[AMDGPU] Add .entry_point back into PAL metadata #124925
Closed
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Member
|
@llvm/pr-subscribers-backend-amdgpu Author: David Stuttard (dstutt) ChangesFull diff: https://github.com/llvm/llvm-project/pull/124925.diff 12 Files Affected:
diff --git a/llvm/lib/Target/AMDGPU/Utils/AMDGPUPALMetadata.cpp b/llvm/lib/Target/AMDGPU/Utils/AMDGPUPALMetadata.cpp
index 1e76bf7056cc81..df919c6b39f084 100644
--- a/llvm/lib/Target/AMDGPU/Utils/AMDGPUPALMetadata.cpp
+++ b/llvm/lib/Target/AMDGPU/Utils/AMDGPUPALMetadata.cpp
@@ -27,6 +27,28 @@
using namespace llvm;
using namespace llvm::AMDGPU;
+// Return the PAL metadata hardware shader stage name.
+static const char *getStageName(CallingConv::ID CC) {
+ switch (CC) {
+ case CallingConv::AMDGPU_PS:
+ return ".ps";
+ case CallingConv::AMDGPU_VS:
+ return ".vs";
+ case CallingConv::AMDGPU_GS:
+ return ".gs";
+ case CallingConv::AMDGPU_ES:
+ return ".es";
+ case CallingConv::AMDGPU_HS:
+ return ".hs";
+ case CallingConv::AMDGPU_LS:
+ return ".ls";
+ case CallingConv::AMDGPU_Gfx:
+ llvm_unreachable("Callable shader has no hardware stage");
+ default:
+ return ".cs";
+ }
+}
+
// Read the PAL metadata from IR metadata, where it was put by the frontend.
void AMDGPUPALMetadata::readFromIR(Module &M) {
auto *NamedMD = M.getNamedMetadata("amdgpu.pal.metadata.msgpack");
@@ -232,8 +254,16 @@ void AMDGPUPALMetadata::setEntryPoint(unsigned CC, StringRef Name) {
if (isLegacy())
return;
// Msgpack format.
+ // Entry point is updated to .entry_point_symbol and is set to the function
+ // name
getHwStage(CC)[".entry_point_symbol"] =
MsgPackDoc.getNode(Name, /*Copy=*/true);
+
+ // Set .entry_point which is defined
+ // to be _amdgpu_<stage> and _amdgpu_cs for non-shader functions
+ std::string EPName("_amdgpu_");
+ EPName += std::string(getStageName(CC) + 1);
+ getHwStage(CC)[".entry_point"] = MsgPackDoc.getNode(EPName, /*Copy=*/true);
}
// Set the number of used vgprs in the metadata. This is an optional
@@ -943,28 +973,6 @@ msgpack::MapDocNode AMDGPUPALMetadata::getGraphicsRegisters() {
return GraphicsRegisters.getMap();
}
-// Return the PAL metadata hardware shader stage name.
