* [PATCH v5 1/6] mm: split SWP_FILE into SWP_ACTIVATED and SWP_FS
2018-08-31 22:36 [PATCH v5 0/6] Btrfs: implement swap file support Omar Sandoval
@ 2018-08-31 22:36 ` Omar Sandoval
2018-08-31 22:36 ` [PATCH v5 2/6] mm: export add_swap_extent() Omar Sandoval
` (5 subsequent siblings)
6 siblings, 0 replies; 9+ messages in thread
From: Omar Sandoval @ 2018-08-31 22:36 UTC (permalink / raw)
To: linux-btrfs; +Cc: kernel-team
From: Omar Sandoval <osandov@fb.com>
The SWP_FILE flag serves two purposes: to make swap_{read,write}page()
go through the filesystem, and to make swapoff() call
->swap_deactivate(). For Btrfs, we want the latter but not the former,
so split this flag into two. This makes us always call
->swap_deactivate() if ->swap_activate() succeeded, not just if it
didn't add any swap extents itself.
This also resolves the issue of the very misleading name of SWP_FILE,
which is only used for swap files over NFS.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Omar Sandoval <osandov@fb.com>
---
include/linux/swap.h | 13 +++++++------
mm/page_io.c | 6 +++---
mm/swapfile.c | 13 ++++++++-----
3 files changed, 18 insertions(+), 14 deletions(-)
diff --git a/include/linux/swap.h b/include/linux/swap.h
index 8e2c11e692ba..0fda0aa743f0 100644
--- a/include/linux/swap.h
+++ b/include/linux/swap.h
@@ -167,13 +167,14 @@ enum {
SWP_SOLIDSTATE = (1 << 4), /* blkdev seeks are cheap */
SWP_CONTINUED = (1 << 5), /* swap_map has count continuation */
SWP_BLKDEV = (1 << 6), /* its a block device */
- SWP_FILE = (1 << 7), /* set after swap_activate success */
- SWP_AREA_DISCARD = (1 << 8), /* single-time swap area discards */
- SWP_PAGE_DISCARD = (1 << 9), /* freed swap page-cluster discards */
- SWP_STABLE_WRITES = (1 << 10), /* no overwrite PG_writeback pages */
- SWP_SYNCHRONOUS_IO = (1 << 11), /* synchronous IO is efficient */
+ SWP_ACTIVATED = (1 << 7), /* set after swap_activate success */
+ SWP_FS = (1 << 8), /* swap file goes through fs */
+ SWP_AREA_DISCARD = (1 << 9), /* single-time swap area discards */
+ SWP_PAGE_DISCARD = (1 << 10), /* freed swap page-cluster discards */
+ SWP_STABLE_WRITES = (1 << 11), /* no overwrite PG_writeback pages */
+ SWP_SYNCHRONOUS_IO = (1 << 12), /* synchronous IO is efficient */
/* add others here before... */
- SWP_SCANNING = (1 << 12), /* refcount in scan_swap_map */
+ SWP_SCANNING = (1 << 13), /* refcount in scan_swap_map */
};
#define SWAP_CLUSTER_MAX 32UL
diff --git a/mm/page_io.c b/mm/page_io.c
index aafd19ec1db4..e8653c368069 100644
--- a/mm/page_io.c
+++ b/mm/page_io.c
@@ -283,7 +283,7 @@ int __swap_writepage(struct page *page, struct writeback_control *wbc,
struct swap_info_struct *sis = page_swap_info(page);
VM_BUG_ON_PAGE(!PageSwapCache(page), page);
- if (sis->flags & SWP_FILE) {
+ if (sis->flags & SWP_FS) {
struct kiocb kiocb;
struct file *swap_file = sis->swap_file;
struct address_space *mapping = swap_file->f_mapping;
@@ -365,7 +365,7 @@ int swap_readpage(struct page *page, bool synchronous)
goto out;
}
- if (sis->flags & SWP_FILE) {
+ if (sis->flags & SWP_FS) {
struct file *swap_file = sis->swap_file;
struct address_space *mapping = swap_file->f_mapping;
@@ -423,7 +423,7 @@ int swap_set_page_dirty(struct page *page)
{
struct swap_info_struct *sis = page_swap_info(page);
- if (sis->flags & SWP_FILE) {
+ if (sis->flags & SWP_FS) {
struct address_space *mapping = sis->swap_file->f_mapping;
VM_BUG_ON_PAGE(!PageSwapCache(page), page);
diff --git a/mm/swapfile.c b/mm/swapfile.c
index d954b71c4f9c..d3f95833d12e 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -989,7 +989,7 @@ int get_swap_pages(int n_goal, swp_entry_t swp_entries[], int entry_size)
goto nextsi;
}
if (size == SWAPFILE_CLUSTER) {
- if (!(si->flags & SWP_FILE))
+ if (!(si->flags & SWP_FS))
n_ret = swap_alloc_cluster(si, swp_entries);
} else
n_ret = scan_swap_map_slots(si, SWAP_HAS_CACHE,
@@ -2310,12 +2310,13 @@ static void destroy_swap_extents(struct swap_info_struct *sis)
kfree(se);
}
- if (sis->flags & SWP_FILE) {
+ if (sis->flags & SWP_ACTIVATED) {
struct file *swap_file = sis->swap_file;
struct address_space *mapping = swap_file->f_mapping;
- sis->flags &= ~SWP_FILE;
