2015-02-02 22:51:24

by Jaegeuk Kim

[permalink] [raw]
Subject: [PATCH 1/2] f2fs: keep PagePrivate during releasepage

If PagePrivate is removed by releasepage, f2fs loses counting dirty pages.

e.g., try_to_release_page will not release page when the page is dirty,
but our releasepage removes PagePrivate.

[<ffffffff81188d75>] try_to_release_page+0x35/0x50
[<ffffffff811996f9>] invalidate_inode_pages2_range+0x2f9/0x3b0
[<ffffffffa02a7f54>] ? truncate_blocks+0x384/0x4d0 [f2fs]
[<ffffffffa02b7583>] ? f2fs_direct_IO+0x283/0x290 [f2fs]
[<ffffffffa02b7fb0>] ? get_data_block_fiemap+0x20/0x20 [f2fs]
[<ffffffff8118aa53>] generic_file_direct_write+0x163/0x170
[<ffffffff8118ad06>] __generic_file_write_iter+0x2a6/0x350
[<ffffffff8118adef>] generic_file_write_iter+0x3f/0xb0
[<ffffffff81203081>] new_sync_write+0x81/0xb0
[<ffffffff81203837>] vfs_write+0xb7/0x1f0
[<ffffffff81204459>] SyS_write+0x49/0xb0
[<ffffffff817c286d>] system_call_fastpath+0x16/0x1b

Signed-off-by: Jaegeuk Kim <[email protected]>
---
fs/f2fs/checkpoint.c | 4 ++++
fs/f2fs/data.c | 4 ++++
fs/f2fs/node.c | 4 ++++
3 files changed, 12 insertions(+)

diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c
index f7cdcad..470fa58 100644
--- a/fs/f2fs/checkpoint.c
+++ b/fs/f2fs/checkpoint.c
@@ -321,6 +321,10 @@ static void f2fs_invalidate_meta_page(struct page *page, unsigned int offset,

static int f2fs_release_meta_page(struct page *page, gfp_t wait)
{
+ /* If this is dirty page, keep PagePrivate */
+ if (PageDirty(page))
+ return 0;
+
ClearPagePrivate(page);
return 1;
}
diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c
index 6d9e6e4..27dc3fd 100644
--- a/fs/f2fs/data.c
+++ b/fs/f2fs/data.c
@@ -1130,6 +1130,10 @@ static void f2fs_invalidate_data_page(struct page *page, unsigned int offset,

static int f2fs_release_data_page(struct page *page, gfp_t wait)
{
+ /* If this is dirty page, keep PagePrivate */
+ if (PageDirty(page))
+ return 0;
+
ClearPagePrivate(page);
return 1;
}
diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
index 05e6faa..9feda38 100644
--- a/fs/f2fs/node.c
+++ b/fs/f2fs/node.c
@@ -1381,6 +1381,10 @@ static void f2fs_invalidate_node_page(struct page *page, unsigned int offset,

static int f2fs_release_node_page(struct page *page, gfp_t wait)
{
+ /* If this is dirty page, keep PagePrivate */
+ if (PageDirty(page))
+ return 0;
+
ClearPagePrivate(page);
return 1;
}
--
2.1.1


2015-02-02 22:51:31

by Jaegeuk Kim

[permalink] [raw]
Subject: [PATCH 2/2] f2fs: show the number of writeback pages in stat

This patch adds the # of writeback pages in stat info.

Signed-off-by: Jaegeuk Kim <[email protected]>
---
fs/f2fs/debug.c | 5 +++--
fs/f2fs/f2fs.h | 2 +-
2 files changed, 4 insertions(+), 3 deletions(-)

