Re: [PATCH v2 4/6] btrfs: Introduce mount time chunk <-> dev extent mapping check

[Qu Wenruo <quwenruo.btrfs@gmx.com>]

[-- Attachment #1.1: Type: text/plain, Size: 11973 bytes --]



On 2019/1/14 下午7:09, Filipe Manana wrote:
> On Wed, Aug 1, 2018 at 3:39 AM Qu Wenruo <wqu@suse.com> wrote:
>>
>> This patch will introduce chunk <-> dev extent mapping check, to protect
>> us against invalid dev extents or chunks.
>>
>> Since chunk mapping is the fundamental infrastructure of btrfs, extra
>> check at mount time could prevent a lot of unexpected behavior (BUG_ON).
>>
>> Reported-by: Xu Wen <wen.xu@gatech.edu>
>> Link: https://bugzilla.kernel.org/show_bug.cgi?id=200403
>> Link: https://bugzilla.kernel.org/show_bug.cgi?id=200407
>> Signed-off-by: Qu Wenruo <wqu@suse.com>
> 
> Btw, this makes at least one test case from btrfs-progs fail:
> 
> root 17:12:02 /home/fdmanana/git/hub/btrfs-progs/tests ((v4.19.1))>
> TEST=021\* ./misc-tests.sh
>     [TEST/misc]   021-image-multi-devices
> failed: mount /dev/loop2 /home/fdmanana/git/hub/btrfs-progs/tests//mnt
> test failed for case 021-image-multi-devices

That is fixed by the following commits already in devel:
9996feb94d btrfs-progs: misc-tests/021: Do extra btrfs check before mounting
a1a98ee7a8 btrfs-progs: image: Remove all existing dev extents for later
rebuild
e6c1fa297a btrfs-progs: volumes: Refactor btrfs_alloc_dev_extent() into
two functions
9a65b425bb btrfs-progs: image: Fix block group item flags when restoring
multi-device image to single device
ca73162b48 btrfs-progs: image: Refactor fixup_devices() to
fixup_chunks_and_devices()

And they are pretty early detected and merged, just after v4.19.1.