- mapping->a_ops->swap_deactivate(swap_file);
+ sis->flags &= ~SWP_ACTIVATED;
+ if (mapping->a_ops->swap_deactivate)
+ mapping->a_ops->swap_deactivate(swap_file);
}
}
@@ -2411,8 +2412,10 @@ static int setup_swap_extents(struct swap_info_struct *sis, sector_t *span)
if (mapping->a_ops->swap_activate) {
ret = mapping->a_ops->swap_activate(sis, swap_file, span);
+ if (ret >= 0)
+ sis->flags |= SWP_ACTIVATED;
if (!ret) {
- sis->flags |= SWP_FILE;
+ sis->flags |= SWP_FS;
ret = add_swap_extent(sis, 0, sis->max, 0);
*span = sis->pages;
}
--
2.18.0
^ permalink raw reply related [flat|nested] 9+ messages in thread* [PATCH v5 2/6] mm: export add_swap_extent()
2018-08-31 22:36 [PATCH v5 0/6] Btrfs: implement swap file support Omar Sandoval
2018-08-31 22:36 ` [PATCH v5 1/6] mm: split SWP_FILE into SWP_ACTIVATED and SWP_FS Omar Sandoval
@ 2018-08-31 22:36 ` Omar Sandoval
2018-08-31 22:36 ` [PATCH v5 3/6] vfs: update swap_{,de}activate documentation Omar Sandoval
` (4 subsequent siblings)
6 siblings, 0 replies; 9+ messages in thread
From: Omar Sandoval @ 2018-08-31 22:36 UTC (permalink / raw)
To: linux-btrfs; +Cc: kernel-team
From: Omar Sandoval <osandov@fb.com>
Btrfs will need this for swap file support.
Signed-off-by: Omar Sandoval <osandov@fb.com>
---
mm/swapfile.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/mm/swapfile.c b/mm/swapfile.c
index d3f95833d12e..51cb30de17bc 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -2365,6 +2365,7 @@ add_swap_extent(struct swap_info_struct *sis, unsigned long start_page,
list_add_tail(&new_se->list, &sis->first_swap_extent.list);
return 1;
}
+EXPORT_SYMBOL_GPL(add_swap_extent);
/*
* A `swap extent' is a simple thing which maps a contiguous range of pages
--
2.18.0
^ permalink raw reply related [flat|nested] 9+ messages in thread* [PATCH v5 3/6] vfs: update swap_{,de}activate documentation
2018-08-31 22:36 [PATCH v5 0/6] Btrfs: implement swap file support Omar Sandoval
2018-08-31 22:36 ` [PATCH v5 1/6] mm: split SWP_FILE into SWP_ACTIVATED and SWP_FS Omar Sandoval
2018-08-31 22:36 ` [PATCH v5 2/6] mm: export add_swap_extent() Omar Sandoval
@ 2018-08-31 22:36 ` Omar Sandoval
2018-08-31 22:36 ` [PATCH v5 4/6] Btrfs: prevent ioctls from interfering with a swap file Omar Sandoval
` (3 subsequent siblings)
6 siblings, 0 replies; 9+ messages in thread
From: Omar Sandoval @ 2018-08-31 22:36 UTC (permalink / raw)
To: linux-btrfs; +Cc: kernel-team
From: Omar Sandoval <osandov@fb.com>
The documentation for these functions is wrong in several ways:
- swap_activate() is called with the inode locked
- swap_activate() takes a swap_info_struct * and a sector_t *
- swap_activate() can also return a positive number of extents it added
itself
- swap_deactivate() does not return anything
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Omar Sandoval <osandov@fb.com>
---
Documentation/filesystems/Locking | 17 +++++++----------
Documentation/filesystems/vfs.txt | 12 ++++++++----
2 files changed, 15 insertions(+), 14 deletions(-)
diff --git a/Documentation/filesystems/Locking b/Documentation/filesystems/Locking
index efea228ccd8a..b970c8c2ee22 100644
--- a/Documentation/filesystems/Locking
+++ b/Documentation/filesystems/Locking
@@ -210,8 +210,9 @@ prototypes:
int (*launder_page)(struct page *);
int (*is_partially_uptodate)(struct page *, unsigned long, unsigned long);
int (*error_remove_page)(struct address_space *, struct page *);
- int (*swap_activate)(struct file *);
- int (*swap_deactivate)(struct file *);
+ int (*swap_activate)(struct swap_info_struct *, struct file *,
+ sector_t *);
+ void (*swap_deactivate)(struct file *);
locking rules:
All except set_page_dirty and freepage may block
@@ -235,8 +236,8 @@ putback_page: yes
launder_page: yes
is_partially_uptodate: yes
error_remove_page: yes
-swap_activate: no
-swap_deactivate: no
+swap_activate: yes
+swap_deactivate: no
->write_begin(), ->write_end() and ->readpage() may be called from
the request handler (/dev/loop).
@@ -333,14 +334,10 @@ cleaned, or an error value if not. Note that in order to prevent the page
getting mapped back in and redirtied, it needs to be kept locked
across the entire operation.
