pkgs/raven-rhel8
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

llvm: 19.1.7

Browse Source
This commit is contained in:
Raven 2025-02-09 14:19:45 +06:00
parent dfed922410
commit 1b61b6800d
18 changed files with 6285 additions and 331 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

19
modular/libclc/libclc.spec
View File

@ -2,7 +2,7 @@
%global debug_package %{nil}
%global shortname clc
%global libclc_version 19.1.2
%global libclc_version 19.1.7
#global rc_ver 4
%global libclc_srcdir libclc-%{libclc_version}%{?rc_ver:-rc%{rc_ver}}.src
@ -94,7 +94,22 @@ export CFLAGS="%{build_cflags} -D__extern_always_inline=inline"
%{_includedir}/%{shortname}
%changelog
* Wed Oct 18 2024 Raven <raven@sysadmins.ws> - 19.1.2-1
* Wed Jan 15 2025 Raven <raven@sysadmins.ws> - 19.1.7-1
- Update to LLVM 19.1.7
* Thu Dec 19 2024 Raven <raven@sysadmins.ws> - 19.1.6-1
- Update to LLVM 19.1.6
* Fri Dec 13 2024 Raven <raven@sysadmins.ws> - 19.1.5-1
- Update to LLVM 19.1.5
* Wed Nov 13 2024 Raven <raven@sysadmins.ws> - 19.1.4-1
- Update to LLVM 19.1.4
* Thu Oct 31 2024 Raven <raven@sysadmins.ws> - 19.1.3-1
- Update to LLVM 19.1.3
* Fri Oct 18 2024 Raven <raven@sysadmins.ws> - 19.1.2-1
- Update to 19.1.2
* Wed Oct 2 2024 Raven <raven@sysadmins.ws> - 19.1.1-1

17
modular/libcxx/libcxx.spec
View File

@ -9,7 +9,7 @@
%endif
%global maj_ver 19
%global libcxx_version %{maj_ver}.1.1
%global libcxx_version %{maj_ver}.1.7
#global rc_ver 4
%global libcxx_srcdir libcxx-%{libcxx_version}%{?rc_ver:-rc%{rc_ver}}.src
%global libcxxabi_srcdir libcxxabi-%{libcxx_version}%{?rc_ver:-rc%{rc_ver}}.src
@ -258,6 +258,21 @@ rm %{buildroot}%{_pkgdocdir}/html/.buildinfo
%doc %{_pkgdocdir}/html
%changelog
* Wed Jan 15 2025 Raven <raven@sysadmins.ws> - 19.1.7-1
- Update to LLVM 19.1.7
* Thu Dec 19 2024 Raven <raven@sysadmins.ws> - 19.1.6-1
- Update to LLVM 19.1.6
* Fri Dec 13 2024 Raven <raven@sysadmins.ws> - 19.1.5-1
- Update to LLVM 19.1.5
* Wed Nov 13 2024 Raven <raven@sysadmins.ws> - 19.1.4-1
- Update to LLVM 19.1.4
* Fri Oct 31 2024 Raven <raven@sysadmins.ws> - 19.1.3-1
- Update to LLVM 19.1.3
* Thu Sep 19 2024 Timm Bäder <tbaeder@redhat.com> - 19.1.0-1
- Update to 19.1.0

29
modular/llvm/0001-18-Always-build-shared-libs-for-LLD.patch Normal file
View File

@ -0,0 +1,29 @@
From b1c60d7fa322a2d208556087df9e7ef94bfbffb8 Mon Sep 17 00:00:00 2001
From: Nikita Popov <npopov@redhat.com>
Date: Wed, 8 May 2024 12:30:36 +0900
Subject: [PATCH] Always build shared libs for LLD
We don't want to enable BUILD_SHARED_LIBS for the whole build,
but we do want to build lld libraries.
---
lld/cmake/modules/AddLLD.cmake | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/lld/cmake/modules/AddLLD.cmake b/lld/cmake/modules/AddLLD.cmake
index 2ee066b41535..270c03f096ac 100644
--- a/lld/cmake/modules/AddLLD.cmake
+++ b/lld/cmake/modules/AddLLD.cmake
@@ -7,9 +7,8 @@ macro(add_lld_library name)
""
""
${ARGN})
- if(ARG_SHARED)
- set(ARG_ENABLE_SHARED SHARED)
- endif()
+ # Always build shared libs for LLD.
+ set(ARG_ENABLE_SHARED SHARED)
llvm_add_library(${name} ${ARG_ENABLE_SHARED} ${ARG_UNPARSED_ARGUMENTS})
set_target_properties(${name} PROPERTIES FOLDER "lld libraries")
--
2.44.0

30
modular/llvm/0001-19-Always-build-shared-libs-for-LLD.patch Normal file
View File

@ -0,0 +1,30 @@
From 69faadbc396000bfa60c722f6fb9c0fc3fb2daf0 Mon Sep 17 00:00:00 2001
From: Nikita Popov <npopov@redhat.com>
Date: Wed, 8 May 2024 12:30:36 +0900
Subject: [PATCH] Always build shared libs for LLD
We don't want to enable BUILD_SHARED_LIBS for the whole build,
but we do want to build lld libraries.
---
lld/cmake/modules/AddLLD.cmake | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/lld/cmake/modules/AddLLD.cmake b/lld/cmake/modules/AddLLD.cmake
index 9f2684b6f933..743ec87814a2 100644
--- a/lld/cmake/modules/AddLLD.cmake
+++ b/lld/cmake/modules/AddLLD.cmake
@@ -7,9 +7,8 @@ macro(add_lld_library name)
""
""
${ARGN})
- if(ARG_SHARED)
- set(ARG_ENABLE_SHARED SHARED)
- endif()
+ # Always build shared libs for LLD.
+ set(ARG_ENABLE_SHARED SHARED)
llvm_add_library(${name} ${ARG_ENABLE_SHARED} ${ARG_UNPARSED_ARGUMENTS})
if (NOT LLVM_INSTALL_TOOLCHAIN_ONLY)
--
2.45.1

41
modular/llvm/0001-19-Remove-myst_parser-dependency-for-RHEL.patch
View File

@ -1,41 +0,0 @@
From d8742e9b361e5fd6fee2298b8ea0aeb4671ec05a Mon Sep 17 00:00:00 2001
From: Nikita Popov <npopov@redhat.com>
Date: Wed, 22 May 2024 09:39:26 +0200
Subject: [PATCH] Remove myst_parser dependency for RHEL
---
clang/docs/conf.py | 3 +--
llvm/docs/conf.py | 3 +--
2 files changed, 2 insertions(+), 4 deletions(-)
diff --git a/clang/docs/conf.py b/clang/docs/conf.py
index 4cee382a718f..d2e2198e05d4 100644
--- a/clang/docs/conf.py
+++ b/clang/docs/conf.py
@@ -43,8 +43,7 @@ try:
extensions.append("myst_parser")
except ImportError:
- if not tags.has("builder-man"):
- raise
+ pass
# The encoding of source files.
diff --git a/llvm/docs/conf.py b/llvm/docs/conf.py
index 7f2ed5309606..354a41f11280 100644
--- a/llvm/docs/conf.py
+++ b/llvm/docs/conf.py
@@ -36,8 +36,7 @@ try:
extensions.append("myst_parser")
except ImportError:
- if not tags.has("builder-man"):
- raise
+ pass
# Automatic anchors for markdown titles
from llvm_slug import make_slug
--
2.44.0

43
modular/llvm/0001-20-Remove-myst_parser-dependency-for-RHEL.patch
View File

@ -1,43 +0,0 @@
From 50cd36c2156d375a6d50f661908b460fbbd22e78 Mon Sep 17 00:00:00 2001
From: Nikita Popov <npopov@redhat.com>
Date: Wed, 22 May 2024 09:39:26 +0200
Subject: [PATCH] Remove myst_parser dependency for RHEL
---
clang/docs/conf.py | 3 +--
llvm/docs/conf.py | 5 +----
2 files changed, 2 insertions(+), 6 deletions(-)
diff --git a/clang/docs/conf.py b/clang/docs/conf.py
index 4cee382a718f..d2e2198e05d4 100644
--- a/clang/docs/conf.py
+++ b/clang/docs/conf.py
@@ -43,8 +43,7 @@ try:
extensions.append("myst_parser")
except ImportError:
- if not tags.has("builder-man"):
- raise
+ pass
# The encoding of source files.
diff --git a/llvm/docs/conf.py b/llvm/docs/conf.py
index d9fa6961032b..e38c009a457d 100644
--- a/llvm/docs/conf.py
+++ b/llvm/docs/conf.py
@@ -36,10 +36,7 @@ try:
extensions.append("myst_parser")
except ImportError:
- if not tags.has("builder-man"):
- raise
-else:
- myst_enable_extensions = ["substitution"]
+ pass
# Automatic anchors for markdown titles
myst_heading_anchors = 6
--
2.46.0

129
modular/llvm/0001-CMake-Add-missing-dependency-108461.patch Normal file
View File

@ -0,0 +1,129 @@
From 17ff6161b83e6a5e86fcb6a13c5551bba1438405 Mon Sep 17 00:00:00 2001
From: Erick Ochoa <github@ceci-nest-pas.me>
Date: Thu, 12 Sep 2024 21:16:58 -0400
Subject: [PATCH 1/9] [CMake] Add missing dependency (#108461)
The [`mlir-capi-execution-engine-test` test
executable](https://github.com/llvm/llvm-project/blob/main/mlir/test/CAPI/CMakeLists.txt#L26-L34)
```cmake
if(MLIR_ENABLE_EXECUTION_ENGINE)
_add_capi_test_executable(mlir-capi-execution-engine-test
execution_engine.c
LINK_LIBS PRIVATE
MLIRCAPIConversion
MLIRCAPIExecutionEngine
MLIRCAPIRegisterEverything
)
endif()
```
is run by lit tests, but it is not properly listed as a dependency. It
is added in places conditionally across the file
[`tests/CMakeLists.txt`](https://github.com/llvm/llvm-project/blob/main/mlir/test/CMakeLists.txt#L130-L143)
```cmake
# The native target may not be enabled, in this case we won't
# run tests that involves executing on the host: do not build
# useless binaries.
if(LLVM_ENABLE_PIC AND TARGET ${LLVM_NATIVE_ARCH})
list(APPEND MLIR_TEST_DEPENDS
mlir-cpu-runner
llc
mlir_async_runtime
mlir-capi-execution-engine-test
mlir_c_runner_utils
mlir_runner_utils
mlir_float16_utils
)
endif()
```
But this condition is not the same as the one where the test executable
is added. [It has been reported on discord that the following error
occurred:](https://discord.com/channels/636084430946959380/642426447167881246/1283811636725022730)
```
FAIL: MLIR :: CAPI/execution_engine.c (2 of 2121)
******************** TEST 'MLIR :: CAPI/execution_engine.c' FAILED ********************
Exit Code: 127
Command Output (stdout):
--
# RUN: at line 10
/usr/bin/mlir-capi-execution-engine-test 2>&1 | /usr/bin/FileCheck /builddir/build/BUILD/mlir-19.1.0_rc4-build/mlir-19.1.0-rc4.src/test/CAPI/execution_engine.c
# executed command: /usr/bin/mlir-capi-execution-engine-test
# .---command stderr------------
# | '/usr/bin/mlir-capi-execution-engine-test': command not found
# `-----------------------------
```
This error will not be deterministic and is dependent on the order in
which tools are built. If by any chance,
`mlir-capi-execution-engine-test` is built before the lit tests run,
then nothing will happen. But lit tests can be run before
`mlir-capi-execution-engine-test` is built.
This patch adds the `mlir-capi-execution-engine` to the
`MLIR_TEST_DEPENDS` list when the `MLIR_ENABLE_EXECUTION_ENGINE` flag is
present.
Happy to make changes like:
* removing `mlir-capi-execution-engine-test` from the other place where
it is included in the tests
* and merge and sort alphabetically these two commands
```cmake
set(MLIR_TEST_DEPENDS
FileCheck count not split-file
mlir-capi-ir-test
mlir-capi-irdl-test
mlir-capi-llvm-test
mlir-capi-pass-test
mlir-capi-quant-test
mlir-capi-rewrite-test
mlir-capi-sparse-tensor-test
mlir-capi-transform-test
mlir-capi-transform-interpreter-test
mlir-capi-translation-test
mlir-linalg-ods-yaml-gen
mlir-lsp-server
mlir-opt
mlir-query
mlir-reduce
mlir-tblgen
mlir-translate
tblgen-lsp-server
tblgen-to-irdl
)
set(MLIR_TEST_DEPENDS ${MLIR_TEST_DEPENDS}
mlir-capi-pdl-test
mlir-pdll-lsp-server
mlir-pdll
)
```
Co-authored-by: Erick Ochoa <erick@ceci-nest-pas.me>
---
mlir/test/CMakeLists.txt | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/mlir/test/CMakeLists.txt b/mlir/test/CMakeLists.txt
index df95e5db11f1..4d2d738b734e 100644
--- a/mlir/test/CMakeLists.txt
+++ b/mlir/test/CMakeLists.txt
@@ -150,6 +150,10 @@ if(MLIR_ENABLE_CUDA_RUNNER)
list(APPEND MLIR_TEST_DEPENDS mlir_cuda_runtime)
endif()
+if(MLIR_ENABLE_EXECUTION_ENGINE)
+ list(APPEND MLIR_TEST_DEPENDS mlir-capi-execution-engine-test)
+endif()
+
if(MLIR_ENABLE_ROCM_RUNNER)
list(APPEND MLIR_TEST_DEPENDS mlir_rocm_runtime)
endif()
--
2.46.0

41
modular/llvm/0001-Driver-Give-devtoolset-path-precedence-over-Installe.patch Normal file
View File

@ -0,0 +1,41 @@
From 73d3b4047d757ef35850e2cef38285b96be82f0f Mon Sep 17 00:00:00 2001
From: Nikita Popov <npopov@redhat.com>
Date: Tue, 23 May 2023 12:17:29 +0200
Subject: [PATCH] [Driver] Give devtoolset path precedence over InstalledDir
This is a followup to the change from c5fe10f365247c3dd9416b7ec8bad73a60b5946e.
While that commit correctly adds the bindir from devtoolset to the
path, the driver dir / install dir still comes first. This means
we'll still end up picking /usr/bin/ld rather than the one from
devtoolset.
Unfortunately, I don't see any way to test this. In the environment
the tests are run, this would only result in a behavior difference
if there is an ld binary present in the LLVM build directory, which
isn't the case.
Differential Revision: https://reviews.llvm.org/D151203
---
clang/lib/Driver/ToolChains/Linux.cpp | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/clang/lib/Driver/ToolChains/Linux.cpp b/clang/lib/Driver/ToolChains/Linux.cpp
index 853ff99d9fe5..aecabb46d4b9 100644
--- a/clang/lib/Driver/ToolChains/Linux.cpp
+++ b/clang/lib/Driver/ToolChains/Linux.cpp
@@ -244,9 +244,9 @@ Linux::Linux(const Driver &D, const llvm::Triple &Triple, const ArgList &Args)
// With devtoolset on RHEL, we want to add a bin directory that is relative
// to the detected gcc install, because if we are using devtoolset gcc then
// we want to use other tools from devtoolset (e.g. ld) instead of the
- // standard system tools.
- PPaths.push_back(Twine(GCCInstallation.getParentLibPath() +
- "/../bin").str());
+ // standard system tools. This should take precedence over InstalledDir.
+ PPaths.insert(PPaths.begin(),
+ Twine(GCCInstallation.getParentLibPath() + "/../bin").str());
if (Arch == llvm::Triple::arm || Arch == llvm::Triple::thumb)
ExtraOpts.push_back("-X");
--
2.40.1

80
modular/llvm/0001-mlir-Specify-deps-via-LLVM_LINK_COMPONENTS.patch Normal file
View File

