Remap tree
v8 cover Remap tree
This is version 8 of the patch series for the new logical remapping tree feature - see the previous cover letters for more information including the rationale:
* RFC: https://lore.kernel.org/all/[email protected]/
* Version 1: https://lore.kernel.org/all/[email protected]/
* Version 2: https://lore.kernel.org/all/[email protected]/
* Version 3: https://lore.kernel.org/all/[email protected]/
* Version 4: https://lore.kernel.org/all/[email protected]/
* Version 5: https://lore.kernel.org/all/[email protected]/
* Version 6: https://lore.kernel.org/all/[email protected]/
* Version 7: https://lore.kernel.org/all/[email protected]/
Changes since version 7:
* renamed struct btrfs_remap to struct btrfs_remap_item
* renamed BTRFS_BLOCK_GROUP_FLAGS_REMAP to BTRFS_BLOCK_GROUP_FLAGS_METADATA_REMAP
* added unlikelies
* renamed new commit_* fields in struct btrfs_block_group to last_*, and added new patch renaming existing commit_used to last_used to match
* merged do_copy() into copy_remapped_data()
* initialized on-stack struct btrfs_remap_items
* fixed comments
* added other minor changes as suggested by David Sterba
Mark Harmstone (17):
btrfs: add definitions and constants for remap-tree
btrfs: add METADATA_REMAP chunk type
btrfs: allow remapped chunks to have zero stripes
btrfs: remove remapped block groups from the free-space tree
btrfs: don't add metadata items for the remap tree to the extent tree
btrfs: rename struct btrfs_block_group field commit_used to last_used
btrfs: add extended version of struct block_group_item
btrfs: allow mounting filesystems with remap-tree incompat flag
btrfs: redirect I/O for remapped block groups
btrfs: handle deletions from remapped block group
btrfs: handle setting up relocation of block group with remap-tree
btrfs: move existing remaps before relocating block group
btrfs: replace identity remaps with actual remaps when doing
relocations
btrfs: add do_remap param to btrfs_discard_extent()
btrfs: allow balancing remap tree
btrfs: handle discarding fully-remapped block groups
btrfs: populate fully_remapped_bgs_list on mount
fs/btrfs/Kconfig | 2 + fs/btrfs/accessors.h | 30 + fs/btrfs/bio.c | 3 +- fs/btrfs/bio.h | 3 + fs/btrfs/block-group.c | 323 ++++-- fs/btrfs/block-group.h | 29 +- fs/btrfs/block-rsv.c | 9 + fs/btrfs/block-rsv.h | 1 + fs/btrfs/discard.c | 57 +- fs/btrfs/disk-io.c | 130 ++- fs/btrfs/extent-tree.c | 151 ++- fs/btrfs/extent-tree.h | 4 +- fs/btrfs/free-space-cache.c | 59 +- fs/btrfs/free-space-cache.h | 1 + fs/btrfs/free-space-tree.c | 4 +- fs/btrfs/free-space-tree.h | 5 +- fs/btrfs/fs.h | 10 +- fs/btrfs/inode.c | 2 +- fs/btrfs/locking.c | 1 + fs/btrfs/relocation.c | 1885 +++++++++++++++++++++++++++++-- fs/btrfs/relocation.h | 18 + fs/btrfs/space-info.c | 22 +- fs/btrfs/sysfs.c | 4 + fs/btrfs/transaction.c | 7 + fs/btrfs/tree-checker.c | 94 +- fs/btrfs/tree-checker.h | 5 + fs/btrfs/volumes.c | 355 +++++- fs/btrfs/volumes.h | 18 +- include/uapi/linux/btrfs.h | 1 + include/uapi/linux/btrfs_tree.h | 34 +- 30 files changed, 2991 insertions(+), 276 deletions(-)
Signature · 1 line
2.51.2
PATCH v8 1/17 btrfs: add definitions and constants for remap-tree
Add an incompat flag for the new remap-tree feature, and the constants and definitions needed to support it.
--- fs/btrfs/accessors.h | 4 ++++ fs/btrfs/locking.c | 1 + fs/btrfs/sysfs.c | 2 ++ fs/btrfs/tree-checker.c | 6 ++---- fs/btrfs/tree-checker.h | 5 +++++ fs/btrfs/volumes.c | 1 + include/uapi/linux/btrfs.h | 1 + include/uapi/linux/btrfs_tree.h | 17 +++++++++++++++++ 8 files changed, 33 insertions(+), 4 deletions(-)
fs/btrfs/accessors.h +4 -0
@@ -1010,6 +1010,10 @@ BTRFS_SETGET_STACK_FUNCS(stack_verity_descriptor_encryption,BTRFS_SETGET_STACK_FUNCS (stack_verity_descriptor_size, struct btrfs_verity_descriptor_item, size, 64); +BTRFS_SETGET_FUNCS(remap_address, struct btrfs_remap_item, address, 64); +BTRFS_SETGET_STACK_FUNCS(stack_remap_address, struct btrfs_remap_item, + address,64 ); +/* Cast into the data area of the leaf. */ #define btrfs_item_ptr (leaf, slot, type) \ ((type *)(btrfs_item_nr_offset(leaf, 0) + btrfs_item_offset(leaf, slot)))
fs/btrfs/locking.c +1 -0
@@ -73,6 +73,7 @@ static struct btrfs_lockdep_keyset { { .id = BTRFS_FREE_SPACE_TREE_OBJECTID, DEFINE_NAME("free-space") }, { .id = BTRFS_BLOCK_GROUP_TREE_OBJECTID, DEFINE_NAME("block-group") }, { .id = BTRFS_RAID_STRIPE_TREE_OBJECTID, DEFINE_NAME("raid-stripe") }, + { .id = BTRFS_REMAP_TREE_OBJECTID, DEFINE_NAME("remap") }, { .id = 0, DEFINE_NAME("tree") }, };
fs/btrfs/sysfs.c +2 -0
@@ -291,6 +291,7 @@ BTRFS_FEAT_ATTR_COMPAT_RO(free_space_tree, FREE_SPACE_TREE);BTRFS_FEAT_ATTR_COMPAT_RO (block_group_tree,BLOCK_GROUP_TREE );BTRFS_FEAT_ATTR_INCOMPAT (raid1c34,RAID1C34 );BTRFS_FEAT_ATTR_INCOMPAT (simple_quota,SIMPLE_QUOTA ); +BTRFS_FEAT_ATTR_INCOMPAT (remap_tree,REMAP_TREE );#ifdef CONFIG_BLK_DEV_ZONED BTRFS_FEAT_ATTR_INCOMPAT (zoned,ZONED );#endif
@@ -331,6 +332,7 @@ static struct attribute *btrfs_supported_feature_attrs[] = {#ifdef CONFIG_BTRFS_EXPERIMENTAL BTRFS_FEAT_ATTR_PTR (extent_tree_v2),BTRFS_FEAT_ATTR_PTR (raid_stripe_tree), +BTRFS_FEAT_ATTR_PTR (remap_tree), #endif #ifdef CONFIG_FS_VERITY BTRFS_FEAT_ATTR_PTR(verity),
fs/btrfs/tree-checker.c +2 -4
@@ -913,12 +913,10 @@ int btrfs_check_chunk_valid(const struct btrfs_fs_info *fs_info, length, btrfs_stripe_nr_to_offset(U32_MAX));return -EUCLEAN ; } -if (unlikely (type & ~(BTRFS_BLOCK_GROUP_TYPE_MASK | -BTRFS_BLOCK_GROUP_PROFILE_MASK ))) { +if (unlikely (type & ~BTRFS_BLOCK_GROUP_VALID )) {chunk_err (fs_info, leaf, chunk, logical,"unrecognized chunk type: 0x%llx" , - ~(BTRFS_BLOCK_GROUP_TYPE_MASK | -BTRFS_BLOCK_GROUP_PROFILE_MASK ) & type); + type & ~BTRFS_BLOCK_GROUP_VALID );return -EUCLEAN ; }
fs/btrfs/tree-checker.h +5 -0
@@ -57,6 +57,11 @@ enum btrfs_tree_block_status { BTRFS_TREE_BLOCK_WRITTEN_NOT_SET, }; + +#define BTRFS_BLOCK_GROUP_VALID (BTRFS_BLOCK_GROUP_TYPE_MASK | \ + BTRFS_BLOCK_GROUP_PROFILE_MASK | \ + BTRFS_BLOCK_GROUP_REMAPPED) +/* * Exported simply for btrfs-progs which wants to have the * btrfs_tree_block_status return codes.
fs/btrfs/volumes.c +1 -0
@@ -231,6 +231,7 @@ void btrfs_describe_block_groups(u64 bg_flags, char *buf, u32 size_buf)DESCRIBE_FLAG (BTRFS_BLOCK_GROUP_DATA ,"data" );DESCRIBE_FLAG (BTRFS_BLOCK_GROUP_SYSTEM ,"system" );DESCRIBE_FLAG (BTRFS_BLOCK_GROUP_METADATA ,"metadata" ); +DESCRIBE_FLAG (BTRFS_BLOCK_GROUP_REMAPPED ,"remapped" );DESCRIBE_FLAG (BTRFS_AVAIL_ALLOC_BIT_SINGLE ,"single" );for (i =0 ; i <BTRFS_NR_RAID_TYPES ; i++)
include/uapi/linux/btrfs.h +1 -0
@@ -336,6 +336,7 @@ struct btrfs_ioctl_fs_info_args {#define BTRFS_FEATURE_INCOMPAT_EXTENT_TREE_V2 (1ULL << 13)#define BTRFS_FEATURE_INCOMPAT_RAID_STRIPE_TREE (1ULL << 14)#define BTRFS_FEATURE_INCOMPAT_SIMPLE_QUOTA (1ULL << 16) +#define BTRFS_FEATURE_INCOMPAT_REMAP_TREE (1ULL << 17)struct btrfs_ioctl_feature_flags {__u64 compat_flags;
include/uapi/linux/btrfs_tree.h +17 -0
@@ -76,6 +76,9 @@/* Tracks RAID stripes in block groups. */ #define BTRFS_RAID_STRIPE_TREE_OBJECTID 12ULL +/* Holds details of remapped addresses after relocation. */ +#define BTRFS_REMAP_TREE_OBJECTID 13ULL +/* device stats in the device tree */ #define BTRFS_DEV_STATS_OBJECTID 0ULL
@@ -282,6 +285,10 @@#define BTRFS_RAID_STRIPE_KEY 230 +#define BTRFS_IDENTITY_REMAP_KEY 234 +#define BTRFS_REMAP_KEY 235 +#define BTRFS_REMAP_BACKREF_KEY 236 +/* * Records the overall state of the qgroups. * There's only one instance of this key present,
@@ -1161,6 +1168,7 @@ struct btrfs_dev_replace_item {#define BTRFS_BLOCK_GROUP_RAID6 (1ULL << 8)#define BTRFS_BLOCK_GROUP_RAID1C3 (1ULL << 9)#define BTRFS_BLOCK_GROUP_RAID1C4 (1ULL << 10) +#define BTRFS_BLOCK_GROUP_REMAPPED (1ULL << 11)#define BTRFS_BLOCK_GROUP_RESERVED (BTRFS_AVAIL_ALLOC_BIT_SINGLE | \ BTRFS_SPACE_INFO_GLOBAL_RSV)
@@ -1323,4 +1331,13 @@ struct btrfs_verity_descriptor_item {__u8 encryption; } __attribute__ ((__packed__)); +/* +* For a range identified by a BTRFS_REMAP_KEY item in the remap tree, gives +* the address that the start of the range will get remapped to. This +* structure is also shared by BTRFS_REMAP_BACKREF_KEY. +*/ +struct btrfs_remap_item { +__le64 address; +} __attribute__ ((__packed__ )); +#endif /* _BTRFS_CTREE_H_ */
Signature · 1 line
2.51.2
PATCH v8 2/17 btrfs: add METADATA_REMAP chunk type
In reply to Mark HarmstoneAdd a new METADATA_REMAP chunk type, which is a metadata chunk that holds the remap tree.
This is needed for bootstrapping purposes: the remap tree can't itself be remapped, and must be relocated the existing way, by COWing every leaf. The remap tree can't go in the SYSTEM chunk as space there is limited, because a copy of the chunk item gets placed in the superblock.
The changes in fs/btrfs/volumes.h are because we're adding a new block group type bit after the profile bits, and so can no longer rely on the const_ilog2 trick.
The sizing to 32MB per chunk, matching the SYSTEM chunk, is an estimate here, we can adjust it later if it proves to be too big or too small. This works out to be ~500,000 remap items, which for a 4KB block size covers ~2GB of remapped data in the worst case and ~500TB in the best case.
--- fs/btrfs/block-rsv.c | 9 +++++++++ fs/btrfs/block-rsv.h | 1 + fs/btrfs/disk-io.c | 1 + fs/btrfs/fs.h | 2 ++ fs/btrfs/space-info.c | 13 ++++++++++++- fs/btrfs/sysfs.c | 2 ++ fs/btrfs/tree-checker.c | 13 +++++++++++-- fs/btrfs/volumes.c | 3 +++ fs/btrfs/volumes.h | 10 +++++++++- include/uapi/linux/btrfs_tree.h | 4 +++- 10 files changed, 53 insertions(+), 5 deletions(-)
fs/btrfs/block-rsv.c +9 -0
@@ -419,6 +419,9 @@ void btrfs_init_root_block_rsv(struct btrfs_root *root)case BTRFS_TREE_LOG_OBJECTID : root->block_rsv = &fs_info->treelog_rsv;break ; +case BTRFS_REMAP_TREE_OBJECTID : + root->block_rsv = &fs_info->remap_block_rsv; +break ;default : root->block_rsv =NULL ;break ;
@@ -432,6 +435,10 @@ void btrfs_init_global_block_rsv(struct btrfs_fs_info *fs_info) space_info =btrfs_find_space_info (fs_info,BTRFS_BLOCK_GROUP_SYSTEM ); fs_info->chunk_block_rsv.space_info = space_info; + space_info =btrfs_find_space_info (fs_info, +BTRFS_BLOCK_GROUP_METADATA_REMAP ); + fs_info->remap_block_rsv.space_info = space_info; + space_info =btrfs_find_space_info (fs_info,BTRFS_BLOCK_GROUP_METADATA ); fs_info->global_block_rsv.space_info = space_info; fs_info->trans_block_rsv.space_info = space_info;
@@ -458,6 +465,8 @@ void btrfs_release_global_block_rsv(struct btrfs_fs_info *fs_info)WARN_ON (fs_info->trans_block_rsv.reserved >0 );WARN_ON (fs_info->chunk_block_rsv.size >0 );WARN_ON (fs_info->chunk_block_rsv.reserved >0 ); +WARN_ON (fs_info->remap_block_rsv.size >0 ); +WARN_ON (fs_info->remap_block_rsv.reserved >0 );WARN_ON (fs_info->delayed_block_rsv.size >0 );WARN_ON (fs_info->delayed_block_rsv.reserved >0 );WARN_ON (fs_info->delayed_refs_rsv.reserved >0 );
fs/btrfs/block-rsv.h +1 -0
@@ -22,6 +22,7 @@ enum btrfs_rsv_type {BTRFS_BLOCK_RSV_DELALLOC ,BTRFS_BLOCK_RSV_TRANS ,BTRFS_BLOCK_RSV_CHUNK , +BTRFS_BLOCK_RSV_REMAP ,BTRFS_BLOCK_RSV_DELOPS ,BTRFS_BLOCK_RSV_DELREFS , BTRFS_BLOCK_RSV_TREELOG,
fs/btrfs/disk-io.c +1 -0
@@ -2773,6 +2773,7 @@ void btrfs_init_fs_info(struct btrfs_fs_info *fs_info) BTRFS_BLOCK_RSV_GLOBAL);btrfs_init_block_rsv (&fs_info->trans_block_rsv,BTRFS_BLOCK_RSV_TRANS );btrfs_init_block_rsv (&fs_info->chunk_block_rsv,BTRFS_BLOCK_RSV_CHUNK ); +btrfs_init_block_rsv (&fs_info->remap_block_rsv,BTRFS_BLOCK_RSV_REMAP );btrfs_init_block_rsv (&fs_info->treelog_rsv,BTRFS_BLOCK_RSV_TREELOG );btrfs_init_block_rsv (&fs_info->empty_block_rsv,BTRFS_BLOCK_RSV_EMPTY ); btrfs_init_block_rsv(&fs_info->delayed_block_rsv,
fs/btrfs/fs.h +2 -0
@@ -501,6 +501,8 @@ struct btrfs_fs_info {struct btrfs_block_rsv trans_block_rsv;/* Block reservation for chunk tree */ struct btrfs_block_rsv chunk_block_rsv; +/* Block reservation for remap tree. */ +struct btrfs_block_rsv remap_block_rsv;/* Block reservation for delayed operations */ struct btrfs_block_rsv delayed_block_rsv;/* Block reservation for delayed refs */
fs/btrfs/space-info.c +12 -1
@@ -215,7 +215,7 @@ static u64 calc_chunk_size(const struct btrfs_fs_info *fs_info, u64 flags)if (flags &BTRFS_BLOCK_GROUP_DATA )return BTRFS_MAX_DATA_CHUNK_SIZE ; -else if (flags &BTRFS_BLOCK_GROUP_SYSTEM ) +else if (flags & (BTRFS_BLOCK_GROUP_SYSTEM |BTRFS_BLOCK_GROUP_METADATA_REMAP ))return SZ_32M ;/* Handle BTRFS_BLOCK_GROUP_METADATA */
@@ -344,6 +344,8 @@ int btrfs_init_space_info(struct btrfs_fs_info *fs_info)if (mixed) { flags =BTRFS_BLOCK_GROUP_METADATA |BTRFS_BLOCK_GROUP_DATA ; ret =create_space_info (fs_info, flags); +if (ret) + goto out; }else { flags =BTRFS_BLOCK_GROUP_METADATA ; ret =create_space_info (fs_info, flags);
@@ -352,7 +354,15 @@ int btrfs_init_space_info(struct btrfs_fs_info *fs_info) flags =BTRFS_BLOCK_GROUP_DATA ; ret =create_space_info (fs_info, flags); +if (ret) + goto out; + } + +if (features &BTRFS_FEATURE_INCOMPAT_REMAP_TREE ) { + flags =BTRFS_BLOCK_GROUP_METADATA_REMAP ; + ret =create_space_info (fs_info, flags); } + out:return ret; }
@@ -607,6 +617,7 @@ static void dump_global_block_rsv(struct btrfs_fs_info *fs_info)DUMP_BLOCK_RSV (fs_info, global_block_rsv);DUMP_BLOCK_RSV (fs_info, trans_block_rsv);DUMP_BLOCK_RSV (fs_info, chunk_block_rsv); +DUMP_BLOCK_RSV (fs_info, remap_block_rsv);DUMP_BLOCK_RSV (fs_info, delayed_block_rsv);DUMP_BLOCK_RSV (fs_info, delayed_refs_rsv); }
fs/btrfs/sysfs.c +2 -0
@@ -1972,6 +1972,8 @@ static const char *alloc_name(struct btrfs_space_info *space_info)case BTRFS_BLOCK_GROUP_SYSTEM :ASSERT (space_info->subgroup_id ==BTRFS_SUB_GROUP_PRIMARY );return "system" ; +case BTRFS_BLOCK_GROUP_METADATA_REMAP : +return "metadata-remap" ;default :WARN_ON (1 );return "invalid-combination" ;
fs/btrfs/tree-checker.c +11 -2
@@ -748,17 +748,26 @@ static int check_block_group_item(struct extent_buffer *leaf,return -EUCLEAN ; } +if (unlikely (flags &BTRFS_BLOCK_GROUP_METADATA_REMAP && + !btrfs_fs_incompat (fs_info,REMAP_TREE ))) { +block_group_err (leaf, slot, +"invalid flags, have 0x%llx (METADATA_REMAP flag set) but no remap-tree incompat flag" , + flags); +return -EUCLEAN ; + } + type = flags &BTRFS_BLOCK_GROUP_TYPE_MASK ;if (unlikely (type !=BTRFS_BLOCK_GROUP_DATA && type !=BTRFS_BLOCK_GROUP_METADATA && type !=BTRFS_BLOCK_GROUP_SYSTEM && + type !=BTRFS_BLOCK_GROUP_METADATA_REMAP && type != (BTRFS_BLOCK_GROUP_METADATA |BTRFS_BLOCK_GROUP_DATA ))) {block_group_err (leaf, slot, -"invalid type, have 0x%llx (%lu bits set) expect either 0x%llx, 0x%llx, 0x%llx or 0x%llx" , +"invalid type, have 0x%llx (%lu bits set) expect either 0x%llx, 0x%llx, 0x%llx, 0x%llx or 0x%llx" , type,hweight64 (type),BTRFS_BLOCK_GROUP_DATA ,BTRFS_BLOCK_GROUP_METADATA , -BTRFS_BLOCK_GROUP_SYSTEM , +BTRFS_BLOCK_GROUP_SYSTEM ,BTRFS_BLOCK_GROUP_METADATA_REMAP ,BTRFS_BLOCK_GROUP_METADATA |BTRFS_BLOCK_GROUP_DATA );return -EUCLEAN ; }
fs/btrfs/volumes.c +3 -0
@@ -231,6 +231,9 @@ void btrfs_describe_block_groups(u64 bg_flags, char *buf, u32 size_buf)DESCRIBE_FLAG (BTRFS_BLOCK_GROUP_DATA ,"data" );DESCRIBE_FLAG (BTRFS_BLOCK_GROUP_SYSTEM ,"system" );DESCRIBE_FLAG (BTRFS_BLOCK_GROUP_METADATA ,"metadata" ); +/* Block groups containing the remap tree. */ +DESCRIBE_FLAG (BTRFS_BLOCK_GROUP_METADATA_REMAP ,"metadata-remap" ); +/* Block group that has been remapped. */ DESCRIBE_FLAG (BTRFS_BLOCK_GROUP_REMAPPED ,"remapped" );DESCRIBE_FLAG (BTRFS_AVAIL_ALLOC_BIT_SINGLE ,"single" );
fs/btrfs/volumes.h +9 -1
@@ -58,7 +58,6 @@ static_assert(ilog2(BTRFS_STRIPE_LEN) == BTRFS_STRIPE_LEN_SHIFT);*/ static_assert (const_ffs (BTRFS_BLOCK_GROUP_RAID0 ) <const_ffs (BTRFS_BLOCK_GROUP_PROFILE_MASK & ~BTRFS_BLOCK_GROUP_RAID0 )); -static_assert (ilog2 (BTRFS_BLOCK_GROUP_RAID0 ) >ilog2 (BTRFS_BLOCK_GROUP_TYPE_MASK ));/* ilog2() can handle both constants and variables */ #define BTRFS_BG_FLAG_TO_INDEX (profile) \
@@ -80,6 +79,15 @@ enum btrfs_raid_types { BTRFS_NR_RAID_TYPES }; +static_assert (BTRFS_RAID_RAID0 ==1 ); +static_assert (BTRFS_RAID_RAID1 ==2 ); +static_assert (BTRFS_RAID_DUP ==3 ); +static_assert (BTRFS_RAID_RAID10 ==4 ); +static_assert (BTRFS_RAID_RAID5 ==5 ); +static_assert (BTRFS_RAID_RAID6 ==6 ); +static_assert (BTRFS_RAID_RAID1C3 ==7 ); +static_assert (BTRFS_RAID_RAID1C4 ==8 ); +/* * Use sequence counter to get consistent device stat data on * 32-bit processors.
