Code fails to unroll when compiled for ps_5_0

[billgan1024]

Type: Bug

vec4 operator*(mat4 m, vec4 v)
{
   return mul(m, v);
}
mat4 operator*(mat4 a, mat4 b)
{
   return mul(a, b);
}
vec3 operator*(mat3 m, vec3 v)
{
   return mul(m, v);
}
mat3 operator*(mat3 a, mat3 b)
{
   return mul(a, b);
}
void psShadow(vec4 svp: SV_Position, vec3 normal: normal, vec2 uv: uv, vec3 pos: pos,
              Texture2D diffuse, Texture2D<float> shadow[4],
              SamplerState sampler, SamplerComparisonState cmp,
              ConstantBuffer<vec3> light, 
              ConstantBuffer<mat4[4]> cascades,
              out vec4 target: SV_Target)
{
   normal = normalize(normal);
   float threshold = 0.25;
   float lambert = (1 - threshold) * ((dot(normal, light) + 1) / 2) + threshold;

   float inLight = 1;
   if (dot(normal, light) > 0)
   {
      for (int i = 0; i < 4; i++)
      {
         vec3 ndc = (cascades[i] * vec4(pos, 1)).xyz;
         if (abs(ndc.x) < 1 && abs(ndc.y) < 1 && ndc.z > 0 && ndc.z < 1)
         {
            inLight = shadow[i].SampleCmp(cmp, vec2((ndc.x + 1) / 2, 1 - (ndc.y + 1) / 2), ndc.z);
            break;
         }
      }
   }
   target = lambert * lerp(0.65, 1, inLight) * diffuse.Sample(sampler, uv);
}

compile with
slangc -matrix-layout-row-major -profile ps_5_0 shaders.hlsl -entry psShadow -o out/psShadow.dxbc gives
fxc 47.0: shaders.hlsl(177): error X3511: forced to unroll loop, but unrolling failed.

Here is the generated HLSL:

#pragma pack_matrix(row_major)
#ifdef SLANG_HLSL_ENABLE_NVAPI
#include "nvHLSLExtns.h"
#endif

#ifndef __DXC_VERSION_MAJOR
// warning X3557: loop doesn't seem to do anything, forcing loop to unroll
#pragma warning(disable: 3557)
#endif

Texture2D<float4> entryPointParams_diffuse_0 : register(t0);

Texture2D<float> entryPointParams_shadow_0[int(4)] : register(t1);

SamplerState entryPointParams_sampler_0 : register(s0);

SamplerComparisonState entryPointParams_cmp_0 : register(s1);

cbuffer entryPointParams_light_0 : register(b0)
{
   float3 entryPointParams_light_0;
}

cbuffer entryPointParams_cascades_0 : register(b1)
{
   float4x4 entryPointParams_cascades_0[int(4)];
}

float4 x2A_0(float4x4 m_0, float4 v_0)
{
   return mul(m_0, v_0);
}

void psShadow(float4 svp_0: SV_Position, float3 normal_0: normal, float2 uv_0: uv, float3 pos_0: pos, out float4 target_0: SV_Target)
{

   float2 _S1 = uv_0;

   float3 _S2 = pos_0;

   float _S3 = dot(normalize(normal_0), entryPointParams_light_0);

   float lambert_0 = 0.75f * ((_S3 + 1.0f) / 2.0f) + 0.25f;

   float inLight_0;

   if (_S3 > 0.0f)
   {

      int i_0 = int(0);

      for (;;)
      {

         if (i_0 < int(4))
         {
         }
         else
         {

            inLight_0 = 1.0f;

            break;
         }
         float3 ndc_0 = x2A_0(entryPointParams_cascades_0[i_0], float4(_S2, 1.0f)).xyz;
         float _S4 = ndc_0.x;

         bool _S5;

         if ((abs(_S4)) < 1.0f)
         {

            _S5 = (abs(ndc_0.y)) < 1.0f;
         }
         else
         {

            _S5 = false;
         }

         bool _S6;

