diff options
author | Chris Mason <chris.mason@fusionio.com> | 2013-02-20 14:06:05 -0500 |
---|---|---|
committer | Chris Mason <chris.mason@fusionio.com> | 2013-02-20 14:06:05 -0500 |
commit | e942f883bc6651d50be139477baf6fb0eed3d5bb (patch) | |
tree | e1d19783e9c8b42198a69c17c9719fb90f302847 /fs/btrfs | |
parent | b2c6b3e0611c58fbeb6b9c0892b6249f7bdfaf6b (diff) | |
parent | 0e4e02636611dbf89a2f36320a32054f9936d6cb (diff) |
Merge branch 'raid56-experimental' into for-linus-3.9
Signed-off-by: Chris Mason <chris.mason@fusionio.com>
Conflicts:
fs/btrfs/ctree.h
fs/btrfs/extent-tree.c
fs/btrfs/inode.c
fs/btrfs/volumes.c
Diffstat (limited to 'fs/btrfs')
-rw-r--r-- | fs/btrfs/Kconfig | 3 | ||||
-rw-r--r-- | fs/btrfs/Makefile | 2 | ||||
-rw-r--r-- | fs/btrfs/compression.c | 4 | ||||
-rw-r--r-- | fs/btrfs/ctree.h | 44 | ||||
-rw-r--r-- | fs/btrfs/delayed-ref.h | 9 | ||||
-rw-r--r-- | fs/btrfs/disk-io.c | 62 | ||||
-rw-r--r-- | fs/btrfs/disk-io.h | 7 | ||||
-rw-r--r-- | fs/btrfs/extent-tree.c | 156 | ||||
-rw-r--r-- | fs/btrfs/extent_io.c | 40 | ||||
-rw-r--r-- | fs/btrfs/extent_io.h | 2 | ||||
-rw-r--r-- | fs/btrfs/free-space-cache.c | 50 | ||||
-rw-r--r-- | fs/btrfs/inode.c | 18 | ||||
-rw-r--r-- | fs/btrfs/raid56.c | 2080 | ||||
-rw-r--r-- | fs/btrfs/raid56.h | 51 | ||||
-rw-r--r-- | fs/btrfs/scrub.c | 8 | ||||
-rw-r--r-- | fs/btrfs/transaction.c | 9 | ||||
-rw-r--r-- | fs/btrfs/volumes.c | 380 | ||||
-rw-r--r-- | fs/btrfs/volumes.h | 9 |
18 files changed, 2814 insertions, 120 deletions
diff --git a/fs/btrfs/Kconfig b/fs/btrfs/Kconfig index d33f01c08b60..5f583c8a36d0 100644 --- a/fs/btrfs/Kconfig +++ b/fs/btrfs/Kconfig @@ -6,6 +6,9 @@ config BTRFS_FS select ZLIB_DEFLATE select LZO_COMPRESS select LZO_DECOMPRESS + select RAID6_PQ + select XOR_BLOCKS + help Btrfs is a new filesystem with extents, writable snapshotting, support for multiple devices and many more features. diff --git a/fs/btrfs/Makefile b/fs/btrfs/Makefile index 7df3e0f0ee51..3932224f99e9 100644 --- a/fs/btrfs/Makefile +++ b/fs/btrfs/Makefile @@ -8,7 +8,7 @@ btrfs-y += super.o ctree.o extent-tree.o print-tree.o root-tree.o dir-item.o \ extent_io.o volumes.o async-thread.o ioctl.o locking.o orphan.o \ export.o tree-log.o free-space-cache.o zlib.o lzo.o \ compression.o delayed-ref.o relocation.o delayed-inode.o scrub.o \ - reada.o backref.o ulist.o qgroup.o send.o dev-replace.o + reada.o backref.o ulist.o qgroup.o send.o dev-replace.o raid56.o btrfs-$(CONFIG_BTRFS_FS_POSIX_ACL) += acl.o btrfs-$(CONFIG_BTRFS_FS_CHECK_INTEGRITY) += check-integrity.o diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index 94ab2f80e7e3..15b94089abc4 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -372,7 +372,7 @@ int btrfs_submit_compressed_write(struct inode *inode, u64 start, page = compressed_pages[pg_index]; page->mapping = inode->i_mapping; if (bio->bi_size) - ret = io_tree->ops->merge_bio_hook(page, 0, + ret = io_tree->ops->merge_bio_hook(WRITE, page, 0, PAGE_CACHE_SIZE, bio, 0); else @@ -655,7 +655,7 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio, page->index = em_start >> PAGE_CACHE_SHIFT; if (comp_bio->bi_size) - ret = tree->ops->merge_bio_hook(page, 