@ -0,0 +1,80 @@
From c8f93f1958de1f59222a89dd64a573f91105e135 Mon Sep 17 00:00:00 2001
From: Konrad Kleine <kkleine@redhat.com>
Date: Tue, 3 Dec 2024 20:53:29 +0100
Subject: [PATCH] [mlir] Specify deps via `LLVM_LINK_COMPONENTS`
This specifies the dependencies to link against with
`LLVM_LINK_COMPONENTS` for the
`mlir/test/Target/LLVM/MLIRTargetLLVMTests` binary.
Before, the dependencies where directly added to the
`target_link_libraries()` call which caused the problems I describe
next:
When doing a build of LLVM with MLIR I want to link against
`libLLVM.so` instead of statically linking `libLLVMSupport.a`.
MLIR on the other side seems to statically link against
`libLLVMSupport.a` because when I link to the shared library `libLLVM.so` I get:
```
CommandLine Error: Option 'aarch64-ptrauth-auth-checks' registered more than once!
```
This error indicates that the `Support` library is linked twice in the `MLIRTargetLLVMTest` binary.
Here's the creation of the `MLIRTargetLLVMTest` binary before (Notice the
`libLLVMSupport.a`):
```
[6535/6847] : && /usr/bin/clang++ -O2 -flto=thin -ffat-lto-objects -fexceptions -g -grecord-gcc-switches -pipe -Wall -Werror=format-security -Wp,-U_FORTIFY_SOURCE,-D_FORTIFY_SOURCE=3 -Wp,-D_GLIBCXX_ASSERTIONS --config=/usr/lib/rpm/redhat/redhat-hardened-clang.cfg -fstack-protector-strong -mbranch-protection=standard -fasynchronous-unwind-tables -D_DEFAULT_SOURCE -Dasm=__asm__ -fPIC -fno-semantic-interposition -fvisibility-inlines-hidden -Werror=date-time -Werror=unguarded-availability-new -Wall -Wextra -Wno-unused-parameter -Wwrite-strings -Wcast-qual -Wmissing-field-initializers -pedantic -Wno-long-long -Wc++98-compat-extra-semi -Wimplicit-fallthrough -Wcovered-switch-default -Wno-noexcept-type -Wnon-virtual-dtor -Wdelete-non-virtual-dtor -Wsuggest-override -Wstring-conversion -Wmisleading-indentation -Wctad-maybe-unsupported -fdiagnostics-color -ffunction-sections -fdata-sections -Wundef -Werror=mismatched-tags -O2 -g -DNDEBUG -Wl,-z,relro -Wl,--as-needed -Wl,-z,pack-relative-relocs -Wl,-z,now --config=/usr/lib/rpm/redhat/redhat-hardened-clang-ld.cfg -flto=thin -ffat-lto-objects -Wl,--build-id=sha1 -Wl,--gc-sections -fno-lto tools/mlir/unittests/Target/LLVM/CMakeFiles/MLIRTargetLLVMTests.dir/SerializeNVVMTarget.cpp.o tools/mlir/unittests/Target/LLVM/CMakeFiles/MLIRTargetLLVMTests.dir/SerializeROCDLTarget.cpp.o tools/mlir/unittests/Target/LLVM/CMakeFiles/MLIRTargetLLVMTests.dir/SerializeToLLVMBitcode.cpp.o -o tools/mlir/unittests/Target/LLVM/MLIRTargetLLVMTests -Wl,-rpath,/builddir/build/BUILD/llvm-19.1.3-build/llvm-project-19.1.3.src/llvm/redhat-linux-build/lib64 lib64/libllvm_gtest_main.a lib64/libllvm_gtest.a lib64/libMLIRTargetLLVM.a lib64/libMLIRNVVMTarget.a lib64/libMLIRROCDLTarget.a lib64/libMLIRGPUDialect.a lib64/libMLIRNVVMDialect.a lib64/libMLIRLLVMDialect.a lib64/libMLIRLLVMToLLVMIRTranslation.a lib64/libMLIRBuiltinToLLVMIRTranslation.a lib64/libMLIRNVVMToLLVMIRTranslation.a lib64/libMLIRROCDLToLLVMIRTranslation.a lib64/libMLIRGPUToLLVMIRTranslation.a lib64/libLLVMAArch64CodeGen.a lib64/libLLVMAArch64Desc.a lib64/libLLVMAArch64Info.a -lpthread lib64/libMLIRTargetLLVM.a lib64/libMLIRROCDLDialect.a lib64/libMLIRExecutionEngineUtils.a lib64/libMLIRGPUDialect.a lib64/libMLIRMemRefDialect.a lib64/libMLIRArithUtils.a lib64/libMLIRDialectUtils.a lib64/libMLIRComplexDialect.a lib64/libMLIRArithAttrToLLVMConversion.a lib64/libMLIRArithDialect.a lib64/libMLIRCastInterfaces.a lib64/libMLIRDialect.a lib64/libMLIRInferIntRangeCommon.a lib64/libMLIRUBDialect.a lib64/libMLIRShapedOpInterfaces.a lib64/libMLIRTargetLLVMIRExport.a lib64/libMLIRDLTIDialect.a lib64/libMLIRLLVMIRTransforms.a lib64/libMLIRNVVMDialect.a lib64/libMLIRLLVMDialect.a lib64/libMLIRFuncDialect.a lib64/libMLIRTransforms.a lib64/libMLIRMemorySlotInterfaces.a lib64/libMLIRCopyOpInterface.a lib64/libMLIRRuntimeVerifiableOpInterface.a lib64/libMLIRTranslateLib.a lib64/libMLIRParser.a lib64/libMLIRBytecodeReader.a lib64/libMLIRAsmParser.a lib64/libMLIRTransformUtils.a lib64/libMLIRSubsetOpInterface.a lib64/libMLIRValueBoundsOpInterface.a lib64/libMLIRDestinationStyleOpInterface.a lib64/libMLIRRewrite.a lib64/libMLIRRewritePDL.a lib64/libMLIRPDLToPDLInterp.a lib64/libMLIRPass.a lib64/libMLIRAnalysis.a lib64/libMLIRControlFlowInterfaces.a lib64/libMLIRInferIntRangeInterface.a lib64/libMLIRCallInterfaces.a lib64/libMLIRDataLayoutInterfaces.a lib64/libMLIRViewLikeInterface.a lib64/libMLIRLoopLikeInterface.a lib64/libMLIRPresburger.a lib64/libMLIRPDLInterpDialect.a lib64/libMLIRFunctionInterfaces.a lib64/libMLIRPDLDialect.a lib64/libMLIRSideEffectInterfaces.a lib64/libMLIRInferTypeOpInterface.a lib64/libMLIRIR.a lib64/libMLIRSupport.a lib64/libLLVM.so.19.1 lib64/libLLVMAArch64Utils.a lib64/libLLVMAsmPrinter.a lib64/libLLVMCFGuard.a lib64/libLLVMGlobalISel.a lib64/libLLVMSelectionDAG.a lib64/libLLVMCodeGen.a lib64/libLLVMScalarOpts.a lib64/libLLVMAggressiveInstCombine.a lib64/libLLVMInstCombine.a lib64/libLLVMBitWriter.a lib64/libLLVMObjCARCOpts.a lib64/libLLVMCodeGenTypes.a lib64/libLLVMTarget.a lib64/libLLVMVectorize.a lib64/libLLVMTransformUtils.a lib64/libLLVMAnalysis.a lib64/libLLVMProfileData.a lib64/libLLVMSymbolize.a lib64/libLLVMDebugInfoDWARF.a lib64/libLLVMDebugInfoPDB.a lib64/libLLVMObject.a lib64/libLLVMMCParser.a lib64/libLLVMMC.a lib64/libLLVMIRReader.a lib64/libLLVMBitReader.a lib64/libLLVMAsmParser.a lib64/libLLVMTextAPI.a lib64/libLLVMDebugInfoCodeView.a lib64/libLLVMDebugInfoMSF.a lib64/libLLVMDebugInfoBTF.a lib64/libLLVMCore.a lib64/libLLVMBinaryFormat.a lib64/libLLVMRemarks.a lib64/libLLVMBitstreamReader.a lib64/libLLVMTargetParser.a lib64/libLLVMSupport.a lib64/libLLVMDemangle.a -lrt -ldl -lm /usr/lib64/libz.so /usr/lib64/libzstd.so && :
```
Here's the full error:
```
[24/25] cd /builddir/build/BUILD/llvm-19.1.3-build/llvm-project-19.1.3.src/llvm/redhat-linux-build/tools/mlir/test && /usr/bin/python3 /builddir/build/BUILD/llvm-19.1.3-build/llvm-project-19.1.3.src/llvm/redhat-linux-build/./bin/llvm-lit -vv /builddir/build/BUILD/llvm-19.1.3-build/llvm-project-19.1.3.src/llvm/redhat-linux-build/tools/mlir/test
: CommandLine Error: Option 'aarch64-ptrauth-auth-checks' registered more than once!
LLVM ERROR: inconsistency in registered CommandLine options
llvm-lit: /builddir/build/BUILD/llvm-19.1.3-build/llvm-project-19.1.3.src/llvm/utils/lit/lit/formats/googletest.py:38: warning: unable to discover google-tests in '/builddir/build/BUILD/llvm-19.1.3-build/llvm-project-19.1.3.src/llvm/redhat-linux-build/tools/mlir/unittests/Target/LLVM/./MLIRTargetLLVMTests': Command '['/builddir/build/BUILD/llvm-19.1.3-build/llvm-project-19.1.3.src/llvm/redhat-linux-build/tools/mlir/unittests/Target/LLVM/./MLIRTargetLLVMTests', '--gtest_list_tests', '--gtest_filter=-*DISABLED_*']' died with <Signals.SIGABRT: 6>.. Process output: b''
error: filter did not match any tests (of 2704 discovered). Use '--allow-empty-runs' to suppress this error.
FAILED: tools/mlir/test/CMakeFiles/check-mlir /builddir/build/BUILD/llvm-19.1.3-build/llvm-project-19.1.3.src/llvm/redhat-linux-build/tools/mlir/test/CMakeFiles/check-mlir
```
Here's the CMake invocation:
```
/usr/bin/cmake -S . -B redhat-linux-build -DCMAKE_C_FLAGS_RELEASE:STRING=-DNDEBUG -DCMAKE_CXX_FLAGS_RELEASE:STRING=-DNDEBUG -DCMAKE_Fortran_FLAGS_RELEASE:STRING=-DNDEBUG -DCMAKE_VERBOSE_MAKEFILE:BOOL=ON -DCMAKE_INSTALL_DO_STRIP:BOOL=OFF -DCMAKE_INSTALL_PREFIX:PATH=/usr -DINCLUDE_INSTALL_DIR:PATH=/usr/include -DLIB_INSTALL_DIR:PATH=/usr/lib64 -DSYSCONF_INSTALL_DIR:PATH=/etc -DSHARE_INSTALL_PREFIX:PATH=/usr/share -DLIB_SUFFIX=64 -DBUILD_SHARED_LIBS:BOOL=ON -G Ninja '' -DCLANG_BUILD_EXAMPLES:BOOL=OFF -DCLANG_CONFIG_FILE_SYSTEM_DIR=/etc/clang/ -DCLANG_DEFAULT_PIE_ON_LINUX=OFF -DCLANG_DEFAULT_UNWINDLIB=libgcc -DCLANG_ENABLE_ARCMT:BOOL=ON -DCLANG_ENABLE_STATIC_ANALYZER:BOOL=ON -DCLANG_INCLUDE_DOCS:BOOL=ON -DCLANG_INCLUDE_TESTS:BOOL=ON -DCLANG_LINK_CLANG_DYLIB=ON -DCLANG_PLUGIN_SUPPORT:BOOL=ON '-DCLANG_REPOSITORY_STRING=Fedora 19.1.3-5.fc42' -DLLVM_EXTERNAL_CLANG_TOOLS_EXTRA_SOURCE_DIR=../clang-tools-extra -DCLANG_RESOURCE_DIR=../lib/clang/19 -DCOMPILER_RT_INCLUDE_TESTS:BOOL=OFF -DCOMPILER_RT_INSTALL_PATH=/usr/lib/clang/19 -DLLVM_ENABLE_DOXYGEN:BOOL=OFF -DLLVM_ENABLE_SPHINX:BOOL=ON -DLLVM_BUILD_DOCS:BOOL=ON -DSPHINX_EXECUTABLE=/usr/bin/sphinx-build-3 -DSPHINX_OUTPUT_HTML:BOOL=OFF -DSPHINX_OUTPUT_MAN:BOOL=ON -DSPHINX_WARNINGS_AS_ERRORS=OFF -DLLDB_DISABLE_CURSES:BOOL=OFF -DLLDB_DISABLE_LIBEDIT:BOOL=OFF -DLLDB_DISABLE_PYTHON:BOOL=OFF -DLLDB_ENFORCE_STRICT_TEST_REQUIREMENTS:BOOL=ON -DLLVM_APPEND_VC_REV:BOOL=OFF -DLLVM_BUILD_EXAMPLES:BOOL=OFF -DLLVM_BUILD_EXTERNAL_COMPILER_RT:BOOL=ON -DLLVM_BUILD_LLVM_DYLIB:BOOL=ON -DLLVM_BUILD_RUNTIME:BOOL=ON -DLLVM_BUILD_TOOLS:BOOL=ON -DLLVM_BUILD_UTILS:BOOL=ON -DLLVM_COMMON_CMAKE_UTILS=/usr/share/llvm/cmake -DLLVM_DEFAULT_TARGET_TRIPLE=aarch64-redhat-linux-gnu -DLLVM_DYLIB_COMPONENTS=all -DLLVM_ENABLE_EH=ON -DLLVM_ENABLE_FFI:BOOL=ON -DLLVM_ENABLE_LIBCXX:BOOL=OFF -DLLVM_ENABLE_PER_TARGET_RUNTIME_DIR=ON '-DLLVM_ENABLE_PROJECTS=clang;clang-tools-extra;lld;lldb;mlir' -DLLVM_ENABLE_RTTI:BOOL=ON '-DLLVM_ENABLE_RUNTIMES=compiler-rt;openmp;offload' -DLLVM_ENABLE_ZLIB:BOOL=ON -DLLVM_EXPERIMENTAL_TARGETS_TO_BUILD=AVR -DLLVM_INCLUDE_BENCHMARKS=OFF -DLLVM_INCLUDE_EXAMPLES:BOOL=ON -DLLVM_INCLUDE_TOOLS:BOOL=ON -DLLVM_INCLUDE_UTILS:BOOL=ON -DLLVM_INSTALL_TOOLCHAIN_ONLY:BOOL=OFF -DLLVM_INSTALL_UTILS:BOOL=ON -DLLVM_LINK_LLVM_DYLIB:BOOL=ON -DLLVM_PARALLEL_LINK_JOBS=1 -DLLVM_TARGETS_TO_BUILD=all -DLLVM_TOOLS_INSTALL_DIR:PATH=bin -DLLVM_UNREACHABLE_OPTIMIZE:BOOL=OFF -DLLVM_USE_PERF:BOOL=ON -DLLVM_UTILS_INSTALL_DIR:PATH=bin -DMLIR_INCLUDE_DOCS:BOOL=ON -DMLIR_INCLUDE_TESTS:BOOL=ON -DMLIR_INCLUDE_INTEGRATION_TESTS:BOOL=OFF -DMLIR_INSTALL_AGGREGATE_OBJECTS=OFF -DMLIR_BUILD_MLIR_C_DYLIB=ON -DMLIR_ENABLE_BINDINGS_PYTHON:BOOL=ON -DOPENMP_INSTALL_LIBDIR=lib64 -DLIBOMP_INSTALL_ALIASES=OFF -DLLVM_BUILD_TESTS:BOOL=ON -DLLVM_INCLUDE_TESTS:BOOL=ON -DLLVM_INSTALL_GTEST:BOOL=ON -DLLVM_LIT_ARGS=-vv -DLLVM_UNITTEST_LINK_FLAGS=-fno-lto -DBUILD_SHARED_LIBS:BOOL=OFF -DCMAKE_BUILD_TYPE=RelWithDebInfo -DCMAKE_INSTALL_PREFIX=/usr -DENABLE_LINKER_BUILD_ID:BOOL=ON -DOFFLOAD_INSTALL_LIBDIR=lib64 -DPython3_EXECUTABLE=/usr/bin/python3 -DCMAKE_SKIP_INSTALL_RPATH:BOOL=ON -DPPC_LINUX_DEFAULT_IEEELONGDOUBLE=ON -DLLVM_LIBDIR_SUFFIX=64 -DLLVM_BINUTILS_INCDIR=/usr/include -DLLVM_VERSION_SUFFIX=
```
---
mlir/unittests/Target/LLVM/CMakeLists.txt | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/mlir/unittests/Target/LLVM/CMakeLists.txt b/mlir/unittests/Target/LLVM/CMakeLists.txt
index 6d612548a94c..0c61d222dedf 100644
--- a/mlir/unittests/Target/LLVM/CMakeLists.txt
+++ b/mlir/unittests/Target/LLVM/CMakeLists.txt
@@ -1,11 +1,11 @@
+set(LLVM_LINK_COMPONENTS nativecodegen)
+
add_mlir_unittest(MLIRTargetLLVMTests
SerializeNVVMTarget.cpp
SerializeROCDLTarget.cpp
SerializeToLLVMBitcode.cpp
)
-llvm_map_components_to_libnames(llvm_libs nativecodegen)
-
target_link_libraries(MLIRTargetLLVMTests
PRIVATE
MLIRTargetLLVM
@@ -19,7 +19,6 @@ target_link_libraries(MLIRTargetLLVMTests
MLIRNVVMToLLVMIRTranslation
MLIRROCDLToLLVMIRTranslation
MLIRGPUToLLVMIRTranslation
- ${llvm_libs}
)
if (DEFINED LLVM_NATIVE_TARGET)
--
2.46.0

