glsl/lower_instruction: handle denorms and overflow in ldexp correctly
authorNicolai Hähnle <nicolai.haehnle@amd.com>
Fri, 15 Sep 2017 14:39:31 +0000 (16:39 +0200)
committerNicolai Hähnle <nicolai.haehnle@amd.com>
Fri, 29 Sep 2017 10:07:08 +0000 (12:07 +0200)
commit93bf9c114b7c54e4faf342810bd848527b7d0a80
treed055c1eccc72f587248a11e7de91c8028a6987e3
parenta208cd7ae4c1613dfd9acafa6046c1cd0be4911f
glsl/lower_instruction: handle denorms and overflow in ldexp correctly

GLSL ES requires both, and while GLSL explicitly doesn't require correct
overflow handling, it does appear to require handling input inf/denorms
correctly.

Fixes dEQP-GLES31.functional.shaders.builtin_functions.precision.ldexp.*

Cc: mesa-stable@lists.freedesktop.org
Acked-by: Matt Turner <mattst88@gmail.com>
Acked-by: Marek Olšák <marek.olsak@amd.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
src/compiler/glsl/lower_instructions.cpp