[AMDGPU] calling conventions for AMDPAL OS type

Summary:
This commit adds comments on how the AMDPAL OS type overloads the
existing AMDGPU_ calling conventions used by Mesa, and adds a couple of
new ones.

Reviewers: arsenm, nhaehnle, dstuttard

Subscribers: mehdi_amini, kzhuravl, wdng, yaxunl, t-tye, llvm-commits

Differential Revision: https://reviews.llvm.org/D37752

llvm-svn: 314502

include/llvm/IR/CallingConv.h

@@ -183,16 +183,18 @@ namespace CallingConv {
     /// which have an "optimized" convention to preserve registers.
     AVR_BUILTIN = 86,
 
-    /// Calling convention used for Mesa vertex shaders.
+    /// Calling convention used for Mesa vertex shaders, or AMDPAL last shader
+    /// stage before rasterization (vertex shader if tessellation and geometry
+    /// are not in use, or otherwise copy shader if one is needed).
     AMDGPU_VS = 87,
 
-    /// Calling convention used for Mesa geometry shaders.
+    /// Calling convention used for Mesa/AMDPAL geometry shaders.
     AMDGPU_GS = 88,
 
-    /// Calling convention used for Mesa pixel shaders.
+    /// Calling convention used for Mesa/AMDPAL pixel shaders.
     AMDGPU_PS = 89,
 
-    /// Calling convention used for Mesa compute shaders.
+    /// Calling convention used for Mesa/AMDPAL compute shaders.
     AMDGPU_CS = 90,
 
     /// Calling convention for AMDGPU code object kernels.
@@ -201,14 +203,23 @@ namespace CallingConv {
     /// Register calling convention used for parameters transfer optimization
     X86_RegCall = 92,
 
-    /// Calling convention used for Mesa hull shaders. (= tessellation control
+    /// Calling convention used for Mesa/AMDPAL hull shaders (= tessellation
-    /// shaders)
+    /// control shaders).
     AMDGPU_HS = 93,
 
     /// Calling convention used for special MSP430 rtlib functions
     /// which have an "optimized" convention using additional registers.
     MSP430_BUILTIN = 94,
 
+    /// Calling convention used for AMDPAL vertex shader if tessellation is in
+    /// use.
+    AMDGPU_LS = 95,
+
+    /// Calling convention used for AMDPAL shader stage before geometry shader
+    /// if geometry is in use. So either the domain (= tessellation evaluation)
+    /// shader if tessellation is in use, or otherwise the vertex shader.
+    AMDGPU_ES = 96,
+
     /// The highest possible calling convention ID. Must be some 2^k - 1.
     MaxID = 1023
   };

lib/AsmParser/LLLexer.cpp

@@ -601,7 +601,9 @@ lltok::Kind LLLexer::LexIdentifier() {
   KEYWORD(hhvm_ccc);
   KEYWORD(cxx_fast_tlscc);
   KEYWORD(amdgpu_vs);
+  KEYWORD(amdgpu_ls);
   KEYWORD(amdgpu_hs);
+  KEYWORD(amdgpu_es);
   KEYWORD(amdgpu_gs);
   KEYWORD(amdgpu_ps);
   KEYWORD(amdgpu_cs);

lib/AsmParser/LLParser.cpp

@@ -1692,7 +1692,9 @@ void LLParser::ParseOptionalDLLStorageClass(unsigned &Res) {
 ///   ::= 'hhvm_ccc'
 ///   ::= 'cxx_fast_tlscc'
 ///   ::= 'amdgpu_vs'
+///   ::= 'amdgpu_ls'
 ///   ::= 'amdgpu_hs'
+///   ::= 'amdgpu_es'
 ///   ::= 'amdgpu_gs'
 ///   ::= 'amdgpu_ps'
 ///   ::= 'amdgpu_cs'
@@ -1734,7 +1736,9 @@ bool LLParser::ParseOptionalCallingConv(unsigned &CC) {
   case lltok::kw_hhvm_ccc:       CC = CallingConv::HHVM_C; break;
   case lltok::kw_cxx_fast_tlscc: CC = CallingConv::CXX_FAST_TLS; break;
   case lltok::kw_amdgpu_vs:      CC = CallingConv::AMDGPU_VS; break;
+  case lltok::kw_amdgpu_ls:      CC = CallingConv::AMDGPU_LS; break;
   case lltok::kw_amdgpu_hs:      CC = CallingConv::AMDGPU_HS; break;
+  case lltok::kw_amdgpu_es:      CC = CallingConv::AMDGPU_ES; break;
   case lltok::kw_amdgpu_gs:      CC = CallingConv::AMDGPU_GS; break;
   case lltok::kw_amdgpu_ps:      CC = CallingConv::AMDGPU_PS; break;
   case lltok::kw_amdgpu_cs:      CC = CallingConv::AMDGPU_CS; break;