- ->swap_activate will be called with a non-zero argument on
-files backing (non block device backed) swapfiles. A return value
-of zero indicates success, in which case this file can be used for
-backing swapspace. The swapspace operations will be proxied to the
-address space operations.
+ ->swap_activate is called from sys_swapon() with the inode locked.
->swap_deactivate() will be called in the sys_swapoff()
-path after ->swap_activate() returned success.
+path after ->swap_activate() returned success. The inode is not locked.
----------------------- file_lock_operations ------------------------------
prototypes:
diff --git a/Documentation/filesystems/vfs.txt b/Documentation/filesystems/vfs.txt
index 4b2084d0f1fb..40d6d6d4b76b 100644
--- a/Documentation/filesystems/vfs.txt
+++ b/Documentation/filesystems/vfs.txt
@@ -652,8 +652,9 @@ struct address_space_operations {
unsigned long);
void (*is_dirty_writeback) (struct page *, bool *, bool *);
int (*error_remove_page) (struct mapping *mapping, struct page *page);
- int (*swap_activate)(struct file *);
- int (*swap_deactivate)(struct file *);
+ int (*swap_activate)(struct swap_info_struct *, struct file *,
+ sector_t *);
+ void (*swap_deactivate)(struct file *);
};
writepage: called by the VM to write a dirty page to backing store.
@@ -830,8 +831,11 @@ struct address_space_operations {
swap_activate: Called when swapon is used on a file to allocate
space if necessary and pin the block lookup information in
- memory. A return value of zero indicates success,
- in which case this file can be used to back swapspace.
+ memory. If this returns zero, the swap system will call the address
+ space operations ->readpage() and ->direct_IO(). Alternatively, this
+ may call add_swap_extent() and return the number of extents added, in
+ which case the swap system will use the provided blocks directly
+ instead of going through the filesystem.
swap_deactivate: Called during swapoff on files where swap_activate
was successful.
--
2.18.0
^ permalink raw reply related [flat|nested] 9+ messages in thread* [PATCH v5 4/6] Btrfs: prevent ioctls from interfering with a swap file
2018-08-31 22:36 [PATCH v5 0/6] Btrfs: implement swap file support Omar Sandoval
` (2 preceding siblings ...)
2018-08-31 22:36 ` [PATCH v5 3/6] vfs: update swap_{,de}activate documentation Omar Sandoval
@ 2018-08-31 22:36 ` Omar Sandoval
2018-08-31 22:36 ` [PATCH v5 5/6] Btrfs: rename get_chunk_map() and make it non-static Omar Sandoval
` (2 subsequent siblings)
6 siblings, 0 replies; 9+ messages in thread
From: Omar Sandoval @ 2018-08-31 22:36 UTC (permalink / raw)
To: linux-btrfs; +Cc: kernel-team
From: Omar Sandoval <osandov@fb.com>
When a swap file is active, we must make sure that the extents of the
file are not moved and that they don't become shared. That means that
the following are not safe:
- chattr +c (enable compression)
- reflink
- dedupe
- snapshot
- defrag
- balance
- device remove/replace/resize
Don't allow those to happen on an active swap file. Balance and device
remove/replace/resize in particular are disallowed entirely; in the
future, we can relax this so that relocation skips/errors out only on
chunks containing an active swap file.
Note that we don't have to worry about chattr -C (disable nocow), which
we ignore for non-empty files, because an active swapfile must be
non-empty and can't be truncated. We also don't have to worry about
autodefrag because it's only done on COW files. Truncate and fallocate
are already taken care of by the generic code. Device add doesn't do
relocation so it's not an issue, either.
Signed-off-by: Omar Sandoval <osandov@fb.com>
---
fs/btrfs/ctree.h | 6 ++++++
fs/btrfs/disk-io.c | 3 +++
fs/btrfs/ioctl.c | 51 ++++++++++++++++++++++++++++++++++++++++++----
fs/btrfs/volumes.c | 6 ++++++
4 files changed, 62 insertions(+), 4 deletions(-)
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 53af9f5253f4..1c767a6394ae 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -1121,6 +1121,9 @@ struct btrfs_fs_info {
u32 sectorsize;
u32 stripesize;
+ /* Number of active swapfiles */
+ atomic_t nr_swapfiles;
+
#ifdef CONFIG_BTRFS_FS_REF_VERIFY
spinlock_t ref_verify_lock;
struct rb_root block_tree;
@@ -1285,6 +1288,9 @@ struct btrfs_root {
spinlock_t qgroup_meta_rsv_lock;
u64 qgroup_meta_rsv_pertrans;
u64 qgroup_meta_rsv_prealloc;
+
+ /* Number of active swapfiles */
+ atomic_t nr_swapfiles;
};
struct btrfs_file_private {