include/uapi/linux/btrfs_tree.h +3 -1
@@ -1169,12 +1169,14 @@ struct btrfs_dev_replace_item {#define BTRFS_BLOCK_GROUP_RAID1C3 (1ULL << 9)#define BTRFS_BLOCK_GROUP_RAID1C4 (1ULL << 10)#define BTRFS_BLOCK_GROUP_REMAPPED (1ULL << 11) +#define BTRFS_BLOCK_GROUP_METADATA_REMAP (1ULL << 12)#define BTRFS_BLOCK_GROUP_RESERVED (BTRFS_AVAIL_ALLOC_BIT_SINGLE | \ BTRFS_SPACE_INFO_GLOBAL_RSV)#define BTRFS_BLOCK_GROUP_TYPE_MASK (BTRFS_BLOCK_GROUP_DATA | \ BTRFS_BLOCK_GROUP_SYSTEM | \ - BTRFS_BLOCK_GROUP_METADATA) + BTRFS_BLOCK_GROUP_METADATA | \ + BTRFS_BLOCK_GROUP_METADATA_REMAP)#define BTRFS_BLOCK_GROUP_PROFILE_MASK (BTRFS_BLOCK_GROUP_RAID0 | \ BTRFS_BLOCK_GROUP_RAID1 | \
Signature · 1 line
2.51.2
PATCH v8 3/17 btrfs: allow remapped chunks to have zero stripes
In reply to Mark HarmstoneWhen a chunk has been fully remapped, we are going to set its num_stripes to 0, as it will no longer represent a physical location on disk.
Change tree-checker to allow for this, and fix read_one_chunk() to avoid a divide by zero.
--- fs/btrfs/tree-checker.c | 65 ++++++++++++++++++++++++++++------------- fs/btrfs/volumes.c | 7 ++++- 2 files changed, 51 insertions(+), 21 deletions(-)
fs/btrfs/tree-checker.c +45 -20
@@ -816,6 +816,41 @@ static void chunk_err(const struct btrfs_fs_info *fs_info,va_end (args); } +static bool valid_stripe_count (u64 profile,u16 num_stripes, +u16 sub_stripes) +{ +switch (profile) { +case BTRFS_BLOCK_GROUP_RAID0 : +return true; +case BTRFS_BLOCK_GROUP_RAID10 : +return sub_stripes == + btrfs_raid_array[BTRFS_RAID_RAID10 ].sub_stripes; +case BTRFS_BLOCK_GROUP_RAID1 : +return num_stripes == + btrfs_raid_array[BTRFS_RAID_RAID1 ].devs_min; +case BTRFS_BLOCK_GROUP_RAID1C3 : +return num_stripes == + btrfs_raid_array[BTRFS_RAID_RAID1C3 ].devs_min; +case BTRFS_BLOCK_GROUP_RAID1C4 : +return num_stripes == + btrfs_raid_array[BTRFS_RAID_RAID1C4 ].devs_min; +case BTRFS_BLOCK_GROUP_RAID5 : +return num_stripes >= + btrfs_raid_array[BTRFS_RAID_RAID5 ].devs_min; +case BTRFS_BLOCK_GROUP_RAID6 : +return num_stripes >= + btrfs_raid_array[BTRFS_RAID_RAID6 ].devs_min; +case BTRFS_BLOCK_GROUP_DUP : +return num_stripes == + btrfs_raid_array[BTRFS_RAID_DUP ].dev_stripes; +case 0 :/* SINGLE */ +return num_stripes == + btrfs_raid_array[BTRFS_RAID_SINGLE ].dev_stripes; +default : +BUG (); + } +} +/* * The common chunk check which could also work on super block sys chunk array. *
@@ -839,6 +874,7 @@ int btrfs_check_chunk_valid(const struct btrfs_fs_info *fs_info,u64 features;u32 chunk_sector_size;bool mixed = false; +bool remapped;int raid_index;int nparity;int ncopies;
@@ -862,12 +898,14 @@ int btrfs_check_chunk_valid(const struct btrfs_fs_info *fs_info, ncopies = btrfs_raid_array[raid_index].ncopies; nparity = btrfs_raid_array[raid_index].nparity; -if (unlikely (!num_stripes)) { + remapped = type &BTRFS_BLOCK_GROUP_REMAPPED ; + +if (unlikely (!remapped && !num_stripes)) {chunk_err (fs_info, leaf, chunk, logical,"invalid chunk num_stripes, have %u" , num_stripes);return -EUCLEAN ; } -if (unlikely (num_stripes < ncopies)) { +if (unlikely (num_stripes !=0 && num_stripes < ncopies)) {chunk_err (fs_info, leaf, chunk, logical,"invalid chunk num_stripes < ncopies, have %u < %d" , num_stripes, ncopies);
@@ -965,22 +1003,9 @@ int btrfs_check_chunk_valid(const struct btrfs_fs_info *fs_info, } } -if (unlikely ((type &BTRFS_BLOCK_GROUP_RAID10 && - sub_stripes != btrfs_raid_array[BTRFS_RAID_RAID10 ].sub_stripes) || - (type &BTRFS_BLOCK_GROUP_RAID1 && - num_stripes != btrfs_raid_array[BTRFS_RAID_RAID1 ].devs_min) || - (type &BTRFS_BLOCK_GROUP_RAID1C3 && - num_stripes != btrfs_raid_array[BTRFS_RAID_RAID1C3 ].devs_min) || - (type &BTRFS_BLOCK_GROUP_RAID1C4 && - num_stripes != btrfs_raid_array[BTRFS_RAID_RAID1C4 ].devs_min) || - (type &BTRFS_BLOCK_GROUP_RAID5 && - num_stripes < btrfs_raid_array[BTRFS_RAID_RAID5 ].devs_min) || - (type &BTRFS_BLOCK_GROUP_RAID6 && - num_stripes < btrfs_raid_array[BTRFS_RAID_RAID6 ].devs_min) || - (type &BTRFS_BLOCK_GROUP_DUP && - num_stripes != btrfs_raid_array[BTRFS_RAID_DUP ].dev_stripes) || - ((type &BTRFS_BLOCK_GROUP_PROFILE_MASK ) ==0 && - num_stripes != btrfs_raid_array[BTRFS_RAID_SINGLE ].dev_stripes))) { +if (!remapped && + !valid_stripe_count (type &BTRFS_BLOCK_GROUP_PROFILE_MASK , + num_stripes, sub_stripes)) { chunk_err(fs_info, leaf, chunk, logical, "invalid num_stripes:sub_stripes %u:%u for profile %llu", num_stripes, sub_stripes,
@@ -1004,11 +1029,11 @@ static int check_leaf_chunk_item(struct extent_buffer *leaf,struct btrfs_fs_info *fs_info = leaf->fs_info;int num_stripes; -if (unlikely (btrfs_item_size (leaf, slot) <sizeof (struct btrfs_chunk ))) { +if (unlikely (btrfs_item_size (leaf, slot) < offsetof(struct btrfs_chunk , stripe))) {chunk_err (fs_info, leaf, chunk, key->offset,"invalid chunk item size: have %u expect [%zu, %u)" ,btrfs_item_size (leaf, slot), -sizeof (struct btrfs_chunk ), + offsetof(struct btrfs_chunk , stripe),BTRFS_LEAF_DATA_SIZE (fs_info));return -EUCLEAN ; }
fs/btrfs/volumes.c +6 -1
@@ -7045,7 +7045,12 @@ static int read_one_chunk(struct btrfs_key *key, struct extent_buffer *leaf,*/ map->sub_stripes = btrfs_raid_array[index].sub_stripes; map->verified_stripes =0 ; - map->stripe_size =btrfs_calc_stripe_length (map); + +if (num_stripes >0 ) + map->stripe_size =btrfs_calc_stripe_length (map); +else + map->stripe_size =0 ; +for (i =0 ; i < num_stripes; i++) { map->stripes[i].physical =btrfs_stripe_offset_nr (leaf, chunk, i);
Signature · 1 line
2.51.2
PATCH v8 4/17 btrfs: remove remapped block groups from the free-space tree
In reply to Mark HarmstoneNo new allocations can be done from block groups that have the REMAPPED flag set, so there's no value in their having entries in the free-space tree.
Prevent a search through the free-space tree being scheduled for such a block group, and prevent any additions to the in-memory free-space tree.
--- fs/btrfs/block-group.c | 19 ++++++++++++++++--- fs/btrfs/free-space-cache.c | 3 +++ 2 files changed, 19 insertions(+), 3 deletions(-)
fs/btrfs/block-group.c +16 -3
@@ -933,6 +933,13 @@ int btrfs_cache_block_group(struct btrfs_block_group *cache, bool wait)if (btrfs_is_zoned (fs_info))return 0 ; +/* +* No allocations can be done from remapped block groups, so they have +* no entries in the free-space tree. +*/ +if (cache->flags &BTRFS_BLOCK_GROUP_REMAPPED ) +return 0 ; + caching_ctl =kzalloc (sizeof (*caching_ctl),GFP_NOFS );if (!caching_ctl)return -ENOMEM ;
@@ -1246,10 +1253,16 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans,* deletes the block group item from the extent tree, allowing for * another task to attempt to create another block group with the same * item key (and failing with -EEXIST and a transaction abort). +* +* If the REMAPPED flag has been set the block group's free space +* has already been removed, so we can skip the call to +* btrfs_remove_block_group_free_space(). */ - ret =btrfs_remove_block_group_free_space (trans, block_group); -if (ret) - goto out; +if (!(block_group->flags &BTRFS_BLOCK_GROUP_REMAPPED )) { + ret =btrfs_remove_block_group_free_space (trans, block_group); +if (ret) + goto out; + } ret =remove_block_group_item (trans, path, block_group);if (ret <0 )
fs/btrfs/free-space-cache.c +3 -0
@@ -2756,6 +2756,9 @@ int btrfs_add_free_space(struct btrfs_block_group *block_group, {enum btrfs_trim_state trim_state =BTRFS_TRIM_STATE_UNTRIMMED ; +if (block_group->flags &BTRFS_BLOCK_GROUP_REMAPPED ) +return 0 ; +if (btrfs_is_zoned (block_group->fs_info))return __btrfs_add_free_space_zoned (block_group, bytenr, size, true);
Signature · 1 line
2.51.2
PATCH v8 5/17 btrfs: don't add metadata items for the remap tree to the extent tree
In reply to Mark HarmstoneThere is the following potential problem with the remap tree and delayed refs:
* Remapped extent freed in a delayed ref, which removes an entry from the remap tree
* Remap tree now small enough to fit in a single leaf
* Corruption as we now have a level-0 block with a level-1 metadata item in the extent tree
One solution to this would be to rework the remap tree code so that it operates via delayed refs. But as we're hoping to remove cow-only metadata items in the future anyway, change things so that the remap tree doesn't have any entries in the extent tree. This also has the benefit of reducing write amplification.
We also make it so that the clear_cache mount option is a no-op, as with the extent tree v2, as the free-space tree can no longer be recreated from the extent tree.
Finally disable relocating the remap tree itself, which is added back in a later patch. As it is we would get corruption as the traditional relocation method walks the extent tree, and we're removing its metadata items.