diff --git a/fs/f2fs/debug.c b/fs/f2fs/debug.c
index 0f721f6..ac2bd8e 100644
--- a/fs/f2fs/debug.c
+++ b/fs/f2fs/debug.c
@@ -40,6 +40,7 @@ static void update_general_status(struct f2fs_sb_info *sbi)
si->ndirty_dirs = sbi->n_dirty_dirs;
si->ndirty_meta = get_pages(sbi, F2FS_DIRTY_META);
si->inmem_pages = get_pages(sbi, F2FS_INMEM_PAGES);
+ si->wb_pages = get_pages(sbi, F2FS_WRITEBACK);
si->total_count = (int)sbi->user_block_count / sbi->blocks_per_seg;
si->rsvd_segs = reserved_segments(sbi);
si->overp_segs = overprovision_segments(sbi);
@@ -266,8 +267,8 @@ static int stat_show(struct seq_file *s, void *v)
seq_printf(s, "\nExtent Hit Ratio: %d / %d\n",
si->hit_ext, si->total_ext);
seq_puts(s, "\nBalancing F2FS Async:\n");
- seq_printf(s, " - inmem: %4d\n",
- si->inmem_pages);
+ seq_printf(s, " - inmem: %4d, wb: %4d\n",
+ si->inmem_pages, si->wb_pages);
seq_printf(s, " - nodes: %4d in %4d\n",
si->ndirty_node, si->node_pages);
seq_printf(s, " - dents: %4d in dirs:%4d\n",
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index d89dc5d0..d005939c3 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -1550,7 +1550,7 @@ struct f2fs_stat_info {
int ndirty_node, ndirty_dent, ndirty_dirs, ndirty_meta;
int nats, dirty_nats, sits, dirty_sits, fnids;
int total_count, utilization;
- int bg_gc, inline_inode, inline_dir, inmem_pages;
+ int bg_gc, inline_inode, inline_dir, inmem_pages, wb_pages;
unsigned int valid_count, valid_node_count, valid_inode_count;
unsigned int bimodal, avg_vblocks;
int util_free, util_valid, util_invalid;
--
2.1.1

2015-02-05 07:51:40

by Chao Yu

[permalink] [raw]
Subject: RE: [f2fs-dev] [PATCH 1/2] f2fs: keep PagePrivate during releasepage

> -----Original Message-----
> From: Jaegeuk Kim [mailto:[email protected]]
> Sent: Tuesday, February 03, 2015 6:51 AM
> To: [email protected]; [email protected];
> [email protected]
> Cc: Jaegeuk Kim
> Subject: [f2fs-dev] [PATCH 1/2] f2fs: keep PagePrivate during releasepage
>
> If PagePrivate is removed by releasepage, f2fs loses counting dirty pages.
>
> e.g., try_to_release_page will not release page when the page is dirty,
> but our releasepage removes PagePrivate.
>
> [<ffffffff81188d75>] try_to_release_page+0x35/0x50
> [<ffffffff811996f9>] invalidate_inode_pages2_range+0x2f9/0x3b0
> [<ffffffffa02a7f54>] ? truncate_blocks+0x384/0x4d0 [f2fs]
> [<ffffffffa02b7583>] ? f2fs_direct_IO+0x283/0x290 [f2fs]
> [<ffffffffa02b7fb0>] ? get_data_block_fiemap+0x20/0x20 [f2fs]
> [<ffffffff8118aa53>] generic_file_direct_write+0x163/0x170
> [<ffffffff8118ad06>] __generic_file_write_iter+0x2a6/0x350
> [<ffffffff8118adef>] generic_file_write_iter+0x3f/0xb0
> [<ffffffff81203081>] new_sync_write+0x81/0xb0
> [<ffffffff81203837>] vfs_write+0xb7/0x1f0
> [<ffffffff81204459>] SyS_write+0x49/0xb0
> [<ffffffff817c286d>] system_call_fastpath+0x16/0x1b
>
> Signed-off-by: Jaegeuk Kim <[email protected]>

Reviewed-by: Chao Yu <[email protected]>

2015-02-05 07:53:32

by Chao Yu

[permalink] [raw]
Subject: RE: [f2fs-dev] [PATCH 2/2] f2fs: show the number of writeback pages in stat

> -----Original Message-----
> From: Jaegeuk Kim [mailto:[email protected]]
> Sent: Tuesday, February 03, 2015 6:51 AM
> To: [email protected]; [email protected];
> [email protected]
> Cc: Jaegeuk Kim
> Subject: [f2fs-dev] [PATCH 2/2] f2fs: show the number of writeback pages in stat
>
> This patch adds the # of writeback pages in stat info.
>
> Signed-off-by: Jaegeuk Kim <[email protected]>

Reviewed-by: Chao Yu <[email protected]>