77
modular/llvm/0001-mlir-python-Reuse-the-library-directory.patch Normal file
View File

@ -0,0 +1,77 @@
From 1d043550c7601ca776b0318b9b913e5ecf325baf Mon Sep 17 00:00:00 2001
From: Tulio Magno Quites Machado Filho <tuliom@redhat.com>
Date: Tue, 5 Sep 2023 10:35:37 -0300
Subject: [PATCH] Reuse the library directory
Prefer to get the path to the shared libraries from config.llvm_shlib_dir.
Fallback to the previous path only if config.llvm_shlib_dir is not
defined.
This ensures the test will pass regardless of the build configuration
used downstream.
---
mlir/test/lit.cfg.py | 1 +
mlir/test/python/execution_engine.py | 12 +++++++-----
2 files changed, 8 insertions(+), 5 deletions(-)
diff --git a/mlir/test/lit.cfg.py b/mlir/test/lit.cfg.py
--- a/mlir/test/lit.cfg.py 2024-08-05 10:40:33.000000000 +0200
+++ b/mlir/test/lit.cfg.py 2024-08-14 19:17:55.160470316 +0200
@@ -164,6 +164,7 @@
ToolSubst("transform-opt-ch2", unresolved="ignore"),
ToolSubst("transform-opt-ch3", unresolved="ignore"),
ToolSubst("transform-opt-ch4", unresolved="ignore"),
+ ToolSubst("%llvm_shlib_dir", config.llvm_shlib_dir, unresolved="ignore"),
ToolSubst("mlir-transform-opt", unresolved="ignore"),
ToolSubst("%mlir_lib_dir", config.mlir_lib_dir, unresolved="ignore"),
ToolSubst("%mlir_src_dir", config.mlir_src_root, unresolved="ignore"),
diff -ruN mlir-19.1.0-rc2.src.orig/test/python/execution_engine.py mlir-19.1.0-rc2.src/test/python/execution_engine.py
--- a/mlir/test/python/execution_engine.py 2024-08-05 10:40:33.000000000 +0200
+++ b/mlir/test/python/execution_engine.py 2024-08-14 19:17:20.822219824 +0200
@@ -1,4 +1,4 @@
-# RUN: %PYTHON %s 2>&1 | FileCheck %s
+# RUN: env LLVM_SHLIB_DIR=%llvm_shlib_dir %PYTHON %s 2>&1 | FileCheck %s
# REQUIRES: host-supports-jit
import gc, sys, os, tempfile
from mlir.ir import *
@@ -6,6 +6,9 @@
from mlir.execution_engine import *
from mlir.runtime import *
+_DEFAULT_LIB_DIR = "../../../../lib"
+LIB_DIR = os.getenv("LLVM_SHLIB_DIR", _DEFAULT_LIB_DIR)
+
# Log everything to stderr and flush so that we have a unified stream to match
# errors/info emitted by MLIR to stderr.
@@ -613,6 +616,7 @@
shared_libs = [
"../../../../bin/mlir_runner_utils.dll",
"../../../../bin/mlir_c_runner_utils.dll",
+
]
elif sys.platform == "darwin":
shared_libs = [
@@ -621,8 +625,9 @@
]
else:
shared_libs = [
- "../../../../lib/libmlir_runner_utils.so",
- "../../../../lib/libmlir_c_runner_utils.so",
+ LIB_DIR + "/libmlir_runner_utils.so",
+ LIB_DIR + "/libmlir_c_runner_utils.so",
+
]
execution_engine = ExecutionEngine(
@@ -664,8 +669,8 @@
]
else:
shared_libs = [
- "../../../../lib/libmlir_runner_utils.so",
- "../../../../lib/libmlir_c_runner_utils.so",
+ LIB_DIR + "/libmlir_runner_utils.so",
+ LIB_DIR + "/libmlir_c_runner_utils.so",
]
execution_engine = ExecutionEngine(

62
modular/llvm/0001-openmp-Add-option-to-disable-tsan-tests-111548.patch Normal file
View File

@ -0,0 +1,62 @@
From b2edeb58b8cb3268acee425cd52b406eb60a8095 Mon Sep 17 00:00:00 2001
From: Nikita Popov <npopov@redhat.com>
Date: Wed, 9 Oct 2024 11:29:30 +0200
Subject: [PATCH] [openmp] Add option to disable tsan tests (#111548)
This adds a OPENMP_TEST_ENABLE_TSAN option that allows to override
whether tests using tsan will be enabled. The option defaults to the
existing auto-detection.
The background here is
https://github.com/llvm/llvm-project/issues/111492, where we have some
systems where tsan doesn't work, but we do still want to build it and
run tests that don't use tsan.
---
openmp/cmake/OpenMPTesting.cmake | 3 +++
openmp/tools/archer/tests/CMakeLists.txt | 2 +-
openmp/tools/archer/tests/lit.site.cfg.in | 2 +-
3 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/openmp/cmake/OpenMPTesting.cmake b/openmp/cmake/OpenMPTesting.cmake
index c67ad8b1cbd9..14cc5c67d84c 100644
--- a/openmp/cmake/OpenMPTesting.cmake
+++ b/openmp/cmake/OpenMPTesting.cmake
@@ -163,6 +163,9 @@ else()
set(OPENMP_TEST_COMPILER_HAS_OMIT_FRAME_POINTER_FLAGS 1)
endif()
+set(OPENMP_TEST_ENABLE_TSAN "${OPENMP_TEST_COMPILER_HAS_TSAN_FLAGS}" CACHE BOOL
+ "Whether to enable tests using tsan")
+
# Function to set compiler features for use in lit.
function(update_test_compiler_features)
set(FEATURES "[")
diff --git a/openmp/tools/archer/tests/CMakeLists.txt b/openmp/tools/archer/tests/CMakeLists.txt
index 5de91148fa4b..412c7d63725e 100644
--- a/openmp/tools/archer/tests/CMakeLists.txt
+++ b/openmp/tools/archer/tests/CMakeLists.txt
@@ -28,7 +28,7 @@ macro(pythonize_bool var)
endmacro()
pythonize_bool(LIBARCHER_HAVE_LIBATOMIC)
-pythonize_bool(OPENMP_TEST_COMPILER_HAS_TSAN_FLAGS)
+pythonize_bool(OPENMP_TEST_ENABLE_TSAN)
set(ARCHER_TSAN_TEST_DEPENDENCE "")
if(TARGET tsan)
diff --git a/openmp/tools/archer/tests/lit.site.cfg.in b/openmp/tools/archer/tests/lit.site.cfg.in
index 55edfde9738e..ddcb7b8bc3a5 100644
--- a/openmp/tools/archer/tests/lit.site.cfg.in
+++ b/openmp/tools/archer/tests/lit.site.cfg.in
@@ -12,7 +12,7 @@ config.omp_library_dir = "@LIBOMP_LIBRARY_DIR@"
config.omp_header_dir = "@LIBOMP_INCLUDE_DIR@"
config.operating_system = "@CMAKE_SYSTEM_NAME@"
config.has_libatomic = @LIBARCHER_HAVE_LIBATOMIC@
-config.has_tsan = @OPENMP_TEST_COMPILER_HAS_TSAN_FLAGS@
+config.has_tsan = @OPENMP_TEST_ENABLE_TSAN@
config.test_archer_flags = "@LIBARCHER_TEST_FLAGS@"
config.libarcher_obj_root = "@CMAKE_CURRENT_BINARY_DIR@"
--
2.46.0

205
modular/llvm/0001-openmp-Use-core_siblings_list-if-physical_package_id.patch Normal file
View File

@ -0,0 +1,205 @@
From 5fb4d7f6079a76b2907ccc8c53c7c509c30a3dca Mon Sep 17 00:00:00 2001
From: Nikita Popov <npopov@redhat.com>
Date: Thu, 10 Oct 2024 12:47:33 +0000
Subject: [PATCH] [openmp] Use core_siblings_list if physical_package_id not
available
On powerpc, physical_package_id may not be available. Currently,
this causes openmp to fall back to flat topology and various
affinity tests fail.
Fix this by parsing core_siblings_list to deterimine which cpus
belong to the same socket. This matches what the testing code
does. The code to parse the CPU list format thankfully already
exists.
Fixes https://github.com/llvm/llvm-project/issues/111809.
---
openmp/runtime/src/kmp_affinity.cpp | 100 +++++++++++++------
openmp/runtime/test/affinity/kmp-hw-subset.c | 2 +-
2 files changed, 72 insertions(+), 30 deletions(-)
diff --git a/openmp/runtime/src/kmp_affinity.cpp b/openmp/runtime/src/kmp_affinity.cpp
index cf5cad04eb57..c3d5ecf1345e 100644
--- a/openmp/runtime/src/kmp_affinity.cpp
+++ b/openmp/runtime/src/kmp_affinity.cpp
@@ -1589,15 +1589,13 @@ kmp_str_buf_t *__kmp_affinity_str_buf_mask(kmp_str_buf_t *buf,
return buf;
}
-// Return (possibly empty) affinity mask representing the offline CPUs
-// Caller must free the mask
-kmp_affin_mask_t *__kmp_affinity_get_offline_cpus() {
- kmp_affin_mask_t *offline;
- KMP_CPU_ALLOC(offline);
- KMP_CPU_ZERO(offline);
+static kmp_affin_mask_t *__kmp_parse_cpu_list(const char *path) {
+ kmp_affin_mask_t *mask;
+ KMP_CPU_ALLOC(mask);
+ KMP_CPU_ZERO(mask);
#if KMP_OS_LINUX
int n, begin_cpu, end_cpu;
- kmp_safe_raii_file_t offline_file;
+ kmp_safe_raii_file_t file;
auto skip_ws = [](FILE *f) {
int c;
do {
@@ -1606,29 +1604,29 @@ kmp_affin_mask_t *__kmp_affinity_get_offline_cpus() {
if (c != EOF)
ungetc(c, f);
};
- // File contains CSV of integer ranges representing the offline CPUs
+ // File contains CSV of integer ranges representing the CPUs
// e.g., 1,2,4-7,9,11-15
- int status = offline_file.try_open("/sys/devices/system/cpu/offline", "r");
+ int status = file.try_open(path, "r");
if (status != 0)
- return offline;
- while (!feof(offline_file)) {
- skip_ws(offline_file);
- n = fscanf(offline_file, "%d", &begin_cpu);
+ return mask;
+ while (!feof(file)) {
+ skip_ws(file);
+ n = fscanf(file, "%d", &begin_cpu);
if (n != 1)
break;
- skip_ws(offline_file);
- int c = fgetc(offline_file);
+ skip_ws(file);
+ int c = fgetc(file);
if (c == EOF || c == ',') {
// Just single CPU
end_cpu = begin_cpu;
} else if (c == '-') {
// Range of CPUs
- skip_ws(offline_file);
- n = fscanf(offline_file, "%d", &end_cpu);
+ skip_ws(file);
+ n = fscanf(file, "%d", &end_cpu);
if (n != 1)
break;
- skip_ws(offline_file);
- c = fgetc(offline_file); // skip ','
+ skip_ws(file);
+ c = fgetc(file); // skip ','
} else {
// Syntax problem
break;
@@ -1638,13 +1636,19 @@ kmp_affin_mask_t *__kmp_affinity_get_offline_cpus() {
end_cpu >= __kmp_xproc || begin_cpu > end_cpu) {
continue;
}
- // Insert [begin_cpu, end_cpu] into offline mask
+ // Insert [begin_cpu, end_cpu] into mask
for (int cpu = begin_cpu; cpu <= end_cpu; ++cpu) {
- KMP_CPU_SET(cpu, offline);
+ KMP_CPU_SET(cpu, mask);
}
}
#endif
- return offline;
+ return mask;
+}
+
+// Return (possibly empty) affinity mask representing the offline CPUs
+// Caller must free the mask
+kmp_affin_mask_t *__kmp_affinity_get_offline_cpus() {
+ return __kmp_parse_cpu_list("/sys/devices/system/cpu/offline");
}
// Return the number of available procs
@@ -3175,6 +3179,37 @@ static inline const char *__kmp_cpuinfo_get_envvar() {
return envvar;
}
+static bool __kmp_package_id_from_core_siblings_list(unsigned **threadInfo,
+ unsigned num_avail,
+ unsigned idx) {
+ if (!KMP_AFFINITY_CAPABLE())
+ return false;
+
+ char path[256];
+ KMP_SNPRINTF(path, sizeof(path),
+ "/sys/devices/system/cpu/cpu%u/topology/core_siblings_list",
+ threadInfo[idx][osIdIndex]);
+ kmp_affin_mask_t *siblings = __kmp_parse_cpu_list(path);
+ for (unsigned i = 0; i < num_avail; ++i) {
+ unsigned cpu_id = threadInfo[i][osIdIndex];
+ KMP_ASSERT(cpu_id < __kmp_affin_mask_size * CHAR_BIT);
+ if (!KMP_CPU_ISSET(cpu_id, siblings))
+ continue;
+ if (threadInfo[i][pkgIdIndex] == UINT_MAX) {
+ // Arbitrarily pick the first index we encounter, it only matters that
+ // the value is the same for all siblings.
+ threadInfo[i][pkgIdIndex] = idx;
+ } else if (threadInfo[i][pkgIdIndex] != idx) {
+ // Contradictory sibling lists.
+ KMP_CPU_FREE(siblings);
+ return false;
+ }
+ }
+ KMP_ASSERT(threadInfo[idx][pkgIdIndex] != UINT_MAX);
+ KMP_CPU_FREE(siblings);
+ return true;
+}
+
// Parse /proc/cpuinfo (or an alternate file in the same format) to obtain the
// affinity map. On AIX, the map is obtained through system SRAD (Scheduler
// Resource Allocation Domain).
@@ -3550,18 +3585,13 @@ static bool __kmp_affinity_create_cpuinfo_map(int *line,
return false;
}
- // Check for missing fields. The osId field must be there, and we
- // currently require that the physical id field is specified, also.
+ // Check for missing fields. The osId field must be there. The physical
+ // id field will be checked later.
if (threadInfo[num_avail][osIdIndex] == UINT_MAX) {
CLEANUP_THREAD_INFO;
*msg_id = kmp_i18n_str_MissingProcField;
return false;
}
- if (threadInfo[0][pkgIdIndex] == UINT_MAX) {
- CLEANUP_THREAD_INFO;
- *msg_id = kmp_i18n_str_MissingPhysicalIDField;
- return false;
- }
// Skip this proc if it is not included in the machine model.
if (KMP_AFFINITY_CAPABLE() &&
@@ -3591,6 +3621,18 @@ static bool __kmp_affinity_create_cpuinfo_map(int *line,
}
*line = 0;
+ // At least on powerpc, Linux may return -1 for physical_package_id. Try
+ // to reconstruct topology from core_siblings_list in that case.
+ for (i = 0; i < num_avail; ++i) {
+ if (threadInfo[i][pkgIdIndex] == UINT_MAX) {
+ if (!__kmp_package_id_from_core_siblings_list(threadInfo, num_avail, i)) {
+ CLEANUP_THREAD_INFO;
+ *msg_id = kmp_i18n_str_MissingPhysicalIDField;
+ return false;
+ }
+ }
+ }
+
#if KMP_MIC && REDUCE_TEAM_SIZE
unsigned teamSize = 0;
#endif // KMP_MIC && REDUCE_TEAM_SIZE
diff --git a/openmp/runtime/test/affinity/kmp-hw-subset.c b/openmp/runtime/test/affinity/kmp-hw-subset.c
index 606fcdfbada9..0b49969bd3b1 100644
--- a/openmp/runtime/test/affinity/kmp-hw-subset.c
+++ b/openmp/runtime/test/affinity/kmp-hw-subset.c
@@ -25,7 +25,7 @@ static int compare_hw_subset_places(const place_list_t *openmp_places,
expected_per_place = nthreads_per_core;
} else {
expected_total = nsockets;
- expected_per_place = ncores_per_socket;
+ expected_per_place = ncores_per_socket * nthreads_per_core;
}
if (openmp_places->num_places != expected_total) {
fprintf(stderr, "error: KMP_HW_SUBSET did not half each resource layer!\n");
--
2.47.0