0, + ret = tree->ops->merge_bio_hook(READ, page, 0, PAGE_CACHE_SIZE, comp_bio, 0); else diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 1679051f4d39..3dcedfe4f759 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -506,6 +506,7 @@ struct btrfs_super_block { #define BTRFS_FEATURE_INCOMPAT_BIG_METADATA (1ULL << 5) #define BTRFS_FEATURE_INCOMPAT_EXTENDED_IREF (1ULL << 6) +#define BTRFS_FEATURE_INCOMPAT_RAID56 (1ULL << 7) #define BTRFS_FEATURE_COMPAT_SUPP 0ULL #define BTRFS_FEATURE_COMPAT_RO_SUPP 0ULL @@ -515,6 +516,7 @@ struct btrfs_super_block { BTRFS_FEATURE_INCOMPAT_MIXED_GROUPS | \ BTRFS_FEATURE_INCOMPAT_BIG_METADATA | \ BTRFS_FEATURE_INCOMPAT_COMPRESS_LZO | \ + BTRFS_FEATURE_INCOMPAT_RAID56 | \ BTRFS_FEATURE_INCOMPAT_EXTENDED_IREF) /* @@ -956,6 +958,8 @@ struct btrfs_dev_replace_item { #define BTRFS_BLOCK_GROUP_RAID1 (1ULL << 4) #define BTRFS_BLOCK_GROUP_DUP (1ULL << 5) #define BTRFS_BLOCK_GROUP_RAID10 (1ULL << 6) +#define BTRFS_BLOCK_GROUP_RAID5 (1 << 7) +#define BTRFS_BLOCK_GROUP_RAID6 (1 << 8) #define BTRFS_BLOCK_GROUP_RESERVED BTRFS_AVAIL_ALLOC_BIT_SINGLE enum btrfs_raid_types { @@ -964,6 +968,8 @@ enum btrfs_raid_types { BTRFS_RAID_DUP, BTRFS_RAID_RAID0, BTRFS_RAID_SINGLE, + BTRFS_RAID_RAID5, + BTRFS_RAID_RAID6, BTRFS_NR_RAID_TYPES }; @@ -973,6 +979,8 @@ enum btrfs_raid_types { #define BTRFS_BLOCK_GROUP_PROFILE_MASK (BTRFS_BLOCK_GROUP_RAID0 | \ BTRFS_BLOCK_GROUP_RAID1 | \ + BTRFS_BLOCK_GROUP_RAID5 | \ + BTRFS_BLOCK_GROUP_RAID6 | \ BTRFS_BLOCK_GROUP_DUP | \ BTRFS_BLOCK_GROUP_RAID10) /* @@ -1197,6 +1205,10 @@ struct btrfs_block_group_cache { u64 flags; u64 sectorsize; u64 cache_generation; + + /* for raid56, this is a full stripe, without parity */ + unsigned long full_stripe_len; + unsigned int ro:1; unsigned int dirty:1; unsigned int iref:1; @@ -1242,6 +1254,23 @@ enum btrfs_orphan_cleanup_state { ORPHAN_CLEANUP_DONE = 2, }; +/* used by the raid56 code to lock stripes for read/modify/write */ +struct btrfs_stripe_hash { + struct list_head hash_list; + wait_queue_head_t wait; + spinlock_t lock; +}; + +/* used by the raid56 code to lock stripes for read/modify/write */ +struct btrfs_stripe_hash_table { + struct list_head stripe_cache; + spinlock_t cache_lock; + int cache_size; + struct btrfs_stripe_hash table[]; +}; + +#define BTRFS_STRIPE_HASH_TABLE_BITS 11 + /* fs_info */ struct reloc_control; struct btrfs_device; @@ -1341,6 +1370,13 @@ struct btrfs_fs_info { struct mutex cleaner_mutex; struct mutex chunk_mutex; struct mutex volume_mutex; + + /* this is used during read/modify/write to make sure + * no two ios are trying to mod the same stripe at the same + * time + */ + struct btrfs_stripe_hash_table *stripe_hash_table; + /* * this protects the ordered operations list only while we are * processing all of the entries on it. This way we make @@ -1423,6 +1459,8 @@ struct btrfs_fs_info { struct btrfs_workers flush_workers; struct btrfs_workers endio_workers; struct btrfs_workers endio_meta_workers; + struct btrfs_workers endio_raid56_workers; + struct btrfs_workers rmw_workers; struct btrfs_workers endio_meta_write_workers; struct btrfs_workers endio_write_workers; struct btrfs_workers