<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=us-ascii">
<style type="text/css" style="display:none;"> P {margin-top:0;margin-bottom:0;} </style>
</head>
<body dir="ltr">
<div style="font-family: Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
+Scott</div>
<hr style="display:inline-block;width:98%" tabindex="-1">
<div id="divRplyFwdMsg" dir="ltr"><font face="Calibri, sans-serif" style="font-size:11pt" color="#000000"><b>From:</b> Zaara Syeda <syzaara@ca.ibm.com><br>
<b>Sent:</b> Tuesday, July 10, 2018 2:40 PM<br>
<b>To:</b> Zaara Syeda<br>
<b>Cc:</b> kzhuravl_dev@outlook.com; llvm-commits@lists.llvm.org<br>
<b>Subject:</b> Re: [llvm] r336697 - AMDGPU: Make hidden argument metadata consistent with</font>
<div> </div>
</div>
<meta content="text/html; charset=utf-8">
<div>
<div class="x_socmaildefaultfont" dir="ltr" style="font-family:Arial,Helvetica,sans-serif; font-size:10.5pt">
<div dir="ltr">Opps sorry, I meant to reply regarding commit <a href="https://reviews.llvm.org/rL336707">
https://reviews.llvm.org/rL336707</a>.</div>
<div dir="ltr"> </div>
<blockquote dir="ltr" style="border-left:solid #aaaaaa 2px; margin-left:5px; padding-left:5px; direction:ltr; margin-right:0px">
----- Original message -----<br>
From: Zaara Syeda via llvm-commits <llvm-commits@lists.llvm.org><br>
Sent by: "llvm-commits" <llvm-commits-bounces@lists.llvm.org><br>
To: kzhuravl_dev@outlook.com<br>
Cc: llvm-commits@lists.llvm.org<br>
Subject: Re: [llvm] r336697 - AMDGPU: Make hidden argument metadata consistent with<br>
Date: Tue, Jul 10, 2018 2:11 PM<br>
<div dir="ltr" style="font-family:Arial,Helvetica,sans-serif; font-size:10.5pt">
<div dir="ltr">I believe this patch is causing failure on: <a href="http://lab.llvm.org:8011/builders/clang-ppc64le-linux-multistage/builds/7287" target="_blank">
http://lab.llvm.org:8011/builders/clang-ppc64le-linux-multistage/builds/7287</a></div>
<div dir="ltr"><br>
<a href="http://lab.llvm.org:8011/builders/clang-ppc64le-linux-multistage/builds/7287/steps/build%20stage%201/logs/stdio" target="_blank">http://lab.llvm.org:8011/builders/clang-ppc64le-linux-multistage/builds/7287/steps/build%20stage%201/logs/stdio</a></div>
<div dir="ltr"> </div>
<blockquote dir="ltr" style="border-left:solid #aaaaaa 2px; margin-left:5px; padding-left:5px; direction:ltr; margin-right:0px">
----- Original message -----<br>
From: Konstantin Zhuravlyov via llvm-commits <llvm-commits@lists.llvm.org><br>
Sent by: "llvm-commits" <llvm-commits-bounces@lists.llvm.org><br>
To: llvm-commits@lists.llvm.org<br>
Cc:<br>
Subject: [llvm] r336697 - AMDGPU: Make hidden argument metadata consistent with<br>
Date: Tue, Jul 10, 2018 12:17 PM<br>
<div><font size="2" face="Default Monospace,Courier New,Courier,monospace">Author: kzhuravl<br>
Date: Tue Jul 10 09:12:51 2018<br>
New Revision: 336697<br>
<br>
URL: <a href="http://llvm.org/viewvc/llvm-project?rev=336697&view=rev" target="_blank">
http://llvm.org/viewvc/llvm-project?rev=336697&view=rev</a><br>
Log:<br>
AMDGPU: Make hidden argument metadata consistent with<br>
amdgpu-implicitarg-num-bytes attribute<br>
<br>
Differential Revision: <a href="https://reviews.llvm.org/D49096" target="_blank">
https://reviews.llvm.org/D49096</a><br>
<br>
<br>
Modified:<br>
llvm/trunk/lib/Target/AMDGPU/MCTargetDesc/AMDGPUHSAMetadataStreamer.cpp<br>
llvm/trunk/lib/Target/AMDGPU/MCTargetDesc/AMDGPUHSAMetadataStreamer.h<br>
llvm/trunk/test/CodeGen/AMDGPU/hsa-metadata-enqueue-kernel.ll<br>
llvm/trunk/test/CodeGen/AMDGPU/hsa-metadata-from-llvm-ir-full.ll<br>
llvm/trunk/test/CodeGen/AMDGPU/hsa-metadata-hidden-args.ll<br>
<br>
Modified: llvm/trunk/lib/Target/AMDGPU/MCTargetDesc/AMDGPUHSAMetadataStreamer.cpp<br>
URL: <a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/AMDGPU/MCTargetDesc/AMDGPUHSAMetadataStreamer.cpp?rev=336697&r1=336696&r2=336697&view=diff" target="_blank">
http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/AMDGPU/MCTargetDesc/AMDGPUHSAMetadataStreamer.cpp?rev=336697&r1=336696&r2=336697&view=diff</a><br>
==============================================================================<br>
--- llvm/trunk/lib/Target/AMDGPU/MCTargetDesc/AMDGPUHSAMetadataStreamer.cpp (original)<br>
+++ llvm/trunk/lib/Target/AMDGPU/MCTargetDesc/AMDGPUHSAMetadataStreamer.cpp Tue Jul 10 09:12:51 2018<br>
