panfrost: Stop using sparse_array for batch BOs

Iterating over a util_sparse_array is very expensive; replace this
with a standard dynarray.

Using the sparse 'nodearray' datastructure instead was tested, but
found to be slower in some cases.

Part-of: <!16988>
49 jobs for !16988 with unsparse in 24 minutes and 39 seconds (queued for 10 seconds)
latest merge request