endio_freespace_worker; @@ -3490,9 +3528,9 @@ int btrfs_writepages(struct address_space *mapping, struct writeback_control *wbc); int btrfs_create_subvol_root(struct btrfs_trans_handle *trans, struct btrfs_root *new_root, u64 new_dirid); -int btrfs_merge_bio_hook(struct page *page, unsigned long offset, - size_t size, struct bio *bio, unsigned long bio_flags); - +int btrfs_merge_bio_hook(int rw, struct page *page, unsigned long offset, + size_t size, struct bio *bio, + unsigned long bio_flags); int btrfs_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf); int btrfs_readpage(struct file *file, struct page *page); void btrfs_evict_inode(struct inode *inode); diff --git a/fs/btrfs/delayed-ref.h b/fs/btrfs/delayed-ref.h index 7939149f8f27..f75fcaf79aeb 100644 --- a/fs/btrfs/delayed-ref.h +++ b/fs/btrfs/delayed-ref.h @@ -132,6 +132,15 @@ struct btrfs_delayed_ref_root { unsigned long num_heads_ready; /* + * bumped when someone is making progress on the delayed + * refs, so that other procs know they are just adding to + * contention intead of helping + */ + atomic_t procs_running_refs; + atomic_t ref_seq; + wait_queue_head_t wait; + + /* * set when the tree is flushing before a transaction commit, * used by the throttling code to decide if new updates need * to be run right away diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 779b401cd952..eb7c14308521 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -46,6 +46,7 @@ #include "check-integrity.h" #include "rcu-string.h" #include "dev-replace.h" +#include "raid56.h" #ifdef CONFIG_X86 #include <asm/cpufeature.h> @@ -640,8 +641,15 @@ err: btree_readahead_hook(root, eb, eb->start, ret); } - if (ret) + if (ret) { + /* + * our io error hook is going to dec the io pages + * again, we have to make sure it has something + * to decrement + */ + atomic_inc(&eb->io_pages); clear_extent_buffer_uptodate(eb); + } free_extent_buffer(eb); out: return ret; @@ -655,6 +663,7 @@ static int btree_io_failed_hook(struct page *page, int failed_mirror) eb = (struct extent_buffer *)page->private; set_bit(EXTENT_BUFFER_IOERR, &eb->bflags); eb->read_mirror = failed_mirror; + atomic_dec(&eb->io_pages); if (test_and_clear_bit(EXTENT_BUFFER_READAHEAD, &eb->bflags)) btree_readahead_hook(root, eb, eb->start, -EIO); return -EIO; /* we fixed nothing */ @@ -671,17 +680,23 @@ static void end_workqueue_bio(struct bio *bio, int err) end_io_wq->work.flags = 0; if (bio->bi_rw & REQ_WRITE) { - if (end_io_wq->metadata == 1) + if (end_io_wq->metadata == BTRFS_WQ_ENDIO_METADATA) btrfs_queue_worker(&fs_info->endio_meta_write_workers, &end_io_wq->work); - else if (end_io_wq->metadata == 2) + else if (end_io_wq->metadata == BTRFS_WQ_ENDIO_FREE_SPACE) btrfs_queue_worker(&fs_info->endio_freespace_worker, &end_io_wq->work); + else if (end_io_wq->metadata == BTRFS_WQ_ENDIO_RAID56) + btrfs_queue_worker(&fs_info->endio_raid56_workers, + &end_io_wq->work); else btrfs_queue_worker(&fs_info->endio_write_workers, &end_io_wq->work); } else { - if (end_io_wq->metadata) + if (end_io_wq->metadata == BTRFS_WQ_ENDIO_RAID56) + btrfs_queue_worker(&fs_info->endio_raid56_workers, + &end_io_wq->work); + else if (end_io_wq->metadata) btrfs_queue_worker(&fs_info->endio_meta_workers, &end_io_wq->work); else @@ -696,6 +711,7 @@ static void end_workqueue_bio(struct bio *bio, int err) * 0 - if data * 1 - if normal metadta * 2 - if writing to the free space cache area + * 3 - raid parity work */ int btrfs_bio_wq_end_io(struct btrfs_fs_info *info, struct bio *bio, int metadata) @@ -2179,6 +2195,12 @@ int open_ctree(struct super_block *sb, init_waitqueue_head(&fs_info->transaction_blocked_wait); init_waitqueue_head(&fs_info->async_submit_wait); + ret = btrfs_alloc_stripe_hash_table(fs_info); + if (ret) { + err = -ENOMEM; + goto fail_alloc; + } + __setup_root(4096, 4096, 4096, 4096, tree_root, fs_info, BTRFS_ROOT_TREE_OBJECTID); @@ -2349,6 +2371,12 @@ int open_ctree(struct super_block *sb, btrfs_init_workers(&fs_info->endio_meta_write_workers, "endio-meta-write", fs_info->thread_pool_size, &fs_info->generic_worker); + btrfs_init_workers(&fs_info->endio_raid56_workers, + "endio-raid56", fs_info->thread_pool_size, + &fs_info->generic_worker); + btrfs_init_workers(&fs_info->rmw_workers, + "rmw", fs_info->thread_pool_size, + &fs_info->generic_worker); btrfs_init_workers(&fs_info->endio_write_workers, "endio-write", fs_info->thread_pool_size, &fs_info->generic_worker); @@ -2367,6 +2395,8 @@ int open_ctree(struct super_block *sb, */ fs_info->endio_workers.idle_thresh = 4; fs_info->endio_meta_workers.idle_thresh = 4; + fs_info->endio_raid56_workers.idle_thresh = 4; + fs_info->rmw_workers.idle_thresh = 2; fs_info->endio_write_workers.idle_thresh = 2; fs_info->endio_meta_write_workers.idle_thresh = 2; @@ -2383,6 +2413,8 @@ int open_ctree(struct super_block *sb, ret |= btrfs_start_workers(&fs_info->fixup_workers); ret |= btrfs_start_workers(&fs_info->endio_workers); ret |= btrfs_start_workers(&fs_info->endio_meta_workers); + ret |= btrfs_start_workers(&fs_info->rmw_workers); + ret |= btrfs_start_workers(&fs_info->endio_raid56_workers); ret |= btrfs_start_workers(&fs_info->endio_meta_write_workers); ret |= btrfs_start_workers(&fs_info->endio_write_workers); ret |= btrfs_start_workers(&fs_info->endio_freespace_worker); @@ -2726,6 +2758,8 @@ fail_sb_buffer: btrfs_stop_workers(&fs_info->workers); btrfs_stop_workers(&fs_info->endio_workers); btrfs_stop_workers(&fs_info->endio_meta_workers); + btrfs_stop_workers(&fs_info->endio_raid56_workers); + btrfs_stop_workers(&fs_info->rmw_workers); btrfs_stop_workers(&fs_info->endio_meta_write_workers); btrfs_stop_workers(&fs_info->endio_write_workers); btrfs_stop_workers(&fs_info->endio_freespace_worker); @@ -2747,6 +2781,7 @@ fail_bdi: fail_srcu: cleanup_srcu_struct(&fs_info->subvol_srcu); fail: + btrfs_free_stripe_hash_table(fs_info); btrfs_close_devices(fs_info->fs_devices); return err; @@ -3094,11 +3129,16 @@ int btrfs_calc_num_tolerated_disk_barrier_failures( ((flags & BTRFS_BLOCK_GROUP_PROFILE_MASK) == 0))) num_tolerated_disk_barrier_failures = 0; - else if (num_tolerated_disk_barrier_failures > 1 - && - (flags & (BTRFS_BLOCK_GROUP_RAID1 | - BTRFS_BLOCK_GROUP_RAID10))) - num_tolerated_disk_barrier_failures = 1; + else if (num_tolerated_disk_barrier_failures > 1) { + if (flags & (BTRFS_BLOCK_GROUP_RAID1 | + BTRFS_BLOCK_GROUP_RAID5 | + BTRFS_BLOCK_GROUP_RAID10)) { + num_tolerated_disk_barrier_failures = 1; + } else if (flags & + BTRFS_BLOCK_GROUP_RAID5) { + num_tolerated_disk_barrier_failures = 2; + } + } } } up_read(&sinfo->groups_sem); @@ -3402,6 +3442,8 @@ int close_ctree(struct btrfs_root *root) btrfs_stop_workers(&fs_info->workers); btrfs_stop_workers(&fs_info->endio_workers); btrfs_stop_workers(&fs_info->endio_meta_workers); + btrfs_stop_workers(&fs_info->endio_raid56_workers); + btrfs_stop_workers(&fs_info->rmw_workers); btrfs_stop_workers(&fs_info->endio_meta_write_workers); btrfs_stop_workers(&fs_info->endio_write_workers); btrfs_stop_workers(&fs_info->endio_freespace_worker); @@ -3424,6 +3466,8 @@ int close_ctree(struct btrfs_root *root) bdi_destroy(&fs_info->bdi); cleanup_srcu_struct(&fs_info->subvol_srcu); + btrfs_free_stripe_hash_table(fs_info); + return 0; } diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h index 305c33efb0e3..034d7dc552b2 100644 --- a/fs/btrfs/disk-io.h +++ b/fs/btrfs/disk-io.h @@ -25,6 +25,13 @@ #define BTRFS_SUPER_MIRROR_MAX 3 #define BTRFS_SUPER_MIRROR_SHIFT 12 +enum { + BTRFS_WQ_ENDIO_DATA = 0, + BTRFS_WQ_ENDIO_METADATA = 1, + BTRFS_WQ_ENDIO_FREE_SPACE = 2, + BTRFS_WQ_ENDIO_RAID56 = 3, +}; + static inline u64 btrfs_sb_offset(int mirror) { u64 start = 16 * 1024; diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index 5cd44e239595..b3ecca447ddf 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -31,6 +31,7 @@ #include "print-tree.h" #include "transaction.h" #include "volumes.h" +#include "raid56.h" #include "locking.h" #include "free-space-cache.h" #include "math.h" @@ -1852,6 +1853,8 @@ static int btrfs_discard_extent(struct btrfs_root *root, u64 bytenr, *actual_bytes = discarded_bytes; + if (ret == -EOPNOTSUPP) + ret = 0; return ret; } @@ -2440,6 +2443,16 @@ int btrfs_delayed_refs_qgroup_accounting(struct btrfs_trans_handle *trans, 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 @@ -2474,6 +2487,44 @@ int btrfs_run_delayed_refs(struct btrfs_trans_handle *trans, 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); @@ -2482,10 +2533,6 @@ again: 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) @@ -2508,9 +2555,12 @@ again: btrfs_release_ref_cluster(&cluster); 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) @@ -2579,6 +2629,11 @@ again: 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; @@ -3284,6 +3339,7 @@ u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags) 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 @@ -3300,30 +3356,32 @@ u64 btrfs_reduce_alloc_profile(struct btrfs_root *root, u64 flags) } 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; - } + 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_RAID1) && - (flags & BTRFS_BLOCK_GROUP_RAID10)) { - flags &= ~BTRFS_BLOCK_GROUP_RAID1; - } - - 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) @@ -3347,6 +3405,7 @@ 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; @@ -3355,7 +3414,8 @@ u64 btrfs_get_alloc_profile(struct btrfs_root *root, int data) else flags = BTRFS_BLOCK_GROUP_METADATA; - return get_alloc_profile(root, flags); + ret = get_alloc_profile(root, flags); + return