Thanks,
Qu





> 
> dmesg/syslog has:
> 
> [432229.206699] BTRFS error (device loop0): dev extent physical offset
> 22020096 devid 1 has no corresponding chunk
> [432229.207497] BTRFS error (device loop0): failed to find devid 1
> [432229.208281] BTRFS error (device loop0): failed to verify dev
> extents against chunks: -117
> [432229.246286] BTRFS error (device loop0): open_ctree failed
> 
> Thanks.
> 
>> ---
>>  fs/btrfs/disk-io.c |   7 ++
>>  fs/btrfs/volumes.c | 183 +++++++++++++++++++++++++++++++++++++++++++++
>>  fs/btrfs/volumes.h |   2 +
>>  3 files changed, 192 insertions(+)
>>
>> diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
>> index 205092dc9390..068ca7498e94 100644
>> --- a/fs/btrfs/disk-io.c
>> +++ b/fs/btrfs/disk-io.c
>> @@ -3075,6 +3075,13 @@ int open_ctree(struct super_block *sb,
>>         fs_info->generation = generation;
>>         fs_info->last_trans_committed = generation;
>>
>> +       ret = btrfs_verify_dev_extents(fs_info);
>> +       if (ret) {
>> +               btrfs_err(fs_info,
>> +                         "failed to verify dev extents against chunks: %d",
>> +                         ret);
>> +               goto fail_block_groups;
>> +       }
>>         ret = btrfs_recover_balance(fs_info);
>>         if (ret) {
>>                 btrfs_err(fs_info, "failed to recover balance: %d", ret);
>> diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
>> index e6a8e4aabc66..467a589854fa 100644
>> --- a/fs/btrfs/volumes.c
>> +++ b/fs/btrfs/volumes.c
>> @@ -6440,6 +6440,7 @@ static int read_one_chunk(struct btrfs_fs_info *fs_info, struct btrfs_key *key,
>>         map->stripe_len = btrfs_chunk_stripe_len(leaf, chunk);
>>         map->type = btrfs_chunk_type(leaf, chunk);
>>         map->sub_stripes = btrfs_chunk_sub_stripes(leaf, chunk);
>> +       map->verified_stripes = 0;
>>         for (i = 0; i < num_stripes; i++) {
>>                 map->stripes[i].physical =
>>                         btrfs_stripe_offset_nr(leaf, chunk, i);
>> @@ -7295,3 +7296,185 @@ void btrfs_reset_fs_info_ptr(struct btrfs_fs_info *fs_info)
>>                 fs_devices = fs_devices->seed;
>>         }
>>  }
>> +
>> +static u64 calc_stripe_length(u64 type, u64 chunk_len, int num_stripes)
>> +{
>> +       int index = btrfs_bg_flags_to_raid_index(type);
>> +       int ncopies = btrfs_raid_array[index].ncopies;
>> +       int data_stripes;
>> +
>> +       switch (type & BTRFS_BLOCK_GROUP_PROFILE_MASK) {
>> +       case BTRFS_BLOCK_GROUP_RAID5:
>> +               data_stripes = num_stripes - 1;
>> +               break;
>> +       case BTRFS_BLOCK_GROUP_RAID6:
>> +               data_stripes = num_stripes - 2;
>> +               break;
>> +       default:
>> +               data_stripes = num_stripes / ncopies;
>> +               break;
>> +       }
>> +       return div_u64(chunk_len, data_stripes);
>> +}
>> +static int verify_one_dev_extent(struct btrfs_fs_info *fs_info,
>> +                                u64 chunk_offset, u64 devid,
>> +                                u64 physical_offset, u64 physical_len)
>> +{
>> +       struct extent_map_tree *em_tree = &fs_info->mapping_tree.map_tree;
>> +       struct extent_map *em;
>> +       struct map_lookup *map;
>> +       u64 stripe_len;
>> +       bool found = false;
>> +       int ret = 0;
>> +       int i;
>> +
>> +       read_lock(&em_tree->lock);
>> +       em = lookup_extent_mapping(em_tree, chunk_offset, 1);
>> +       read_unlock(&em_tree->lock);
>> +
>> +       if (!em) {
>> +               ret = -EUCLEAN;
>> +               btrfs_err(fs_info,
>> +               "dev extent (%llu, %llu) doesn't have corresponding chunk",
>> +                         devid, physical_offset);
>> +               goto out;
>> +       }
>> +
>> +       map = em->map_lookup;
>> +       stripe_len = calc_stripe_length(map->type, em->len, map->num_stripes);
>> +       if (physical_len != stripe_len) {
>> +               btrfs_err(fs_info,
>> +"dev extent (%llu, %llu) length doesn't match with chunk %llu, have %llu expect %llu",
>> +                         devid, physical_offset, em->start, physical_len,
>> +                         stripe_len);
>> +               ret = -EUCLEAN;
>> +               goto out;
>> +       }
>> +
>> +       for (i = 0; i < map->num_stripes; i++) {
>> +               if (map->stripes[i].dev->devid == devid &&
>> +                   map->stripes[i].physical == physical_offset) {
>> +                       found = true;
>> +                       if (map->verified_stripes >= map->num_stripes) {
>> +                               btrfs_err(fs_info,
>> +                       "too many dev extent for chunk %llu is detected",
>> +                                         em->start);
>> +                               ret = -EUCLEAN;
>> +                               goto out;
>> +                       }
>> +                       map->verified_stripes++;
>> +                       break;
>> +               }
>> +       }
>> +       if (!found) {
>> +               ret = -EUCLEAN;
