nilfs2: fix gc failure on volumes keeping numerous snapshots

This resolves the following failure of nilfs2 cleaner daemon:

 nilfs_cleanerd[20670]: cannot clean segments: No such file or directory
 nilfs_cleanerd[20670]: shutdown

When creating thousands of snapshots, the cleaner daemon had rarely died
as above due to an error returned from the kernel code.

After applying the recent patch which fixed memory allocation problems in
ioctl (Message-Id: <20081215.155840.105124170.ryusuke@osrg.net>), the
problem gets more frequent.

It turned out to be a bug of nilfs_ioctl_wrap_copy function and one of its
callback routines to read out information of snapshots; if the
nilfs_ioctl_wrap_copy function divided a large read request into multiple
requests, the second and later requests have failed since a restart
position on snapshot meta data was not properly set forward.

It's a deficiency of the callback interface that cannot pass the restart
position among multiple requests.  This patch fixes the issue by allowing
nilfs_ioctl_wrap_copy and snapshot read functions to exchange a position
argument.

Signed-off-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

1 files changed, 14 insertions, 10 deletions

diff --git a/fs/nilfs2/cpfile.c b/fs/nilfs2/cpfile.c
index 82462acd06e..a4c9550fd77 100644
--- a/fs/nilfs2/cpfile.c
+++ b/fs/nilfs2/cpfile.c
@@ -422,20 +422,20 @@ static ssize_t nilfs_cpfile_do_get_cpinfo(struct inode *cpfile, __u64 cno,
 	return ret;
 }
 
-static ssize_t nilfs_cpfile_do_get_ssinfo(struct inode *cpfile, __u64 cno,
+static ssize_t nilfs_cpfile_do_get_ssinfo(struct inode *cpfile, __u64 *cnop,
 					  struct nilfs_cpinfo *ci, size_t nci)
 {
 	struct buffer_head *bh;
 	struct nilfs_cpfile_header *header;
 	struct nilfs_checkpoint *cp;
-	__u64 curr, next;
+	__u64 curr = *cnop, next;
 	unsigned long curr_blkoff, next_blkoff;
 	void *kaddr;
 	int n, ret;
 
 	down_read(&NILFS_MDT(cpfile)->mi_sem);
 
-	if (cno == 0) {
+	if (curr == 0) {
 		ret = nilfs_cpfile_get_header_block(cpfile, &bh);
 		if (ret < 0)
 			goto out;
@@ -448,8 +448,11 @@ static ssize_t nilfs_cpfile_do_get_ssinfo(struct inode *cpfile, __u64 cno,
 			ret = 0;
 			goto out;
 		}
-	} else
-		curr = cno;
+	} else if (unlikely(curr == ~(__u64)0)) {
+		ret = 0;
+		goto out;
+	}
+
 	curr_blkoff = nilfs_cpfile_get_blkoff(cpfile, curr);
 	ret = nilfs_cpfile_get_checkpoint_block(cpfile, curr, 0, &bh);
 	if (ret < 0)
@@ -461,7 +464,7 @@ static ssize_t nilfs_cpfile_do_get_ssinfo(struct inode *cpfile, __u64 cno,
 		nilfs_cpfile_checkpoint_to_cpinfo(cpfile, cp, &ci[n]);
 		next = le64_to_cpu(cp->cp_snapshot_list.ssl_next);
 		if (next == 0) {
-			curr = next;
+			curr = ~(__u64)0; /* Terminator */
 			n++;
 			break;
 		}
@@ -480,6 +483,7 @@ static ssize_t nilfs_cpfile_do_get_ssinfo(struct inode *cpfile, __u64 cno,
 	}
 	kunmap_atomic(kaddr, KM_USER0);
 	brelse(bh);
+	*cnop = curr;
 	ret = n;
 
  out:
@@ -494,15 +498,15 @@ static ssize_t nilfs_cpfile_do_get_ssinfo(struct inode *cpfile, __u64 cno,
  * @ci:
  * @nci:
  */
-ssize_t nilfs_cpfile_get_cpinfo(struct inode *cpfile,
-				__u64 cno, int mode,
+
+ssize_t nilfs_cpfile_get_cpinfo(struct inode *cpfile, __u64 *cnop, int mode,
 				struct nilfs_cpinfo *ci, size_t nci)
 {
 	switch (mode) {
 	case NILFS_CHECKPOINT:
-		return nilfs_cpfile_do_get_cpinfo(cpfile, cno, ci, nci);
+		return nilfs_cpfile_do_get_cpinfo(cpfile, *cnop, ci, nci);
 	case NILFS_SNAPSHOT:
-		return nilfs_cpfile_do_get_ssinfo(cpfile, cno, ci, nci);
+		return nilfs_cpfile_do_get_ssinfo(cpfile, cnop, ci, nci);
 	default:
 		return -EINVAL;
 	}