[PATCH v2 0/8] Fix and cleanups to jbd2

[PATCH v2 0/8] Fix and cleanups to jbd2

[Kemeng Shi]

v1->v2:
-Move escape handle to futher improve
jbd2_journal_write_metadata_buffer.

This series contains a fix and some random cleanups to jbd2. More
details can be found in respective patches. Thanks!

Kemeng Shi (8):
  jbd2: correctly compare tids with tid_geq function in
    jbd2_fc_begin_commit
  jbd2: remove dead check in journal_alloc_journal_head
  jbd2: remove unused return value of jbd2_fc_release_bufs
  jbd2: remove unneeded kmap for jh_in->b_frozen_data in
    jbd2_journal_write_metadata_buffer
  jbd2: remove unneeded done_copy_out variable in
    jbd2_journal_write_metadata_buffer
  ext4: move escape handle to futher improve
    jbd2_journal_write_metadata_buffer
  jbd2: correct comment jbd2_mark_journal_empty
  jbd2: remove unneeded check of ret in jbd2_fc_get_buf

 fs/jbd2/journal.c    | 87 +++++++++++++++++++-------------------------
 include/linux/jbd2.h |  2 +-
 2 files changed, 39 insertions(+), 50 deletions(-)

-- 
2.30.0

[PATCH v2 1/8] jbd2: correctly compare tids with tid_geq function in jbd2_fc_begin_commit

[Kemeng Shi]

Use tid_geq to compare tids to work over sequence number wraps.

Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
Reviewed-by: Jan Kara <jack@suse.cz>
---
 fs/jbd2/journal.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index 1ebf2393bfb7..da5a56d700f1 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -710,7 +710,7 @@ int jbd2_fc_begin_commit(journal_t *journal, tid_t tid)
 		return -EINVAL;
 
 	write_lock(&journal->j_state_lock);
-	if (tid <= journal->j_commit_sequence) {
+	if (tid_geq(journal->j_commit_sequence, tid)) {
 		write_unlock(&journal->j_state_lock);
 		return -EALREADY;
 	}
-- 
2.30.0

[PATCH v2 2/8] jbd2: remove dead check in journal_alloc_journal_head

[Kemeng Shi]

We will alloc journal_head with __GFP_NOFAIL anyway, test for failure
is pointless.

Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
Reviewed-by: Jan Kara <jack@suse.cz>
---
 fs/jbd2/journal.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index da5a56d700f1..b5d02de1ffff 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -2866,8 +2866,7 @@ static struct journal_head *journal_alloc_journal_head(void)
 		ret = kmem_cache_zalloc(jbd2_journal_head_cache,
 				GFP_NOFS | __GFP_NOFAIL);
 	}
-	if (ret)
-		spin_lock_init(&ret->b_state_lock);
+	spin_lock_init(&ret->b_state_lock);
 	return ret;
 }
 
-- 
2.30.0

[PATCH v2 3/8] jbd2: remove unused return value of jbd2_fc_release_bufs

[Kemeng Shi]

Remove unused return value of jbd2_fc_release_bufs.

Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
Reviewed-by: Jan Kara <jack@suse.cz>
---
 fs/jbd2/journal.c    | 4 +---
 include/linux/jbd2.h | 2 +-
 2 files changed, 2 insertions(+), 4 deletions(-)

diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index b5d02de1ffff..312c7575b54f 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -903,7 +903,7 @@ int jbd2_fc_wait_bufs(journal_t *journal, int num_blks)
 }
 EXPORT_SYMBOL(jbd2_fc_wait_bufs);
 
-int jbd2_fc_release_bufs(journal_t *journal)
+void jbd2_fc_release_bufs(journal_t *journal)
 {
 	struct buffer_head *bh;
 	int i, j_fc_off;
@@ -917,8 +917,6 @@ int jbd2_fc_release_bufs(journal_t *journal)
 		put_bh(bh);
 		journal->j_fc_wbuf[i] = NULL;
 	}
-
-	return 0;
 }
 EXPORT_SYMBOL(jbd2_fc_release_bufs);
 
diff --git a/include/linux/jbd2.h b/include/linux/jbd2.h
index b900c642210c..735229e8ad17 100644
--- a/include/linux/jbd2.h
+++ b/include/linux/jbd2.h
@@ -1665,7 +1665,7 @@ int jbd2_fc_get_buf(journal_t *journal, struct buffer_head **bh_out);
 int jbd2_submit_inode_data(journal_t *journal, struct jbd2_inode *jinode);
 int jbd2_wait_inode_data(journal_t *journal, struct jbd2_inode *jinode);
 int jbd2_fc_wait_bufs(journal_t *journal, int num_blks);
-int jbd2_fc_release_bufs(journal_t *journal);
+void jbd2_fc_release_bufs(journal_t *journal);
 
 /*
  * is_journal_abort
-- 
2.30.0

[PATCH v2 4/8] jbd2: remove unneeded kmap for jh_in->b_frozen_data in jbd2_journal_write_metadata_buffer

[Kemeng Shi]

Remove kmap for page of b_frozen_data from jbd2_alloc() which always
provides an address from the direct kernel mapping.

Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
Reviewed-by: Jan Kara <jack@suse.cz>
---
 fs/jbd2/journal.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index 312c7575b54f..9c1ffb0dc740 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -352,12 +352,13 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
 		done_copy_out = 1;
 		new_folio = virt_to_folio(jh_in->b_frozen_data);
 		new_offset = offset_in_folio(new_folio, jh_in->b_frozen_data);
+		mapped_data = jh_in->b_frozen_data;
 	} else {
 		new_folio = bh_in->b_folio;
 		new_offset = offset_in_folio(new_folio, bh_in->b_data);
+		mapped_data = kmap_local_folio(new_folio, new_offset);
 	}
 
-	mapped_data = kmap_local_folio(new_folio, new_offset);
 	/*
 	 * Fire data frozen trigger if data already wasn't frozen.  Do this
 	 * before checking for escaping, as the trigger may modify the magic
@@ -373,7 +374,8 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
 	 */
 	if (*((__be32 *)mapped_data) == cpu_to_be32(JBD2_MAGIC_NUMBER))
 		do_escape = 1;
-	kunmap_local(mapped_data);
+	if (!jh_in->b_frozen_data)
+		kunmap_local(mapped_data);
 
 	/*
 	 * Do we need to do a data copy?
-- 
2.30.0

[PATCH v2 5/8] jbd2: remove unneeded done_copy_out variable in jbd2_journal_write_metadata_buffer

[Kemeng Shi]

It's more intuitive to use jh_in->b_frozen_data directly instead of
done_copy_out variable. Simply remove unneeded done_copy_out variable
and use b_frozen_data instead.

Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
---
 fs/jbd2/journal.c | 21 ++++++++-------------
 1 file changed, 8 insertions(+), 13 deletions(-)

diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index 9c1ffb0dc740..f17d05bc61df 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -318,7 +318,6 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
 				  struct buffer_head **bh_out,
 				  sector_t blocknr)
 {
-	int done_copy_out = 0;
 	int do_escape = 0;
 	char *mapped_data;
 	struct buffer_head *new_bh;
@@ -349,7 +348,6 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
 	 * we use that version of the data for the commit.
 	 */
 	if (jh_in->b_frozen_data) {
-		done_copy_out = 1;
 		new_folio = virt_to_folio(jh_in->b_frozen_data);
 		new_offset = offset_in_folio(new_folio, jh_in->b_frozen_data);
 		mapped_data = jh_in->b_frozen_data;
@@ -357,17 +355,15 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
 		new_folio = bh_in->b_folio;
 		new_offset = offset_in_folio(new_folio, bh_in->b_data);
 		mapped_data = kmap_local_folio(new_folio, new_offset);
-	}
-
-	/*
-	 * Fire data frozen trigger if data already wasn't frozen.  Do this
-	 * before checking for escaping, as the trigger may modify the magic
-	 * offset.  If a copy-out happens afterwards, it will have the correct
-	 * data in the buffer.
-	 */
-	if (!done_copy_out)
+		/*
+		 * Fire data frozen trigger if data already wasn't frozen. Do
+		 * this before checking for escaping, as the trigger may modify
+		 * the magic offset.  If a copy-out happens afterwards, it will
+		 * have the correct data in the buffer.
+		 */
 		jbd2_buffer_frozen_trigger(jh_in, mapped_data,
 					   jh_in->b_triggers);
+	}
 
 	/*
 	 * Check for escaping
@@ -380,7 +376,7 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
 	/*
 	 * Do we need to do a data copy?
 	 */
-	if (do_escape && !done_copy_out) {
+	if (do_escape && !jh_in->b_frozen_data) {
 		char *tmp;
 
 		spin_unlock(&jh_in->b_state_lock);
@@ -408,7 +404,6 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
 copy_done:
 		new_folio = virt_to_folio(jh_in->b_frozen_data);
 		new_offset = offset_in_folio(new_folio, jh_in->b_frozen_data);
-		done_copy_out = 1;
 	}
 
 	/*
-- 
2.30.0

[PATCH v2 6/8] ext4: move escape handle to futher improve jbd2_journal_write_metadata_buffer

[Kemeng Shi]

Move escape handle to futher improve code readability and remove some
repeat check.

Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
Suggested-by: Jan Kara <jack@suse.cz>
---
 fs/jbd2/journal.c | 49 +++++++++++++++++++++++------------------------
 1 file changed, 24 insertions(+), 25 deletions(-)

diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index f17d05bc61df..85273fb1accb 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -281,6 +281,16 @@ static void journal_kill_thread(journal_t *journal)
 	write_unlock(&journal->j_state_lock);
 }
 
+static inline bool jbd2_data_needs_escaping(char *data)
+{
+	return *((__be32 *)data) == cpu_to_be32(JBD2_MAGIC_NUMBER);
+}
+
+static inline void jbd2_data_do_escape(char *data)
+{
+	*((unsigned int *)data) = 0;
+}
+
 /*
  * jbd2_journal_write_metadata_buffer: write a metadata buffer to the journal.
  *
@@ -319,7 +329,6 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
 				  sector_t blocknr)
 {
 	int do_escape = 0;
-	char *mapped_data;
 	struct buffer_head *new_bh;
 	struct folio *new_folio;
 	unsigned int new_offset;
@@ -350,8 +359,13 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
 	if (jh_in->b_frozen_data) {
 		new_folio = virt_to_folio(jh_in->b_frozen_data);
 		new_offset = offset_in_folio(new_folio, jh_in->b_frozen_data);
-		mapped_data = jh_in->b_frozen_data;
+		do_escape = jbd2_data_needs_escaping(jh_in->b_frozen_data);
+		if (do_escape)
+			jbd2_data_do_escape(jh_in->b_frozen_data);
 	} else {
+		char *tmp;
+		char *mapped_data;
+
 		new_folio = bh_in->b_folio;
 		new_offset = offset_in_folio(new_folio, bh_in->b_data);
 		mapped_data = kmap_local_folio(new_folio, new_offset);
@@ -363,21 +377,13 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
 		 */
 		jbd2_buffer_frozen_trigger(jh_in, mapped_data,
 					   jh_in->b_triggers);
-	}
-
-	/*
-	 * Check for escaping
-	 */
-	if (*((__be32 *)mapped_data) == cpu_to_be32(JBD2_MAGIC_NUMBER))
-		do_escape = 1;
-	if (!jh_in->b_frozen_data)
+		do_escape = jbd2_data_needs_escaping(mapped_data);
 		kunmap_local(mapped_data);
-
-	/*
-	 * Do we need to do a data copy?
-	 */
-	if (do_escape && !jh_in->b_frozen_data) {
-		char *tmp;
+		/*
+		 * Do we need to do a data copy?
+		 */
+		if (!do_escape)
+			goto escape_done;
 
 		spin_unlock(&jh_in->b_state_lock);
 		tmp = jbd2_alloc(bh_in->b_size, GFP_NOFS);
@@ -404,17 +410,10 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
 copy_done:
 		new_folio = virt_to_folio(jh_in->b_frozen_data);
 		new_offset = offset_in_folio(new_folio, jh_in->b_frozen_data);
+		jbd2_data_do_escape(jh_in->b_frozen_data);
 	}
 
-	/*
-	 * Did we need to do an escaping?  Now we've done all the
-	 * copying, we can finally do so.
-	 * b_frozen_data is from jbd2_alloc() which always provides an
-	 * address from the direct kernels mapping.
-	 */
-	if (do_escape)
-		*((unsigned int *)jh_in->b_frozen_data) = 0;
-
+escape_done:
 	folio_set_bh(new_bh, new_folio, new_offset);
 	new_bh->b_size = bh_in->b_size;
 	new_bh->b_bdev = journal->j_dev;
-- 
2.30.0

[PATCH v2 7/8] jbd2: correct comment jbd2_mark_journal_empty

[Kemeng Shi]

After jbd2_mark_journal_empty, journal log is supposed to be empty.

Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
Reviewed-by: Jan Kara <jack@suse.cz>
---
 fs/jbd2/journal.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index 85273fb1accb..e89d777ded34 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -1938,7 +1938,7 @@ static void jbd2_mark_journal_empty(journal_t *journal, blk_opf_t write_flags)
 	if (had_fast_commit)
 		jbd2_set_feature_fast_commit(journal);
 
-	/* Log is no longer empty */
+	/* Log is empty */
 	write_lock(&journal->j_state_lock);
 	journal->j_flags |= JBD2_FLUSHED;
 	write_unlock(&journal->j_state_lock);
-- 
2.30.0

[PATCH v2 8/8] jbd2: remove unneeded check of ret in jbd2_fc_get_buf

[Kemeng Shi]

Simply return -EINVAL if j_fc_off is invalid to avoid repeated check of
ret.

Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
Reviewed-by: Jan Kara <jack@suse.cz>
---
 fs/jbd2/journal.c | 8 ++------
 1 file changed, 2 insertions(+), 6 deletions(-)

diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
index e89d777ded34..9d0f6735f8e6 100644
--- a/fs/jbd2/journal.c
+++ b/fs/jbd2/journal.c
@@ -841,12 +841,8 @@ int jbd2_fc_get_buf(journal_t *journal, struct buffer_head **bh_out)
 		fc_off = journal->j_fc_off;
 		blocknr = journal->j_fc_first + fc_off;
 		journal->j_fc_off++;
-	} else {
-		ret = -EINVAL;
-	}
-
-	if (ret)
-		return ret;
+	} else
+		return -EINVAL;
 
 	ret = jbd2_journal_bmap(journal, blocknr, &pblock);
 	if (ret)
-- 
2.30.0

Re: [PATCH v2 6/8] ext4: move escape handle to futher improve jbd2_journal_write_metadata_buffer

[Jan Kara]

On Wed 31-07-24 14:22:45, Kemeng Shi wrote:
> Move escape handle to futher improve code readability and remove some
> repeat check.
> 
> Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
> Suggested-by: Jan Kara <jack@suse.cz>

Looks good. Feel free to add:

Reviewed-by: Jan Kara <jack@suse.cz>

								Honza

> ---
>  fs/jbd2/journal.c | 49 +++++++++++++++++++++++------------------------
>  1 file changed, 24 insertions(+), 25 deletions(-)
> 
> diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
> index f17d05bc61df..85273fb1accb 100644
> --- a/fs/jbd2/journal.c
> +++ b/fs/jbd2/journal.c
> @@ -281,6 +281,16 @@ static void journal_kill_thread(journal_t *journal)
>  	write_unlock(&journal->j_state_lock);
>  }
>  
> +static inline bool jbd2_data_needs_escaping(char *data)
> +{
> +	return *((__be32 *)data) == cpu_to_be32(JBD2_MAGIC_NUMBER);
> +}
> +
> +static inline void jbd2_data_do_escape(char *data)
> +{
> +	*((unsigned int *)data) = 0;
> +}
> +
>  /*
>   * jbd2_journal_write_metadata_buffer: write a metadata buffer to the journal.
>   *
> @@ -319,7 +329,6 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
>  				  sector_t blocknr)
>  {
>  	int do_escape = 0;
> -	char *mapped_data;
>  	struct buffer_head *new_bh;
>  	struct folio *new_folio;
>  	unsigned int new_offset;
> @@ -350,8 +359,13 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
>  	if (jh_in->b_frozen_data) {
>  		new_folio = virt_to_folio(jh_in->b_frozen_data);
>  		new_offset = offset_in_folio(new_folio, jh_in->b_frozen_data);
> -		mapped_data = jh_in->b_frozen_data;
> +		do_escape = jbd2_data_needs_escaping(jh_in->b_frozen_data);
> +		if (do_escape)
> +			jbd2_data_do_escape(jh_in->b_frozen_data);
>  	} else {
> +		char *tmp;
> +		char *mapped_data;
> +
>  		new_folio = bh_in->b_folio;
>  		new_offset = offset_in_folio(new_folio, bh_in->b_data);
>  		mapped_data = kmap_local_folio(new_folio, new_offset);
> @@ -363,21 +377,13 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
>  		 */
>  		jbd2_buffer_frozen_trigger(jh_in, mapped_data,
>  					   jh_in->b_triggers);
> -	}
> -
> -	/*
> -	 * Check for escaping
> -	 */
> -	if (*((__be32 *)mapped_data) == cpu_to_be32(JBD2_MAGIC_NUMBER))
> -		do_escape = 1;
> -	if (!jh_in->b_frozen_data)
> +		do_escape = jbd2_data_needs_escaping(mapped_data);
>  		kunmap_local(mapped_data);
> -
> -	/*
> -	 * Do we need to do a data copy?
> -	 */
> -	if (do_escape && !jh_in->b_frozen_data) {
> -		char *tmp;
> +		/*
> +		 * Do we need to do a data copy?
> +		 */
> +		if (!do_escape)
> +			goto escape_done;
>  
>  		spin_unlock(&jh_in->b_state_lock);
>  		tmp = jbd2_alloc(bh_in->b_size, GFP_NOFS);
> @@ -404,17 +410,10 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
>  copy_done:
>  		new_folio = virt_to_folio(jh_in->b_frozen_data);
>  		new_offset = offset_in_folio(new_folio, jh_in->b_frozen_data);
> +		jbd2_data_do_escape(jh_in->b_frozen_data);
>  	}
>  
> -	/*
> -	 * Did we need to do an escaping?  Now we've done all the
> -	 * copying, we can finally do so.
> -	 * b_frozen_data is from jbd2_alloc() which always provides an
> -	 * address from the direct kernels mapping.
> -	 */
> -	if (do_escape)
> -		*((unsigned int *)jh_in->b_frozen_data) = 0;
> -
> +escape_done:
>  	folio_set_bh(new_bh, new_folio, new_offset);
>  	new_bh->b_size = bh_in->b_size;
>  	new_bh->b_bdev = journal->j_dev;
> -- 
> 2.30.0
> 
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR

Re: [PATCH v2 5/8] jbd2: remove unneeded done_copy_out variable in jbd2_journal_write_metadata_buffer

[Jan Kara]

On Wed 31-07-24 14:22:44, Kemeng Shi wrote:
> It's more intuitive to use jh_in->b_frozen_data directly instead of
> done_copy_out variable. Simply remove unneeded done_copy_out variable
> and use b_frozen_data instead.
> 
> Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>

Looks good. Feel free to add:

Reviewed-by: Jan Kara <jack@suse.cz>

								Honza

> ---
>  fs/jbd2/journal.c | 21 ++++++++-------------
>  1 file changed, 8 insertions(+), 13 deletions(-)
> 
> diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
> index 9c1ffb0dc740..f17d05bc61df 100644
> --- a/fs/jbd2/journal.c
> +++ b/fs/jbd2/journal.c
> @@ -318,7 +318,6 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
>  				  struct buffer_head **bh_out,
>  				  sector_t blocknr)
>  {
> -	int done_copy_out = 0;
>  	int do_escape = 0;
>  	char *mapped_data;
>  	struct buffer_head *new_bh;
> @@ -349,7 +348,6 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
>  	 * we use that version of the data for the commit.
>  	 */
>  	if (jh_in->b_frozen_data) {
> -		done_copy_out = 1;
>  		new_folio = virt_to_folio(jh_in->b_frozen_data);
>  		new_offset = offset_in_folio(new_folio, jh_in->b_frozen_data);
>  		mapped_data = jh_in->b_frozen_data;
> @@ -357,17 +355,15 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
>  		new_folio = bh_in->b_folio;
>  		new_offset = offset_in_folio(new_folio, bh_in->b_data);
>  		mapped_data = kmap_local_folio(new_folio, new_offset);
> -	}
> -
> -	/*
> -	 * Fire data frozen trigger if data already wasn't frozen.  Do this
> -	 * before checking for escaping, as the trigger may modify the magic
> -	 * offset.  If a copy-out happens afterwards, it will have the correct
> -	 * data in the buffer.
> -	 */
> -	if (!done_copy_out)
> +		/*
> +		 * Fire data frozen trigger if data already wasn't frozen. Do
> +		 * this before checking for escaping, as the trigger may modify
> +		 * the magic offset.  If a copy-out happens afterwards, it will
> +		 * have the correct data in the buffer.
> +		 */
>  		jbd2_buffer_frozen_trigger(jh_in, mapped_data,
>  					   jh_in->b_triggers);
> +	}
>  
>  	/*
>  	 * Check for escaping
> @@ -380,7 +376,7 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
>  	/*
>  	 * Do we need to do a data copy?
>  	 */
> -	if (do_escape && !done_copy_out) {
> +	if (do_escape && !jh_in->b_frozen_data) {
>  		char *tmp;
>  
>  		spin_unlock(&jh_in->b_state_lock);
> @@ -408,7 +404,6 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
>  copy_done:
>  		new_folio = virt_to_folio(jh_in->b_frozen_data);
>  		new_offset = offset_in_folio(new_folio, jh_in->b_frozen_data);
> -		done_copy_out = 1;
>  	}
>  
>  	/*
> -- 
> 2.30.0
> 
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR