/snap needs ino cache */
@@ -490,17 +491,17 @@ again:
/* Just to make sure we have enough space */
prealloc += 8 * PAGE_CACHE_SIZE;

- ret = btrfs_delalloc_reserve_space(inode, prealloc);
+ ret = btrfs_delalloc_reserve_space(inode, prealloc, &flag);
if (ret)
goto out_put;

ret = btrfs_prealloc_file_range_trans(inode, trans, 0, 0, prealloc,
prealloc, prealloc, &alloc_hint);
if (ret) {
- btrfs_delalloc_release_space(inode, prealloc);
+ btrfs_delalloc_release_space(inode, prealloc, flag);
goto out_put;
}
- btrfs_free_reserved_data_space(inode, prealloc);
+ btrfs_free_reserved_data_space(inode, prealloc, flag);

ret = btrfs_write_out_ino_cache(root, trans, path);
out_put:
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 9b31b3b..096f97f 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -57,6 +57,7 @@
#include "free-space-cache.h"
#include "inode-map.h"
#include "backref.h"
+#include "hot_relocate.h"

struct btrfs_iget_args {
u64 ino;
@@ -106,6 +107,27 @@ static struct extent_map *create_pinned_em(struct inode *inode, u64 start,

static int btrfs_dirty_inode(struct inode *inode);

+static int get_chunk_type(struct inode *inode, u64 start, u64 end)
+{
+ int hot, cold, ret = 1;
+
+ hot = test_range_bit(&BTRFS_I(inode)->io_tree,
+ start, end, EXTENT_HOT, 1, NULL);
+ cold = test_range_bit(&BTRFS_I(inode)->io_tree,
+ start, end, EXTENT_COLD, 1, NULL);
+
+ WARN_ON(hot && cold);
+
+ if (hot)
+ ret = 2;
+ else if (cold)
+ ret = 1;
+ else
+ WARN_ON(1);
+
+ return ret;
+}
+
static int btrfs_init_inode_security(struct btrfs_trans_handle *trans,
struct inode *inode, struct inode *dir,
const struct qstr *qstr)
@@ -859,13 +881,14 @@ static noinline int __cow_file_range(struct btrfs_trans_handle *trans,
{
u64 alloc_hint = 0;
u64 num_bytes;
- unsigned long ram_size;
+ unsigned long ram_size, hot_flag = 0;
u64 disk_num_bytes;
u64 cur_alloc_size;
u64 blocksize = root->sectorsize;
struct btrfs_key ins;
struct extent_map *em;
struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree;
+ int chunk_type = 1;
int ret = 0;

BUG_ON(btrfs_is_free_space_inode(inode));
@@ -873,6 +896,7 @@ static noinline int __cow_file_range(struct btrfs_trans_handle *trans,
num_bytes = ALIGN(end - start + 1, blocksize);
num_bytes = max(blocksize, num_bytes);
disk_num_bytes = num_bytes;
+ ret = 0;

/* if this is a small write inside eof, kick off defrag */
if (num_bytes < 64 * 1024 &&
@@ -892,7 +916,8 @@ static noinline int __cow_file_range(struct btrfs_trans_handle *trans,
EXTENT_CLEAR_DELALLOC |
EXTENT_CLEAR_DIRTY |
EXTENT_SET_WRITEBACK |
- EXTENT_END_WRITEBACK);
+ EXTENT_END_WRITEBACK |
+ hot_flag);

*nr_written = *nr_written +
(end - start + PAGE_CACHE_SIZE) / PAGE_CACHE_SIZE;
@@ -914,9 +939,25 @@ static noinline int __cow_file_range(struct btrfs_trans_handle *trans,
unsigned long op;

cur_alloc_size = disk_num_bytes;
+
+ /*
+ * Use COW operations to move hot data to SSD and cold data
+ * back to rotating disk. Sets chunk_type to 1 to indicate
+ * to write to BTRFS_BLOCK_GROUP_DATA or 2 to indicate
+ * BTRFS_BLOCK_GROUP_DATA_NONROT.
+ */
+ if (btrfs_test_opt(root, HOT_MOVE)) {
+ chunk_type = get_chunk_type(inode, start,
+ start + cur_alloc_size - 1);
+ if (chunk_type == 2)
+ hot_flag = EXTENT_CLEAR_HOT;
+ else
+ hot_flag = EXTENT_CLEAR_COLD;
+ }
+
ret = btrfs_reserve_extent(trans, root, cur_alloc_size,
root->sectorsize, 0, alloc_hint,
- &ins, 1);
+ &ins, chunk_type);
if (ret < 0) {
btrfs_abort_transaction(trans, root, ret);
goto out_unlock;
@@ -982,7 +1023,7 @@ static noinline int __cow_file_range(struct btrfs_trans_handle *trans,
*/
op = unlock ? EXTENT_CLEAR_UNLOCK_PAGE : 0;
op |= EXTENT_CLEAR_UNLOCK | EXTENT_CLEAR_DELALLOC |
- EXTENT_SET_PRIVATE2;
+ EXTENT_SET_PRIVATE2 | hot_flag;

