diff mbox series

[1/2] ocfs2: use jbd2_inode dirty range scoping

Message ID 1562914972-97318-1-git-send-email-joseph.qi@linux.alibaba.com
State Superseded
Headers show
Series [1/2] ocfs2: use jbd2_inode dirty range scoping | expand

Commit Message

Joseph Qi July 12, 2019, 7:02 a.m. UTC
commit 6ba0e7dc64a5 ("jbd2: introduce jbd2_inode dirty range scoping")
allow us scoping each of the inode dirty ranges associated with a given
transaction, and ext4 already does this way.
Now let's also use the newly introduced jbd2_inode dirty range scoping
to prevent us from waiting forever when trying to complete a journal
transaction in ocfs2.

Signed-off-by: Joseph Qi <joseph.qi@linux.alibaba.com>
---
 fs/ocfs2/alloc.c   |  4 +++-
 fs/ocfs2/aops.c    |  6 ++++--
 fs/ocfs2/file.c    | 10 +++++++---
 fs/ocfs2/journal.h |  6 ++++--
 4 files changed, 18 insertions(+), 8 deletions(-)

Comments

Changwei Ge July 12, 2019, 9:45 a.m. UTC | #1
Hi Joseph,


Originally, ocfs2_jbd2_file_inode() is a wrapper of jbd2 routine 
jbd2_journal_file_inode() which has been renamed

by Jan Kara long ago. (41617e1a8dec9fe082ba5dec26bacb154eb55482)


So how about we change ocfs2_jbd2_file_inode to 
ocfs2_jbd2_inode_add_write() this time within your patch?


