From 00cd455bd50c6b16b2f72a6d2726de3d3818c7f5 Mon Sep 17 00:00:00 2001 From: Zack Rusin Date: Thu, 11 Jul 2013 12:16:06 -0400 Subject: [PATCH] gallium: fixup definitions of the rsq and sqrt GLSL spec says that rsq is undefined for src<=0, but the D3D10 spec says it needs to be a NaN, so lets stop taking an absolute value of the source which completely breaks that behavior. For the gl program we can simply insert an extra abs instrunction which produces the desired behavior there. Signed-off-by: Zack Rusin Reviewed-by: Roland Scheidegger Reviewed-by: Brian Paul --- .../auxiliary/gallivm/lp_bld_tgsi_action.c | 5 ----- src/gallium/auxiliary/tgsi/tgsi_exec.c | 16 ++++++++-------- src/gallium/docs/source/tgsi.rst | 6 +++--- src/mesa/state_tracker/st_mesa_to_tgsi.c | 6 ++++-- 4 files changed, 15 insertions(+), 18 deletions(-) diff --git a/src/gallium/auxiliary/gallivm/lp_bld_tgsi_action.c b/src/gallium/auxiliary/gallivm/lp_bld_tgsi_action.c index 68bd1243ece..e99c8ef134e 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_tgsi_action.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_tgsi_action.c @@ -633,8 +633,6 @@ rsq_emit( struct lp_build_tgsi_context * bld_base, struct lp_build_emit_data * emit_data) { - emit_data->args[0] = lp_build_emit_llvm_unary(bld_base, TGSI_OPCODE_ABS, - emit_data->args[0]); if (bld_base->rsq_action.emit) { bld_base->rsq_action.emit(&bld_base->rsq_action, bld_base, emit_data); } else { @@ -1349,9 +1347,6 @@ rcp_emit_cpu( } /* Reciprical squareroot (CPU Only) */ - -/* This is not the same as TGSI_OPCODE_RSQ, which requres the argument to be - * greater than or equal to 0 */ static void recip_sqrt_emit_cpu( const struct lp_build_tgsi_action * action, diff --git a/src/gallium/auxiliary/tgsi/tgsi_exec.c b/src/gallium/auxiliary/tgsi/tgsi_exec.c index bed08522fcd..1f8e62d6f3e 100644 --- a/src/gallium/auxiliary/tgsi/tgsi_exec.c +++ b/src/gallium/auxiliary/tgsi/tgsi_exec.c @@ -339,20 +339,20 @@ micro_rsq(union tgsi_exec_channel *dst, assert(src->f[2] != 0.0f); assert(src->f[3] != 0.0f); #endif - dst->f[0] = 1.0f / sqrtf(fabsf(src->f[0])); - dst->f[1] = 1.0f / sqrtf(fabsf(src->f[1])); - dst->f[2] = 1.0f / sqrtf(fabsf(src->f[2])); - dst->f[3] = 1.0f / sqrtf(fabsf(src->f[3])); + dst->f[0] = 1.0f / sqrtf(src->f[0]); + dst->f[1] = 1.0f / sqrtf(src->f[1]); + dst->f[2] = 1.0f / sqrtf(src->f[2]); + dst->f[3] = 1.0f / sqrtf(src->f[3]); } static void micro_sqrt(union tgsi_exec_channel *dst, const union tgsi_exec_channel *src) { - dst->f[0] = sqrtf(fabsf(src->f[0])); - dst->f[1] = sqrtf(fabsf(src->f[1])); - dst->f[2] = sqrtf(fabsf(src->f[2])); - dst->f[3] = sqrtf(fabsf(src->f[3])); + dst->f[0] = sqrtf(src->f[0]); + dst->f[1] = sqrtf(src->f[1]); + dst->f[2] = sqrtf(src->f[2]); + dst->f[3] = sqrtf(src->f[3]); } static void diff --git a/src/gallium/docs/source/tgsi.rst b/src/gallium/docs/source/tgsi.rst index 13daa62ab62..4d26c465579 100644 --- a/src/gallium/docs/source/tgsi.rst +++ b/src/gallium/docs/source/tgsi.rst @@ -94,16 +94,16 @@ This instruction replicates its result. .. opcode:: RSQ - Reciprocal Square Root -This instruction replicates its result. +This instruction replicates its result. The results are undefined for src <= 0. .. math:: - dst = \frac{1}{\sqrt{|src.x|}} + dst = \frac{1}{\sqrt{src.x}} .. opcode:: SQRT - Square Root -This instruction replicates its result. +This instruction replicates its result. The results are undefined for src < 0. .. math:: diff --git a/src/mesa/state_tracker/st_mesa_to_tgsi.c b/src/mesa/state_tracker/st_mesa_to_tgsi.c index 7f306973f3d..67c6f53f638 100644 --- a/src/mesa/state_tracker/st_mesa_to_tgsi.c +++ b/src/mesa/state_tracker/st_mesa_to_tgsi.c @@ -615,8 +615,6 @@ translate_opcode( unsigned op ) return TGSI_OPCODE_RCP; case OPCODE_RET: return TGSI_OPCODE_RET; - case OPCODE_RSQ: - return TGSI_OPCODE_RSQ; case OPCODE_SCS: return TGSI_OPCODE_SCS; case OPCODE_SEQ: @@ -756,6 +754,10 @@ compile_instruction( emit_ddy( t, dst[0], &inst->SrcReg[0] ); break; + case OPCODE_RSQ: + ureg_RSQ( ureg, dst[0], ureg_abs(src[0]) ); + break; + default: ureg_insn( ureg, translate_opcode( inst->Opcode ), -- 2.30.2