@@ -14,7 +14,7 @@<br>
//===----------------------------------------------------------------------===//<br>
<br>
#include "AMDGPUHSAMetadataStreamer.h"<br>
-#include "AMDGPU.h"<br>
+#include "Utils/AMDGPUBaseInfo.h"<br>
#include "llvm/ADT/StringSwitch.h"<br>
#include "llvm/IR/Constants.h"<br>
#include "llvm/IR/Module.h"<br>
@@ -255,36 +255,7 @@ void MetadataStreamer::emitKernelArgs(co<br>
for (auto &Arg : Func.args())<br>
emitKernelArg(Arg);<br>
<br>
- // TODO: What about other languages?<br>
- if (!Func.getParent()->getNamedMetadata("opencl.ocl.version"))<br>
- return;<br>
-<br>
- auto &DL = Func.getParent()->getDataLayout();<br>
- auto Int64Ty = Type::getInt64Ty(Func.getContext());<br>
-<br>
- emitKernelArg(DL, Int64Ty, ValueKind::HiddenGlobalOffsetX);<br>
- emitKernelArg(DL, Int64Ty, ValueKind::HiddenGlobalOffsetY);<br>
- emitKernelArg(DL, Int64Ty, ValueKind::HiddenGlobalOffsetZ);<br>
-<br>
- auto Int8PtrTy = Type::getInt8PtrTy(Func.getContext(),<br>
- AMDGPUASI.GLOBAL_ADDRESS);<br>
-<br>
- // Emit "printf buffer" argument if printf is used, otherwise emit dummy<br>
- // "none" argument.<br>
- if (Func.getParent()->getNamedMetadata("llvm.printf.fmts"))<br>
- emitKernelArg(DL, Int8PtrTy, ValueKind::HiddenPrintfBuffer);<br>
- else<br>
- emitKernelArg(DL, Int8PtrTy, ValueKind::HiddenNone);<br>
-<br>
- // Emit "default queue" and "completion action" arguments if enqueue kernel is<br>
- // used, otherwise emit dummy "none" arguments.<br>
- if (Func.hasFnAttribute("calls-enqueue-kernel")) {<br>
- emitKernelArg(DL, Int8PtrTy, ValueKind::HiddenDefaultQueue);<br>
- emitKernelArg(DL, Int8PtrTy, ValueKind::HiddenCompletionAction);<br>
- } else {<br>
- emitKernelArg(DL, Int8PtrTy, ValueKind::HiddenNone);<br>
- emitKernelArg(DL, Int8PtrTy, ValueKind::HiddenNone);<br>
- }<br>
+ emitHiddenKernelArgs(Func);<br>
}<br>
<br>
void MetadataStreamer::emitKernelArg(const Argument &Arg) {<br>
@@ -378,6 +349,48 @@ void MetadataStreamer::emitKernelArg(con<br>
}<br>
}<br>
<br>
+void MetadataStreamer::emitHiddenKernelArgs(const Function &Func) {<br>
+ int HiddenArgNumBytes =<br>
+ getIntegerAttribute(Func, "amdgpu-implicitarg-num-bytes", 0);<br>
+<br>
+ if (!HiddenArgNumBytes)<br>
+ return;<br>
+<br>
+ auto &DL = Func.getParent()->getDataLayout();<br>
+ auto Int64Ty = Type::getInt64Ty(Func.getContext());<br>
+<br>
+ if (HiddenArgNumBytes >= 8)<br>
+ emitKernelArg(DL, Int64Ty, ValueKind::HiddenGlobalOffsetX);<br>
+ if (HiddenArgNumBytes >= 16)<br>
+ emitKernelArg(DL, Int64Ty, ValueKind::HiddenGlobalOffsetY);<br>
+ if (HiddenArgNumBytes >= 24)<br>
+ emitKernelArg(DL, Int64Ty, ValueKind::HiddenGlobalOffsetZ);<br>
+<br>
+ auto Int8PtrTy = Type::getInt8PtrTy(Func.getContext(),<br>
+ AMDGPUASI.GLOBAL_ADDRESS);<br>
+<br>
+ // Emit "printf buffer" argument if printf is used, otherwise emit dummy<br>
+ // "none" argument.<br>
+ if (HiddenArgNumBytes >= 32) {<br>
+ if (Func.getParent()->getNamedMetadata("llvm.printf.fmts"))<br>
+ emitKernelArg(DL, Int8PtrTy, ValueKind::HiddenPrintfBuffer);<br>
+ else<br>
+ emitKernelArg(DL, Int8PtrTy, ValueKind::HiddenNone);<br>
+ }<br>
+<br>
+ // Emit "default queue" and "completion action" arguments if enqueue kernel is<br>
+ // used, otherwise emit dummy "none" arguments.<br>
+ if (HiddenArgNumBytes >= 48) {<br>
+ if (Func.hasFnAttribute("calls-enqueue-kernel")) {<br>
+ emitKernelArg(DL, Int8PtrTy, ValueKind::HiddenDefaultQueue);<br>
+ emitKernelArg(DL, Int8PtrTy, ValueKind::HiddenCompletionAction);<br>
+ } else {<br>
+ emitKernelArg(DL, Int8PtrTy, ValueKind::HiddenNone);<br>
+ emitKernelArg(DL, Int8PtrTy, ValueKind::HiddenNone);<br>
+ }<br>
+ }<br>
+}<br>
+<br>
void MetadataStreamer::begin(const Module &Mod) {<br>
AMDGPUASI = getAMDGPUAS(Mod);<br>
emitVersion();<br>
<br>
Modified: llvm/trunk/lib/Target/AMDGPU/MCTargetDesc/AMDGPUHSAMetadataStreamer.h<br>
URL: <a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/AMDGPU/MCTargetDesc/AMDGPUHSAMetadataStreamer.h?rev=336697&r1=336696&r2=336697&view=diff" target="_blank">
http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/AMDGPU/MCTargetDesc/AMDGPUHSAMetadataStreamer.h?rev=336697&r1=336696&r2=336697&view=diff</a><br>
