From: Filipe Manana <[email protected]>
[ Upstream commit e383e158ed1b6abc2d2d3e6736d77a46393f80fa ]
When logging an inode and we require to copy items from subvolume leaves
to the log tree, we clone each subvolume leaf and than use that clone to
copy items to the log tree. This is required to avoid possible deadlocks
as stated in commit 796787c978ef ("btrfs: do not modify log tree while
holding a leaf from fs tree locked").
The cloning requires allocating an extent buffer (struct extent_buffer)
and then allocating pages (folios) to attach to the extent buffer. This
may be slow in case we are under memory pressure, and since we are doing
the cloning while holding a read lock on a subvolume leaf, it means we
can be blocking other operations on that leaf for significant periods of
time, which can increase latency on operations like creating other files,
renaming files, etc. Similarly because we're under a log transaction, we
may also cause extra delay on other tasks doing an fsync, because syncing
the log requires waiting for tasks that joined a log transaction to exit
the transaction.
So to improve this, for any inode logging operation that needs to copy
items from a subvolume leaf ("full sync" or "copy everything" bit set
in the inode), preallocate a dummy extent buffer before locking any
extent buffer from the subvolume tree, and even before joining a log
transaction, add it to the log context and then use it when we need to
copy items from a subvolume leaf to the log tree. This avoids making
other operations get extra latency when waiting to lock a subvolume
leaf that is used during inode logging and we are under heavy memory
pressure.
The following test script with bonnie++ was used to test this:
$ cat test.sh
#!/bin/bash
DEV=/dev/sdh
MNT=/mnt/sdh
MOUNT_OPTIONS="-o ssd"
MEMTOTAL_BYTES=`free -b | grep Mem: | awk '{ print $2 }'`
NR_DIRECTORIES=20
NR_FILES=20480
DATASET_SIZE=$((MEMTOTAL_BYTES * 2 / 1048576))
DIRECTORY_SIZE=$((MEMTOTAL_BYTES * 2 / NR_FILES))
NR_FILES=$((NR_FILES / 1024))
echo "performance" | \
tee /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor
umount $DEV &> /dev/null
mkfs.btrfs -f $MKFS_OPTIONS $DEV
mount $MOUNT_OPTIONS $DEV $MNT
bonnie++ -u root -d $MNT \
-n $NR_FILES:$DIRECTORY_SIZE:$DIRECTORY_SIZE:$NR_DIRECTORIES \
-r 0 -s $DATASET_SIZE -b
umount $MNT
The results of this test on a 8G VM running a non-debug kernel (Debian's
default kernel config), were the following.
Before this change:
Version 2.00a ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Name:Size etc /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
debian0 7501M 376k 99 1.4g 96 117m 14 1510k 99 2.5g 95 +++++ +++
Latency 35068us 24976us 2944ms 30725us 71770us 26152us
Version 2.00a ------Sequential Create------ --------Random Create--------
debian0 -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
20:384100:384100/20 20480 32 20480 58 20480 48 20480 39 20480 56 20480 61
Latency 411ms 11914us 119ms 617ms 10296us 110ms
After this change:
Version 2.00a ------Sequential Output------ --Sequential Input- --Random-
-Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Name:Size etc /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
debian0 7501M 375k 99 1.4g 97 117m 14 1546k 99 2.3g 98 +++++ +++
Latency 35975us 20945us 2144ms 10297us 2217us 6004us
Version 2.00a ------Sequential Create------ --------Random Create--------
debian0 -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
20:384100:384100/20 20480 35 20480 58 20480 48 20480 40 20480 57 20480 59
Latency 320ms 11237us 77779us 518ms 6470us 86389us
Reviewed-by: Josef Bacik <[email protected]>
Signed-off-by: Filipe Manana <[email protected]>
Reviewed-by: David Sterba <[email protected]>
Signed-off-by: David Sterba <[email protected]>
Signed-off-by: Sasha Levin <[email protected]>
---
fs/btrfs/file.c | 12 ++++++
fs/btrfs/tree-log.c | 93 +++++++++++++++++++++++++++------------------
fs/btrfs/tree-log.h | 25 ++++++++++++
3 files changed, 94 insertions(+), 36 deletions(-)
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index 38dfcac476099..77c3b4567b91a 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -1912,6 +1912,8 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
goto out_release_extents;
}
+ btrfs_init_log_ctx_scratch_eb(&ctx);
+
/*
* We use start here because we will need to wait on the IO to complete
* in btrfs_sync_log, which could require joining a transaction (for
@@ -1931,6 +1933,15 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
trans->in_fsync = true;
ret = btrfs_log_dentry_safe(trans, dentry, &ctx);
+ /*
+ * Scratch eb no longer needed, release before syncing log or commit
+ * transaction, to avoid holding unnecessary memory during such long
+ * operations.
