NTFS: Fix bug in mft record writing where we forgot to set the device in

      the buffers when mapping them after the VM had discarded them.
      Thanks to Martin MOKREJŠ for the bug report.

Signed-off-by: Anton Altaparmakov <aia21@cantab.net>

fs/ntfs/ChangeLog

@@ -174,6 +174,9 @@ ToDo/Notes:
 	  fact that the vfs and ntfs inodes are one struct in memory to find
 	  the ntfs inode in memory if present.  Also, the ntfs inode has its
 	  own locking so it does not matter if the vfs inode is locked.
+	- Fix bug in mft record writing where we forgot to set the device in
+	  the buffers when mapping them after the VM had discarded them
+	  Thanks to Martin MOKREJŠ for the bug report.
 
 2.1.22 - Many bug and race fixes and error handling improvements.
 

fs/ntfs/mft.c

@@ -533,6 +533,7 @@ int ntfs_sync_mft_mirror(ntfs_volume *vol, const unsigned long mft_no,
 			LCN lcn;
 			unsigned int vcn_ofs;
 
+			bh->b_bdev = vol->sb->s_bdev;
 			/* Obtain the vcn and offset of the current block. */
 			vcn = ((VCN)mft_no << vol->mft_record_size_bits) +
 					(block_start - m_start);
@@ -725,6 +726,7 @@ int write_mft_record_nolock(ntfs_inode *ni, MFT_RECORD *m, int sync)
 			LCN lcn;
 			unsigned int vcn_ofs;
 
+			bh->b_bdev = vol->sb->s_bdev;
 			/* Obtain the vcn and offset of the current block. */
 			vcn = ((VCN)ni->mft_no << vol->mft_record_size_bits) +
 					(block_start - m_start);