==============================================================================<br>
--- llvm/trunk/lib/Target/AMDGPU/MCTargetDesc/AMDGPUHSAMetadataStreamer.h (original)<br>
+++ llvm/trunk/lib/Target/AMDGPU/MCTargetDesc/AMDGPUHSAMetadataStreamer.h Tue Jul 10 09:12:51 2018<br>
@@ -73,6 +73,8 @@ private:<br>
StringRef BaseTypeName = "", StringRef AccQual = "",<br>
StringRef TypeQual = "");<br>
<br>
+ void emitHiddenKernelArgs(const Function &Func);<br>
+<br>
public:<br>
MetadataStreamer() = default;<br>
~MetadataStreamer() = default;<br>
<br>
Modified: llvm/trunk/test/CodeGen/AMDGPU/hsa-metadata-enqueue-kernel.ll<br>
URL: <a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/AMDGPU/hsa-metadata-enqueue-kernel.ll?rev=336697&r1=336696&r2=336697&view=diff" target="_blank">
http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/AMDGPU/hsa-metadata-enqueue-kernel.ll?rev=336697&r1=336696&r2=336697&view=diff</a><br>
==============================================================================<br>
--- llvm/trunk/test/CodeGen/AMDGPU/hsa-metadata-enqueue-kernel.ll (original)<br>
+++ llvm/trunk/test/CodeGen/AMDGPU/hsa-metadata-enqueue-kernel.ll Tue Jul 10 09:12:51 2018<br>
@@ -32,7 +32,7 @@<br>
; CHECK-NEXT: ValueType: I64<br>
; CHECK-NOT: ValueKind: HiddenDefaultQueue<br>
; CHECK-NOT: ValueKind: HiddenCompletionAction<br>
-define amdgpu_kernel void @test_non_enqueue_kernel_caller(i8 %a)<br>
+define amdgpu_kernel void @test_non_enqueue_kernel_caller(i8 %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !3<br>
!kernel_arg_base_type !3 !kernel_arg_type_qual !4 {<br>
ret void<br>
@@ -77,13 +77,14 @@ define amdgpu_kernel void @test_non_enqu<br>
; CHECK-NEXT: ValueKind: HiddenCompletionAction<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_enqueue_kernel_caller(i8 %a) #0<br>
+define amdgpu_kernel void @test_enqueue_kernel_caller(i8 %a) #1<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !3<br>
!kernel_arg_base_type !3 !kernel_arg_type_qual !4 {<br>
ret void<br>
}<br>
<br>
-attributes #0 = { "calls-enqueue-kernel" }<br>
+attributes #0 = { "amdgpu-implicitarg-num-bytes"="48" }<br>
+attributes #1 = { "calls-enqueue-kernel" "amdgpu-implicitarg-num-bytes"="48" }<br>
<br>
!1 = !{i32 0}<br>
!2 = !{!"none"}<br>
@@ -93,5 +94,4 @@ attributes #0 = { "calls-enqueue-kernel"<br>
!opencl.ocl.version = !{!90}<br>
!90 = !{i32 2, i32 0}<br>
<br>
-<br>
; PARSER: AMDGPU HSA Metadata Parser Test: PASS<br>
<br>
Modified: llvm/trunk/test/CodeGen/AMDGPU/hsa-metadata-from-llvm-ir-full.ll<br>
URL: <a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/AMDGPU/hsa-metadata-from-llvm-ir-full.ll?rev=336697&r1=336696&r2=336697&view=diff" target="_blank">
http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/AMDGPU/hsa-metadata-from-llvm-ir-full.ll?rev=336697&r1=336696&r2=336697&view=diff</a><br>
==============================================================================<br>
--- llvm/trunk/test/CodeGen/AMDGPU/hsa-metadata-from-llvm-ir-full.ll (original)<br>
+++ llvm/trunk/test/CodeGen/AMDGPU/hsa-metadata-from-llvm-ir-full.ll Tue Jul 10 09:12:51 2018<br>
@@ -54,7 +54,7 @@<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
; CHECK-NOT: ValueKind: HiddenDefaultQueue<br>
; CHECK-NOT: ValueKind: HiddenCompletionAction<br>
-define amdgpu_kernel void @test_char(i8 %a)<br>
+define amdgpu_kernel void @test_char(i8 %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !9<br>
!kernel_arg_base_type !9 !kernel_arg_type_qual !4 {<br>
ret void<br>
@@ -89,7 +89,7 @@ define amdgpu_kernel void @test_char(i8<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_ushort2(<2 x i16> %a)<br>
+define amdgpu_kernel void @test_ushort2(<2 x i16> %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !10<br>
!kernel_arg_base_type !10 !kernel_arg_type_qual !4 {<br>
ret void<br>
@@ -124,7 +124,7 @@ define amdgpu_kernel void @test_ushort2(<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_int3(<3 x i32> %a)<br>
+define amdgpu_kernel void @test_int3(<3 x i32> %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !11<br>
!kernel_arg_base_type !11 !kernel_arg_type_qual !4 {<br>
ret void<br>
@@ -159,7 +159,7 @@ define amdgpu_kernel void @test_int3(<3<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_ulong4(<4 x i64> %a)<br>