+ */
+ if (ctx.scratch_eb) {
+ free_extent_buffer(ctx.scratch_eb);
+ ctx.scratch_eb = NULL;
+ }
btrfs_release_log_ctx_extents(&ctx);
if (ret < 0) {
/* Fallthrough and commit/free transaction. */
@@ -2006,6 +2017,7 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
ret = btrfs_commit_transaction(trans);
out:
+ free_extent_buffer(ctx.scratch_eb);
ASSERT(list_empty(&ctx.list));
ASSERT(list_empty(&ctx.conflict_inodes));
err = file_check_and_advance_wb_err(file);
diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
index 331fc7429952f..761b13b3d3421 100644
--- a/fs/btrfs/tree-log.c
+++ b/fs/btrfs/tree-log.c
@@ -3619,6 +3619,30 @@ static int flush_dir_items_batch(struct btrfs_trans_handle *trans,
return ret;
}
+static int clone_leaf(struct btrfs_path *path, struct btrfs_log_ctx *ctx)
+{
+ const int slot = path->slots[0];
+
+ if (ctx->scratch_eb) {
+ copy_extent_buffer_full(ctx->scratch_eb, path->nodes[0]);
+ } else {
+ ctx->scratch_eb = btrfs_clone_extent_buffer(path->nodes[0]);
+ if (!ctx->scratch_eb)
+ return -ENOMEM;
+ }
+
+ btrfs_release_path(path);
+ path->nodes[0] = ctx->scratch_eb;
+ path->slots[0] = slot;
+ /*
+ * Add extra ref to scratch eb so that it is not freed when callers
+ * release the path, so we can reuse it later if needed.
+ */
+ atomic_inc(&ctx->scratch_eb->refs);
+
+ return 0;
+}
+
static int process_dir_items_leaf(struct btrfs_trans_handle *trans,
struct btrfs_inode *inode,
struct btrfs_path *path,
@@ -3633,23 +3657,20 @@ static int process_dir_items_leaf(struct btrfs_trans_handle *trans,
bool last_found = false;
int batch_start = 0;
int batch_size = 0;
- int i;
+ int ret;
/*
* We need to clone the leaf, release the read lock on it, and use the
* clone before modifying the log tree. See the comment at copy_items()
* about why we need to do this.
*/
- src = btrfs_clone_extent_buffer(path->nodes[0]);
- if (!src)
- return -ENOMEM;
+ ret = clone_leaf(path, ctx);
+ if (ret < 0)
+ return ret;
- i = path->slots[0];
- btrfs_release_path(path);
- path->nodes[0] = src;
- path->slots[0] = i;
+ src = path->nodes[0];
- for (; i < nritems; i++) {
+ for (int i = path->slots[0]; i < nritems; i++) {
struct btrfs_dir_item *di;
struct btrfs_key key;
int ret;
@@ -4259,17 +4280,16 @@ static noinline int copy_items(struct btrfs_trans_handle *trans,
struct btrfs_path *dst_path,
struct btrfs_path *src_path,
int start_slot, int nr, int inode_only,
- u64 logged_isize)
+ u64 logged_isize, struct btrfs_log_ctx *ctx)
{
struct btrfs_root *log = inode->root->log_root;
struct btrfs_file_extent_item *extent;
struct extent_buffer *src;
- int ret = 0;
+ int ret;
struct btrfs_key *ins_keys;
u32 *ins_sizes;
struct btrfs_item_batch batch;
char *ins_data;
- int i;
int dst_index;
const bool skip_csum = (inode->flags & BTRFS_INODE_NODATASUM);
const u64 i_size = i_size_read(&inode->vfs_inode);
@@ -4302,14 +4322,11 @@ static noinline int copy_items(struct btrfs_trans_handle *trans,
* while the other is holding the delayed node's mutex and wants to
* write lock the same subvolume leaf for flushing delayed items.
