Merge branch 'for-chris' of git://git.kernel.org/pub/scm/linux/kernel/git/arne/btrfs...
authorChris Mason <chris.mason@oracle.com>
Mon, 23 May 2011 10:30:52 +0000 (06:30 -0400)
committerChris Mason <chris.mason@oracle.com>
Mon, 23 May 2011 10:30:52 +0000 (06:30 -0400)
Conflicts:
fs/btrfs/Makefile
fs/btrfs/ctree.h
fs/btrfs/volumes.h

Signed-off-by: Chris Mason <chris.mason@oracle.com>
12 files changed:
1  2 
fs/btrfs/Makefile
fs/btrfs/ctree.h
fs/btrfs/disk-io.c
fs/btrfs/file-item.c
fs/btrfs/inode.c
fs/btrfs/ioctl.c
fs/btrfs/relocation.c
fs/btrfs/scrub.c
fs/btrfs/transaction.c
fs/btrfs/tree-log.c
fs/btrfs/volumes.c
fs/btrfs/volumes.h

index a8411c22313d743c8ac222b03f7e58caf185d528,8fda3133c1b81ca3703d41001ae570149d6c8a8a..9b72dcf1cd258bd2e7733079b399ddd6413e2826
@@@ -7,4 -7,4 +7,4 @@@ btrfs-y += super.o ctree.o extent-tree.
           extent_map.o sysfs.o struct-funcs.o xattr.o ordered-data.o \
           extent_io.o volumes.o async-thread.o ioctl.o locking.o orphan.o \
           export.o tree-log.o acl.o free-space-cache.o zlib.o lzo.o \
-          compression.o delayed-ref.o relocation.o delayed-inode.o
 -         compression.o delayed-ref.o relocation.o scrub.o
++         compression.o delayed-ref.o relocation.o delayed-inode.o scrub.o
index e7d40791ec9f26f84a41280262bb8056066a23e0,ee904666b766682b7d422a60124ab8feecc70014..0d50c57caf5df9a2d80dc9dfce332afad72af798
@@@ -1083,10 -1083,19 +1089,21 @@@ struct btrfs_fs_info 
  
        void *bdev_holder;
  
+       /* private scrub information */
+       struct mutex scrub_lock;
+       atomic_t scrubs_running;
+       atomic_t scrub_pause_req;
+       atomic_t scrubs_paused;
+       atomic_t scrub_cancel_req;
+       wait_queue_head_t scrub_pause_wait;
+       struct rw_semaphore scrub_super_lock;
+       int scrub_workers_refcnt;
+       struct btrfs_workers scrub_workers;
        /* filesystem state */
        u64 fs_state;
 +
 +      struct btrfs_delayed_root *delayed_root;
  };
  
  /*
index a2eb3a3755db5fef162093b4e2d58652f2b25bc0,e48e8095c61f88766709307d13b621cf660483dd..29a2475b9b1408be2062f373cfe7652fcf3eb2de
@@@ -1646,14 -1772,18 +1646,25 @@@ struct btrfs_root *open_ctree(struct su
  
        INIT_LIST_HEAD(&fs_info->ordered_extents);
        spin_lock_init(&fs_info->ordered_extent_lock);
 +      fs_info->delayed_root = kmalloc(sizeof(struct btrfs_delayed_root),
 +                                      GFP_NOFS);
 +      if (!fs_info->delayed_root) {
 +              err = -ENOMEM;
 +              goto fail_iput;
 +      }
 +      btrfs_init_delayed_root(fs_info->delayed_root);
  
+       mutex_init(&fs_info->scrub_lock);
+       atomic_set(&fs_info->scrubs_running, 0);
+       atomic_set(&fs_info->scrub_pause_req, 0);
+       atomic_set(&fs_info->scrubs_paused, 0);
+       atomic_set(&fs_info->scrub_cancel_req, 0);
+       init_waitqueue_head(&fs_info->scrub_pause_wait);
+       init_rwsem(&fs_info->scrub_super_lock);
+       fs_info->scrub_workers_refcnt = 0;
+       btrfs_init_workers(&fs_info->scrub_workers, "scrub",
+                          fs_info->thread_pool_size, &fs_info->generic_worker);
        sb->s_blocksize = 4096;
        sb->s_blocksize_bits = blksize_bits(4096);
        sb->s_bdi = &fs_info->bdi;
Simple merge
Simple merge
Simple merge
Simple merge
index 0000000000000000000000000000000000000000,a31f2a9bd2e20f43f9be9e6ccabc98ef807164ea..87a2f1273136253ac2d5a32279c1ff805c069063
mode 000000,100644..100644
--- /dev/null
@@@ -1,0 -1,1368 +1,1368 @@@
 -      rw |= REQ_SYNC | REQ_UNPLUG;
+ /*
+  * Copyright (C) 2011 STRATO.  All rights reserved.
+  *
+  * This program is free software; you can redistribute it and/or
+  * modify it under the terms of the GNU General Public
+  * License v2 as published by the Free Software Foundation.
+  *
+  * This program is distributed in the hope that it will be useful,
+  * but WITHOUT ANY WARRANTY; without even the implied warranty of
+  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+  * General Public License for more details.
+  *
+  * You should have received a copy of the GNU General Public
+  * License along with this program; if not, write to the
+  * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+  * Boston, MA 021110-1307, USA.
+  */
+ #include <linux/sched.h>
+ #include <linux/pagemap.h>
+ #include <linux/writeback.h>
+ #include <linux/blkdev.h>
+ #include <linux/rbtree.h>
+ #include <linux/slab.h>
+ #include <linux/workqueue.h>
+ #include "ctree.h"
+ #include "volumes.h"
+ #include "disk-io.h"
+ #include "ordered-data.h"
+ /*
+  * This is only the first step towards a full-features scrub. It reads all
+  * extent and super block and verifies the checksums. In case a bad checksum
+  * is found or the extent cannot be read, good data will be written back if
+  * any can be found.
+  *
+  * Future enhancements:
+  *  - To enhance the performance, better read-ahead strategies for the
+  *    extent-tree can be employed.
