diff options
Diffstat (limited to 'fs/btrfs/volumes.c')
| -rw-r--r-- | fs/btrfs/volumes.c | 1582 |
1 files changed, 1314 insertions, 268 deletions
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 0f5ebb72a5ea..5989a92236f7 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -25,6 +25,7 @@ #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" @@ -33,9 +34,12 @@ #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" +#include "math.h" +#include "dev-replace.h" static int init_first_rw_device(struct btrfs_trans_handle *trans, struct btrfs_root *root, @@ -71,6 +75,19 @@ static void free_fs_devices(struct btrfs_fs_devices *fs_devices) kfree(fs_devices); } +static void btrfs_kobject_uevent(struct block_device *bdev, + enum kobject_action action) +{ + int ret; + + ret = kobject_uevent(&disk_to_dev(bdev->bd_disk)->kobj, action); + if (ret) + pr_warn("Sending event '%d' to kobject: '%s' (%p): failed\n", + action, + kobject_name(&disk_to_dev(bdev->bd_disk)->kobj), + &disk_to_dev(bdev->bd_disk)->kobj); +} + void btrfs_cleanup_fs_uuids(void) { struct btrfs_fs_devices *fs_devices; @@ -108,6 +125,44 @@ static noinline struct btrfs_fs_devices *find_fsid(u8 *fsid) return NULL; } +static int +btrfs_get_bdev_and_sb(const char *device_path, fmode_t flags, void *holder, + int flush, struct block_device **bdev, + struct buffer_head **bh) +{ + int ret; + + *bdev = blkdev_get_by_path(device_path, flags, holder); + + if (IS_ERR(*bdev)) { + ret = PTR_ERR(*bdev); + printk(KERN_INFO "btrfs: open %s failed\n", device_path); + goto error; + } + + if (flush) + filemap_write_and_wait((*bdev)->bd_inode->i_mapping); + ret = set_blocksize(*bdev, 4096); + if (ret) { + blkdev_put(*bdev, flags); + goto error; + } + invalidate_bdev(*bdev); + *bh = btrfs_read_dev_super(*bdev); + if (!*bh) { + ret = -EINVAL; + blkdev_put(*bdev, flags); + goto error; + } + + return 0; + +error: + *bdev = NULL; + *bh = NULL; + return ret; +} + static void requeue_list(struct btrfs_pending_bios *pending_bios, struct bio *head, struct bio *tail) { @@ -467,7 +522,8 @@ error: return ERR_PTR(-ENOMEM); } -void btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices) +void btrfs_close_extra_devices(struct btrfs_fs_info *fs_info, + struct btrfs_fs_devices *fs_devices, int step) { struct btrfs_device *device, *next; @@ -480,8 +536,9 @@ again: /* This is the initialized path, it is safe to release the devices. */ list_for_each_entry_safe(device, next, &fs_devices->devices, dev_list) { if (device->in_fs_metadata) { - if (!latest_transid || - device->generation > latest_transid) { + if (!device->is_tgtdev_for_dev_replace && + (!latest_transid || + device->generation > latest_transid)) { latest_devid = device->devid; latest_transid = device->generation; latest_bdev = device->bdev; @@ -489,6 +546,21 @@ again: continue; } + if (device->devid == BTRFS_DEV_REPLACE_DEVID) { + /* + * In the first step, keep the device which has + * the correct fsid and the devid that is used + * for the dev_replace procedure. + * In the second step, the dev_replace state is + * read from the device tree and it is known + * whether the procedure is really active or + * not, which means whether this device is + * used or whether it should be removed. + */ + if (step == 0 || device->is_tgtdev_for_dev_replace) { + continue; + } + } if (device->bdev) { blkdev_put(device->bdev, device->mode); device->bdev = NULL; @@ -497,7 +569,8 @@ again: if (device->writeable) { list_del_init(&device->dev_alloc_list); device->writeable = 0; - fs_devices->rw_devices--; + if (!device->is_tgtdev_for_dev_replace) + fs_devices->rw_devices--; } list_del_init(&device->dev_list); fs_devices->num_devices--; @@ -555,7 +628,7 @@ static int __btrfs_close_devices(struct btrfs_fs_devices *fs_devices) if (device->bdev) fs_devices->open_devices--; - if (device->writeable) { + if (device->writeable && !device->is_tgtdev_for_dev_replace) { list_del_init(&device->dev_alloc_list); fs_devices->rw_devices--; } @@ -577,6 +650,7 @@ static int __btrfs_close_devices(struct btrfs_fs_devices *fs_devices) new_device->writeable = 0; new_device->in_fs_metadata = 0; new_device->can_discard = 0; + spin_lock_init(&new_device->io_lock); list_replace_rcu(&device->dev_list, &new_device->dev_list); call_rcu(&device->rcu, free_device); @@ -610,6 +684,12 @@ int btrfs_close_devices(struct btrfs_fs_devices *fs_devices) __btrfs_close_devices(fs_devices); free_fs_devices(fs_devices); } + /* + * Wait for rcu kworkers under __btrfs_close_devices + * to finish all blkdev_puts so device is really + * free when umount is done. + */ + rcu_barrier(); return ret; } @@ -637,18 +717,10 @@ static int __btrfs_open_devices(struct btrfs_fs_devices *fs_devices, if (!device->name) continue; - bdev = blkdev_get_by_path(device->name->str, flags, holder); - if (IS_ERR(bdev)) { - printk(KERN_INFO "btrfs: open %s failed\n", device->name->str); - goto error; - } - filemap_write_and_wait(bdev->bd_inode->i_mapping); - invalidate_bdev(bdev); - set_blocksize(bdev, 4096); - - bh = btrfs_read_dev_super(bdev); - if (!bh) - goto error_close; + ret = btrfs_get_bdev_and_sb(device->name->str, flags, holder, 1, + &bdev, &bh); + if (ret) + continue; disk_super = (struct btrfs_super_block *)bh->b_data; devid = btrfs_stack_device_id(&disk_super->dev_item); @@ -687,7 +759,7 @@ static int __btrfs_open_devices(struct btrfs_fs_devices *fs_devices, fs_devices->rotating = 1; fs_devices->open_devices++; - if (device->writeable) { + if (device->writeable && !device->is_tgtdev_for_dev_replace) { fs_devices->rw_devices++; list_add(&device->dev_alloc_list, &fs_devices->alloc_list); @@ -697,9 +769,7 @@ static int __btrfs_open_devices(struct btrfs_fs_devices *fs_devices, error_brelse: brelse(bh); -error_close: blkdev_put(bdev, flags); -error: continue; } if (fs_devices->open_devices == 0) { @@ -732,18 +802,35 @@ int btrfs_open_devices(struct btrfs_fs_devices *fs_devices, return ret; } +/* + * Look for a btrfs signature on a device. This may be called out of the mount path + * and we are not allowed to call set_blocksize during the scan. The superblock + * is read via pagecache + */ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder, struct btrfs_fs_devices **fs_devices_ret) { struct btrfs_super_block *disk_super; struct block_device *bdev; - struct buffer_head *bh; - int ret; + struct page *page; + void *p; + int ret = -EINVAL; u64 devid; u64 transid; u64 total_devices; + u64 bytenr; + pgoff_t index; + /* + * we would like to check all the supers, but that would make + * a btrfs mount succeed after a mkfs from a different FS. + * So, we need to add a special mount option to scan for + * later supers, using BTRFS_SUPER_MIRROR_MAX instead + */ + bytenr = btrfs_sb_offset(0); flags |= FMODE_EXCL; + mutex_lock(&uuid_mutex); + bdev = blkdev_get_by_path(path, flags, holder); if (IS_ERR(bdev)) { @@ -751,33 +838,62 @@ int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder, goto error; } - mutex_lock(&uuid_mutex); - ret = set_blocksize(bdev, 4096); - if (ret) - goto error_close; - bh = btrfs_read_dev_super(bdev); - if (!bh) { - ret = -EINVAL; - goto error_close; - } - disk_super = (struct btrfs_super_block *)bh->b_data; + /* make sure our super fits in the device */ + if (bytenr + PAGE_CACHE_SIZE >= i_size_read(bdev->bd_inode)) + goto error_bdev_put; + + /* make sure our super fits in the page */ + if (sizeof(*disk_super) > PAGE_CACHE_SIZE) + goto error_bdev_put; + + /* make sure our super doesn't straddle pages on disk */ + index = bytenr >> PAGE_CACHE_SHIFT; + if ((bytenr + sizeof(*disk_super) - 1) >> PAGE_CACHE_SHIFT != index) + goto error_bdev_put; + + /* pull in the page with our super */ + page = read_cache_page_gfp(bdev->bd_inode->i_mapping, + index, GFP_NOFS); + + if (IS_ERR_OR_NULL(page)) + goto error_bdev_put; + + p = kmap(page); + + /* align our pointer to the offset of the super block */ + disk_super = p + (bytenr & ~PAGE_CACHE_MASK); + + if (btrfs_super_bytenr(disk_super) != bytenr || + disk_super->magic != cpu_to_le64(BTRFS_MAGIC)) + goto error_unmap; + devid = btrfs_stack_device_id(&disk_super->dev_item); transid = btrfs_super_generation(disk_super); total_devices = btrfs_super_num_devices(disk_super); - if (disk_super->label[0]) + + if (disk_super->label[0]) { + if (disk_super->label[BTRFS_LABEL_SIZE - 1]) + disk_super->label[BTRFS_LABEL_SIZE - 1] = '\0'; printk(KERN_INFO "device label %s ", disk_super->label); - else + } else { printk(KERN_INFO "device fsid %pU ", disk_super->fsid); + } + printk(KERN_CONT "devid %llu transid %llu %s\n", (unsigned long long)devid, (unsigned long long)transid, path); + ret = device_list_add(path, disk_super, devid, fs_devices_ret); if (!ret && fs_devices_ret) (*fs_devices_ret)->total_devices = total_devices; - brelse(bh); -error_close: - mutex_unlock(&uuid_mutex); + +error_unmap: + kunmap(page); + page_cache_release(page); + +error_bdev_put: blkdev_put(bdev, flags); error: + mutex_unlock(&uuid_mutex); return ret; } @@ -796,7 +912,7 @@ int btrfs_account_dev_extents_size(struct btrfs_device *device, u64 start, *length = 0; - if (start >= device->total_bytes) + if (start >= device->total_bytes || device->is_tgtdev_for_dev_replace) return 0; path = btrfs_alloc_path(); @@ -913,7 +1029,7 @@ int find_free_dev_extent(struct btrfs_device *device, u64 num_bytes, max_hole_size = 0; hole_size = 0; - if (search_start >= search_end) { + if (search_start >= search_end || device->is_tgtdev_for_dev_replace) { ret = -ENOSPC; goto error; } @@ -1096,6 +1212,7 @@ int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans, struct btrfs_key key; WARN_ON(!device->in_fs_metadata); + WARN_ON(device->is_tgtdev_for_dev_replace); path = btrfs_alloc_path(); if (!path) return -ENOMEM; @@ -1321,31 +1438,57 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) u64 devid; u64 num_devices; u8 *dev_uuid; + unsigned seq; int ret = 0; bool clear_super = false; mutex_lock(&uuid_mutex); - all_avail = root->fs_info->avail_data_alloc_bits | - root->fs_info->avail_system_alloc_bits | - root->fs_info->avail_metadata_alloc_bits; + do { + seq = read_seqbegin(&root->fs_info->profiles_lock); + + all_avail = root->fs_info->avail_data_alloc_bits | + root->fs_info->avail_system_alloc_bits | + root->fs_info->avail_metadata_alloc_bits; + } while (read_seqretry(&root->fs_info->profiles_lock, seq)); - if ((all_avail & BTRFS_BLOCK_GROUP_RAID10) && - root->fs_info->fs_devices->num_devices <= 4) { + num_devices = root->fs_info->fs_devices->num_devices; + btrfs_dev_replace_lock(&root->fs_info->dev_replace); + if (btrfs_dev_replace_is_ongoing(&root->fs_info->dev_replace)) { + WARN_ON(num_devices < 1); + num_devices--; + } + btrfs_dev_replace_unlock(&root->fs_info->dev_replace); + + if ((all_avail & BTRFS_BLOCK_GROUP_RAID10) && num_devices <= 4) { printk(KERN_ERR "btrfs: unable to go below four devices " "on raid10\n"); ret = -EINVAL; goto out; } - if ((all_avail & BTRFS_BLOCK_GROUP_RAID1) && - root->fs_info->fs_devices->num_devices <= 2) { + if ((all_avail & BTRFS_BLOCK_GROUP_RAID1) && num_devices <= 2) { printk(KERN_ERR "btrfs: unable to go below two " "devices on raid1\n"); ret = -EINVAL; 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; @@ -1357,7 +1500,9 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) * is held. */ list_for_each_entry(tmp, devices, dev_list) { - if (tmp->in_fs_metadata && !tmp->bdev) { + if (tmp->in_fs_metadata && + !tmp->is_tgtdev_for_dev_replace && + !tmp->bdev) { device = tmp; break; } @@ -1371,24 +1516,16 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) goto out; } } else { - bdev = blkdev_get_by_path(device_path, FMODE_READ | FMODE_EXCL, - root->fs_info->bdev_holder); - if (IS_ERR(bdev)) { - ret = PTR_ERR(bdev); + ret = btrfs_get_bdev_and_sb(device_path, + FMODE_WRITE | FMODE_EXCL, + root->fs_info->bdev_holder, 0, + &bdev, &bh); + if (ret) goto out; - } - - set_blocksize(bdev, 4096); - invalidate_bdev(bdev); - bh = btrfs_read_dev_super(bdev); - if (!bh) { - ret = -EINVAL; - goto error_close; - } disk_super = (struct btrfs_super_block *)bh->b_data; devid = btrfs_stack_device_id(&disk_super->dev_item); dev_uuid = disk_super->dev_item.uuid; - device = btrfs_find_device(root, devid, dev_uuid, + device = btrfs_find_device(root->fs_info, devid, dev_uuid, disk_super->fsid); if (!device) { ret = -ENOENT; @@ -1396,6 +1533,12 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) } } + if (device->is_tgtdev_for_dev_replace) { + pr_err("btrfs: unable to remove the dev_replace target dev\n"); + ret = -EINVAL; + goto error_brelse; + } + if (device->writeable && root->fs_info->fs_devices->rw_devices == 1) { printk(KERN_ERR "btrfs: unable to remove the only writeable " "device\n"); @@ -1415,6 +1558,11 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) if (ret) goto error_undo; + /* + * TODO: the superblock still includes this device in its num_devices + * counter although write_all_supers() is not locked out. This + * could give a filesystem state which requires a degraded mount. + */ ret = btrfs_rm_dev_item(root->fs_info->chunk_root, device); if (ret) goto error_undo; @@ -1425,7 +1573,7 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) spin_unlock(&root->fs_info->free_chunk_lock); device->in_fs_metadata = 0; - btrfs_scrub_cancel_dev(root, device); + btrfs_scrub_cancel_dev(root->fs_info, device); /* * the device list mutex makes sure that we don't change @@ -1482,7 +1630,7 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) * at this point, the device is zero sized. We want to * remove it from the devices list and zero out the old super */ - if (clear_super) { + if (clear_super && disk_super) { /* make sure this device isn't detected as part of * the FS anymore */ @@ -1493,9 +1641,12 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) ret = 0; + /* Notify udev that device has changed */ + if (bdev) + btrfs_kobject_uevent(bdev, KOBJ_CHANGE); + error_brelse: brelse(bh); -error_close: if (bdev) blkdev_put(bdev, FMODE_READ | FMODE_EXCL); out: @@ -1512,6 +1663,112 @@ error_undo: goto error_brelse; } +void btrfs_rm_dev_replace_srcdev(struct btrfs_fs_info *fs_info, + struct btrfs_device *srcdev) +{ + WARN_ON(!mutex_is_locked(&fs_info->fs_devices->device_list_mutex)); + list_del_rcu(&srcdev->dev_list); + list_del_rcu(&srcdev->dev_alloc_list); + fs_info->fs_devices->num_devices--; + if (srcdev->missing) { + fs_info->fs_devices->missing_devices--; + fs_info->fs_devices->rw_devices++; + } + if (srcdev->can_discard) + fs_info->fs_devices->num_can_discard--; + if (srcdev->bdev) + fs_info->fs_devices->open_devices--; + + call_rcu(&srcdev->rcu, free_device); +} + +void btrfs_destroy_dev_replace_tgtdev(struct btrfs_fs_info *fs_info, + struct btrfs_device *tgtdev) +{ + struct btrfs_device *next_device; + + WARN_ON(!tgtdev); + mutex_lock(&fs_info->fs_devices->device_list_mutex); + if (tgtdev->bdev) { + btrfs_scratch_superblock(tgtdev); + fs_info->fs_devices->open_devices--; + } + fs_info->fs_devices->num_devices--; + if (tgtdev->can_discard) + fs_info->fs_devices->num_can_discard++; + + next_device = list_entry(fs_info->fs_devices->devices.next, + struct btrfs_device, dev_list); + if (tgtdev->bdev == fs_info->sb->s_bdev) + fs_info->sb->s_bdev = next_device->bdev; + if (tgtdev->bdev == fs_info->fs_devices->latest_bdev) + fs_info->fs_devices->latest_bdev = next_device->bdev; + list_del_rcu(&tgtdev->dev_list); + + call_rcu(&tgtdev->rcu, free_device); + + mutex_unlock(&fs_info->fs_devices->device_list_mutex); +} + +int btrfs_find_device_by_path(struct btrfs_root *root, char *device_path, + struct btrfs_device **device) +{ + int ret = 0; + struct btrfs_super_block *disk_super; + u64 devid; + u8 *dev_uuid; + struct block_device *bdev; + struct buffer_head *bh; + + *device = NULL; + ret = btrfs_get_bdev_and_sb(device_path, FMODE_READ, + root->fs_info->bdev_holder, 0, &bdev, &bh); + if (ret) + return ret; + disk_super = (struct btrfs_super_block *)bh->b_data; + devid = btrfs_stack_device_id(&disk_super->dev_item); + dev_uuid = disk_super->dev_item.uuid; + *device = btrfs_find_device(root->fs_info, devid, dev_uuid, + disk_super->fsid); + brelse(bh); + if (!*device) + ret = -ENOENT; + blkdev_put(bdev, FMODE_READ); + return ret; +} + +int btrfs_find_device_missing_or_by_path(struct btrfs_root *root, + char *device_path, + struct btrfs_device **device) +{ + *device = NULL; + if (strcmp(device_path, "missing") == 0) { + struct list_head *devices; + struct btrfs_device *tmp; + + devices = &root->fs_info->fs_devices->devices; + /* + * It is safe to read the devices since the volume_mutex + * is held by the caller. + */ + list_for_each_entry(tmp, devices, dev_list) { + if (tmp->in_fs_metadata && !tmp->bdev) { + *device = tmp; + break; + } + } + + if (!*device) { + pr_err("btrfs: no missing device found\n"); + return -ENOENT; + } + + return 0; + } else { + return btrfs_find_device_by_path(root, device_path, device); + } +} + /* * does all the dirty work required for changing file system's UUID. */ @@ -1630,7 +1887,8 @@ next_slot: read_extent_buffer(leaf, fs_uuid, (unsigned long)btrfs_device_fsid(dev_item), BTRFS_UUID_SIZE); - device = btrfs_find_device(root, devid, dev_uuid, fs_uuid); + device = btrfs_find_device(root->fs_info, devid, dev_uuid, + fs_uuid); BUG_ON(!device); /* Logic error */ if (device->fs_devices->seeding) { @@ -1678,16 +1936,17 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path) filemap_write_and_wait(bdev->bd_inode->i_mapping); devices = &root->fs_info->fs_devices->devices; - /* - * we have the volume lock, so we don't need the extra - * device list mutex while reading the list here. - */ + + mutex_lock(&root->fs_info->fs_devices->device_list_mutex); list_for_each_entry(device, devices, dev_list) { if (device->bdev == bdev) { ret = -EEXIST; + mutex_unlock( + &root->fs_info->fs_devices->device_list_mutex); goto error; } } + mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); device = kzalloc(sizeof(*device), GFP_NOFS); if (!device) { @@ -1737,6 +1996,7 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path) device->dev_root = root->fs_info->dev_root; device->bdev = bdev; device->in_fs_metadata = 1; + device->is_tgtdev_for_dev_replace = 0; device->mode = FMODE_EXCL; set_blocksize(device->bdev, 4096); @@ -1844,6 +2104,98 @@ error: return ret; } +int btrfs_init_dev_replace_tgtdev(struct btrfs_root *root, char *device_path, + struct btrfs_device **device_out) +{ + struct request_queue *q; + struct btrfs_device *device; + struct block_device *bdev; + struct btrfs_fs_info *fs_info = root->fs_info; + struct list_head *devices; + struct rcu_string *name; + int ret = 0; + + *device_out = NULL; + if (fs_info->fs_devices->seeding) + return -EINVAL; + + bdev = blkdev_get_by_path(device_path, FMODE_WRITE | FMODE_EXCL, + fs_info->bdev_holder); + if (IS_ERR(bdev)) + return PTR_ERR(bdev); + + filemap_write_and_wait(bdev->bd_inode->i_mapping); + + devices = &fs_info->fs_devices->devices; + list_for_each_entry(device, devices, dev_list) { + if (device->bdev == bdev) { + ret = -EEXIST; + goto error; + } + } + + device = kzalloc(sizeof(*device), GFP_NOFS); + if (!device) { + ret = -ENOMEM; + goto error; + } + + name = rcu_string_strdup(device_path, GFP_NOFS); + if (!name) { + kfree(device); + ret = -ENOMEM; + goto error; + } + rcu_assign_pointer(device->name, name); + + q = bdev_get_queue(bdev); + if (blk_queue_discard(q)) + device->can_discard = 1; + mutex_lock(&root->fs_info->fs_devices->device_list_mutex); + device->writeable = 1; + device->work.func = pending_bios_fn; + generate_random_uuid(device->uuid); + device->devid = BTRFS_DEV_REPLACE_DEVID; + spin_lock_init(&device->io_lock); + device->generation = 0; + device->io_width = root->sectorsize; + device->io_align = root->sectorsize; + device->sector_size = root->sectorsize; + device->total_bytes = i_size_read(bdev->bd_inode); + device->disk_total_bytes = device->total_bytes; + device->dev_root = fs_info->dev_root; + device->bdev = bdev; + device->in_fs_metadata = 1; + device->is_tgtdev_for_dev_replace = 1; + device->mode = FMODE_EXCL; + set_blocksize(device->bdev, 4096); + device->fs_devices = fs_info->fs_devices; + list_add(&device->dev_list, &fs_info->fs_devices->devices); + fs_info->fs_devices->num_devices++; + fs_info->fs_devices->open_devices++; + if (device->can_discard) + fs_info->fs_devices->num_can_discard++; + mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); + + *device_out = device; + return