         if (_S5)
         {

            _S6 = (ndc_0.z) > 0.0f;
         }
         else
         {

            _S6 = false;
         }

         bool _S7;

         if (_S6)
         {

            _S7 = (ndc_0.z) < 1.0f;
         }
         else
         {

            _S7 = false;
         }

         if (_S7)
         {

            inLight_0 = entryPointParams_shadow_0[i_0].SampleCmp(entryPointParams_cmp_0, float2((_S4 + 1.0f) / 2.0f, 1.0f - (ndc_0.y + 1.0f) / 2.0f), ndc_0.z);

            break;
         }

         i_0 = i_0 + int(1);
      }
   }
   else
   {

      inLight_0 = 1.0f;
   }

   target_0 = lambert_0 * lerp(0.64999997615814209f, 1.0f, inLight_0) * entryPointParams_diffuse_0.Sample(entryPointParams_sampler_0, _S1);
   return;
}

the for loop should be simplified to make fxc work properly.

Extension version: 2.0.3
VS Code version: Code 1.107.0 (618725e67565b290ba4da6fe2d29f8fa1d4e3622, 2025-12-10T07:43:47.883Z)
OS version: Windows_NT x64 10.0.19045
Modes: Unsupported
Remote OS version: Linux x64 6.6.87.2-microsoft-standard-WSL2
Remote OS version: Linux x64 6.6.87.2-microsoft-standard-WSL2

System Info

Item	Value
CPUs	12th Gen Intel(R) Core(TM) i9-12900K (24 x 3187)
GPU Status	2d_canvas: enabled direct_rendering_display_compositor: disabled_off_ok gpu_compositing: enabled multiple_raster_threads: enabled_on opengl: enabled_on rasterization: enabled raw_draw: disabled_off_ok skia_graphite: disabled_off trees_in_viz: disabled_off video_decode: enabled video_encode: enabled webgl: enabled webgl2: enabled webgpu: enabled webnn: disabled_off
Load (avg)	undefined
Memory (System)	63.75GB (48.71GB free)
Process Argv	--crash-reporter-id 0a3edcfc-5c4b-4c1f-a092-6afdf9563f4b
Screen Reader	no
VM	0%

Item	Value
Remote	WSL: Ubuntu
OS	Linux x64 6.6.87.2-microsoft-standard-WSL2
CPUs	12th Gen Intel(R) Core(TM) i9-12900K (24 x 0)
Memory (System)	31.22GB (28.75GB free)
VM	0%

Item	Value
Remote	WSL: Ubuntu
OS	Linux x64 6.6.87.2-microsoft-standard-WSL2
CPUs	12th Gen Intel(R) Core(TM) i9-12900K (24 x 0)
Memory (System)	31.22GB (28.75GB free)
VM	0%

A/B Experiments

vsliv368:30146709
binariesv615:30325510
nativeloc1:31344060
dwcopilot:31170013
dwoutputs:31242946
copilot_t_ci:31333650
e5gg6876:31282496
pythonrdcb7:31342333
6518g693:31436602
aj953862:31281341
nes-set-on:31351930
6abeh943:31336334
envsactivate1:31353494
cloudbuttont:31379625
todos-1:31405332
42190218_ostrepl:31403339
trigger-command-fix:31379601
use-responses-api:31390855
6j29d282:31407606
2a0ce269:31428710
ec5jj548:31422691
terminalsuggestenabled:31431119
cmp-ext-treat:31426748
cp_cls_c_966_ss:31435507
copilot6169-t3000:31431383
c0683394:31419495
478ah919:31426797
ge8j1254_inline_auto_hint_haiku:31427726
a5gib710:31434435
38bie571_auto:31426784
request_with_suggest:31435827
rename_enabled:31436409
7a04d226_do_not_restore_last_panel_session:31434268
anthropic_thinking_t:31432745
4g4b7572_control_gpt_5_mini:31434516
406hc587_ask_agent:31428393
cp_cls_c_1081:31433293
copilot-nes-callisto-trt:31434412