f2fs: skip ->writepages for {mete,node}_inode during recovery

PATCH In review Chao Yu 2 participants 5 messages Jun 29, 2017 – Jul 5, 2017 F2FS file system

lore

Share & tools

lore

Chao Yu

author maintainer

2017-06-29 15:22 UTC #0 lore raw

PATCH 1/2 f2fs: skip ->writepages for {mete,node}_inode during recovery

From: Chao Yu <[email protected]>

Skip ->writepages in prior to ->writepage for {meta,node}_inode during recovery, hence unneeded loop in ->writepages can be avoided.

Moreover, check SBI_POR_DOING earlier while writebacking pages.

Signed-off-by Chao Yu <[email protected]>

---
 fs/f2fs/checkpoint.c |  3 +++
 fs/f2fs/data.c       | 13 +++++++------
 fs/f2fs/node.c       |  3 +++
 3 files changed, 13 insertions(+), 6 deletions(-)

fs/f2fs/checkpoint.c +3 -0

index 12559a4b6c24..954917d582f8 100644
@@ -269,6 +269,9 @@ static int f2fs_write_meta_pages(struct address_space *mapping,
 	struct f2fs_sb_info *sbi = F2FS_M_SB(mapping);
 	long diff, written;
 
+	if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
+		goto skip_write;
+
 	/* collect a number of dirty meta pages and write together */
 	if (wbc->for_kupdate ||
 		get_pages(sbi, F2FS_DIRTY_META) < nr_pages_to_skip(sbi, META))

fs/f2fs/data.c +7 -6

index 9141bd19a902..d58b81213a86 100644
@@ -1493,6 +1493,9 @@ static int __write_data_page(struct page *page, bool *submitted,
 
 	trace_f2fs_writepage(page, DATA);
 
+	if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
+		goto redirty_out;
+
 	if (page->index < end_index)
 		goto write;

@@ -1506,8 +1509,6 @@ static int __write_data_page(struct page *page, bool *submitted,
 
 	zero_user_segment(page, offset, PAGE_SIZE);
 write:
-	if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
-		goto redirty_out;
 	if (f2fs_is_drop_cache(inode))
 		goto out;
 	/* we should not write 0'th page having journal header */

@@ -1755,6 +1756,10 @@ static int f2fs_write_data_pages(struct address_space *mapping,
 	if (!get_dirty_pages(inode) && wbc->sync_mode == WB_SYNC_NONE)
 		return 0;
 
+	/* during POR, we don't need to trigger writepage at all. */
+	if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
+		goto skip_write;
+
 	if (S_ISDIR(inode->i_mode) && wbc->sync_mode == WB_SYNC_NONE &&
 			get_dirty_pages(inode) < nr_pages_to_skip(sbi, DATA) &&
 			available_free_memory(sbi, DIRTY_DENTS))

@@ -1764,10 +1769,6 @@ static int f2fs_write_data_pages(struct address_space *mapping,
 	if (is_inode_flag_set(inode, FI_DO_DEFRAG))
 		goto skip_write;
 
-	/* during POR, we don't need to trigger writepage at all. */
-	if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
-		goto skip_write;
-
 	trace_f2fs_writepages(mapping->host, wbc, DATA);
 
 	/* to avoid spliting IOs due to mixed WB_SYNC_ALL and WB_SYNC_NONE */

fs/f2fs/node.c +3 -0

index f6f46be139f4..fd57ffd88508 100644
@@ -1687,6 +1687,9 @@ static int f2fs_write_node_pages(struct address_space *mapping,
 	struct blk_plug plug;
 	long diff;
 
+	if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
+		goto skip_write;
+
 	/* balancing f2fs's metadata in background */
 	f2fs_balance_fs_bg(sbi);

Signature · 6 lines

2.13.0.90.g1eb437020


------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot

Chao Yu

author maintainer

2017-06-29 15:23 UTC #1 lore raw

PATCH 2/2 f2fs: reuse original mount option in ->remount_fs

From: Chao Yu <[email protected]>

Don't clear old mount option before parse new option during ->remount_fs like other generic filesystems.