+  *  - In case an unrepairable extent is encountered, track which files are
+  *    affected and report them
+  *  - In case of a read error on files with nodatasum, map the file and read
+  *    the extent to trigger a writeback of the good copy
+  *  - track and record media errors, throw out bad devices
+  *  - add a mode to also read unallocated space
+  *  - make the prefetch cancellable
+  */
+ struct scrub_bio;
+ struct scrub_page;
+ struct scrub_dev;
+ static void scrub_bio_end_io(struct bio *bio, int err);
+ static void scrub_checksum(struct btrfs_work *work);
+ static int scrub_checksum_data(struct scrub_dev *sdev,
+                              struct scrub_page *spag, void *buffer);
+ static int scrub_checksum_tree_block(struct scrub_dev *sdev,
+                                    struct scrub_page *spag, u64 logical,
+                                    void *buffer);
+ static int scrub_checksum_super(struct scrub_bio *sbio, void *buffer);
+ static int scrub_fixup_check(struct scrub_bio *sbio, int ix);
+ static void scrub_fixup_end_io(struct bio *bio, int err);
+ static int scrub_fixup_io(int rw, struct block_device *bdev, sector_t sector,
+                         struct page *page);
+ static void scrub_fixup(struct scrub_bio *sbio, int ix);
+ #define SCRUB_PAGES_PER_BIO   16      /* 64k per bio */
+ #define SCRUB_BIOS_PER_DEV    16      /* 1 MB per device in flight */
+ struct scrub_page {
+       u64                     flags;  /* extent flags */
+       u64                     generation;
+       u64                     mirror_num;
+       int                     have_csum;
+       u8                      csum[BTRFS_CSUM_SIZE];
+ };
+ struct scrub_bio {
+       int                     index;
+       struct scrub_dev        *sdev;
+       struct bio              *bio;
+       int                     err;
+       u64                     logical;
+       u64                     physical;
+       struct scrub_page       spag[SCRUB_PAGES_PER_BIO];
+       u64                     count;
+       int                     next_free;
+       struct btrfs_work       work;
+ };
+ struct scrub_dev {
+       struct scrub_bio        *bios[SCRUB_BIOS_PER_DEV];
+       struct btrfs_device     *dev;
+       int                     first_free;
+       int                     curr;
+       atomic_t                in_flight;
+       spinlock_t              list_lock;
+       wait_queue_head_t       list_wait;
+       u16                     csum_size;
+       struct list_head        csum_list;
+       atomic_t                cancel_req;
+       int                     readonly;
+       /*
+        * statistics
+        */
+       struct btrfs_scrub_progress stat;
+       spinlock_t              stat_lock;
+ };
+ static void scrub_free_csums(struct scrub_dev *sdev)
+ {
+       while (!list_empty(&sdev->csum_list)) {
+               struct btrfs_ordered_sum *sum;
+               sum = list_first_entry(&sdev->csum_list,
+                                      struct btrfs_ordered_sum, list);
+               list_del(&sum->list);
+               kfree(sum);
+       }
+ }
+ static noinline_for_stack void scrub_free_dev(struct scrub_dev *sdev)
+ {
+       int i;
+       int j;
+       struct page *last_page;
+       if (!sdev)
+               return;
+       for (i = 0; i < SCRUB_BIOS_PER_DEV; ++i) {
+               struct scrub_bio *sbio = sdev->bios[i];
+               struct bio *bio;
+               if (!sbio)
+                       break;
+               bio = sbio->bio;
+               if (bio) {
+                       last_page = NULL;
+                       for (j = 0; j < bio->bi_vcnt; ++j) {
+                               if (bio->bi_io_vec[j].bv_page == last_page)
+                                       continue;
+                               last_page = bio->bi_io_vec[j].bv_page;
+                               __free_page(last_page);
+                       }
+                       bio_put(bio);
+               }
+               kfree(sbio);
+       }
+       scrub_free_csums(sdev);
+       kfree(sdev);
+ }
+ static noinline_for_stack
+ struct scrub_dev *scrub_setup_dev(struct btrfs_device *dev)
+ {
+       struct scrub_dev *sdev;
+       int             i;
+       int             j;
+       int             ret;
+       struct btrfs_fs_info *fs_info = dev->dev_root->fs_info;
+       sdev = kzalloc(sizeof(*sdev), GFP_NOFS);
+       if (!sdev)
+               goto nomem;
+       sdev->dev = dev;
+       for (i = 0; i < SCRUB_BIOS_PER_DEV; ++i) {
+               struct bio *bio;
+               struct scrub_bio *sbio;
+               sbio = kzalloc(sizeof(*sbio), GFP_NOFS);
+               if (!sbio)
+                       goto nomem;
+               sdev->bios[i] = sbio;
+               bio = bio_kmalloc(GFP_NOFS, SCRUB_PAGES_PER_BIO);
+               if (!bio)
+                       goto nomem;
+               sbio->index = i;
+               sbio->sdev = sdev;
+               sbio->bio = bio;
+               sbio->count = 0;
+               sbio->work.func = scrub_checksum;
+               bio->bi_private = sdev->bios[i];
+               bio->bi_end_io = scrub_bio_end_io;
+               bio->bi_sector = 0;
+               bio->bi_bdev = dev->bdev;
+               bio->bi_size = 0;
+               for (j = 0; j < SCRUB_PAGES_PER_BIO; ++j) {
+                       struct page *page;
+                       page = alloc_page(GFP_NOFS);
+                       if (!page)
+                               goto nomem;
+                       ret = bio_add_page(bio, page, PAGE_SIZE, 0);
+                       if (!ret)
+                               goto nomem;
+               }
+               WARN_ON(bio->bi_vcnt != SCRUB_PAGES_PER_BIO);
+               if (i != SCRUB_BIOS_PER_DEV-1)
+                       sdev->bios[i]->next_free = i + 1;
+                else
+                       sdev->bios[i]->next_free = -1;
+       }
+       sdev->first_free = 0;
+       sdev->curr = -1;
+       atomic_set(&sdev->in_flight, 0);
