f2fs: avoid to wait all the node blocks during fsync

Previously, f2fs_sync_file() waits for all the node blocks to be written.
But, we don't need to do that, but wait only the inode-related node blocks.

This patch adds wait_on_node_pages_writeback() in which waits inode-related
node blocks that are on writeback.

Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>

3 files changed, 44 insertions, 2 deletions

diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index 6aaefdbaaac0..625eb4befad4 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -1026,6 +1026,7 @@ void get_node_info(struct f2fs_sb_info *, nid_t, struct node_info *);
 int get_dnode_of_data(struct dnode_of_data *, pgoff_t, int);
 int truncate_inode_blocks(struct inode *, pgoff_t);
 int truncate_xattr_node(struct inode *, struct page *);
+int wait_on_node_pages_writeback(struct f2fs_sb_info *, nid_t);
 int remove_inode_page(struct inode *);
 struct page *new_inode_page(struct inode *, const struct qstr *);
 struct page *new_node_page(struct dnode_of_data *, unsigned int, struct page *);
diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
index 58ed19aaa4c6..7d714f4972d5 100644
--- a/fs/f2fs/file.c
+++ b/fs/f2fs/file.c
@@ -189,8 +189,9 @@ int f2fs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
                         if (ret)
                                 goto out;
                 }
-                filemap_fdatawait_range(sbi->node_inode->i_mapping,
-                                                        0, LONG_MAX);
+                ret = wait_on_node_pages_writeback(sbi, inode->i_ino);
+                if (ret)
+                        goto out;
                 ret = blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL, NULL);
         }
 out:
diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
index 8e331d5b9fcd..b527ed4c0225 100644
--- a/fs/f2fs/node.c
+++ b/fs/f2fs/node.c
@@ -1148,6 +1148,46 @@ continue_unlock:
         return nwritten;
 }
 
+int wait_on_node_pages_writeback(struct f2fs_sb_info *sbi, nid_t ino)
+{
+        struct address_space *mapping = sbi->node_inode->i_mapping;
+        pgoff_t index = 0, end = LONG_MAX;
+        struct pagevec pvec;
+        int nr_pages;
+        int ret2 = 0, ret = 0;
+
+        pagevec_init(&pvec, 0);
+        while ((index <= end) &&
+                        (nr_pages = pagevec_lookup_tag(&pvec, mapping, &index,
+                        PAGECACHE_TAG_WRITEBACK,
+                        min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1)) != 0) {
+                unsigned i;
+
+                for (i = 0; i < nr_pages; i++) {
+                        struct page *page = pvec.pages[i];
+
+                        /* until radix tree lookup accepts end_index */
+                        if (page->index > end)
+                                continue;
+
+                        if (ino && ino_of_node(page) == ino)
+                                wait_on_page_writeback(page);
+                        if (TestClearPageError(page))
+                                ret = -EIO;
+                }
+                pagevec_release(&pvec);
+                cond_resched();
+        }
+
+        if (test_and_clear_bit(AS_ENOSPC, &mapping->flags))
+                ret2 = -ENOSPC;
+        if (test_and_clear_bit(AS_EIO, &mapping->flags))
+                ret2 = -EIO;
+        if (!ret)
+                ret = ret2;
+        return ret;
+}
+
 static int f2fs_write_node_page(struct page *page,
                                 struct writeback_control *wbc)
 {