extent_clear_unlock_delalloc(inode, &BTRFS_I(inode)->io_tree,
start, start + ram_size - 1,
@@ -1006,7 +1047,8 @@ out_unlock:
EXTENT_CLEAR_DELALLOC |
EXTENT_CLEAR_DIRTY |
EXTENT_SET_WRITEBACK |
- EXTENT_END_WRITEBACK);
+ EXTENT_END_WRITEBACK |
+ hot_flag);

goto out;
}
@@ -1600,8 +1642,12 @@ static void btrfs_clear_bit_hook(struct inode *inode,
btrfs_delalloc_release_metadata(inode, len);

if (root->root_key.objectid != BTRFS_DATA_RELOC_TREE_OBJECTID
- && do_list)
- btrfs_free_reserved_data_space(inode, len);
+ && do_list) {
+ int flag = TYPE_ROT;
+ if ((state->state & EXTENT_HOT) && (*bits & EXTENT_HOT))
+ flag = TYPE_NONROT;
+ btrfs_free_reserved_data_space(inode, len, flag);
+ }

__percpu_counter_add(&root->fs_info->delalloc_bytes, -len,
root->fs_info->delalloc_batch);
@@ -1796,6 +1842,7 @@ static void btrfs_writepage_fixup_worker(struct btrfs_work *work)
u64 page_start;
u64 page_end;
int ret;
+ int flag = TYPE_ROT;

fixup = container_of(work, struct btrfs_writepage_fixup, work);
page = fixup->page;
@@ -1827,7 +1874,7 @@ again:
goto again;
}

- ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE);
+ ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE, &flag);
if (ret) {
mapping_set_error(page->mapping, ret);
end_extent_writepage(page, ret, page_start, page_end);
@@ -1835,6 +1882,10 @@ again:
goto out;
}

+ if (btrfs_test_opt(BTRFS_I(inode)->root, HOT_MOVE))
+ set_extent_hot(inode, page_start, page_end,
+ &cached_state, flag, 0);
+
btrfs_set_extent_delalloc(inode, page_start, page_end, &cached_state);
ClearPageChecked(page);
set_page_dirty(page);
@@ -4282,20 +4333,21 @@ int btrfs_truncate_page(struct inode *inode, loff_t from, loff_t len,
struct page *page;
gfp_t mask = btrfs_alloc_write_mask(mapping);
int ret = 0;
+ int flag = TYPE_ROT;
u64 page_start;
u64 page_end;

if ((offset & (blocksize - 1)) == 0 &&
(!len || ((len & (blocksize - 1)) == 0)))
goto out;
- ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE);
+ ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE, &flag);
if (ret)
goto out;

again:
page = find_or_create_page(mapping, index, mask);
if (!page) {
- btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE);
+ btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE, flag);
ret = -ENOMEM;
goto out;
}
@@ -4337,6 +4389,10 @@ again:
EXTENT_DO_ACCOUNTING | EXTENT_DEFRAG,
0, 0, &cached_state, GFP_NOFS);

+ if (btrfs_test_opt(root, HOT_MOVE))
+ set_extent_hot(inode, page_start, page_end,
+ &cached_state, flag, 0);
+
ret = btrfs_set_extent_delalloc(inode, page_start, page_end,
&cached_state);
if (ret) {
@@ -4363,7 +4419,7 @@ again:

out_unlock:
if (ret)
- btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE);
+ btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE, flag);
unlock_page(page);
page_cache_release(page);
out:
@@ -7353,6 +7409,7 @@ static ssize_t btrfs_direct_IO(int rw, struct kiocb *iocb,
struct inode *inode = file->f_mapping->host;
size_t count = 0;
int flags = 0;
+ int flag = TYPE_ROT;
bool wakeup = true;
bool relock = false;
ssize_t ret;
@@ -7375,7 +7432,7 @@ static ssize_t btrfs_direct_IO(int rw, struct kiocb *iocb,
mutex_unlock(&inode->i_mutex);
relock = true;
}
- ret = btrfs_delalloc_reserve_space(inode, count);
+ ret = btrfs_delalloc_reserve_space(inode, count, &flag);
if (ret)
goto out;
} else if (unlikely(test_bit(BTRFS_INODE_READDIO_NEED_LOCK,
@@ -7391,10 +7448,10 @@ static ssize_t btrfs_direct_IO(int rw, struct kiocb *iocb,
btrfs_submit_direct, flags);
if (rw & WRITE) {
if (ret < 0 && ret != -EIOCBQUEUED)
- btrfs_delalloc_release_space(inode, count);
+ btrfs_delalloc_release_space(inode, count, flag);
else if (ret >= 0 && (size_t)ret < count)
btrfs_delalloc_release_space(inode,
- count - (size_t)ret);
+ count - (size_t)ret, flag);
else
btrfs_delalloc_release_metadata(inode, 0);
}
@@ -7573,11 +7630,12 @@ int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
loff_t size;
int ret;
int reserved = 0;
+ int flag = TYPE_ROT;
u64 page_start;
u64 page_end;