+       atomic_set(&sdev->cancel_req, 0);
+       sdev->csum_size = btrfs_super_csum_size(&fs_info->super_copy);
+       INIT_LIST_HEAD(&sdev->csum_list);
+       spin_lock_init(&sdev->list_lock);
+       spin_lock_init(&sdev->stat_lock);
+       init_waitqueue_head(&sdev->list_wait);
+       return sdev;
+ nomem:
+       scrub_free_dev(sdev);
+       return ERR_PTR(-ENOMEM);
+ }
+ /*
+  * scrub_recheck_error gets called when either verification of the page
+  * failed or the bio failed to read, e.g. with EIO. In the latter case,
+  * recheck_error gets called for every page in the bio, even though only
+  * one may be bad
+  */
+ static void scrub_recheck_error(struct scrub_bio *sbio, int ix)
+ {
+       if (sbio->err) {
+               if (scrub_fixup_io(READ, sbio->sdev->dev->bdev,
+                                  (sbio->physical + ix * PAGE_SIZE) >> 9,
+                                  sbio->bio->bi_io_vec[ix].bv_page) == 0) {
+                       if (scrub_fixup_check(sbio, ix) == 0)
+                               return;
+               }
+       }
+       scrub_fixup(sbio, ix);
+ }
+ static int scrub_fixup_check(struct scrub_bio *sbio, int ix)
+ {
+       int ret = 1;
+       struct page *page;
+       void *buffer;
+       u64 flags = sbio->spag[ix].flags;
+       page = sbio->bio->bi_io_vec[ix].bv_page;
+       buffer = kmap_atomic(page, KM_USER0);
+       if (flags & BTRFS_EXTENT_FLAG_DATA) {
+               ret = scrub_checksum_data(sbio->sdev,
+                                         sbio->spag + ix, buffer);
+       } else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
+               ret = scrub_checksum_tree_block(sbio->sdev,
+                                               sbio->spag + ix,
+                                               sbio->logical + ix * PAGE_SIZE,
+                                               buffer);
+       } else {
+               WARN_ON(1);
+       }
+       kunmap_atomic(buffer, KM_USER0);
+       return ret;
+ }
+ static void scrub_fixup_end_io(struct bio *bio, int err)
+ {
+       complete((struct completion *)bio->bi_private);
+ }
+ static void scrub_fixup(struct scrub_bio *sbio, int ix)
+ {
+       struct scrub_dev *sdev = sbio->sdev;
+       struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info;
+       struct btrfs_mapping_tree *map_tree = &fs_info->mapping_tree;
+       struct btrfs_multi_bio *multi = NULL;
+       u64 logical = sbio->logical + ix * PAGE_SIZE;
+       u64 length;
+       int i;
+       int ret;
+       DECLARE_COMPLETION_ONSTACK(complete);
+       if ((sbio->spag[ix].flags & BTRFS_EXTENT_FLAG_DATA) &&
+           (sbio->spag[ix].have_csum == 0)) {
+               /*
+                * nodatasum, don't try to fix anything
+                * FIXME: we can do better, open the inode and trigger a
+                * writeback
+                */
+               goto uncorrectable;
+       }
+       length = PAGE_SIZE;
+       ret = btrfs_map_block(map_tree, REQ_WRITE, logical, &length,
+                             &multi, 0);
+       if (ret || !multi || length < PAGE_SIZE) {
+               printk(KERN_ERR
+                      "scrub_fixup: btrfs_map_block failed us for %llu\n",
+                      (unsigned long long)logical);
+               WARN_ON(1);
+               return;
+       }
+       if (multi->num_stripes == 1)
+               /* there aren't any replicas */
+               goto uncorrectable;
+       /*
+        * first find a good copy
+        */
+       for (i = 0; i < multi->num_stripes; ++i) {
+               if (i == sbio->spag[ix].mirror_num)
+                       continue;
+               if (scrub_fixup_io(READ, multi->stripes[i].dev->bdev,
+                                  multi->stripes[i].physical >> 9,
+                                  sbio->bio->bi_io_vec[ix].bv_page)) {
+                       /* I/O-error, this is not a good copy */
+                       continue;
+               }
+               if (scrub_fixup_check(sbio, ix) == 0)
+                       break;
+       }
+       if (i == multi->num_stripes)
+               goto uncorrectable;
+       if (!sdev->readonly) {
+               /*
+                * bi_io_vec[ix].bv_page now contains good data, write it back
+                */
+               if (scrub_fixup_io(WRITE, sdev->dev->bdev,
+                                  (sbio->physical + ix * PAGE_SIZE) >> 9,
+                                  sbio->bio->bi_io_vec[ix].bv_page)) {
+                       /* I/O-error, writeback failed, give up */
+                       goto uncorrectable;
+               }
+       }
+       kfree(multi);
+       spin_lock(&sdev->stat_lock);
+       ++sdev->stat.corrected_errors;
+       spin_unlock(&sdev->stat_lock);
+       if (printk_ratelimit())
+               printk(KERN_ERR "btrfs: fixed up at %llu\n",
+                      (unsigned long long)logical);
+       return;
+ uncorrectable:
+       kfree(multi);
+       spin_lock(&sdev->stat_lock);
+       ++sdev->stat.uncorrectable_errors;
+       spin_unlock(&sdev->stat_lock);
+       if (printk_ratelimit())
+               printk(KERN_ERR "btrfs: unable to fixup at %llu\n",
+                        (unsigned long long)logical);
+ }
+ static int scrub_fixup_io(int rw, struct block_device *bdev, sector_t sector,
+                        struct page *page)
+ {
+       struct bio *bio = NULL;
+       int ret;
+       DECLARE_COMPLETION_ONSTACK(complete);
+       /* we are going to wait on this IO */
 -              btrfs_release_path(root, path);
++      rw |= REQ_SYNC;
+       bio = bio_alloc(GFP_NOFS, 1);
+       bio->bi_bdev = bdev;
+       bio->bi_sector = sector;
+       bio_add_page(bio, page, PAGE_SIZE, 0);
+       bio->bi_end_io = scrub_fixup_end_io;
+       bio->bi_private = &complete;
+       submit_bio(rw, bio);
+       wait_for_completion(&complete);
+       ret = !test_bit(BIO_UPTODATE, &bio->bi_flags);
+       bio_put(bio);
+       return ret;
+ }
+ static void scrub_bio_end_io(struct bio *bio, int err)