86
modular/llvm/0001-profile-Use-base-vaddr-for-__llvm_write_binary_ids-n.patch Normal file
View File

@ -0,0 +1,86 @@
From ccc2b792e57d632bc887b226a4e7f0a8189eab8b Mon Sep 17 00:00:00 2001
From: Josh Stone <jistone@redhat.com>
Date: Mon, 4 Nov 2024 16:37:49 -0800
Subject: [PATCH] [profile] Use base+vaddr for `__llvm_write_binary_ids` note
pointers
This function is always examining its own ELF headers in memory, but it
was trying to use conditions between examining files or memory, and it
wasn't accounting for LOAD offsets at runtime. This is especially bad if
a loaded segment has additional padding that's not in the file offsets.
Now we do a first scan of the program headers to figure out the runtime
base address based on `PT_PHDR` and/or `PT_DYNAMIC` (else assume zero),
similar to libc's `do_start`. Then each `PT_NOTE` pointer is simply the
base plus the segments's `pt_vaddr`, which includes LOAD offsets.
Fixes #114605
---
.../lib/profile/InstrProfilingPlatformLinux.c | 40 ++++++++-----------
1 file changed, 16 insertions(+), 24 deletions(-)
diff --git a/compiler-rt/lib/profile/InstrProfilingPlatformLinux.c b/compiler-rt/lib/profile/InstrProfilingPlatformLinux.c
index e2c06d51e0c6..c365129a0768 100644
--- a/compiler-rt/lib/profile/InstrProfilingPlatformLinux.c
+++ b/compiler-rt/lib/profile/InstrProfilingPlatformLinux.c
@@ -194,41 +194,33 @@ static int WriteBinaryIds(ProfDataWriter *Writer, const ElfW(Nhdr) * Note,
*/
COMPILER_RT_VISIBILITY int __llvm_write_binary_ids(ProfDataWriter *Writer) {
extern const ElfW(Ehdr) __ehdr_start __attribute__((visibility("hidden")));
+ extern ElfW(Dyn) _DYNAMIC[] __attribute__((weak, visibility("hidden")));
+
const ElfW(Ehdr) *ElfHeader = &__ehdr_start;
const ElfW(Phdr) *ProgramHeader =
(const ElfW(Phdr) *)((uintptr_t)ElfHeader + ElfHeader->e_phoff);
+ /* Compute the added base address in case of position-independent code. */
+ uintptr_t Base = 0;
+ for (uint32_t I = 0; I < ElfHeader->e_phnum; I++) {
+ if (ProgramHeader[I].p_type == PT_PHDR)
+ Base = (uintptr_t)ProgramHeader - ProgramHeader[I].p_vaddr;
+ if (ProgramHeader[I].p_type == PT_DYNAMIC && _DYNAMIC)
+ Base = (uintptr_t)_DYNAMIC - ProgramHeader[I].p_vaddr;
+ }
+
int TotalBinaryIdsSize = 0;
- uint32_t I;
/* Iterate through entries in the program header. */
- for (I = 0; I < ElfHeader->e_phnum; I++) {
+ for (uint32_t I = 0; I < ElfHeader->e_phnum; I++) {
/* Look for the notes segment in program header entries. */
if (ProgramHeader[I].p_type != PT_NOTE)
continue;
/* There can be multiple notes segment, and examine each of them. */
- const ElfW(Nhdr) * Note;
- const ElfW(Nhdr) * NotesEnd;
- /*
- * When examining notes in file, use p_offset, which is the offset within
- * the elf file, to find the start of notes.
- */
- if (ProgramHeader[I].p_memsz == 0 ||
- ProgramHeader[I].p_memsz == ProgramHeader[I].p_filesz) {
- Note = (const ElfW(Nhdr) *)((uintptr_t)ElfHeader +
- ProgramHeader[I].p_offset);
- NotesEnd = (const ElfW(Nhdr) *)((const char *)(Note) +
- ProgramHeader[I].p_filesz);
- } else {
- /*
- * When examining notes in memory, use p_vaddr, which is the address of
- * section after loaded to memory, to find the start of notes.
- */
- Note =
- (const ElfW(Nhdr) *)((uintptr_t)ElfHeader + ProgramHeader[I].p_vaddr);
- NotesEnd =
- (const ElfW(Nhdr) *)((const char *)(Note) + ProgramHeader[I].p_memsz);
- }
+ const ElfW(Nhdr) *Note =
+ (const ElfW(Nhdr) *)(Base + ProgramHeader[I].p_vaddr);
+ const ElfW(Nhdr) *NotesEnd =
+ (const ElfW(Nhdr) *)((const char *)(Note) + ProgramHeader[I].p_memsz);
int BinaryIdsSize = WriteBinaryIds(Writer, Note, NotesEnd);
if (TotalBinaryIdsSize == -1)
--
2.47.0