sb_start_pagefault(inode->i_sb);
- ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE);
+ ret = btrfs_delalloc_reserve_space(inode, PAGE_CACHE_SIZE, &flag);
if (!ret) {
ret = file_update_time(vma->vm_file);
reserved = 1;
@@ -7635,6 +7693,10 @@ again:
EXTENT_DO_ACCOUNTING | EXTENT_DEFRAG,
0, 0, &cached_state, GFP_NOFS);

+ if (btrfs_test_opt(root, HOT_MOVE))
+ set_extent_hot(inode, page_start, page_end,
+ &cached_state, flag, 0);
+
ret = btrfs_set_extent_delalloc(inode, page_start, page_end,
&cached_state);
if (ret) {
@@ -7674,7 +7736,7 @@ out_unlock:
}
unlock_page(page);
out:
- btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE);
+ btrfs_delalloc_release_space(inode, PAGE_CACHE_SIZE, flag);
out_noreserve:
sb_end_pagefault(inode->i_sb);
return ret;
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 0de4a2f..91da5ae 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -56,6 +56,7 @@
#include "rcu-string.h"
#include "send.h"
#include "dev-replace.h"
+#include "hot_relocate.h"

/* Mask out flags that are inappropriate for the given type of inode. */
static inline __u32 btrfs_mask_flags(umode_t mode, __u32 flags)
@@ -1001,6 +1002,7 @@ static int cluster_pages_for_defrag(struct inode *inode,
int ret;
int i;
int i_done;
+ int flag = TYPE_ROT;
struct btrfs_ordered_extent *ordered;
struct extent_state *cached_state = NULL;
struct extent_io_tree *tree;
@@ -1013,7 +1015,7 @@ static int cluster_pages_for_defrag(struct inode *inode,
page_cnt = min_t(u64, (u64)num_pages, (u64)file_end - start_index + 1);

ret = btrfs_delalloc_reserve_space(inode,
- page_cnt << PAGE_CACHE_SHIFT);
+ page_cnt << PAGE_CACHE_SHIFT, &flag);
if (ret)
return ret;
i_done = 0;
@@ -1101,9 +1103,12 @@ again:
BTRFS_I(inode)->outstanding_extents++;
spin_unlock(&BTRFS_I(inode)->lock);
btrfs_delalloc_release_space(inode,
- (page_cnt - i_done) << PAGE_CACHE_SHIFT);
+ (page_cnt - i_done) << PAGE_CACHE_SHIFT, flag);
}

+ if (btrfs_test_opt(BTRFS_I(inode)->root, HOT_MOVE))
+ set_extent_hot(inode, page_start, page_end - 1,
+ &cached_state, flag, 0);

set_extent_defrag(&BTRFS_I(inode)->io_tree, page_start, page_end - 1,
&cached_state, GFP_NOFS);
@@ -1126,7 +1131,8 @@ out:
unlock_page(pages[i]);
page_cache_release(pages[i]);
}
- btrfs_delalloc_release_space(inode, page_cnt << PAGE_CACHE_SHIFT);
+ btrfs_delalloc_release_space(inode,
+ page_cnt << PAGE_CACHE_SHIFT, flag);
return ret;

}
@@ -3021,8 +3027,9 @@ static long btrfs_ioctl_space_info(struct btrfs_root *root, void __user *arg)
u64 types[] = {BTRFS_BLOCK_GROUP_DATA,
BTRFS_BLOCK_GROUP_SYSTEM,
BTRFS_BLOCK_GROUP_METADATA,
- BTRFS_BLOCK_GROUP_DATA | BTRFS_BLOCK_GROUP_METADATA};
- int num_types = 4;
+ BTRFS_BLOCK_GROUP_DATA | BTRFS_BLOCK_GROUP_METADATA,
+ BTRFS_BLOCK_GROUP_DATA_NONROT};
+ int num_types = 5;
int alloc_size;
int ret = 0;
u64 slot_count = 0;
diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 704a1b8..62c5897 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -31,6 +31,7 @@
#include "async-thread.h"
#include "free-space-cache.h"
#include "inode-map.h"
+#include "hot_relocate.h"

