[Mesa-dev] [PATCH 1/3] nir/algebraic: Separate ffma lowering from fusing

[Jason Ekstrand]

The i965 driver has its own pass for fusing mul+add combinations that's
much smarter than what nir_opt_algebraic can do so we don't want to get the
nir_opt_algebraic one just because we didn't set lower_ffma.
---
 src/compiler/nir/nir.h                      | 1 +
 src/compiler/nir/nir_opt_algebraic.py       | 2 +-
 src/gallium/drivers/freedreno/ir3/ir3_nir.c | 1 +
 3 files changed, 3 insertions(+), 1 deletion(-)

diff --git a/src/compiler/nir/nir.h b/src/compiler/nir/nir.h
index 8a616d4..ba9d311 100644
--- a/src/compiler/nir/nir.h
+++ b/src/compiler/nir/nir.h
@@ -1624,6 +1624,7 @@ typedef struct nir_function {
 typedef struct nir_shader_compiler_options {
    bool lower_fdiv;
    bool lower_ffma;
+   bool fuse_ffma;
    bool lower_flrp32;
    /** Lowers flrp when it does not support doubles */
    bool lower_flrp64;
diff --git a/src/compiler/nir/nir_opt_algebraic.py b/src/compiler/nir/nir_opt_algebraic.py
index 0a95725..f8db2b6 100644
--- a/src/compiler/nir/nir_opt_algebraic.py
+++ b/src/compiler/nir/nir_opt_algebraic.py
@@ -108,7 +108,7 @@ optimizations = [
    (('~fadd at 32', a, ('fmul',         c , ('fadd', b, ('fneg', a)))), ('flrp', a, b, c), '!options->lower_flrp32'),
    (('~fadd at 64', a, ('fmul',         c , ('fadd', b, ('fneg', a)))), ('flrp', a, b, c), '!options->lower_flrp64'),
    (('ffma', a, b, c), ('fadd', ('fmul', a, b), c), 'options->lower_ffma'),
-   (('~fadd', ('fmul', a, b), c), ('ffma', a, b, c), '!options->lower_ffma'),
+   (('~fadd', ('fmul', a, b), c), ('ffma', a, b, c), 'options->fuse_ffma'),
    # Comparison simplifications
    (('~inot', ('flt', a, b)), ('fge', a, b)),
    (('~inot', ('fge', a, b)), ('flt', a, b)),
diff --git a/src/gallium/drivers/freedreno/ir3/ir3_nir.c b/src/gallium/drivers/freedreno/ir3/ir3_nir.c
index 99c2ecb..5662927 100644
--- a/src/gallium/drivers/freedreno/ir3/ir3_nir.c
+++ b/src/gallium/drivers/freedreno/ir3/ir3_nir.c
@@ -44,6 +44,7 @@ ir3_tgsi_to_nir(const struct tgsi_token *tokens)
 			.lower_scmp = true,
 			.lower_flrp32 = true,
 			.lower_ffract = true,
+			.fuse_ffma = true,
 			.native_integers = true,
 			.vertex_id_zero_based = true,
 			.lower_extract_byte = true,
-- 
2.5.0.400.gff86faf