893
modular/llvm/18-99273.patch Normal file
View File

@ -0,0 +1,893 @@
From 91052169960477fbc39169c10f9fae3bec732510 Mon Sep 17 00:00:00 2001
From: Carl Ritson <carl.ritson@amd.com>
Date: Wed, 17 Jul 2024 15:07:42 +0900
Subject: [PATCH 1/3] [AMDGPU] Implement workaround for GFX11.5 export priority
On GFX11.5 shaders having completed exports need to execute/wait
at a lower priority than shaders still executing exports.
Add code to maintain normal priority of 2 for shaders that export
and drop to priority 0 after exports.
---
llvm/lib/Target/AMDGPU/AMDGPU.td | 15 +-
.../lib/Target/AMDGPU/GCNHazardRecognizer.cpp | 112 ++++++
llvm/lib/Target/AMDGPU/GCNHazardRecognizer.h | 1 +
llvm/lib/Target/AMDGPU/GCNSubtarget.h | 3 +
.../AMDGPU/required-export-priority.ll | 344 ++++++++++++++++++
.../AMDGPU/required-export-priority.mir | 293 +++++++++++++++
6 files changed, 765 insertions(+), 3 deletions(-)
create mode 100644 llvm/test/CodeGen/AMDGPU/required-export-priority.ll
create mode 100644 llvm/test/CodeGen/AMDGPU/required-export-priority.mir
diff --git a/llvm/lib/Target/AMDGPU/AMDGPU.td b/llvm/lib/Target/AMDGPU/AMDGPU.td
index dfc8eaea66f7b..14fcf6a210a78 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPU.td
+++ b/llvm/lib/Target/AMDGPU/AMDGPU.td
@@ -947,6 +947,12 @@ def FeatureHasRestrictedSOffset : SubtargetFeature<"restricted-soffset",
"Has restricted SOffset (immediate not supported)."
>;
+def FeatureRequiredExportPriority : SubtargetFeature<"required-export-priority",
+ "HasRequiredExportPriority",
+ "true",
+ "Export priority must be explicitly manipulated on GFX11.5"
+>;
+
//===------------------------------------------------------------===//
// Subtarget Features (options and debugging)
//===------------------------------------------------------------===//
@@ -1597,14 +1603,16 @@ def FeatureISAVersion11_5_0 : FeatureSet<
!listconcat(FeatureISAVersion11_Common.Features,
[FeatureSALUFloatInsts,
FeatureDPPSrc1SGPR,
- FeatureVGPRSingleUseHintInsts])>;
+ FeatureVGPRSingleUseHintInsts,
+ FeatureRequiredExportPriority])>;
def FeatureISAVersion11_5_1 : FeatureSet<
!listconcat(FeatureISAVersion11_Common.Features,
[FeatureSALUFloatInsts,
FeatureDPPSrc1SGPR,
FeatureVGPRSingleUseHintInsts,
- FeatureGFX11FullVGPRs])>;
+ FeatureGFX11FullVGPRs,
+ FeatureRequiredExportPriority])>;
def FeatureISAVersion12 : FeatureSet<
[FeatureGFX12,
diff --git a/llvm/lib/Target/AMDGPU/GCNHazardRecognizer.cpp b/llvm/lib/Target/AMDGPU/GCNHazardRecognizer.cpp
index a402fc6d7e611..a8b171aa82840 100644
--- a/llvm/lib/Target/AMDGPU/GCNHazardRecognizer.cpp
+++ b/llvm/lib/Target/AMDGPU/GCNHazardRecognizer.cpp
@@ -14,6 +14,7 @@
#include "GCNSubtarget.h"
#include "MCTargetDesc/AMDGPUMCTargetDesc.h"
#include "SIMachineFunctionInfo.h"
+#include "llvm/CodeGen/MachineFrameInfo.h"
#include "llvm/CodeGen/MachineFunction.h"
#include "llvm/CodeGen/ScheduleDAG.h"
#include "llvm/TargetParser/TargetParser.h"
@@ -1104,6 +1105,7 @@ void GCNHazardRecognizer::fixHazards(MachineInstr *MI) {
fixWMMAHazards(MI);
fixShift64HighRegBug(MI);
fixVALUMaskWriteHazard(MI);
+ fixRequiredExportPriority(MI);
}
bool GCNHazardRecognizer::fixVcmpxPermlaneHazards(MachineInstr *MI) {
@@ -2895,3 +2897,113 @@ bool GCNHazardRecognizer::fixVALUMaskWriteHazard(MachineInstr *MI) {
return true;
}
+
+static bool ensureEntrySetPrio(MachineFunction *MF, int Priority,
+ const SIInstrInfo &TII) {
+ MachineBasicBlock &EntryMBB = MF->front();
+ if (EntryMBB.begin() != EntryMBB.end()) {
+ auto &EntryMI = *EntryMBB.begin();
+ if (EntryMI.getOpcode() == AMDGPU::S_SETPRIO &&
+ EntryMI.getOperand(0).getImm() >= Priority)
+ return false;
+ }
+
+ BuildMI(EntryMBB, EntryMBB.begin(), DebugLoc(), TII.get(AMDGPU::S_SETPRIO))
+ .addImm(Priority);
+ return true;
+}
+
+bool GCNHazardRecognizer::fixRequiredExportPriority(MachineInstr *MI) {
+ if (!ST.hasRequiredExportPriority())
+ return false;
+
+ // Assume the following shader types will never have exports,
+ // and avoid adding or adjusting S_SETPRIO.
+ MachineBasicBlock *MBB = MI->getParent();
+ MachineFunction *MF = MBB->getParent();
+ auto CC = MF->getFunction().getCallingConv();
+ switch (CC) {
+ case CallingConv::AMDGPU_CS:
+ case CallingConv::AMDGPU_CS_Chain:
+ case CallingConv::AMDGPU_CS_ChainPreserve:
+ case CallingConv::AMDGPU_KERNEL:
+ return false;
+ default:
+ break;
+ }
+
+ const int MaxPriority = 3;
+ const int NormalPriority = 2;
+ const int PostExportPriority = 0;
+
+ auto It = MI->getIterator();
+ switch (MI->getOpcode()) {
+ case AMDGPU::S_ENDPGM:
+ case AMDGPU::S_ENDPGM_SAVED:
+ case AMDGPU::S_ENDPGM_ORDERED_PS_DONE:
+ case AMDGPU::SI_RETURN_TO_EPILOG:
+ // Ensure shader with calls raises priority at entry.
+ // This ensures correct priority if exports exist in callee.
+ if (MF->getFrameInfo().hasCalls())
+ return ensureEntrySetPrio(MF, NormalPriority, TII);
+ return false;
+ case AMDGPU::S_SETPRIO: {
+ // Raise minimum priority unless in workaround.
+ auto &PrioOp = MI->getOperand(0);
+ int Prio = PrioOp.getImm();
+ bool InWA = (Prio == PostExportPriority) &&
+ (It != MBB->begin() && TII.isEXP(*std::prev(It)));
+ if (InWA || Prio >= NormalPriority)
+ return false;
+ PrioOp.setImm(std::min(Prio + NormalPriority, MaxPriority));
+ return true;
+ }
+ default:
+ if (!TII.isEXP(*MI))
+ return false;
+ break;
+ }
+
+ // Check entry priority at each export (as there will only be a few).
+ // Note: amdgpu_gfx can only be a callee, so defer to caller setprio.
+ bool Changed = false;
+ if (CC != CallingConv::AMDGPU_Gfx)
+ Changed = ensureEntrySetPrio(MF, NormalPriority, TII);
+
+ auto NextMI = std::next(It);
+ bool EndOfShader = false;
+ if (NextMI != MBB->end()) {
+ // Only need WA at end of sequence of exports.
+ if (TII.isEXP(*NextMI))
+ return Changed;
+ // Assume appropriate S_SETPRIO after export means WA already applied.
+ if (NextMI->getOpcode() == AMDGPU::S_SETPRIO &&
+ NextMI->getOperand(0).getImm() == PostExportPriority)
+ return Changed;
+ EndOfShader = NextMI->getOpcode() == AMDGPU::S_ENDPGM;
+ }
+
+ const DebugLoc &DL = MI->getDebugLoc();
+
+ // Lower priority.
+ BuildMI(*MBB, NextMI, DL, TII.get(AMDGPU::S_SETPRIO))
+ .addImm(PostExportPriority);
+
+ if (!EndOfShader) {
+ // Wait for exports to complete.
+ BuildMI(*MBB, NextMI, DL, TII.get(AMDGPU::S_WAITCNT_EXPCNT))
+ .addReg(AMDGPU::SGPR_NULL)
+ .addImm(0);
+ }
+
+ BuildMI(*MBB, NextMI, DL, TII.get(AMDGPU::S_NOP)).addImm(0);
+ BuildMI(*MBB, NextMI, DL, TII.get(AMDGPU::S_NOP)).addImm(0);
+
+ if (!EndOfShader) {
+ // Return to normal (higher) priority.
+ BuildMI(*MBB, NextMI, DL, TII.get(AMDGPU::S_SETPRIO))
+ .addImm(NormalPriority);
+ }
+
+ return true;
+}
diff --git a/llvm/lib/Target/AMDGPU/GCNHazardRecognizer.h b/llvm/lib/Target/AMDGPU/GCNHazardRecognizer.h
index 3ccca527c626b..f2a64ab48e180 100644
--- a/llvm/lib/Target/AMDGPU/GCNHazardRecognizer.h
+++ b/llvm/lib/Target/AMDGPU/GCNHazardRecognizer.h
@@ -107,6 +107,7 @@ class GCNHazardRecognizer final : public ScheduleHazardRecognizer {
bool fixWMMAHazards(MachineInstr *MI);
bool fixShift64HighRegBug(MachineInstr *MI);
bool fixVALUMaskWriteHazard(MachineInstr *MI);
+ bool fixRequiredExportPriority(MachineInstr *MI);
int checkMAIHazards(MachineInstr *MI);
int checkMAIHazards908(MachineInstr *MI);
diff --git a/llvm/lib/Target/AMDGPU/GCNSubtarget.h b/llvm/lib/Target/AMDGPU/GCNSubtarget.h
index e5817594a4521..def89c785b855 100644
--- a/llvm/lib/Target/AMDGPU/GCNSubtarget.h
+++ b/llvm/lib/Target/AMDGPU/GCNSubtarget.h
@@ -238,6 +238,7 @@ class GCNSubtarget final : public AMDGPUGenSubtargetInfo,
bool HasVOPDInsts = false;
bool HasVALUTransUseHazard = false;
bool HasForceStoreSC0SC1 = false;
+ bool HasRequiredExportPriority = false;
// Dummy feature to use for assembler in tablegen.
bool FeatureDisable = false;
@@ -1282,6 +1283,8 @@ class GCNSubtarget final : public AMDGPUGenSubtargetInfo,
bool hasRestrictedSOffset() const { return HasRestrictedSOffset; }
+ bool hasRequiredExportPriority() const { return HasRequiredExportPriority; }
+
/// \returns true if the target uses LOADcnt/SAMPLEcnt/BVHcnt, DScnt/KMcnt
/// and STOREcnt rather than VMcnt, LGKMcnt and VScnt respectively.
bool hasExtendedWaitCounts() const { return getGeneration() >= GFX12; }
diff --git a/llvm/test/CodeGen/AMDGPU/required-export-priority.ll b/llvm/test/CodeGen/AMDGPU/required-export-priority.ll
new file mode 100644
index 0000000000000..377902f3f0d1a
--- /dev/null
+++ b/llvm/test/CodeGen/AMDGPU/required-export-priority.ll
@@ -0,0 +1,344 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
+; RUN: llc -mtriple=amdgcn -mcpu=gfx1150 -amdgpu-enable-vopd=0 -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
+
+define amdgpu_ps void @test_export_zeroes_f32() #0 {
+; GCN-LABEL: test_export_zeroes_f32:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_setprio 2
+; GCN-NEXT: v_mov_b32_e32 v0, 0
+; GCN-NEXT: exp mrt0 off, off, off, off
+; GCN-NEXT: exp mrt0 off, off, off, off done
+; GCN-NEXT: s_setprio 0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_endpgm
+ call void @llvm.amdgcn.exp.f32(i32 0, i32 0, float 0.0, float 0.0, float 0.0, float 0.0, i1 false, i1 false)
+ call void @llvm.amdgcn.exp.f32(i32 0, i32 0, float 0.0, float 0.0, float 0.0, float 0.0, i1 true, i1 false)
+ ret void
+}
+
+define amdgpu_ps void @test_export_en_src0_f32() #0 {
+; GCN-LABEL: test_export_en_src0_f32:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_setprio 2
+; GCN-NEXT: v_mov_b32_e32 v0, 4.0
+; GCN-NEXT: v_mov_b32_e32 v1, 0.5
+; GCN-NEXT: v_mov_b32_e32 v2, 2.0
+; GCN-NEXT: v_mov_b32_e32 v3, 1.0
+; GCN-NEXT: exp mrt0 v3, off, off, off done
+; GCN-NEXT: s_setprio 0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_endpgm
+ call void @llvm.amdgcn.exp.f32(i32 0, i32 1, float 1.0, float 2.0, float 0.5, float 4.0, i1 true, i1 false)
+ ret void
+}
+
+define amdgpu_gs void @test_export_gs() #0 {
+; GCN-LABEL: test_export_gs:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_setprio 2
+; GCN-NEXT: v_mov_b32_e32 v0, 4.0
+; GCN-NEXT: v_mov_b32_e32 v1, 0.5
+; GCN-NEXT: v_mov_b32_e32 v2, 2.0
+; GCN-NEXT: v_mov_b32_e32 v3, 1.0
+; GCN-NEXT: exp mrt0 off, v2, off, off done
+; GCN-NEXT: s_setprio 0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_endpgm
+ call void @llvm.amdgcn.exp.f32(i32 0, i32 2, float 1.0, float 2.0, float 0.5, float 4.0, i1 true, i1 false)
+ ret void
+}
+
+define amdgpu_hs void @test_export_hs() #0 {
+; GCN-LABEL: test_export_hs:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_setprio 2
+; GCN-NEXT: v_mov_b32_e32 v0, 4.0
+; GCN-NEXT: v_mov_b32_e32 v1, 0.5
+; GCN-NEXT: v_mov_b32_e32 v2, 2.0
+; GCN-NEXT: v_mov_b32_e32 v3, 1.0
+; GCN-NEXT: exp mrt0 off, v2, off, off done
+; GCN-NEXT: s_setprio 0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_endpgm
+ call void @llvm.amdgcn.exp.f32(i32 0, i32 2, float 1.0, float 2.0, float 0.5, float 4.0, i1 true, i1 false)
+ ret void
+}
+
+define amdgpu_gfx void @test_export_gfx(float %v) #0 {
+; GCN-LABEL: test_export_gfx:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GCN-NEXT: v_mov_b32_e32 v1, 4.0
+; GCN-NEXT: v_mov_b32_e32 v2, 0.5
+; GCN-NEXT: v_mov_b32_e32 v3, 2.0
+; GCN-NEXT: exp mrt0 off, v3, off, off done
+; GCN-NEXT: s_setprio 0
+; GCN-NEXT: s_waitcnt_expcnt null, 0x0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_setprio 2
+; GCN-NEXT: s_waitcnt expcnt(0)
+; GCN-NEXT: s_setpc_b64 s[30:31]
+ call void @llvm.amdgcn.exp.f32(i32 0, i32 2, float %v, float 2.0, float 0.5, float 4.0, i1 true, i1 false)
+ ret void
+}
+
+define amdgpu_cs void @test_export_cs() #0 {
+; GCN-LABEL: test_export_cs:
+; GCN: ; %bb.0:
+; GCN-NEXT: v_mov_b32_e32 v0, 4.0
+; GCN-NEXT: v_mov_b32_e32 v1, 0.5
+; GCN-NEXT: v_mov_b32_e32 v2, 2.0
+; GCN-NEXT: v_mov_b32_e32 v3, 1.0
+; GCN-NEXT: exp mrt0 off, v2, off, off done
+; GCN-NEXT: s_endpgm
+ call void @llvm.amdgcn.exp.f32(i32 0, i32 2, float 1.0, float 2.0, float 0.5, float 4.0, i1 true, i1 false)
+ ret void
+}
+
+define amdgpu_kernel void @test_export_kernel() #0 {
+; GCN-LABEL: test_export_kernel:
+; GCN: ; %bb.0:
+; GCN-NEXT: v_mov_b32_e32 v0, 4.0
+; GCN-NEXT: v_mov_b32_e32 v1, 0.5
+; GCN-NEXT: v_mov_b32_e32 v2, 2.0
+; GCN-NEXT: v_mov_b32_e32 v3, 1.0
+; GCN-NEXT: exp mrt0 off, v2, off, off done
+; GCN-NEXT: s_endpgm
+ call void @llvm.amdgcn.exp.f32(i32 0, i32 2, float 1.0, float 2.0, float 0.5, float 4.0, i1 true, i1 false)
+ ret void
+}
+
+define amdgpu_gfx void @test_no_export_gfx(float %v) #0 {
+; GCN-LABEL: test_no_export_gfx:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_waitcnt vmcnt(0) expcnt(0) lgkmcnt(0)
+; GCN-NEXT: s_setpc_b64 s[30:31]
+ ret void
+}
+
+define amdgpu_ps void @test_no_export_ps(float %v) #0 {
+; GCN-LABEL: test_no_export_ps:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_endpgm
+ ret void
+}
+
+define amdgpu_ps void @test_if_export_f32(i32 %flag, float %x, float %y, float %z, float %w) #0 {
+; GCN-LABEL: test_if_export_f32:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_setprio 2
+; GCN-NEXT: s_mov_b32 s0, exec_lo
+; GCN-NEXT: v_cmpx_ne_u32_e32 0, v0
+; GCN-NEXT: s_cbranch_execz .LBB9_2
+; GCN-NEXT: ; %bb.1: ; %exp
+; GCN-NEXT: exp mrt0 v1, v2, v3, v4
+; GCN-NEXT: s_setprio 0
+; GCN-NEXT: s_waitcnt_expcnt null, 0x0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_setprio 2
+; GCN-NEXT: .LBB9_2: ; %end
+; GCN-NEXT: s_endpgm
+ %cc = icmp eq i32 %flag, 0
+ br i1 %cc, label %end, label %exp
+
+exp:
+ call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %x, float %y, float %z, float %w, i1 false, i1 false)
+ br label %end
+
+end:
+ ret void
+}
+
+define amdgpu_ps void @test_if_export_vm_f32(i32 %flag, float %x, float %y, float %z, float %w) #0 {
+; GCN-LABEL: test_if_export_vm_f32:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_setprio 2
+; GCN-NEXT: s_mov_b32 s0, exec_lo
+; GCN-NEXT: v_cmpx_ne_u32_e32 0, v0
+; GCN-NEXT: s_cbranch_execz .LBB10_2
+; GCN-NEXT: ; %bb.1: ; %exp
+; GCN-NEXT: exp mrt0 v1, v2, v3, v4
+; GCN-NEXT: s_setprio 0
+; GCN-NEXT: s_waitcnt_expcnt null, 0x0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_setprio 2
+; GCN-NEXT: .LBB10_2: ; %end
+; GCN-NEXT: s_endpgm
+ %cc = icmp eq i32 %flag, 0
+ br i1 %cc, label %end, label %exp
+
+exp:
+ call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %x, float %y, float %z, float %w, i1 false, i1 true)
+ br label %end
+
+end:
+ ret void
+}
+
+define amdgpu_ps void @test_if_export_done_f32(i32 %flag, float %x, float %y, float %z, float %w) #0 {
+; GCN-LABEL: test_if_export_done_f32:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_setprio 2
+; GCN-NEXT: s_mov_b32 s0, exec_lo
+; GCN-NEXT: v_cmpx_ne_u32_e32 0, v0
+; GCN-NEXT: s_cbranch_execz .LBB11_2
+; GCN-NEXT: ; %bb.1: ; %exp
+; GCN-NEXT: exp mrt0 v1, v2, v3, v4 done
+; GCN-NEXT: s_setprio 0
+; GCN-NEXT: s_waitcnt_expcnt null, 0x0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_setprio 2
+; GCN-NEXT: .LBB11_2: ; %end