/*
* backref_node, mapping_node and tree_block start with this
@@ -2938,12 +2939,13 @@ int prealloc_file_extent_cluster(struct inode *inode,
u64 num_bytes;
int nr = 0;
int ret = 0;
+ int flag = TYPE_ROT;

BUG_ON(cluster->start != cluster->boundary[0]);
mutex_lock(&inode->i_mutex);

ret = btrfs_check_data_free_space(inode, cluster->end +
- 1 - cluster->start);
+ 1 - cluster->start, &flag);
if (ret)
goto out;

@@ -2965,7 +2967,7 @@ int prealloc_file_extent_cluster(struct inode *inode,
nr++;
}
btrfs_free_reserved_data_space(inode, cluster->end +
- 1 - cluster->start);
+ 1 - cluster->start, flag);
out:
mutex_unlock(&inode->i_mutex);
return ret;
diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index 09fb9d2..c10477b 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -58,6 +58,7 @@
#include "rcu-string.h"
#include "dev-replace.h"
#include "free-space-cache.h"
+#include "hot_relocate.h"

#define CREATE_TRACE_POINTS
#include <trace/events/btrfs.h>
@@ -1520,7 +1521,8 @@ static int btrfs_statfs(struct dentry *dentry, struct kstatfs *buf)
mutex_lock(&fs_info->chunk_mutex);
rcu_read_lock();
list_for_each_entry_rcu(found, head, list) {
- if (found->flags & BTRFS_BLOCK_GROUP_DATA) {
+ if ((found->flags & BTRFS_BLOCK_GROUP_DATA) ||
+ (found->flags & BTRFS_BLOCK_GROUP_DATA_NONROT)) {
total_free_data += found->disk_total - found->disk_used;
total_free_data -=
btrfs_account_ro_block_groups_free_space(found);
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index 0e925ce..29e416d 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -1451,6 +1451,9 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path)
all_avail = root->fs_info->avail_data_alloc_bits |
root->fs_info->avail_system_alloc_bits |
root->fs_info->avail_metadata_alloc_bits;
+ if (btrfs_test_opt(root, HOT_MOVE))
+ all_avail |=
+ root->fs_info->avail_data_nonrot_alloc_bits;
} while (read_seqretry(&root->fs_info->profiles_lock, seq));

num_devices = root->fs_info->fs_devices->num_devices;
@@ -3729,7 +3732,8 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
devs_increment = btrfs_raid_array[index].devs_increment;
ncopies = btrfs_raid_array[index].ncopies;

- if (type & BTRFS_BLOCK_GROUP_DATA) {
+ if (type & BTRFS_BLOCK_GROUP_DATA ||
+ type & BTRFS_BLOCK_GROUP_DATA_NONROT) {
max_stripe_size = 1024 * 1024 * 1024;
max_chunk_size = 10 * max_stripe_size;
} else if (type & BTRFS_BLOCK_GROUP_METADATA) {
@@ -3768,9 +3772,30 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
struct btrfs_device *device;
u64 max_avail;
u64 dev_offset;
+ int dev_rot;
+ int skip = 0;

device = list_entry(cur, struct btrfs_device, dev_alloc_list);

+ /*
+ * If HOT_MOVE is set, the chunk type being allocated
+ * determines which disks the data may be allocated on.
+ * This can cause problems if, for example, the data alloc
+ * profile is RAID0 and there are only two devices, 1 SSD +
+ * 1 HDD. All allocations to BTRFS_BLOCK_GROUP_DATA_NONROT
+ * in this config will return -ENOSPC as the allocation code
+ * can't find allowable space for the second stripe.
+ */
+ dev_rot = !blk_queue_nonrot(bdev_get_queue(device->bdev));
+ if (btrfs_test_opt(extent_root, HOT_MOVE)) {
+ int ret1 = type & (BTRFS_BLOCK_GROUP_DATA |
+ BTRFS_BLOCK_GROUP_METADATA |
+ BTRFS_BLOCK_GROUP_SYSTEM) && !dev_rot;
+ int ret2 = type & BTRFS_BLOCK_GROUP_DATA_NONROT && dev_rot;
+ if (ret1 || ret2)
+ skip = 1;
+ }
+
cur = cur->next;

if (!device->writeable) {
@@ -3779,7 +3804,7 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
continue;
}

- if (!device->in_fs_metadata ||
+ if (skip || !device->in_fs_metadata ||
device->is_tgtdev_for_dev_replace)
continue;

--
1.7.11.7

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/