+ {
+       struct scrub_bio *sbio = bio->bi_private;
+       struct scrub_dev *sdev = sbio->sdev;
+       struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info;
+       sbio->err = err;
+       btrfs_queue_worker(&fs_info->scrub_workers, &sbio->work);
+ }
+ static void scrub_checksum(struct btrfs_work *work)
+ {
+       struct scrub_bio *sbio = container_of(work, struct scrub_bio, work);
+       struct scrub_dev *sdev = sbio->sdev;
+       struct page *page;
+       void *buffer;
+       int i;
+       u64 flags;
+       u64 logical;
+       int ret;
+       if (sbio->err) {
+               for (i = 0; i < sbio->count; ++i)
+                       scrub_recheck_error(sbio, i);
+               sbio->bio->bi_flags &= ~(BIO_POOL_MASK - 1);
+               sbio->bio->bi_flags |= 1 << BIO_UPTODATE;
+               sbio->bio->bi_phys_segments = 0;
+               sbio->bio->bi_idx = 0;
+               for (i = 0; i < sbio->count; i++) {
+                       struct bio_vec *bi;
+                       bi = &sbio->bio->bi_io_vec[i];
+                       bi->bv_offset = 0;
+                       bi->bv_len = PAGE_SIZE;
+               }
+               spin_lock(&sdev->stat_lock);
+               ++sdev->stat.read_errors;
+               spin_unlock(&sdev->stat_lock);
+               goto out;
+       }
+       for (i = 0; i < sbio->count; ++i) {
+               page = sbio->bio->bi_io_vec[i].bv_page;
+               buffer = kmap_atomic(page, KM_USER0);
+               flags = sbio->spag[i].flags;
+               logical = sbio->logical + i * PAGE_SIZE;
+               ret = 0;
+               if (flags & BTRFS_EXTENT_FLAG_DATA) {
+                       ret = scrub_checksum_data(sdev, sbio->spag + i, buffer);
+               } else if (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK) {
+                       ret = scrub_checksum_tree_block(sdev, sbio->spag + i,
+                                                       logical, buffer);
+               } else if (flags & BTRFS_EXTENT_FLAG_SUPER) {
+                       BUG_ON(i);
+                       (void)scrub_checksum_super(sbio, buffer);
+               } else {
+                       WARN_ON(1);
+               }
+               kunmap_atomic(buffer, KM_USER0);
+               if (ret)
+                       scrub_recheck_error(sbio, i);
+       }
+ out:
+       spin_lock(&sdev->list_lock);
+       sbio->next_free = sdev->first_free;
+       sdev->first_free = sbio->index;
+       spin_unlock(&sdev->list_lock);
+       atomic_dec(&sdev->in_flight);
+       wake_up(&sdev->list_wait);
+ }
+ static int scrub_checksum_data(struct scrub_dev *sdev,
+                              struct scrub_page *spag, void *buffer)
+ {
+       u8 csum[BTRFS_CSUM_SIZE];
+       u32 crc = ~(u32)0;
+       int fail = 0;
+       struct btrfs_root *root = sdev->dev->dev_root;
+       if (!spag->have_csum)
+               return 0;
+       crc = btrfs_csum_data(root, buffer, crc, PAGE_SIZE);
+       btrfs_csum_final(crc, csum);
+       if (memcmp(csum, spag->csum, sdev->csum_size))
+               fail = 1;
+       spin_lock(&sdev->stat_lock);
+       ++sdev->stat.data_extents_scrubbed;
+       sdev->stat.data_bytes_scrubbed += PAGE_SIZE;
+       if (fail)
+               ++sdev->stat.csum_errors;
+       spin_unlock(&sdev->stat_lock);
+       return fail;
+ }
+ static int scrub_checksum_tree_block(struct scrub_dev *sdev,
+                                    struct scrub_page *spag, u64 logical,
+                                    void *buffer)
+ {
+       struct btrfs_header *h;
+       struct btrfs_root *root = sdev->dev->dev_root;
+       struct btrfs_fs_info *fs_info = root->fs_info;
+       u8 csum[BTRFS_CSUM_SIZE];
+       u32 crc = ~(u32)0;
+       int fail = 0;
+       int crc_fail = 0;
+       /*
+        * we don't use the getter functions here, as we
+        * a) don't have an extent buffer and
+        * b) the page is already kmapped
+        */
+       h = (struct btrfs_header *)buffer;
+       if (logical != le64_to_cpu(h->bytenr))
+               ++fail;
+       if (spag->generation != le64_to_cpu(h->generation))
+               ++fail;
+       if (memcmp(h->fsid, fs_info->fsid, BTRFS_UUID_SIZE))
+               ++fail;
+       if (memcmp(h->chunk_tree_uuid, fs_info->chunk_tree_uuid,
+                  BTRFS_UUID_SIZE))
+               ++fail;
+       crc = btrfs_csum_data(root, buffer + BTRFS_CSUM_SIZE, crc,
+                             PAGE_SIZE - BTRFS_CSUM_SIZE);
+       btrfs_csum_final(crc, csum);
+       if (memcmp(csum, h->csum, sdev->csum_size))
+               ++crc_fail;
+       spin_lock(&sdev->stat_lock);
+       ++sdev->stat.tree_extents_scrubbed;
+       sdev->stat.tree_bytes_scrubbed += PAGE_SIZE;
+       if (crc_fail)
+               ++sdev->stat.csum_errors;
+       if (fail)
+               ++sdev->stat.verify_errors;
+       spin_unlock(&sdev->stat_lock);
+       return fail || crc_fail;
+ }
+ static int scrub_checksum_super(struct scrub_bio *sbio, void *buffer)
+ {
+       struct btrfs_super_block *s;
+       u64 logical;
+       struct scrub_dev *sdev = sbio->sdev;
+       struct btrfs_root *root = sdev->dev->dev_root;
+       struct btrfs_fs_info *fs_info = root->fs_info;
+       u8 csum[BTRFS_CSUM_SIZE];
+       u32 crc = ~(u32)0;
+       int fail = 0;
+       s = (struct btrfs_super_block *)buffer;
+       logical = sbio->logical;
+       if (logical != le64_to_cpu(s->bytenr))
+               ++fail;
+       if (sbio->spag[0].generation != le64_to_cpu(s->generation))
+               ++fail;
+       if (memcmp(s->fsid, fs_info->fsid, BTRFS_UUID_SIZE))
+               ++fail;
+       crc = btrfs_csum_data(root, buffer + BTRFS_CSUM_SIZE, crc,
+                             PAGE_SIZE - BTRFS_CSUM_SIZE);
+       btrfs_csum_final(crc, csum);
+       if (memcmp(csum, s->csum, sbio->sdev->csum_size))
+               ++fail;
+       if (fail) {
+               /*
+                * if we find an error in a super block, we just report it.