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 5124c15705ce..50ee5cd3efae 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -1187,6 +1187,7 @@ static void __setup_root(struct btrfs_root *root, struct btrfs_fs_info *fs_info,
atomic_set(&root->log_batch, 0);
refcount_set(&root->refs, 1);
atomic_set(&root->will_be_snapshotted, 0);
+ atomic_set(&root->nr_swapfiles, 0);
root->log_transid = 0;
root->log_transid_committed = -1;
root->last_log_commit = 0;
@@ -2781,6 +2782,8 @@ int open_ctree(struct super_block *sb,
fs_info->sectorsize = 4096;
fs_info->stripesize = 4096;
+ atomic_set(&fs_info->nr_swapfiles, 0);
+
ret = btrfs_alloc_stripe_hash_table(fs_info);
if (ret) {
err = ret;
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 63600dc2ac4c..cc230dcd32a4 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -290,6 +290,11 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg)
} else if (fsflags & FS_COMPR_FL) {
const char *comp;
+ if (IS_SWAPFILE(inode)) {
+ ret = -ETXTBSY;
+ goto out_unlock;
+ }
+
binode->flags |= BTRFS_INODE_COMPRESS;
binode->flags &= ~BTRFS_INODE_NOCOMPRESS;
@@ -751,6 +756,12 @@ static int create_snapshot(struct btrfs_root *root, struct inode *dir,
if (!test_bit(BTRFS_ROOT_REF_COWS, &root->state))
return -EINVAL;
+ if (atomic_read(&root->nr_swapfiles)) {
+ btrfs_info(fs_info,
+ "cannot snapshot subvolume with active swapfile");
+ return -ETXTBSY;
+ }
+
pending_snapshot = kzalloc(sizeof(*pending_snapshot), GFP_KERNEL);
if (!pending_snapshot)
return -ENOMEM;
@@ -1487,9 +1498,13 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
}
inode_lock(inode);
- if (do_compress)
- BTRFS_I(inode)->defrag_compress = compress_type;
- ret = cluster_pages_for_defrag(inode, pages, i, cluster);
+ if (IS_SWAPFILE(inode)) {
+ ret = -ETXTBSY;
+ } else {
+ if (do_compress)
+ BTRFS_I(inode)->defrag_compress = compress_type;
+ ret = cluster_pages_for_defrag(inode, pages, i, cluster);
+ }
if (ret < 0) {
inode_unlock(inode);
goto out_ra;
@@ -1585,6 +1600,12 @@ static noinline int btrfs_ioctl_resize(struct file *file,
return BTRFS_ERROR_DEV_EXCL_RUN_IN_PROGRESS;
}
+ if (atomic_read(&fs_info->nr_swapfiles)) {
+ btrfs_info(fs_info, "cannot resize with active swapfile");
+ ret = -ETXTBSY;
+ goto out;
+ }
+
vol_args = memdup_user(arg, sizeof(*vol_args));
if (IS_ERR(vol_args)) {
ret = PTR_ERR(vol_args);
@@ -3538,6 +3559,11 @@ static int btrfs_extent_same(struct inode *src, u64 loff, u64 olen,
goto out_unlock;
}
+ if (IS_SWAPFILE(src) || IS_SWAPFILE(dst)) {
+ ret = -ETXTBSY;
+ goto out_unlock;
+ }
+
tail_len = olen % BTRFS_MAX_DEDUPE_LEN;
chunk_count = div_u64(olen, BTRFS_MAX_DEDUPE_LEN);
if (chunk_count == 0)
@@ -4234,6 +4260,11 @@ static noinline int btrfs_clone_files(struct file *file, struct file *file_src,
goto out_unlock;
}
+ if (IS_SWAPFILE(src) || IS_SWAPFILE(inode)) {
+ ret = -ETXTBSY;
+ goto out_unlock;
+ }
+
/* determine range to clone */
ret = -EINVAL;
if (off + len > src->i_size || off + len < off)
@@ -4712,7 +4743,13 @@ static long btrfs_ioctl_dev_replace(struct btrfs_fs_info *fs_info,
if (test_and_set_bit(BTRFS_FS_EXCL_OP, &fs_info->flags)) {
ret = BTRFS_ERROR_DEV_EXCL_RUN_IN_PROGRESS;
} else {
- ret = btrfs_dev_replace_by_ioctl(fs_info, p);
+ if (atomic_read(&fs_info->nr_swapfiles)) {
+ btrfs_info(fs_info,
+ "cannot replace device with active swapfile");
+ ret = -ETXTBSY;
+ } else {
+ ret = btrfs_dev_replace_by_ioctl(fs_info, p);
+ }
clear_bit(BTRFS_FS_EXCL_OP, &fs_info->flags);
}
break;
@@ -4972,6 +5009,12 @@ static long btrfs_ioctl_balance(struct file *file, void __user *arg)
locked:
BUG_ON(!test_bit(BTRFS_FS_EXCL_OP, &fs_info->flags));
+ if (atomic_read(&fs_info->nr_swapfiles)) {
+ btrfs_info(fs_info, "cannot balance with active swapfile");
+ ret = -ETXTBSY;
+ goto out_unlock;
+ }
+
if (arg) {
bargs = memdup_user(arg, sizeof(*bargs));
if (IS_ERR(bargs)) {
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index da86706123ff..c03ef5322689 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -1863,6 +1863,12 @@ int btrfs_rm_device(struct btrfs_fs_info *fs_info, const char *device_path,
u64 num_devices;
int ret = 0;
+ if (atomic_read(&fs_info->nr_swapfiles)) {
+ btrfs_info(fs_info,
+ "cannot remove device with active swapfile");
+ return -ETXTBSY;
+ }
+
mutex_lock(&uuid_mutex);
num_devices = fs_devices->num_devices;
--
2.18.0
^ permalink raw reply related [flat|nested] 9+ messages in thread* [PATCH v5 5/6] Btrfs: rename get_chunk_map() and make it non-static
2018-08-31 22:36 [PATCH v5 0/6] Btrfs: implement swap file support Omar Sandoval
` (3 preceding siblings ...)