+define amdgpu_kernel void @test_ulong4(<4 x i64> %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !12<br>
!kernel_arg_base_type !12 !kernel_arg_type_qual !4 {<br>
ret void<br>
@@ -194,7 +194,7 @@ define amdgpu_kernel void @test_ulong4(<<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_half8(<8 x half> %a)<br>
+define amdgpu_kernel void @test_half8(<8 x half> %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !13<br>
!kernel_arg_base_type !13 !kernel_arg_type_qual !4 {<br>
ret void<br>
@@ -229,7 +229,7 @@ define amdgpu_kernel void @test_half8(<8<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_float16(<16 x float> %a)<br>
+define amdgpu_kernel void @test_float16(<16 x float> %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !14<br>
!kernel_arg_base_type !14 !kernel_arg_type_qual !4 {<br>
ret void<br>
@@ -264,7 +264,7 @@ define amdgpu_kernel void @test_float16(<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_double16(<16 x double> %a)<br>
+define amdgpu_kernel void @test_double16(<16 x double> %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !15<br>
!kernel_arg_base_type !15 !kernel_arg_type_qual !4 {<br>
ret void<br>
@@ -300,7 +300,7 @@ define amdgpu_kernel void @test_double16<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_pointer(i32 addrspace(1)* %a)<br>
+define amdgpu_kernel void @test_pointer(i32 addrspace(1)* %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !16<br>
!kernel_arg_base_type !16 !kernel_arg_type_qual !4 {<br>
ret void<br>
@@ -336,7 +336,7 @@ define amdgpu_kernel void @test_pointer(<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_image(%opencl.image2d_t addrspace(1)* %a)<br>
+define amdgpu_kernel void @test_image(%opencl.image2d_t addrspace(1)* %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !17<br>
!kernel_arg_base_type !17 !kernel_arg_type_qual !4 {<br>
ret void<br>
@@ -371,7 +371,7 @@ define amdgpu_kernel void @test_image(%o<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_sampler(i32 %a)<br>
+define amdgpu_kernel void @test_sampler(i32 %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !18<br>
!kernel_arg_base_type !18 !kernel_arg_type_qual !4 {<br>
ret void<br>
@@ -407,7 +407,7 @@ define amdgpu_kernel void @test_sampler(<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_queue(%opencl.queue_t addrspace(1)* %a)<br>
+define amdgpu_kernel void @test_queue(%opencl.queue_t addrspace(1)* %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !19<br>
!kernel_arg_base_type !19 !kernel_arg_type_qual !4 {<br>
ret void<br>
@@ -443,7 +443,7 @@ define amdgpu_kernel void @test_queue(%o<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_struct(%struct.A addrspace(5)* byval %a)<br>
+define amdgpu_kernel void @test_struct(%struct.A addrspace(5)* byval %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !20<br>
!kernel_arg_base_type !20 !kernel_arg_type_qual !4 {<br>
ret void<br>
@@ -478,7 +478,7 @@ define amdgpu_kernel void @test_struct(%<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_i128(i128 %a)<br>
+define amdgpu_kernel void @test_i128(i128 %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !21<br>
!kernel_arg_base_type !21 !kernel_arg_type_qual !4 {<br>
ret void<br>
@@ -527,7 +527,7 @@ define amdgpu_kernel void @test_i128(i12<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_multi_arg(i32 %a, <2 x i16> %b, <3 x i8> %c)<br>
+define amdgpu_kernel void @test_multi_arg(i32 %a, <2 x i16> %b, <3 x i8> %c) #0<br>
!kernel_arg_addr_space !22 !kernel_arg_access_qual !23 !kernel_arg_type !24<br>
!kernel_arg_base_type !24 !kernel_arg_type_qual !25 {<br>
ret void<br>
@@ -582,7 +582,7 @@ define amdgpu_kernel void @test_multi_ar<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
define amdgpu_kernel void @test_addr_space(i32 addrspace(1)* %g,<br>
i32 addrspace(4)* %c,<br>
- i32 addrspace(3)* %l)<br>
+ i32 addrspace(3)* %l) #0<br>
!kernel_arg_addr_space !50 !kernel_arg_access_qual !23 !kernel_arg_type !51<br>
!kernel_arg_base_type !51 !kernel_arg_type_qual !25 {<br>
ret void<br>
@@ -640,7 +640,7 @@ define amdgpu_kernel void @test_addr_spa<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
define amdgpu_kernel void @test_type_qual(i32 addrspace(1)* %a,<br>
i32 addrspace(1)* %b,<br>
- %opencl.pipe_t addrspace(1)* %c)<br>