+                * They will get written with the next transaction commit
+                * anyway
+                */
+               spin_lock(&sdev->stat_lock);
+               ++sdev->stat.super_errors;
+               spin_unlock(&sdev->stat_lock);
+       }
+       return fail;
+ }
+ static int scrub_submit(struct scrub_dev *sdev)
+ {
+       struct scrub_bio *sbio;
+       if (sdev->curr == -1)
+               return 0;
+       sbio = sdev->bios[sdev->curr];
+       sbio->bio->bi_sector = sbio->physical >> 9;
+       sbio->bio->bi_size = sbio->count * PAGE_SIZE;
+       sbio->bio->bi_next = NULL;
+       sbio->bio->bi_flags |= 1 << BIO_UPTODATE;
+       sbio->bio->bi_comp_cpu = -1;
+       sbio->bio->bi_bdev = sdev->dev->bdev;
+       sbio->err = 0;
+       sdev->curr = -1;
+       atomic_inc(&sdev->in_flight);
+       submit_bio(0, sbio->bio);
+       return 0;
+ }
+ static int scrub_page(struct scrub_dev *sdev, u64 logical, u64 len,
+                     u64 physical, u64 flags, u64 gen, u64 mirror_num,
+                     u8 *csum, int force)
+ {
+       struct scrub_bio *sbio;
+ again:
+       /*
+        * grab a fresh bio or wait for one to become available
+        */
+       while (sdev->curr == -1) {
+               spin_lock(&sdev->list_lock);
+               sdev->curr = sdev->first_free;
+               if (sdev->curr != -1) {
+                       sdev->first_free = sdev->bios[sdev->curr]->next_free;
+                       sdev->bios[sdev->curr]->next_free = -1;
+                       sdev->bios[sdev->curr]->count = 0;
+                       spin_unlock(&sdev->list_lock);
+               } else {
+                       spin_unlock(&sdev->list_lock);
+                       wait_event(sdev->list_wait, sdev->first_free != -1);
+               }
+       }
+       sbio = sdev->bios[sdev->curr];
+       if (sbio->count == 0) {
+               sbio->physical = physical;
+               sbio->logical = logical;
+       } else if (sbio->physical + sbio->count * PAGE_SIZE != physical) {
+               scrub_submit(sdev);
+               goto again;
+       }
+       sbio->spag[sbio->count].flags = flags;
+       sbio->spag[sbio->count].generation = gen;
+       sbio->spag[sbio->count].have_csum = 0;
+       sbio->spag[sbio->count].mirror_num = mirror_num;
+       if (csum) {
+               sbio->spag[sbio->count].have_csum = 1;
+               memcpy(sbio->spag[sbio->count].csum, csum, sdev->csum_size);
+       }
+       ++sbio->count;
+       if (sbio->count == SCRUB_PAGES_PER_BIO || force)
+               scrub_submit(sdev);
+       return 0;
+ }
+ static int scrub_find_csum(struct scrub_dev *sdev, u64 logical, u64 len,
+                          u8 *csum)
+ {
+       struct btrfs_ordered_sum *sum = NULL;
+       int ret = 0;
+       unsigned long i;
+       unsigned long num_sectors;
+       u32 sectorsize = sdev->dev->dev_root->sectorsize;
+       while (!list_empty(&sdev->csum_list)) {
+               sum = list_first_entry(&sdev->csum_list,
+                                      struct btrfs_ordered_sum, list);
+               if (sum->bytenr > logical)
+                       return 0;
+               if (sum->bytenr + sum->len > logical)
+                       break;
+               ++sdev->stat.csum_discards;
+               list_del(&sum->list);
+               kfree(sum);
+               sum = NULL;
+       }
+       if (!sum)
+               return 0;
+       num_sectors = sum->len / sectorsize;
+       for (i = 0; i < num_sectors; ++i) {
+               if (sum->sums[i].bytenr == logical) {
+                       memcpy(csum, &sum->sums[i].sum, sdev->csum_size);
+                       ret = 1;
+                       break;
+               }
+       }
+       if (ret && i == num_sectors - 1) {
+               list_del(&sum->list);
+               kfree(sum);
+       }
+       return ret;
+ }
+ /* scrub extent tries to collect up to 64 kB for each bio */
+ static int scrub_extent(struct scrub_dev *sdev, u64 logical, u64 len,
+                       u64 physical, u64 flags, u64 gen, u64 mirror_num)
+ {
+       int ret;
+       u8 csum[BTRFS_CSUM_SIZE];
+       while (len) {
+               u64 l = min_t(u64, len, PAGE_SIZE);
+               int have_csum = 0;
+               if (flags & BTRFS_EXTENT_FLAG_DATA) {
+                       /* push csums to sbio */
+                       have_csum = scrub_find_csum(sdev, logical, l, csum);
+                       if (have_csum == 0)
+                               ++sdev->stat.no_csum;
+               }
+               ret = scrub_page(sdev, logical, l, physical, flags, gen,
+                                mirror_num, have_csum ? csum : NULL, 0);
+               if (ret)
+                       return ret;
+               len -= l;
+               logical += l;
+               physical += l;
+       }
+       return 0;
+ }
+ static noinline_for_stack int scrub_stripe(struct scrub_dev *sdev,
+       struct map_lookup *map, int num, u64 base, u64 length)
+ {
+       struct btrfs_path *path;
+       struct btrfs_fs_info *fs_info = sdev->dev->dev_root->fs_info;
+       struct btrfs_root *root = fs_info->extent_root;
+       struct btrfs_root *csum_root = fs_info->csum_root;
+       struct btrfs_extent_item *extent;
+       u64 flags;
+       int ret;
+       int slot;
+       int i;
+       u64 nstripes;
+       int start_stripe;
+       struct extent_buffer *l;
+       struct btrfs_key key;
+       u64 physical;
+       u64 logical;
+       u64 generation;
+       u64 mirror_num;
+       u64 increment = map->stripe_len;
+       u64 offset;
+       nstripes = length;
+       offset = 0;
+       do_div(nstripes, map->stripe_len);
+       if (map->type & BTRFS_BLOCK_GROUP_RAID0) {
+               offset = map->stripe_len * num;
+               increment = map->stripe_len * map->num_stripes;
+               mirror_num = 0;
+       } else if (map->type & BTRFS_BLOCK_GROUP_RAID10) {
+               int factor = map->num_stripes / map->sub_stripes;
+               offset = map->stripe_len * (num / map->sub_stripes);
+               increment = map->stripe_len * factor;
+               mirror_num = num % map->sub_stripes;
+       } else if (map->type & BTRFS_BLOCK_GROUP_RAID1) {
+               increment = map->stripe_len;
+               mirror_num = num % map->num_stripes;
+       } else if (map->type & BTRFS_BLOCK_GROUP_DUP) {
+               increment = map->stripe_len;
+               mirror_num = num % map->num_stripes;
+       } else {
+               increment = map->stripe_len;
+               mirror_num = 0;
+       }
+       path = btrfs_alloc_path();
+       if (!path)
+               return -ENOMEM;
+       path->reada = 2;
+       path->search_commit_root = 1;
+       path->skip_locking = 1;
+       /*
+        * find all extents for each stripe and just read them to get
+        * them into the page cache
+        * FIXME: we can do better. build a more intelligent prefetching
+        */
+       logical = base + offset;
+       physical = map->stripes[num].physical;
+       ret = 0;
+       for (i = 0; i < nstripes; ++i) {
+               key.objectid = logical;
+               key.type = BTRFS_EXTENT_ITEM_KEY;
+               key.offset = (u64)0;
+               ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+               if (ret < 0)
+                       goto out;
+               l = path->nodes[0];
+               slot = path->slots[0];
+               btrfs_item_key_to_cpu(l, &key, slot);
+               if (key.objectid != logical) {
+                       ret = btrfs_previous_item(root, path, 0,
+                                                 BTRFS_EXTENT_ITEM_KEY);
+                       if (ret < 0)
+                               goto out;
+               }
+               while (1) {
+                       l = path->nodes[0];
+                       slot = path->slots[0];
+                       if (slot >= btrfs_header_nritems(l)) {
+                               ret = btrfs_next_leaf(root, path);
+                               if (ret == 0)
+                                       continue;
+                               if (ret < 0)
+                                       goto out;
+                               break;
+                       }
+                       btrfs_item_key_to_cpu(l, &key, slot);
+                       if (key.objectid >= logical + map->stripe_len)
+                               break;
+                       path->slots[0]++;
+               }
 -              btrfs_release_path(root, path);
++              btrfs_release_path(path);
+               logical += increment;
+               physical += map->stripe_len;
+               cond_resched();
+       }
+       /*
+        * collect all data csums for the stripe to avoid seeking during
+        * the scrub. This might currently (crc32) end up to be about 1MB
+        */
+       start_stripe = 0;
+ again:
+       logical = base + offset + start_stripe * increment;
+       for (i = start_stripe; i < nstripes; ++i) {
+               ret = btrfs_lookup_csums_range(csum_root, logical,
+                                              logical + map->stripe_len - 1,
+                                              &sdev->csum_list, 1);
+               if (ret)
+                       goto out;
+               logical += increment;
+               cond_resched();
+       }
+       /*
+        * now find all extents for each stripe and scrub them
+        */
+       logical = base + offset + start_stripe * increment;
+       physical = map->stripes[num].physical + start_stripe * map->stripe_len;
+       ret = 0;
+       for (i = start_stripe; i < nstripes; ++i) {
+               /*
+                * canceled?