2018-08-31 22:36 ` [PATCH v5 4/6] Btrfs: prevent ioctls from interfering with a swap file Omar Sandoval
@ 2018-08-31 22:36 ` Omar Sandoval
2018-08-31 22:36 ` [PATCH v5 6/6] Btrfs: support swap files Omar Sandoval
2018-09-06 11:59 ` [PATCH v5 0/6] Btrfs: implement swap file support David Sterba
6 siblings, 0 replies; 9+ messages in thread
From: Omar Sandoval @ 2018-08-31 22:36 UTC (permalink / raw)
To: linux-btrfs; +Cc: kernel-team
From: Omar Sandoval <osandov@fb.com>
The Btrfs swap code is going to need it, so give it a btrfs_ prefix and
make it non-static.
Signed-off-by: Omar Sandoval <osandov@fb.com>
---
fs/btrfs/volumes.c | 22 +++++++++++-----------
fs/btrfs/volumes.h | 9 +++++++++
2 files changed, 20 insertions(+), 11 deletions(-)
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index c03ef5322689..0aa8aff6774b 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -2712,8 +2712,8 @@ static int btrfs_del_sys_chunk(struct btrfs_fs_info *fs_info, u64 chunk_offset)
return ret;
}
-static struct extent_map *get_chunk_map(struct btrfs_fs_info *fs_info,
- u64 logical, u64 length)
+struct extent_map *btrfs_get_chunk_map(struct btrfs_fs_info *fs_info,
+ u64 logical, u64 length)
{
struct extent_map_tree *em_tree;
struct extent_map *em;
@@ -2750,7 +2750,7 @@ int btrfs_remove_chunk(struct btrfs_trans_handle *trans, u64 chunk_offset)
int i, ret = 0;
struct btrfs_fs_devices *fs_devices = fs_info->fs_devices;
- em = get_chunk_map(fs_info, chunk_offset, 1);
+ em = btrfs_get_chunk_map(fs_info, chunk_offset, 1);
if (IS_ERR(em)) {
/*
* This is a logic error, but we don't want to just rely on the
@@ -4884,7 +4884,7 @@ int btrfs_finish_chunk_alloc(struct btrfs_trans_handle *trans,
int i = 0;
int ret = 0;
- em = get_chunk_map(fs_info, chunk_offset, chunk_size);
+ em = btrfs_get_chunk_map(fs_info, chunk_offset, chunk_size);
if (IS_ERR(em))
return PTR_ERR(em);
@@ -5026,7 +5026,7 @@ int btrfs_chunk_readonly(struct btrfs_fs_info *fs_info, u64 chunk_offset)
int miss_ndevs = 0;
int i;
- em = get_chunk_map(fs_info, chunk_offset, 1);
+ em = btrfs_get_chunk_map(fs_info, chunk_offset, 1);
if (IS_ERR(em))
return 1;
@@ -5086,7 +5086,7 @@ int btrfs_num_copies(struct btrfs_fs_info *fs_info, u64 logical, u64 len)
struct map_lookup *map;
int ret;
- em = get_chunk_map(fs_info, logical, len);
+ em = btrfs_get_chunk_map(fs_info, logical, len);
if (IS_ERR(em))
/*
* We could return errors for these cases, but that could get
@@ -5132,7 +5132,7 @@ unsigned long btrfs_full_stripe_len(struct btrfs_fs_info *fs_info,
struct map_lookup *map;
unsigned long len = fs_info->sectorsize;
- em = get_chunk_map(fs_info, logical, len);
+ em = btrfs_get_chunk_map(fs_info, logical, len);
if (!WARN_ON(IS_ERR(em))) {
map = em->map_lookup;
@@ -5149,7 +5149,7 @@ int btrfs_is_parity_mirror(struct btrfs_fs_info *fs_info, u64 logical, u64 len)
struct map_lookup *map;
int ret = 0;
- em = get_chunk_map(fs_info, logical, len);
+ em = btrfs_get_chunk_map(fs_info, logical, len);
if(!WARN_ON(IS_ERR(em))) {
map = em->map_lookup;
@@ -5308,7 +5308,7 @@ static int __btrfs_map_block_for_discard(struct btrfs_fs_info *fs_info,
/* discard always return a bbio */
ASSERT(bbio_ret);
- em = get_chunk_map(fs_info, logical, length);
+ em = btrfs_get_chunk_map(fs_info, logical, length);
if (IS_ERR(em))
return PTR_ERR(em);
@@ -5634,7 +5634,7 @@ static int __btrfs_map_block(struct btrfs_fs_info *fs_info,
return __btrfs_map_block_for_discard(fs_info, logical,
*length, bbio_ret);
- em = get_chunk_map(fs_info, logical, *length);
+ em = btrfs_get_chunk_map(fs_info, logical, *length);
if (IS_ERR(em))
return PTR_ERR(em);
@@ -5933,7 +5933,7 @@ int btrfs_rmap_block(struct btrfs_fs_info *fs_info, u64 chunk_start,
u64 rmap_len;
int i, j, nr = 0;
- em = get_chunk_map(fs_info, chunk_start, 1);
+ em = btrfs_get_chunk_map(fs_info, chunk_start, 1);
if (IS_ERR(em))
return -EIO;
diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h
index 23e9285d88de..d68c8a05a774 100644
--- a/fs/btrfs/volumes.h
+++ b/fs/btrfs/volumes.h
@@ -465,6 +465,15 @@ unsigned long btrfs_full_stripe_len(struct btrfs_fs_info *fs_info,
int btrfs_finish_chunk_alloc(struct btrfs_trans_handle *trans,
u64 chunk_offset, u64 chunk_size);
int btrfs_remove_chunk(struct btrfs_trans_handle *trans, u64 chunk_offset);
+/**
+ * btrfs_get_chunk_map() - Find the mapping containing the given logical extent.