-static const char *getStageName(CallingConv::ID CC) {
- switch (CC) {
- case CallingConv::AMDGPU_PS:
- return ".ps";
- case CallingConv::AMDGPU_VS:
- return ".vs";
- case CallingConv::AMDGPU_GS:
- return ".gs";
- case CallingConv::AMDGPU_ES:
- return ".es";
- case CallingConv::AMDGPU_HS:
- return ".hs";
- case CallingConv::AMDGPU_LS:
- return ".ls";
- case CallingConv::AMDGPU_Gfx:
- llvm_unreachable("Callable shader has no hardware stage");
- default:
- return ".cs";
- }
-}
-
msgpack::DocNode &AMDGPUPALMetadata::refHwStage() {
auto &N =
MsgPackDoc.getRoot()
diff --git a/llvm/test/CodeGen/AMDGPU/amdpal-cs.ll b/llvm/test/CodeGen/AMDGPU/amdpal-cs.ll
index 0818f607da0a5c..96775f4763e34b 100644
--- a/llvm/test/CodeGen/AMDGPU/amdpal-cs.ll
+++ b/llvm/test/CodeGen/AMDGPU/amdpal-cs.ll
@@ -8,6 +8,7 @@
; GCN-NEXT: amdpal.pipelines:
; GCN-NEXT: - .hardware_stages:
; GCN-NEXT: .cs:
+; GCN-NEXT: .entry_point: _amdgpu_cs
; GCN-NEXT: .entry_point_symbol: cs_amdpal
; GCN-NEXT: .scratch_memory_size: 0
; GCN: .registers:
diff --git a/llvm/test/CodeGen/AMDGPU/amdpal-es.ll b/llvm/test/CodeGen/AMDGPU/amdpal-es.ll
index e37d22c7df372d..1379246c32573c 100644
--- a/llvm/test/CodeGen/AMDGPU/amdpal-es.ll
+++ b/llvm/test/CodeGen/AMDGPU/amdpal-es.ll
@@ -7,6 +7,7 @@
; GCN-NEXT: amdpal.pipelines:
; GCN-NEXT: - .hardware_stages:
; GCN-NEXT: .es:
+; GCN-NEXT: .entry_point: _amdgpu_es
; GCN-NEXT: .entry_point_symbol: es_amdpal
; GCN-NEXT: .scratch_memory_size: 0
; GCN: .registers:
diff --git a/llvm/test/CodeGen/AMDGPU/amdpal-gs.ll b/llvm/test/CodeGen/AMDGPU/amdpal-gs.ll
index d847f75a5c0939..1fba34a50094f0 100644
--- a/llvm/test/CodeGen/AMDGPU/amdpal-gs.ll
+++ b/llvm/test/CodeGen/AMDGPU/amdpal-gs.ll
@@ -8,6 +8,7 @@
; GCN-NEXT: amdpal.pipelines:
; GCN-NEXT: - .hardware_stages:
; GCN-NEXT: .gs:
+; GCN-NEXT: .entry_point: _amdgpu_gs
; GCN-NEXT: .entry_point_symbol: gs_amdpal
; GCN-NEXT: .scratch_memory_size: 0
; GCN: .registers:
diff --git a/llvm/test/CodeGen/AMDGPU/amdpal-hs.ll b/llvm/test/CodeGen/AMDGPU/amdpal-hs.ll
index 74f5f440c99d79..53c6b95f073515 100644
--- a/llvm/test/CodeGen/AMDGPU/amdpal-hs.ll
+++ b/llvm/test/CodeGen/AMDGPU/amdpal-hs.ll
@@ -8,6 +8,7 @@
; GCN-NEXT: amdpal.pipelines:
; GCN-NEXT: - .hardware_stages:
; GCN-NEXT: .hs:
+; GCN-NEXT: .entry_point: _amdgpu_hs
; GCN-NEXT: .entry_point_symbol: hs_amdpal
; GCN-NEXT: .scratch_memory_size: 0
; GCN: .registers:
diff --git a/llvm/test/CodeGen/AMDGPU/amdpal-ls.ll b/llvm/test/CodeGen/AMDGPU/amdpal-ls.ll
index 287cc1201a3c58..ebe753134a42d5 100644
--- a/llvm/test/CodeGen/AMDGPU/amdpal-ls.ll
+++ b/llvm/test/CodeGen/AMDGPU/amdpal-ls.ll