*/
- src = btrfs_clone_extent_buffer(src_path->nodes[0]);
- if (!src)
- return -ENOMEM;
+ ret = clone_leaf(src_path, ctx);
+ if (ret < 0)
+ return ret;
- i = src_path->slots[0];
- btrfs_release_path(src_path);
- src_path->nodes[0] = src;
- src_path->slots[0] = i;
+ src = src_path->nodes[0];
ins_data = kmalloc(nr * sizeof(struct btrfs_key) +
nr * sizeof(u32), GFP_NOFS);
@@ -4324,7 +4341,7 @@ static noinline int copy_items(struct btrfs_trans_handle *trans,
batch.nr = 0;
dst_index = 0;
- for (i = 0; i < nr; i++) {
+ for (int i = 0; i < nr; i++) {
const int src_slot = start_slot + i;
struct btrfs_root *csum_root;
struct btrfs_ordered_sum *sums;
@@ -4431,7 +4448,7 @@ static noinline int copy_items(struct btrfs_trans_handle *trans,
goto out;
dst_index = 0;
- for (i = 0; i < nr; i++) {
+ for (int i = 0; i < nr; i++) {
const int src_slot = start_slot + i;
const int dst_slot = dst_path->slots[0] + dst_index;
struct btrfs_key key;
@@ -4704,7 +4721,8 @@ static int log_one_extent(struct btrfs_trans_handle *trans,
*/
static int btrfs_log_prealloc_extents(struct btrfs_trans_handle *trans,
struct btrfs_inode *inode,
- struct btrfs_path *path)
+ struct btrfs_path *path,
+ struct btrfs_log_ctx *ctx)
{
struct btrfs_root *root = inode->root;
struct btrfs_key key;
@@ -4770,7 +4788,7 @@ static int btrfs_log_prealloc_extents(struct btrfs_trans_handle *trans,
if (slot >= btrfs_header_nritems(leaf)) {
if (ins_nr > 0) {
ret = copy_items(trans, inode, dst_path, path,
- start_slot, ins_nr, 1, 0);
+ start_slot, ins_nr, 1, 0, ctx);
if (ret < 0)
goto out;
ins_nr = 0;
@@ -4820,7 +4838,7 @@ static int btrfs_log_prealloc_extents(struct btrfs_trans_handle *trans,
}
if (ins_nr > 0)
ret = copy_items(trans, inode, dst_path, path,
- start_slot, ins_nr, 1, 0);
+ start_slot, ins_nr, 1, 0, ctx);
out:
btrfs_release_path(path);
btrfs_free_path(dst_path);
@@ -4899,7 +4917,7 @@ static int btrfs_log_changed_extents(struct btrfs_trans_handle *trans,
write_unlock(&tree->lock);
if (!ret)
- ret = btrfs_log_prealloc_extents(trans, inode, path);
+ ret = btrfs_log_prealloc_extents(trans, inode, path, ctx);
if (ret)
return ret;
@@ -4980,7 +4998,8 @@ static int logged_inode_size(struct btrfs_root *log, struct btrfs_inode *inode,
static int btrfs_log_all_xattrs(struct btrfs_trans_handle *trans,
struct btrfs_inode *inode,
struct btrfs_path *path,
- struct btrfs_path *dst_path)
+ struct btrfs_path *dst_path,
+ struct btrfs_log_ctx *ctx)
{
struct btrfs_root *root = inode->root;
int ret;
@@ -5009,7 +5028,7 @@ static int btrfs_log_all_xattrs(struct btrfs_trans_handle *trans,
if (slot >= nritems) {
if (ins_nr > 0) {
ret = copy_items(trans, inode, dst_path, path,
- start_slot, ins_nr, 1, 0);
+ start_slot, ins_nr, 1, 0, ctx);
if (ret < 0)
return ret;
ins_nr = 0;
@@ -5035,7 +5054,7 @@ static int btrfs_log_all_xattrs(struct btrfs_trans_handle *trans,
}
if (ins_nr > 0) {
ret = copy_items(trans, inode, dst_path, path,