--- fs/btrfs/disk-io.c | 3 +++ fs/btrfs/extent-tree.c | 31 ++++++++++++++++++++++++++++++- fs/btrfs/volumes.c | 3 +++ 3 files changed, 36 insertions(+), 1 deletion(-)
fs/btrfs/disk-io.c +3 -0
@@ -3007,6 +3007,9 @@ int btrfs_start_pre_rw_mount(struct btrfs_fs_info *fs_info)if (btrfs_fs_incompat (fs_info,EXTENT_TREE_V2 ))btrfs_warn (fs_info,"'clear_cache' option is ignored with extent tree v2" ); +else if (btrfs_fs_incompat (fs_info,REMAP_TREE )) +btrfs_warn (fs_info, +"'clear_cache' option is ignored with remap tree" );else rebuild_free_space_tree = true; } else if (btrfs_fs_compat_ro(fs_info, FREE_SPACE_TREE) &&
fs/btrfs/extent-tree.c +30 -1
@@ -1553,6 +1553,28 @@ static void free_head_ref_squota_rsv(struct btrfs_fs_info *fs_info, BTRFS_QGROUP_RSV_DATA); } +static int drop_remap_tree_ref (struct btrfs_trans_handle *trans, +const struct btrfs_delayed_ref_node *node) +{ +u64 bytenr = node->bytenr; +u64 num_bytes = node->num_bytes; +int ret; + + ret =btrfs_add_to_free_space_tree (trans, bytenr, num_bytes); +if (unlikely (ret)) { +btrfs_abort_transaction (trans, ret); +return ret; + } + + ret =btrfs_update_block_group (trans, bytenr, num_bytes, false); +if (unlikely (ret)) { +btrfs_abort_transaction (trans, ret); +return ret; + } + +return 0 ; +} + static int run_delayed_data_ref(struct btrfs_trans_handle *trans, struct btrfs_delayed_ref_head *href, const struct btrfs_delayed_ref_node *node,
@@ -1747,7 +1769,10 @@ static int run_delayed_tree_ref(struct btrfs_trans_handle *trans, }else if (node->action ==BTRFS_ADD_DELAYED_REF ) { ret =__btrfs_inc_extent_ref (trans, node, extent_op); }else if (node->action ==BTRFS_DROP_DELAYED_REF ) { - ret =__btrfs_free_extent (trans, href, node, extent_op); +if (node->ref_root ==BTRFS_REMAP_TREE_OBJECTID ) + ret =drop_remap_tree_ref (trans, node); +else + ret =__btrfs_free_extent (trans, href, node, extent_op); }else {BUG (); }
@@ -4886,6 +4911,9 @@ static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,int level =btrfs_delayed_ref_owner (node);bool skinny_metadata =btrfs_fs_incompat (fs_info,SKINNY_METADATA ); +if (unlikely (node->ref_root ==BTRFS_REMAP_TREE_OBJECTID )) + goto skip; + extent_key.objectid = node->bytenr;if (skinny_metadata) {/* The owner of a tree block is the level. */
@@ -4938,6 +4966,7 @@ static int alloc_reserved_tree_block(struct btrfs_trans_handle *trans,btrfs_free_path (path); +skip:return alloc_reserved_extent (trans, node->bytenr, fs_info->nodesize); }
fs/btrfs/volumes.c +3 -0
@@ -3970,6 +3970,9 @@ static bool should_balance_chunk(struct extent_buffer *leaf, struct btrfs_chunkstruct btrfs_balance_args *bargs =NULL ;u64 chunk_type =btrfs_chunk_type (leaf, chunk); +if (chunk_type &BTRFS_BLOCK_GROUP_METADATA_REMAP ) +return false; +/* type filter */ if (!((chunk_type &BTRFS_BLOCK_GROUP_TYPE_MASK ) & (bctl->flags &BTRFS_BALANCE_TYPE_MASK ))) {
Signature · 1 line
2.51.2
PATCH v8 6/17 btrfs: rename struct btrfs_block_group field commit_used to last_used
In reply to Mark HarmstoneRename the field commit_used in struct btrfs_block_group to last_used, for clarity and consistency with the similar fields we're about to add. It's not obvious that commit_flags means "flags as of the last commit" rather than "flags related to a commit".
--- fs/btrfs/block-group.c | 24 ++++++++++++------------ fs/btrfs/block-group.h | 4 ++-- 2 files changed, 14 insertions(+), 14 deletions(-)
fs/btrfs/block-group.c +12 -12
@@ -2388,7 +2388,7 @@ static int read_one_block_group(struct btrfs_fs_info *info, cache->length = key->offset; cache->used =btrfs_stack_block_group_used (bgi); - cache->commit_used = cache->used; + cache->last_used = cache->used; cache->flags =btrfs_stack_block_group_flags (bgi); cache->global_root_id =btrfs_stack_block_group_chunk_objectid (bgi); cache->space_info =btrfs_find_space_info (info, cache->flags);
@@ -2667,7 +2667,7 @@ static int insert_block_group_item(struct btrfs_trans_handle *trans,struct btrfs_block_group_item bgi;struct btrfs_root *root =btrfs_block_group_root (fs_info);struct btrfs_key key; -u64 old_commit_used; +u64 old_last_used;int ret;spin_lock (&block_group->lock);
@@ -2675,8 +2675,8 @@ static int insert_block_group_item(struct btrfs_trans_handle *trans,btrfs_set_stack_block_group_chunk_objectid (&bgi, block_group->global_root_id);btrfs_set_stack_block_group_flags (&bgi, block_group->flags); - old_commit_used = block_group->commit_used; - block_group->commit_used = block_group->used; + old_last_used = block_group->last_used; + block_group->last_used = block_group->used; key.objectid = block_group->start; key.type =BTRFS_BLOCK_GROUP_ITEM_KEY ; key.offset = block_group->length;
@@ -2685,7 +2685,7 @@ static int insert_block_group_item(struct btrfs_trans_handle *trans, ret =btrfs_insert_item (trans, root, &key, &bgi,sizeof (bgi));if (ret <0 ) {spin_lock (&block_group->lock); - block_group->commit_used = old_commit_used; + block_group->last_used = old_last_used;spin_unlock (&block_group->lock); }
@@ -3139,7 +3139,7 @@ static int update_block_group_item(struct btrfs_trans_handle *trans,struct extent_buffer *leaf;struct btrfs_block_group_item bgi;struct btrfs_key key; -u64 old_commit_used; +u64 old_last_used;u64 used;/*
@@ -3149,14 +3149,14 @@ static int update_block_group_item(struct btrfs_trans_handle *trans,* may be changed. */ spin_lock (&cache->lock); - old_commit_used = cache->commit_used; + old_last_used = cache->last_used; used = cache->used;/* No change in used bytes, can safely skip it. */ -if (cache->commit_used == used) { +if (cache->last_used == used) {spin_unlock (&cache->lock);return 0 ; } - cache->commit_used = used; + cache->last_used = used;spin_unlock (&cache->lock); key.objectid = cache->start;
@@ -3180,17 +3180,17 @@ static int update_block_group_item(struct btrfs_trans_handle *trans, fail:btrfs_release_path (path);/* -* We didn't update the block group item, need to revert commit_used +* We didn't update the block group item, need to revert last_used * unless the block group item didn't exist yet - this is to prevent a * race with a concurrent insertion of the block group item, with * insert_block_group_item(), that happened just after we attempted to -* update. In that case we would reset commit_used to 0 just after the +* update. In that case we would reset last_used to 0 just after the * insertion set it to a value greater than 0 - if the block group later * becomes with 0 used bytes, we would incorrectly skip its update. */ if (ret <0 && ret != -ENOENT ) {spin_lock (&cache->lock); - cache->commit_used = old_commit_used; + cache->last_used = old_last_used;spin_unlock (&cache->lock); }return ret;
fs/btrfs/block-group.h +2 -2
@@ -132,10 +132,10 @@ struct btrfs_block_group {/* * The last committed used bytes of this block group, if the above @used -* is still the same as @commit_used, we don't need to update block +* is still the same as @last_used, we don't need to update block * group item of this block group. */ -u64 commit_used; +u64 last_used;/* * If the free space extent count exceeds this number, convert the block * group to bitmaps.
Signature · 1 line
2.51.2
PATCH v8 7/17 btrfs: add extended version of struct block_group_item
In reply to Mark HarmstoneAdd a struct btrfs_block_group_item_v2, which is used in the block group tree if the remap-tree incompat flag is set.
This adds two new fields to the block group item: remap_bytes and identity_remap_count.
remap_bytes records the amount of data that's physically within this block group, but nominally in another, remapped block group. This is necessary because this data will need to be moved first if this block group is itself relocated. If remap_bytes > 0, this is an indicator to the relocation thread that it will need to search the remap-tree for backrefs. A block group must also have remap_bytes == 0 before it can be dropped.
identity_remap_count records how many identity remap items are located in the remap tree for this block group. When relocation is begun for this block group, this is set to the number of holes in the free-space tree for this range. As identity remaps are converted into actual remaps by the relocation process, this number is decreased. Once it reaches 0, either because of relocation or because extents have been deleted, the block group has been fully remapped and its chunk's device extents are removed.
--- fs/btrfs/accessors.h | 20 +++++++ fs/btrfs/block-group.c | 100 ++++++++++++++++++++++++-------- fs/btrfs/block-group.h | 14 ++++- fs/btrfs/discard.c | 2 +- fs/btrfs/tree-checker.c | 10 +++- include/uapi/linux/btrfs_tree.h | 8 +++ 6 files changed, 126 insertions(+), 28 deletions(-)
fs/btrfs/accessors.h +20 -0
@@ -240,6 +240,26 @@ BTRFS_SETGET_FUNCS(block_group_flags, struct btrfs_block_group_item, flags, 64);BTRFS_SETGET_STACK_FUNCS (stack_block_group_flags, struct btrfs_block_group_item, flags, 64); +/* struct btrfs_block_group_item_v2 */ +BTRFS_SETGET_STACK_FUNCS(stack_block_group_v2_used, struct btrfs_block_group_item_v2, + used,64 ); +BTRFS_SETGET_FUNCS(block_group_v2_used, struct btrfs_block_group_item_v2, used, 64); +BTRFS_SETGET_STACK_FUNCS (stack_block_group_v2_chunk_objectid, + struct btrfs_block_group_item_v2, chunk_objectid, 64); +BTRFS_SETGET_FUNCS (block_group_v2_chunk_objectid, + struct btrfs_block_group_item_v2, chunk_objectid, 64); +BTRFS_SETGET_STACK_FUNCS (stack_block_group_v2_flags, + struct btrfs_block_group_item_v2, flags, 64); +BTRFS_SETGET_FUNCS(block_group_v2_flags, struct btrfs_block_group_item_v2, flags, 64); +BTRFS_SETGET_STACK_FUNCS (stack_block_group_v2_remap_bytes, + struct btrfs_block_group_item_v2, remap_bytes, 64); +BTRFS_SETGET_FUNCS(block_group_v2_remap_bytes, struct btrfs_block_group_item_v2, + remap_bytes,64 ); +BTRFS_SETGET_STACK_FUNCS (stack_block_group_v2_identity_remap_count, + struct btrfs_block_group_item_v2, identity_remap_count, 32); +BTRFS_SETGET_FUNCS(block_group_v2_identity_remap_count, struct btrfs_block_group_item_v2, + identity_remap_count,32 ); +/* struct btrfs_free_space_info */ BTRFS_SETGET_FUNCS(free_space_extent_count, struct btrfs_free_space_info, extent_count,32 );
fs/btrfs/block-group.c +76 -24
190 lines not on this page · raw message
fs/btrfs/block-group.h +13 -1
@@ -129,6 +129,8 @@ struct btrfs_block_group {u64 flags;u64 cache_generation;u64 global_root_id; +u64 remap_bytes; +u32 identity_remap_count;/* * The last committed used bytes of this block group, if the above @used
@@ -136,6 +138,15 @@ struct btrfs_block_group {* group item of this block group. */ u64 last_used; +/* +* The last committed remap_bytes value of this block group. +*/ +u64 last_remap_bytes; +/* +* The last commited identity_remap_count value of this block group. +*/ +u32 last_identity_remap_count; +/* * If the free space extent count exceeds this number, convert the block * group to bitmaps.
@@ -282,7 +293,8 @@ static inline bool btrfs_is_block_group_used(const struct btrfs_block_group *bg) {lockdep_assert_held (&bg->lock); -return (bg->used >0 || bg->reserved >0 || bg->pinned >0 ); +return (bg->used >0 || bg->reserved >0 || bg->pinned >0 || + bg->remap_bytes >0 ); }static inline bool btrfs_is_block_group_data_only (const struct btrfs_block_group *block_group)
fs/btrfs/discard.c +1 -1
@@ -373,7 +373,7 @@ void btrfs_discard_queue_work(struct btrfs_discard_ctl *discard_ctl,if (!block_group || !btrfs_test_opt (block_group->fs_info,DISCARD_ASYNC ))return ; -if (block_group->used ==0 ) +if (block_group->used ==0 && block_group->remap_bytes ==0 )add_to_discard_unused_list (discard_ctl, block_group);else add_to_discard_list (discard_ctl, block_group);
fs/btrfs/tree-checker.c +8 -2
@@ -688,6 +688,7 @@ static int check_block_group_item(struct extent_buffer *leaf,u64 chunk_objectid;u64 flags;u64 type; +size_t exp_size;/* * Here we don't really care about alignment since extent allocator can
@@ -699,10 +700,15 @@ static int check_block_group_item(struct extent_buffer *leaf,return -EUCLEAN ; } -if (unlikely (item_size !=sizeof (bgi))) { +if (btrfs_fs_incompat (fs_info,REMAP_TREE )) + exp_size =sizeof (struct btrfs_block_group_item_v2 ); +else + exp_size =sizeof (struct btrfs_block_group_item ); + +if (unlikely (item_size != exp_size)) {block_group_err (leaf, slot,"invalid item size, have %u expect %zu" , - item_size,sizeof (bgi)); + item_size, exp_size);return -EUCLEAN ; }
include/uapi/linux/btrfs_tree.h +8 -0
@@ -1229,6 +1229,14 @@ struct btrfs_block_group_item {__le64 flags; } __attribute__ ((__packed__)); +struct btrfs_block_group_item_v2 { +__le64 used; +__le64 chunk_objectid; +__le64 flags; +__le64 remap_bytes; +__le32 identity_remap_count; +} __attribute__ ((__packed__ )); +struct btrfs_free_space_info {__le32 extent_count;__le32 flags;
Signature · 1 line
2.51.2
PATCH v8 8/17 btrfs: allow mounting filesystems with remap-tree incompat flag
In reply to Mark HarmstoneIf we encounter a filesystem with the remap-tree incompat flag set, valdiate its compatibility with the other flags, and load the remap tree using the values that have been added to the superblock.
The remap-tree feature depends on the free space tere, but no-holes and block-group-tree have been made dependencies to reduce the testing matrix. Similarly I'm not aware of any reason why mixed-bg and zoned would be incompatible with remap-tree, but this is blocked for the time being until it can be fully tested.
--- fs/btrfs/Kconfig | 2 + fs/btrfs/accessors.h | 6 ++ fs/btrfs/disk-io.c | 111 ++++++++++++++++++++++++++++---- fs/btrfs/extent-tree.c | 2 + fs/btrfs/fs.h | 4 +- fs/btrfs/transaction.c | 7 ++ include/uapi/linux/btrfs_tree.h | 5 +- 7 files changed, 122 insertions(+), 15 deletions(-)
fs/btrfs/Kconfig +2 -0
@@ -115,4 +115,6 @@ config BTRFS_EXPERIMENTAL - large folio support + - remap-tree - logical address remapping tree + If unsure, say N.
fs/btrfs/accessors.h +6 -0
@@ -883,6 +883,12 @@ BTRFS_SETGET_STACK_FUNCS(super_uuid_tree_generation, struct btrfs_super_block, uuid_tree_generation, 64); BTRFS_SETGET_STACK_FUNCS(super_nr_global_roots, struct btrfs_super_block, nr_global_roots,64 ); +BTRFS_SETGET_STACK_FUNCS(super_remap_root, struct btrfs_super_block, + remap_root,64 ); +BTRFS_SETGET_STACK_FUNCS(super_remap_root_generation, struct btrfs_super_block, + remap_root_generation,64 ); +BTRFS_SETGET_STACK_FUNCS(super_remap_root_level, struct btrfs_super_block, + remap_root_level,8 );/* struct btrfs_file_extent_item */ BTRFS_SETGET_STACK_FUNCS(stack_file_extent_type, struct btrfs_file_extent_item,
fs/btrfs/disk-io.c +98 -13
161 lines not on this page · raw message
fs/btrfs/extent-tree.c +2 -0
@@ -2589,6 +2589,8 @@ static u64 get_alloc_profile_by_root(struct btrfs_root *root, int data) flags =BTRFS_BLOCK_GROUP_DATA ; else if (root == fs_info->chunk_root) flags =BTRFS_BLOCK_GROUP_SYSTEM ; + else if (root == fs_info->remap_root) + flags =BTRFS_BLOCK_GROUP_METADATA_REMAP ; else flags =BTRFS_BLOCK_GROUP_METADATA ;
fs/btrfs/fs.h +3 -1
@@ -307,7 +307,8 @@ enum {#define BTRFS_FEATURE_INCOMPAT_SUPP \ (BTRFS_FEATURE_INCOMPAT_SUPP_STABLE | \ BTRFS_FEATURE_INCOMPAT_RAID_STRIPE_TREE | \ - BTRFS_FEATURE_INCOMPAT_EXTENT_TREE_V2) + BTRFS_FEATURE_INCOMPAT_EXTENT_TREE_V2 | \ + BTRFS_FEATURE_INCOMPAT_REMAP_TREE)#else
@@ -467,6 +468,7 @@ struct btrfs_fs_info {struct btrfs_root *data_reloc_root;struct btrfs_root *block_group_root;struct btrfs_root *stripe_root; +struct btrfs_root *remap_root;/* The log root tree is a directory of all the other log roots */ struct btrfs_root *log_root_tree;
fs/btrfs/transaction.c +7 -0
@@ -1967,6 +1967,13 @@ static void update_super_roots(struct btrfs_fs_info *fs_info) super->cache_generation =0 ;if (test_bit (BTRFS_FS_UPDATE_UUID_TREE_GEN , &fs_info->flags)) super->uuid_tree_generation = root_item->generation; + +if (btrfs_fs_incompat (fs_info,REMAP_TREE )) { + root_item = &fs_info->remap_root->root_item; + super->remap_root = root_item->bytenr; + super->remap_root_generation = root_item->generation; + super->remap_root_level = root_item->level; + } }int btrfs_transaction_blocked (struct btrfs_fs_info *info)
include/uapi/linux/btrfs_tree.h +4 -1
@@ -721,9 +721,12 @@ struct btrfs_super_block {__u8 metadata_uuid[BTRFS_FSID_SIZE ];__u64 nr_global_roots; +__le64 remap_root; +__le64 remap_root_generation; +__u8 remap_root_level;/* Future expansion */ -__le64 reserved[27 ]; +__u8 reserved[199 ];__u8 sys_chunk_array[BTRFS_SYSTEM_CHUNK_ARRAY_SIZE ];struct btrfs_root_backup super_roots[BTRFS_NUM_BACKUP_ROOTS ];
Signature · 1 line
2.51.2
PATCH v8 9/17 btrfs: redirect I/O for remapped block groups
In reply to Mark HarmstoneChange btrfs_map_block() so that if the block group has the REMAPPED flag set, we call btrfs_translate_remap() to obtain a new address.
btrfs_translate_remap() searches the remap tree for a range corresponding to the logical address passed to btrfs_map_block(). If it is within an identity remap, this part of the block group hasn't yet been relocated, and so we use the existing address.
If it is within an actual remap, we subtract the start of the remap range and add the address of its destination, contained in the item's payload.