+                */
+               if (atomic_read(&fs_info->scrub_cancel_req) ||
+                   atomic_read(&sdev->cancel_req)) {
+                       ret = -ECANCELED;
+                       goto out;
+               }
+               /*
+                * check to see if we have to pause
+                */
+               if (atomic_read(&fs_info->scrub_pause_req)) {
+                       /* push queued extents */
+                       scrub_submit(sdev);
+                       wait_event(sdev->list_wait,
+                                  atomic_read(&sdev->in_flight) == 0);
+                       atomic_inc(&fs_info->scrubs_paused);
+                       wake_up(&fs_info->scrub_pause_wait);
+                       mutex_lock(&fs_info->scrub_lock);
+                       while (atomic_read(&fs_info->scrub_pause_req)) {
+                               mutex_unlock(&fs_info->scrub_lock);
+                               wait_event(fs_info->scrub_pause_wait,
+                                  atomic_read(&fs_info->scrub_pause_req) == 0);
+                               mutex_lock(&fs_info->scrub_lock);
+                       }
+                       atomic_dec(&fs_info->scrubs_paused);
+                       mutex_unlock(&fs_info->scrub_lock);
+                       wake_up(&fs_info->scrub_pause_wait);
+                       scrub_free_csums(sdev);
+                       start_stripe = i;
+                       goto again;
+               }
+               key.objectid = logical;
+               key.type = BTRFS_EXTENT_ITEM_KEY;
+               key.offset = (u64)0;
+               ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+               if (ret < 0)
+                       goto out;
+               l = path->nodes[0];
+               slot = path->slots[0];
+               btrfs_item_key_to_cpu(l, &key, slot);
+               if (key.objectid != logical) {
+                       ret = btrfs_previous_item(root, path, 0,
+                                                 BTRFS_EXTENT_ITEM_KEY);
+                       if (ret < 0)
+                               goto out;
+               }
+               while (1) {
+                       l = path->nodes[0];
+                       slot = path->slots[0];
+                       if (slot >= btrfs_header_nritems(l)) {
+                               ret = btrfs_next_leaf(root, path);
+                               if (ret == 0)
+                                       continue;
+                               if (ret < 0)
+                                       goto out;
+                               break;
+                       }
+                       btrfs_item_key_to_cpu(l, &key, slot);
+                       if (key.objectid + key.offset <= logical)
+                               goto next;
+                       if (key.objectid >= logical + map->stripe_len)
+                               break;
+                       if (btrfs_key_type(&key) != BTRFS_EXTENT_ITEM_KEY)
+                               goto next;
+                       extent = btrfs_item_ptr(l, slot,
+                                               struct btrfs_extent_item);
+                       flags = btrfs_extent_flags(l, extent);
+                       generation = btrfs_extent_generation(l, extent);
+                       if (key.objectid < logical &&
+                           (flags & BTRFS_EXTENT_FLAG_TREE_BLOCK)) {
+                               printk(KERN_ERR
+                                      "btrfs scrub: tree block %llu spanning "
+                                      "stripes, ignored. logical=%llu\n",
+                                      (unsigned long long)key.objectid,
+                                      (unsigned long long)logical);
+                               goto next;
+                       }
+                       /*
+                        * trim extent to this stripe
+                        */
+                       if (key.objectid < logical) {
+                               key.offset -= logical - key.objectid;
+                               key.objectid = logical;
+                       }
+                       if (key.objectid + key.offset >
+                           logical + map->stripe_len) {
+                               key.offset = logical + map->stripe_len -
+                                            key.objectid;
+                       }
+                       ret = scrub_extent(sdev, key.objectid, key.offset,
+                                          key.objectid - logical + physical,
+                                          flags, generation, mirror_num);
+                       if (ret)
+                               goto out;
+ next:
+                       path->slots[0]++;
+               }
 -                      btrfs_release_path(root, path);
++              btrfs_release_path(path);
+               logical += increment;
+               physical += map->stripe_len;
+               spin_lock(&sdev->stat_lock);
+               sdev->stat.last_physical = physical;
+               spin_unlock(&sdev->stat_lock);
+       }
+       /* push queued extents */
+       scrub_submit(sdev);
+ out:
+       btrfs_free_path(path);
+       return ret < 0 ? ret : 0;
+ }
+ static noinline_for_stack int scrub_chunk(struct scrub_dev *sdev,
+       u64 chunk_tree, u64 chunk_objectid, u64 chunk_offset, u64 length)
+ {
+       struct btrfs_mapping_tree *map_tree =
+               &sdev->dev->dev_root->fs_info->mapping_tree;
+       struct map_lookup *map;
+       struct extent_map *em;
+       int i;
+       int ret = -EINVAL;
+       read_lock(&map_tree->map_tree.lock);
+       em = lookup_extent_mapping(&map_tree->map_tree, chunk_offset, 1);
+       read_unlock(&map_tree->map_tree.lock);
+       if (!em)
+               return -EINVAL;
+       map = (struct map_lookup *)em->bdev;
+       if (em->start != chunk_offset)
+               goto out;
+       if (em->len < length)
+               goto out;
+       for (i = 0; i < map->num_stripes; ++i) {
+               if (map->stripes[i].dev == sdev->dev) {
+                       ret = scrub_stripe(sdev, map, i, chunk_offset, length);
+                       if (ret)
+                               goto out;
