Commit 2291dab2 authored by Dave Chinner's avatar Dave Chinner Committed by Dave Chinner
Browse files

xfs: Always flush caches when integrity is required 



There is no reason anymore for not issuing device integrity
operations when teh filesystem requires ordering or data integrity
guarantees. We should always issue cache flushes and FUA writes
where necessary and let the underlying storage optimise them as
necessary for correct integrity operation.

Signed-Off-By: default avatarDave Chinner <dchinner@redhat.com>
Reviewed-by: default avatarChristoph Hellwig <hch@lst.de>
Signed-off-by: default avatarDave Chinner <david@fromorbit.com>
parent 2e1d2337

fs/xfs/xfs_buf.c

+1 −2
Original line number Original line Diff line number Diff line
@@ -1711,7 +1711,6 @@ xfs_free_buftarg( 
	percpu_counter_destroy(&btp->bt_io_count);

	percpu_counter_destroy(&btp->bt_io_count);

	list_lru_destroy(&btp->bt_lru);

	list_lru_destroy(&btp->bt_lru);





	if (mp->m_flags & XFS_MOUNT_BARRIER)

	xfs_blkdev_issue_flush(btp);

	xfs_blkdev_issue_flush(btp);





	kmem_free(btp);

	kmem_free(btp);

fs/xfs/xfs_file.c

+12 −17
Original line number Original line Diff line number Diff line
@@ -149,19 +149,16 @@ xfs_file_fsync( 




	xfs_iflags_clear(ip, XFS_ITRUNCATED);

	xfs_iflags_clear(ip, XFS_ITRUNCATED);





	if (mp->m_flags & XFS_MOUNT_BARRIER) {

	/*

	/*

		 * If we have an RT and/or log subvolume we need to make sure

	 * If we have an RT and/or log subvolume we need to make sure to flush

		 * to flush the write cache the device used for file data

	 * the write cache the device used for file data first.  This is to

		 * first.  This is to ensure newly written file data make

	 * ensure newly written file data make it to disk before logging the new

		 * it to disk before logging the new inode size in case of

	 * inode size in case of an extending write.

		 * an extending write.

	 */

	 */

	if (XFS_IS_REALTIME_INODE(ip))

	if (XFS_IS_REALTIME_INODE(ip))

		xfs_blkdev_issue_flush(mp->m_rtdev_targp);

		xfs_blkdev_issue_flush(mp->m_rtdev_targp);

	else if (mp->m_logdev_targp != mp->m_ddev_targp)

	else if (mp->m_logdev_targp != mp->m_ddev_targp)

		xfs_blkdev_issue_flush(mp->m_ddev_targp);

		xfs_blkdev_issue_flush(mp->m_ddev_targp);

	}





	/*

	/*

	 * All metadata updates are logged, which means that we just have to

	 * All metadata updates are logged, which means that we just have to
@@ -196,10 +193,8 @@ xfs_file_fsync( 
	 * an already allocated file and thus do not have any metadata to

	 * an already allocated file and thus do not have any metadata to

	 * commit.

	 * commit.

	 */

	 */

	if ((mp->m_flags & XFS_MOUNT_BARRIER) &&

	if (!log_flushed && !XFS_IS_REALTIME_INODE(ip) &&

	    mp->m_logdev_targp == mp->m_ddev_targp &&

	    mp->m_logdev_targp == mp->m_ddev_targp)

	    !XFS_IS_REALTIME_INODE(ip) &&

	    !log_flushed)

		xfs_blkdev_issue_flush(mp->m_ddev_targp);

		xfs_blkdev_issue_flush(mp->m_ddev_targp);





	return error;

	return error;

fs/xfs/xfs_log.c

+16 −23
Original line number Original line Diff line number Diff line
@@ -1862,26 +1862,21 @@ xlog_sync( 




	bp->b_io_length = BTOBB(count);

	bp->b_io_length = BTOBB(count);

	bp->b_fspriv = iclog;

	bp->b_fspriv = iclog;

	bp->b_flags &= ~(XBF_FUA | XBF_FLUSH);

	bp->b_flags &= ~XBF_FLUSH;

	bp->b_flags |= (XBF_ASYNC | XBF_SYNCIO | XBF_WRITE);

	bp->b_flags |= (XBF_ASYNC | XBF_SYNCIO | XBF_WRITE | XBF_FUA);



	if (log->l_mp->m_flags & XFS_MOUNT_BARRIER) {

		bp->b_flags |= XBF_FUA;





	/*

	/*

		 * Flush the data device before flushing the log to make

	 * Flush the data device before flushing the log to make sure all meta

		 * sure all meta data written back from the AIL actually made

	 * data written back from the AIL actually made it to disk before

		 * it to disk before stamping the new log tail LSN into the

	 * stamping the new log tail LSN into the log buffer.  For an external

		 * log buffer.  For an external log we need to issue the

	 * log we need to issue the flush explicitly, and unfortunately

		 * flush explicitly, and unfortunately synchronously here;

	 * synchronously here; for an internal log we can simply use the block

		 * for an internal log we can simply use the block layer

	 * layer state machine for preflushes.

		 * state machine for preflushes.

	 */

	 */

	if (log->l_mp->m_logdev_targp != log->l_mp->m_ddev_targp)

	if (log->l_mp->m_logdev_targp != log->l_mp->m_ddev_targp)

		xfs_blkdev_issue_flush(log->l_mp->m_ddev_targp);

		xfs_blkdev_issue_flush(log->l_mp->m_ddev_targp);

	else

	else

		bp->b_flags |= XBF_FLUSH;

		bp->b_flags |= XBF_FLUSH;

	}





	ASSERT(XFS_BUF_ADDR(bp) <= log->l_logBBsize-1);

	ASSERT(XFS_BUF_ADDR(bp) <= log->l_logBBsize-1);

	ASSERT(XFS_BUF_ADDR(bp) + BTOBB(count) <= log->l_logBBsize);

	ASSERT(XFS_BUF_ADDR(bp) + BTOBB(count) <= log->l_logBBsize);
@@ -1906,10 +1901,8 @@ xlog_sync( 
		xfs_buf_associate_memory(bp,

		xfs_buf_associate_memory(bp,

				(char *)&iclog->ic_header + count, split);

				(char *)&iclog->ic_header + count, split);

		bp->b_fspriv = iclog;

		bp->b_fspriv = iclog;

		bp->b_flags &= ~(XBF_FUA | XBF_FLUSH);

		bp->b_flags &= ~XBF_FLUSH;

		bp->b_flags |= (XBF_ASYNC | XBF_SYNCIO | XBF_WRITE);

		bp->b_flags |= (XBF_ASYNC | XBF_SYNCIO | XBF_WRITE | XBF_FUA);

		if (log->l_mp->m_flags & XFS_MOUNT_BARRIER)

			bp->b_flags |= XBF_FUA;





		ASSERT(XFS_BUF_ADDR(bp) <= log->l_logBBsize-1);

		ASSERT(XFS_BUF_ADDR(bp) <= log->l_logBBsize-1);

		ASSERT(XFS_BUF_ADDR(bp) + BTOBB(count) <= log->l_logBBsize);

		ASSERT(XFS_BUF_ADDR(bp) + BTOBB(count) <= log->l_logBBsize);