diff mbox

[1/2] ocfs2: make metadata estimation accurate and clear

Message ID 63ADC13FD55D6546B7DECE290D39E373F290CEEE@H3CMLB12-EX.srv.huawei-3com.com (mailing list archive)
State New, archived
Headers show

Commit Message

Changwei Ge Jan. 8, 2018, 1:36 a.m. UTC
Current code assume that ::w_unwritten_list always has only one item on.
This is not right and hard to get understood.
So improve how to count unwritten item.

Reported-by: John Lightsey <john@nixnuts.net>
Signed-off-by: Changwei Ge <ge.changwei@h3c.com>
---
  fs/ocfs2/aops.c | 4 +++-
  1 file changed, 3 insertions(+), 1 deletion(-)

Comments

Andrew Morton Jan. 9, 2018, 1:23 a.m. UTC | #1
On Mon, 8 Jan 2018 01:36:05 +0000 Changwei Ge <ge.changwei@h3c.com> wrote:

> Current code assume that ::w_unwritten_list always has only one item on.
> This is not right and hard to get understood.
> So improve how to count unwritten item.

I get a large number of patch rejects applying this (and patch 2) to
current mainline.  What kernel version are these against?
Changwei Ge Jan. 9, 2018, 1:49 a.m. UTC | #2
Hi Andrew

On 2018/1/9 9:24, Andrew Morton wrote:
> On Mon, 8 Jan 2018 01:36:05 +0000 Changwei Ge <ge.changwei@h3c.com> wrote:
> 
>> Current code assume that ::w_unwritten_list always has only one item on.
>> This is not right and hard to get understood.
>> So improve how to count unwritten item.
> 
> I get a large number of patch rejects applying this (and patch 2) to
> current mainline.  What kernel version are these against?
> 
> 

These two patches are against 4.15-rc4 (1291a0d5049dbc06baaaf66a9ff3f53db493b19b).
If you have any preferable kernel version, please let me know.

Thanks,
Changwei
diff mbox

Patch

diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
index d151632..256986a 100644
--- a/fs/ocfs2/aops.c
+++ b/fs/ocfs2/aops.c
@@ -797,6 +797,7 @@  struct ocfs2_write_ctxt {
  	struct ocfs2_cached_dealloc_ctxt w_dealloc;
  
  	struct list_head		w_unwritten_list;
+	unsigned int			w_unwritten_count;
  };
  
  void ocfs2_unlock_and_free_pages(struct page **pages, int num_pages)
@@ -1386,6 +1387,7 @@  static int ocfs2_unwritten_check(struct inode *inode,
  	desc->c_clear_unwritten = 0;
  	list_add_tail(&new->ue_ip_node, &oi->ip_unwritten_list);
  	list_add_tail(&new->ue_node, &wc->w_unwritten_list);
+	wc->w_unwritten_count++;
  	new = NULL;
  unlock:
  	spin_unlock(&oi->ip_lock);
@@ -2256,7 +2258,7 @@  static int ocfs2_dio_wr_get_block(struct inode *inode, sector_t iblock,
  		ue->ue_phys = desc->c_phys;
  
  		list_splice_tail_init(&wc->w_unwritten_list, &dwc->dw_zero_list);
-		dwc->dw_zero_count++;
+		dwc->dw_zero_count += wc->w_unwritten_count;
  	}
  
  	ret = ocfs2_write_end_nolock(inode->i_mapping, pos, len, len, wc);