ret; } /* @@ -3530,8 +3590,10 @@ static u64 get_system_chunk_thresh(struct btrfs_root *root, u64 type) { 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; @@ -3706,7 +3768,9 @@ static int can_overcommit(struct btrfs_root *root, /* * 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 | @@ -5539,10 +5603,14 @@ int btrfs_free_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root, 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; } @@ -5599,8 +5667,12 @@ int __get_raid_index(u64 flags) return BTRFS_RAID_DUP; else if (flags & BTRFS_BLOCK_GROUP_RAID0) return BTRFS_RAID_RAID0; - else - return BTRFS_RAID_SINGLE; + else if (flags & BTRFS_BLOCK_GROUP_RAID5) + return BTRFS_RAID_RAID5; + else if (flags & BTRFS_BLOCK_GROUP_RAID6) + return BTRFS_RAID_RAID6; + + return BTRFS_RAID_SINGLE; /* BTRFS_BLOCK_GROUP_SINGLE */ } static int get_block_group_index(struct btrfs_block_group_cache *cache) @@ -5743,6 +5815,8 @@ search: 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; /* @@ -5771,6 +5845,7 @@ have_block_group: * lets look there */ if (last_ptr) { + unsigned long aligned_cluster; /* * the refill lock keeps out other * people trying to start a new cluster @@ -5837,11 +5912,15 @@ refill_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 @@ -5912,7 +5991,8 @@ unclustered_alloc: 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 > @@ -7284,6 +7364,7 @@ static u64 update_block_group_flags(struct btrfs_root *root, u64 flags) 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) { @@ -7837,7 +7918,9 @@ int btrfs_read_block_groups(struct btrfs_root *root) 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); /* @@ -7891,6 +7974,8 @@ int btrfs_read_block_groups(struct btrfs_root *root) 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; /* @@ -7966,6 +8051,9 @@ int btrfs_make_block_group(struct btrfs_trans_handle *trans, 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); diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 5c00d6aeae75..66f999b97cbb 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -1895,13 +1895,11 @@ static int free_io_failure(struct inode *inode, struct io_failure_record *rec, if (ret) err = ret; - if (did_repair) { - ret = clear_extent_bits(&BTRFS_I(inode)->io_tree, rec->start, - rec->start + rec->len - 1, - EXTENT_DAMAGED, GFP_NOFS); - if (ret && !err) - err = ret; - } + ret = clear_extent_bits(&BTRFS_I(inode)->io_tree, rec->start, + rec->start + rec->len - 1, + EXTENT_DAMAGED, GFP_NOFS); + if (ret && !err) + err = ret; kfree(rec); return err; @@ -1932,10 +1930,15 @@ int repair_io_failure(struct btrfs_fs_info *fs_info, u64 start, u64 map_length = 0; u64 sector; struct btrfs_bio *bbio = NULL; + struct btrfs_mapping_tree *map_tree = &fs_info->mapping_tree; int ret; BUG_ON(!mirror_num); + /* we can't repair anything in raid56 yet */ + if (btrfs_is_parity_mirror(map_tree, logical, length, mirror_num)) + return 0; + bio = bio_alloc(GFP_NOFS, 1); if (!bio) return -EIO; @@ -2052,6 +2055,7 @@ static int clean_io_failure(u64 start, struct page *page) failrec->failed_mirror); did_repair = !ret; } + ret = 0; } out: @@ -2487,13 +2491,13 @@ static int __must_check submit_one_bio(int