Thanks,

Changwei


On 2019/7/12 3:02 下午, Joseph Qi wrote:
> commit 6ba0e7dc64a5 ("jbd2: introduce jbd2_inode dirty range scoping")
> allow us scoping each of the inode dirty ranges associated with a given
> transaction, and ext4 already does this way.
> Now let's also use the newly introduced jbd2_inode dirty range scoping
> to prevent us from waiting forever when trying to complete a journal
> transaction in ocfs2.
>
> Signed-off-by: Joseph Qi <joseph.qi@linux.alibaba.com>
> ---
>   fs/ocfs2/alloc.c   |  4 +++-
>   fs/ocfs2/aops.c    |  6 ++++--
>   fs/ocfs2/file.c    | 10 +++++++---
>   fs/ocfs2/journal.h |  6 ++++--
>   4 files changed, 18 insertions(+), 8 deletions(-)
>
> diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c
> index d1348fc..2a58ca4 100644
> --- a/fs/ocfs2/alloc.c
> +++ b/fs/ocfs2/alloc.c
> @@ -6792,6 +6792,8 @@ void ocfs2_map_and_dirty_page(struct inode *inode, handle_t *handle,
>   			      struct page *page, int zero, u64 *phys)
>   {
>   	int ret, partial = 0;
> +	loff_t start_byte = ((loff_t)page->index << PAGE_SHIFT) + from;
> +	loff_t length = to - from;
>   
>   	ret = ocfs2_map_page_blocks(page, phys, inode, from, to, 0);
>   	if (ret)
> @@ -6811,7 +6813,7 @@ void ocfs2_map_and_dirty_page(struct inode *inode, handle_t *handle,
>   	if (ret < 0)
>   		mlog_errno(ret);
>   	else if (ocfs2_should_order_data(inode)) {
> -		ret = ocfs2_jbd2_file_inode(handle, inode);
> +		ret = ocfs2_jbd2_file_inode(handle, inode, start_byte, length);
>   		if (ret < 0)
>   			mlog_errno(ret);
>   	}
> diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
> index a4c905d..bbb508a 100644
> --- a/fs/ocfs2/aops.c
> +++ b/fs/ocfs2/aops.c
> @@ -942,7 +942,7 @@ static void ocfs2_write_failure(struct inode *inode,
>   
>   		if (tmppage && page_has_buffers(tmppage)) {
>   			if (ocfs2_should_order_data(inode))
> -				ocfs2_jbd2_file_inode(wc->w_handle, inode);
> +				ocfs2_jbd2_file_inode(wc->w_handle, inode, user_pos, user_len);
>   
>   			block_commit_write(tmppage, from, to);
>   		}
> @@ -2024,7 +2024,9 @@ int ocfs2_write_end_nolock(struct address_space *mapping,
>   
>   		if (page_has_buffers(tmppage)) {
>   			if (handle && ocfs2_should_order_data(inode))
> -				ocfs2_jbd2_file_inode(handle, inode);
> +				ocfs2_jbd2_file_inode(handle, inode,
> +						      ((loff_t)tmppage->index << PAGE_SHIFT) + from,
> +						      to - from);
>   			block_commit_write(tmppage, from, to);
>   		}
>   	}
> diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
> index 4435df3..43e6c28 100644
> --- a/fs/ocfs2/file.c
> +++ b/fs/ocfs2/file.c
> @@ -706,7 +706,9 @@ static int ocfs2_extend_allocation(struct inode *inode, u32 logical_start,
>    * Thus, we need to explicitly order the zeroed pages.
>    */
>   static handle_t *ocfs2_zero_start_ordered_transaction(struct inode *inode,
> -						struct buffer_head *di_bh)
> +						      struct buffer_head *di_bh,
> +						      loff_t start_bytes,
> +						      loff_t length)
>   {
>   	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
>   	handle_t *handle = NULL;
> @@ -722,7 +724,7 @@ static handle_t *ocfs2_zero_start_ordered_transaction(struct inode *inode,
>   		goto out;
>   	}
>   
> -	ret = ocfs2_jbd2_file_inode(handle, inode);
> +	ret = ocfs2_jbd2_file_inode(handle, inode, start_bytes, length);
>   	if (ret < 0) {
>   		mlog_errno(ret);
>   		goto out;
> @@ -761,7 +763,9 @@ static int ocfs2_write_zero_page(struct inode *inode, u64 abs_from,
>   	BUG_ON(abs_to > (((u64)index + 1) << PAGE_SHIFT));
>   	BUG_ON(abs_from & (inode->i_blkbits - 1));
>   
> -	handle = ocfs2_zero_start_ordered_transaction(inode, di_bh);
> +	handle = ocfs2_zero_start_ordered_transaction(inode, di_bh,
> +						      abs_from,
> +						      abs_to - abs_from);
>   	if (IS_ERR(handle)) {
>   		ret = PTR_ERR(handle);
>   		goto out;
> diff --git a/fs/ocfs2/journal.h b/fs/ocfs2/journal.h
> index c0fe6ed..932e6a8 100644
> --- a/fs/ocfs2/journal.h
> +++ b/fs/ocfs2/journal.h
> @@ -603,9 +603,11 @@ static inline int ocfs2_calc_tree_trunc_credits(struct super_block *sb,
>   	return credits;
>   }
>   
> -static inline int ocfs2_jbd2_file_inode(handle_t *handle, struct inode *inode)
> +static inline int ocfs2_jbd2_file_inode(handle_t *handle, struct inode *inode,
> +					loff_t start_byte, loff_t length)
>   {
> -	return jbd2_journal_inode_add_write(handle, &OCFS2_I(inode)->ip_jinode);
> +	return jbd2_journal_inode_ranged_write(handle, &OCFS2_I(inode)->ip_jinode,
> +					       start_byte, length);
>   }
>   
>   static inline int ocfs2_begin_ordered_truncate(struct inode *inode,
Joseph Qi July 12, 2019, 1:32 p.m. UTC | #2
Hi Changwei,

On 19/7/12 17:45, Changwei Ge wrote:
> Hi Joseph,
> 
> 
> Originally, ocfs2_jbd2_file_inode() is a wrapper of jbd2 routine jbd2_journal_file_inode() which has been renamed
> 
> by Jan Kara long ago. (41617e1a8dec9fe082ba5dec26bacb154eb55482)
> 
> 
> So how about we change ocfs2_jbd2_file_inode to ocfs2_jbd2_inode_add_write() this time within your patch?

Sure, I'll make this change along with other opinions in v2.

