Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1412263

[PATCH 6/8] nilfs2: emit error message when I/O error is detected

Show all headers | View raw

When nilfs returned -EIO as an error code, it's not always clear if it
came from the underlying block device or not.  This will mend the
issue by having low level I/O routines of nilfs output an error
message when they detected an I/O error.

Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
---
 fs/nilfs2/btree.c   | 3 +++
 fs/nilfs2/gcinode.c | 9 ++++++++-
 fs/nilfs2/mdt.c     | 6 +++++-
 3 files changed, 16 insertions(+), 2 deletions(-)

diff --git a/fs/nilfs2/btree.c b/fs/nilfs2/btree.c
index 4816c5a..493e699 100644
--- a/fs/nilfs2/btree.c
+++ b/fs/nilfs2/btree.c
@@ -515,6 +515,9 @@ static int __nilfs_btree_get_block(const struct nilfs_bmap *btree, __u64 ptr,
 
  out_no_wait:
 	if (!buffer_uptodate(bh)) {
+		nilfs_msg(btree->b_inode->i_sb, KERN_ERR,
+			  "I/O error reading b-tree node block (ino=%lu, blocknr=%llu)",
+			  btree->b_inode->i_ino, (unsigned long long)ptr);
 		brelse(bh);
 		return -EIO;
 	}
diff --git a/fs/nilfs2/gcinode.c b/fs/nilfs2/gcinode.c
index 693aded..b94db89 100644
--- a/fs/nilfs2/gcinode.c
+++ b/fs/nilfs2/gcinode.c
@@ -147,8 +147,15 @@ int nilfs_gccache_submit_read_node(struct inode *inode, sector_t pbn,
 int nilfs_gccache_wait_and_mark_dirty(struct buffer_head *bh)
 {
 	wait_on_buffer(bh);
-	if (!buffer_uptodate(bh))
+	if (!buffer_uptodate(bh)) {
+		struct inode *inode = bh->b_page->mapping->host;
+
+		nilfs_msg(inode->i_sb, KERN_ERR,
+			  "I/O error reading %s block for GC (ino=%lu, vblocknr=%llu)",
+			  buffer_nilfs_node(bh) ? "node" : "data",
+			  inode->i_ino, (unsigned long long)bh->b_blocknr);
 		return -EIO;
+	}
 	if (buffer_dirty(bh))
 		return -EEXIST;
 
diff --git a/fs/nilfs2/mdt.c b/fs/nilfs2/mdt.c
index 3417d85..27ba7c7 100644
--- a/fs/nilfs2/mdt.c
+++ b/fs/nilfs2/mdt.c
@@ -206,8 +206,12 @@ static int nilfs_mdt_read_block(struct inode *inode, unsigned long block,
 
  out_no_wait:
 	err = -EIO;
-	if (!buffer_uptodate(first_bh))
+	if (!buffer_uptodate(first_bh)) {
+		nilfs_msg(inode->i_sb, KERN_ERR,
+			  "I/O error reading meta-data file (ino=%lu, block-offset=%lu)",
+			  inode->i_ino, block);
 		goto failed_bh;
+	}
  out:
 	*out_bh = first_bh;
 	return 0;
-- 
1.8.3.1

Back to linux.kernel | Previous | Next — Previous in thread | Find similar | Unroll thread

Thread

[PATCH 0/8] nilfs2 updates Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> - 2016-06-02 16:10 +0200
  [PATCH 7/8] nilfs2: do not use yield() Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> - 2016-06-02 16:10 +0200
  [PATCH 3/8] nilfs2: embed a back pointer to super block instance in nilfs object Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> - 2016-06-02 16:10 +0200
  [PATCH 8/8] nilfs2: refactor parser of snapshot mount option Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> - 2016-06-02 16:10 +0200
  [PATCH 2/8] nilfs2: add nilfs_msg() message interface Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> - 2016-06-02 16:10 +0200
  [PATCH 5/8] nilfs2: replace nilfs_warning() with nilfs_msg() Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> - 2016-06-02 16:10 +0200
  [PATCH 6/8] nilfs2: emit error message when I/O error is detected Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp> - 2016-06-02 16:10 +0200

csiph-web