intel/compiler: Introduce backend_shader method to propagate IR changes to analysis...
authorFrancisco Jerez <currojerez@riseup.net>
Sun, 13 Mar 2016 02:50:24 +0000 (18:50 -0800)
committerMatt Turner <mattst88@gmail.com>
Fri, 6 Mar 2020 18:20:32 +0000 (10:20 -0800)
commitd966a6b4c4684bc02647a8fdc69a6c88e5ed00c2
treed0d4da56e4b463bfb3ddb06a9150fbc4879cc2df
parent03eb46f4a74c8df3de6785ffe18e968b876469b8
intel/compiler: Introduce backend_shader method to propagate IR changes to analysis passes

The invalidate_analysis() method knows what analysis passes there are
in the back-end and calls their invalidate() method to report changes
in the IR.  For the moment it just calls invalidate_live_intervals()
(which will eventually be fully replaced by this function) if anything
changed.

This makes all optimization passes invalidate DEPENDENCY_EVERYTHING,
which is clearly far from ideal -- The dependency classes passed to
invalidate_analysis() will be refined in a future commit.

Reviewed-by: Matt Turner <mattst88@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4012>
24 files changed:
src/intel/compiler/brw_dead_control_flow.cpp
src/intel/compiler/brw_fs.cpp
src/intel/compiler/brw_fs.h
src/intel/compiler/brw_fs_cmod_propagation.cpp
src/intel/compiler/brw_fs_combine_constants.cpp
src/intel/compiler/brw_fs_copy_propagation.cpp
src/intel/compiler/brw_fs_cse.cpp
src/intel/compiler/brw_fs_dead_code_eliminate.cpp
src/intel/compiler/brw_fs_lower_pack.cpp
src/intel/compiler/brw_fs_lower_regioning.cpp
src/intel/compiler/brw_fs_reg_allocate.cpp
src/intel/compiler/brw_fs_register_coalesce.cpp
src/intel/compiler/brw_fs_sel_peephole.cpp
src/intel/compiler/brw_predicated_break.cpp
src/intel/compiler/brw_schedule_instructions.cpp
src/intel/compiler/brw_shader.cpp
src/intel/compiler/brw_shader.h
src/intel/compiler/brw_vec4.cpp
src/intel/compiler/brw_vec4.h
src/intel/compiler/brw_vec4_cmod_propagation.cpp
src/intel/compiler/brw_vec4_copy_propagation.cpp
src/intel/compiler/brw_vec4_cse.cpp
src/intel/compiler/brw_vec4_dead_code_eliminate.cpp
src/intel/compiler/brw_vec4_reg_allocate.cpp