@@ -7,6 +7,7 @@
; GCN-NEXT: amdpal.pipelines:
; GCN-NEXT: - .hardware_stages:
; GCN-NEXT: .ls:
+; GCN-NEXT: .entry_point: _amdgpu_ls
; GCN-NEXT: .entry_point_symbol: ls_amdpal
; GCN-NEXT: .scratch_memory_size: 0
; GCN: .registers:
diff --git a/llvm/test/CodeGen/AMDGPU/amdpal-psenable.ll b/llvm/test/CodeGen/AMDGPU/amdpal-psenable.ll
index e1767182c359df..32f19e2af32ec7 100644
--- a/llvm/test/CodeGen/AMDGPU/amdpal-psenable.ll
+++ b/llvm/test/CodeGen/AMDGPU/amdpal-psenable.ll
@@ -11,6 +11,7 @@
; GCN-NEXT: amdpal.pipelines:
; GCN-NEXT: - .hardware_stages:
; GCN-NEXT: .ps:
+; GCN-NEXT: .entry_point: _amdgpu_ps
; GCN-NEXT: .entry_point_symbol: amdpal_psenable
; GCN-NEXT: .scratch_memory_size: 0
; GCN: .registers:
diff --git a/llvm/test/CodeGen/AMDGPU/amdpal-vs.ll b/llvm/test/CodeGen/AMDGPU/amdpal-vs.ll
index b225d978601ab8..853d221ee3aabd 100644
--- a/llvm/test/CodeGen/AMDGPU/amdpal-vs.ll
+++ b/llvm/test/CodeGen/AMDGPU/amdpal-vs.ll
@@ -8,6 +8,7 @@
; GCN-NEXT: amdpal.pipelines:
; GCN-NEXT: - .hardware_stages:
; GCN-NEXT: .vs:
+; GCN-NEXT: .entry_point: _amdgpu_vs
; GCN-NEXT: .entry_point_symbol: vs_amdpal
; GCN-NEXT: .scratch_memory_size: 0
; GCN: .registers:
diff --git a/llvm/test/CodeGen/AMDGPU/amdpal.ll b/llvm/test/CodeGen/AMDGPU/amdpal.ll
index 97fcf0606b5b78..171df029615ed2 100644
--- a/llvm/test/CodeGen/AMDGPU/amdpal.ll
+++ b/llvm/test/CodeGen/AMDGPU/amdpal.ll
@@ -86,6 +86,7 @@ declare void @llvm.amdgcn.raw.ptr.buffer.store.f32(float, ptr addrspace(8), i32,
; PAL-NEXT: amdpal.pipelines:
; PAL-NEXT: - .hardware_stages:
; PAL-NEXT: .cs:
+; PAL-NEXT: .entry_point: _amdgpu_cs
; PAL-NEXT: .entry_point_symbol: scratch2_cs
; PAL-NEXT: .scratch_memory_size: 0x10
; PAL-NEXT: .sgpr_count: 0x
diff --git a/llvm/test/CodeGen/AMDGPU/elf-notes.ll b/llvm/test/CodeGen/AMDGPU/elf-notes.ll
index e91bed464136f8..b205678bd90850 100644
--- a/llvm/test/CodeGen/AMDGPU/elf-notes.ll
+++ b/llvm/test/CodeGen/AMDGPU/elf-notes.ll
@@ -66,6 +66,7 @@
; OSABI-PAL-ELF: amdpal.pipelines:
; OSABI-PAL-ELF: - .hardware_stages:
; OSABI-PAL-ELF: .cs:
+; OSABI-PAL-ELF: .entry_point: _amdgpu_cs
; OSABI-PAL-ELF: .entry_point_symbol: elf_notes
; OSABI-PAL-ELF: .scratch_memory_size: 0
; OSABI-PAL-ELF: .sgpr_count: 96
diff --git a/llvm/test/CodeGen/AMDGPU/pal-metadata-3.0.ll b/llvm/test/CodeGen/AMDGPU/pal-metadata-3.0.ll
index f89341d539a0f6..7536e83a9da6bd 100644
--- a/llvm/test/CodeGen/AMDGPU/pal-metadata-3.0.ll
+++ b/llvm/test/CodeGen/AMDGPU/pal-metadata-3.0.ll
@@ -53,6 +53,7 @@
; CHECK-NEXT: .cs:
; CHECK-NEXT: .checksum_value: 0x9444d7d0
; CHECK-NEXT: .debug_mode: false