--- fs/btrfs/relocation.c | 54 +++++++++++++++++++++++++++++++++++++++++++ fs/btrfs/relocation.h | 2 ++ fs/btrfs/volumes.c | 19 +++++++++++++++ 3 files changed, 75 insertions(+)
fs/btrfs/relocation.c +54 -0
@@ -3859,6 +3859,60 @@ static const char *stage_to_string(enum reloc_stage stage)return "unknown" ; } +int btrfs_translate_remap (struct btrfs_fs_info *fs_info,u64 *logical, +u64 *length) +{ +int ret; +struct btrfs_key key, found_key; +struct extent_buffer *leaf; +struct btrfs_remap_item *remap; +BTRFS_PATH_AUTO_FREE (path); + + path =btrfs_alloc_path (); +if (!path) +return -ENOMEM ; + + key.objectid = *logical; + key.type = (u8)-1 ; + key.offset = (u64)-1 ; + + ret =btrfs_search_slot (NULL , fs_info->remap_root, &key, path, +0 ,0 ); +if (ret <0 ) +return ret; + + leaf = path->nodes[0 ]; + +if (path->slots[0 ] ==0 ) +return -ENOENT ; + + path->slots[0 ]--; + +btrfs_item_key_to_cpu (leaf, &found_key, path->slots[0 ]); + +if (found_key.type !=BTRFS_REMAP_KEY && + found_key.type !=BTRFS_IDENTITY_REMAP_KEY ) { +return -ENOENT ; + } + +if (found_key.objectid > *logical || + found_key.objectid + found_key.offset <= *logical) { +return -ENOENT ; + } + +if (*logical + *length > found_key.objectid + found_key.offset) + *length = found_key.objectid + found_key.offset - *logical; + +if (found_key.type ==BTRFS_IDENTITY_REMAP_KEY ) +return 0 ; + + remap = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_remap_item); + + *logical +=btrfs_remap_address (leaf, remap) - found_key.objectid; + +return 0 ; +} +/* * function to relocate all extents in a block group. */
fs/btrfs/relocation.h +2 -0
@@ -31,5 +31,7 @@ int btrfs_should_cancel_balance(const struct btrfs_fs_info *fs_info);struct btrfs_root *find_reloc_root (struct btrfs_fs_info *fs_info,u64 bytenr);bool btrfs_should_ignore_reloc_root (const struct btrfs_root *root);u64 btrfs_get_reloc_bg_bytenr (const struct btrfs_fs_info *fs_info); +int btrfs_translate_remap (struct btrfs_fs_info *fs_info,u64 *logical, +u64 *length);#endif
fs/btrfs/volumes.c +19 -0
@@ -6584,6 +6584,25 @@ int btrfs_map_block(struct btrfs_fs_info *fs_info, enum btrfs_map_op op,if (IS_ERR (map))return PTR_ERR (map); +if (map->type &BTRFS_BLOCK_GROUP_REMAPPED ) { +u64 new_logical = logical; + + ret =btrfs_translate_remap (fs_info, &new_logical, length); +if (ret) +return ret; + +if (new_logical != logical) { +btrfs_free_chunk_map (map); + + map =btrfs_get_chunk_map (fs_info, new_logical, + *length); +if (IS_ERR (map)) +return PTR_ERR (map); + + logical = new_logical; + } + } + num_copies =btrfs_chunk_map_num_copies (map);if (io_geom.mirror_num > num_copies)return -EINVAL ;
Signature · 1 line
2.51.2
PATCH v8 10/17 btrfs: handle deletions from remapped block group
In reply to Mark HarmstoneHandle the case where we free an extent from a block group that has the REMAPPED flag set. Because the remap tree is orthogonal to the extent tree, for data this may be within any number of identity remaps or actual remaps. If we're freeing a metadata node, this will be wholly inside one or the other.
btrfs_remove_extent_from_remap_tree() searches the remap tree for the remaps that cover the range in question, then calls remove_range_from_remap_tree() for each one, to punch a hole in the remap and adjust the free-space tree.
For an identity remap, remove_range_from_remap_tree() will adjust the block group's identity_remap_count if this changes. If it reaches zero we mark the block group as fully remapped.
For an identity remap, remove_range_from_remap_tree() will adjust the block group's identity_remap_count if this changes. If it reaches zero we mark the block group as fully remapped.
Fully remapped block groups have their chunk stripes removed and their device extents freed, which makes the disk space available again to the chunk allocator. This happens asynchronously: in the cleaner thread for sync discard and nodiscard, and (in a later patch) in the discard worker for async discard.
--- fs/btrfs/block-group.c | 98 ++++++--- fs/btrfs/block-group.h | 4 + fs/btrfs/disk-io.c | 6 + fs/btrfs/extent-tree.c | 98 ++++++++- fs/btrfs/extent-tree.h | 2 + fs/btrfs/fs.h | 4 +- fs/btrfs/relocation.c | 453 +++++++++++++++++++++++++++++++++++++++++ fs/btrfs/relocation.h | 5 + fs/btrfs/volumes.c | 57 ++++-- fs/btrfs/volumes.h | 6 + 10 files changed, 678 insertions(+), 55 deletions(-)
fs/btrfs/block-group.c +72 -26
@@ -1067,6 +1067,29 @@ static int remove_block_group_item(struct btrfs_trans_handle *trans,return btrfs_del_item (trans, root, path); } +void btrfs_remove_bg_from_sinfo (struct btrfs_block_group *bg) +{ +int factor =btrfs_bg_type_to_factor (bg->flags); + +spin_lock (&bg->space_info->lock); + +if (btrfs_test_opt (bg->fs_info,ENOSPC_DEBUG )) { +WARN_ON (bg->space_info->total_bytes < bg->length); +WARN_ON (bg->space_info->bytes_readonly + < bg->length - bg->zone_unusable); +WARN_ON (bg->space_info->bytes_zone_unusable + < bg->zone_unusable); +WARN_ON (bg->space_info->disk_total < bg->length * factor); + } + bg->space_info->total_bytes -= bg->length; + bg->space_info->bytes_readonly -= (bg->length - bg->zone_unusable); +btrfs_space_info_update_bytes_zone_unusable (bg->space_info, + -bg->zone_unusable); + bg->space_info->disk_total -= bg->length * factor; + +spin_unlock (&bg->space_info->lock); +} +int btrfs_remove_block_group (struct btrfs_trans_handle *trans,struct btrfs_chunk_map *map) {
@@ -1078,7 +1101,6 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans,struct kobject *kobj =NULL ;int ret;int index; -int factor;struct btrfs_caching_control *caching_ctl =NULL ;bool remove_map;bool remove_rsv = false;
@@ -1087,7 +1109,7 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans,if (!block_group)return -ENOENT ; -BUG_ON (!block_group->ro); +BUG_ON (!block_group->ro && !(block_group->flags &BTRFS_BLOCK_GROUP_REMAPPED ));trace_btrfs_remove_block_group (block_group);/*
@@ -1099,7 +1121,6 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans, block_group->length); index =btrfs_bg_flags_to_raid_index (block_group->flags); - factor =btrfs_bg_type_to_factor (block_group->flags);/* make sure this block group isn't part of an allocation cluster */ cluster = &fs_info->data_alloc_cluster;
@@ -1223,26 +1244,11 @@ int btrfs_remove_block_group(struct btrfs_trans_handle *trans,spin_lock (&block_group->space_info->lock);list_del_init (&block_group->ro_list); - -if (btrfs_test_opt (fs_info,ENOSPC_DEBUG )) { -WARN_ON (block_group->space_info->total_bytes - < block_group->length); -WARN_ON (block_group->space_info->bytes_readonly - < block_group->length - block_group->zone_unusable); -WARN_ON (block_group->space_info->bytes_zone_unusable - < block_group->zone_unusable); -WARN_ON (block_group->space_info->disk_total - < block_group->length * factor); - } - block_group->space_info->total_bytes -= block_group->length; - block_group->space_info->bytes_readonly -= - (block_group->length - block_group->zone_unusable); -btrfs_space_info_update_bytes_zone_unusable (block_group->space_info, - -block_group->zone_unusable); - block_group->space_info->disk_total -= block_group->length * factor; -spin_unlock (&block_group->space_info->lock); +if (!(block_group->flags &BTRFS_BLOCK_GROUP_REMAPPED )) +btrfs_remove_bg_from_sinfo (block_group); +/* * Remove the free space for the block group from the free space tree * and the block group's item from the extent tree before marking the
@@ -1576,8 +1582,10 @@ void btrfs_delete_unused_bgs(struct btrfs_fs_info *fs_info)spin_lock (&space_info->lock);spin_lock (&block_group->lock); -if (btrfs_is_block_group_used (block_group) || block_group->ro || -list_is_singular (&block_group->list)) { +if (btrfs_is_block_group_used (block_group) || + (block_group->ro && !(block_group->flags &BTRFS_BLOCK_GROUP_REMAPPED )) || +list_is_singular (&block_group->list) || +test_bit (BLOCK_GROUP_FLAG_FULLY_REMAPPED , &block_group->runtime_flags)) {/* * We want to bail if we made new allocations or have * outstanding allocations in this block group. We do
@@ -1618,9 +1626,10 @@ void btrfs_delete_unused_bgs(struct btrfs_fs_info *fs_info)* needing to allocate extents from the block group. */ used =btrfs_space_info_used (space_info, true); -if ((space_info->total_bytes - block_group->length < used && - block_group->zone_unusable < block_group->length) || -has_unwritten_metadata (block_group)) { +if (((space_info->total_bytes - block_group->length < used && + block_group->zone_unusable < block_group->length) || +has_unwritten_metadata (block_group)) && + !(block_group->flags &BTRFS_BLOCK_GROUP_REMAPPED )) {/* * Add a reference for the list, compensate for the ref * drop under the "next" label for the
@@ -1785,6 +1794,12 @@ void btrfs_mark_bg_unused(struct btrfs_block_group *bg)btrfs_get_block_group (bg);trace_btrfs_add_unused_block_group (bg);list_add_tail (&bg->bg_list, &fs_info->unused_bgs); + }else if (bg->flags &BTRFS_BLOCK_GROUP_REMAPPED && + bg->identity_remap_count ==0 ) { +/* +* Leave fully remapped block groups on the +* fully_remapped_bgs list. +*/ }else if (!test_bit (BLOCK_GROUP_FLAG_NEW , &bg->runtime_flags)) {/* Pull out the block group from the reclaim_bgs list. */ trace_btrfs_add_unused_block_group (bg);
@@ -4594,6 +4609,14 @@ int btrfs_free_block_groups(struct btrfs_fs_info *info)list_del_init (&block_group->bg_list);btrfs_put_block_group (block_group); } + +while (!list_empty (&info->fully_remapped_bgs)) { + block_group =list_first_entry (&info->fully_remapped_bgs, + struct btrfs_block_group, + bg_list); +list_del_init (&block_group->bg_list); +btrfs_put_block_group (block_group); + }spin_unlock (&info->unused_bgs_lock);spin_lock (&info->zone_active_bgs_lock);
@@ -4781,3 +4804,26 @@ bool btrfs_block_group_should_use_size_class(const struct btrfs_block_group *bg)return false;return true; } + +void btrfs_mark_bg_fully_remapped (struct btrfs_block_group *bg, +struct btrfs_trans_handle *trans) +{ +struct btrfs_fs_info *fs_info = trans->fs_info; + +spin_lock (&fs_info->unused_bgs_lock); + +/* +* The block group might already be on the unused_bgs list, remove it +* if it is. It'll get readded after the async discard worker finishes, +* or in btrfs_handle_fully_remapped_bgs() if we're not using async +* discard. +*/ +if (!list_empty (&bg->bg_list)) +list_del (&bg->bg_list); +else +btrfs_get_block_group (bg); + +list_add_tail (&bg->bg_list, &fs_info->fully_remapped_bgs); + +spin_unlock (&fs_info->unused_bgs_lock); +}
fs/btrfs/block-group.h +4 -0
@@ -92,6 +92,7 @@ enum btrfs_block_group_flags {* transaction. */ BLOCK_GROUP_FLAG_NEW , + BLOCK_GROUP_FLAG_FULLY_REMAPPED, }; enum btrfs_caching_type {
@@ -336,6 +337,7 @@ int btrfs_add_new_free_space(struct btrfs_block_group *block_group,struct btrfs_trans_handle *btrfs_start_trans_remove_block_group (struct btrfs_fs_info *fs_info,const u64 chunk_offset); +void btrfs_remove_bg_from_sinfo (struct btrfs_block_group *bg);int btrfs_remove_block_group (struct btrfs_trans_handle *trans,struct btrfs_chunk_map *map);void btrfs_delete_unused_bgs (struct btrfs_fs_info *fs_info);
@@ -407,5 +409,7 @@ int btrfs_use_block_group_size_class(struct btrfs_block_group *bg,enum btrfs_block_group_size_class size_class, bool force_wrong_size_class);bool btrfs_block_group_should_use_size_class (const struct btrfs_block_group *bg); +void btrfs_mark_bg_fully_remapped (struct btrfs_block_group *bg, +struct btrfs_trans_handle *trans);#endif /* BTRFS_BLOCK_GROUP_H */
fs/btrfs/disk-io.c +6 -0
@@ -1495,6 +1495,10 @@ static int cleaner_kthread(void *arg)*/ btrfs_run_defrag_inodes (fs_info); +if (btrfs_fs_incompat (fs_info,REMAP_TREE ) && + !btrfs_test_opt (fs_info,DISCARD_ASYNC )) +btrfs_handle_fully_remapped_bgs (fs_info); +/* * Acquires fs_info->reclaim_bgs_lock to avoid racing * with relocation (btrfs_relocate_chunk) and relocation
@@ -2835,6 +2839,7 @@ void btrfs_init_fs_info(struct btrfs_fs_info *fs_info)INIT_LIST_HEAD (&fs_info->tree_mod_seq_list);INIT_LIST_HEAD (&fs_info->unused_bgs);INIT_LIST_HEAD (&fs_info->reclaim_bgs); +INIT_LIST_HEAD (&fs_info->fully_remapped_bgs);INIT_LIST_HEAD (&fs_info->zone_active_bgs);#ifdef CONFIG_BTRFS_DEBUG INIT_LIST_HEAD (&fs_info->allocated_roots);
@@ -2890,6 +2895,7 @@ void btrfs_init_fs_info(struct btrfs_fs_info *fs_info)mutex_init (&fs_info->chunk_mutex);mutex_init (&fs_info->transaction_kthread_mutex);mutex_init (&fs_info->cleaner_mutex); +mutex_init (&fs_info->remap_mutex);mutex_init (&fs_info->ro_block_group_mutex);init_rwsem (&fs_info->commit_root_sem);init_rwsem (&fs_info->cleanup_work_sem);
fs/btrfs/extent-tree.c +92 -6
132 lines not on this page · raw message
fs/btrfs/extent-tree.h +2 -0
@@ -163,5 +163,7 @@ void btrfs_error_unpin_extent_range(struct btrfs_fs_info *fs_info, u64 start, u6int btrfs_discard_extent (struct btrfs_fs_info *fs_info,u64 bytenr,u64 num_bytes,u64 *actual_bytes);int btrfs_trim_fs (struct btrfs_fs_info *fs_info,struct fstrim_range *range); +void btrfs_handle_fully_remapped_bgs (struct btrfs_fs_info *fs_info); +int btrfs_complete_bg_remapping (struct btrfs_block_group *bg);#endif
fs/btrfs/fs.h +3 -1
@@ -579,6 +579,7 @@ struct btrfs_fs_info {struct mutex transaction_kthread_mutex;struct mutex cleaner_mutex;struct mutex chunk_mutex; +struct mutex remap_mutex;/* * This is taken to make sure we don't set block groups ro after the
@@ -832,10 +833,11 @@ struct btrfs_fs_info {struct list_head reclaim_bgs;int bg_reclaim_threshold; -/* Protects the lists unused_bgs and reclaim_bgs. */ +/* Protects the lists unused_bgs, reclaim_bgs, and fully_remapped_bgs. */ spinlock_t unused_bgs_lock;/* Protected by unused_bgs_lock. */ struct list_head unused_bgs; +struct list_head fully_remapped_bgs;struct mutex unused_bg_unpin_mutex;/* Protect block groups that are going to be deleted */ struct mutex reclaim_bgs_lock;
fs/btrfs/relocation.c +453 -0
468 lines not on this page · raw message
fs/btrfs/relocation.h +5 -0
@@ -33,5 +33,10 @@ bool btrfs_should_ignore_reloc_root(const struct btrfs_root *root);u64 btrfs_get_reloc_bg_bytenr (const struct btrfs_fs_info *fs_info);int btrfs_translate_remap (struct btrfs_fs_info *fs_info,u64 *logical,u64 *length); +int btrfs_remove_extent_from_remap_tree (struct btrfs_trans_handle *trans, +struct btrfs_path *path, +u64 bytenr,u64 num_bytes); +int btrfs_last_identity_remap_gone (struct btrfs_chunk_map *chunk_map, +struct btrfs_block_group *bg);#endif
fs/btrfs/volumes.c +35 -22
109 lines not on this page · raw message
fs/btrfs/volumes.h +6 -0
@@ -794,6 +794,8 @@ u64 btrfs_calc_stripe_length(const struct btrfs_chunk_map *map);int btrfs_nr_parity_stripes (u64 type);int btrfs_chunk_alloc_add_chunk_item (struct btrfs_trans_handle *trans,struct btrfs_block_group *bg); +int btrfs_remove_dev_extents (struct btrfs_trans_handle *trans, +struct btrfs_chunk_map *map);int btrfs_remove_chunk (struct btrfs_trans_handle *trans,u64 chunk_offset);#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS
@@ -905,6 +907,10 @@ bool btrfs_repair_one_zone(struct btrfs_fs_info *fs_info, u64 logical);bool btrfs_pinned_by_swapfile (struct btrfs_fs_info *fs_info,void *ptr);const u8 *btrfs_sb_fsid_ptr (const struct btrfs_super_block *sb); +int btrfs_update_device (struct btrfs_trans_handle *trans, +struct btrfs_device *device); +void btrfs_chunk_map_device_clear_bits (struct btrfs_chunk_map *map, +unsigned int bits);#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS struct btrfs_io_context *alloc_btrfs_io_context(struct btrfs_fs_info *fs_info,
Signature · 1 line
2.51.2
PATCH v8 11/17 btrfs: handle setting up relocation of block group with remap-tree
In reply to Mark HarmstoneHandle the preliminary work for relocating a block group in a filesystem with the remap-tree flag set.
