Mesa (master): freedreno/ir3/cf: handle widening too

[GitLab Mirror]

Module: Mesa
Branch: master
Commit: e73a8a9703b3be23ece12c1092185926522b6c4d
URL:    http://cgit.freedesktop.org/mesa/mesa/commit/?id=e73a8a9703b3be23ece12c1092185926522b6c4d

Author: Rob Clark <robdclark at chromium.org>
Date:   Mon Mar 16 06:47:05 2020 -0700

freedreno/ir3/cf: handle widening too

We can also fold f16->f32 conversions.

Signed-off-by: Rob Clark <robdclark at chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4423>

---

 src/freedreno/ir3/ir3_cf.c | 38 ++++++++++++++++++++++++++++++++------
 1 file changed, 32 insertions(+), 6 deletions(-)

diff --git a/src/freedreno/ir3/ir3_cf.c b/src/freedreno/ir3/ir3_cf.c
index 2c2dc06c27a..636ed384d79 100644
--- a/src/freedreno/ir3/ir3_cf.c
+++ b/src/freedreno/ir3/ir3_cf.c
@@ -26,11 +26,26 @@
 static bool
 is_fp16_conv(struct ir3_instruction *instr)
 {
-	if (instr->opc == OPC_MOV &&
-			instr->cat1.src_type == TYPE_F32 &&
+	if (instr->opc != OPC_MOV)
+		return false;
+
+	struct ir3_register *dst = instr->regs[0];
+	struct ir3_register *src = instr->regs[1];
+
+	/* disallow conversions that cannot be folded into
+	 * alu instructions:
+	 */
+	if (dst->flags & (IR3_REG_EVEN | IR3_REG_POS_INF))
+		return false;
+
+	if (instr->cat1.src_type == TYPE_F32 &&
 			instr->cat1.dst_type == TYPE_F16)
 		return true;
 
+	if (instr->cat1.src_type == TYPE_F16 &&
+			instr->cat1.dst_type == TYPE_F32)
+		return true;
+
 	return false;
 }
 
@@ -77,8 +92,15 @@ try_conversion_folding(struct ir3 *ir, struct ir3_instruction *conv)
 	if (!is_alu(src))
 		return;
 
+	/* avoid folding f2f32(f2f16) together, in cases where this is legal to
+	 * do (glsl) nir should have handled that for us already:
+	 */
+	if (is_fp16_conv(src))
+		return;
+
 	switch (src->opc) {
 	case OPC_SEL_B32:
+	case OPC_SEL_B16:
 	case OPC_MAX_F:
 	case OPC_MIN_F:
 	case OPC_SIGN_F:
@@ -97,18 +119,22 @@ try_conversion_folding(struct ir3 *ir, struct ir3_instruction *conv)
 			 * change the dst type to F32 to get the right behavior, since we
 			 * could be moving a float with a u32.u32 move.
 			 */
-			src->cat1.dst_type = TYPE_F16;
-			src->cat1.src_type = TYPE_F32;
+			src->cat1.dst_type = conv->cat1.dst_type;
+			src->cat1.src_type = conv->cat1.src_type;
 		} else {
 			/* Otherwise, for typechanging movs, we can just change the dst
 			 * type to F16 to collaps the two conversions.  For example
 			 * cov.s32f32 follwed by cov.f32f16 becomes cov.s32f16.
 			 */
-			src->cat1.dst_type = TYPE_F16;
+			src->cat1.dst_type = conv->cat1.dst_type;
 		}
 	}
 
-	src->regs[0]->flags |= IR3_REG_HALF;
+	if (conv->regs[0]->flags & IR3_REG_HALF) {
+		src->regs[0]->flags |= IR3_REG_HALF;
+	} else {
+		src->regs[0]->flags &= ~IR3_REG_HALF;
+	}
 
 	rewrite_uses(ir, conv, src);
 }