+ %opencl.pipe_t addrspace(1)* %c) #0<br>
!kernel_arg_addr_space !22 !kernel_arg_access_qual !23 !kernel_arg_type !51<br>
!kernel_arg_base_type !51 !kernel_arg_type_qual !70 {<br>
ret void<br>
@@ -694,7 +694,7 @@ define amdgpu_kernel void @test_type_qua<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
define amdgpu_kernel void @test_access_qual(%opencl.image1d_t addrspace(1)* %ro,<br>
%opencl.image2d_t addrspace(1)* %wo,<br>
- %opencl.image3d_t addrspace(1)* %rw)<br>
+ %opencl.image3d_t addrspace(1)* %rw) #0<br>
!kernel_arg_addr_space !60 !kernel_arg_access_qual !61 !kernel_arg_type !62<br>
!kernel_arg_base_type !62 !kernel_arg_type_qual !25 {<br>
ret void<br>
@@ -731,7 +731,7 @@ define amdgpu_kernel void @test_access_q<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_vec_type_hint_half(i32 %a)<br>
+define amdgpu_kernel void @test_vec_type_hint_half(i32 %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !3<br>
!kernel_arg_base_type !3 !kernel_arg_type_qual !4 !vec_type_hint !26 {<br>
ret void<br>
@@ -768,7 +768,7 @@ define amdgpu_kernel void @test_vec_type<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_vec_type_hint_float(i32 %a)<br>
+define amdgpu_kernel void @test_vec_type_hint_float(i32 %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !3<br>
!kernel_arg_base_type !3 !kernel_arg_type_qual !4 !vec_type_hint !27 {<br>
ret void<br>
@@ -805,7 +805,7 @@ define amdgpu_kernel void @test_vec_type<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_vec_type_hint_double(i32 %a)<br>
+define amdgpu_kernel void @test_vec_type_hint_double(i32 %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !3<br>
!kernel_arg_base_type !3 !kernel_arg_type_qual !4 !vec_type_hint !28 {<br>
ret void<br>
@@ -842,7 +842,7 @@ define amdgpu_kernel void @test_vec_type<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_vec_type_hint_char(i32 %a)<br>
+define amdgpu_kernel void @test_vec_type_hint_char(i32 %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !3<br>
!kernel_arg_base_type !3 !kernel_arg_type_qual !4 !vec_type_hint !29 {<br>
ret void<br>
@@ -879,7 +879,7 @@ define amdgpu_kernel void @test_vec_type<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_vec_type_hint_short(i32 %a)<br>
+define amdgpu_kernel void @test_vec_type_hint_short(i32 %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !3<br>
!kernel_arg_base_type !3 !kernel_arg_type_qual !4 !vec_type_hint !30 {<br>
ret void<br>
@@ -916,7 +916,7 @@ define amdgpu_kernel void @test_vec_type<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_vec_type_hint_long(i32 %a)<br>
+define amdgpu_kernel void @test_vec_type_hint_long(i32 %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !3<br>
!kernel_arg_base_type !3 !kernel_arg_type_qual !4 !vec_type_hint !31 {<br>
ret void<br>
@@ -953,7 +953,7 @@ define amdgpu_kernel void @test_vec_type<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_vec_type_hint_unknown(i32 %a)<br>
+define amdgpu_kernel void @test_vec_type_hint_unknown(i32 %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !3<br>
!kernel_arg_base_type !3 !kernel_arg_type_qual !4 !vec_type_hint !32 {<br>
ret void<br>
@@ -991,7 +991,7 @@ define amdgpu_kernel void @test_vec_type<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_reqd_wgs_vec_type_hint(i32 %a)<br>
+define amdgpu_kernel void @test_reqd_wgs_vec_type_hint(i32 %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !3<br>
!kernel_arg_base_type !3 !kernel_arg_type_qual !4 !vec_type_hint !5<br>
!reqd_work_group_size !6 {<br>
@@ -1030,7 +1030,7 @@ define amdgpu_kernel void @test_reqd_wgs<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_wgs_hint_vec_type_hint(i32 %a)<br>
+define amdgpu_kernel void @test_wgs_hint_vec_type_hint(i32 %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !3<br>
!kernel_arg_base_type !3 !kernel_arg_type_qual !4 !vec_type_hint !7<br>
!work_group_size_hint !8 {<br>
@@ -1067,7 +1067,7 @@ define amdgpu_kernel void @test_wgs_hint<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_arg_ptr_to_ptr(i32 addrspace(5)* addrspace(1)* %a)<br>
+define amdgpu_kernel void @test_arg_ptr_to_ptr(i32 addrspace(5)* addrspace(1)* %a) #0<br>
!kernel_arg_addr_space !81 !kernel_arg_access_qual !2 !kernel_arg_type !80<br>