If the block group is SYSTEM btrfs_relocate_block_group() proceeds as it does already, as bootstrapping issues mean that these block groups have to be processed the existing way. Similarly with METADATA_REMAP blocks, which are dealt with in a later patch.
Otherwise we walk the free-space tree for the block group in question, recording any holes. These get converted into identity remaps and placed in the remap tree, and the block group's REMAPPED flag is set. From now on no new allocations are possible within this block group, and any I/O to it will be funnelled through btrfs_translate_remap(). We store the number of identity remaps in identity_remap_count, so that we know when we've removed the last one and the block group is fully remapped.
The change in btrfs_read_roots() is because data relocations no longer rely on the data reloc tree as a hidden subvolume in which to do snapshots.
(Thanks to Sun YangKai for his suggestions.)
--- fs/btrfs/block-group.c | 6 +- fs/btrfs/block-group.h | 4 + fs/btrfs/free-space-tree.c | 4 +- fs/btrfs/free-space-tree.h | 5 +- fs/btrfs/relocation.c | 516 +++++++++++++++++++++++++++++++++---- fs/btrfs/relocation.h | 11 + fs/btrfs/space-info.c | 9 +- fs/btrfs/volumes.c | 89 ++++--- 8 files changed, 551 insertions(+), 93 deletions(-)
fs/btrfs/block-group.c +5 -1
@@ -2405,6 +2405,7 @@ static int read_one_block_group(struct btrfs_fs_info *info, cache->used =btrfs_stack_block_group_v2_used (bgi); cache->last_used = cache->used; cache->flags =btrfs_stack_block_group_v2_flags (bgi); + cache->last_flags = cache->flags; cache->global_root_id =btrfs_stack_block_group_v2_chunk_objectid (bgi); cache->space_info =btrfs_find_space_info (info, cache->flags); cache->remap_bytes =btrfs_stack_block_group_v2_remap_bytes (bgi);
@@ -2714,6 +2715,7 @@ static int insert_block_group_item(struct btrfs_trans_handle *trans, block_group->last_remap_bytes = block_group->remap_bytes; block_group->last_identity_remap_count = block_group->identity_remap_count; + block_group->last_flags = block_group->flags; key.objectid = block_group->start; key.type =BTRFS_BLOCK_GROUP_ITEM_KEY ; key.offset = block_group->length;
@@ -3202,13 +3204,15 @@ static int update_block_group_item(struct btrfs_trans_handle *trans,/* No change in values, can safely skip it. */ if (cache->last_used == used && cache->last_remap_bytes == remap_bytes && - cache->last_identity_remap_count == identity_remap_count) { + cache->last_identity_remap_count == identity_remap_count && + cache->last_flags == cache->flags) {spin_unlock (&cache->lock);return 0 ; } cache->last_used = used; cache->last_remap_bytes = remap_bytes; cache->last_identity_remap_count = identity_remap_count; + cache->last_flags = cache->flags;spin_unlock (&cache->lock); key.objectid = cache->start;
fs/btrfs/block-group.h +4 -0
@@ -147,6 +147,10 @@ struct btrfs_block_group {* The last commited identity_remap_count value of this block group. */ u32 last_identity_remap_count; +/* +* The last committed flags value for this block group. +*/ +u64 last_flags;/* * If the free space extent count exceeds this number, convert the block
fs/btrfs/free-space-tree.c +1 -3
@@ -21,8 +21,7 @@ static int __add_block_group_free_space(struct btrfs_trans_handle *trans,struct btrfs_block_group *block_group, struct btrfs_path *path); -static struct btrfs_root *btrfs_free_space_root ( -struct btrfs_block_group *block_group) +struct btrfs_root *btrfs_free_space_root (struct btrfs_block_group *block_group) {struct btrfs_key key = { .objectid =BTRFS_FREE_SPACE_TREE_OBJECTID ,
@@ -93,7 +92,6 @@ static int add_new_free_space_info(struct btrfs_trans_handle *trans,return 0 ; } -EXPORT_FOR_TESTS struct btrfs_free_space_info *btrfs_search_free_space_info( struct btrfs_trans_handle *trans, struct btrfs_block_group *block_group,
fs/btrfs/free-space-tree.h +3 -2
@@ -36,12 +36,13 @@ int btrfs_add_to_free_space_tree(struct btrfs_trans_handle *trans,int btrfs_remove_from_free_space_tree (struct btrfs_trans_handle *trans,u64 start,u64 size);int btrfs_delete_orphan_free_space_entries (struct btrfs_fs_info *fs_info); - -#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS struct btrfs_free_space_info *btrfs_search_free_space_info (struct btrfs_trans_handle *trans,struct btrfs_block_group *block_group,struct btrfs_path *path,int cow); +struct btrfs_root *btrfs_free_space_root (struct btrfs_block_group *block_group); + +#ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS int __btrfs_add_to_free_space_tree (struct btrfs_trans_handle *trans,struct btrfs_block_group *block_group,struct btrfs_path *path,u64 start,u64 size);
fs/btrfs/relocation.c +463 -53
580 lines not on this page · raw message
fs/btrfs/relocation.h +11 -0
@@ -12,6 +12,17 @@ struct btrfs_trans_handle;struct btrfs_ordered_extent ;struct btrfs_pending_snapshot ; +static inline bool should_relocate_using_remap_tree (struct btrfs_block_group *bg) +{ +if (!btrfs_fs_incompat (bg->fs_info,REMAP_TREE )) +return false; + +if (bg->flags & (BTRFS_BLOCK_GROUP_SYSTEM |BTRFS_BLOCK_GROUP_METADATA_REMAP )) +return false; + +return true; +} +int btrfs_relocate_block_group (struct btrfs_fs_info *fs_info,u64 group_start,bool verbose);int btrfs_init_reloc_root (struct btrfs_trans_handle *trans,struct btrfs_root *root);
fs/btrfs/space-info.c +7 -2
@@ -376,8 +376,13 @@ void btrfs_add_bg_to_space_info(struct btrfs_fs_info *info, factor =btrfs_bg_type_to_factor (block_group->flags);spin_lock (&space_info->lock); - space_info->total_bytes += block_group->length; - space_info->disk_total += block_group->length * factor; + +if (!(block_group->flags &BTRFS_BLOCK_GROUP_REMAPPED ) || + block_group->identity_remap_count !=0 ) { + space_info->total_bytes += block_group->length; + space_info->disk_total += block_group->length * factor; + } + space_info->bytes_used += block_group->used; space_info->disk_used += block_group->used * factor; space_info->bytes_readonly += block_group->bytes_super;
fs/btrfs/volumes.c +57 -32
116 lines not on this page · raw message
Signature · 1 line
2.51.2
PATCH v8 12/17 btrfs: move existing remaps before relocating block group
In reply to Mark HarmstoneIf when relocating a block group we find that remap_bytes > 0 in its block group item, that means that it has been the destination block group for another that has been remapped.
We need to seach the remap tree for any remap backrefs within this range, and move the data to a third block group. This is because otherwise btrfs_translate_remap() could end up following an unbounded chain of remaps, which would only get worse over time.
We only relocate one block group at a time, so remap_bytes will only ever go down while we are doing this. Once we're finished we set the REMAPPED flag on the block group, which will permanently prevent any other data from being moved to within it.
--- fs/btrfs/bio.c | 3 +- fs/btrfs/bio.h | 3 + fs/btrfs/extent-tree.c | 6 +- fs/btrfs/relocation.c | 481 +++++++++++++++++++++++++++++++++++++++++ 4 files changed, 490 insertions(+), 3 deletions(-)
fs/btrfs/bio.c +2 -1
@@ -826,7 +826,8 @@ static bool btrfs_submit_chunk(struct btrfs_bio *bbio, int mirror_num)*/ if (!(inode->flags &BTRFS_INODE_NODATASUM ) && !test_bit (BTRFS_FS_STATE_NO_DATA_CSUMS , &fs_info->fs_state) && - !btrfs_is_data_reloc_root (inode->root)) { + !btrfs_is_data_reloc_root (inode->root) && + !bbio->is_remap) {if (should_async_write (bbio) &&btrfs_wq_submit_bio (bbio, bioc, &smap, mirror_num)) goto done;
fs/btrfs/bio.h +3 -0
@@ -90,6 +90,9 @@ struct btrfs_bio {*/ bool is_scrub:1; +/* Whether the bio is coming from copy_remapped_data_io(). */ + bool is_remap:1; +/* Whether the csum generation for data write is async. */ bool async_csum:1;
fs/btrfs/extent-tree.c +4 -2
@@ -4557,7 +4557,8 @@ static noinline int find_free_extent(struct btrfs_root *root, block_group->cached != BTRFS_CACHE_NO) {down_read (&space_info->groups_sem);if (list_empty (&block_group->list) || - block_group->ro) { + block_group->ro || + block_group->flags &BTRFS_BLOCK_GROUP_REMAPPED ) {/* * someone is removing this block group, * we can't jump into the have_block_group
@@ -4591,7 +4592,8 @@ static noinline int find_free_extent(struct btrfs_root *root, ffe_ctl->hinted = false;/* If the block group is read-only, we can skip it entirely. */ -if (unlikely (block_group->ro)) { +if (unlikely (block_group->ro || + block_group->flags &BTRFS_BLOCK_GROUP_REMAPPED )) {if (ffe_ctl->for_treelog)btrfs_clear_treelog_bg (block_group);if (ffe_ctl->for_data_reloc)
fs/btrfs/relocation.c +481 -0
493 lines not on this page · raw message
Signature · 1 line
2.51.2
PATCH v8 13/17 btrfs: replace identity remaps with actual remaps when doing relocations
In reply to Mark HarmstoneAdd a function do_remap_tree_reloc(), which does the actual work of doing a relocation using the remap tree.
In a loop we call do_remap_reloc_trans(), which searches for the first identity remap for the block group. We call btrfs_reserve_extent() to find space elsewhere for it, and read the data into memory and write it to the new location. We then carve out the identity remap and replace it with an actual remap, which points to the new location in which to look.
Once the last identity remap has been removed we call last_identity_remap_gone(), which, as with deletions, removes the chunk's stripes and device extents.
--- fs/btrfs/relocation.c | 337 ++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 337 insertions(+)
fs/btrfs/relocation.c +337 -0
@@ -4633,6 +4633,61 @@ static int create_remap_tree_entries(struct btrfs_trans_handle *trans,return ret; } +static int find_next_identity_remap (struct btrfs_trans_handle *trans, +struct btrfs_path *path,u64 bg_end, +u64 last_start,u64 *start, +u64 *length) +{ +int ret; +struct btrfs_key key, found_key; +struct btrfs_root *remap_root = trans->fs_info->remap_root; +struct extent_buffer *leaf; + + key.objectid = last_start; + key.type =BTRFS_IDENTITY_REMAP_KEY ; + key.offset =0 ; + + ret =btrfs_search_slot (trans, remap_root, &key, path,0 ,0 ); +if (ret <0 ) + goto out; + + leaf = path->nodes[0 ]; +while (true) { +if (path->slots[0 ] >=btrfs_header_nritems (leaf)) { + ret =btrfs_next_leaf (remap_root, path); + +if (ret !=0 ) { +if (ret ==1 ) + ret = -ENOENT ; + goto out; + } + + leaf = path->nodes[0 ]; + } + +btrfs_item_key_to_cpu (leaf, &found_key, path->slots[0 ]); + +if (found_key.objectid >= bg_end) { + ret = -ENOENT ; + goto out; + } + +if (found_key.type ==BTRFS_IDENTITY_REMAP_KEY ) { + *start = found_key.objectid; + *length = found_key.offset; + ret =0 ; + goto out; + } + + path->slots[0 ]++; + } + +out: +btrfs_release_path (path); + +return ret; +} +static int remove_chunk_stripes (struct btrfs_trans_handle *trans,struct btrfs_chunk_map *chunk_map,struct btrfs_path *path)
@@ -4779,6 +4834,96 @@ static void adjust_identity_remap_count(struct btrfs_trans_handle *trans,btrfs_mark_bg_fully_remapped (bg, trans); } +static int add_remap_entry (struct btrfs_trans_handle *trans, +struct btrfs_path *path, +struct btrfs_block_group *src_bg,u64 old_addr, +u64 new_addr,u64 length) +{ +struct btrfs_fs_info *fs_info = trans->fs_info; +struct btrfs_key key, new_key; +int ret; +int identity_count_delta =0 ; + + key.objectid = old_addr; + key.type = (u8)-1 ; + key.offset = (u64)-1 ; + + ret =btrfs_search_slot (trans, fs_info->remap_root, &key, path,-1 ,1 ); +if (ret <0 ) + goto end; + +if (path->slots[0 ] ==0 ) { + ret = -ENOENT ; + goto end; + } + + path->slots[0 ]--; + +btrfs_item_key_to_cpu (path->nodes[0 ], &key, path->slots[0 ]); + +if (key.type !=BTRFS_IDENTITY_REMAP_KEY || + key.objectid > old_addr || + key.objectid + key.offset <= old_addr) { + ret = -ENOENT ; + goto end; + } + +/* Shorten or delete identity mapping entry. */ + +if (key.objectid == old_addr) { + ret =btrfs_del_item (trans, fs_info->remap_root, path); +if (ret) + goto end; + + identity_count_delta--; + }else { + new_key.objectid = key.objectid; + new_key.type =BTRFS_IDENTITY_REMAP_KEY ; + new_key.offset = old_addr - key.objectid; + +btrfs_set_item_key_safe (trans, path, &new_key); + } + +btrfs_release_path (path); + +/* Create new remap entry. */ + + ret =add_remap_item (trans, path, new_addr, length, old_addr); +if (ret) + goto end; + +/* Add entry for remainder of identity mapping, if necessary. */ + +if (key.objectid + key.offset != old_addr + length) { + new_key.objectid = old_addr + length; + new_key.type =BTRFS_IDENTITY_REMAP_KEY ; + new_key.offset = key.objectid + key.offset - old_addr - length; + + ret =btrfs_insert_empty_item (trans, fs_info->remap_root, + path, &new_key,0 ); +if (ret) + goto end; + +btrfs_release_path (path); + + identity_count_delta++; + } + +/* Add backref. */ + + ret =add_remap_backref_item (trans, path, new_addr, length, old_addr); +if (ret) + goto end; + +if (identity_count_delta !=0 ) +adjust_identity_remap_count (trans, src_bg, identity_count_delta); + +end: +btrfs_release_path (path); + +return ret; +} +static int mark_chunk_remapped (struct btrfs_trans_handle *trans,struct btrfs_path *path,uint64_t start) {
@@ -4828,6 +4973,190 @@ static int mark_chunk_remapped(struct btrfs_trans_handle *trans,return ret; } +static int do_remap_reloc_trans (struct btrfs_fs_info *fs_info, +struct btrfs_block_group *src_bg, +struct btrfs_path *path,u64 *last_start) +{ +struct btrfs_trans_handle *trans; +struct btrfs_root *extent_root; +struct btrfs_key ins; +struct btrfs_block_group *dest_bg =NULL ; +u64 start, remap_length, length, new_addr, min_size; +int ret; +bool no_more = false; +bool is_data = src_bg->flags &BTRFS_BLOCK_GROUP_DATA ; +bool made_reservation = false, bg_needs_free_space; +struct btrfs_space_info *sinfo = src_bg->space_info; + + extent_root =btrfs_extent_root (fs_info, src_bg->start); + + trans =btrfs_start_transaction (extent_root,0 ); +if (IS_ERR (trans)) +return PTR_ERR (trans); + +mutex_lock (&fs_info->remap_mutex); + + ret =find_next_identity_remap (trans, path, src_bg->start + src_bg->length, + *last_start, &start, &remap_length); +if (ret == -ENOENT ) { + no_more = true; + goto next; + }else if (ret) { +mutex_unlock (&fs_info->remap_mutex); +btrfs_end_transaction (trans); +return ret; + } + +/* Try to reserve enough space for block. */ + +spin_lock (&sinfo->lock); +btrfs_space_info_update_bytes_may_use (sinfo, remap_length); +spin_unlock (&sinfo->lock); + +if (is_data) + min_size = fs_info->sectorsize; +else + min_size = fs_info->nodesize; + +/* +* We're using btrfs_reserve_extent() to allocate a contiguous +* logical address range, but this will become a remap item rather than +* an extent in the extent tree. +* +* Short allocations are fine: it means that we chop off the beginning +* of the identity remap that we're processing, and will tackle the +* rest of it the next time round. +*/ + ret =btrfs_reserve_extent (fs_info->fs_root, remap_length, + remap_length, min_size, +0 ,0 , &ins, is_data, false); +if (ret) { +spin_lock (&sinfo->lock); +btrfs_space_info_update_bytes_may_use (sinfo, -remap_length); +spin_unlock (&sinfo->lock); + +mutex_unlock (&fs_info->remap_mutex); +btrfs_end_transaction (trans); +return ret; + } + + made_reservation = true; + + new_addr = ins.objectid; + length = ins.offset; + +if (!is_data && !IS_ALIGNED (length, fs_info->nodesize)) { +u64 new_length =ALIGN_DOWN (length, fs_info->nodesize); + +btrfs_free_reserved_extent (fs_info, new_addr + new_length, + length - new_length,0 ); + + length = new_length; + } + + dest_bg =btrfs_lookup_block_group (fs_info, new_addr); + +mutex_lock (&dest_bg->free_space_lock); + bg_needs_free_space =test_bit (BLOCK_GROUP_FLAG_NEEDS_FREE_SPACE , + &dest_bg->runtime_flags); +mutex_unlock (&dest_bg->free_space_lock); + +if (bg_needs_free_space) { + ret =btrfs_add_block_group_free_space (trans, dest_bg); +if (ret) + goto fail; + } + + ret =copy_remapped_data (fs_info, start, new_addr, length); +if (ret) + goto fail; + + ret =btrfs_remove_from_free_space_tree (trans, new_addr, length); +if (ret) + goto fail; + + ret =add_remap_entry (trans, path, src_bg, start, new_addr, length); +if (ret) { +btrfs_add_to_free_space_tree (trans, new_addr, length); + goto fail; + } + +adjust_block_group_remap_bytes (trans, dest_bg, length); +btrfs_free_reserved_bytes (dest_bg, length,0 ); + +spin_lock (&sinfo->lock); + sinfo->bytes_readonly += length; +spin_unlock (&sinfo->lock); + +next: +if (dest_bg) +btrfs_put_block_group (dest_bg); + +if (made_reservation) +btrfs_dec_block_group_reservations (fs_info, new_addr); + +mutex_unlock (&fs_info->remap_mutex); + +if (src_bg->identity_remap_count ==0 ) { +bool mark_fully_remapped = false; + +spin_lock (&src_bg->lock); + +if (!test_bit (BLOCK_GROUP_FLAG_FULLY_REMAPPED , &src_bg->runtime_flags)) { + mark_fully_remapped = true; +set_bit (BLOCK_GROUP_FLAG_FULLY_REMAPPED , + &src_bg->runtime_flags); + } + +spin_unlock (&src_bg->lock); + +if (mark_fully_remapped) +btrfs_mark_bg_fully_remapped (src_bg, trans); + } + + ret =btrfs_end_transaction (trans); +if (ret) +return ret; + +if (no_more) +return 1 ; + + *last_start = start; + +return 0 ; + +fail: +if (dest_bg) +btrfs_put_block_group (dest_bg); + +btrfs_free_reserved_extent (fs_info, new_addr, length,0 ); + +mutex_unlock (&fs_info->remap_mutex); +btrfs_end_transaction (trans); + +return ret; +} + +static int do_remap_reloc (struct btrfs_fs_info *fs_info, +struct btrfs_path *path,struct btrfs_block_group *bg) +{ +u64 last_start; +int ret; + + last_start = bg->start; + +while (true) { + ret =do_remap_reloc_trans (fs_info, bg, path, &last_start); +if (ret) { +if (ret ==1 ) + ret =0 ; +break ; + } + } + +return ret; +} +int btrfs_translate_remap (struct btrfs_fs_info *fs_info,u64 *logical,u64 *length) {
@@ -5121,6 +5450,14 @@ int btrfs_relocate_block_group(struct btrfs_fs_info *fs_info, u64 group_start, } ret =start_block_group_remapping (fs_info, path, bg); +if (ret) + goto out; + + ret =do_remap_reloc (fs_info, path, rc->block_group); +if (ret) + goto out; + +btrfs_delete_unused_bgs (fs_info); }else { ret =do_nonremap_reloc (fs_info, verbose, rc); }
Signature · 1 line
2.51.2
PATCH v8 14/17 btrfs: add do_remap param to btrfs_discard_extent()
In reply to Mark Harmstonebtrfs_discard_extent() can be called either when an extent is removed or from walking the free-space tree. With a remapped block group these two things are no longer equivalent: the extent's addresses are remapped, while the free-space tree exclusively uses underlying addresses.