Thanks,
Joseph
> 
> 
> Thanks,
> 
> Changwei
> 
> 
> On 2019/7/12 3:02 下午, Joseph Qi wrote:
>> commit 6ba0e7dc64a5 ("jbd2: introduce jbd2_inode dirty range scoping")
>> allow us scoping each of the inode dirty ranges associated with a given
>> transaction, and ext4 already does this way.
>> Now let's also use the newly introduced jbd2_inode dirty range scoping
>> to prevent us from waiting forever when trying to complete a journal
>> transaction in ocfs2.
>>
>> Signed-off-by: Joseph Qi <joseph.qi@linux.alibaba.com>
>> ---
>>   fs/ocfs2/alloc.c   |  4 +++-
>>   fs/ocfs2/aops.c    |  6 ++++--
>>   fs/ocfs2/file.c    | 10 +++++++---
>>   fs/ocfs2/journal.h |  6 ++++--
>>   4 files changed, 18 insertions(+), 8 deletions(-)
>>
>> diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c
>> index d1348fc..2a58ca4 100644
>> --- a/fs/ocfs2/alloc.c
>> +++ b/fs/ocfs2/alloc.c
>> @@ -6792,6 +6792,8 @@ void ocfs2_map_and_dirty_page(struct inode *inode, handle_t *handle,
>>                     struct page *page, int zero, u64 *phys)
>>   {
>>       int ret, partial = 0;
>> +    loff_t start_byte = ((loff_t)page->index << PAGE_SHIFT) + from;
>> +    loff_t length = to - from;
>>         ret = ocfs2_map_page_blocks(page, phys, inode, from, to, 0);
>>       if (ret)
>> @@ -6811,7 +6813,7 @@ void ocfs2_map_and_dirty_page(struct inode *inode, handle_t *handle,
>>       if (ret < 0)
>>           mlog_errno(ret);
>>       else if (ocfs2_should_order_data(inode)) {
>> -        ret = ocfs2_jbd2_file_inode(handle, inode);
>> +        ret = ocfs2_jbd2_file_inode(handle, inode, start_byte, length);
>>           if (ret < 0)
>>               mlog_errno(ret);
>>       }
>> diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
>> index a4c905d..bbb508a 100644
>> --- a/fs/ocfs2/aops.c
>> +++ b/fs/ocfs2/aops.c
>> @@ -942,7 +942,7 @@ static void ocfs2_write_failure(struct inode *inode,
>>             if (tmppage && page_has_buffers(tmppage)) {
>>               if (ocfs2_should_order_data(inode))
>> -                ocfs2_jbd2_file_inode(wc->w_handle, inode);
>> +                ocfs2_jbd2_file_inode(wc->w_handle, inode, user_pos, user_len);
>>                 block_commit_write(tmppage, from, to);
>>           }
>> @@ -2024,7 +2024,9 @@ int ocfs2_write_end_nolock(struct address_space *mapping,
>>             if (page_has_buffers(tmppage)) {
>>               if (handle && ocfs2_should_order_data(inode))
>> -                ocfs2_jbd2_file_inode(handle, inode);
>> +                ocfs2_jbd2_file_inode(handle, inode,
>> +                              ((loff_t)tmppage->index << PAGE_SHIFT) + from,
>> +                              to - from);
>>               block_commit_write(tmppage, from, to);
>>           }
>>       }
>> diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
>> index 4435df3..43e6c28 100644
>> --- a/fs/ocfs2/file.c
>> +++ b/fs/ocfs2/file.c
>> @@ -706,7 +706,9 @@ static int ocfs2_extend_allocation(struct inode *inode, u32 logical_start,
>>    * Thus, we need to explicitly order the zeroed pages.
>>    */
>>   static handle_t *ocfs2_zero_start_ordered_transaction(struct inode *inode,
>> -                        struct buffer_head *di_bh)
>> +                              struct buffer_head *di_bh,
>> +                              loff_t start_bytes,
>> +                              loff_t length)
>>   {
>>       struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
>>       handle_t *handle = NULL;
>> @@ -722,7 +724,7 @@ static handle_t *ocfs2_zero_start_ordered_transaction(struct inode *inode,
>>           goto out;
>>       }
>>   -    ret = ocfs2_jbd2_file_inode(handle, inode);
>> +    ret = ocfs2_jbd2_file_inode(handle, inode, start_bytes, length);
>>       if (ret < 0) {
>>           mlog_errno(ret);
>>           goto out;
>> @@ -761,7 +763,9 @@ static int ocfs2_write_zero_page(struct inode *inode, u64 abs_from,
>>       BUG_ON(abs_to > (((u64)index + 1) << PAGE_SHIFT));
>>       BUG_ON(abs_from & (inode->i_blkbits - 1));
>>   -    handle = ocfs2_zero_start_ordered_transaction(inode, di_bh);
>> +    handle = ocfs2_zero_start_ordered_transaction(inode, di_bh,
>> +                              abs_from,
>> +                              abs_to - abs_from);
>>       if (IS_ERR(handle)) {
>>           ret = PTR_ERR(handle);
>>           goto out;
>> diff --git a/fs/ocfs2/journal.h b/fs/ocfs2/journal.h
>> index c0fe6ed..932e6a8 100644
>> --- a/fs/ocfs2/journal.h
>> +++ b/fs/ocfs2/journal.h
>> @@ -603,9 +603,11 @@ static inline int ocfs2_calc_tree_trunc_credits(struct super_block *sb,
>>       return credits;
>>   }
>>   -static inline int ocfs2_jbd2_file_inode(handle_t *handle, struct inode *inode)
>> +static inline int ocfs2_jbd2_file_inode(handle_t *handle, struct inode *inode,
>> +                    loff_t start_byte, loff_t length)
>>   {
>> -    return jbd2_journal_inode_add_write(handle, &OCFS2_I(inode)->ip_jinode);
>> +    return jbd2_journal_inode_ranged_write(handle, &OCFS2_I(inode)->ip_jinode,
>> +                           start_byte, length);
>>   }
>>     static inline int ocfs2_begin_ordered_truncate(struct inode *inode,
Ross Zwisler July 12, 2019, 9:39 p.m. UTC | #3
On Fri, Jul 12, 2019 at 03:02:51PM +0800, Joseph Qi wrote:
> commit 6ba0e7dc64a5 ("jbd2: introduce jbd2_inode dirty range scoping")
> allow us scoping each of the inode dirty ranges associated with a given
> transaction, and ext4 already does this way.
> Now let's also use the newly introduced jbd2_inode dirty range scoping
> to prevent us from waiting forever when trying to complete a journal
> transaction in ocfs2.
> 
> Signed-off-by: Joseph Qi <joseph.qi@linux.alibaba.com>
 
> diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
> index a4c905d..bbb508a 100644
> --- a/fs/ocfs2/aops.c
> +++ b/fs/ocfs2/aops.c
> @@ -942,7 +942,7 @@ static void ocfs2_write_failure(struct inode *inode,
>  
>  		if (tmppage && page_has_buffers(tmppage)) {
>  			if (ocfs2_should_order_data(inode))
> -				ocfs2_jbd2_file_inode(wc->w_handle, inode);
> +				ocfs2_jbd2_file_inode(wc->w_handle, inode, user_pos, user_len);

Line longer than 80 characters, should be wrapped.

> @@ -2024,7 +2024,9 @@ int ocfs2_write_end_nolock(struct address_space *mapping,
>  
>  		if (page_has_buffers(tmppage)) {
>  			if (handle && ocfs2_should_order_data(inode))
> -				ocfs2_jbd2_file_inode(handle, inode);
> +				ocfs2_jbd2_file_inode(handle, inode,
> +						      ((loff_t)tmppage->index << PAGE_SHIFT) + from,

Line longer than 80 characters, should be wrapped.

> diff --git a/fs/ocfs2/journal.h b/fs/ocfs2/journal.h
> index c0fe6ed..932e6a8 100644
> --- a/fs/ocfs2/journal.h
> +++ b/fs/ocfs2/journal.h
> @@ -603,9 +603,11 @@ static inline int ocfs2_calc_tree_trunc_credits(struct super_block *sb,
>  	return credits;
>  }
>  
> -static inline int ocfs2_jbd2_file_inode(handle_t *handle, struct inode *inode)
> +static inline int ocfs2_jbd2_file_inode(handle_t *handle, struct inode *inode,
> +					loff_t start_byte, loff_t length)
>  {
> -	return jbd2_journal_inode_add_write(handle, &OCFS2_I(inode)->ip_jinode);
> +	return jbd2_journal_inode_ranged_write(handle, &OCFS2_I(inode)->ip_jinode,
> +					       start_byte, length);
>  }

Perhaps ocfs2_jbd2_ranged_write() would be more informative if you're renaming
this function?