+ * @logical: Logical block offset in bytes.
+ * @length: Length of extent in bytes.
+ *
+ * Return: Chunk mapping or ERR_PTR.
+ */
+struct extent_map *btrfs_get_chunk_map(struct btrfs_fs_info *fs_info,
+ u64 logical, u64 length);
static inline void btrfs_dev_stat_inc(struct btrfs_device *dev,
int index)
--
2.18.0
^ permalink raw reply related [flat|nested] 9+ messages in thread* [PATCH v5 6/6] Btrfs: support swap files
2018-08-31 22:36 [PATCH v5 0/6] Btrfs: implement swap file support Omar Sandoval
` (4 preceding siblings ...)
2018-08-31 22:36 ` [PATCH v5 5/6] Btrfs: rename get_chunk_map() and make it non-static Omar Sandoval
@ 2018-08-31 22:36 ` Omar Sandoval
2018-09-06 11:59 ` [PATCH v5 0/6] Btrfs: implement swap file support David Sterba
6 siblings, 0 replies; 9+ messages in thread
From: Omar Sandoval @ 2018-08-31 22:36 UTC (permalink / raw)
To: linux-btrfs; +Cc: kernel-team
From: Omar Sandoval <osandov@fb.com>
Implement the swap file a_ops on Btrfs. Activation needs to make sure
that the file can be used as a swap file, which currently means it must
be fully allocated as nocow with no compression on one device. It also
sets up the swap extents directly with add_swap_extent(), so export it.
Signed-off-by: Omar Sandoval <osandov@fb.com>
---
fs/btrfs/inode.c | 232 +++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 232 insertions(+)
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 9357a19d2bff..c0409e632768 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -27,6 +27,7 @@
#include <linux/uio.h>
#include <linux/magic.h>
#include <linux/iversion.h>
+#include <linux/swap.h>
#include <asm/unaligned.h>
#include "ctree.h"
#include "disk-io.h"
@@ -10437,6 +10438,235 @@ void btrfs_set_range_writeback(struct extent_io_tree *tree, u64 start, u64 end)
}
}
+struct btrfs_swap_info {
+ u64 start;
+ u64 block_start;
+ u64 block_len;
+ u64 lowest_ppage;
+ u64 highest_ppage;
+ unsigned long nr_pages;
+ int nr_extents;
+};
+
+static int btrfs_add_swap_extent(struct swap_info_struct *sis,
+ struct btrfs_swap_info *bsi)
+{
+ unsigned long nr_pages;
+ u64 first_ppage, first_ppage_reported, next_ppage;
+ int ret;
+
+ first_ppage = ALIGN(bsi->block_start, PAGE_SIZE) >> PAGE_SHIFT;
+ next_ppage = ALIGN_DOWN(bsi->block_start + bsi->block_len,
+ PAGE_SIZE) >> PAGE_SHIFT;
+
+ if (first_ppage >= next_ppage)
+ return 0;
+ nr_pages = next_ppage - first_ppage;
+
+ first_ppage_reported = first_ppage;
+ if (bsi->start == 0)
+ first_ppage_reported++;
+ if (bsi->lowest_ppage > first_ppage_reported)
+ bsi->lowest_ppage = first_ppage_reported;
+ if (bsi->highest_ppage < (next_ppage - 1))
+ bsi->highest_ppage = next_ppage - 1;
+
+ ret = add_swap_extent(sis, bsi->nr_pages, nr_pages, first_ppage);
+ if (ret < 0)
+ return ret;
+ bsi->nr_extents += ret;
+ bsi->nr_pages += nr_pages;
+ return 0;
+}
+
+static int btrfs_swap_activate(struct swap_info_struct *sis, struct file *file,
+ sector_t *span)
+{
+ struct inode *inode = file_inode(file);
+ struct btrfs_fs_info *fs_info = BTRFS_I(inode)->root->fs_info;
+ struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+ struct extent_state *cached_state = NULL;
+ struct extent_map *em = NULL;
+ struct btrfs_device *device = NULL;
+ struct btrfs_swap_info bsi = {
+ .lowest_ppage = (sector_t)-1ULL,
+ };
+ int ret = 0;
+ u64 isize = inode->i_size;
+ u64 start;
+
+ /*
+ * If the swap file was just created, make sure delalloc is done. If the
+ * file changes again after this, the user is doing something stupid and
+ * we don't really care.