ret; + +error: + blkdev_put(bdev, FMODE_EXCL); + return ret; +} + +void btrfs_init_dev_replace_tgtdev_for_resume(struct btrfs_fs_info *fs_info, + struct btrfs_device *tgtdev) +{ + WARN_ON(fs_info->fs_devices->rw_devices == 0); + tgtdev->io_width = fs_info->dev_root->sectorsize; + tgtdev->io_align = fs_info->dev_root->sectorsize; + tgtdev->sector_size = fs_info->dev_root->sectorsize; + tgtdev->dev_root = fs_info->dev_root; + tgtdev->in_fs_metadata = 1; +} + static noinline int btrfs_update_device(struct btrfs_trans_handle *trans, struct btrfs_device *device) { @@ -1900,7 +2252,8 @@ static int __btrfs_grow_device(struct btrfs_trans_handle *trans, if (!device->writeable) return -EACCES; - if (new_size <= device->total_bytes) + if (new_size <= device->total_bytes || + device->is_tgtdev_for_dev_replace) return -EINVAL; btrfs_set_super_total_bytes(super_copy, old_total + diff); @@ -2032,7 +2385,11 @@ static int btrfs_relocate_chunk(struct btrfs_root *root, return ret; trans = btrfs_start_transaction(root, 0); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + btrfs_std_error(root->fs_info, ret); + return ret; + } lock_chunks(root); @@ -2338,18 +2695,6 @@ static int chunk_profiles_filter(u64 chunk_type, return 1; } -static u64 div_factor_fine(u64 num, int factor) -{ - if (factor <= 0) - return 0; - if (factor >= 100) - return num; - - num *= factor; - do_div(num, 100); - return num; -} - static int chunk_usage_filter(struct btrfs_fs_info *fs_info, u64 chunk_offset, struct btrfs_balance_args *bargs) { @@ -2360,7 +2705,14 @@ static int chunk_usage_filter(struct btrfs_fs_info *fs_info, u64 chunk_offset, 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 = 1; + 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; @@ -2402,11 +2754,15 @@ static int chunk_drange_filter(struct extent_buffer *leaf, 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); @@ -2514,15 +2870,6 @@ static int should_balance_chunk(struct btrfs_root *root, return 1; } -static u64 div_factor(u64 num, int factor) -{ - if (factor == 10) - return num; - num *= factor; - do_div(num, 10); - return num; -} - static int __btrfs_balance(struct btrfs_fs_info *fs_info) { struct btrfs_balance_control *bctl = fs_info->balance_ctl; @@ -2550,7 +2897,8 @@ static int __btrfs_balance(struct btrfs_fs_info *fs_info) size_to_free = div_factor(old_size, 1); size_to_free = min(size_to_free, (u64)1 * 1024 * 1024); if (!device->writeable || - device->total_bytes - device->bytes_used > size_to_free) + device->total_bytes - device->bytes_used > size_to_free || + device->is_tgtdev_for_dev_replace) continue; ret = btrfs_shrink_device(device, old_size - size_to_free); @@ -2712,7 +3060,10 @@ static void __cancel_balance(struct btrfs_fs_info *fs_info) unset_balance_control(fs_info); ret = del_balance_item(fs_info->tree_root); - BUG_ON(ret); + if (ret) + btrfs_std_error(fs_info, ret); + + atomic_set(&fs_info->mutually_exclusive_operation_running, 0); } void update_ioctl_balance_args(struct btrfs_fs_info *fs_info, int lock, @@ -2728,6 +3079,8 @@ int btrfs_balance(struct btrfs_balance_control *bctl, u64 allowed; int mixed = 0; int ret; + u64 num_devices; + unsigned seq; if (btrfs_fs_closing(fs_info) || atomic_read(&fs_info->balance_pause_req) || @@ -2756,14 +3109,23 @@ int btrfs_balance(struct btrfs_balance_control *bctl, } } + num_devices = fs_info->fs_devices->num_devices; + btrfs_dev_replace_lock(&fs_info->dev_replace); + if (btrfs_dev_replace_is_ongoing(&fs_info->dev_replace)) { + BUG_ON(num_devices < 1); + num_devices--; + } + btrfs_dev_replace_unlock(&fs_info->dev_replace); allowed = BTRFS_AVAIL_ALLOC_BIT_SINGLE; - if (fs_info->fs_devices->num_devices == 1) + if (num_devices == 1) allowed |= BTRFS_BLOCK_GROUP_DUP; - else if (fs_info->fs_devices->num_devices < 4) + else if (num_devices < 4) 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) || @@ -2803,23 +3165,29 @@ int btrfs_balance(struct btrfs_balance_control *bctl, /* allow to reduce meta or sys integrity only if force set */ allowed = BTRFS_BLOCK_GROUP_DUP | BTRFS_BLOCK_GROUP_RAID1 | - BTRFS_BLOCK_GROUP_RAID10; - if (((bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT) && - (fs_info->avail_system_alloc_bits & allowed) && - !(bctl->sys.target & allowed)) || - ((bctl->meta.flags & BTRFS_BALANCE_ARGS_CONVERT) && - (fs_info->avail_metadata_alloc_bits & allowed) && - !(bctl->meta.target & allowed))) { - if (bctl->flags & BTRFS_BALANCE_FORCE) { - printk(KERN_INFO "btrfs: force reducing metadata " - "integrity\n"); - } else { - printk(KERN_ERR "btrfs: balance will reduce metadata " - "integrity, use force if you want this\n"); - ret = -EINVAL; - goto out; + BTRFS_BLOCK_GROUP_RAID10 | + BTRFS_BLOCK_GROUP_RAID5 | + BTRFS_BLOCK_GROUP_RAID6; + do { + seq = read_seqbegin(&fs_info->profiles_lock); + + if (((bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT) && + (fs_info->avail_system_alloc_bits & allowed) && + !(bctl->sys.target & allowed)) || + ((bctl->meta.flags & BTRFS_BALANCE_ARGS_CONVERT) && + (fs_info->avail_metadata_alloc_bits & allowed) && + !(bctl->meta.target & allowed))) { + if (bctl->flags & BTRFS_BALANCE_FORCE) { + printk(KERN_INFO "btrfs: force reducing metadata " + "integrity\n"); + } else { + printk(KERN_ERR "btrfs: balance will reduce metadata " + "integrity, use force if you want this\n"); + ret = -EINVAL; + goto out; + } } - } + } while (read_seqretry(&fs_info->profiles_lock, seq)); if (bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT) { int num_tolerated_disk_barrier_failures; @@ -2863,6 +3231,11 @@ int btrfs_balance(struct btrfs_balance_control *bctl, mutex_lock(&fs_info->balance_mutex); atomic_dec(&fs_info->balance_running); + 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 (bargs) { memset(bargs, 0, sizeof(*bargs)); update_ioctl_balance_args(fs_info, 0, bargs); @@ -2873,19 +3246,16 @@ int btrfs_balance(struct btrfs_balance_control *bctl, __cancel_balance(fs_info); } - if (bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT) { - fs_info->num_tolerated_disk_barrier_failures = - btrfs_calc_num_tolerated_disk_barrier_failures(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; } @@ -2977,6 +3347,8 @@ int btrfs_recover_balance(struct btrfs_fs_info *fs_info) 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); @@ -3080,7 +3452,7 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size) u64 old_size = device->total_bytes; u64 diff = device->total_bytes - new_size; - if (new_size >= device->total_bytes) + if (device->is_tgtdev_for_dev_replace) return -EINVAL; path = btrfs_alloc_path(); @@ -3235,6 +3607,87 @@ static int btrfs_cmp_device_info(const void *a, const void *b) return 0; } +struct btrfs_raid_attr btrfs_raid_array[BTRFS_NR_RAID_TYPES] = { + [BTRFS_RAID_RAID10] = { + .sub_stripes = 2, + .dev_stripes = 1, + .devs_max = 0, /* 0 == as many as possible */ + .devs_min = 4, + .devs_increment = 2, + .ncopies = 2, + }, + [BTRFS_RAID_RAID1] = { + .sub_stripes = 1, + .dev_stripes = 1, + .devs_max = 2, + .devs_min = 2, + .devs_increment = 2, + .ncopies = 2, + }, + [BTRFS_RAID_DUP] = { + .sub_stripes = 1, + .dev_stripes = 2, + .devs_max = 1, + .devs_min = 1, + .devs_increment = 1, + .ncopies = 2, + }, + [BTRFS_RAID_RAID0] = { + .sub_stripes = 1, + .dev_stripes = 1, + .devs_max = 0, + .devs_min = 2, + .devs_increment = 1, + .ncopies = 1, + }, + [BTRFS_RAID_SINGLE] = { + .sub_stripes = 1, + .dev_stripes = 1, + .devs_max = 1, + .devs_min = 1, + .devs_increment = 1, + .ncopies = 1, + }, + [BTRFS_RAID_RAID5] = { + .sub_stripes = 1, + .dev_stripes = 1, + .devs_max = 0, + .devs_min = 2, + .devs_increment = 1, + .ncopies = 2, + }, + [BTRFS_RAID_RAID6] = { + .sub_stripes = 1, + .dev_stripes = 1, + .devs_max = 0, + .devs_min = 3, + .devs_increment = 1, + .ncopies = 3, + }, +}; + +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, @@ -3250,6 +3703,8 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, 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 */ @@ -3261,46 +3716,25 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, u64 max_chunk_size; u64 stripe_size; u64 num_bytes; + u64 raid_stripe_len = BTRFS_STRIPE_LEN; int ndevs; int i; int j; + int index; BUG_ON(!alloc_profile_is_valid(type, 0)); if (list_empty(&fs_devices->alloc_list)) return -ENOSPC; - sub_stripes = 1; - dev_stripes = 1; - devs_increment = 1; - ncopies = 1; - devs_max = 0; /* 0 == as many as possible */ - devs_min = 1; + index = __get_raid_index(type); - /* - * define the properties of each RAID type. - * FIXME: move this to a global table and use it in all RAID - * calculation code - */ - if (type & (BTRFS_BLOCK_GROUP_DUP)) { - dev_stripes = 2; - ncopies = 2; - devs_max = 1; - } else if (type & (BTRFS_BLOCK_GROUP_RAID0)) { - devs_min = 2; - } else if (type & (BTRFS_BLOCK_GROUP_RAID1)) { - devs_increment = 2; - ncopies = 2; - devs_max = 2; - devs_min = 2; - } else if (type & (BTRFS_BLOCK_GROUP_RAID10)) { - sub_stripes = 2; - devs_increment = 2; - ncopies = 2; - devs_min = 4; - } else { - devs_max = 1; - } + sub_stripes = btrfs_raid_array[index].sub_stripes; + dev_stripes = btrfs_raid_array[index].dev_stripes; + devs_max = btrfs_raid_array[index].devs_max; + devs_min = btrfs_raid_array[index].devs_min; + devs_increment = btrfs_raid_array[index].devs_increment; + ncopies = btrfs_raid_array[index].ncopies; if (type & BTRFS_BLOCK_GROUP_DATA) { max_stripe_size = 1024 * 1024 * 1024; @@ -3347,13 +3781,13 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, cur = cur->next; if (!device->writeable) { - printk(KERN_ERR + WARN(1, KERN_ERR "btrfs: read-only device in alloc_list\n"); - WARN_ON(1); continue; } - if (!device->in_fs_metadata) + if (!device->in_fs_metadata || + device->is_tgtdev_for_dev_replace) continue; if (device->total_bytes > device->bytes_used) @@ -3377,6 +3811,11 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, if (max_avail < BTRFS_STRIPE_LEN * dev_stripes) continue; + if (ndevs == fs_devices->rw_devices) { + WARN(1, "%s: found more than %llu devices\n", + __func__, fs_devices->rw_devices); + break; + } devices_info[ndevs].dev_offset = dev_offset; devices_info[ndevs].max_avail = max_avail; devices_info[ndevs].total_avail = total_avail; @@ -3407,16 +3846,48 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, stripe_size = devices_info[ndevs-1].max_avail; num_stripes = ndevs * dev_stripes; - if (stripe_size * ndevs > max_chunk_size * ncopies) { - stripe_size = max_chunk_size * ncopies; - do_div(stripe_size, ndevs); + /* + * this will have to be fixed for RAID1 and RAID10 over + * more drives + */ + data_stripes = num_stripes / ncopies; + + 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; + } + + /* + * Use the number of data stripes to figure out how big this chunk + * is really going to be in terms of logical address space, + * and compare that answer with the max chunk size + */ + if (stripe_size * data_stripes > max_chunk_size) { + u64 mask = (1ULL << 24) - 1; + stripe_size = max_chunk_size; + do_div(stripe_size, data_stripes); + + /* bump the answer up to a 16MB boundary */ + stripe_size = (stripe_size + mask) & ~mask; + + /* but don't go higher than the limits we found + * while searching for free extents + */ + if (stripe_size > devices_info[ndevs-1].max_avail) + stripe_size = devices_info[ndevs-1].max_avail; } 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) { @@ -3434,14 +3905,14 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, } } 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; @@ -3463,15 +3934,10 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, write_lock(&em_tree->lock); ret = add_extent_mapping(em_tree, em); write_unlock(&em_tree->lock); - free_extent_map(em); - if (ret) - goto error; - - ret = btrfs_make_block_group(trans, extent_root, 0, type, - BTRFS_FIRST_CHUNK_TREE_OBJECTID, - start, num_bytes); - if (ret) + if (ret) { + free_extent_map(em); goto error; + } for (i = 0; i < map->num_stripes; ++i) { struct btrfs_device *device; @@ -3484,15 +3950,44 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, info->chunk_root->root_key.objectid, BTRFS_FIRST_CHUNK_TREE_OBJECTID, start, dev_offset, stripe_size); - if (ret) { - btrfs_abort_transaction(trans, extent_root, ret); - goto error; - } + if (ret) + goto error_dev_extent; } + ret = btrfs_make_block_group(trans, extent_root, 0, type, + BTRFS_FIRST_CHUNK_TREE_OBJECTID, + start, num_bytes); + if (ret) { + i = map->num_stripes - 1; + goto error_dev_extent; + } + + free_extent_map(em); + check_raid56_incompat_flag(extent_root->fs_info, type); + kfree(devices_info); return 0; +error_dev_extent: + for (; i >= 0; i--) { + struct btrfs_device *device; + int err; + + device = map->stripes[i].dev; + err = btrfs_free_dev_extent(trans, device, start); + if (err) { + btrfs_abort_transaction(trans, extent_root, err); + break; + } + } + write_lock(&em_tree->lock); + remove_extent_mapping(em_tree, em); + write_unlock(&em_tree->lock); + + /* One for our allocation */ + free_extent_map(em); + /* One for the tree reference */ + free_extent_map(em); error: kfree(map); kfree(devices_info); @@ -3632,10 +4127,7 @@ static noinline int init_first_rw_device(struct btrfs_trans_handle *trans, if (ret) return ret; - alloc_profile = BTRFS_BLOCK_GROUP_METADATA | - fs_info->avail_metadata_alloc_bits; - alloc_profile = btrfs_reduce_alloc_profile(root, alloc_profile); - + alloc_profile = btrfs_get_alloc_profile(extent_root, 0); ret = __btrfs_alloc_chunk(trans, extent_root, &map, &chunk_size, &stripe_size, chunk_offset, alloc_profile); if (ret) @@ -3643,10 +4135,7 @@ static noinline int init_first_rw_device(struct btrfs_trans_handle *trans, sys_chunk_offset = chunk_offset + chunk_size; - alloc_profile = BTRFS_BLOCK_GROUP_SYSTEM | - fs_info->avail_system_alloc_bits; - alloc_profile = btrfs_reduce_alloc_profile(root, alloc_profile); - + alloc_profile = btrfs_get_alloc_profile(fs_info->chunk_root, 0); ret = __btrfs_alloc_chunk(trans, extent_root, &sys_map, &sys_chunk_size, &sys_stripe_size, sys_chunk_offset, alloc_profile); @@ -3740,8 +4229,9 @@ void btrfs_mapping_tree_free(struct btrfs_mapping_tree *tree) } } -int btrfs_num_copies(struct btrfs_mapping_tree *map_tree, u64 logical, u64 len) +int btrfs_num_copies(struct btrfs_fs_info *fs_info, u64 logical, u64 len) { + struct btrfs_mapping_tree *map_tree = &fs_info->mapping_tree; struct extent_map *em; struct map_lookup *map; struct extent_map_tree *em_tree = &map_tree->map_tree; @@ -3758,35 +4248,142 @@ int btrfs_num_copies(struct btrfs_mapping_tree *map_tree, u64 logical, u64 len) 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); + + btrfs_dev_replace_lock(&fs_info->dev_replace); + if (btrfs_dev_replace_is_ongoing(&fs_info->dev_replace)) + ret++; + btrfs_dev_replace_unlock(&fs_info->dev_replace); + + 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 map_lookup *map, int first, int num, - int optimal) +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) { int i; - if (map->stripes[optimal].dev->bdev) - return optimal; - for (i = first; i < first + num; i++) { - if (map->stripes[i].dev->bdev) - return i; + int tolerance; + struct btrfs_device *srcdev; + + if (dev_replace_is_ongoing && + fs_info->dev_replace.cont_reading_from_srcdev_mode == + BTRFS_DEV_REPLACE_ITEM_CONT_READING_FROM_SRCDEV_MODE_AVOID) + srcdev = fs_info->dev_replace.srcdev; + else + srcdev = NULL; + + /* + * try to avoid the drive that is the source drive for a + * dev-replace procedure, only choose it if no other non-missing + * mirror is available + */ + for (tolerance = 0; tolerance < 2; tolerance++) { + if (map->stripes[optimal].dev->bdev && + (tolerance || map->stripes[optimal].dev != srcdev)) + return optimal; + for (i = first; i < first + num; i++) { + if (map->stripes[i].dev->bdev && + (tolerance || map->stripes[i].dev != srcdev)) + return i; + } } + /* we couldn't find one that doesn't fail. Just return something * and the io error handling code will clean up eventually */ return optimal; } -static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, +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; + struct btrfs_mapping_tree *map_tree = &fs_info->mapping_tree; struct extent_map_tree *em_tree = &map_tree->map_tree; u64 offset; u64 stripe_offset; @@ -3794,12 +4391,20 @@ static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, 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_stripes; int max_errors = 0; struct btrfs_bio *bbio = NULL; + struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace; + int dev_replace_is_ongoing = 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); @@ -3819,59 +4424,178 @@ static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, 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; + btrfs_dev_replace_lock(dev_replace); + dev_replace_is_ongoing = btrfs_dev_replace_is_ongoing(dev_replace); + if (!dev_replace_is_ongoing) + btrfs_dev_replace_unlock(dev_replace); + + if (dev_replace_is_ongoing && mirror_num == map->num_stripes + 1 && + !(rw & (REQ_WRITE | REQ_DISCARD | REQ_GET_READ_MIRRORS)) && + dev_replace->tgtdev != NULL) { + /* + * in dev-replace case, for repair case (that's the only + * case where the mirror is selected explicitly when + * calling btrfs_map_block), blocks left of the left cursor + * can also be read from the target drive. + * For REQ_GET_READ_MIRRORS, the target drive is added as + * the last one to the array of stripes. For READ, it also + * needs to be supported using the same mirror number. + * If the requested block is not left of the left cursor, + * EIO is returned. This can happen because btrfs_num_copies() + * returns one more in the dev-replace case. + */ + u64 tmp_length = *length; + struct btrfs_bio *tmp_bbio = NULL; + int tmp_num_stripes; + u64 srcdev_devid = dev_replace->srcdev->devid; + int index_srcdev = 0; + int found = 0; + u64 physical_of_found = 0; + + ret = __btrfs_map_block(fs_info, REQ_GET_READ_MIRRORS, + logical, &tmp_length, &tmp_bbio, 0, NULL); + if (ret) { + WARN_ON(tmp_bbio != NULL); + goto out; + } + + tmp_num_stripes = tmp_bbio->num_stripes; + if (mirror_num > tmp_num_stripes) { + /* + * REQ_GET_READ_MIRRORS does not contain this + * mirror, that means that the requested area + * is not left of the left cursor + */ + ret = -EIO; + kfree(tmp_bbio); + goto out; + } + + /* + * process the rest of the function using the mirror_num + * of the source drive. Therefore look it up first. + * At the end, patch the device pointer to the one of the + * target drive. + */ + for (i = 0; i < tmp_num_stripes; i++) { + if (tmp_bbio->stripes[i].dev->devid == srcdev_devid) { + /* + * In case of DUP, in order to keep it + * simple, only add the mirror with the + * lowest physical address + */ + if (found && + physical_of_found <= + tmp_bbio->stripes[i].physical) + continue; + index_srcdev = i; + found = 1; + physical_of_found = + tmp_bbio->stripes[i].physical; + } + } + + if (found) { + mirror_num = index_srcdev + 1; + patch_the_first_stripe_for_dev_replace = 1; + physical_to_patch_in_first_stripe = physical_of_found; + } else { + WARN_ON(1); + ret = -EIO; + kfree(tmp_bbio); + goto out; + } + + kfree(tmp_bbio); + } else if (mirror_num > map->num_stripes) { + mirror_num = 0; + } + num_stripes = 1; stripe_index = 0; stripe_nr_orig = stripe_nr; - stripe_nr_end = (offset + *length + map->stripe_len - 1) & - (~(map->stripe_len - 1)); + stripe_nr_end = ALIGN(offset + *length, map->stripe_len); 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, stripe_nr_end - stripe_nr_orig); stripe_index = do_div(stripe_nr, map->num_stripes); } else if (map->type & BTRFS_BLOCK_GROUP_RAID1) { - if (rw & (REQ_WRITE | REQ_DISCARD)) + if (rw & (REQ_WRITE | REQ_DISCARD | REQ_GET_READ_MIRRORS)) num_stripes = map->num_stripes; else if (mirror_num) stripe_index = mirror_num - 1; else { - stripe_index = find_live_mirror(map, 0, + stripe_index = find_live_mirror(fs_info, map, 0, map->num_stripes, - current->pid % map->num_stripes); + current->pid % map->num_stripes, + dev_replace_is_ongoing); mirror_num = stripe_index + 1; } } else if (map->type & BTRFS_BLOCK_GROUP_DUP) { - if (rw & (REQ_WRITE | REQ_DISCARD)) { + if (rw & (REQ_WRITE | REQ_DISCARD | REQ_GET_READ_MIRRORS)) { num_stripes = map->num_stripes; } else if (mirror_num) { stripe_index = mirror_num - 1; @@ -3885,7 +4609,7 @@ static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, stripe_index = do_div(stripe_nr, factor); stripe_index *= map->sub_stripes; - if (rw & REQ_WRITE) + if (rw & (REQ_WRITE | REQ_GET_READ_MIRRORS)) num_stripes = map->sub_stripes; else if (rw & REQ_DISCARD) num_stripes = min_t(u64, map->sub_stripes * @@ -3895,11 +4619,72 @@ static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, stripe_index += mirror_num - 1; else { int old_stripe_index = stripe_index; - stripe_index = find_live_mirror(map, stripe_index, + stripe_index = find_live_mirror(fs_info, map, + stripe_index, map->sub_stripes, stripe_index + - current->pid % map->sub_stripes); + current->pid % map->sub_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 @@ -3911,7 +4696,14 @@ static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, } BUG_ON(stripe_index >= map->num_stripes); - bbio = kzalloc(btrfs_bio_size(num_stripes), GFP_NOFS); + num_alloc_stripes = num_stripes; + if (dev_replace_is_ongoing) { + if (rw & (REQ_WRITE | REQ_DISCARD)) + num_alloc_stripes <<= 1; + if (rw & REQ_GET_READ_MIRRORS) + num_alloc_stripes++; + } + bbio = kzalloc(btrfs_bio_size(num_alloc_stripes), GFP_NOFS); if (!bbio) { ret = -ENOMEM; goto out; @@ -3998,11 +4790,95 @@ static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, } } - if (rw & REQ_WRITE) { + 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; + } + } + + if (dev_replace_is_ongoing && (rw & (REQ_WRITE | REQ_DISCARD)) && + dev_replace->tgtdev != NULL) { + int index_where_to_add; + u64 srcdev_devid = dev_replace->srcdev->devid; + + /* + * duplicate the write operations while the dev replace + * procedure is running. Since the copying of the old disk + * to the new disk takes place at run time while the + * filesystem is mounted writable, the regular write + * operations to the old disk have to be duplicated to go + * to the new disk as well. + * Note that device->missing is handled by the caller, and + * that the write to the old disk is already set up in the + * stripes array. + */ + index_where_to_add = num_stripes; + for (i = 0; i < num_stripes; i++) { + if (bbio->stripes[i].dev->devid == srcdev_devid) { + /* write to new disk, too */ + struct btrfs_bio_stripe *new = + bbio->stripes + index_where_to_add; + struct btrfs_bio_stripe *old = + bbio->stripes + i; + + new->physical = old->physical; + new->length = old->length; + new->dev = dev_replace->tgtdev; + index_where_to_add++; + max_errors++; + } + } + num_stripes = index_where_to_add; + } else if (dev_replace_is_ongoing && (rw & REQ_GET_READ_MIRRORS) && + dev_replace->tgtdev != NULL) { + u64 srcdev_devid = dev_replace->srcdev->devid; + int index_srcdev = 0; + int found = 0; + u64 physical_of_found = 0; + + /* + * During the dev-replace procedure, the target drive can + * also be used to read data in case it is needed to repair + * a corrupt block elsewhere. This is possible if the + * requested area is left of the left cursor. In this area, + * the target drive is a full copy of the source drive. + */ + for (i = 0; i < num_stripes; i++) { + if (bbio->stripes[i].dev->devid == srcdev_devid) { + /* + * In case of DUP, in order to keep it + * simple, only add the mirror with the + * lowest physical address + */ + if (found && + physical_of_found <= + bbio->stripes[i].physical) + continue; + index_srcdev = i; + found = 1; + physical_of_found = bbio->stripes[i].physical; + } + } + if (found) { + u64 length = map->stripe_len; + + if (physical_of_found + length <= + dev_replace->cursor_left) { + struct btrfs_bio_stripe *tgtdev_stripe = + bbio->stripes + num_stripes; + + tgtdev_stripe->physical = physical_of_found; + tgtdev_stripe->length = + bbio->stripes[index_srcdev].length; + tgtdev_stripe->dev = dev_replace->tgtdev; + + num_stripes++; + } } } @@ -4010,17 +4886,35 @@ static int __btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, bbio->num_stripes = num_stripes; bbio->max_errors = max_errors; bbio->mirror_num = mirror_num; + + /* + * this is the case that REQ_READ && dev_replace_is_ongoing && + * mirror_num == num_stripes + 1 && dev_replace target drive is + * available as a mirror + */ + if (patch_the_first_stripe_for_dev_replace && num_stripes > 0) { + WARN_ON(num_stripes > 1); + bbio->stripes[0].dev = dev_replace->tgtdev; + 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); free_extent_map(em); return ret; } -int btrfs_map_block(struct btrfs_mapping_tree *map_tree, int rw, +int btrfs_map_block(struct btrfs_fs_info *fs_info, int rw, u64 logical, u64 *length, struct btrfs_bio **bbio_ret, int mirror_num) { - return __btrfs_map_block(map_tree, rw, logical, length, bbio_ret, - mirror_num); + return __btrfs_map_block(fs_info, rw, logical, length, bbio_ret, + mirror_num, NULL); } int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree, @@ -4034,6 +4928,7 @@ 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); @@ -4044,10 +4939,17 @@ int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree, 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 */ @@ -4067,8 +4969,11 @@ int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree, 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) @@ -4082,7 +4987,7 @@ int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree, *logical = buf; *naddrs = nr; - *stripe_len = map->stripe_len; + *stripe_len = rmap_len; free_extent_map(em); return 0; @@ -4156,7 +5061,7 @@ static void btrfs_end_bio(struct bio *bio, int err) 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; @@ -4190,13 +5095,18 @@ struct async_sched { * 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); @@ -4238,30 +5148,154 @@ static noinline void schedule_bio(struct btrfs_root *root, &device->work); } +static int bio_size_ok(struct block_device *bdev, struct bio *bio, + sector_t sector) +{ + struct bio_vec *prev; + struct request_queue *q = bdev_get_queue(bdev); + unsigned short max_sectors = queue_max_sectors(q); + struct bvec_merge_data bvm = { + .bi_bdev = bdev, + .bi_sector = sector, + .bi_rw = bio->bi_rw, + }; + + if (bio->bi_vcnt == 0) { + WARN_ON(1); + return 1; + } + + prev = &bio->bi_io_vec[bio->bi_vcnt - 1]; + if ((bio->bi_size >> 9) > max_sectors) + return 0; + + if (!q->merge_bvec_fn) + return 1; + + bvm.bi_size = bio->bi_size - prev->bv_len; + if (q->merge_bvec_fn(q, &bvm, prev) < prev->bv_len) + return 0; + return 1; +} + +static void submit_stripe_bio(struct btrfs_root *root, struct btrfs_bio *bbio, + struct bio *bio, u64 physical, int dev_nr, + int rw, int async) +{ + struct btrfs_device *dev = bbio->stripes[dev_nr].dev; + + bio->bi_private = bbio; + bio->bi_private = merge_stripe_index_into_bio_private( + bio->bi_private, (unsigned int)dev_nr); + bio->bi_end_io = btrfs_end_bio; + bio->bi_sector = physical >> 9; +#ifdef DEBUG + { + struct rcu_string *name; + + rcu_read_lock(); + name = rcu_dereference(dev->name); + pr_debug("btrfs_map_bio: rw %d, sector=%llu, dev=%lu " + "(%s id %llu), size=%u\n", rw, + (u64)bio->bi_sector, (u_long)dev->bdev->bd_dev, + name->str, dev->devid, bio->bi_size); + rcu_read_unlock(); + } +#endif + bio->bi_bdev = dev->bdev; + if (async) + btrfs_schedule_bio(root, dev, rw, bio); + else + btrfsic_submit_bio(rw, bio); +} + +static int breakup_stripe_bio(struct btrfs_root *root, struct btrfs_bio *bbio, + struct bio *first_bio, struct btrfs_device *dev, + int dev_nr, int rw, int async) +{ + struct bio_vec *bvec = first_bio->bi_io_vec; + struct bio *bio; + int nr_vecs = bio_get_nr_vecs(dev->bdev); + u64 physical = bbio->stripes[dev_nr].physical; + +again: + bio = btrfs_bio_alloc(dev->bdev, physical >> 9, nr_vecs, GFP_NOFS); + if (!bio) + return -ENOMEM; + + while (bvec <= (first_bio->bi_io_vec + first_bio->bi_vcnt - 1)) { + if (bio_add_page(bio, bvec->bv_page, bvec->bv_len, + bvec->bv_offset) < bvec->bv_len) { + u64 len = bio->bi_size; + + atomic_inc(&bbio->stripes_pending); + submit_stripe_bio(root, bbio, bio, physical, dev_nr, + rw, async); + physical += len; + goto again; + } + bvec++; + } + + submit_stripe_bio(root, bbio, bio, physical, dev_nr, rw, async); + return 0; +} + +static void bbio_error(struct btrfs_bio *bbio, struct bio *bio, u64 logical) +{ + atomic_inc(&bbio->error); + if (atomic_dec_and_test(&bbio->stripes_pending)) { + bio->bi_private = bbio->private; + bio->bi_end_io = bbio->end_io; + bio->bi_bdev = (struct block_device *) + (unsigned long)bbio->mirror_num; + bio->bi_sector = logical >> 9; + kfree(bbio); + bio_endio(bio, -EIO); + } +} + int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio, int mirror_num, int async_submit) { - struct btrfs_mapping_tree *map_tree; struct btrfs_device *dev; struct bio *first_bio = 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; struct btrfs_bio *bbio = NULL; length = bio->bi_size; - map_tree = &root->fs_info->mapping_tree; map_length = length; - ret = btrfs_map_block(map_tree, rw, logical, &map_length, &bbio, - mirror_num); + 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, @@ -4270,58 +5304,49 @@ int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio, 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)) { + bbio_error(bbio, first_bio, logical); + dev_nr++; + continue; + } + + /* + * Check and see if we're ok with this bio based on it's size + * and offset with the given device. + */ + if (!bio_size_ok(dev->bdev, first_bio, + bbio->stripes[dev_nr].physical >> 9)) { + ret = breakup_stripe_bio(root, bbio, first_bio, dev, + dev_nr, rw, async_submit); + BUG_ON(ret); + dev_nr++; + continue; + } + if (dev_nr < total_devs - 1) { bio = bio_clone(first_bio, GFP_NOFS); BUG_ON(!bio); /* -ENOMEM */ } else { bio = first_bio; } - bio->bi_private = bbio; - bio->bi_private = merge_stripe_index_into_bio_private( - bio->bi_private, (unsigned int)dev_nr); - bio->bi_end_io = btrfs_end_bio; - bio->bi_sector = bbio->stripes[dev_nr].physical >> 9; - dev = bbio->stripes[dev_nr].dev; - if (dev && dev->bdev && (rw != WRITE || dev->writeable)) { -#ifdef DEBUG - struct rcu_string *name; - - rcu_read_lock(); - name = rcu_dereference(dev->name); - pr_debug("btrfs_map_bio: rw %d, secor=%llu, dev=%lu " - "(%s id %llu), size=%u\n", rw, - (u64)bio->bi_sector, (u_long)dev->bdev->bd_dev, - name->str, dev->devid, bio->bi_size); - rcu_read_unlock(); -#endif - bio->bi_bdev = dev->bdev; - if (async_submit) - schedule_bio(root, dev, rw, bio); - else - btrfsic_submit_bio(rw, bio); - } else { - bio->bi_bdev = root->fs_info->fs_devices->latest_bdev; - bio->bi_sector = logical >> 9; - bio_endio(bio, -EIO); - } + + submit_stripe_bio(root, bbio, bio, + bbio->stripes[dev_nr].physical, dev_nr, rw, + async_submit); dev_nr++; } return 0; } -struct btrfs_device *btrfs_find_device(struct btrfs_root *root, u64 devid, +struct btrfs_device *btrfs_find_device(struct btrfs_fs_info *fs_info, u64 devid, u8 *uuid, u8 *fsid) { struct btrfs_device *device; struct btrfs_fs_devices *cur_devices; - cur_devices = root->fs_info->fs_devices; + cur_devices = fs_info->fs_devices; while (cur_devices) { if (!fsid || !memcmp(cur_devices->fsid, fsid, BTRFS_UUID_SIZE)) { @@ -4402,6 +5427,7 @@ static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key, em->bdev = (struct block_device *)map; em->start = logical; em->len = length; + em->orig_start = 0; em->block_start = 0; em->block_len = em->len; @@ -4419,8 +5445,8 @@ static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key, read_extent_buffer(leaf, uuid, (unsigned long) btrfs_stripe_dev_uuid_nr(chunk, i), BTRFS_UUID_SIZE); - map->stripes[i].dev = btrfs_find_device(root, devid, uuid, - NULL); + map->stripes[i].dev = btrfs_find_device(root->fs_info, devid, + uuid, NULL); if (!map->stripes[i].dev && !btrfs_test_opt(root, DEGRADED)) { kfree(map); free_extent_map(em); @@ -4461,6 +5487,8 @@ static void fill_device_from_item(struct extent_buffer *leaf, device->io_align = btrfs_device_io_align(leaf, dev_item); device->io_width = btrfs_device_io_width(leaf, dev_item); device->sector_size = btrfs_device_sector_size(leaf, dev_item); + WARN_ON(device->devid == BTRFS_DEV_REPLACE_DEVID); + device->is_tgtdev_for_dev_replace = 0; ptr = (unsigned long)btrfs_device_uuid(dev_item); read_extent_buffer(leaf, device->uuid, ptr, BTRFS_UUID_SIZE); @@ -4538,7 +5566,7 @@ static int read_one_dev(struct btrfs_root *root, return ret; } - device = btrfs_find_device(root, devid, dev_uuid, fs_uuid); + device = btrfs_find_device(root->fs_info, devid, dev_uuid, fs_uuid); if (!device || !device->bdev) { if (!btrfs_test_opt(root, DEGRADED)) return -EIO; @@ -4571,7 +5599,7 @@ static int read_one_dev(struct btrfs_root *root, fill_device_from_item(leaf, dev_item, device); device->dev_root = root->fs_info->dev_root; device->in_fs_metadata = 1; - if (device->writeable) { + if (device->writeable && !device->is_tgtdev_for_dev_replace) { device->fs_devices->total_rw_bytes += device->total_bytes; spin_lock(&root->fs_info->free_chunk_lock); root->fs_info->free_chunk_space += device->total_bytes - @@ -4930,7 +5958,7 @@ int btrfs_get_dev_stats(struct btrfs_root *root, int i; mutex_lock(&fs_devices->device_list_mutex); - dev = btrfs_find_device(root, stats->devid, NULL, NULL); + dev = btrfs_find_device(root->fs_info, stats->devid, NULL, NULL); mutex_unlock(&fs_devices->device_list_mutex); if (!dev) { @@ -4958,3 +5986,21 @@ int btrfs_get_dev_stats(struct btrfs_root *root, stats->nr_items = BTRFS_DEV_STAT_VALUES_MAX; return 0; } + +int btrfs_scratch_superblock(struct btrfs_device *device) +{ + struct buffer_head *bh; + struct btrfs_super_block *disk_super; + + bh = btrfs_read_dev_super(device->bdev); + if (!bh) + return -EINVAL; + disk_super = (struct btrfs_super_block *)bh->b_data; + + memset(&disk_super->magic, 0, sizeof(disk_super->magic)); + set_buffer_dirty(bh); + sync_dirty_buffer(bh); + brelse(bh); + + return 0; +} |