- start_slot, ins_nr, 1, 0);
+ start_slot, ins_nr, 1, 0, ctx);
if (ret < 0)
return ret;
}
@@ -5847,7 +5866,7 @@ static int copy_inode_items_to_log(struct btrfs_trans_handle *trans,
}
ret = copy_items(trans, inode, dst_path, path,
ins_start_slot, ins_nr,
- inode_only, logged_isize);
+ inode_only, logged_isize, ctx);
if (ret < 0)
return ret;
ins_nr = 0;
@@ -5866,7 +5885,7 @@ static int copy_inode_items_to_log(struct btrfs_trans_handle *trans,
goto next_slot;
ret = copy_items(trans, inode, dst_path, path,
ins_start_slot,
- ins_nr, inode_only, logged_isize);
+ ins_nr, inode_only, logged_isize, ctx);
if (ret < 0)
return ret;
ins_nr = 0;
@@ -5883,7 +5902,7 @@ static int copy_inode_items_to_log(struct btrfs_trans_handle *trans,
}
ret = copy_items(trans, inode, dst_path, path, ins_start_slot,
- ins_nr, inode_only, logged_isize);
+ ins_nr, inode_only, logged_isize, ctx);
if (ret < 0)
return ret;
ins_nr = 1;
@@ -5898,7 +5917,7 @@ static int copy_inode_items_to_log(struct btrfs_trans_handle *trans,
if (ins_nr) {
ret = copy_items(trans, inode, dst_path, path,
ins_start_slot, ins_nr, inode_only,
- logged_isize);
+ logged_isize, ctx);
if (ret < 0)
return ret;
ins_nr = 0;
@@ -5923,7 +5942,7 @@ static int copy_inode_items_to_log(struct btrfs_trans_handle *trans,
}
if (ins_nr) {
ret = copy_items(trans, inode, dst_path, path, ins_start_slot,
- ins_nr, inode_only, logged_isize);
+ ins_nr, inode_only, logged_isize, ctx);
if (ret)
return ret;
}
@@ -5934,7 +5953,7 @@ static int copy_inode_items_to_log(struct btrfs_trans_handle *trans,
* lock the same leaf with btrfs_log_prealloc_extents() below.
*/
btrfs_release_path(path);
- ret = btrfs_log_prealloc_extents(trans, inode, dst_path);
+ ret = btrfs_log_prealloc_extents(trans, inode, dst_path, ctx);
}
return ret;
@@ -6526,7 +6545,7 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans,
btrfs_release_path(path);
btrfs_release_path(dst_path);
- ret = btrfs_log_all_xattrs(trans, inode, path, dst_path);
+ ret = btrfs_log_all_xattrs(trans, inode, path, dst_path, ctx);
if (ret)
goto out_unlock;
xattrs_logged = true;
@@ -6553,7 +6572,7 @@ static int btrfs_log_inode(struct btrfs_trans_handle *trans,
* BTRFS_INODE_COPY_EVERYTHING set.
*/
if (!xattrs_logged && inode->logged_trans < trans->transid) {
- ret = btrfs_log_all_xattrs(trans, inode, path, dst_path);
+ ret = btrfs_log_all_xattrs(trans, inode, path, dst_path, ctx);
if (ret)
goto out_unlock;
btrfs_release_path(path);
@@ -7502,6 +7521,7 @@ void btrfs_log_new_name(struct btrfs_trans_handle *trans,
btrfs_init_log_ctx(&ctx, &inode->vfs_inode);
ctx.logging_new_name = true;
+ btrfs_init_log_ctx_scratch_eb(&ctx);
/*
* We don't care about the return value. If we fail to log the new name
* then we know the next attempt to sync the log will fallback to a full
@@ -7510,6 +7530,7 @@ void btrfs_log_new_name(struct btrfs_trans_handle *trans,
* inconsistent state after a rename operation.