+; GCN-NEXT: s_endpgm
+ %cc = icmp eq i32 %flag, 0
+ br i1 %cc, label %end, label %exp
+
+exp:
+ call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %x, float %y, float %z, float %w, i1 true, i1 false)
+ br label %end
+
+end:
+ ret void
+}
+
+define amdgpu_ps void @test_if_export_vm_done_f32(i32 %flag, float %x, float %y, float %z, float %w) #0 {
+; GCN-LABEL: test_if_export_vm_done_f32:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_setprio 2
+; GCN-NEXT: s_mov_b32 s0, exec_lo
+; GCN-NEXT: v_cmpx_ne_u32_e32 0, v0
+; GCN-NEXT: s_cbranch_execz .LBB12_2
+; GCN-NEXT: ; %bb.1: ; %exp
+; GCN-NEXT: exp mrt0 v1, v2, v3, v4 done
+; GCN-NEXT: s_setprio 0
+; GCN-NEXT: s_waitcnt_expcnt null, 0x0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_setprio 2
+; GCN-NEXT: .LBB12_2: ; %end
+; GCN-NEXT: s_endpgm
+ %cc = icmp eq i32 %flag, 0
+ br i1 %cc, label %end, label %exp
+
+exp:
+ call void @llvm.amdgcn.exp.f32(i32 0, i32 15, float %x, float %y, float %z, float %w, i1 true, i1 true)
+ br label %end
+
+end:
+ ret void
+}
+
+define amdgpu_ps void @test_export_pos_before_param_across_load(i32 %idx) #0 {
+; GCN-LABEL: test_export_pos_before_param_across_load:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_setprio 2
+; GCN-NEXT: buffer_load_b32 v0, v0, s[0:3], 0 offen
+; GCN-NEXT: v_mov_b32_e32 v1, 0
+; GCN-NEXT: v_mov_b32_e32 v2, 1.0
+; GCN-NEXT: v_mov_b32_e32 v3, 0.5
+; GCN-NEXT: s_waitcnt vmcnt(0)
+; GCN-NEXT: exp pos0 v1, v1, v1, v0 done
+; GCN-NEXT: exp invalid_target_32 v2, v2, v2, v2
+; GCN-NEXT: exp invalid_target_33 v2, v2, v2, v3
+; GCN-NEXT: s_setprio 0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_endpgm
+ call void @llvm.amdgcn.exp.f32(i32 32, i32 15, float 1.0, float 1.0, float 1.0, float 1.0, i1 false, i1 false)
+ call void @llvm.amdgcn.exp.f32(i32 33, i32 15, float 1.0, float 1.0, float 1.0, float 0.5, i1 false, i1 false)
+ %load = call float @llvm.amdgcn.raw.ptr.buffer.load.f32(ptr addrspace(8) undef, i32 %idx, i32 0, i32 0)
+ call void @llvm.amdgcn.exp.f32(i32 12, i32 15, float 0.0, float 0.0, float 0.0, float %load, i1 true, i1 false)
+ ret void
+}
+
+define amdgpu_ps void @test_export_across_store_load(i32 %idx, float %v) #0 {
+; GCN-LABEL: test_export_across_store_load:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_setprio 2
+; GCN-NEXT: v_mov_b32_e32 v2, 24
+; GCN-NEXT: v_cmp_eq_u32_e32 vcc_lo, 1, v0
+; GCN-NEXT: s_delay_alu instid0(VALU_DEP_2)
+; GCN-NEXT: v_cndmask_b32_e64 v0, v2, 8, vcc_lo
+; GCN-NEXT: v_mov_b32_e32 v2, 0
+; GCN-NEXT: scratch_store_b32 v0, v1, off
+; GCN-NEXT: scratch_load_b32 v0, off, off
+; GCN-NEXT: v_mov_b32_e32 v1, 1.0
+; GCN-NEXT: exp pos0 v2, v2, v2, v1 done
+; GCN-NEXT: s_setprio 0
+; GCN-NEXT: s_waitcnt_expcnt null, 0x0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_setprio 2
+; GCN-NEXT: s_waitcnt vmcnt(0)
+; GCN-NEXT: exp invalid_target_32 v0, v2, v1, v2
+; GCN-NEXT: exp invalid_target_33 v0, v2, v1, v2
+; GCN-NEXT: s_setprio 0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_nop 0
+; GCN-NEXT: s_endpgm
+ %data0 = alloca <4 x float>, align 8, addrspace(5)
+ %data1 = alloca <4 x float>, align 8, addrspace(5)
+ %cmp = icmp eq i32 %idx, 1
+ %data = select i1 %cmp, ptr addrspace(5) %data0, ptr addrspace(5) %data1
+ store float %v, ptr addrspace(5) %data, align 8
+ call void @llvm.amdgcn.exp.f32(i32 12, i32 15, float 0.0, float 0.0, float 0.0, float 1.0, i1 true, i1 false)
+ %load0 = load float, ptr addrspace(5) %data0, align 8
+ call void @llvm.amdgcn.exp.f32(i32 32, i32 15, float %load0, float 0.0, float 1.0, float 0.0, i1 false, i1 false)
+ call void @llvm.amdgcn.exp.f32(i32 33, i32 15, float %load0, float 0.0, float 1.0, float 0.0, i1 false, i1 false)
+ ret void
+}
+
+define amdgpu_ps void @test_export_in_callee(float %v) #0 {
+; GCN-LABEL: test_export_in_callee:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_setprio 2
+; GCN-NEXT: s_getpc_b64 s[0:1]
+; GCN-NEXT: s_add_u32 s0, s0, test_export_gfx@gotpcrel32@lo+4
+; GCN-NEXT: s_addc_u32 s1, s1, test_export_gfx@gotpcrel32@hi+12
+; GCN-NEXT: v_add_f32_e32 v0, 1.0, v0
+; GCN-NEXT: s_load_b64 s[0:1], s[0:1], 0x0
+; GCN-NEXT: s_mov_b32 s32, 0
+; GCN-NEXT: s_waitcnt lgkmcnt(0)
+; GCN-NEXT: s_swappc_b64 s[30:31], s[0:1]
+; GCN-NEXT: s_endpgm
+ %x = fadd float %v, 1.0
+ call void @test_export_gfx(float %x)
+ ret void
+}
+
+define amdgpu_ps void @test_export_in_callee_prio(float %v) #0 {
+; GCN-LABEL: test_export_in_callee_prio:
+; GCN: ; %bb.0:
+; GCN-NEXT: s_setprio 2
+; GCN-NEXT: s_mov_b32 s32, 0
+; GCN-NEXT: v_add_f32_e32 v0, 1.0, v0
+; GCN-NEXT: s_setprio 2
+; GCN-NEXT: s_getpc_b64 s[0:1]
+; GCN-NEXT: s_add_u32 s0, s0, test_export_gfx@gotpcrel32@lo+4
+; GCN-NEXT: s_addc_u32 s1, s1, test_export_gfx@gotpcrel32@hi+12
+; GCN-NEXT: s_load_b64 s[0:1], s[0:1], 0x0
+; GCN-NEXT: s_waitcnt lgkmcnt(0)
+; GCN-NEXT: s_swappc_b64 s[30:31], s[0:1]
+; GCN-NEXT: s_endpgm
+ %x = fadd float %v, 1.0
+ call void @llvm.amdgcn.s.setprio(i16 0)
+ call void @test_export_gfx(float %x)
+ ret void
+}
+
+declare void @llvm.amdgcn.exp.f32(i32, i32, float, float, float, float, i1, i1) #1
+declare void @llvm.amdgcn.exp.i32(i32, i32, i32, i32, i32, i32, i1, i1) #1
+declare float @llvm.amdgcn.raw.ptr.buffer.load.f32(ptr addrspace(8), i32, i32, i32) #2
+declare void @llvm.amdgcn.s.setprio(i16)
+
+attributes #0 = { nounwind }
+attributes #1 = { nounwind inaccessiblememonly }
+attributes #2 = { nounwind readnone }
diff --git a/llvm/test/CodeGen/AMDGPU/required-export-priority.mir b/llvm/test/CodeGen/AMDGPU/required-export-priority.mir
new file mode 100644
index 0000000000000..eee04468036e5
--- /dev/null
+++ b/llvm/test/CodeGen/AMDGPU/required-export-priority.mir
@@ -0,0 +1,293 @@
+# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py UTC_ARGS: --version 5
+# RUN: llc -mtriple=amdgcn -mcpu=gfx1150 -mattr=-wavefrontsize32,+wavefrontsize64 -run-pass=post-RA-hazard-rec -verify-machineinstrs %s -o - | FileCheck -check-prefixes=GFX1150 %s
+
+--- |
+ define amdgpu_ps void @end_of_shader() {
+ ret void
+ }
+ define amdgpu_ps void @end_of_shader_return_to_epilogue() {
+ ret void
+ }
+ define amdgpu_ps void @end_of_block() {
+ ret void
+ }
+ define amdgpu_ps void @start_of_block() {
+ ret void
+ }
+ define amdgpu_ps void @block_of_exports() {
+ ret void
+ }
+ define amdgpu_ps void @sparse_exports() {
+ ret void
+ }
+ define amdgpu_ps void @existing_setprio_1() {
+ ret void
+ }
+ define amdgpu_ps void @existing_setprio_2() {
+ ret void
+ }
+...
+
+---
+name: end_of_shader
+tracksRegLiveness: true
+liveins:
+ - { reg: '$vgpr0' }
+body: |
+ bb.0:
+ liveins: $vgpr0
+ ; GFX1150-LABEL: name: end_of_shader
+ ; GFX1150: liveins: $vgpr0
+ ; GFX1150-NEXT: {{ $}}
+ ; GFX1150-NEXT: S_SETPRIO 2
+ ; GFX1150-NEXT: EXP_DONE 0, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ ; GFX1150-NEXT: S_SETPRIO 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_ENDPGM 0
+ EXP_DONE 0, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ S_ENDPGM 0
+...
+
+---
+name: end_of_shader_return_to_epilogue
+tracksRegLiveness: true
+liveins:
+ - { reg: '$vgpr0' }
+body: |
+ bb.0:
+ liveins: $vgpr0
+ ; GFX1150-LABEL: name: end_of_shader_return_to_epilogue
+ ; GFX1150: liveins: $vgpr0
+ ; GFX1150-NEXT: {{ $}}
+ ; GFX1150-NEXT: S_SETPRIO 2
+ ; GFX1150-NEXT: EXP_DONE 0, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ ; GFX1150-NEXT: S_SETPRIO 0
+ ; GFX1150-NEXT: S_WAITCNT_EXPCNT $sgpr_null, 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_SETPRIO 2
+ ; GFX1150-NEXT: SI_RETURN_TO_EPILOG $vgpr0
+ EXP_DONE 0, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ SI_RETURN_TO_EPILOG $vgpr0
+...
+
+---
+name: end_of_block
+tracksRegLiveness: true
+liveins:
+ - { reg: '$vgpr0' }
+body: |
+ ; GFX1150-LABEL: name: end_of_block
+ ; GFX1150: bb.0:
+ ; GFX1150-NEXT: successors: %bb.1(0x80000000)
+ ; GFX1150-NEXT: liveins: $vgpr0
+ ; GFX1150-NEXT: {{ $}}
+ ; GFX1150-NEXT: S_SETPRIO 2
+ ; GFX1150-NEXT: EXP_DONE 0, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ ; GFX1150-NEXT: S_SETPRIO 0
+ ; GFX1150-NEXT: S_WAITCNT_EXPCNT $sgpr_null, 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_SETPRIO 2
+ ; GFX1150-NEXT: {{ $}}
+ ; GFX1150-NEXT: bb.1:
+ ; GFX1150-NEXT: S_ENDPGM 0
+ bb.0:
+ liveins: $vgpr0
+ EXP_DONE 0, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+
+ bb.1:
+ S_ENDPGM 0
+...
+
+---
+name: start_of_block
+tracksRegLiveness: true
+liveins:
+ - { reg: '$vgpr0' }
+body: |
+ ; GFX1150-LABEL: name: start_of_block
+ ; GFX1150: bb.0:
+ ; GFX1150-NEXT: successors: %bb.1(0x80000000)
+ ; GFX1150-NEXT: liveins: $vgpr0
+ ; GFX1150-NEXT: {{ $}}
+ ; GFX1150-NEXT: S_SETPRIO 2
+ ; GFX1150-NEXT: {{ $}}
+ ; GFX1150-NEXT: bb.1:
+ ; GFX1150-NEXT: successors: %bb.2(0x80000000)
+ ; GFX1150-NEXT: liveins: $vgpr0
+ ; GFX1150-NEXT: {{ $}}
+ ; GFX1150-NEXT: EXP_DONE 0, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ ; GFX1150-NEXT: S_SETPRIO 0
+ ; GFX1150-NEXT: S_WAITCNT_EXPCNT $sgpr_null, 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_SETPRIO 2
+ ; GFX1150-NEXT: {{ $}}
+ ; GFX1150-NEXT: bb.2:
+ ; GFX1150-NEXT: S_ENDPGM 0
+ bb.0:
+ liveins: $vgpr0
+
+ bb.1:
+ liveins: $vgpr0
+ EXP_DONE 0, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+
+ bb.2:
+ S_ENDPGM 0
+...
+
+---
+name: block_of_exports
+tracksRegLiveness: true
+liveins:
+ - { reg: '$vgpr0' }
+body: |
+ bb.0:
+ liveins: $vgpr0
+ ; GFX1150-LABEL: name: block_of_exports
+ ; GFX1150: liveins: $vgpr0
+ ; GFX1150-NEXT: {{ $}}
+ ; GFX1150-NEXT: S_SETPRIO 2
+ ; GFX1150-NEXT: EXP 2, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ ; GFX1150-NEXT: EXP 1, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ ; GFX1150-NEXT: EXP_DONE 0, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ ; GFX1150-NEXT: S_SETPRIO 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_ENDPGM 0
+ EXP 2, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ EXP 1, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ EXP_DONE 0, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ S_ENDPGM 0
+...
+
+---
+name: sparse_exports
+tracksRegLiveness: true
+liveins:
+ - { reg: '$vgpr0' }
+body: |
+ bb.0:
+ liveins: $vgpr0
+ ; GFX1150-LABEL: name: sparse_exports
+ ; GFX1150: liveins: $vgpr0
+ ; GFX1150-NEXT: {{ $}}
+ ; GFX1150-NEXT: S_SETPRIO 2
+ ; GFX1150-NEXT: EXP 2, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ ; GFX1150-NEXT: S_SETPRIO 0
+ ; GFX1150-NEXT: S_WAITCNT_EXPCNT $sgpr_null, 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_SETPRIO 2
+ ; GFX1150-NEXT: $vgpr0 = V_AND_B32_e32 1, $vgpr0, implicit $exec
+ ; GFX1150-NEXT: EXP 1, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ ; GFX1150-NEXT: S_SETPRIO 0
+ ; GFX1150-NEXT: S_WAITCNT_EXPCNT $sgpr_null, 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_SETPRIO 2
+ ; GFX1150-NEXT: $vgpr0 = V_OR_B32_e32 2, $vgpr0, implicit $exec
+ ; GFX1150-NEXT: EXP_DONE 0, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ ; GFX1150-NEXT: S_SETPRIO 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_ENDPGM 0
+ EXP 2, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ $vgpr0 = V_AND_B32_e32 1, $vgpr0, implicit $exec
+ EXP 1, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ $vgpr0 = V_OR_B32_e32 2, $vgpr0, implicit $exec
+ EXP_DONE 0, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ S_ENDPGM 0
+...
+
+---
+name: existing_setprio_1
+tracksRegLiveness: true
+liveins:
+ - { reg: '$vgpr0' }
+body: |
+ ; GFX1150-LABEL: name: existing_setprio_1
+ ; GFX1150: bb.0:
+ ; GFX1150-NEXT: successors: %bb.1(0x80000000)
+ ; GFX1150-NEXT: liveins: $vgpr0
+ ; GFX1150-NEXT: {{ $}}
+ ; GFX1150-NEXT: S_SETPRIO 2
+ ; GFX1150-NEXT: $vgpr0 = V_AND_B32_e32 1, $vgpr0, implicit $exec
+ ; GFX1150-NEXT: {{ $}}
+ ; GFX1150-NEXT: bb.1:
+ ; GFX1150-NEXT: successors: %bb.2(0x80000000)
+ ; GFX1150-NEXT: liveins: $vgpr0
+ ; GFX1150-NEXT: {{ $}}
+ ; GFX1150-NEXT: S_SETPRIO 3
+ ; GFX1150-NEXT: $vgpr0 = V_OR_B32_e32 2, $vgpr0, implicit $exec
+ ; GFX1150-NEXT: S_SETPRIO 2
+ ; GFX1150-NEXT: {{ $}}
+ ; GFX1150-NEXT: bb.2:
+ ; GFX1150-NEXT: successors: %bb.3(0x80000000)
+ ; GFX1150-NEXT: liveins: $vgpr0
+ ; GFX1150-NEXT: {{ $}}
+ ; GFX1150-NEXT: S_SETPRIO 3
+ ; GFX1150-NEXT: $vgpr0 = V_OR_B32_e32 3, $vgpr0, implicit $exec
+ ; GFX1150-NEXT: S_SETPRIO 2
+ ; GFX1150-NEXT: {{ $}}
+ ; GFX1150-NEXT: bb.3:
+ ; GFX1150-NEXT: liveins: $vgpr0
+ ; GFX1150-NEXT: {{ $}}
+ ; GFX1150-NEXT: EXP 1, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ ; GFX1150-NEXT: EXP_DONE 0, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ ; GFX1150-NEXT: S_SETPRIO 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_ENDPGM 0
+ bb.0:
+ liveins: $vgpr0
+ $vgpr0 = V_AND_B32_e32 1, $vgpr0, implicit $exec
+
+ bb.1:
+ liveins: $vgpr0
+ S_SETPRIO 3
+ $vgpr0 = V_OR_B32_e32 2, $vgpr0, implicit $exec
+ S_SETPRIO 0
+
+ bb.2:
+ liveins: $vgpr0
+ S_SETPRIO 1
+ $vgpr0 = V_OR_B32_e32 3, $vgpr0, implicit $exec
+ S_SETPRIO 0
+
+ bb.3:
+ liveins: $vgpr0
+ EXP 1, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ EXP_DONE 0, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ S_ENDPGM 0
+...
+
+---
+name: existing_setprio_2
+tracksRegLiveness: true
+liveins:
+ - { reg: '$vgpr0' }
+body: |
+ bb.0:
+ liveins: $vgpr0
+ ; GFX1150-LABEL: name: existing_setprio_2
+ ; GFX1150: liveins: $vgpr0
+ ; GFX1150-NEXT: {{ $}}
+ ; GFX1150-NEXT: S_SETPRIO 3
+ ; GFX1150-NEXT: EXP 1, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ ; GFX1150-NEXT: EXP_DONE 0, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ ; GFX1150-NEXT: S_SETPRIO 0
+ ; GFX1150-NEXT: S_WAITCNT_EXPCNT $sgpr_null, 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_NOP 0
+ ; GFX1150-NEXT: S_SETPRIO 2
+ ; GFX1150-NEXT: S_SETPRIO 3
+ ; GFX1150-NEXT: S_ENDPGM 0
+ S_SETPRIO 3
+ EXP 1, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ EXP_DONE 0, $vgpr0, $vgpr0, $vgpr0, $vgpr0, -1, -1, 15, implicit $exec
+ S_SETPRIO 3
+ S_ENDPGM 0
+...
From 8ea44e65f2c19facff751aeb2ac960f907fb210f Mon Sep 17 00:00:00 2001
From: Carl Ritson <carl.ritson@amd.com>
Date: Wed, 17 Jul 2024 16:18:02 +0900
Subject: [PATCH 2/3] Remove -verify-machineinstrs from test.
---
llvm/test/CodeGen/AMDGPU/required-export-priority.ll | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/llvm/test/CodeGen/AMDGPU/required-export-priority.ll b/llvm/test/CodeGen/AMDGPU/required-export-priority.ll
index 377902f3f0d1a..ebc209bd4d451 100644
--- a/llvm/test/CodeGen/AMDGPU/required-export-priority.ll
+++ b/llvm/test/CodeGen/AMDGPU/required-export-priority.ll
@@ -1,5 +1,5 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
-; RUN: llc -mtriple=amdgcn -mcpu=gfx1150 -amdgpu-enable-vopd=0 -verify-machineinstrs < %s | FileCheck -check-prefix=GCN %s
+; RUN: llc -mtriple=amdgcn -mcpu=gfx1150 -amdgpu-enable-vopd=0 < %s | FileCheck -check-prefix=GCN %s
define amdgpu_ps void @test_export_zeroes_f32() #0 {
; GCN-LABEL: test_export_zeroes_f32:

