Different parts of our codebase disagree on whether spatial coordinates/dimensions are given in pixels or blocks (clumps), which differ by a constant factor for block-compressed formats. This disagreement manifests as incorrect results accessing block-compressed formats (either linear or tiled).
To resolve this, define the public tiling routines to take their coordinates in blocks, not pixels, and translate the pipe_box to blocks when transferring regions. This matches the strategy used by v3d where the translation code was copied from, and it neatly avoids special cases for linear vs tiled operation.
Signed-off-by: Alyssa Rosenzweig email@example.com Reported-by: Icecream95 Cc: mesa-stable