*/
btrfs_log_inode_parent(trans, inode, parent, LOG_INODE_EXISTS, &ctx);
+ free_extent_buffer(ctx.scratch_eb);
ASSERT(list_empty(&ctx.conflict_inodes));
out:
/*
diff --git a/fs/btrfs/tree-log.h b/fs/btrfs/tree-log.h
index a550a8a375cd1..af219e8840d28 100644
--- a/fs/btrfs/tree-log.h
+++ b/fs/btrfs/tree-log.h
@@ -36,6 +36,15 @@ struct btrfs_log_ctx {
struct list_head conflict_inodes;
int num_conflict_inodes;
bool logging_conflict_inodes;
+ /*
+ * Used for fsyncs that need to copy items from the subvolume tree to
+ * the log tree (full sync flag set or copy everything flag set) to
+ * avoid allocating a temporary extent buffer while holding a lock on
+ * an extent buffer of the subvolume tree and under the log transaction.
+ * Also helps to avoid allocating and freeing a temporary extent buffer
+ * in case we need to process multiple leaves from the subvolume tree.
+ */
+ struct extent_buffer *scratch_eb;
};
static inline void btrfs_init_log_ctx(struct btrfs_log_ctx *ctx,
@@ -53,6 +62,22 @@ static inline void btrfs_init_log_ctx(struct btrfs_log_ctx *ctx,
INIT_LIST_HEAD(&ctx->conflict_inodes);
ctx->num_conflict_inodes = 0;
ctx->logging_conflict_inodes = false;
+ ctx->scratch_eb = NULL;
+}
+
+static inline void btrfs_init_log_ctx_scratch_eb(struct btrfs_log_ctx *ctx)
+{
+ struct btrfs_inode *inode = BTRFS_I(ctx->inode);
+
+ if (!test_bit(BTRFS_INODE_NEEDS_FULL_SYNC, &inode->runtime_flags) &&
+ !test_bit(BTRFS_INODE_COPY_EVERYTHING, &inode->runtime_flags))
+ return;
+
+ /*
+ * Don't care about allocation failure. This is just for optimization,
+ * if we fail to allocate here, we will try again later if needed.
+ */
+ ctx->scratch_eb = alloc_dummy_extent_buffer(inode->root->fs_info, 0);
}
static inline void btrfs_release_log_ctx_extents(struct btrfs_log_ctx *ctx)
--
2.43.0
On Fri, Mar 29, 2024 at 08:25:55AM -0400, Sasha Levin wrote:
> From: Filipe Manana <[email protected]>
>
> [ Upstream commit e383e158ed1b6abc2d2d3e6736d77a46393f80fa ]
>
> When logging an inode and we require to copy items from subvolume leaves
> to the log tree, we clone each subvolume leaf and than use that clone to
> copy items to the log tree. This is required to avoid possible deadlocks
> as stated in commit 796787c978ef ("btrfs: do not modify log tree while
> holding a leaf from fs tree locked").
>
> The cloning requires allocating an extent buffer (struct extent_buffer)
> and then allocating pages (folios) to attach to the extent buffer. This
> may be slow in case we are under memory pressure, and since we are doing
> the cloning while holding a read lock on a subvolume leaf, it means we
> can be blocking other operations on that leaf for significant periods of
> time, which can increase latency on operations like creating other files,
> renaming files, etc. Similarly because we're under a log transaction, we
> may also cause extra delay on other tasks doing an fsync, because syncing
> the log requires waiting for tasks that joined a log transaction to exit
> the transaction.
>
> So to improve this, for any inode logging operation that needs to copy
> items from a subvolume leaf ("full sync" or "copy everything" bit set
> in the inode), preallocate a dummy extent buffer before locking any
> extent buffer from the subvolume tree, and even before joining a log
> transaction, add it to the log context and then use it when we need to
> copy items from a subvolume leaf to the log tree. This avoids making
> other operations get extra latency when waiting to lock a subvolume
> leaf that is used during inode logging and we are under heavy memory
> pressure.