rw, struct bio *bio, return ret; } -static int merge_bio(struct extent_io_tree *tree, struct page *page, +static int merge_bio(int rw, struct extent_io_tree *tree, struct page *page, unsigned long offset, size_t size, struct bio *bio, unsigned long bio_flags) { int ret = 0; if (tree->ops && tree->ops->merge_bio_hook) - ret = tree->ops->merge_bio_hook(page, offset, size, bio, + ret = tree->ops->merge_bio_hook(rw, page, offset, size, bio, bio_flags); BUG_ON(ret < 0); return ret; @@ -2528,7 +2532,7 @@ static int submit_extent_page(int rw, struct extent_io_tree *tree, sector; if (prev_bio_flags != bio_flags || !contig || - merge_bio(tree, page, offset, page_size, bio, bio_flags) || + merge_bio(rw, tree, page, offset, page_size, bio, bio_flags) || bio_add_page(bio, page, page_size, offset) < page_size) { ret = submit_one_bio(rw, bio, mirror_num, prev_bio_flags); @@ -4162,6 +4166,7 @@ static inline void btrfs_release_extent_buffer(struct extent_buffer *eb) static void check_buffer_tree_ref(struct extent_buffer *eb) { + int refs; /* the ref bit is tricky. We have to make sure it is set * if we have the buffer dirty. Otherwise the * code to free a buffer can end up dropping a dirty @@ -4182,6 +4187,10 @@ static void check_buffer_tree_ref(struct extent_buffer *eb) * So bump the ref count first, then set the bit. If someone * beat us to it, drop the ref we added. */ + refs = atomic_read(&eb->refs); + if (refs >= 2 && test_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags)) + return; + spin_lock(&eb->refs_lock); if (!test_and_set_bit(EXTENT_BUFFER_TREE_REF, &eb->bflags)) atomic_inc(&eb->refs); @@ -4383,9 +4392,20 @@ static int release_extent_buffer(struct extent_buffer *eb, gfp_t mask) void free_extent_buffer(struct extent_buffer *eb) { + int refs; + int old; if (!eb) return; + while (1) { + refs = atomic_read(&eb->refs); + if (refs <= 3) + break; + old = atomic_cmpxchg(&eb->refs, refs, refs - 1); + if (old == refs) + return; + } + spin_lock(&eb->refs_lock); if (atomic_read(&eb->refs) == 2 && test_bit(EXTENT_BUFFER_DUMMY, &eb->bflags)) diff --git a/fs/btrfs/extent_io.h b/fs/btrfs/extent_io.h index ff182322d112..dc81868d975a 100644 --- a/fs/btrfs/extent_io.h +++ b/fs/btrfs/extent_io.h @@ -72,7 +72,7 @@ struct extent_io_ops { int (*writepage_start_hook)(struct page *page, u64 start, u64 end); int (*writepage_io_hook)(struct page *page, u64 start, u64 end); extent_submit_bio_hook_t *submit_bio_hook; - int (*merge_bio_hook)(struct page *page, unsigned long offset, + int (*merge_bio_hook)(int rw, struct page *page, unsigned long offset, size_t size, struct bio *bio, unsigned long bio_flags); int (*readpage_io_failed_hook)(struct page *page, int failed_mirror); diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c index c8090f18c217..1f84fc09c1a8 100644 --- a/fs/btrfs/free-space-cache.c +++ b/fs/btrfs/free-space-cache.c @@ -1465,10 +1465,14 @@ static int search_bitmap(struct btrfs_free_space_ctl *ctl, } 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) @@ -1483,15 +1487,34 @@ find_free_space(struct btrfs_free_space_ctl *ctl, u64 *offset, u64 *bytes) 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 |