mewin/glslang
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

HLSL: handle multiple clip/cull semantic IDs

Browse Source 
HLSL allows several variables to be declared.  There are packing rules involved:
e.g, a float3 and a float1 can be packed into a single array[4], while for a
float3 and another float3, the second one will skip the third array entry to
avoid straddling

This is implements that ability.  Because there can be multiple variables involved,
and the final output array will often be a different type altogether (to fuse
the values into a single destination), a new variable is synthesized, unlike the prior
clip/cull support which used the declared variable.  The new variable name is
taken from one of the declared ones, so the old tests are unchanged.

Several new tests are added to test various packing scenarios.

Only two semantic IDs are supported: 0, and 1, per HLSL rules.  This is
encapsulated in

     static const int maxClipCullRegs = 2;

and the algorithm (probably :) ) generalizes to larger values, although there
are a few issues around how HLSL would pack (e.g, would 4 scalars be packed into
a single HLSL float4 out reg?  Probably, and this algorithm assumes so).
This commit is contained in:
LoopDawg
2017-07-05 11:33:06 -06:00
parent cd52fd5a42
commit 307b6507b3
17 changed files with 2257 additions and 430 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
hlsl/hlslParseHelper.h
View File

@@ -39,6 +39,8 @@
#include "../glslang/MachineIndependent/parseVersions.h"
#include "../glslang/MachineIndependent/ParseHelper.h"
#include <array>
namespace glslang {
class TAttributeMap; // forward declare
@@ -91,7 +93,7 @@ public:
TIntermTyped* handleAssign(const TSourceLoc&, TOperator, TIntermTyped* left, TIntermTyped* right);
TIntermTyped* handleAssignToMatrixSwizzle(const TSourceLoc&, TOperator, TIntermTyped* left, TIntermTyped* right);
TIntermTyped* handleFunctionCall(const TSourceLoc&, TFunction*, TIntermTyped*);
TIntermAggregate* assignClipCullDistance(const TSourceLoc&, TOperator, TIntermTyped* left, TIntermTyped* right);
TIntermAggregate* assignClipCullDistance(const TSourceLoc&, TOperator, int semanticId, TIntermTyped* left, TIntermTyped* right);
void decomposeIntrinsic(const TSourceLoc&, TIntermTyped*& node, TIntermNode* arguments);
void decomposeSampleMethods(const TSourceLoc&, TIntermTyped*& node, TIntermNode* arguments);
void decomposeStructBufferMethods(const TSourceLoc&, TIntermTyped*& node, TIntermNode* arguments);
@@ -308,6 +310,10 @@ protected:
// Finalization step: remove unused buffer blocks from linkage (we don't know until the
// shader is entirely compiled)
void removeUnusedStructBufferCounters();
static bool isClipOrCullDistance(TBuiltInVariable);
static bool isClipOrCullDistance(const TQualifier& qual) { return isClipOrCullDistance(qual.builtIn); }
static bool isClipOrCullDistance(const TType& type) { return isClipOrCullDistance(type.getQualifier()); }
// Pass through to base class after remembering builtin mappings.
using TParseContextBase::trackLinkage;
@@ -432,6 +438,13 @@ protected:
TVariable* gsStreamOutput; // geometry shader stream outputs, for emit (Append method)
TVariable* clipDistanceOutput; // synthesized clip distance output variable (shader might have >1)
TVariable* cullDistanceOutput; // synthesized cull distance output variable (shader might have >1)
static const int maxClipCullRegs = 2;
std::array<int, maxClipCullRegs> clipSemanticNSize; // vector, indexed by clip semantic ID
std::array<int, maxClipCullRegs> cullSemanticNSize; // vector, indexed by cull semantic ID
// This tracks the first (mip level) argument to the .mips[][] operator. Since this can be nested as
// in tx.mips[tx.mips[0][1].x][2], we need a stack. We also track the TSourceLoc for error reporting
// purposes.