+               }
+       }
+ out:
+       free_extent_map(em);
+       return ret;
+ }
+ static noinline_for_stack
+ int scrub_enumerate_chunks(struct scrub_dev *sdev, u64 start, u64 end)
+ {
+       struct btrfs_dev_extent *dev_extent = NULL;
+       struct btrfs_path *path;
+       struct btrfs_root *root = sdev->dev->dev_root;
+       struct btrfs_fs_info *fs_info = root->fs_info;
+       u64 length;
+       u64 chunk_tree;
+       u64 chunk_objectid;
+       u64 chunk_offset;
+       int ret;
+       int slot;
+       struct extent_buffer *l;
+       struct btrfs_key key;
+       struct btrfs_key found_key;
+       struct btrfs_block_group_cache *cache;
+       path = btrfs_alloc_path();
+       if (!path)
+               return -ENOMEM;
+       path->reada = 2;
+       path->search_commit_root = 1;
+       path->skip_locking = 1;
+       key.objectid = sdev->dev->devid;
+       key.offset = 0ull;
+       key.type = BTRFS_DEV_EXTENT_KEY;
+       while (1) {
+               ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
+               if (ret < 0)
+                       goto out;
+               ret = 0;
+               l = path->nodes[0];
+               slot = path->slots[0];
+               btrfs_item_key_to_cpu(l, &found_key, slot);
+               if (found_key.objectid != sdev->dev->devid)
+                       break;
+               if (btrfs_key_type(&key) != BTRFS_DEV_EXTENT_KEY)
+                       break;
+               if (found_key.offset >= end)
+                       break;
+               if (found_key.offset < key.offset)
+                       break;
+               dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent);
+               length = btrfs_dev_extent_length(l, dev_extent);
+               if (found_key.offset + length <= start) {
+                       key.offset = found_key.offset + length;
 -              btrfs_release_path(root, path);
++                      btrfs_release_path(path);
+                       continue;
+               }
+               chunk_tree = btrfs_dev_extent_chunk_tree(l, dev_extent);
+               chunk_objectid = btrfs_dev_extent_chunk_objectid(l, dev_extent);
+               chunk_offset = btrfs_dev_extent_chunk_offset(l, dev_extent);
+               /*
+                * get a reference on the corresponding block group to prevent
+                * the chunk from going away while we scrub it
+                */
+               cache = btrfs_lookup_block_group(fs_info, chunk_offset);
+               if (!cache) {
+                       ret = -ENOENT;
+                       goto out;
+               }
+               ret = scrub_chunk(sdev, chunk_tree, chunk_objectid,
+                                 chunk_offset, length);
+               btrfs_put_block_group(cache);
+               if (ret)
+                       break;
+               key.offset = found_key.offset + length;
++              btrfs_release_path(path);
+       }
+ out:
+       btrfs_free_path(path);
+       return ret;
+ }
+ static noinline_for_stack int scrub_supers(struct scrub_dev *sdev)
+ {
+       int     i;
+       u64     bytenr;
+       u64     gen;
+       int     ret;
+       struct btrfs_device *device = sdev->dev;
+       struct btrfs_root *root = device->dev_root;
+       gen = root->fs_info->last_trans_committed;
+       for (i = 0; i < BTRFS_SUPER_MIRROR_MAX; i++) {
+               bytenr = btrfs_sb_offset(i);
+               if (bytenr + BTRFS_SUPER_INFO_SIZE >= device->total_bytes)
+                       break;
+               ret = scrub_page(sdev, bytenr, PAGE_SIZE, bytenr,
+                                BTRFS_EXTENT_FLAG_SUPER, gen, i, NULL, 1);
+               if (ret)
+                       return ret;
+       }
+       wait_event(sdev->list_wait, atomic_read(&sdev->in_flight) == 0);
+       return 0;
+ }
+ /*
+  * get a reference count on fs_info->scrub_workers. start worker if necessary
+  */
+ static noinline_for_stack int scrub_workers_get(struct btrfs_root *root)
+ {
+       struct btrfs_fs_info *fs_info = root->fs_info;
+       mutex_lock(&fs_info->scrub_lock);
+       if (fs_info->scrub_workers_refcnt == 0)
+               btrfs_start_workers(&fs_info->scrub_workers, 1);
+       ++fs_info->scrub_workers_refcnt;
+       mutex_unlock(&fs_info->scrub_lock);
+       return 0;
+ }
+ static noinline_for_stack void scrub_workers_put(struct btrfs_root *root)
+ {
+       struct btrfs_fs_info *fs_info = root->fs_info;
+       mutex_lock(&fs_info->scrub_lock);
+       if (--fs_info->scrub_workers_refcnt == 0)
+               btrfs_stop_workers(&fs_info->scrub_workers);
+       WARN_ON(fs_info->scrub_workers_refcnt < 0);
+       mutex_unlock(&fs_info->scrub_lock);
+ }
+ int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end,
+                   struct btrfs_scrub_progress *progress, int readonly)
+ {
+       struct scrub_dev *sdev;
+       struct btrfs_fs_info *fs_info = root->fs_info;
+       int ret;
+       struct btrfs_device *dev;
+       if (root->fs_info->closing)
+               return -EINVAL;
+       /*
+        * check some assumptions
+        */
+       if (root->sectorsize != PAGE_SIZE ||
+           root->sectorsize != root->leafsize ||
+           root->sectorsize != root->nodesize) {
+               printk(KERN_ERR "btrfs_scrub: size assumptions fail\n");
+               return -EINVAL;
+       }
+       ret = scrub_workers_get(root);
+       if (ret)
+               return ret;
+       mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
+       dev = btrfs_find_device(root, devid, NULL, NULL);
+       if (!dev || dev->missing) {
+               mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
+               scrub_workers_put(root);
+               return -ENODEV;
+       }
+       mutex_lock(&fs_info->scrub_lock);
+       if (!dev->in_fs_metadata) {
+               mutex_unlock(&fs_info->scrub_lock);
+               mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
+               scrub_workers_put(root);
+               return -ENODEV;
+       }
+       if (dev->scrub_device) {
+               mutex_unlock(&fs_info->scrub_lock);
+               mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
+               scrub_workers_put(root);