+; CHECK-NEXT: .entry_point: _amdgpu_cs
; CHECK-NEXT: .entry_point_symbol: _amdgpu_cs_main
; CHECK-NEXT: .excp_en: 0
; CHECK-NEXT: .float_mode: 0xc0
@@ -109,6 +110,7 @@
; CHECK-NEXT: .wgp_mode: false
; CHECK-NEXT: .gs:
; CHECK-NEXT: .debug_mode: false
+; CHECK-NEXT: .entry_point: _amdgpu_gs
; CHECK-NEXT: .entry_point_symbol: gs_shader
; CHECK-NEXT: .ieee_mode: false
; CHECK-NEXT: .lds_size: 0x200
@@ -120,6 +122,7 @@
; CHECK-NEXT: .wgp_mode: true
; CHECK-NEXT: .hs:
; CHECK-NEXT: .debug_mode: false
+; CHECK-NEXT: .entry_point: _amdgpu_hs
; CHECK-NEXT: .entry_point_symbol: hs_shader
; CHECK-NEXT: .ieee_mode: false
; CHECK-NEXT: .lds_size: 0x1000
@@ -131,6 +134,7 @@
; CHECK-NEXT: .wgp_mode: true
; CHECK-NEXT: .ps:
; CHECK-NEXT: .debug_mode: false
+; CHECK-NEXT: .entry_point: _amdgpu_ps
; CHECK-NEXT: .entry_point_symbol: ps_shader
; CHECK-NEXT: .ieee_mode: false
; CHECK-NEXT: .lds_size: 0
diff --git a/llvm/test/CodeGen/AMDGPU/wave_dispatch_regs.ll b/llvm/test/CodeGen/AMDGPU/wave_dispatch_regs.ll
index e3a6240aac0057..fdc1e6abb051f2 100644
--- a/llvm/test/CodeGen/AMDGPU/wave_dispatch_regs.ll
+++ b/llvm/test/CodeGen/AMDGPU/wave_dispatch_regs.ll
@@ -12,6 +12,7 @@
; GCN-NEXT: amdpal.pipelines:
; GCN-NEXT: - .hardware_stages:
; GCN-NEXT: .cs:
+; GCN-NEXT: .entry_point: _amdgpu_cs
; GCN-NEXT: .entry_point_symbol: _amdgpu_cs_main
; GCN-NEXT: .scratch_memory_size: 0
; SI-NEXT: .sgpr_count: 0x11
|
arsenm
reviewed
Jan 29, 2025
You can test this locally with the following command:git-clang-format --diff 7441e87fe05376782d0ddb90a13e1756eb1b1976 0384f09a74c3b70ff857b7be4dbcd55433e16c30 --extensions cpp -- llvm/lib/Target/AMDGPU/Utils/AMDGPUPALMetadata.cppView the diff from clang-format here.diff --git a/llvm/lib/Target/AMDGPU/Utils/AMDGPUPALMetadata.cpp b/llvm/lib/Target/AMDGPU/Utils/AMDGPUPALMetadata.cpp
index a61bbe13cb..296031e4a0 100644
--- a/llvm/lib/Target/AMDGPU/Utils/AMDGPUPALMetadata.cpp
+++ b/llvm/lib/Target/AMDGPU/Utils/AMDGPUPALMetadata.cpp
@@ -264,7 +264,8 @@ void AMDGPUPALMetadata::setEntryPoint(unsigned CC, StringRef Name) {
SmallString<16> EPName("_amdgpu_");
raw_svector_ostream EPNameOS(EPName);
EPNameOS << getStageName(CC) + 1;
- getHwStage(CC)[".entry_point"] = MsgPackDoc.getNode(EPNameOS.str(), /*Copy=*/true);
+ getHwStage(CC)[".entry_point"] =
+ MsgPackDoc.getNode(EPNameOS.str(), /*Copy=*/true);
}
// Set the number of used vgprs in the metadata. This is an optional
|
Collaborator
Author
|
This PR seems to be stuck on "Processing updates", so going to close and submit a new one. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.