!kernel_arg_base_type !80 !kernel_arg_type_qual !4 {<br>
ret void<br>
@@ -1103,7 +1103,7 @@ define amdgpu_kernel void @test_arg_ptr_<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_arg_struct_contains_ptr(%struct.B addrspace(5)* byval %a)<br>
+define amdgpu_kernel void @test_arg_struct_contains_ptr(%struct.B addrspace(5)* byval %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !82<br>
!kernel_arg_base_type !82 !kernel_arg_type_qual !4 {<br>
ret void<br>
@@ -1138,7 +1138,7 @@ define amdgpu_kernel void @test_arg_stru<br>
; CHECK-NEXT: ValueKind: HiddenPrintfBuffer<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_arg_vector_of_ptr(<2 x i32 addrspace(1)*> %a)<br>
+define amdgpu_kernel void @test_arg_vector_of_ptr(<2 x i32 addrspace(1)*> %a) #0<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !83<br>
!kernel_arg_base_type !83 !kernel_arg_type_qual !4 {<br>
ret void<br>
@@ -1175,7 +1175,7 @@ define amdgpu_kernel void @test_arg_vect<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
define amdgpu_kernel void @test_arg_unknown_builtin_type(<br>
- %opencl.clk_event_t addrspace(1)* %a)<br>
+ %opencl.clk_event_t addrspace(1)* %a) #0<br>
!kernel_arg_addr_space !81 !kernel_arg_access_qual !2 !kernel_arg_type !84<br>
!kernel_arg_base_type !84 !kernel_arg_type_qual !4 {<br>
ret void<br>
@@ -1279,7 +1279,7 @@ define amdgpu_kernel void @test_pointee_<br>
<4 x i8> addrspace(3)* %e,<br>
<8 x i8> addrspace(3)* %f,<br>
<16 x i8> addrspace(3)* %g,<br>
- {} addrspace(3)* %h)<br>
+ {} addrspace(3)* %h) #0<br>
!kernel_arg_addr_space !91 !kernel_arg_access_qual !92 !kernel_arg_type !93<br>
!kernel_arg_base_type !93 !kernel_arg_type_qual !94 {<br>
ret void<br>
@@ -1383,7 +1383,7 @@ define amdgpu_kernel void @test_pointee_<br>
<4 x i8> addrspace(3)* align 256 %e,<br>
<8 x i8> addrspace(3)* align 128 %f,<br>
<16 x i8> addrspace(3)* align 1024 %g,<br>
- {} addrspace(3)* align 16 %h)<br>
+ {} addrspace(3)* align 16 %h) #0<br>
!kernel_arg_addr_space !91 !kernel_arg_access_qual !92 !kernel_arg_type !93<br>
!kernel_arg_base_type !93 !kernel_arg_type_qual !94 {<br>
ret void<br>
@@ -1422,7 +1422,7 @@ define amdgpu_kernel void @test_pointee_<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
define amdgpu_kernel void @__test_block_invoke_kernel(<br>
- <{ i32, i32, i8*, i8 addrspace(1)*, i8 }> %arg) #0<br>
+ <{ i32, i32, i8*, i8 addrspace(1)*, i8 }> %arg) #1<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !110<br>
!kernel_arg_base_type !110 !kernel_arg_type_qual !4 {<br>
ret void<br>
@@ -1467,14 +1467,15 @@ define amdgpu_kernel void @__test_block_<br>
; CHECK-NEXT: ValueKind: HiddenCompletionAction<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test_enqueue_kernel_caller(i8 %a) #1<br>
+define amdgpu_kernel void @test_enqueue_kernel_caller(i8 %a) #2<br>
!kernel_arg_addr_space !1 !kernel_arg_access_qual !2 !kernel_arg_type !9<br>
!kernel_arg_base_type !9 !kernel_arg_type_qual !4 {<br>
ret void<br>
}<br>
<br>
-attributes #0 = { "runtime-handle"="__test_block_invoke_kernel_runtime_handle" }<br>
-attributes #1 = { "calls-enqueue-kernel" }<br>
+attributes #0 = { "amdgpu-implicitarg-num-bytes"="48" }<br>
+attributes #1 = { "amdgpu-implicitarg-num-bytes"="48" "runtime-handle"="__test_block_invoke_kernel_runtime_handle" }<br>
+attributes #2 = { "amdgpu-implicitarg-num-bytes"="48" "calls-enqueue-kernel" }<br>
<br>
!llvm.printf.fmts = !{!100, !101}<br>
<br>
<br>
Modified: llvm/trunk/test/CodeGen/AMDGPU/hsa-metadata-hidden-args.ll<br>
URL: <a href="http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/AMDGPU/hsa-metadata-hidden-args.ll?rev=336697&r1=336696&r2=336697&view=diff" target="_blank">
http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/AMDGPU/hsa-metadata-hidden-args.ll?rev=336697&r1=336696&r2=336697&view=diff</a><br>
==============================================================================<br>
--- llvm/trunk/test/CodeGen/AMDGPU/hsa-metadata-hidden-args.ll (original)<br>
+++ llvm/trunk/test/CodeGen/AMDGPU/hsa-metadata-hidden-args.ll Tue Jul 10 09:12:51 2018<br>
@@ -6,8 +6,168 @@<br>
; CHECK: Version: [ 1, 0 ]<br>
; CHECK: Kernels:<br>
<br>
-; CHECK: - Name: test<br>
-; CHECK: SymbolName: 'test@kd'<br>
+; CHECK: - Name: test0<br>