Add a do_remap parameter to btrfs_discard_extent() and btrfs_map_discard(), saying whether or not the address needs to be run through the remap tree first.
--- fs/btrfs/extent-tree.c | 11 +++++++---- fs/btrfs/extent-tree.h | 2 +- fs/btrfs/free-space-cache.c | 2 +- fs/btrfs/inode.c | 2 +- fs/btrfs/volumes.c | 23 +++++++++++++++++++++-- fs/btrfs/volumes.h | 2 +- 6 files changed, 32 insertions(+), 10 deletions(-)
fs/btrfs/extent-tree.c +7 -4
@@ -1381,7 +1381,7 @@ static int do_discard_extent(struct btrfs_discard_stripe *stripe, u64 *bytes) }int btrfs_discard_extent (struct btrfs_fs_info *fs_info,u64 bytenr, -u64 num_bytes,u64 *actual_bytes) +u64 num_bytes,u64 *actual_bytes,bool do_remap) {int ret =0 ;u64 discarded_bytes =0 ;
@@ -1399,7 +1399,8 @@ int btrfs_discard_extent(struct btrfs_fs_info *fs_info, u64 bytenr,int i; num_bytes = end - cur; - stripes =btrfs_map_discard (fs_info, cur, &num_bytes, &num_stripes); + stripes =btrfs_map_discard (fs_info, cur, &num_bytes, + &num_stripes, do_remap);if (IS_ERR (stripes)) { ret =PTR_ERR (stripes);if (ret == -EOPNOTSUPP )
@@ -2932,7 +2933,8 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans)if (btrfs_test_opt (fs_info,DISCARD_SYNC )) ret =btrfs_discard_extent (fs_info, start, - end +1 - start,NULL ); + end +1 - start,NULL , + true); next_state =btrfs_next_extent_state (unpin, cached_state);btrfs_clear_extent_dirty (unpin, start, end, &cached_state);
@@ -2990,7 +2992,8 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans) ret = -EROFS ;if (!TRANS_ABORTED (trans)) ret =btrfs_discard_extent (fs_info, block_group->start, - block_group->length,NULL ); + block_group->length,NULL , + true);/* * Not strictly necessary to lock, as the block_group should be
fs/btrfs/extent-tree.h +1 -1
@@ -161,7 +161,7 @@ int btrfs_drop_subtree(struct btrfs_trans_handle *trans, struct extent_buffer *parent);void btrfs_error_unpin_extent_range (struct btrfs_fs_info *fs_info,u64 start,u64 end);int btrfs_discard_extent (struct btrfs_fs_info *fs_info,u64 bytenr, -u64 num_bytes,u64 *actual_bytes); +u64 num_bytes,u64 *actual_bytes,bool do_remap);int btrfs_trim_fs (struct btrfs_fs_info *fs_info,struct fstrim_range *range);void btrfs_handle_fully_remapped_bgs (struct btrfs_fs_info *fs_info);int btrfs_complete_bg_remapping (struct btrfs_block_group *bg);
fs/btrfs/free-space-cache.c +1 -1
@@ -3677,7 +3677,7 @@ static int do_trimming(struct btrfs_block_group *block_group, }spin_unlock (&space_info->lock); - ret =btrfs_discard_extent (fs_info, start, bytes, &trimmed); + ret =btrfs_discard_extent (fs_info, start, bytes, &trimmed, false);if (!ret) { *total_trimmed += trimmed; trim_state =BTRFS_TRIM_STATE_TRIMMED ;
fs/btrfs/inode.c +1 -1
@@ -3370,7 +3370,7 @@ int btrfs_finish_one_ordered(struct btrfs_ordered_extent *ordered_extent)btrfs_discard_extent (fs_info, ordered_extent->disk_bytenr, ordered_extent->disk_num_bytes, -NULL ); +NULL , true);btrfs_free_reserved_extent (fs_info, ordered_extent->disk_bytenr, ordered_extent->disk_num_bytes, true);
fs/btrfs/volumes.c +21 -2
@@ -3425,7 +3425,7 @@ static int btrfs_relocate_chunk_finish(struct btrfs_fs_info *fs_info,*/ if (btrfs_is_zoned (fs_info)) { ret =btrfs_discard_extent (fs_info, bg->start, length, -NULL ); +NULL , true);if (ret) btrfs_info(fs_info, "failed to reset zone %llu after relocation",
@@ -6111,7 +6111,7 @@ void btrfs_put_bioc(struct btrfs_io_context *bioc)*/ struct btrfs_discard_stripe *btrfs_map_discard (struct btrfs_fs_info *fs_info,u64 logical,u64 *length_ret, -u32 *num_stripes) +u32 *num_stripes,bool do_remap) {struct btrfs_chunk_map *map;struct btrfs_discard_stripe *stripes;
@@ -6135,6 +6135,25 @@ struct btrfs_discard_stripe *btrfs_map_discard(struct btrfs_fs_info *fs_info,if (IS_ERR (map))return ERR_CAST (map); +if (do_remap && map->type &BTRFS_BLOCK_GROUP_REMAPPED ) { +u64 new_logical = logical; + + ret =btrfs_translate_remap (fs_info, &new_logical, &length); +if (ret) + goto out_free_map; + +if (new_logical != logical) { +btrfs_free_chunk_map (map); + + map =btrfs_get_chunk_map (fs_info, new_logical, + length); +if (IS_ERR (map)) +return ERR_CAST (map); + + logical = new_logical; + } + } +/* we don't discard raid56 yet */ if (map->type &BTRFS_BLOCK_GROUP_RAID56_MASK ) { ret = -EOPNOTSUPP ;
fs/btrfs/volumes.h +1 -1
@@ -732,7 +732,7 @@ int btrfs_map_repair_block(struct btrfs_fs_info *fs_info, u32 length, int mirror_num);struct btrfs_discard_stripe *btrfs_map_discard (struct btrfs_fs_info *fs_info,u64 logical,u64 *length_ret, -u32 *num_stripes); +u32 *num_stripes,bool do_remap);int btrfs_read_sys_array (struct btrfs_fs_info *fs_info);int btrfs_read_chunk_tree (struct btrfs_fs_info *fs_info); struct btrfs_block_group *btrfs_create_chunk(struct btrfs_trans_handle *trans,
Signature · 1 line
2.51.2
PATCH v8 15/17 btrfs: allow balancing remap tree
In reply to Mark HarmstoneBalancing the METADATA_REMAP chunk, i.e. the chunk in which the remap tree lives, is a special case.
We can't use the remap tree itself for this, as then we'd have no way to boostrap it on mount. And we can't use the pre-remap tree code for this as it relies on walking the extent tree, and we're not creating backrefs for METADATA_REMAP chunks.
So instead, if a balance would relocate any METADATA_REMAP block groups, mark those block groups as readonly and COW every leaf of the remap tree.
There's more sophisticated ways of doing this, such as only COWing nodes within a block group that's to be relocated, but they're fiddly and with lots of edge cases. Plus it's not anticipated that a) the number of METADATA_REMAP chunks is going to be particularly large, or b) that users will want to only relocate some of these chunks - the main use case here is to unbreak RAID conversion and device removal.
--- fs/btrfs/volumes.c | 159 +++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 155 insertions(+), 4 deletions(-)
fs/btrfs/volumes.c +155 -4
@@ -4000,8 +4000,11 @@ static bool should_balance_chunk(struct extent_buffer *leaf, struct btrfs_chunkstruct btrfs_balance_args *bargs =NULL ;u64 chunk_type =btrfs_chunk_type (leaf, chunk); -if (chunk_type &BTRFS_BLOCK_GROUP_METADATA_REMAP ) -return false; +/* Treat METADATA_REMAP chunks as METADATA. */ +if (chunk_type &BTRFS_BLOCK_GROUP_METADATA_REMAP ) { + chunk_type &= ~BTRFS_BLOCK_GROUP_METADATA_REMAP ; + chunk_type |=BTRFS_BLOCK_GROUP_METADATA ; + }/* type filter */ if (!((chunk_type & BTRFS_BLOCK_GROUP_TYPE_MASK) &
@@ -4084,6 +4087,113 @@ static bool should_balance_chunk(struct extent_buffer *leaf, struct btrfs_chunkreturn true; } +struct remap_chunk_info { +struct list_head list; +u64 offset; +struct btrfs_block_group *bg; +bool made_ro; +}; + +static int cow_remap_tree (struct btrfs_trans_handle *trans, +struct btrfs_path *path) +{ +struct btrfs_fs_info *fs_info = trans->fs_info; +struct btrfs_key key = {0 }; +int ret; + + ret =btrfs_search_slot (trans, fs_info->remap_root, &key, path,0 ,1 ); +if (ret <0 ) +return ret; + +while (true) { + ret =btrfs_next_leaf (fs_info->remap_root, path); +if (ret <0 ) { +return ret; + }else if (ret >0 ) { + ret =0 ; +break ; + } + +btrfs_item_key_to_cpu (path->nodes[0 ], &key, path->slots[0 ]); + +btrfs_release_path (path); + + ret =btrfs_search_slot (trans, fs_info->remap_root, &key, path, +0 ,1 ); +if (ret <0 ) +break ; + } + +return ret; +} + +static int balance_remap_chunks (struct btrfs_fs_info *fs_info, +struct btrfs_path *path, +struct list_head *chunks) +{ +struct remap_chunk_info *rci, *tmp; +struct btrfs_trans_handle *trans; +int ret; + +list_for_each_entry_safe (rci, tmp, chunks, list) { + rci->bg =btrfs_lookup_block_group (fs_info, rci->offset); +if (!rci->bg) { +list_del (&rci->list); +kfree (rci); +continue ; + } + + ret =btrfs_inc_block_group_ro (rci->bg, false); +if (ret) + goto end; + + rci->made_ro = true; + } + +if (list_empty (chunks)) +return 0 ; + + trans =btrfs_start_transaction (fs_info->remap_root,0 ); +if (IS_ERR (trans)) { + ret =PTR_ERR (trans); + goto end; + } + +mutex_lock (&fs_info->remap_mutex); + + ret =cow_remap_tree (trans, path); + +mutex_unlock (&fs_info->remap_mutex); + +btrfs_release_path (path); + +btrfs_commit_transaction (trans); + +end: +while (!list_empty (chunks)) { +bool is_unused; + + rci = list_first_entry(chunks, struct remap_chunk_info, list); + +spin_lock (&rci->bg->lock); + is_unused = !btrfs_is_block_group_used (rci->bg); +spin_unlock (&rci->bg->lock); + +if (is_unused) +btrfs_mark_bg_unused (rci->bg); + +if (rci->made_ro) +btrfs_dec_block_group_ro (rci->bg); + +btrfs_put_block_group (rci->bg); + +list_del (&rci->list); +kfree (rci); + } + +return ret; +} +static int __btrfs_balance (struct btrfs_fs_info *fs_info) {struct btrfs_balance_control *bctl = fs_info->balance_ctl;
@@ -4106,6 +4216,9 @@ static int __btrfs_balance(struct btrfs_fs_info *fs_info)u32 count_meta =0 ;u32 count_sys =0 ;int chunk_reserved =0 ; +struct remap_chunk_info *rci; +unsigned int num_remap_chunks =0 ; +LIST_HEAD (remap_chunks); path =btrfs_alloc_path ();if (!path) {
@@ -4204,7 +4317,8 @@ static int __btrfs_balance(struct btrfs_fs_info *fs_info) count_data++; else if (chunk_type & BTRFS_BLOCK_GROUP_SYSTEM) count_sys++; - else if (chunk_type & BTRFS_BLOCK_GROUP_METADATA) + else if (chunk_type & (BTRFS_BLOCK_GROUP_METADATA | + BTRFS_BLOCK_GROUP_METADATA_REMAP)) count_meta++; goto loop;
@@ -4224,6 +4338,30 @@ static int __btrfs_balance(struct btrfs_fs_info *fs_info) goto loop; } +/* +* Balancing METADATA_REMAP chunks takes place separately - add +* the details to a list so it can be processed later. +*/ +if (chunk_type &BTRFS_BLOCK_GROUP_METADATA_REMAP ) { +mutex_unlock (&fs_info->reclaim_bgs_lock); + + rci =kmalloc (sizeof (struct remap_chunk_info ), +GFP_NOFS ); +if (!rci) { + ret = -ENOMEM ; + goto error; + } + + rci->offset = found_key.offset; + rci->bg =NULL ; + rci->made_ro = false; +list_add_tail (&rci->list, &remap_chunks); + + num_remap_chunks++; + + goto loop; + } +if (!chunk_reserved) {/* * We may be relocating the only data chunk we have,
@@ -4263,11 +4401,24 @@ static int __btrfs_balance(struct btrfs_fs_info *fs_info) key.offset = found_key.offset -1 ; } +btrfs_release_path (path); +if (counting) { -btrfs_release_path (path); counting = false; goto again; } + +if (!list_empty (&remap_chunks)) { + ret =balance_remap_chunks (fs_info, path, &remap_chunks); +if (ret == -ENOSPC ) + enospc_errors++; + +if (!ret) { +spin_lock (&fs_info->balance_lock); + bctl->stat.completed += num_remap_chunks; +spin_unlock (&fs_info->balance_lock); + } + } error:if (enospc_errors) { btrfs_info(fs_info, "%d enospc errors during balance",
Signature · 1 line
2.51.2
PATCH v8 16/17 btrfs: handle discarding fully-remapped block groups
In reply to Mark HarmstoneDiscard normally works by iterating over the free-space entries of a block group. This doesn't work for fully-remapped block groups, as we removed their free-space entries when we started relocation.
For sync discard, call btrfs_discard_extent() when we commit the transaction in which the last identity remap was removed.
For async discard, add a new function btrfs_trim_fully_remapped_block_group() to be called by the discard worker, which iterates over the block group's range using the normal async discard rules. Once we reach the end, remove the chunk's stripes and device extents to get back its free space.
--- fs/btrfs/block-group.c | 29 ++++++++++--------- fs/btrfs/block-group.h | 1 + fs/btrfs/discard.c | 57 ++++++++++++++++++++++++++++++++----- fs/btrfs/extent-tree.c | 3 ++ fs/btrfs/free-space-cache.c | 36 +++++++++++++++++++++++ fs/btrfs/free-space-cache.h | 1 + 6 files changed, 107 insertions(+), 20 deletions(-)
fs/btrfs/block-group.c +16 -13
@@ -4814,20 +4814,23 @@ void btrfs_mark_bg_fully_remapped(struct btrfs_block_group *bg, {struct btrfs_fs_info *fs_info = trans->fs_info; -spin_lock (&fs_info->unused_bgs_lock); +if (btrfs_test_opt (fs_info,DISCARD_ASYNC )) { +btrfs_discard_queue_work (&fs_info->discard_ctl, bg); + }else { +spin_lock (&fs_info->unused_bgs_lock); -/* -* The block group might already be on the unused_bgs list, remove it -* if it is. It'll get readded after the async discard worker finishes, -* or in btrfs_handle_fully_remapped_bgs() if we're not using async -* discard. -*/ -if (!list_empty (&bg->bg_list)) -list_del (&bg->bg_list); -else -btrfs_get_block_group (bg); +/* +* The block group might already be on the unused_bgs list, +* remove it if it is. It'll get readded after +* btrfs_handle_fully_remapped_bgs() finishes. +*/ +if (!list_empty (&bg->bg_list)) +list_del (&bg->bg_list); +else +btrfs_get_block_group (bg); -list_add_tail (&bg->bg_list, &fs_info->fully_remapped_bgs); +list_add_tail (&bg->bg_list, &fs_info->fully_remapped_bgs); -spin_unlock (&fs_info->unused_bgs_lock); +spin_unlock (&fs_info->unused_bgs_lock); + } }
fs/btrfs/block-group.h +1 -0
@@ -49,6 +49,7 @@ enum btrfs_discard_state {BTRFS_DISCARD_EXTENTS ,BTRFS_DISCARD_BITMAPS ,BTRFS_DISCARD_RESET_CURSOR , + BTRFS_DISCARD_FULLY_REMAPPED, };/*
fs/btrfs/discard.c +50 -7
109 lines not on this page · raw message
fs/btrfs/extent-tree.c +3 -0
@@ -2901,6 +2901,9 @@ void btrfs_handle_fully_remapped_bgs(struct btrfs_fs_info *fs_info)list_del_init (&bg->bg_list);spin_unlock (&fs_info->unused_bgs_lock); +btrfs_discard_extent (fs_info, bg->start, bg->length, +NULL , false); + ret =btrfs_complete_bg_remapping (bg);if (ret) {btrfs_put_block_group (bg);
fs/btrfs/free-space-cache.c +36 -0
@@ -29,6 +29,7 @@#include "file-item.h" #include "file.h" #include "super.h" +#include "relocation.h" #define BITS_PER_BITMAP (PAGE_SIZE * 8UL)#define MAX_CACHE_BYTES_PER_GIG SZ_64K
@@ -3066,6 +3067,11 @@ bool btrfs_is_free_space_trimmed(struct btrfs_block_group *block_group)struct rb_node *node;bool ret = true; +if (block_group->flags &BTRFS_BLOCK_GROUP_REMAPPED && + block_group->identity_remap_count ==0 ) { +return true; + } +spin_lock (&ctl->tree_lock); node =rb_first (&ctl->free_space_offset);
@@ -3834,6 +3840,36 @@ static int trim_no_bitmap(struct btrfs_block_group *block_group,return ret; } +void btrfs_trim_fully_remapped_block_group (struct btrfs_block_group *bg) +{ +struct btrfs_fs_info *fs_info = bg->fs_info; +struct btrfs_discard_ctl *discard_ctl = &fs_info->discard_ctl; +int ret =0 ; +u64 bytes, trimmed; +const u64 max_discard_size =READ_ONCE (discard_ctl->max_discard_size); +u64 end =btrfs_block_group_end (bg); + + bytes = end - bg->discard_cursor; + +if (max_discard_size && + bytes >= (max_discard_size + +BTRFS_ASYNC_DISCARD_MIN_FILTER )) { + bytes = max_discard_size; + } + + ret =btrfs_discard_extent (fs_info, bg->discard_cursor, bytes, &trimmed, + false); +if (ret) +return ; + + bg->discard_cursor += trimmed; + +if (bg->discard_cursor < end) +return ; + +btrfs_complete_bg_remapping (bg); +} +/* * If we break out of trimming a bitmap prematurely, we should reset the * trimming bit. In a rather contrived case, it's possible to race here so
fs/btrfs/free-space-cache.h +1 -0
@@ -166,6 +166,7 @@ int btrfs_trim_block_group_extents(struct btrfs_block_group *block_group,int btrfs_trim_block_group_bitmaps (struct btrfs_block_group *block_group,u64 *trimmed,u64 start,u64 end,u64 minlen,u64 maxlen,bool async); +void btrfs_trim_fully_remapped_block_group (struct btrfs_block_group *bg);bool btrfs_free_space_cache_v1_active (struct btrfs_fs_info *fs_info);int btrfs_set_free_space_cache_v1_active (struct btrfs_fs_info *fs_info,bool active);
Signature · 1 line
2.51.2
PATCH v8 17/17 btrfs: populate fully_remapped_bgs_list on mount
In reply to Mark HarmstoneAdd a function btrfs_populate_fully_remapped_bgs_list() which gets called on mount, which looks for fully remapped block groups (i.e. identity_remap_count == 0) which haven't yet had their chunk stripes and device extents removed.
This happens when a filesystem is unmounted while async discard has not yet finished, as otherwise the data range occupied by the chunk stripes would be permanently unusable.
--- fs/btrfs/block-group.c | 79 +++++++++++++++++++++++++++++++++++++ fs/btrfs/block-group.h | 2 + fs/btrfs/disk-io.c | 9 +++++ fs/btrfs/free-space-cache.c | 18 +++++++++ fs/btrfs/relocation.c | 4 ++ 5 files changed, 112 insertions(+)
fs/btrfs/block-group.c +79 -0
@@ -4815,6 +4815,11 @@ void btrfs_mark_bg_fully_remapped(struct btrfs_block_group *bg,struct btrfs_fs_info *fs_info = trans->fs_info;if (btrfs_test_opt (fs_info,DISCARD_ASYNC )) { +spin_lock (&bg->lock); +set_bit (BLOCK_GROUP_FLAG_STRIPE_REMOVAL_PENDING , + &bg->runtime_flags); +spin_unlock (&bg->lock); +btrfs_discard_queue_work (&fs_info->discard_ctl, bg); }else {spin_lock (&fs_info->unused_bgs_lock);
@@ -4834,3 +4839,77 @@ void btrfs_mark_bg_fully_remapped(struct btrfs_block_group *bg,spin_unlock (&fs_info->unused_bgs_lock); } } + +/* +* Compare the block group and chunk trees, and find any fully-remapped block +* groups which haven't yet had their chunk stripes and device extents removed, +* and put them on the fully_remapped_bgs list so this gets done. +* +* This happens when a block group becomes fully remapped, i.e. its last +* identity mapping is removed, and the volume is unmounted before async +* discard has finished. It's important this gets done as until it is the +* chunk's stripes are dead space. +*/ +int btrfs_populate_fully_remapped_bgs_list (struct btrfs_fs_info *fs_info) +{ +struct rb_node *node_bg, *node_chunk; + + node_bg =rb_first_cached (&fs_info->block_group_cache_tree); + node_chunk =rb_first_cached (&fs_info->mapping_tree); + +while (node_bg && node_chunk) { +struct btrfs_block_group *bg; +struct btrfs_chunk_map *map; + + bg = rb_entry(node_bg, struct btrfs_block_group, cache_node); + map = rb_entry(node_chunk, struct btrfs_chunk_map, rb_node); + +ASSERT (bg->start == map->start); + +if (!(bg->flags &BTRFS_BLOCK_GROUP_REMAPPED )) + goto next; + +if (bg->identity_remap_count !=0 ) + goto next; + +if (map->num_stripes ==0 ) + goto next; + +spin_lock (&fs_info->unused_bgs_lock); + +if (list_empty (&bg->bg_list)) { +btrfs_get_block_group (bg); +list_add_tail (&bg->bg_list, + &fs_info->fully_remapped_bgs); + }else { +list_move_tail (&bg->bg_list, + &fs_info->fully_remapped_bgs); + } + +spin_unlock (&fs_info->unused_bgs_lock); + +/* +* Ideally we'd want to call btrfs_discard_queue_work() here, +* but it'd do nothing as the discard worker hasn't been +* started yet. +* +* The block group will get added to the discard list when +* btrfs_handle_fully_remapped_bgs() gets called, when we +* commit the first transaction. +*/ +if (btrfs_test_opt (fs_info,DISCARD_ASYNC )) { +spin_lock (&bg->lock); +set_bit (BLOCK_GROUP_FLAG_STRIPE_REMOVAL_PENDING , + &bg->runtime_flags); +spin_unlock (&bg->lock); + } + +next: + node_bg =rb_next (node_bg); + node_chunk =rb_next (node_chunk); + } + +ASSERT (!node_bg && !node_chunk); + +return 0 ; +}
fs/btrfs/block-group.h +2 -0
@@ -94,6 +94,7 @@ enum btrfs_block_group_flags {*/ BLOCK_GROUP_FLAG_NEW ,BLOCK_GROUP_FLAG_FULLY_REMAPPED , + BLOCK_GROUP_FLAG_STRIPE_REMOVAL_PENDING, }; enum btrfs_caching_type {
@@ -416,5 +417,6 @@ int btrfs_use_block_group_size_class(struct btrfs_block_group *bg,bool btrfs_block_group_should_use_size_class (const struct btrfs_block_group *bg);void btrfs_mark_bg_fully_remapped (struct btrfs_block_group *bg,struct btrfs_trans_handle *trans); +int btrfs_populate_fully_remapped_bgs_list (struct btrfs_fs_info *fs_info);#endif /* BTRFS_BLOCK_GROUP_H */
fs/btrfs/disk-io.c +9 -0
@@ -3613,6 +3613,15 @@ int __cold open_ctree(struct super_block *sb, struct btrfs_fs_devices *fs_device goto fail_sysfs; } +if (btrfs_fs_incompat (fs_info,REMAP_TREE )) { + ret =btrfs_populate_fully_remapped_bgs_list (fs_info); +if (ret) { +btrfs_err (fs_info, +"failed to populate fully_remapped_bgs list: %d" , ret); + goto fail_sysfs; + } + } +btrfs_zoned_reserve_data_reloc_bg (fs_info);btrfs_free_zone_cache (fs_info);
fs/btrfs/free-space-cache.c +18 -0
@@ -3068,6 +3068,7 @@ bool btrfs_is_free_space_trimmed(struct btrfs_block_group *block_group)bool ret = true;if (block_group->flags &BTRFS_BLOCK_GROUP_REMAPPED && + !test_bit (BLOCK_GROUP_FLAG_STRIPE_REMOVAL_PENDING , &block_group->runtime_flags) && block_group->identity_remap_count ==0 ) {return true; }
@@ -3849,6 +3850,23 @@ void btrfs_trim_fully_remapped_block_group(struct btrfs_block_group *bg)const u64 max_discard_size =READ_ONCE (discard_ctl->max_discard_size);u64 end =btrfs_block_group_end (bg); +if (!test_bit (BLOCK_GROUP_FLAG_STRIPE_REMOVAL_PENDING , &bg->runtime_flags)) { + bg->discard_cursor = end; + +if (bg->used ==0 ) { +spin_lock (&fs_info->unused_bgs_lock); +if (!list_empty (&bg->bg_list)) { +list_del_init (&bg->bg_list); +btrfs_put_block_group (bg); + } +spin_unlock (&fs_info->unused_bgs_lock); + +btrfs_mark_bg_unused (bg); + } + +return ; + } + bytes = end - bg->discard_cursor; if (max_discard_size &&
fs/btrfs/relocation.c +4 -0
@@ -4785,6 +4785,10 @@ int btrfs_last_identity_remap_gone(struct btrfs_chunk_map *chunk_map,btrfs_remove_bg_from_sinfo (bg); +spin_lock (&bg->lock); +clear_bit (BLOCK_GROUP_FLAG_STRIPE_REMOVAL_PENDING , &bg->runtime_flags); +spin_unlock (&bg->lock); + ret =remove_chunk_stripes (trans, chunk_map, path);if (unlikely (ret)) {btrfs_abort_transaction (trans, ret);
Signature · 1 line
2.51.2
Quoted 35 lines · Mark Harmstone
This is version 8 of the patch series for the new logical remapping tree feature - see the previous cover letters for more information including the rationale:
* RFC: https://lore.kernel.org/all/[email protected]/
* Version 1: https://lore.kernel.org/all/[email protected]/
* Version 2: https://lore.kernel.org/all/[email protected]/
* Version 3: https://lore.kernel.org/all/[email protected]/
* Version 4: https://lore.kernel.org/all/[email protected]/
* Version 5: https://lore.kernel.org/all/[email protected]/
* Version 6: https://lore.kernel.org/all/[email protected]/
* Version 7: https://lore.kernel.org/all/[email protected]/Changes since version 7:
* renamed structbtrfs_remapto structbtrfs_remap_item
* renamedBTRFS_BLOCK_GROUP_FLAGS_REMAPtoBTRFS_BLOCK_GROUP_FLAGS_METADATA_REMAP
* added unlikelies
* renamed new commit_* fields in structbtrfs_block_groupto last_*, and added new patch renaming existingcommit_usedtolast_usedto match
* mergeddo_copy()intocopy_remapped_data()
* initialized on-stack structbtrfs_remap_items
* fixed comments
* added other minor changes as suggested by David SterbaMark Harmstone (17):
btrfs: add definitions and constants for remap-tree
btrfs: addMETADATA_REMAPchunk type
btrfs: allow remapped chunks to have zero stripes
btrfs: remove remapped block groups from the free-space tree
btrfs: don't add metadata items for the remap tree to the extent tree
btrfs: rename structbtrfs_block_groupfieldcommit_usedtolast_used
btrfs: add extended version of structblock_group_item
btrfs: allow mounting filesystems with remap-tree incompat flag
btrfs: redirect I/O for remapped block groups
btrfs: handle deletions from remapped block group
Quoted 8 lines
btrfs: handle setting up relocation of block group with remap-tree
btrfs: move existing remaps before relocating block group
btrfs: replace identity remaps with actual remaps when doing
relocations
btrfs: adddo_remapparam tobtrfs_discard_extent()
btrfs: allow balancing remap tree
btrfs: handle discarding fully-remapped block groups
btrfs: populatefully_remapped_bgs_liston mount
Patches have been added to for-next. There were many coding style issues which I've tried to fix. As this is a lot of new code it'll get updated anyway, I realized that for this kind of initial batch the coding style is quite important as we'd have to stick with until some random change touches it. Please have a look for the differences. Thanks.
Quoted 46 lines · David Sterba
Quoted 35 lines · Mark Harmstone
This is version 8 of the patch series for the new logical remapping tree feature - see the previous cover letters for more information including the rationale:
* RFC: https://lore.kernel.org/all/[email protected]/
* Version 1: https://lore.kernel.org/all/[email protected]/
* Version 2: https://lore.kernel.org/all/[email protected]/
* Version 3: https://lore.kernel.org/all/[email protected]/
* Version 4: https://lore.kernel.org/all/[email protected]/
* Version 5: https://lore.kernel.org/all/[email protected]/
* Version 6: https://lore.kernel.org/all/[email protected]/
* Version 7: https://lore.kernel.org/all/[email protected]/Changes since version 7:
* renamed structbtrfs_remapto structbtrfs_remap_item
* renamedBTRFS_BLOCK_GROUP_FLAGS_REMAPtoBTRFS_BLOCK_GROUP_FLAGS_METADATA_REMAP
* added unlikelies
* renamed new commit_* fields in structbtrfs_block_groupto last_*, and added new patch renaming existingcommit_usedtolast_usedto match
* mergeddo_copy()intocopy_remapped_data()
* initialized on-stack structbtrfs_remap_items
* fixed comments
* added other minor changes as suggested by David SterbaMark Harmstone (17):
btrfs: add definitions and constants for remap-tree
btrfs: addMETADATA_REMAPchunk type
btrfs: allow remapped chunks to have zero stripes
btrfs: remove remapped block groups from the free-space tree
btrfs: don't add metadata items for the remap tree to the extent tree
btrfs: rename structbtrfs_block_groupfieldcommit_usedtolast_used
btrfs: add extended version of structblock_group_item
btrfs: allow mounting filesystems with remap-tree incompat flag
btrfs: redirect I/O for remapped block groups
btrfs: handle deletions from remapped block groupQuoted 8 lines
btrfs: handle setting up relocation of block group with remap-tree
btrfs: move existing remaps before relocating block group
btrfs: replace identity remaps with actual remaps when doing
relocations
btrfs: adddo_remapparam tobtrfs_discard_extent()
btrfs: allow balancing remap tree
btrfs: handle discarding fully-remapped block groups
btrfs: populatefully_remapped_bgs_liston mount
Quoted 5 lines
Patches have been added to for-next. There were many coding style issues which I've tried to fix. As this is a lot of new code it'll get updated anyway, I realized that for this kind of initial batch the coding style is quite important as we'd have to stick with until some random change touches it. Please have a look for the differences. Thanks.
This is a huge amount of code and quite critical. Shouldn't we have test cases in fstests to exercise this feature? I didn't see any test cases submitted.
Quoted 1 line
Quoted 47 lines · Filipe Manana
Quoted 46 lines · David Sterba
Quoted 43 lines · Mark Harmstone
This is version 8 of the patch series for the new logical remapping tree feature - see the previous cover letters for more information including the rationale:
* RFC: https://lore.kernel.org/all/[email protected]/
* Version 1: https://lore.kernel.org/all/[email protected]/
* Version 2: https://lore.kernel.org/all/[email protected]/
* Version 3: https://lore.kernel.org/all/[email protected]/
* Version 4: https://lore.kernel.org/all/[email protected]/
* Version 5: https://lore.kernel.org/all/[email protected]/
* Version 6: https://lore.kernel.org/all/[email protected]/
* Version 7: https://lore.kernel.org/all/[email protected]/Changes since version 7:
* renamed structbtrfs_remapto structbtrfs_remap_item
* renamedBTRFS_BLOCK_GROUP_FLAGS_REMAPtoBTRFS_BLOCK_GROUP_FLAGS_METADATA_REMAP
* added unlikelies
* renamed new commit_* fields in structbtrfs_block_groupto last_*, and added new patch renaming existingcommit_usedtolast_usedto match
* mergeddo_copy()intocopy_remapped_data()
* initialized on-stack structbtrfs_remap_items
* fixed comments
* added other minor changes as suggested by David SterbaMark Harmstone (17):
btrfs: add definitions and constants for remap-tree
btrfs: addMETADATA_REMAPchunk type
btrfs: allow remapped chunks to have zero stripes
btrfs: remove remapped block groups from the free-space tree
btrfs: don't add metadata items for the remap tree to the extent tree
btrfs: rename structbtrfs_block_groupfieldcommit_usedtolast_used
btrfs: add extended version of structblock_group_item
btrfs: allow mounting filesystems with remap-tree incompat flag
btrfs: redirect I/O for remapped block groups
btrfs: handle deletions from remapped block group
btrfs: handle setting up relocation of block group with remap-tree
btrfs: move existing remaps before relocating block group
btrfs: replace identity remaps with actual remaps when doing
relocations
btrfs: adddo_remapparam tobtrfs_discard_extent()
btrfs: allow balancing remap tree
btrfs: handle discarding fully-remapped block groups
btrfs: populatefully_remapped_bgs_liston mount
Quoted 5 lines
Quoted 5 lines
Patches have been added to for-next. There were many coding style issues which I've tried to fix. As this is a lot of new code it'll get updated anyway, I realized that for this kind of initial batch the coding style is quite important as we'd have to stick with until some random change touches it. Please have a look for the differences. Thanks.
Thanks David.
