[PATCH] f2fs: avoid SSA corruption when replacing block addresses

[Jaegeuk Kim]

If block addresses are pointed by the previous checkpoint, we should not update
the SSA when replacing block addresses by f2fs_collapse_range.

This patch allocates a new block address if there is an exising block address.

Signed-off-by: Jaegeuk Kim <jaegeuk@xxxxxxxxxx>
---
 fs/f2fs/segment.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index 1d86a35..3c546eb 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -1391,8 +1391,20 @@ void f2fs_replace_block(struct f2fs_sb_info *sbi, struct dnode_of_data *dn,
 {
 	struct f2fs_summary sum;
 
+	/*
+	 * we should not use the existing address to avoid SSA
+	 * corruption.
+	 */
 	set_summary(&sum, dn->nid, dn->ofs_in_node, version);
 
+	if (recover_curseg && old_addr != NULL_ADDR && old_addr != NEW_ADDR) {
+		allocate_data_block(sbi, NULL, dn->data_blkaddr,
+				&dn->data_blkaddr, &sum, CURSEG_WARM_DATA);
+		set_data_blkaddr(dn);
+		f2fs_update_extent_cache(dn);
+		old_addr = dn->data_blkaddr;
+	}
+
 	__f2fs_replace_block(sbi, &sum, old_addr, new_addr, recover_curseg);
 
 	dn->data_blkaddr = new_addr;
-- 
2.1.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/