+; CHECK: SymbolName: 'test0@kd'<br>
+; CHECK: Args:<br>
+; CHECK-NEXT: - Name: r<br>
+; CHECK-NEXT: Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: GlobalBuffer<br>
+; CHECK-NEXT: ValueType: F16<br>
+; CHECK-NEXT: AddrSpaceQual: Global<br>
+; CHECK-NEXT: - Name: a<br>
+; CHECK-NEXT: Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: GlobalBuffer<br>
+; CHECK-NEXT: ValueType: F16<br>
+; CHECK-NEXT: AddrSpaceQual: Global<br>
+; CHECK-NEXT: - Name: b<br>
+; CHECK-NEXT: Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: GlobalBuffer<br>
+; CHECK-NEXT: ValueType: F16<br>
+; CHECK-NEXT: AddrSpaceQual: Global<br>
+; CHECK-NEXT: CodeProps:<br>
+define amdgpu_kernel void @test0(<br>
+ half addrspace(1)* %r,<br>
+ half addrspace(1)* %a,<br>
+ half addrspace(1)* %b) {<br>
+entry:<br>
+ %a.val = load half, half addrspace(1)* %a<br>
+ %b.val = load half, half addrspace(1)* %b<br>
+ %r.val = fadd half %a.val, %b.val<br>
+ store half %r.val, half addrspace(1)* %r<br>
+ ret void<br>
+}<br>
+<br>
+; CHECK: - Name: test8<br>
+; CHECK: SymbolName: 'test8@kd'<br>
+; CHECK: Args:<br>
+; CHECK-NEXT: - Name: r<br>
+; CHECK-NEXT: Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: GlobalBuffer<br>
+; CHECK-NEXT: ValueType: F16<br>
+; CHECK-NEXT: AddrSpaceQual: Global<br>
+; CHECK-NEXT: - Name: a<br>
+; CHECK-NEXT: Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: GlobalBuffer<br>
+; CHECK-NEXT: ValueType: F16<br>
+; CHECK-NEXT: AddrSpaceQual: Global<br>
+; CHECK-NEXT: - Name: b<br>
+; CHECK-NEXT: Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: GlobalBuffer<br>
+; CHECK-NEXT: ValueType: F16<br>
+; CHECK-NEXT: AddrSpaceQual: Global<br>
+; CHECK-NEXT: - Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: HiddenGlobalOffsetX<br>
+; CHECK-NEXT: ValueType: I64<br>
+; CHECK-NEXT: CodeProps:<br>
+define amdgpu_kernel void @test8(<br>
+ half addrspace(1)* %r,<br>
+ half addrspace(1)* %a,<br>
+ half addrspace(1)* %b) #0 {<br>
+entry:<br>
+ %a.val = load half, half addrspace(1)* %a<br>
+ %b.val = load half, half addrspace(1)* %b<br>
+ %r.val = fadd half %a.val, %b.val<br>
+ store half %r.val, half addrspace(1)* %r<br>
+ ret void<br>
+}<br>
+<br>
+; CHECK: - Name: test16<br>
+; CHECK: SymbolName: 'test16@kd'<br>
+; CHECK: Args:<br>
+; CHECK-NEXT: - Name: r<br>
+; CHECK-NEXT: Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: GlobalBuffer<br>
+; CHECK-NEXT: ValueType: F16<br>
+; CHECK-NEXT: AddrSpaceQual: Global<br>
+; CHECK-NEXT: - Name: a<br>
+; CHECK-NEXT: Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: GlobalBuffer<br>
+; CHECK-NEXT: ValueType: F16<br>
+; CHECK-NEXT: AddrSpaceQual: Global<br>
+; CHECK-NEXT: - Name: b<br>
+; CHECK-NEXT: Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: GlobalBuffer<br>
+; CHECK-NEXT: ValueType: F16<br>
+; CHECK-NEXT: AddrSpaceQual: Global<br>
+; CHECK-NEXT: - Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: HiddenGlobalOffsetX<br>
+; CHECK-NEXT: ValueType: I64<br>
+; CHECK-NEXT: - Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: HiddenGlobalOffsetY<br>
+; CHECK-NEXT: ValueType: I64<br>
+; CHECK-NEXT: CodeProps:<br>
+define amdgpu_kernel void @test16(<br>
+ half addrspace(1)* %r,<br>
+ half addrspace(1)* %a,<br>
+ half addrspace(1)* %b) #1 {<br>
+entry:<br>
+ %a.val = load half, half addrspace(1)* %a<br>
+ %b.val = load half, half addrspace(1)* %b<br>
+ %r.val = fadd half %a.val, %b.val<br>
+ store half %r.val, half addrspace(1)* %r<br>
+ ret void<br>
+}<br>
+<br>
+; CHECK: - Name: test24<br>
+; CHECK: SymbolName: 'test24@kd'<br>
+; CHECK: Args:<br>
+; CHECK-NEXT: - Name: r<br>
+; CHECK-NEXT: Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: GlobalBuffer<br>
+; CHECK-NEXT: ValueType: F16<br>
+; CHECK-NEXT: AddrSpaceQual: Global<br>
+; CHECK-NEXT: - Name: a<br>
+; CHECK-NEXT: Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: GlobalBuffer<br>
+; CHECK-NEXT: ValueType: F16<br>
+; CHECK-NEXT: AddrSpaceQual: Global<br>
+; CHECK-NEXT: - Name: b<br>
+; CHECK-NEXT: Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: GlobalBuffer<br>
+; CHECK-NEXT: ValueType: F16<br>
+; CHECK-NEXT: AddrSpaceQual: Global<br>
+; CHECK-NEXT: - Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: HiddenGlobalOffsetX<br>
+; CHECK-NEXT: ValueType: I64<br>