+ */
+ ret = btrfs_wait_ordered_range(inode, 0, (u64)-1);
+ if (ret)
+ return ret;
+
+ /*
+ * The inode is locked, so these flags won't change after we check them.
+ */
+ if (BTRFS_I(inode)->flags & BTRFS_INODE_COMPRESS) {
+ btrfs_err(fs_info, "swapfile must not be compressed");
+ return -EINVAL;
+ }
+ if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW)) {
+ btrfs_err(fs_info, "swapfile must not be copy-on-write");
+ return -EINVAL;
+ }
+
+ /*
+ * Balance or device remove/replace/resize can move stuff around from
+ * under us. The EXCL_OP flag makes sure they aren't running/won't run
+ * concurrently while we are mapping the swap extents, and the fs_info
+ * nr_swapfiles counter prevents them from running while the swap file
+ * is active and moving the extents. Note that this also prevents a
+ * concurrent device add which isn't actually necessary, but it's not
+ * really worth the trouble to allow it.
+ */
+ if (test_and_set_bit(BTRFS_FS_EXCL_OP, &fs_info->flags))
+ return -EBUSY;
+ atomic_inc(&fs_info->nr_swapfiles);
+ /*
+ * Snapshots can create extents which require COW even if NODATACOW is
+ * set. We use this counter to prevent snapshots. We must increment it
+ * before walking the extents because we don't want a concurrent
+ * snapshot to run after we've already checked the extents.
+ */
+ atomic_inc(&BTRFS_I(inode)->root->nr_swapfiles);
+
+ lock_extent_bits(io_tree, 0, isize - 1, &cached_state);
+ start = 0;
+ while (start < isize) {
+ u64 end, logical_block_start, physical_block_start;
+ u64 len = isize - start;
+
+ em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, start, len, 0);
+ if (IS_ERR(em)) {
+ ret = PTR_ERR(em);
+ goto out;
+ }
+ end = extent_map_end(em);
+
+ if (em->block_start == EXTENT_MAP_HOLE) {
+ btrfs_err(fs_info, "swapfile must not have holes");
+ ret = -EINVAL;
+ goto out;
+ }
+ if (em->block_start == EXTENT_MAP_INLINE) {
+ /*
+ * It's unlikely we'll ever actually find ourselves
+ * here, as a file small enough to fit inline won't be
+ * big enough to store more than the swap header, but in
+ * case something changes in the future, let's catch it
+ * here rather than later.
+ */
+ btrfs_err(fs_info, "swapfile must not be inline");
+ ret = -EINVAL;
+ goto out;
+ }
+ if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) {
+ btrfs_err(fs_info, "swapfile must not be compressed");
+ ret = -EINVAL;
+ goto out;
+ }
+
+ logical_block_start = em->block_start + (start - em->start);
+ len = min(len, em->len - (start - em->start));
+ free_extent_map(em);
+ em = NULL;
+
+ ret = can_nocow_extent(inode, start, &len, NULL, NULL, NULL);
+ if (ret < 0) {
+ goto out;
+ } else if (ret) {
+ ret = 0;
+ } else {
+ btrfs_err(fs_info, "swapfile must not be copy-on-write");
+ ret = -EINVAL;
+ goto out;
+ }
+
+ em = btrfs_get_chunk_map(fs_info, logical_block_start, len);
+ if (IS_ERR(em)) {
+ ret = PTR_ERR(em);
+ goto out;
+ }
+
+ if (em->map_lookup->type & BTRFS_BLOCK_GROUP_PROFILE_MASK) {
+ btrfs_err(fs_info, "swapfile must have single data profile");
+ ret = -EINVAL;
+ goto out;
+ }
+
+ if (device == NULL) {
+ device = em->map_lookup->stripes[0].dev;
+ } else if (device != em->map_lookup->stripes[0].dev) {
+ btrfs_err(fs_info, "swapfile must be on one device");
+ ret = -EINVAL;
+ goto out;
+ }
+
+ physical_block_start = (em->map_lookup->stripes[0].physical +
+ (logical_block_start - em->start));
+ len = min(len, em->len - (logical_block_start - em->start));
+ free_extent_map(em);
+ em = NULL;
+
+ if (bsi.block_len &&
+ bsi.block_start + bsi.block_len == physical_block_start) {
+ bsi.block_len += len;
+ } else {
+ if (bsi.block_len) {
+ ret = btrfs_add_swap_extent(sis, &bsi);
+ if (ret)
+ goto out;
+ }
+ bsi.start = start;
+ bsi.block_start = physical_block_start;
+ bsi.block_len = len;
+ }
+
+ start = end;
+ }
+
+ if (bsi.block_len)
+ ret = btrfs_add_swap_extent(sis, &bsi);
+
+out:
+ if (!IS_ERR_OR_NULL(em))
+ free_extent_map(em);
+
+ unlock_extent_cached(io_tree, 0, isize - 1, &cached_state);
+
+ if (ret) {
+ atomic_dec(&BTRFS_I(inode)->root->nr_swapfiles);
+ atomic_dec(&fs_info->nr_swapfiles);
+ }
+
+ clear_bit(BTRFS_FS_EXCL_OP, &fs_info->flags);
+
+ if (ret)
+ return ret;
+
+ if (device)
+ sis->bdev = device->bdev;
+ *span = bsi.highest_ppage - bsi.lowest_ppage + 1;
+ sis->max = bsi.nr_pages;
+ sis->pages = bsi.nr_pages - 1;
+ sis->highest_bit = bsi.nr_pages - 1;
+ return bsi.nr_extents;
+}
+
+static void btrfs_swap_deactivate(struct file *file)
+{
+ struct inode *inode = file_inode(file);
+
+ atomic_dec(&BTRFS_I(inode)->root->nr_swapfiles);
+ atomic_dec(&BTRFS_I(inode)->root->fs_info->nr_swapfiles);
+}
+
static const struct inode_operations btrfs_dir_inode_operations = {
.getattr = btrfs_getattr,
.lookup = btrfs_lookup,
@@ -10514,6 +10744,8 @@ static const struct address_space_operations btrfs_aops = {
.releasepage = btrfs_releasepage,
.set_page_dirty = btrfs_set_page_dirty,
.error_remove_page = generic_error_remove_page,
+ .swap_activate = btrfs_swap_activate,
+ .swap_deactivate = btrfs_swap_deactivate,
};
static const struct address_space_operations btrfs_symlink_aops = {
--
2.18.0
^ permalink raw reply related [flat|nested] 9+ messages in thread* Re: [PATCH v5 0/6] Btrfs: implement swap file support
2018-08-31 22:36 [PATCH v5 0/6] Btrfs: implement swap file support Omar Sandoval
` (5 preceding siblings ...)