Signed-off-by Chao Yu <[email protected]>

---
 fs/f2fs/super.c | 1 -
 1 file changed, 1 deletion(-)

fs/f2fs/super.c +0 -1

index f27c141cd8aa..af472f7968d0 100644
@@ -846,7 +846,6 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
 			clear_sbi_flag(sbi, SBI_NEED_SB_WRITE);
 	}
 
-	sbi->mount_opt.opt = 0;
 	default_options(sbi);
 
 	/* parse mount options */

Signature · 6 lines

2.13.0.90.g1eb437020


------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot

Jaegeuk Kim

maintainer

2017-07-01 07:48 UTC #2 lore raw

Re: [PATCH 2/2] f2fs: reuse original mount option in ->remount_fs

Quoted 4 lines · Chao Yu

From: Chao Yu <[email protected]>
Don't clear old mount option before parse new option during ->remount_fs like other generic filesystems.

So, what happened on your original patch?

commit 26666c8a4366debae30ae37d0688b2bec92d196a
f2fs: fix to clean previous mount option when remount_fs

I think we need to rever the original patch tho.

Thanks,

Quoted 20 lines

Signed-off-by Chao Yu <[email protected]>

---
 fs/f2fs/super.c | 1 -
 1 file changed, 1 deletion(-)

fs/f2fs/super.c +0 -1

index f27c141cd8aa..af472f7968d0 100644
@@ -846,7 +846,6 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
 			clear_sbi_flag(sbi, SBI_NEED_SB_WRITE);
 	}
 
-	sbi->mount_opt.opt = 0;
 	default_options(sbi);
 
 	/* parse mount options */

Signature · 1 line

2.13.0.90.g1eb437020

Chao Yu

author maintainer

2017-07-01 16:14 UTC #3 lore raw

Re: [PATCH 2/2] f2fs: reuse original mount option in ->remount_fs

Quoted 11 lines · Jaegeuk Kim

Quoted 4 lines · Chao Yu
From: Chao Yu <[email protected]>
Don't clear old mount option before parse new option during ->remount_fs like other generic filesystems>
So, what happened on your original patch?
commit 26666c8a4366debae30ae37d0688b2bec92d196a
f2fs: fix to clean previous mount option when remount_fs
I think we need to rever the original patch tho.

I just removed this for passing some quota testcases in xfstests suit, and after that, I didn't get further failures in other testcases.

Let me check this again.

Thanks,

Quoted 23 lines

Thanks,

Quoted 20 lines

Signed-off-by Chao Yu <[email protected]>

---
 fs/f2fs/super.c | 1 -
 1 file changed, 1 deletion(-)

fs/f2fs/super.c +0 -1

index f27c141cd8aa..af472f7968d0 100644
@@ -846,7 +846,6 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
 			clear_sbi_flag(sbi, SBI_NEED_SB_WRITE);
 	}
 
-	sbi->mount_opt.opt = 0;
 	default_options(sbi);
 
 	/* parse mount options */

Signature · 1 line

2.13.0.90.g1eb437020

Chao Yu

author maintainer

2017-07-05 04:05 UTC #4 lore raw

Re: [PATCH 2/2] f2fs: reuse original mount option in ->remount_fs

In reply to Jaegeuk Kim

Quoted 10 lines · Jaegeuk Kim

Quoted 4 lines · Chao Yu
From: Chao Yu <[email protected]>
Don't clear old mount option before parse new option during ->remount_fs like other generic filesystems.
So, what happened on your original patch?
commit 26666c8a4366debae30ae37d0688b2bec92d196a
f2fs: fix to clean previous mount option when remount_fs

I think that patch is wrong, since the manual doesn't describe kernel logic, so I'd like revert that patch, then f2fs ->remount_fs logic is more like generic filesystem.

Thanks,

Quoted 24 lines

I think we need to rever the original patch tho.>
Thanks,

Quoted 20 lines

Signed-off-by Chao Yu <[email protected]>

---
 fs/f2fs/super.c | 1 -
 1 file changed, 1 deletion(-)

fs/f2fs/super.c +0 -1

index f27c141cd8aa..af472f7968d0 100644
@@ -846,7 +846,6 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data)
 			clear_sbi_flag(sbi, SBI_NEED_SB_WRITE);
 	}
 
-	sbi->mount_opt.opt = 0;
 	default_options(sbi);
 
 	/* parse mount options */

Signature · 1 line

2.13.0.90.g1eb437020