+               return -EINPROGRESS;
+       }
+       sdev = scrub_setup_dev(dev);
+       if (IS_ERR(sdev)) {
+               mutex_unlock(&fs_info->scrub_lock);
+               mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
+               scrub_workers_put(root);
+               return PTR_ERR(sdev);
+       }
+       sdev->readonly = readonly;
+       dev->scrub_device = sdev;
+       atomic_inc(&fs_info->scrubs_running);
+       mutex_unlock(&fs_info->scrub_lock);
+       mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
+       down_read(&fs_info->scrub_super_lock);
+       ret = scrub_supers(sdev);
+       up_read(&fs_info->scrub_super_lock);
+       if (!ret)
+               ret = scrub_enumerate_chunks(sdev, start, end);
+       wait_event(sdev->list_wait, atomic_read(&sdev->in_flight) == 0);
+       atomic_dec(&fs_info->scrubs_running);
+       wake_up(&fs_info->scrub_pause_wait);
+       if (progress)
+               memcpy(progress, &sdev->stat, sizeof(*progress));
+       mutex_lock(&fs_info->scrub_lock);
+       dev->scrub_device = NULL;
+       mutex_unlock(&fs_info->scrub_lock);
+       scrub_free_dev(sdev);
+       scrub_workers_put(root);
+       return ret;
+ }
+ int btrfs_scrub_pause(struct btrfs_root *root)
+ {
+       struct btrfs_fs_info *fs_info = root->fs_info;
+       mutex_lock(&fs_info->scrub_lock);
+       atomic_inc(&fs_info->scrub_pause_req);
+       while (atomic_read(&fs_info->scrubs_paused) !=
+              atomic_read(&fs_info->scrubs_running)) {
+               mutex_unlock(&fs_info->scrub_lock);
+               wait_event(fs_info->scrub_pause_wait,
+                          atomic_read(&fs_info->scrubs_paused) ==
+                          atomic_read(&fs_info->scrubs_running));
+               mutex_lock(&fs_info->scrub_lock);
+       }
+       mutex_unlock(&fs_info->scrub_lock);
+       return 0;
+ }
+ int btrfs_scrub_continue(struct btrfs_root *root)
+ {
+       struct btrfs_fs_info *fs_info = root->fs_info;
+       atomic_dec(&fs_info->scrub_pause_req);
+       wake_up(&fs_info->scrub_pause_wait);
+       return 0;
+ }
+ int btrfs_scrub_pause_super(struct btrfs_root *root)
+ {
+       down_write(&root->fs_info->scrub_super_lock);
+       return 0;
+ }
+ int btrfs_scrub_continue_super(struct btrfs_root *root)
+ {
+       up_write(&root->fs_info->scrub_super_lock);
+       return 0;
+ }
+ int btrfs_scrub_cancel(struct btrfs_root *root)
+ {
+       struct btrfs_fs_info *fs_info = root->fs_info;
+       mutex_lock(&fs_info->scrub_lock);
+       if (!atomic_read(&fs_info->scrubs_running)) {
+               mutex_unlock(&fs_info->scrub_lock);
+               return -ENOTCONN;
+       }
+       atomic_inc(&fs_info->scrub_cancel_req);
+       while (atomic_read(&fs_info->scrubs_running)) {
+               mutex_unlock(&fs_info->scrub_lock);
+               wait_event(fs_info->scrub_pause_wait,
+                          atomic_read(&fs_info->scrubs_running) == 0);
+               mutex_lock(&fs_info->scrub_lock);
+       }
+       atomic_dec(&fs_info->scrub_cancel_req);
+       mutex_unlock(&fs_info->scrub_lock);
+       return 0;
+ }
+ int btrfs_scrub_cancel_dev(struct btrfs_root *root, struct btrfs_device *dev)
+ {
+       struct btrfs_fs_info *fs_info = root->fs_info;
+       struct scrub_dev *sdev;
+       mutex_lock(&fs_info->scrub_lock);
+       sdev = dev->scrub_device;
+       if (!sdev) {
+               mutex_unlock(&fs_info->scrub_lock);
+               return -ENOTCONN;
+       }
+       atomic_inc(&sdev->cancel_req);
+       while (dev->scrub_device) {
+               mutex_unlock(&fs_info->scrub_lock);
+               wait_event(fs_info->scrub_pause_wait,
+                          dev->scrub_device == NULL);
+               mutex_lock(&fs_info->scrub_lock);
+       }
+       mutex_unlock(&fs_info->scrub_lock);
+       return 0;
+ }
+ int btrfs_scrub_cancel_devid(struct btrfs_root *root, u64 devid)
+ {
+       struct btrfs_fs_info *fs_info = root->fs_info;
+       struct btrfs_device *dev;
+       int ret;
+       /*
+        * we have to hold the device_list_mutex here so the device
+        * does not go away in cancel_dev. FIXME: find a better solution
+        */
+       mutex_lock(&fs_info->fs_devices->device_list_mutex);
+       dev = btrfs_find_device(root, devid, NULL, NULL);
+       if (!dev) {
+               mutex_unlock(&fs_info->fs_devices->device_list_mutex);
+               return -ENODEV;
+       }
+       ret = btrfs_scrub_cancel_dev(root, dev);
+       mutex_unlock(&fs_info->fs_devices->device_list_mutex);
+       return ret;
+ }
+ int btrfs_scrub_progress(struct btrfs_root *root, u64 devid,
+                        struct btrfs_scrub_progress *progress)
+ {
+       struct btrfs_device *dev;
+       struct scrub_dev *sdev = NULL;
+       mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
+       dev = btrfs_find_device(root, devid, NULL, NULL);
+       if (dev)
+               sdev = dev->scrub_device;
+       if (sdev)
+               memcpy(progress, &sdev->stat, sizeof(*progress));
+       mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
+       return dev ? (sdev ? 0 : -ENOTCONN) : -ENODEV;
+ }
Simple merge
Simple merge
Simple merge
index 05d5d199381adbfb3e65b69a60558f5eedafec8b,f7c20123a1fe173187e3820a963602c3c82c30ac..7c58a7b2d99f0b64407903ecac7665682a5e8b71
@@@ -158,6 -160,24 +161,9 @@@ struct map_lookup 
        struct btrfs_bio_stripe stripes[];
  };
  
 -/* Used to sort the devices by max_avail(descending sort) */
 -int btrfs_cmp_device_free_bytes(const void *dev_info1, const void *dev_info2);
 -
 -/*
 - * sort the devices by max_avail, in which max free extent size of each device
 - * is stored.(Descending Sort)
 - */
 -static inline void btrfs_descending_sort_devices(
 -                                      struct btrfs_device_info *devices,
 -                                      size_t nr_devices)
 -{
 -      sort(devices, nr_devices, sizeof(struct btrfs_device_info),
 -           btrfs_cmp_device_free_bytes, NULL);
 -}
 -
+ #define map_lookup_size(n) (sizeof(struct map_lookup) + \
+                           (sizeof(struct btrfs_bio_stripe) * (n)))
  int btrfs_account_dev_extents_size(struct btrfs_device *device, u64 start,
                                   u64 end, u64 *length);