2022-02-14 09:30:08

by Ritesh Harjani

[permalink] [raw]
Subject: [RFC 1/1] ext4: Improve fast_commit performance and scalability

Currently ext4_fc_commit_dentry_updates() is of quadratic time
complexity, which is causing performance bottlenecks with high
threads/file/dir count with fs_mark.

This patch makes commit dentry updates (and hence ext4_fc_commit()) path
to linear time complexity. Hence improves the performance of workloads
which does fsync on multiple threads/open files one-by-one.

Absolute numbers in avg file creates per sec (from fs_mark in 1K order)
=======================================================================
no. Order without-patch(K) with-patch(K) Diff(%)
1 1 16.90 17.51 +3.60
2 2,2 32.08 31.80 -0.87
3 3,3 53.97 55.01 +1.92
4 4,4 78.94 76.90 -2.58
5 5,5 95.82 95.37 -0.46
6 6,6 87.92 103.38 +17.58
7 6,10 0.73 126.13 +17178.08
8 6,14 2.33 143.19 +6045.49

workload type
==============
For e.g. 7th row order of 6,10 (2^6 == 64 && 2^10 == 1024)
echo /run/riteshh/mnt/{1..64} |sed -E 's/[[:space:]]+/ -d /g' \
| xargs -I {} bash -c "sudo fs_mark -L 100 -D 1024 -n 1024 -s0 -S5 -d {}"

Perf profile
(w/o patches)
=============================
87.15% [kernel] [k] ext4_fc_commit --> Heavy contention/bottleneck
1.98% [kernel] [k] perf_event_interrupt
0.96% [kernel] [k] power_pmu_enable
0.91% [kernel] [k] update_sd_lb_stats.constprop.0
0.67% [kernel] [k] ktime_get

Signed-off-by: Ritesh Harjani <[email protected]>
---
fs/ext4/ext4.h | 2 ++
fs/ext4/fast_commit.c | 64 +++++++++++++++++++++++++++++++------------
fs/ext4/fast_commit.h | 1 +
3 files changed, 50 insertions(+), 17 deletions(-)

diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index bcd3b9bf8069..25242648d8c9 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -1046,6 +1046,8 @@ struct ext4_inode_info {

/* Fast commit related info */

+ /* For tracking dentry create updates */
+ struct list_head i_fc_dilist;
struct list_head i_fc_list; /*
* inodes that need fast commit
* protected by sbi->s_fc_lock.
diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c
index 7964ee34e322..f2bee4cf5648 100644
--- a/fs/ext4/fast_commit.c
+++ b/fs/ext4/fast_commit.c
@@ -199,6 +199,7 @@ void ext4_fc_init_inode(struct inode *inode)
ext4_fc_reset_inode(inode);
ext4_clear_inode_state(inode, EXT4_STATE_FC_COMMITTING);
INIT_LIST_HEAD(&ei->i_fc_list);
+ INIT_LIST_HEAD(&ei->i_fc_dilist);
init_waitqueue_head(&ei->i_fc_wait);
atomic_set(&ei->i_fc_updates, 0);
}
@@ -279,6 +280,8 @@ void ext4_fc_stop_update(struct inode *inode)
void ext4_fc_del(struct inode *inode)
{
struct ext4_inode_info *ei = EXT4_I(inode);
+ struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
+ struct ext4_fc_dentry_update *fc_dentry, *fc_dentry_n;

if (!test_opt2(inode->i_sb, JOURNAL_FAST_COMMIT) ||
(EXT4_SB(inode->i_sb)->s_mount_state & EXT4_FC_REPLAY))
@@ -286,7 +289,7 @@ void ext4_fc_del(struct inode *inode)