>> +               btrfs_err(fs_info,
>> +                       "dev extent (%llu, %llu) has no corresponding chunk",
>> +                       devid, physical_offset);
>> +       }
>> +out:
>> +       free_extent_map(em);
>> +       return ret;
>> +}
>> +
>> +static int verify_chunk_dev_extent_mapping(struct btrfs_fs_info *fs_info)
>> +{
>> +       struct extent_map_tree *em_tree = &fs_info->mapping_tree.map_tree;
>> +       struct extent_map *em;
>> +       struct rb_node *node;
>> +       int ret = 0;
>> +
>> +       read_lock(&em_tree->lock);
>> +       for (node = rb_first(&em_tree->map); node; node = rb_next(node)) {
>> +               em = rb_entry(node, struct extent_map, rb_node);
>> +               if (em->map_lookup->num_stripes !=
>> +                   em->map_lookup->verified_stripes) {
>> +                       btrfs_err(fs_info,
>> +                       "chunk %llu has missing dev extent, have %d expect %d",
>> +                                 em->start, em->map_lookup->verified_stripes,
>> +                                 em->map_lookup->num_stripes);
>> +                       ret = -EUCLEAN;
>> +                       goto out;
>> +               }
>> +       }
>> +out:
>> +       read_unlock(&em_tree->lock);
>> +       return ret;
>> +}
>> +
>> +/*
>> + * Ensure all dev extents are mapped to correct chunk.
>> + * Or later chunk allocation/free would cause unexpected behavior.
>> + *
>> + * NOTE: This will iterate through the whole device tree, which should be
>> + * at the same size level of chunk tree.
>> + * This would increase mount time by a tiny fraction.
>> + */
>> +int btrfs_verify_dev_extents(struct btrfs_fs_info *fs_info)
>> +{
>> +       struct btrfs_path *path;
>> +       struct btrfs_root *root = fs_info->dev_root;
>> +       struct btrfs_key key;
>> +       int ret = 0;
>> +
>> +       key.objectid = 1;
>> +       key.type = BTRFS_DEV_EXTENT_KEY;
>> +       key.offset = 0;
>> +
>> +       path = btrfs_alloc_path();
>> +       if (!path)
>> +               return -ENOMEM;
>> +
>> +       path->reada = READA_FORWARD;
>> +       ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
>> +       if (ret < 0)
>> +               goto out;
>> +
>> +       if (path->slots[0] >= btrfs_header_nritems(path->nodes[0])) {
>> +               ret = btrfs_next_item(root, path);
>> +               if (ret < 0)
>> +                       goto out;
>> +               /* No dev extents at all? Not good */
>> +               if (ret > 0) {
>> +                       ret = -EUCLEAN;
>> +                       goto out;
>> +               }
>> +       }
>> +       while (1) {
>> +               struct extent_buffer *leaf = path->nodes[0];
>> +               struct btrfs_dev_extent *dext;
>> +               int slot = path->slots[0];
>> +               u64 chunk_offset;
>> +               u64 physical_offset;
>> +               u64 physical_len;
>> +               u64 devid;
>> +
>> +               btrfs_item_key_to_cpu(leaf, &key, slot);
>> +               if (key.type != BTRFS_DEV_EXTENT_KEY)
>> +                       break;
>> +               devid = key.objectid;
>> +               physical_offset = key.offset;
>> +
>> +               dext = btrfs_item_ptr(leaf, slot, struct btrfs_dev_extent);
>> +               chunk_offset = btrfs_dev_extent_chunk_offset(leaf, dext);
>> +               physical_len = btrfs_dev_extent_length(leaf, dext);
>> +
>> +               ret = verify_one_dev_extent(fs_info, chunk_offset, devid,
>> +                                           physical_offset, physical_len);
>> +               if (ret < 0)
>> +                       goto out;
>> +               ret = btrfs_next_item(root, path);
>> +               if (ret < 0)
>> +                       goto out;
>> +               if (ret > 0) {
>> +                       ret = 0;
>> +                       break;
>> +               }
>> +       }
>> +
>> +       /* Ensure all chunks have corresponding dev extents */
>> +       ret = verify_chunk_dev_extent_mapping(fs_info);
>> +out:
>> +       btrfs_free_path(path);
>> +       return ret;
>> +}
>> diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h
>> index 6d4f38ad9f5c..4301bf2d0534 100644
>> --- a/fs/btrfs/volumes.h
>> +++ b/fs/btrfs/volumes.h
>> @@ -345,6 +345,7 @@ struct map_lookup {
>>         u64 stripe_len;
>>         int num_stripes;
>>         int sub_stripes;
>> +       int verified_stripes; /* For mount time dev extent verification */
>>         struct btrfs_bio_stripe stripes[];
>>  };
>>
>> @@ -559,5 +560,6 @@ void btrfs_set_fs_info_ptr(struct btrfs_fs_info *fs_info);
>>  void btrfs_reset_fs_info_ptr(struct btrfs_fs_info *fs_info);
>>  bool btrfs_check_rw_degradable(struct btrfs_fs_info *fs_info,
>>                                         struct btrfs_device *failing_dev);
>> +int btrfs_verify_dev_extents(struct btrfs_fs_info *fs_info);
>>
>>  #endif
>> --
>> 2.18.0
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 
> 
> 


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]