lib/AsmParser/LLToken.h

@@ -153,7 +153,9 @@ enum Kind {
   kw_hhvm_ccc,
   kw_cxx_fast_tlscc,
   kw_amdgpu_vs,
+  kw_amdgpu_ls,
   kw_amdgpu_hs,
+  kw_amdgpu_es,
   kw_amdgpu_gs,
   kw_amdgpu_ps,
   kw_amdgpu_cs,

lib/IR/AsmWriter.cpp

@@ -373,7 +373,9 @@ static void PrintCallingConv(unsigned cc, raw_ostream &Out) {
   case CallingConv::HHVM:          Out << "hhvmcc"; break;
   case CallingConv::HHVM_C:        Out << "hhvm_ccc"; break;
   case CallingConv::AMDGPU_VS:     Out << "amdgpu_vs"; break;
+  case CallingConv::AMDGPU_LS:     Out << "amdgpu_ls"; break;
   case CallingConv::AMDGPU_HS:     Out << "amdgpu_hs"; break;
+  case CallingConv::AMDGPU_ES:     Out << "amdgpu_es"; break;
   case CallingConv::AMDGPU_GS:     Out << "amdgpu_gs"; break;
   case CallingConv::AMDGPU_PS:     Out << "amdgpu_ps"; break;
   case CallingConv::AMDGPU_CS:     Out << "amdgpu_cs"; break;

lib/Target/AMDGPU/AMDGPUAliasAnalysis.cpp

@@ -129,8 +129,11 @@ bool AMDGPUAAResult::pointsToConstantMemory(const MemoryLocation &Loc,
     switch (F->getCallingConv()) {
     default:
       return AAResultBase::pointsToConstantMemory(Loc, OrLocal);
-    case CallingConv::AMDGPU_VS:
+    case CallingConv::AMDGPU_LS:
+    case CallingConv::AMDGPU_HS:
+    case CallingConv::AMDGPU_ES:
     case CallingConv::AMDGPU_GS:
+    case CallingConv::AMDGPU_VS:
     case CallingConv::AMDGPU_PS:
     case CallingConv::AMDGPU_CS:
     case CallingConv::AMDGPU_KERNEL:

lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp

@@ -865,10 +865,12 @@ static unsigned getRsrcReg(CallingConv::ID CallConv) {
   switch (CallConv) {
   default: LLVM_FALLTHROUGH;
   case CallingConv::AMDGPU_CS: return R_00B848_COMPUTE_PGM_RSRC1;
+  case CallingConv::AMDGPU_LS: return R_00B528_SPI_SHADER_PGM_RSRC1_LS;
   case CallingConv::AMDGPU_HS: return R_00B428_SPI_SHADER_PGM_RSRC1_HS;
+  case CallingConv::AMDGPU_ES: return R_00B328_SPI_SHADER_PGM_RSRC1_ES;
   case CallingConv::AMDGPU_GS: return R_00B228_SPI_SHADER_PGM_RSRC1_GS;
-  case CallingConv::AMDGPU_PS: return R_00B028_SPI_SHADER_PGM_RSRC1_PS;
   case CallingConv::AMDGPU_VS: return R_00B128_SPI_SHADER_PGM_RSRC1_VS;
+  case CallingConv::AMDGPU_PS: return R_00B028_SPI_SHADER_PGM_RSRC1_PS;
   }
 }
 