>
> The following test script with bonnie++ was used to test this:
>
> $ cat test.sh
> #!/bin/bash
>
> DEV=/dev/sdh
> MNT=/mnt/sdh
> MOUNT_OPTIONS="-o ssd"
>
> MEMTOTAL_BYTES=`free -b | grep Mem: | awk '{ print $2 }'`
> NR_DIRECTORIES=20
> NR_FILES=20480
> DATASET_SIZE=$((MEMTOTAL_BYTES * 2 / 1048576))
> DIRECTORY_SIZE=$((MEMTOTAL_BYTES * 2 / NR_FILES))
> NR_FILES=$((NR_FILES / 1024))
>
> echo "performance" | \
> tee /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor
>
> umount $DEV &> /dev/null
> mkfs.btrfs -f $MKFS_OPTIONS $DEV
> mount $MOUNT_OPTIONS $DEV $MNT
>
> bonnie++ -u root -d $MNT \
> -n $NR_FILES:$DIRECTORY_SIZE:$DIRECTORY_SIZE:$NR_DIRECTORIES \
> -r 0 -s $DATASET_SIZE -b
>
> umount $MNT
>
> The results of this test on a 8G VM running a non-debug kernel (Debian's
> default kernel config), were the following.
>
> Before this change:
>
> Version 2.00a ------Sequential Output------ --Sequential Input- --Random-
> -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
> Name:Size etc /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
> debian0 7501M 376k 99 1.4g 96 117m 14 1510k 99 2.5g 95 +++++ +++
> Latency 35068us 24976us 2944ms 30725us 71770us 26152us
> Version 2.00a ------Sequential Create------ --------Random Create--------
> debian0 -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
> files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
> 20:384100:384100/20 20480 32 20480 58 20480 48 20480 39 20480 56 20480 61
> Latency 411ms 11914us 119ms 617ms 10296us 110ms
>
> After this change:
>
> Version 2.00a ------Sequential Output------ --Sequential Input- --Random-
> -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
> Name:Size etc /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
> debian0 7501M 375k 99 1.4g 97 117m 14 1546k 99 2.3g 98 +++++ +++
> Latency 35975us 20945us 2144ms 10297us 2217us 6004us
> Version 2.00a ------Sequential Create------ --------Random Create--------
> debian0 -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
> files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
> 20:384100:384100/20 20480 35 20480 58 20480 48 20480 40 20480 57 20480 59
> Latency 320ms 11237us 77779us 518ms 6470us 86389us
>
> Reviewed-by: Josef Bacik <[email protected]>
> Signed-off-by: Filipe Manana <[email protected]>
> Reviewed-by: David Sterba <[email protected]>
> Signed-off-by: David Sterba <[email protected]>
> Signed-off-by: Sasha Levin <[email protected]>
This is a performance improvement, how does this qualify for stable? I
read only about notable perfromance fixes but this is not one.
On Tue, Apr 02, 2024 at 03:35:18PM +0200, David Sterba wrote:
>On Fri, Mar 29, 2024 at 08:25:55AM -0400, Sasha Levin wrote:
>> From: Filipe Manana <[email protected]>
>>
>> [ Upstream commit e383e158ed1b6abc2d2d3e6736d77a46393f80fa ]
>>
>> When logging an inode and we require to copy items from subvolume leaves
>> to the log tree, we clone each subvolume leaf and than use that clone to
>> copy items to the log tree. This is required to avoid possible deadlocks
>> as stated in commit 796787c978ef ("btrfs: do not modify log tree while
>> holding a leaf from fs tree locked").
>>
>> The cloning requires allocating an extent buffer (struct extent_buffer)
>> and then allocating pages (folios) to attach to the extent buffer. This
>> may be slow in case we are under memory pressure, and since we are doing
>> the cloning while holding a read lock on a subvolume leaf, it means we
>> can be blocking other operations on that leaf for significant periods of
>> time, which can increase latency on operations like creating other files,
>> renaming files, etc. Similarly because we're under a log transaction, we
>> may also cause extra delay on other tasks doing an fsync, because syncing
>> the log requires waiting for tasks that joined a log transaction to exit
>> the transaction.
>>
>> So to improve this, for any inode logging operation that needs to copy
>> items from a subvolume leaf ("full sync" or "copy everything" bit set
>> in the inode), preallocate a dummy extent buffer before locking any
>> extent buffer from the subvolume tree, and even before joining a log
>> transaction, add it to the log context and then use it when we need to
>> copy items from a subvolume leaf to the log tree. This avoids making
>> other operations get extra latency when waiting to lock a subvolume
>> leaf that is used during inode logging and we are under heavy memory
>> pressure.
>>
>> The following test script with bonnie++ was used to test this:
>>
>> $ cat test.sh
>> #!/bin/bash
>>
>> DEV=/dev/sdh
>> MNT=/mnt/sdh
>> MOUNT_OPTIONS="-o ssd"
>>
>> MEMTOTAL_BYTES=`free -b | grep Mem: | awk '{ print $2 }'`
>> NR_DIRECTORIES=20
>> NR_FILES=20480
>> DATASET_SIZE=$((MEMTOTAL_BYTES * 2 / 1048576))
>> DIRECTORY_SIZE=$((MEMTOTAL_BYTES * 2 / NR_FILES))
>> NR_FILES=$((NR_FILES / 1024))
>>
>> echo "performance" | \
>> tee /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor
>>
>> umount $DEV &> /dev/null
>> mkfs.btrfs -f $MKFS_OPTIONS $DEV
>> mount $MOUNT_OPTIONS $DEV $MNT
>>
>> bonnie++ -u root -d $MNT \
>> -n $NR_FILES:$DIRECTORY_SIZE:$DIRECTORY_SIZE:$NR_DIRECTORIES \
>> -r 0 -s $DATASET_SIZE -b
>>
>> umount $MNT
>>
>> The results of this test on a 8G VM running a non-debug kernel (Debian's
>> default kernel config), were the following.
>>
>> Before this change:
>>
>> Version 2.00a ------Sequential Output------ --Sequential Input- --Random-
>> -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
>> Name:Size etc /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
>> debian0 7501M 376k 99 1.4g 96 117m 14 1510k 99 2.5g 95 +++++ +++
>> Latency 35068us 24976us 2944ms 30725us 71770us 26152us
>> Version 2.00a ------Sequential Create------ --------Random Create--------
>> debian0 -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
>> files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
>> 20:384100:384100/20 20480 32 20480 58 20480 48 20480 39 20480 56 20480 61
>> Latency 411ms 11914us 119ms 617ms 10296us 110ms
>>
>> After this change:
>>
>> Version 2.00a ------Sequential Output------ --Sequential Input- --Random-
>> -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
>> Name:Size etc /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
>> debian0 7501M 375k 99 1.4g 97 117m 14 1546k 99 2.3g 98 +++++ +++
>> Latency 35975us 20945us 2144ms 10297us 2217us 6004us
>> Version 2.00a ------Sequential Create------ --------Random Create--------
>> debian0 -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
>> files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
>> 20:384100:384100/20 20480 35 20480 58 20480 48 20480 40 20480 57 20480 59
>> Latency 320ms 11237us 77779us 518ms 6470us 86389us
>>
>> Reviewed-by: Josef Bacik <[email protected]>
>> Signed-off-by: Filipe Manana <[email protected]>
>> Reviewed-by: David Sterba <[email protected]>
>> Signed-off-by: David Sterba <[email protected]>
>> Signed-off-by: Sasha Levin <[email protected]>
>
>This is a performance improvement, how does this qualify for stable? I
>read only about notable perfromance fixes but this is not one.
No objection to dropping it. Description of the commit states that it
fixes blocking for "significant amount of time".
--
Thanks,
Sasha
On Tue, Apr 02, 2024 at 08:33:55PM -0400, Sasha Levin wrote:
> On Tue, Apr 02, 2024 at 03:35:18PM +0200, David Sterba wrote:
> >On Fri, Mar 29, 2024 at 08:25:55AM -0400, Sasha Levin wrote:
> >> From: Filipe Manana <[email protected]>
> >>
> >> [ Upstream commit e383e158ed1b6abc2d2d3e6736d77a46393f80fa ]
> >>
> >> When logging an inode and we require to copy items from subvolume leaves
> >> to the log tree, we clone each subvolume leaf and than use that clone to
> >> copy items to the log tree. This is required to avoid possible deadlocks
> >> as stated in commit 796787c978ef ("btrfs: do not modify log tree while
> >> holding a leaf from fs tree locked").
> >>
> >> The cloning requires allocating an extent buffer (struct extent_buffer)
> >> and then allocating pages (folios) to attach to the extent buffer. This
> >> may be slow in case we are under memory pressure, and since we are doing
> >> the cloning while holding a read lock on a subvolume leaf, it means we
> >> can be blocking other operations on that leaf for significant periods of
> >> time, which can increase latency on operations like creating other files,
> >> renaming files, etc. Similarly because we're under a log transaction, we
> >> may also cause extra delay on other tasks doing an fsync, because syncing
> >> the log requires waiting for tasks that joined a log transaction to exit
> >> the transaction.