restart:
spin_lock(&EXT4_SB(inode->i_sb)->s_fc_lock);
- if (list_empty(&ei->i_fc_list)) {
+ if (list_empty(&ei->i_fc_list) && list_empty(&ei->i_fc_dilist)) {
spin_unlock(&EXT4_SB(inode->i_sb)->s_fc_lock);
return;
}
@@ -295,7 +298,26 @@ void ext4_fc_del(struct inode *inode)
ext4_fc_wait_committing_inode(inode);
goto restart;
}
- list_del_init(&ei->i_fc_list);
+
+ if (!list_empty(&ei->i_fc_list))
+ list_del_init(&ei->i_fc_list);
+
+ /*
+ * Since this inode is getting removed, let's also remove all FC
+ * dentry create references, since it is not needed to log it anyways.
+ */
+ list_for_each_entry_safe(fc_dentry, fc_dentry_n, &ei->i_fc_dilist, fcd_dilist) {
+ WARN_ON(fc_dentry->fcd_op != EXT4_FC_TAG_CREAT);
+ list_del_init(&fc_dentry->fcd_list);
+ list_del_init(&fc_dentry->fcd_dilist);
+ spin_unlock(&sbi->s_fc_lock);
+
+ if (fc_dentry->fcd_name.name &&
+ fc_dentry->fcd_name.len > DNAME_INLINE_LEN)
+ kfree(fc_dentry->fcd_name.name);
+ kmem_cache_free(ext4_fc_dentry_cachep, fc_dentry);
+ return;
+ }
spin_unlock(&EXT4_SB(inode->i_sb)->s_fc_lock);
}

@@ -427,7 +449,7 @@ static int __track_dentry_update(struct inode *inode, void *arg, bool update)
node->fcd_name.name = node->fcd_iname;
}
node->fcd_name.len = dentry->d_name.len;
-
+ INIT_LIST_HEAD(&node->fcd_dilist);
spin_lock(&sbi->s_fc_lock);
if (sbi->s_journal->j_flags & JBD2_FULL_COMMIT_ONGOING ||
sbi->s_journal->j_flags & JBD2_FAST_COMMIT_ONGOING)
@@ -435,6 +457,18 @@ static int __track_dentry_update(struct inode *inode, void *arg, bool update)
&sbi->s_fc_dentry_q[FC_Q_STAGING]);
else
list_add_tail(&node->fcd_list, &sbi->s_fc_dentry_q[FC_Q_MAIN]);
+
+ /*
+ * This helps us keep a track of all fc_dentry updates which is part of
+ * this ext4 inode. So in case the inode is getting unlinked, before
+ * even we get a chance to fsync, we could remove all fc_dentry
+ * references while evicting the inode in ext4_fc_del().
+ * Also with this, we don't need to loop over all the inodes in
+ * sbi->s_fc_q to get the corresponding inode in
+ * ext4_fc_commit_dentry_updates().
+ */
+ if (dentry_update->op == EXT4_FC_TAG_CREAT)
+ list_add_tail(&node->fcd_dilist, &ei->i_fc_dilist);
spin_unlock(&sbi->s_fc_lock);
mutex_lock(&ei->i_fc_lock);

@@ -954,7 +988,7 @@ __releases(&sbi->s_fc_lock)
struct ext4_sb_info *sbi = EXT4_SB(sb);
struct ext4_fc_dentry_update *fc_dentry, *fc_dentry_n;
struct inode *inode;
- struct ext4_inode_info *ei, *ei_n;
+ struct ext4_inode_info *ei;
int ret;