Aside from these nits, this all looked good to me.  You can add:
Reviewed-by: Ross Zwisler <zwisler@google.com>
diff mbox series

Patch

diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c
index d1348fc..2a58ca4 100644
--- a/fs/ocfs2/alloc.c
+++ b/fs/ocfs2/alloc.c
@@ -6792,6 +6792,8 @@  void ocfs2_map_and_dirty_page(struct inode *inode, handle_t *handle,
 			      struct page *page, int zero, u64 *phys)
 {
 	int ret, partial = 0;
+	loff_t start_byte = ((loff_t)page->index << PAGE_SHIFT) + from;
+	loff_t length = to - from;
 
 	ret = ocfs2_map_page_blocks(page, phys, inode, from, to, 0);
 	if (ret)
@@ -6811,7 +6813,7 @@  void ocfs2_map_and_dirty_page(struct inode *inode, handle_t *handle,
 	if (ret < 0)
 		mlog_errno(ret);
 	else if (ocfs2_should_order_data(inode)) {
-		ret = ocfs2_jbd2_file_inode(handle, inode);
+		ret = ocfs2_jbd2_file_inode(handle, inode, start_byte, length);
 		if (ret < 0)
 			mlog_errno(ret);
 	}
diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
index a4c905d..bbb508a 100644
--- a/fs/ocfs2/aops.c
+++ b/fs/ocfs2/aops.c
@@ -942,7 +942,7 @@  static void ocfs2_write_failure(struct inode *inode,
 
 		if (tmppage && page_has_buffers(tmppage)) {
 			if (ocfs2_should_order_data(inode))
-				ocfs2_jbd2_file_inode(wc->w_handle, inode);
+				ocfs2_jbd2_file_inode(wc->w_handle, inode, user_pos, user_len);
 
 			block_commit_write(tmppage, from, to);
 		}
@@ -2024,7 +2024,9 @@  int ocfs2_write_end_nolock(struct address_space *mapping,
 
 		if (page_has_buffers(tmppage)) {
 			if (handle && ocfs2_should_order_data(inode))
-				ocfs2_jbd2_file_inode(handle, inode);
+				ocfs2_jbd2_file_inode(handle, inode,
+						      ((loff_t)tmppage->index << PAGE_SHIFT) + from,
+						      to - from);
 			block_commit_write(tmppage, from, to);
 		}
 	}
diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
index 4435df3..43e6c28 100644
--- a/fs/ocfs2/file.c
+++ b/fs/ocfs2/file.c
@@ -706,7 +706,9 @@  static int ocfs2_extend_allocation(struct inode *inode, u32 logical_start,
  * Thus, we need to explicitly order the zeroed pages.
  */
 static handle_t *ocfs2_zero_start_ordered_transaction(struct inode *inode,
-						struct buffer_head *di_bh)
+						      struct buffer_head *di_bh,
+						      loff_t start_bytes,
+						      loff_t length)
 {
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 	handle_t *handle = NULL;
@@ -722,7 +724,7 @@  static handle_t *ocfs2_zero_start_ordered_transaction(struct inode *inode,
 		goto out;
 	}
 
-	ret = ocfs2_jbd2_file_inode(handle, inode);
+	ret = ocfs2_jbd2_file_inode(handle, inode, start_bytes, length);
 	if (ret < 0) {
 		mlog_errno(ret);
 		goto out;
@@ -761,7 +763,9 @@  static int ocfs2_write_zero_page(struct inode *inode, u64 abs_from,
 	BUG_ON(abs_to > (((u64)index + 1) << PAGE_SHIFT));
 	BUG_ON(abs_from & (inode->i_blkbits - 1));
 
-	handle = ocfs2_zero_start_ordered_transaction(inode, di_bh);
+	handle = ocfs2_zero_start_ordered_transaction(inode, di_bh,
+						      abs_from,
+						      abs_to - abs_from);
 	if (IS_ERR(handle)) {
 		ret = PTR_ERR(handle);
 		goto out;
diff --git a/fs/ocfs2/journal.h b/fs/ocfs2/journal.h
index c0fe6ed..932e6a8 100644
--- a/fs/ocfs2/journal.h
+++ b/fs/ocfs2/journal.h
@@ -603,9 +603,11 @@  static inline int ocfs2_calc_tree_trunc_credits(struct super_block *sb,
 	return credits;
 }
 
-static inline int ocfs2_jbd2_file_inode(handle_t *handle, struct inode *inode)
+static inline int ocfs2_jbd2_file_inode(handle_t *handle, struct inode *inode,
+					loff_t start_byte, loff_t length)
 {
-	return jbd2_journal_inode_add_write(handle, &OCFS2_I(inode)->ip_jinode);
+	return jbd2_journal_inode_ranged_write(handle, &OCFS2_I(inode)->ip_jinode,
+					       start_byte, length);
 }
 
 static inline int ocfs2_begin_ordered_truncate(struct inode *inode,