lib/Target/AMDGPU/AMDGPUISelLowering.cpp

@@ -848,6 +848,8 @@ CCAssignFn *AMDGPUCallLowering::CCAssignFnForCall(CallingConv::ID CC,
   case CallingConv::AMDGPU_PS:
   case CallingConv::AMDGPU_CS:
   case CallingConv::AMDGPU_HS:
+  case CallingConv::AMDGPU_ES:
+  case CallingConv::AMDGPU_LS:
     return CC_AMDGPU;
   case CallingConv::C:
   case CallingConv::Fast:
@@ -869,6 +871,8 @@ CCAssignFn *AMDGPUCallLowering::CCAssignFnForReturn(CallingConv::ID CC,
   case CallingConv::AMDGPU_PS:
   case CallingConv::AMDGPU_CS:
   case CallingConv::AMDGPU_HS:
+  case CallingConv::AMDGPU_ES:
+  case CallingConv::AMDGPU_LS:
     return RetCC_SI_Shader;
   case CallingConv::C:
   case CallingConv::Fast:

lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp

@@ -491,7 +491,9 @@ static bool isArgPassedInSGPR(const Argument *A) {
   case CallingConv::SPIR_KERNEL:
     return true;
   case CallingConv::AMDGPU_VS:
+  case CallingConv::AMDGPU_LS:
   case CallingConv::AMDGPU_HS:
+  case CallingConv::AMDGPU_ES:
   case CallingConv::AMDGPU_GS:
   case CallingConv::AMDGPU_PS:
   case CallingConv::AMDGPU_CS:

lib/Target/AMDGPU/SIDefines.h

@@ -375,7 +375,9 @@ enum SDWA9EncValues{
 #define   S_00B02C_EXTRA_LDS_SIZE(x)                                  (((x) & 0xFF) << 8)
 #define R_00B128_SPI_SHADER_PGM_RSRC1_VS                                0x00B128
 #define R_00B228_SPI_SHADER_PGM_RSRC1_GS                                0x00B228
+#define R_00B328_SPI_SHADER_PGM_RSRC1_ES                                0x00B328
 #define R_00B428_SPI_SHADER_PGM_RSRC1_HS                                0x00B428
+#define R_00B528_SPI_SHADER_PGM_RSRC1_LS                                0x00B528
 #define R_00B848_COMPUTE_PGM_RSRC1                                      0x00B848
 #define   S_00B028_VGPRS(x)                                           (((x) & 0x3F) << 0)
 #define   S_00B028_SGPRS(x)                                           (((x) & 0x0F) << 6)

lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp

@@ -486,7 +486,9 @@ unsigned getInitialPSInputAddr(const Function &F) {
 bool isShader(CallingConv::ID cc) {
   switch(cc) {
     case CallingConv::AMDGPU_VS:
+    case CallingConv::AMDGPU_LS:
     case CallingConv::AMDGPU_HS:
+    case CallingConv::AMDGPU_ES:
     case CallingConv::AMDGPU_GS:
     case CallingConv::AMDGPU_PS:
     case CallingConv::AMDGPU_CS:
@@ -508,7 +510,9 @@ bool isEntryFunctionCC(CallingConv::ID CC) {
   case CallingConv::AMDGPU_GS:
   case CallingConv::AMDGPU_PS:
   case CallingConv::AMDGPU_CS:
+  case CallingConv::AMDGPU_ES:
   case CallingConv::AMDGPU_HS:
+  case CallingConv::AMDGPU_LS:
     return true;
   default:
     return false;
@@ -744,7 +748,9 @@ bool isArgPassedInSGPR(const Argument *A) {
   case CallingConv::SPIR_KERNEL:
     return true;
   case CallingConv::AMDGPU_VS:
+  case CallingConv::AMDGPU_LS:
   case CallingConv::AMDGPU_HS:
+  case CallingConv::AMDGPU_ES:
   case CallingConv::AMDGPU_GS:
   case CallingConv::AMDGPU_PS:
   case CallingConv::AMDGPU_CS:

test/Bitcode/compatibility.ll

@@ -476,6 +476,14 @@ declare cc93 void @f.cc93()
 ; CHECK: declare amdgpu_hs void @f.cc93()
 declare amdgpu_hs void @f.amdgpu_hs()
 ; CHECK: declare amdgpu_hs void @f.amdgpu_hs()
+declare cc95 void @f.cc95()
+; CHECK: declare amdgpu_ls void @f.cc95()
+declare amdgpu_ls void @f.amdgpu_ls()
+; CHECK: declare amdgpu_ls void @f.amdgpu_ls()
+declare cc96 void @f.cc96()
+; CHECK: declare amdgpu_es void @f.cc96()
+declare amdgpu_es void @f.amdgpu_es()
+; CHECK: declare amdgpu_es void @f.amdgpu_es()
 declare cc1023 void @f.cc1023()
 ; CHECK: declare cc1023 void @f.cc1023()
 

test/CodeGen/AMDGPU/amdpal-cs.ll

@@ -0,0 +1,13 @@
+; RUN: llc -mtriple=amdgcn--amdpal -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=SI -enable-var-scope %s
+; RUN: llc -mtriple=amdgcn--amdpal -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI -enable-var-scope %s
+; RUN: llc -mtriple=amdgcn--amdpal -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GFX9 -enable-var-scope %s
+
+; amdpal compute shader: check for 47176 (COMPUTE_PGM_RSRC1) in .AMDGPU.config
+; GCN-LABEL: .AMDGPU.config
+; GCN: .long  47176
+; GCN-LABEL: {{^}}cs_amdpal:
+define amdgpu_cs half @cs_amdpal(half %arg0) {
+  %add = fadd half %arg0, 1.0
+  ret half %add
+}
+

test/CodeGen/AMDGPU/amdpal-es.ll

@@ -0,0 +1,13 @@
+; RUN: llc -mtriple=amdgcn--amdpal -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=SI %s
+; RUN: llc -mtriple=amdgcn--amdpal -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI %s
+
+; amdpal pixel shader: check for 45864 (SPI_SHADER_PGM_RSRC1_ES) in .AMDGPU.config
+; GCN-LABEL: .AMDGPU.config
+; GCN: .long  45864
+; GCN-LABEL: {{^}}es_amdpal:
+define amdgpu_es half @es_amdpal(half %arg0) {
+  %add = fadd half %arg0, 1.0
+  ret half %add
+}
+
+

test/CodeGen/AMDGPU/amdpal-gs.ll

@@ -0,0 +1,14 @@
+; RUN: llc -mtriple=amdgcn--amdpal -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=SI %s
+; RUN: llc -mtriple=amdgcn--amdpal -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI %s
+; RUN: llc -mtriple=amdgcn--amdpal -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GFX9 -enable-var-scope %s
+
+; amdpal pixel shader: check for 45608 (SPI_SHADER_PGM_RSRC1_GS) in .AMDGPU.config
+; GCN-LABEL: .AMDGPU.config
+; GCN: .long  45608
+; GCN-LABEL: {{^}}gs_amdpal:
+define amdgpu_gs half @gs_amdpal(half %arg0) {
+  %add = fadd half %arg0, 1.0
+  ret half %add
+}
+
+

test/CodeGen/AMDGPU/amdpal-hs.ll

@@ -0,0 +1,14 @@
+; RUN: llc -mtriple=amdgcn--amdpal -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=SI %s
+; RUN: llc -mtriple=amdgcn--amdpal -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI %s
+; RUN: llc -mtriple=amdgcn--amdpal -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GFX9 -enable-var-scope %s
+
+; amdpal pixel shader: check for 46120 (SPI_SHADER_PGM_RSRC1_HS) in .AMDGPU.config
+; GCN-LABEL: .AMDGPU.config
+; GCN: .long  46120
+; GCN-LABEL: {{^}}hs_amdpal:
+define amdgpu_hs half @hs_amdpal(half %arg0) {
+  %add = fadd half %arg0, 1.0
+  ret half %add
+}
+
+

test/CodeGen/AMDGPU/amdpal-ls.ll

@@ -0,0 +1,13 @@
+; RUN: llc -mtriple=amdgcn--amdpal -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=SI %s
+; RUN: llc -mtriple=amdgcn--amdpal -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI %s
+
+; amdpal pixel shader: check for 46376 (SPI_SHADER_PGM_RSRC1_LS) in .AMDGPU.config
+; GCN-LABEL: .AMDGPU.config
+; GCN: .long  46376
+; GCN-LABEL: {{^}}ls_amdpal:
+define amdgpu_ls half @ls_amdpal(half %arg0) {
+  %add = fadd half %arg0, 1.0
+  ret half %add
+}
+
+

test/CodeGen/AMDGPU/amdpal-ps.ll

@@ -0,0 +1,14 @@
+; RUN: llc -mtriple=amdgcn--amdpal -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=SI %s
+; RUN: llc -mtriple=amdgcn--amdpal -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI %s
+; RUN: llc -mtriple=amdgcn--amdpal -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GFX9 -enable-var-scope %s
+
+; amdpal pixel shader: check for 45096 (SPI_SHADER_PGM_RSRC1_PS) in .AMDGPU.config
+; GCN-LABEL: .AMDGPU.config
+; GCN: .long  45096
+; GCN-LABEL: {{^}}ps_amdpal:
+define amdgpu_ps half @ps_amdpal(half %arg0) {
+  %add = fadd half %arg0, 1.0
+  ret half %add
+}
+
+

test/CodeGen/AMDGPU/amdpal-vs.ll

@@ -0,0 +1,14 @@
+; RUN: llc -mtriple=amdgcn--amdpal -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=SI %s
+; RUN: llc -mtriple=amdgcn--amdpal -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI %s
+; RUN: llc -mtriple=amdgcn--amdpal -mcpu=gfx900 -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=GFX9 -enable-var-scope %s
+
+; amdpal pixel shader: check for 45352 (SPI_SHADER_PGM_RSRC1_VS) in .AMDGPU.config
+; GCN-LABEL: .AMDGPU.config
+; GCN: .long  45352
+; GCN-LABEL: {{^}}vs_amdpal:
+define amdgpu_vs half @vs_amdpal(half %arg0) {
+  %add = fadd half %arg0, 1.0
+  ret half %add
+}
+
+

test/CodeGen/AMDGPU/calling-conventions.ll

@@ -76,4 +76,49 @@ define amdgpu_kernel void @call_fastcc() #0 {
   ret void
 }
 
-attributes #0 = { nounwind noinline }
+; Mesa compute shader: check for 47176 (COMPUTE_PGM_RSRC1) in .AMDGPU.config
+; GCN-LABEL: .AMDGPU.config
+; GCN: .long  47176
+; GCN-LABEL: {{^}}cs_mesa:
+define amdgpu_cs half @cs_mesa(half %arg0) {
+  %add = fadd half %arg0, 1.0
+  ret half %add
+}
+
+; Mesa pixel shader: check for 45096 (SPI_SHADER_PGM_RSRC1_PS) in .AMDGPU.config
+; GCN-LABEL: .AMDGPU.config
+; GCN: .long  45096
+; GCN-LABEL: {{^}}ps_mesa:
+define amdgpu_ps half @ps_mesa(half %arg0) {
+  %add = fadd half %arg0, 1.0
+  ret half %add
+}
+
+; Mesa vertex shader: check for 45352 (SPI_SHADER_PGM_RSRC1_VS) in .AMDGPU.config
+; GCN-LABEL: .AMDGPU.config
+; GCN: .long  45352
+; GCN-LABEL: {{^}}vs_mesa:
+define amdgpu_vs half @vs_mesa(half %arg0) {
+  %add = fadd half %arg0, 1.0
+  ret half %add
+}
+
+; Mesa geometry shader: check for 45608 (SPI_SHADER_PGM_RSRC1_GS) in .AMDGPU.config
+; GCN-LABEL: .AMDGPU.config
+; GCN: .long  45608
+; GCN-LABEL: {{^}}gs_mesa:
+define amdgpu_gs half @gs_mesa(half %arg0) {
+  %add = fadd half %arg0, 1.0
+  ret half %add
+}
+
+; Mesa hull shader: check for 46120 (SPI_SHADER_PGM_RSRC1_HS) in .AMDGPU.config
+; GCN-LABEL: .AMDGPU.config
+; GCN: .long  46120
+; GCN-LABEL: {{^}}hs_mesa:
+define amdgpu_hs half @hs_mesa(half %arg0) {
+  %add = fadd half %arg0, 1.0
+  ret half %add
+}
+
+attributes #0 = { nounwind noinline }