if (list_empty(&sbi->s_fc_dentry_q[FC_Q_MAIN]))
@@ -970,21 +1004,16 @@ __releases(&sbi->s_fc_lock)
spin_lock(&sbi->s_fc_lock);
continue;
}
-
- inode = NULL;
- list_for_each_entry_safe(ei, ei_n, &sbi->s_fc_q[FC_Q_MAIN],
- i_fc_list) {
- if (ei->vfs_inode.i_ino == fc_dentry->fcd_ino) {
- inode = &ei->vfs_inode;
- break;
- }
- }
/*
- * If we don't find inode in our list, then it was deleted,
- * in which case, we don't need to record it's create tag.
+ * With fcd_dilist we need not loop in sbi->s_fc_q to get the
+ * corresponding inode pointer
*/
- if (!inode)
- continue;
+ WARN_ON(list_empty(&fc_dentry->fcd_dilist));
+ ei = list_entry(fc_dentry->fcd_dilist.next,
+ struct ext4_inode_info, i_fc_dilist);
+ inode = &ei->vfs_inode;
+ WARN_ON(inode->i_ino != fc_dentry->fcd_ino);
+
spin_unlock(&sbi->s_fc_lock);

/*
@@ -1228,6 +1257,7 @@ static void ext4_fc_cleanup(journal_t *journal, int full, tid_t tid)
struct ext4_fc_dentry_update,
fcd_list);
list_del_init(&fc_dentry->fcd_list);
+ list_del_init(&fc_dentry->fcd_dilist);
spin_unlock(&sbi->s_fc_lock);

if (fc_dentry->fcd_name.name &&
diff --git a/fs/ext4/fast_commit.h b/fs/ext4/fast_commit.h
index 083ad1cb705a..02afa52e8e41 100644
--- a/fs/ext4/fast_commit.h
+++ b/fs/ext4/fast_commit.h
@@ -109,6 +109,7 @@ struct ext4_fc_dentry_update {
struct qstr fcd_name; /* Dirent name */
unsigned char fcd_iname[DNAME_INLINE_LEN]; /* Dirent name string */
struct list_head fcd_list;
+ struct list_head fcd_dilist;
};

struct ext4_fc_stats {
--
2.31.1


2022-02-21 09:54:50

by Ritesh Harjani

[permalink] [raw]
Subject: Re: [RFC 1/1] ext4: Improve fast_commit performance and scalability

Hello Harshad,

Few points below.

