Received-SPF: pass (google.com: domain of linux-ext4+bounces-2590-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.199.223 as permitted sender) client-ip=147.75.199.223;
From: Harshad Shirwadkar <harshadshirwadkar@gmail.com>
To: linux-ext4@vger.kernel.org
Cc: tytso@mit.edu,
	saukad@google.com,
	harshads@google.com,
	Harshad Shirwadkar <harshadshirwadkar@gmail.com>
Subject: [PATCH 02/10] ext4: for committing inode, make ext4_fc_track_inode wait
Date: Mon, 20 May 2024 05:51:45 +0000
Message-ID: <20240520055153.136091-3-harshadshirwadkar@gmail.com>
In-Reply-To: <20240520055153.136091-1-harshadshirwadkar@gmail.com>
References: <20240520055153.136091-1-harshadshirwadkar@gmail.com>
Precedence: bulk
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit

If the inode that's being requested to track using ext4_fc_track_inode
is being committed, then wait until the inode finishes the
commit. Also, add calls to ext4_fc_track_inode at the right places.

With this patch, now calling ext4_reserve_inode_write() results in
inode being tracked for next fast commit. A subtle lock ordering
requirement with i_data_sem (which is documented in the code) requires
that ext4_fc_track_inode() be called before grabbing i_data_sem. So,
this patch also adds explicit ext4_fc_track_inode() calls in places
where i_data_sem grabbed.

Signed-off-by: Harshad Shirwadkar <harshadshirwadkar@gmail.com>
---
 fs/ext4/fast_commit.c | 34 ++++++++++++++++++++++++++++++++++
 fs/ext4/inline.c      |  3 +++
 fs/ext4/inode.c       |  4 ++++
 3 files changed, 41 insertions(+)

diff --git a/fs/ext4/fast_commit.c b/fs/ext4/fast_commit.c
index a1aadebfcd66..fa93ce399440 100644
--- a/fs/ext4/fast_commit.c
+++ b/fs/ext4/fast_commit.c
@@ -581,6 +581,8 @@ static int __track_inode(struct inode *inode, void *arg, bool update)
 
 void ext4_fc_track_inode(handle_t *handle, struct inode *inode)
 {
+	struct ext4_inode_info *ei = EXT4_I(inode);
+	wait_queue_head_t *wq;
 	int ret;
 
 	if (S_ISDIR(inode->i_mode))
@@ -598,6 +600,38 @@ void ext4_fc_track_inode(handle_t *handle, struct inode *inode)
 	if (ext4_test_mount_flag(inode->i_sb, EXT4_MF_FC_INELIGIBLE))
 		return;
 
+	if (!test_opt2(inode->i_sb, JOURNAL_FAST_COMMIT) ||
+	    (EXT4_SB(inode->i_sb)->s_mount_state & EXT4_FC_REPLAY) ||
+		ext4_test_mount_flag(inode->i_sb, EXT4_MF_FC_INELIGIBLE) ||
+		!list_empty(&ei->i_fc_list))
+		return;
+
+	/*
+	 * If we come here, we may sleep while waiting for the inode to
+	 * commit. We shouldn't be holding i_data_sem in write mode when we go
+	 * to sleep since the commit path needs to grab the lock while
+	 * committing the inode.
+	 */
+	WARN_ON(lockdep_is_held_type(&ei->i_data_sem, 1));
+
+	while (ext4_test_inode_state(inode, EXT4_STATE_FC_COMMITTING)) {
+#if (BITS_PER_LONG < 64)
+		DEFINE_WAIT_BIT(wait, &ei->i_state_flags,
+				EXT4_STATE_FC_COMMITTING);
+		wq = bit_waitqueue(&ei->i_state_flags,
+				   EXT4_STATE_FC_COMMITTING);
+#else
+		DEFINE_WAIT_BIT(wait, &ei->i_flags,
+				EXT4_STATE_FC_COMMITTING);
+		wq = bit_waitqueue(&ei->i_flags,
+				   EXT4_STATE_FC_COMMITTING);
+#endif
+		prepare_to_wait(wq, &wait.wq_entry, TASK_UNINTERRUPTIBLE);
+		if (ext4_test_inode_state(inode, EXT4_STATE_FC_COMMITTING))
+			schedule();
+		finish_wait(wq, &wait.wq_entry);
+	}
+
 	ret = ext4_fc_track_template(handle, inode, __track_inode, NULL, 1);
 	trace_ext4_fc_track_inode(handle, inode, ret);
 }
diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c
index d5bd1e3a5d36..bd5778e680c0 100644
--- a/fs/ext4/inline.c
+++ b/fs/ext4/inline.c
@@ -596,6 +596,7 @@ static int ext4_convert_inline_data_to_extent(struct address_space *mapping,
 			goto out;
 	}
 
+	ext4_fc_track_inode(handle, inode);
 	ret = ext4_destroy_inline_data_nolock(handle, inode);
 	if (ret)
 		goto out;
@@ -696,6 +697,7 @@ int ext4_try_to_write_inline_data(struct address_space *mapping,
 		goto convert;
 	}
 
+	ext4_fc_track_inode(handle, inode);
 	ret = ext4_journal_get_write_access(handle, inode->i_sb, iloc.bh,
 					    EXT4_JTR_NONE);
 	if (ret)
@@ -948,6 +950,7 @@ int ext4_da_write_inline_data_begin(struct address_space *mapping,
 		if (ret < 0)
 			goto out_release_page;
 	}
+	ext4_fc_track_inode(handle, inode);
 	ret = ext4_journal_get_write_access(handle, inode->i_sb, iloc.bh,
 					    EXT4_JTR_NONE);
 	if (ret)
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 4bae9ccf5fe0..aa6440992a55 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -607,6 +607,7 @@ int ext4_map_blocks(handle_t *handle, struct inode *inode,
 	 */
 	map->m_flags &= ~EXT4_MAP_FLAGS;
 
+	ext4_fc_track_inode(handle, inode);
 	/*
 	 * New blocks allocate and/or writing to unwritten extent
 	 * will possibly result in updating i_data, so we take
@@ -3978,6 +3979,7 @@ int ext4_punch_hole(struct file *file, loff_t offset, loff_t length)
 	if (stop_block > first_block) {
 		ext4_lblk_t hole_len = stop_block - first_block;
 
+		ext4_fc_track_inode(handle, inode);
 		down_write(&EXT4_I(inode)->i_data_sem);
 		ext4_discard_preallocations(inode);
 
@@ -4131,6 +4133,7 @@ int ext4_truncate(struct inode *inode)
 	if (err)
 		goto out_stop;
 
+	ext4_fc_track_inode(handle, inode);
 	down_write(&EXT4_I(inode)->i_data_sem);
 
 	ext4_discard_preallocations(inode);
@@ -5727,6 +5730,7 @@ ext4_reserve_inode_write(handle_t *handle, struct inode *inode,
 			brelse(iloc->bh);
 			iloc->bh = NULL;
 		}
+		ext4_fc_track_inode(handle, inode);
 	}
 	ext4_std_error(inode->i_sb, err);
 	return err;
-- 
2.45.0.rc1.225.g2a3ae87e7f-goog