+; CHECK-NEXT: - Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: HiddenGlobalOffsetY<br>
+; CHECK-NEXT: ValueType: I64<br>
+; CHECK-NEXT: - Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: HiddenGlobalOffsetZ<br>
+; CHECK-NEXT: ValueType: I64<br>
+; CHECK-NEXT: CodeProps:<br>
+define amdgpu_kernel void @test24(<br>
+ half addrspace(1)* %r,<br>
+ half addrspace(1)* %a,<br>
+ half addrspace(1)* %b) #2 {<br>
+entry:<br>
+ %a.val = load half, half addrspace(1)* %a<br>
+ %b.val = load half, half addrspace(1)* %b<br>
+ %r.val = fadd half %a.val, %b.val<br>
+ store half %r.val, half addrspace(1)* %r<br>
+ ret void<br>
+}<br>
+<br>
+; CHECK: - Name: test32<br>
+; CHECK: SymbolName: 'test32@kd'<br>
; CHECK: Args:<br>
; CHECK-NEXT: - Name: r<br>
; CHECK-NEXT: Size: 8<br>
@@ -44,6 +204,52 @@<br>
; CHECK-NEXT: ValueKind: HiddenNone<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
+; CHECK-NEXT: CodeProps:<br>
+define amdgpu_kernel void @test32(<br>
+ half addrspace(1)* %r,<br>
+ half addrspace(1)* %a,<br>
+ half addrspace(1)* %b) #3 {<br>
+entry:<br>
+ %a.val = load half, half addrspace(1)* %a<br>
+ %b.val = load half, half addrspace(1)* %b<br>
+ %r.val = fadd half %a.val, %b.val<br>
+ store half %r.val, half addrspace(1)* %r<br>
+ ret void<br>
+}<br>
+<br>
+; CHECK: - Name: test48<br>
+; CHECK: SymbolName: 'test48@kd'<br>
+; CHECK: Args:<br>
+; CHECK-NEXT: - Name: r<br>
+; CHECK-NEXT: Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: GlobalBuffer<br>
+; CHECK-NEXT: ValueType: F16<br>
+; CHECK-NEXT: AddrSpaceQual: Global<br>
+; CHECK-NEXT: - Name: a<br>
+; CHECK-NEXT: Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: GlobalBuffer<br>
+; CHECK-NEXT: ValueType: F16<br>
+; CHECK-NEXT: AddrSpaceQual: Global<br>
+; CHECK-NEXT: - Name: b<br>
+; CHECK-NEXT: Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: GlobalBuffer<br>
+; CHECK-NEXT: ValueType: F16<br>
+; CHECK-NEXT: AddrSpaceQual: Global<br>
+; CHECK-NEXT: - Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: HiddenGlobalOffsetX<br>
+; CHECK-NEXT: ValueType: I64<br>
+; CHECK-NEXT: - Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: HiddenGlobalOffsetY<br>
+; CHECK-NEXT: ValueType: I64<br>
+; CHECK-NEXT: - Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: HiddenGlobalOffsetZ<br>
+; CHECK-NEXT: ValueType: I64<br>
; CHECK-NEXT: - Size: 8<br>
; CHECK-NEXT: Align: 8<br>
; CHECK-NEXT: ValueKind: HiddenNone<br>
@@ -54,10 +260,16 @@<br>
; CHECK-NEXT: ValueKind: HiddenNone<br>
; CHECK-NEXT: ValueType: I8<br>
; CHECK-NEXT: AddrSpaceQual: Global<br>
-define amdgpu_kernel void @test(<br>
+; CHECK-NEXT: - Size: 8<br>
+; CHECK-NEXT: Align: 8<br>
+; CHECK-NEXT: ValueKind: HiddenNone<br>
+; CHECK-NEXT: ValueType: I8<br>
+; CHECK-NEXT: AddrSpaceQual: Global<br>
+; CHECK-NEXT: CodeProps:<br>
+define amdgpu_kernel void @test48(<br>
half addrspace(1)* %r,<br>
half addrspace(1)* %a,<br>
- half addrspace(1)* %b) {<br>
+ half addrspace(1)* %b) #4 {<br>
entry:<br>
%a.val = load half, half addrspace(1)* %a<br>
%b.val = load half, half addrspace(1)* %b<br>
@@ -66,5 +278,8 @@ entry:<br>
ret void<br>
}<br>
<br>
-!opencl.ocl.version = !{!0}<br>
-!0 = !{i32 2, i32 0}<br>
+attributes #0 = { "amdgpu-implicitarg-num-bytes"="8" }<br>
+attributes #1 = { "amdgpu-implicitarg-num-bytes"="16" }<br>
+attributes #2 = { "amdgpu-implicitarg-num-bytes"="24" }<br>
+attributes #3 = { "amdgpu-implicitarg-num-bytes"="32" }<br>
+attributes #4 = { "amdgpu-implicitarg-num-bytes"="48" }<br>
<br>
<br>
_______________________________________________<br>
llvm-commits mailing list<br>
llvm-commits@lists.llvm.org<br>
<a href="http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits" target="_blank">http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits</a></font><br>
</div>
</blockquote>
<div dir="ltr"> </div>
</div>
<div><font size="2" face="Default Monospace,Courier New,Courier,monospace">_______________________________________________<br>
llvm-commits mailing list<br>
llvm-commits@lists.llvm.org<br>
<a href="http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits" target="_blank">http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits</a></font></div>
</blockquote>
<div dir="ltr"> </div>
</div>
<br>
</div>
</body>
</html>