632
modular/llvm/2ffded40ed8bfdbc0859bfc3ce86cc129243962b.patch Normal file
View File

@ -0,0 +1,632 @@
From 2ffded40ed8bfdbc0859bfc3ce86cc129243962b Mon Sep 17 00:00:00 2001
From: Konrad Kleine <kkleine@redhat.com>
Date: Dec 05 2024 11:44:07 +0000
Subject: Add MLIR
---
diff --git a/0001-CMake-Add-missing-dependency-108461.patch b/0001-CMake-Add-missing-dependency-108461.patch
new file mode 100644
index 0000000..e6862fd
--- /dev/null
+++ b/0001-CMake-Add-missing-dependency-108461.patch
@@ -0,0 +1,129 @@
+From 17ff6161b83e6a5e86fcb6a13c5551bba1438405 Mon Sep 17 00:00:00 2001
+From: Erick Ochoa <github@ceci-nest-pas.me>
+Date: Thu, 12 Sep 2024 21:16:58 -0400
+Subject: [PATCH 1/9] [CMake] Add missing dependency (#108461)
+
+The [`mlir-capi-execution-engine-test` test
+executable](https://github.com/llvm/llvm-project/blob/main/mlir/test/CAPI/CMakeLists.txt#L26-L34)
+
+```cmake
+if(MLIR_ENABLE_EXECUTION_ENGINE)
+ _add_capi_test_executable(mlir-capi-execution-engine-test
+ execution_engine.c
+ LINK_LIBS PRIVATE
+ MLIRCAPIConversion
+ MLIRCAPIExecutionEngine
+ MLIRCAPIRegisterEverything
+)
+endif()
+```
+
+
+is run by lit tests, but it is not properly listed as a dependency. It
+is added in places conditionally across the file
+[`tests/CMakeLists.txt`](https://github.com/llvm/llvm-project/blob/main/mlir/test/CMakeLists.txt#L130-L143)
+
+```cmake
+# The native target may not be enabled, in this case we won't
+# run tests that involves executing on the host: do not build
+# useless binaries.
+if(LLVM_ENABLE_PIC AND TARGET ${LLVM_NATIVE_ARCH})
+ list(APPEND MLIR_TEST_DEPENDS
+ mlir-cpu-runner
+ llc
+ mlir_async_runtime
+ mlir-capi-execution-engine-test
+ mlir_c_runner_utils
+ mlir_runner_utils
+ mlir_float16_utils
+ )
+endif()
+```
+
+But this condition is not the same as the one where the test executable
+is added. [It has been reported on discord that the following error
+occurred:](https://discord.com/channels/636084430946959380/642426447167881246/1283811636725022730)
+
+```
+FAIL: MLIR :: CAPI/execution_engine.c (2 of 2121)
+******************** TEST 'MLIR :: CAPI/execution_engine.c' FAILED ********************
+Exit Code: 127
+Command Output (stdout):
+--
+# RUN: at line 10
+/usr/bin/mlir-capi-execution-engine-test 2>&1 | /usr/bin/FileCheck /builddir/build/BUILD/mlir-19.1.0_rc4-build/mlir-19.1.0-rc4.src/test/CAPI/execution_engine.c
+# executed command: /usr/bin/mlir-capi-execution-engine-test
+# .---command stderr------------
+# | '/usr/bin/mlir-capi-execution-engine-test': command not found
+# `-----------------------------
+```
+
+This error will not be deterministic and is dependent on the order in
+which tools are built. If by any chance,
+`mlir-capi-execution-engine-test` is built before the lit tests run,
+then nothing will happen. But lit tests can be run before
+`mlir-capi-execution-engine-test` is built.
+
+This patch adds the `mlir-capi-execution-engine` to the
+`MLIR_TEST_DEPENDS` list when the `MLIR_ENABLE_EXECUTION_ENGINE` flag is
+present.
+
+Happy to make changes like:
+* removing `mlir-capi-execution-engine-test` from the other place where
+it is included in the tests
+* and merge and sort alphabetically these two commands
+
+```cmake
+set(MLIR_TEST_DEPENDS
+FileCheck count not split-file
+mlir-capi-ir-test
+mlir-capi-irdl-test
+mlir-capi-llvm-test
+mlir-capi-pass-test
+mlir-capi-quant-test
+mlir-capi-rewrite-test
+mlir-capi-sparse-tensor-test
+mlir-capi-transform-test
+mlir-capi-transform-interpreter-test
+mlir-capi-translation-test
+mlir-linalg-ods-yaml-gen
+mlir-lsp-server
+mlir-opt
+ mlir-query
+ mlir-reduce
+ mlir-tblgen
+ mlir-translate
+ tblgen-lsp-server
+ tblgen-to-irdl
+ )
+
+set(MLIR_TEST_DEPENDS ${MLIR_TEST_DEPENDS}
+ mlir-capi-pdl-test
+ mlir-pdll-lsp-server
+ mlir-pdll
+ )
+```
+
+Co-authored-by: Erick Ochoa <erick@ceci-nest-pas.me>
+---
+ mlir/test/CMakeLists.txt | 4 ++++
+ 1 file changed, 4 insertions(+)
+
+diff --git a/mlir/test/CMakeLists.txt b/mlir/test/CMakeLists.txt
+index df95e5db11f1..4d2d738b734e 100644
+--- a/mlir/test/CMakeLists.txt
++++ b/mlir/test/CMakeLists.txt
+@@ -150,6 +150,10 @@ if(MLIR_ENABLE_CUDA_RUNNER)
+ list(APPEND MLIR_TEST_DEPENDS mlir_cuda_runtime)
+ endif()
+
++if(MLIR_ENABLE_EXECUTION_ENGINE)
++ list(APPEND MLIR_TEST_DEPENDS mlir-capi-execution-engine-test)
++endif()
++
+ if(MLIR_ENABLE_ROCM_RUNNER)
+ list(APPEND MLIR_TEST_DEPENDS mlir_rocm_runtime)
+ endif()
+--
+2.46.0
+
diff --git a/0001-mlir-Specify-deps-via-LLVM_LINK_COMPONENTS.patch b/0001-mlir-Specify-deps-via-LLVM_LINK_COMPONENTS.patch
new file mode 100644
index 0000000..2178800
--- /dev/null
+++ b/0001-mlir-Specify-deps-via-LLVM_LINK_COMPONENTS.patch
@@ -0,0 +1,80 @@
+From c8f93f1958de1f59222a89dd64a573f91105e135 Mon Sep 17 00:00:00 2001
+From: Konrad Kleine <kkleine@redhat.com>
+Date: Tue, 3 Dec 2024 20:53:29 +0100
+Subject: [PATCH] [mlir] Specify deps via `LLVM_LINK_COMPONENTS`
+
+This specifies the dependencies to link against with
+`LLVM_LINK_COMPONENTS` for the
+`mlir/test/Target/LLVM/MLIRTargetLLVMTests` binary.
+
+Before, the dependencies where directly added to the
+`target_link_libraries()` call which caused the problems I describe
+next:
+
+When doing a build of LLVM with MLIR I want to link against
+`libLLVM.so` instead of statically linking `libLLVMSupport.a`.
+MLIR on the other side seems to statically link against
+`libLLVMSupport.a` because when I link to the shared library `libLLVM.so` I get:
+
+```
+CommandLine Error: Option 'aarch64-ptrauth-auth-checks' registered more than once!
+```
+
+This error indicates that the `Support` library is linked twice in the `MLIRTargetLLVMTest` binary.
+
+Here's the creation of the `MLIRTargetLLVMTest` binary before (Notice the
+`libLLVMSupport.a`):
+
+```
+[6535/6847] : && /usr/bin/clang++ -O2 -flto=thin -ffat-lto-objects -fexceptions -g -grecord-gcc-switches -pipe -Wall -Werror=format-security -Wp,-U_FORTIFY_SOURCE,-D_FORTIFY_SOURCE=3 -Wp,-D_GLIBCXX_ASSERTIONS --config=/usr/lib/rpm/redhat/redhat-hardened-clang.cfg -fstack-protector-strong -mbranch-protection=standard -fasynchronous-unwind-tables -D_DEFAULT_SOURCE -Dasm=__asm__ -fPIC -fno-semantic-interposition -fvisibility-inlines-hidden -Werror=date-time -Werror=unguarded-availability-new -Wall -Wextra -Wno-unused-parameter -Wwrite-strings -Wcast-qual -Wmissing-field-initializers -pedantic -Wno-long-long -Wc++98-compat-extra-semi -Wimplicit-fallthrough -Wcovered-switch-default -Wno-noexcept-type -Wnon-virtual-dtor -Wdelete-non-virtual-dtor -Wsuggest-override -Wstring-conversion -Wmisleading-indentation -Wctad-maybe-unsupported -fdiagnostics-color -ffunction-sections -fdata-sections -Wundef -Werror=mismatched-tags -O2 -g -DNDEBUG -Wl,-z,relro -Wl,--as-needed -Wl,-z,pack-relative-relocs -Wl,-z,now --config=/usr/lib/rpm/redhat/redhat-hardened-clang-ld.cfg -flto=thin -ffat-lto-objects -Wl,--build-id=sha1 -Wl,--gc-sections -fno-lto tools/mlir/unittests/Target/LLVM/CMakeFiles/MLIRTargetLLVMTests.dir/SerializeNVVMTarget.cpp.o tools/mlir/unittests/Target/LLVM/CMakeFiles/MLIRTargetLLVMTests.dir/SerializeROCDLTarget.cpp.o tools/mlir/unittests/Target/LLVM/CMakeFiles/MLIRTargetLLVMTests.dir/SerializeToLLVMBitcode.cpp.o -o tools/mlir/unittests/Target/LLVM/MLIRTargetLLVMTests -Wl,-rpath,/builddir/build/BUILD/llvm-19.1.3-build/llvm-project-19.1.3.src/llvm/redhat-linux-build/lib64 lib64/libllvm_gtest_main.a lib64/libllvm_gtest.a lib64/libMLIRTargetLLVM.a lib64/libMLIRNVVMTarget.a lib64/libMLIRROCDLTarget.a lib64/libMLIRGPUDialect.a lib64/libMLIRNVVMDialect.a lib64/libMLIRLLVMDialect.a lib64/libMLIRLLVMToLLVMIRTranslation.a lib64/libMLIRBuiltinToLLVMIRTranslation.a lib64/libMLIRNVVMToLLVMIRTranslation.a lib64/libMLIRROCDLToLLVMIRTranslation.a lib64/libMLIRGPUToLLVMIRTranslation.a lib64/libLLVMAArch64CodeGen.a lib64/libLLVMAArch64Desc.a lib64/libLLVMAArch64Info.a -lpthread lib64/libMLIRTargetLLVM.a lib64/libMLIRROCDLDialect.a lib64/libMLIRExecutionEngineUtils.a lib64/libMLIRGPUDialect.a lib64/libMLIRMemRefDialect.a lib64/libMLIRArithUtils.a lib64/libMLIRDialectUtils.a lib64/libMLIRComplexDialect.a lib64/libMLIRArithAttrToLLVMConversion.a lib64/libMLIRArithDialect.a lib64/libMLIRCastInterfaces.a lib64/libMLIRDialect.a lib64/libMLIRInferIntRangeCommon.a lib64/libMLIRUBDialect.a lib64/libMLIRShapedOpInterfaces.a lib64/libMLIRTargetLLVMIRExport.a lib64/libMLIRDLTIDialect.a lib64/libMLIRLLVMIRTransforms.a lib64/libMLIRNVVMDialect.a lib64/libMLIRLLVMDialect.a lib64/libMLIRFuncDialect.a lib64/libMLIRTransforms.a lib64/libMLIRMemorySlotInterfaces.a lib64/libMLIRCopyOpInterface.a lib64/libMLIRRuntimeVerifiableOpInterface.a lib64/libMLIRTranslateLib.a lib64/libMLIRParser.a lib64/libMLIRBytecodeReader.a lib64/libMLIRAsmParser.a lib64/libMLIRTransformUtils.a lib64/libMLIRSubsetOpInterface.a lib64/libMLIRValueBoundsOpInterface.a lib64/libMLIRDestinationStyleOpInterface.a lib64/libMLIRRewrite.a lib64/libMLIRRewritePDL.a lib64/libMLIRPDLToPDLInterp.a lib64/libMLIRPass.a lib64/libMLIRAnalysis.a lib64/libMLIRControlFlowInterfaces.a lib64/libMLIRInferIntRangeInterface.a lib64/libMLIRCallInterfaces.a lib64/libMLIRDataLayoutInterfaces.a lib64/libMLIRViewLikeInterface.a lib64/libMLIRLoopLikeInterface.a lib64/libMLIRPresburger.a lib64/libMLIRPDLInterpDialect.a lib64/libMLIRFunctionInterfaces.a lib64/libMLIRPDLDialect.a lib64/libMLIRSideEffectInterfaces.a lib64/libMLIRInferTypeOpInterface.a lib64/libMLIRIR.a lib64/libMLIRSupport.a lib64/libLLVM.so.19.1 lib64/libLLVMAArch64Utils.a lib64/libLLVMAsmPrinter.a lib64/libLLVMCFGuard.a lib64/libLLVMGlobalISel.a lib64/libLLVMSelectionDAG.a lib64/libLLVMCodeGen.a lib64/libLLVMScalarOpts.a lib64/libLLVMAggressiveInstCombine.a lib64/libLLVMInstCombine.a lib64/libLLVMBitWriter.a lib64/libLLVMObjCARCOpts.a lib64/libLLVMCodeGenTypes.a lib64/libLLVMTarget.a lib64/libLLVMVectorize.a lib64/libLLVMTransformUtils.a lib64/libLLVMAnalysis.a lib64/libLLVMProfileData.a lib64/libLLVMSymbolize.a lib64/libLLVMDebugInfoDWARF.a lib64/libLLVMDebugInfoPDB.a lib64/libLLVMObject.a lib64/libLLVMMCParser.a lib64/libLLVMMC.a lib64/libLLVMIRReader.a lib64/libLLVMBitReader.a lib64/libLLVMAsmParser.a lib64/libLLVMTextAPI.a lib64/libLLVMDebugInfoCodeView.a lib64/libLLVMDebugInfoMSF.a lib64/libLLVMDebugInfoBTF.a lib64/libLLVMCore.a lib64/libLLVMBinaryFormat.a lib64/libLLVMRemarks.a lib64/libLLVMBitstreamReader.a lib64/libLLVMTargetParser.a lib64/libLLVMSupport.a lib64/libLLVMDemangle.a -lrt -ldl -lm /usr/lib64/libz.so /usr/lib64/libzstd.so && :
+```
+
+Here's the full error:
+
+```
+[24/25] cd /builddir/build/BUILD/llvm-19.1.3-build/llvm-project-19.1.3.src/llvm/redhat-linux-build/tools/mlir/test && /usr/bin/python3 /builddir/build/BUILD/llvm-19.1.3-build/llvm-project-19.1.3.src/llvm/redhat-linux-build/./bin/llvm-lit -vv /builddir/build/BUILD/llvm-19.1.3-build/llvm-project-19.1.3.src/llvm/redhat-linux-build/tools/mlir/test
+: CommandLine Error: Option 'aarch64-ptrauth-auth-checks' registered more than once!
+LLVM ERROR: inconsistency in registered CommandLine options
+llvm-lit: /builddir/build/BUILD/llvm-19.1.3-build/llvm-project-19.1.3.src/llvm/utils/lit/lit/formats/googletest.py:38: warning: unable to discover google-tests in '/builddir/build/BUILD/llvm-19.1.3-build/llvm-project-19.1.3.src/llvm/redhat-linux-build/tools/mlir/unittests/Target/LLVM/./MLIRTargetLLVMTests': Command '['/builddir/build/BUILD/llvm-19.1.3-build/llvm-project-19.1.3.src/llvm/redhat-linux-build/tools/mlir/unittests/Target/LLVM/./MLIRTargetLLVMTests', '--gtest_list_tests', '--gtest_filter=-*DISABLED_*']' died with <Signals.SIGABRT: 6>.. Process output: b''