On 22/02/17 09:27PM, Ritesh Harjani wrote:
> On 22/02/16 03:25PM, harshad shirwadkar wrote:
> > Thanks for the patch Ritesh. Some questions / comments inlined:
>
> Thanks a lot for reviewing this :)
>
> >
> > On Sun, 13 Feb 2022 at 19:57, Ritesh Harjani <[email protected]> wrote:
> > >
> > > Currently ext4_fc_commit_dentry_updates() is of quadratic time
> > > complexity, which is causing performance bottlenecks with high
> > > threads/file/dir count with fs_mark.
> > >
> > > This patch makes commit dentry updates (and hence ext4_fc_commit()) path
> > > to linear time complexity. Hence improves the performance of workloads
> > > which does fsync on multiple threads/open files one-by-one.
> > >
> > > Absolute numbers in avg file creates per sec (from fs_mark in 1K order)
> > > =======================================================================
> > > no. Order without-patch(K) with-patch(K) Diff(%)
> > > 1 1 16.90 17.51 +3.60
> > > 2 2,2 32.08 31.80 -0.87
> > > 3 3,3 53.97 55.01 +1.92
> > > 4 4,4 78.94 76.90 -2.58
> > > 5 5,5 95.82 95.37 -0.46
> > > 6 6,6 87.92 103.38 +17.58
> > > 7 6,10 0.73 126.13 +17178.08
> > > 8 6,14 2.33 143.19 +6045.49
> > >
> > > workload type
> > > ==============
> > > For e.g. 7th row order of 6,10 (2^6 == 64 && 2^10 == 1024)
> > > echo /run/riteshh/mnt/{1..64} |sed -E 's/[[:space:]]+/ -d /g' \
> > > | xargs -I {} bash -c "sudo fs_mark -L 100 -D 1024 -n 1024 -s0 -S5 -d {}"
> > >
> > > Perf profile
> > > (w/o patches)
> > > =============================
> > > 87.15% [kernel] [k] ext4_fc_commit --> Heavy contention/bottleneck
> > > 1.98% [kernel] [k] perf_event_interrupt
> > > 0.96% [kernel] [k] power_pmu_enable
> > > 0.91% [kernel] [k] update_sd_lb_stats.constprop.0
> > > 0.67% [kernel] [k] ktime_get
> > >
> > > Signed-off-by: Ritesh Harjani <[email protected]>
> > > ---
> > > fs/ext4/ext4.h | 2 ++
> > > fs/ext4/fast_commit.c | 64 +++++++++++++++++++++++++++++++------------
> > > fs/ext4/fast_commit.h | 1 +
> > > 3 files changed, 50 insertions(+), 17 deletions(-)
> > >
> > > diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
> > > index bcd3b9bf8069..25242648d8c9 100644
> > > --- a/fs/ext4/ext4.h
> > > +++ b/fs/ext4/ext4.h
> > > @@ -1046,6 +1046,8 @@ struct ext4_inode_info {
> > >
> > > /* Fast commit related info */
> > >
> > > + /* For tracking dentry create updates */
> > > + struct list_head i_fc_dilist;
> > The only case in which this list will have multiple entries if hard
> > links are created on this inode right? I think that's probably a very
>
> So I too had this thought on my mind later. But then I ended up coding the old way
> only.
>
> Ok, so it seems it is only when the first time an inode is created we
> will have a EXT4_FC_TAG_CREAT. When we are creating a hard link that's actually
> a EXT4_FC_TAG_LINK.
> So I think there shouldn't be any case where we have more than one fc_dentry for
> the same inode. Your thoughts?
>
>
> > rare scenario and we can just fallback to full commits. That might
> > simplify this patch a bit. Basically if you do that then fc_dentry
> > would directly store a pointer to the inode and the inode can store a
> > pointer to the "CREAT" fc_dentry object. That way we don't have to do
> > list traversals in fc_del and fc_commit. But barring a few fixes, what
> > you have here is fine too. So I'll leave it up to you to decide what
> > you want to do.
>
> Yes, you are right. If there is only a single fc_dentry object for any given
> inode, then we can store back pointers in each of those to point to their
> respective inode and fc_dentry objects.
>
> I will try and change this in next revision then.

I tried this approach. But it soon becomes messy the moment we think of adding
a inode pointer in fc_dentry object. Once that happens, we also want to
remove fcd_ino, since we have a inode and then we need to take care of a lot of
corner cases where inode will be lost since we don't hold any references.

whereas,

If we still use the struct list_head abstraction then we don't have to add a lot
of careful if else checking and things are much clear to understand.

What I would like to do now is send out a v2 addressing your review comments.
So I am not doing any list traversal now, but just popping the list first entry
and adding a warn_on if list does not becomes empty.

Could you then please take a look at v2 and suggest if you are fine with the
patch/design. We can take it from there then.