> >>
> >> So to improve this, for any inode logging operation that needs to copy
> >> items from a subvolume leaf ("full sync" or "copy everything" bit set
> >> in the inode), preallocate a dummy extent buffer before locking any
> >> extent buffer from the subvolume tree, and even before joining a log
> >> transaction, add it to the log context and then use it when we need to
> >> copy items from a subvolume leaf to the log tree. This avoids making
> >> other operations get extra latency when waiting to lock a subvolume
> >> leaf that is used during inode logging and we are under heavy memory
> >> pressure.
> >>
> >> The following test script with bonnie++ was used to test this:
> >>
> >> $ cat test.sh
> >> #!/bin/bash
> >>
> >> DEV=/dev/sdh
> >> MNT=/mnt/sdh
> >> MOUNT_OPTIONS="-o ssd"
> >>
> >> MEMTOTAL_BYTES=`free -b | grep Mem: | awk '{ print $2 }'`
> >> NR_DIRECTORIES=20
> >> NR_FILES=20480
> >> DATASET_SIZE=$((MEMTOTAL_BYTES * 2 / 1048576))
> >> DIRECTORY_SIZE=$((MEMTOTAL_BYTES * 2 / NR_FILES))
> >> NR_FILES=$((NR_FILES / 1024))
> >>
> >> echo "performance" | \
> >> tee /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor
> >>
> >> umount $DEV &> /dev/null
> >> mkfs.btrfs -f $MKFS_OPTIONS $DEV
> >> mount $MOUNT_OPTIONS $DEV $MNT
> >>
> >> bonnie++ -u root -d $MNT \
> >> -n $NR_FILES:$DIRECTORY_SIZE:$DIRECTORY_SIZE:$NR_DIRECTORIES \
> >> -r 0 -s $DATASET_SIZE -b
> >>
> >> umount $MNT
> >>
> >> The results of this test on a 8G VM running a non-debug kernel (Debian's
> >> default kernel config), were the following.
> >>
> >> Before this change:
> >>
> >> Version 2.00a ------Sequential Output------ --Sequential Input- --Random-
> >> -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
> >> Name:Size etc /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
> >> debian0 7501M 376k 99 1.4g 96 117m 14 1510k 99 2.5g 95 +++++ +++
> >> Latency 35068us 24976us 2944ms 30725us 71770us 26152us
> >> Version 2.00a ------Sequential Create------ --------Random Create--------
> >> debian0 -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
> >> files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
> >> 20:384100:384100/20 20480 32 20480 58 20480 48 20480 39 20480 56 20480 61
> >> Latency 411ms 11914us 119ms 617ms 10296us 110ms
> >>
> >> After this change:
> >>
> >> Version 2.00a ------Sequential Output------ --Sequential Input- --Random-
> >> -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
> >> Name:Size etc /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
> >> debian0 7501M 375k 99 1.4g 97 117m 14 1546k 99 2.3g 98 +++++ +++
> >> Latency 35975us 20945us 2144ms 10297us 2217us 6004us
> >> Version 2.00a ------Sequential Create------ --------Random Create--------
> >> debian0 -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
> >> files:max:min /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
> >> 20:384100:384100/20 20480 35 20480 58 20480 48 20480 40 20480 57 20480 59
> >> Latency 320ms 11237us 77779us 518ms 6470us 86389us
> >>
> >> Reviewed-by: Josef Bacik <[email protected]>
> >> Signed-off-by: Filipe Manana <[email protected]>
> >> Reviewed-by: David Sterba <[email protected]>
> >> Signed-off-by: David Sterba <[email protected]>
> >> Signed-off-by: Sasha Levin <[email protected]>
> >
> >This is a performance improvement, how does this qualify for stable? I
> >read only about notable perfromance fixes but this is not one.
>
> No objection to dropping it. Description of the commit states that it
> fixes blocking for "significant amount of time".
I see, that would make sense as keyword for stable backport, though it
applies under heavy memory pressure so not a regular workload where I'd
consider it for stable right away. A system under load will block on
many allocations, from that perspective the patch may not make any
difference.