xfs: don't bounce the iolock between free_{eof,cow}blocks

Since xfs_inode_free_eofblocks and xfs_inode_free_cowblocks are now
internal static functions, we can save ourselves a cycling of the iolock
by passing the lock state out to xfs_blockgc_scan_inode and letting it
do all the unlocking.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>