-ritesh

>
> > > struct list_head i_fc_list; /*
> > > * inodes that need fast commit
> > > * protected by sbi->s_fc_lock.
> > > diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c
> > > index 7964ee34e322..f2bee4cf5648 100644
> > > --- a/fs/ext4/fast_commit.c
> > > +++ b/fs/ext4/fast_commit.c
> > > @@ -199,6 +199,7 @@ void ext4_fc_init_inode(struct inode *inode)
> > > ext4_fc_reset_inode(inode);
> > > ext4_clear_inode_state(inode, EXT4_STATE_FC_COMMITTING);
> > > INIT_LIST_HEAD(&ei->i_fc_list);
> > > + INIT_LIST_HEAD(&ei->i_fc_dilist);
> > > init_waitqueue_head(&ei->i_fc_wait);
> > > atomic_set(&ei->i_fc_updates, 0);
> > > }
> > > @@ -279,6 +280,8 @@ void ext4_fc_stop_update(struct inode *inode)
> > > void ext4_fc_del(struct inode *inode)
> > > {
> > > struct ext4_inode_info *ei = EXT4_I(inode);
> > > + struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
> > > + struct ext4_fc_dentry_update *fc_dentry, *fc_dentry_n;
> > >
> > > if (!test_opt2(inode->i_sb, JOURNAL_FAST_COMMIT) ||
> > > (EXT4_SB(inode->i_sb)->s_mount_state & EXT4_FC_REPLAY))
> > > @@ -286,7 +289,7 @@ void ext4_fc_del(struct inode *inode)
> > >
> > > restart:
> > > spin_lock(&EXT4_SB(inode->i_sb)->s_fc_lock);
> > > - if (list_empty(&ei->i_fc_list)) {
> > > + if (list_empty(&ei->i_fc_list) && list_empty(&ei->i_fc_dilist)) {
> > > spin_unlock(&EXT4_SB(inode->i_sb)->s_fc_lock);
> > > return;
> > > }
> > > @@ -295,7 +298,26 @@ void ext4_fc_del(struct inode *inode)
> > > ext4_fc_wait_committing_inode(inode);
> > > goto restart;
> > > }
> > > - list_del_init(&ei->i_fc_list);
> > > +
> > > + if (!list_empty(&ei->i_fc_list))
> > > + list_del_init(&ei->i_fc_list);
> > > +
> > > + /*
> > > + * Since this inode is getting removed, let's also remove all FC
> > > + * dentry create references, since it is not needed to log it anyways.
> > > + */
> > > + list_for_each_entry_safe(fc_dentry, fc_dentry_n, &ei->i_fc_dilist, fcd_dilist) {
> > > + WARN_ON(fc_dentry->fcd_op != EXT4_FC_TAG_CREAT);
> > > + list_del_init(&fc_dentry->fcd_list);
> > > + list_del_init(&fc_dentry->fcd_dilist);
> > > + spin_unlock(&sbi->s_fc_lock);
> > > +
> > > + if (fc_dentry->fcd_name.name &&
> > > + fc_dentry->fcd_name.len > DNAME_INLINE_LEN)
> > > + kfree(fc_dentry->fcd_name.name);
> > > + kmem_cache_free(ext4_fc_dentry_cachep, fc_dentry);
> > > + return;
> > Shouldn't we continue and remove all nodes in ei->i_fc_dilist?
>
> Yes, I guess this survived, since we anyway have only one entry in the list
> always. But thanks for catching.
>
> > > + }
> > > spin_unlock(&EXT4_SB(inode->i_sb)->s_fc_lock);
> > > }
> > >
> > > @@ -427,7 +449,7 @@ static int __track_dentry_update(struct inode *inode, void *arg, bool update)
> > > node->fcd_name.name = node->fcd_iname;
> > > }
> > > node->fcd_name.len = dentry->d_name.len;
> > > -
> > > + INIT_LIST_HEAD(&node->fcd_dilist);
> > > spin_lock(&sbi->s_fc_lock);
> > > if (sbi->s_journal->j_flags & JBD2_FULL_COMMIT_ONGOING ||
> > > sbi->s_journal->j_flags & JBD2_FAST_COMMIT_ONGOING)
> > > @@ -435,6 +457,18 @@ static int __track_dentry_update(struct inode *inode, void *arg, bool update)
> > > &sbi->s_fc_dentry_q[FC_Q_STAGING]);
> > > else
> > > list_add_tail(&node->fcd_list, &sbi->s_fc_dentry_q[FC_Q_MAIN]);
> > > +
> > > + /*