Quoted 3 lines
This is a huge amount of code and quite critical. Shouldn't we have test cases in fstests to exercise this feature? I didn't see any test cases submitted.
It is, but it's a no-op if the incompat flag set isn't set.
There will be fstests for this before I propose taking it out of EXPERIMENTAL.
Re: [PATCH v8 15/17] btrfs: allow balancing remap tree
In reply to Mark HarmstoneQuoted 10 lines · Mark Harmstone
Balancing the
METADATA_REMAPchunk, i.e. the chunk in which the remap tree lives, is a special case.We can't use the remap tree itself for this, as then we'd have no way to boostrap it on mount. And we can't use the pre-remap tree code for this as it relies on walking the extent tree, and we're not creating backrefs for
METADATA_REMAPchunks.So instead, if a balance would relocate any
METADATA_REMAPblock groups, mark those block groups as readonly and COW every leaf of the remap tree.
Hi everyone,
I ran my AI patch review prompts on linux-next, and this one was flagged. As
we add more btrfs specifics we'll probably find some other fun, but this one seems right to me:
Quoted 4 lines
fs/btrfs/volumes.c +0
-0
[ ... ]
Quoted 8 lines
fs/btrfs/volumes.c +8
-0+static int balance_remap_chunks (struct btrfs_fs_info *fs_info,struct btrfs_path *path, +struct list_head *chunks) +{ +struct remap_chunk_info *rci, *tmp; +struct btrfs_trans_handle *trans; +int ret; + +list_for_each_entry_safe (rci, tmp, chunks, list) {
fs/btrfs/volumes.c static bool should_balance_chunk(struct extent_buffer *leaf, struct btrfs_chunk · line 4148
Quoted 10 lines
fs/btrfs/volumes.c +10
-0+ rci->bg =btrfs_lookup_block_group (fs_info, rci->offset); +if (!rci->bg) { +list_del (&rci->list); +kfree (rci); +continue ; + } + + ret =btrfs_inc_block_group_ro (rci->bg, false); +if (ret) + goto end;
If btrfs_inc_block_group_ro() fails here, the remaining entries in the list still have rci->bg set to NULL (from the initialization in __btrfs_balance()). The goto jumps to the cleanup loop below.
Quoted 3 lines
fs/btrfs/volumes.c +3
-0+ + rci->made_ro = true; + }
[ ... ]
fs/btrfs/volumes.c static bool should_balance_chunk(struct extent_buffer *leaf, struct btrfs_chunk · line 4178
Quoted 7 lines
fs/btrfs/volumes.c +7
-0+end: +while (!list_empty (chunks)) { +bool is_unused; + + rci = list_first_entry(chunks, struct remap_chunk_info, list); + +spin_lock (&rci->bg->lock);
^^^^^^^^
Can this dereference a NULL pointer? When the loop above exits early via goto end, the remaining rci entries in the list have rci->bg == NULL. The cleanup loop then calls spin_lock(&rci->bg->lock) on these entries, which would crash the kernel.
Quoted 17 lines
fs/btrfs/volumes.c +17
-0+ is_unused = !btrfs_is_block_group_used (rci->bg); +spin_unlock (&rci->bg->lock); + +if (is_unused) +btrfs_mark_bg_unused (rci->bg); + +if (rci->made_ro) +btrfs_dec_block_group_ro (rci->bg); + +btrfs_put_block_group (rci->bg); + +list_del (&rci->list); +kfree (rci); + } + +return ret; +}
[ ... ]
fs/btrfs/volumes.c static int __btrfs_balance(struct btrfs_fs_info *fs_info) · line 4356
Quoted 2 lines
fs/btrfs/volumes.c +2
-0+ rci->offset = found_key.offset; + rci->bg =NULL ;
^^^^
This is where rci->bg is initialized to NULL. If balance_remap_chunks() fails partway through processing the list, these unprocessed entries still have bg == NULL when the cleanup loop runs.
[ ... ]
Re: [PATCH v8 12/17] btrfs: move existing remaps before relocating block group
In reply to Mark HarmstoneQuoted 9 lines · Mark Harmstone
If when relocating a block group we find that
remap_bytes> 0 in its block group item, that means that it has been the destination block group for another that has been remapped.We need to seach the remap tree for any remap backrefs within this range, and move the data to a third block group. This is because otherwise
btrfs_translate_remap()could end up following an unbounded chain of remaps, which would only get worse over time.
Quoted 4 lines
We only relocate one block group at a time, so
remap_byteswill only ever go down while we are doing this. Once we're finished we set the REMAPPED flag on the block group, which will permanently prevent any other data from being moved to within it.
Hi everyone,
One more AI review from linux-next, this looks valid to me:
Quoted 3 lines
fs/btrfs/relocation.c +0
-0
[ ... ]
Quoted 5 lines
fs/btrfs/relocation.c +5
-0+static int move_existing_remap (struct btrfs_fs_info *fs_info, +struct btrfs_path *path, +struct btrfs_block_group *bg,u64 new_addr, +u64 length,u64 old_addr) +{
[ ... ]
fs/btrfs/relocation.c static void adjust_block_group_remap_bytes(struct btrfs_trans_handle *trans, · line 4338
Quoted 12 lines
fs/btrfs/relocation.c +12
-0+ dest_bg =btrfs_lookup_block_group (fs_info, dest_addr); + +adjust_block_group_remap_bytes (trans, dest_bg, dest_length); + +mutex_lock (&dest_bg->free_space_lock); + bg_needs_free_space =test_bit (BLOCK_GROUP_FLAG_NEEDS_FREE_SPACE , + &dest_bg->runtime_flags); +mutex_unlock (&dest_bg->free_space_lock); +btrfs_put_block_group (dest_bg); + +if (bg_needs_free_space) { + ret =btrfs_add_block_group_free_space (trans, dest_bg);
^^^^^^^
Is there a use-after-free here? btrfs_put_block_group() is called above, which can free the block group if the refcount drops to zero. Then btrfs_add_block_group_free_space() dereferences dest_bg by calling mutex_lock(&block_group->free_space_lock).
Perhaps the btrfs_put_block_group() call should be moved after the conditional block that uses dest_bg?
Quoted 3 lines
fs/btrfs/relocation.c +3
-0+if (unlikely (ret)) + goto end; + }
[ ... ]
Quoted 24 lines
fs/btrfs/relocation.c +24
-0+static int move_existing_remaps (struct btrfs_fs_info *fs_info, +struct btrfs_block_group *bg, +struct btrfs_path *path) +{ +int ret; +struct btrfs_key key; +struct extent_buffer *leaf; +struct btrfs_remap_item *remap; +u64 old_addr; + +/* Look for backrefs in remap tree. */ +while (bg->remap_bytes >0 ) { + key.objectid = bg->start; + key.type =BTRFS_REMAP_BACKREF_KEY ; + key.offset =0 ; + + ret =btrfs_search_slot (NULL , fs_info->remap_root, &key, path,0 ,0 ); +if (ret <0 ) +return ret; + + leaf = path->nodes[0 ]; + +if (path->slots[0 ] >=btrfs_header_nritems (leaf)) { + ret =btrfs_next_leaf (fs_info->remap_root, path);
[ ... ]
fs/btrfs/relocation.c static void adjust_block_group_remap_bytes(struct btrfs_trans_handle *trans, · line 4418
Quoted 7 lines
fs/btrfs/relocation.c +7
-0+ leaf = path->nodes[0 ]; + } + +btrfs_item_key_to_cpu (leaf, &key, path->slots[0 ]); + +if (key.type !=BTRFS_REMAP_BACKREF_KEY ) { + path->slots[0 ]++;
When the slot is incremented here because the key type does not match, should the key be re-read from the new slot? The code below reads the remap item from path->slots[0] (the incremented slot), but then passes key.objectid and key.offset (from the old slot) to move_existing_remap().
If the key at the new slot has different objectid/offset values than the previous slot, this would pass mismatched address and length values to move_existing_remap().
Quoted 3 lines
fs/btrfs/relocation.c +3
-0+ +if (path->slots[0 ] >=btrfs_header_nritems (leaf)) { + ret =btrfs_next_leaf (fs_info->remap_root, path);
[ ... ]
Quoted 11 lines
fs/btrfs/relocation.c +11
-0+ leaf = path->nodes[0 ]; + } + } + + remap = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_remap_item); + old_addr =btrfs_remap_address (leaf, remap); + +btrfs_release_path (path); + + ret =move_existing_remap (fs_info, path, bg, key.objectid, + key.offset, old_addr);
[ ... ]
Re: [PATCH v8 10/17] btrfs: handle deletions from remapped block group
In reply to Mark HarmstoneQuoted 19 lines · Mark Harmstone
Handle the case where we free an extent from a block group that has the REMAPPED flag set. Because the remap tree is orthogonal to the extent tree, for data this may be within any number of identity remaps or actual remaps. If we're freeing a metadata node, this will be wholly inside one or the other.
btrfs_remove_extent_from_remap_tree()searches the remap tree for the remaps that cover the range in question, then callsremove_range_from_remap_tree()for each one, to punch a hole in the remap and adjust the free-space tree.For an identity remap,
remove_range_from_remap_tree()will adjust the block group'sidentity_remap_countif this changes. If it reaches zero we mark the block group as fully remapped.For an identity remap,
remove_range_from_remap_tree()will adjust the block group'sidentity_remap_countif this changes. If it reaches zero we mark the block group as fully remapped.
Quoted 5 lines
Fully remapped block groups have their chunk stripes removed and their device extents freed, which makes the disk space available again to the chunk allocator. This happens asynchronously: in the cleaner thread for sync discard and nodiscard, and (in a later patch) in the discard worker for async discard.
Hi everyone,
Another AI review from my recent linux-next run:
Quoted 3 lines
fs/btrfs/relocation.c +0
-0
[ ... ]
Quoted 22 lines
fs/btrfs/relocation.c +22
-0+int btrfs_last_identity_remap_gone (struct btrfs_chunk_map *chunk_map, +struct btrfs_block_group *bg) +{ +struct btrfs_fs_info *fs_info = bg->fs_info; +struct btrfs_trans_handle *trans; +int ret; +unsigned int num_items; +BTRFS_PATH_AUTO_FREE (path); + + path =btrfs_alloc_path (); +if (!path) +return -ENOMEM ; + +/* +* One item for each entry we're removing in the dev extents tree, and +* another for each device. DUP chunks are all on one device, +* everything else has one device per stripe. +*/ +if (bg->flags &BTRFS_BLOCK_GROUP_DUP ) + num_items = chunk_map->num_stripes +1 ; +else + num_items =2 * chunk_map->num_stripes;
fs/btrfs/relocation.c static const char *stage_to_string(enum reloc_stage stage) · line 3971
Quoted 10 lines
fs/btrfs/relocation.c +10
-0+ + trans =btrfs_start_transaction_fallback_global_rsv (fs_info->tree_root, num_items); +if (IS_ERR (trans)) +return PTR_ERR (trans); + + ret =btrfs_remove_dev_extents (trans, chunk_map); +if (unlikely (ret)) { +btrfs_abort_transaction (trans, ret); +return ret; + }
Can this leak the transaction handle? btrfs_abort_transaction() marks the transaction as aborted but does not end it. __btrfs_abort_transaction() has a comment explaining cleanup happens in btrfs_end_transaction().
fs/btrfs/relocation.c static const char *stage_to_string(enum reloc_stage stage) · line 3981
Quoted 9 lines
fs/btrfs/relocation.c +9
-0+ +mutex_lock (&trans->fs_info->chunk_mutex); +for (unsigned int i =0 ; i < chunk_map->num_stripes; i++) { + ret =btrfs_update_device (trans, chunk_map->stripes[i].dev); +if (unlikely (ret)) { +mutex_unlock (&trans->fs_info->chunk_mutex); +btrfs_abort_transaction (trans, ret); +return ret; + }
Same question here about leaking the transaction handle.
Quoted 4 lines
fs/btrfs/relocation.c +4
-0+ } +mutex_unlock (&trans->fs_info->chunk_mutex); + +write_lock (&trans->fs_info->mapping_tree_lock);
fs/btrfs/relocation.c static const char *stage_to_string(enum reloc_stage stage) · line 3996
Quoted 10 lines
fs/btrfs/relocation.c +10
-0+btrfs_chunk_map_device_clear_bits (chunk_map,CHUNK_ALLOCATED ); +write_unlock (&trans->fs_info->mapping_tree_lock); + +btrfs_remove_bg_from_sinfo (bg); + + ret =remove_chunk_stripes (trans, chunk_map, path); +if (unlikely (ret)) { +btrfs_abort_transaction (trans, ret); +return ret; + }
And here as well - does this return path need btrfs_end_transaction()?
Quoted 7 lines
fs/btrfs/relocation.c +7
-0+ + ret =btrfs_commit_transaction (trans); +if (ret) +return ret; + +return 0 ; +}
[ ... ]
Quoted 5 lines
fs/btrfs/relocation.c +5
-0+static int remove_range_from_remap_tree (struct btrfs_trans_handle *trans, +struct btrfs_path *path, +struct btrfs_block_group *bg, +u64 hole_start,u64 hole_length) +{
[ ... ]
fs/btrfs/relocation.c u64 btrfs_get_reloc_bg_bytenr(const struct btrfs_fs_info *fs_info) · line 4826
Quoted 6 lines
fs/btrfs/relocation.c +6
-0+if (!is_identity_remap) { +struct btrfs_block_group *dest_bg; + + dest_bg =btrfs_lookup_block_group (fs_info, new_addr); +adjust_block_group_remap_bytes (trans, dest_bg, -overlap_length); +btrfs_put_block_group (dest_bg);
btrfs_lookup_block_group() can return NULL if no block group is found for the address. Every other caller checks for NULL before using the result. Is a NULL check needed here before passing dest_bg to adjust_block_group_remap_bytes()? That function does spin_lock(&bg->lock) which would crash on a NULL pointer.
fs/btrfs/relocation.c u64 btrfs_get_reloc_bg_bytenr(const struct btrfs_fs_info *fs_info) · line 4833
Quoted 6 lines
fs/btrfs/relocation.c +6
-0+ ret =btrfs_add_to_free_space_tree (trans, + hole_start - remap_start + new_addr, + overlap_length); +if (ret) +return ret; + }
[ ... ]
This is not a bug, but the commit message appears to have a duplicated paragraph about identity remaps:
For an identity remap, remove_range_from_remap_tree() will adjust the block group's identity_remap_count if this changes. If it reaches zero we mark the block group as fully remapped.
For an identity remap, remove_range_from_remap_tree() will adjust the block group's identity_remap_count if this changes. If it reaches zero we mark the block group as fully remapped.
Re: [PATCH v8 09/17] btrfs: redirect I/O for remapped block groups
In reply to Mark HarmstoneQuoted 12 lines · Mark Harmstone
Change
btrfs_map_block()so that if the block group has the REMAPPED flag set, we callbtrfs_translate_remap()to obtain a new address.
btrfs_translate_remap()searches the remap tree for a range corresponding to the logical address passed tobtrfs_map_block(). If it is within an identity remap, this part of the block group hasn't yet been relocated, and so we use the existing address.If it is within an actual remap, we subtract the start of the remap range and add the address of its destination, contained in the item's payload.
Hi everyone,
Another AI review from linux-next runs:
fs/btrfs/volumes.c struct btrfs_io_geometry io_geom = { 0 }; · line 6594
Quoted 13 lines
fs/btrfs/volumes.c +6
-0@@ -6586,6 +6586,24 @@ struct btrfs_io_geometry io_geom = { 0 };if (IS_ERR (map))return PTR_ERR (map); +if (map->type &BTRFS_BLOCK_GROUP_REMAPPED ) { +u64 new_logical = logical; + + ret =btrfs_translate_remap (fs_info, &new_logical, length); +if (ret) +return ret;
^^^^^^^^^^
When btrfs_translate_remap() returns an error, does this leak the chunk map? The map was obtained via btrfs_get_chunk_map() which increments a refcount, and btrfs_get_chunk_map() documents that callers are responsible for dropping the reference. The other error paths in this function use goto out, which calls btrfs_free_chunk_map(map).
Quoted 11 lines
fs/btrfs/volumes.c +11
-0+ +if (new_logical != logical) { +btrfs_free_chunk_map (map); + + map =btrfs_get_chunk_map (fs_info, new_logical, *length); +if (IS_ERR (map)) +return PTR_ERR (map); + + logical = new_logical; + } + }
[ ... ]
Re: [PATCH v8 09/17] btrfs: redirect I/O for remapped block groups
Claude's right, and I made this mistake because I was copying the "return -EINVAL" a few lines lower down, which also leaks. I'll patch them both.
Quoted 52 lines · Chris Mason
Quoted 12 lines · Mark Harmstone
Change
btrfs_map_block()so that if the block group has the REMAPPED flag set, we callbtrfs_translate_remap()to obtain a new address.
btrfs_translate_remap()searches the remap tree for a range corresponding to the logical address passed tobtrfs_map_block(). If it is within an identity remap, this part of the block group hasn't yet been relocated, and so we use the existing address.If it is within an actual remap, we subtract the start of the remap range and add the address of its destination, contained in the item's payload.
Hi everyone,
Another AI review from linux-next runs:
fs/btrfs/volumes.c struct btrfs_io_geometry io_geom = { 0 }; · line 6594
Quoted 13 lines
fs/btrfs/volumes.c +6
-0@@ -6586,6 +6586,24 @@ struct btrfs_io_geometry io_geom = { 0 };if (IS_ERR (map))return PTR_ERR (map); +if (map->type &BTRFS_BLOCK_GROUP_REMAPPED ) { +u64 new_logical = logical; + + ret =btrfs_translate_remap (fs_info, &new_logical, length); +if (ret) +return ret;^^^^^^^^^^When
btrfs_translate_remap()returns an error, does this leak the chunk map? The map was obtained viabtrfs_get_chunk_map()which increments a refcount, andbtrfs_get_chunk_map()documents that callers are responsible for dropping the reference. The other error paths in this function use goto out, which callsbtrfs_free_chunk_map(map).Quoted 11 lines
fs/btrfs/volumes.c +11
-0+ +if (new_logical != logical) { +btrfs_free_chunk_map (map); + + map =btrfs_get_chunk_map (fs_info, new_logical, *length); +if (IS_ERR (map)) +return PTR_ERR (map); + + logical = new_logical; + } + }[ ... ]