2018-08-31 22:36 ` [PATCH v5 6/6] Btrfs: support swap files Omar Sandoval
@ 2018-09-06 11:59 ` David Sterba
2018-09-06 18:08 ` Omar Sandoval
6 siblings, 1 reply; 9+ messages in thread
From: David Sterba @ 2018-09-06 11:59 UTC (permalink / raw)
To: Omar Sandoval; +Cc: linux-btrfs, kernel-team
On Fri, Aug 31, 2018 at 03:36:35PM -0700, Omar Sandoval wrote:
> This series implements swap file support for Btrfs.
>
> Changes since v4 [1]:
>
> - Added a kernel doc for btrfs_get_chunk_map()
> - Got rid of "Btrfs: push EXCL_OP set into btrfs_rm_device()"
> - Made activate error messages more clear and consistent
> - Changed clear vs unlock order in activate error case
> - Added "mm: export add_swap_extent()" as a separate patch
> - Added a btrfs_wait_ordered_range() at the beginning of
> btrfs_swap_activate() to catch newly created files
> - Added some Reviewed-bys from Nikolay
>
> I took a stab at adding support for balance when a swap file is active,
> but it's a major pain: we need to mark block groups which contain swap
> file extents, check the block group counter in relocate/scrub, then
> unmark the block groups when the swap file is deactivated, which gets
> really messy because the file can grow while it is an active swap file.
> If this is a deal breaker, I can work something out, but I don't think
> it's worth the trouble.
I'm afraid it is a deal breaker. Unlike dev-replace or resize, balance
is used more often so switching off the swap file for the duration of
the operation is administration pain.
If it's possible to constrain the swap file further, like no growing
that you mention, or mandatory preallocation or similar, then I hope it
would make it possible to implement in a sane way.
^ permalink raw reply [flat|nested] 9+ messages in thread* Re: [PATCH v5 0/6] Btrfs: implement swap file support
2018-09-06 11:59 ` [PATCH v5 0/6] Btrfs: implement swap file support David Sterba
@ 2018-09-06 18:08 ` Omar Sandoval
0 siblings, 0 replies; 9+ messages in thread
From: Omar Sandoval @ 2018-09-06 18:08 UTC (permalink / raw)
To: dsterba, linux-btrfs, kernel-team
On Thu, Sep 06, 2018 at 01:59:54PM +0200, David Sterba wrote:
> On Fri, Aug 31, 2018 at 03:36:35PM -0700, Omar Sandoval wrote:
> > This series implements swap file support for Btrfs.
> >
> > Changes since v4 [1]:
> >
> > - Added a kernel doc for btrfs_get_chunk_map()
> > - Got rid of "Btrfs: push EXCL_OP set into btrfs_rm_device()"
> > - Made activate error messages more clear and consistent
> > - Changed clear vs unlock order in activate error case
> > - Added "mm: export add_swap_extent()" as a separate patch
> > - Added a btrfs_wait_ordered_range() at the beginning of
> > btrfs_swap_activate() to catch newly created files
> > - Added some Reviewed-bys from Nikolay
> >
> > I took a stab at adding support for balance when a swap file is active,
> > but it's a major pain: we need to mark block groups which contain swap
> > file extents, check the block group counter in relocate/scrub, then
> > unmark the block groups when the swap file is deactivated, which gets
> > really messy because the file can grow while it is an active swap file.
> > If this is a deal breaker, I can work something out, but I don't think
> > it's worth the trouble.
>
> I'm afraid it is a deal breaker. Unlike dev-replace or resize, balance
> is used more often so switching off the swap file for the duration of
> the operation is administration pain.
>
> If it's possible to constrain the swap file further, like no growing
> that you mention, or mandatory preallocation or similar, then I hope it
> would make it possible to implement in a sane way.
Alright, I'll have another go.
^ permalink raw reply [flat|nested] 9+ messages in thread