> > > + * This helps us keep a track of all fc_dentry updates which is part of
> > > + * this ext4 inode. So in case the inode is getting unlinked, before
> > > + * even we get a chance to fsync, we could remove all fc_dentry
> > > + * references while evicting the inode in ext4_fc_del().
> > > + * Also with this, we don't need to loop over all the inodes in
> > > + * sbi->s_fc_q to get the corresponding inode in
> > > + * ext4_fc_commit_dentry_updates().
> > > + */
> > > + if (dentry_update->op == EXT4_FC_TAG_CREAT)
> > > + list_add_tail(&node->fcd_dilist, &ei->i_fc_dilist);
> > > spin_unlock(&sbi->s_fc_lock);
> > > mutex_lock(&ei->i_fc_lock);
> > >
> > > @@ -954,7 +988,7 @@ __releases(&sbi->s_fc_lock)
> > > struct ext4_sb_info *sbi = EXT4_SB(sb);
> > > struct ext4_fc_dentry_update *fc_dentry, *fc_dentry_n;
> > > struct inode *inode;
> > > - struct ext4_inode_info *ei, *ei_n;
> > > + struct ext4_inode_info *ei;
> > > int ret;
> > >
> > > if (list_empty(&sbi->s_fc_dentry_q[FC_Q_MAIN]))
> > > @@ -970,21 +1004,16 @@ __releases(&sbi->s_fc_lock)
> > > spin_lock(&sbi->s_fc_lock);
> > > continue;
> > > }
> > > -
> > > - inode = NULL;
> > > - list_for_each_entry_safe(ei, ei_n, &sbi->s_fc_q[FC_Q_MAIN],
> > > - i_fc_list) {
> > > - if (ei->vfs_inode.i_ino == fc_dentry->fcd_ino) {
> > > - inode = &ei->vfs_inode;
> > > - break;
> > > - }
> > > - }
> > > /*
> > > - * If we don't find inode in our list, then it was deleted,
> > > - * in which case, we don't need to record it's create tag.
> > > + * With fcd_dilist we need not loop in sbi->s_fc_q to get the
> > > + * corresponding inode pointer
> > > */
> > > - if (!inode)
> > > - continue;
> > > + WARN_ON(list_empty(&fc_dentry->fcd_dilist));
> > > + ei = list_entry(fc_dentry->fcd_dilist.next,
> > > + struct ext4_inode_info, i_fc_dilist);
> > I think we want "fc_dentry->fcd_ilist.prev" here right? We are
> > sequentially traversing all the nodes in the list from first to last.
> > Given that I think the inode is the prev of any node that you
> > encounter in the list.
>
> Not that this will be relevant in the next iteration. But doesn't matter right,
> next and prev both will have pointer to inode (since it is a circular doubly
> linked list)? And we are talking about fcd_dilist right?
>
> -ritesh
>
> >
> > - Harshad
> > > + inode = &ei->vfs_inode;
> > > + WARN_ON(inode->i_ino != fc_dentry->fcd_ino);
> > > +
> > > spin_unlock(&sbi->s_fc_lock);
> > >
> > > /*
> > > @@ -1228,6 +1257,7 @@ static void ext4_fc_cleanup(journal_t *journal, int full, tid_t tid)
> > > struct ext4_fc_dentry_update,
> > > fcd_list);
> > > list_del_init(&fc_dentry->fcd_list);
> > > + list_del_init(&fc_dentry->fcd_dilist);
> > > spin_unlock(&sbi->s_fc_lock);
> > >
> > > if (fc_dentry->fcd_name.name &&
> > > diff --git a/fs/ext4/fast_commit.h b/fs/ext4/fast_commit.h
> > > index 083ad1cb705a..02afa52e8e41 100644
> > > --- a/fs/ext4/fast_commit.h
> > > +++ b/fs/ext4/fast_commit.h
> > > @@ -109,6 +109,7 @@ struct ext4_fc_dentry_update {
> > > struct qstr fcd_name; /* Dirent name */
> > > unsigned char fcd_iname[DNAME_INLINE_LEN]; /* Dirent name string */
> > > struct list_head fcd_list;
> > > + struct list_head fcd_dilist;
> > > };
> > >
> > > struct ext4_fc_stats {
> > > --
> > > 2.31.1
> > >