+error: filter did not match any tests (of 2704 discovered). Use '--allow-empty-runs' to suppress this error.
+FAILED: tools/mlir/test/CMakeFiles/check-mlir /builddir/build/BUILD/llvm-19.1.3-build/llvm-project-19.1.3.src/llvm/redhat-linux-build/tools/mlir/test/CMakeFiles/check-mlir
+```
+
+Here's the CMake invocation:
+
+```
+/usr/bin/cmake -S . -B redhat-linux-build -DCMAKE_C_FLAGS_RELEASE:STRING=-DNDEBUG -DCMAKE_CXX_FLAGS_RELEASE:STRING=-DNDEBUG -DCMAKE_Fortran_FLAGS_RELEASE:STRING=-DNDEBUG -DCMAKE_VERBOSE_MAKEFILE:BOOL=ON -DCMAKE_INSTALL_DO_STRIP:BOOL=OFF -DCMAKE_INSTALL_PREFIX:PATH=/usr -DINCLUDE_INSTALL_DIR:PATH=/usr/include -DLIB_INSTALL_DIR:PATH=/usr/lib64 -DSYSCONF_INSTALL_DIR:PATH=/etc -DSHARE_INSTALL_PREFIX:PATH=/usr/share -DLIB_SUFFIX=64 -DBUILD_SHARED_LIBS:BOOL=ON -G Ninja '' -DCLANG_BUILD_EXAMPLES:BOOL=OFF -DCLANG_CONFIG_FILE_SYSTEM_DIR=/etc/clang/ -DCLANG_DEFAULT_PIE_ON_LINUX=OFF -DCLANG_DEFAULT_UNWINDLIB=libgcc -DCLANG_ENABLE_ARCMT:BOOL=ON -DCLANG_ENABLE_STATIC_ANALYZER:BOOL=ON -DCLANG_INCLUDE_DOCS:BOOL=ON -DCLANG_INCLUDE_TESTS:BOOL=ON -DCLANG_LINK_CLANG_DYLIB=ON -DCLANG_PLUGIN_SUPPORT:BOOL=ON '-DCLANG_REPOSITORY_STRING=Fedora 19.1.3-5.fc42' -DLLVM_EXTERNAL_CLANG_TOOLS_EXTRA_SOURCE_DIR=../clang-tools-extra -DCLANG_RESOURCE_DIR=../lib/clang/19 -DCOMPILER_RT_INCLUDE_TESTS:BOOL=OFF -DCOMPILER_RT_INSTALL_PATH=/usr/lib/clang/19 -DLLVM_ENABLE_DOXYGEN:BOOL=OFF -DLLVM_ENABLE_SPHINX:BOOL=ON -DLLVM_BUILD_DOCS:BOOL=ON -DSPHINX_EXECUTABLE=/usr/bin/sphinx-build-3 -DSPHINX_OUTPUT_HTML:BOOL=OFF -DSPHINX_OUTPUT_MAN:BOOL=ON -DSPHINX_WARNINGS_AS_ERRORS=OFF -DLLDB_DISABLE_CURSES:BOOL=OFF -DLLDB_DISABLE_LIBEDIT:BOOL=OFF -DLLDB_DISABLE_PYTHON:BOOL=OFF -DLLDB_ENFORCE_STRICT_TEST_REQUIREMENTS:BOOL=ON -DLLVM_APPEND_VC_REV:BOOL=OFF -DLLVM_BUILD_EXAMPLES:BOOL=OFF -DLLVM_BUILD_EXTERNAL_COMPILER_RT:BOOL=ON -DLLVM_BUILD_LLVM_DYLIB:BOOL=ON -DLLVM_BUILD_RUNTIME:BOOL=ON -DLLVM_BUILD_TOOLS:BOOL=ON -DLLVM_BUILD_UTILS:BOOL=ON -DLLVM_COMMON_CMAKE_UTILS=/usr/share/llvm/cmake -DLLVM_DEFAULT_TARGET_TRIPLE=aarch64-redhat-linux-gnu -DLLVM_DYLIB_COMPONENTS=all -DLLVM_ENABLE_EH=ON -DLLVM_ENABLE_FFI:BOOL=ON -DLLVM_ENABLE_LIBCXX:BOOL=OFF -DLLVM_ENABLE_PER_TARGET_RUNTIME_DIR=ON '-DLLVM_ENABLE_PROJECTS=clang;clang-tools-extra;lld;lldb;mlir' -DLLVM_ENABLE_RTTI:BOOL=ON '-DLLVM_ENABLE_RUNTIMES=compiler-rt;openmp;offload' -DLLVM_ENABLE_ZLIB:BOOL=ON -DLLVM_EXPERIMENTAL_TARGETS_TO_BUILD=AVR -DLLVM_INCLUDE_BENCHMARKS=OFF -DLLVM_INCLUDE_EXAMPLES:BOOL=ON -DLLVM_INCLUDE_TOOLS:BOOL=ON -DLLVM_INCLUDE_UTILS:BOOL=ON -DLLVM_INSTALL_TOOLCHAIN_ONLY:BOOL=OFF -DLLVM_INSTALL_UTILS:BOOL=ON -DLLVM_LINK_LLVM_DYLIB:BOOL=ON -DLLVM_PARALLEL_LINK_JOBS=1 -DLLVM_TARGETS_TO_BUILD=all -DLLVM_TOOLS_INSTALL_DIR:PATH=bin -DLLVM_UNREACHABLE_OPTIMIZE:BOOL=OFF -DLLVM_USE_PERF:BOOL=ON -DLLVM_UTILS_INSTALL_DIR:PATH=bin -DMLIR_INCLUDE_DOCS:BOOL=ON -DMLIR_INCLUDE_TESTS:BOOL=ON -DMLIR_INCLUDE_INTEGRATION_TESTS:BOOL=OFF -DMLIR_INSTALL_AGGREGATE_OBJECTS=OFF -DMLIR_BUILD_MLIR_C_DYLIB=ON -DMLIR_ENABLE_BINDINGS_PYTHON:BOOL=ON -DOPENMP_INSTALL_LIBDIR=lib64 -DLIBOMP_INSTALL_ALIASES=OFF -DLLVM_BUILD_TESTS:BOOL=ON -DLLVM_INCLUDE_TESTS:BOOL=ON -DLLVM_INSTALL_GTEST:BOOL=ON -DLLVM_LIT_ARGS=-vv -DLLVM_UNITTEST_LINK_FLAGS=-fno-lto -DBUILD_SHARED_LIBS:BOOL=OFF -DCMAKE_BUILD_TYPE=RelWithDebInfo -DCMAKE_INSTALL_PREFIX=/usr -DENABLE_LINKER_BUILD_ID:BOOL=ON -DOFFLOAD_INSTALL_LIBDIR=lib64 -DPython3_EXECUTABLE=/usr/bin/python3 -DCMAKE_SKIP_INSTALL_RPATH:BOOL=ON -DPPC_LINUX_DEFAULT_IEEELONGDOUBLE=ON -DLLVM_LIBDIR_SUFFIX=64 -DLLVM_BINUTILS_INCDIR=/usr/include -DLLVM_VERSION_SUFFIX=
+```
+---
+ mlir/unittests/Target/LLVM/CMakeLists.txt | 5 ++---
+ 1 file changed, 2 insertions(+), 3 deletions(-)
+
+diff --git a/mlir/unittests/Target/LLVM/CMakeLists.txt b/mlir/unittests/Target/LLVM/CMakeLists.txt
+index 6d612548a94c..0c61d222dedf 100644
+--- a/mlir/unittests/Target/LLVM/CMakeLists.txt
++++ b/mlir/unittests/Target/LLVM/CMakeLists.txt
+@@ -1,11 +1,11 @@
++set(LLVM_LINK_COMPONENTS nativecodegen)
++
+ add_mlir_unittest(MLIRTargetLLVMTests
+ SerializeNVVMTarget.cpp
+ SerializeROCDLTarget.cpp
+ SerializeToLLVMBitcode.cpp
+ )
+
+-llvm_map_components_to_libnames(llvm_libs nativecodegen)
+-
+ target_link_libraries(MLIRTargetLLVMTests
+ PRIVATE
+ MLIRTargetLLVM
+@@ -19,7 +19,6 @@ target_link_libraries(MLIRTargetLLVMTests
+ MLIRNVVMToLLVMIRTranslation
+ MLIRROCDLToLLVMIRTranslation
+ MLIRGPUToLLVMIRTranslation
+- ${llvm_libs}
+ )
+
+ if (DEFINED LLVM_NATIVE_TARGET)
+--
+2.46.0
+
diff --git a/0001-mlir-python-Reuse-the-library-directory.patch b/0001-mlir-python-Reuse-the-library-directory.patch
new file mode 100644
index 0000000..7590347
--- /dev/null
+++ b/0001-mlir-python-Reuse-the-library-directory.patch
@@ -0,0 +1,77 @@
+From 1d043550c7601ca776b0318b9b913e5ecf325baf Mon Sep 17 00:00:00 2001
+From: Tulio Magno Quites Machado Filho <tuliom@redhat.com>
+Date: Tue, 5 Sep 2023 10:35:37 -0300
+Subject: [PATCH] Reuse the library directory
+
+Prefer to get the path to the shared libraries from config.llvm_shlib_dir.
+Fallback to the previous path only if config.llvm_shlib_dir is not
+defined.
+
+This ensures the test will pass regardless of the build configuration
+used downstream.
+---
+ mlir/test/lit.cfg.py | 1 +
+ mlir/test/python/execution_engine.py | 12 +++++++-----
+ 2 files changed, 8 insertions(+), 5 deletions(-)
+
+diff --git a/mlir/test/lit.cfg.py b/mlir/test/lit.cfg.py
+--- a/mlir/test/lit.cfg.py 2024-08-05 10:40:33.000000000 +0200
++++ b/mlir/test/lit.cfg.py 2024-08-14 19:17:55.160470316 +0200
+@@ -164,6 +164,7 @@
+ ToolSubst("transform-opt-ch2", unresolved="ignore"),
+ ToolSubst("transform-opt-ch3", unresolved="ignore"),
+ ToolSubst("transform-opt-ch4", unresolved="ignore"),
++ ToolSubst("%llvm_shlib_dir", config.llvm_shlib_dir, unresolved="ignore"),
+ ToolSubst("mlir-transform-opt", unresolved="ignore"),
+ ToolSubst("%mlir_lib_dir", config.mlir_lib_dir, unresolved="ignore"),
+ ToolSubst("%mlir_src_dir", config.mlir_src_root, unresolved="ignore"),
+diff -ruN mlir-19.1.0-rc2.src.orig/test/python/execution_engine.py mlir-19.1.0-rc2.src/test/python/execution_engine.py
+--- a/mlir/test/python/execution_engine.py 2024-08-05 10:40:33.000000000 +0200
++++ b/mlir/test/python/execution_engine.py 2024-08-14 19:17:20.822219824 +0200
+@@ -1,4 +1,4 @@
+-# RUN: %PYTHON %s 2>&1 | FileCheck %s
++# RUN: env LLVM_SHLIB_DIR=%llvm_shlib_dir %PYTHON %s 2>&1 | FileCheck %s
+ # REQUIRES: host-supports-jit
+ import gc, sys, os, tempfile
+ from mlir.ir import *
+@@ -6,6 +6,9 @@
+ from mlir.execution_engine import *
+ from mlir.runtime import *
+
++_DEFAULT_LIB_DIR = "../../../../lib"
++LIB_DIR = os.getenv("LLVM_SHLIB_DIR", _DEFAULT_LIB_DIR)
++
+
+ # Log everything to stderr and flush so that we have a unified stream to match
+ # errors/info emitted by MLIR to stderr.
+@@ -613,6 +616,7 @@
+ shared_libs = [
+ "../../../../bin/mlir_runner_utils.dll",
+ "../../../../bin/mlir_c_runner_utils.dll",
++
+ ]
+ elif sys.platform == "darwin":
+ shared_libs = [
+@@ -621,8 +625,9 @@
+ ]
+ else:
+ shared_libs = [
+- "../../../../lib/libmlir_runner_utils.so",
+- "../../../../lib/libmlir_c_runner_utils.so",
++ LIB_DIR + "/libmlir_runner_utils.so",
++ LIB_DIR + "/libmlir_c_runner_utils.so",
++
+ ]
+
+ execution_engine = ExecutionEngine(
+@@ -664,8 +669,8 @@
+ ]
+ else:
+ shared_libs = [
+- "../../../../lib/libmlir_runner_utils.so",
+- "../../../../lib/libmlir_c_runner_utils.so",
++ LIB_DIR + "/libmlir_runner_utils.so",
++ LIB_DIR + "/libmlir_c_runner_utils.so",
+ ]
+
+ execution_engine = ExecutionEngine(
diff --git a/Makefile b/Makefile
index 0d91301..83904c9 100644
--- a/Makefile
+++ b/Makefile
@@ -89,7 +89,7 @@ edit-last-failing-script: get-last-run-script
######### Re-run the last failing script from mock
.PHONY: mockbuild-rerun-last-script
-## Re-runs the last failing or running script of your release/mock mockbuild.
+## Re-runs the last failing or running script of your release/snapshot mock mockbuild.
mockbuild-rerun-last-script: get-last-run-script
mock --root=$(MOCK_CHROOT) --shell 'sh -e $(last_run_script)'
diff --git a/llvm.spec b/llvm.spec
index cea5ced..e2ce86f 100644
--- a/llvm.spec
+++ b/llvm.spec
@@ -41,6 +41,16 @@
%bcond_without lldb
%endif
+%if %{without compat_build} && 0%{?fedora}
+%ifarch %{ix86}
+%bcond_with mlir
+%else
+%bcond_without mlir
+%endif
+%else
+%bcond_with mlir
+%endif
+
# Disable LTO on x86 and riscv in order to reduce memory consumption.
%ifarch %ix86 riscv64
%bcond_with lto_build
@@ -178,11 +188,15 @@
#endregion LLDB globals
#endregion globals
+#region MLIR globals
+%global pkg_name_mlir mlir%{pkg_suffix}
+#endregion MLIR globals
+
#region packages
#region main package
Name: %{pkg_name_llvm}
Version: %{maj_ver}.%{min_ver}.%{patch_ver}%{?rc_ver:~rc%{rc_ver}}%{?llvm_snapshot_version_suffix:~%{llvm_snapshot_version_suffix}}
-Release: 1%{?dist}
+Release: 2%{?dist}
Summary: The Low Level Virtual Machine
License: Apache-2.0 WITH LLVM-exception OR NCSA
@@ -254,6 +268,15 @@ Patch103: 0001-Workaround-a-bug-in-ORC-on-ppc64le.patch
# this might no longer be needed.
Patch104: 0001-Driver-Give-devtoolset-path-precedence-over-Installe.patch
+#region MLIR patches
+# See https://github.com/llvm/llvm-project/pull/108579
+Patch1904: 0001-mlir-python-Reuse-the-library-directory.patch
+# See https://github.com/llvm/llvm-project/pull/108461
+Patch1905: 0001-CMake-Add-missing-dependency-108461.patch
+# See https://github.com/llvm/llvm-project/pull/118542
+Patch1906: 0001-mlir-Specify-deps-via-LLVM_LINK_COMPONENTS.patch
+#endregion MLIR patches
+
#region LLD patches
Patch1800: 0001-18-Always-build-shared-libs-for-LLD.patch
Patch1902: 0001-19-Always-build-shared-libs-for-LLD.patch
@@ -363,6 +386,12 @@ BuildRequires: perl(Term::ANSIColor)
BuildRequires: perl(Text::ParseWords)
BuildRequires: perl(Sys::Hostname)
+%if %{with mlir}
+BuildRequires: python%{python3_pkgversion}-numpy
+BuildRequires: python%{python3_pkgversion}-pybind11
+BuildRequires: python%{python3_pkgversion}-pyyaml
+%endif
+
BuildRequires: graphviz
# This is required because we need "ps" when running LLDB tests
@@ -761,6 +790,48 @@ Obsoletes: python3-lldb < 18.9
The package contains the LLDB Python module.
%endif
#endregion LLDB packages
+
+#region MLIR packages
+%if %{with mlir}
+%package -n %{pkg_name_mlir}
+Summary: Multi-Level Intermediate Representation Overview
+License: Apache-2.0 WITH LLVM-exception
+URL: http://mlir.llvm.org
+
+%description -n %{pkg_name_mlir}
+The MLIR project is a novel approach to building reusable and extensible
+compiler infrastructure. MLIR aims to address software fragmentation,
+improve compilation for heterogeneous hardware, significantly reduce
+the cost of building domain specific compilers, and aid in connecting
+existing compilers together.
+
+%package -n %{pkg_name_mlir}-static
+Summary: MLIR static files
+Requires: %{pkg_name_mlir}%{?_isa} = %{version}-%{release}
+
+%description -n %{pkg_name_mlir}-static
+MLIR static files.
+
+%package -n %{pkg_name_mlir}-devel
+Summary: MLIR development files
+Requires: %{pkg_name_mlir}%{?_isa} = %{version}-%{release}
+Requires: %{pkg_name_mlir}-static%{?_isa} = %{version}-%{release}
+
+%description -n %{pkg_name_mlir}-devel
+MLIR development files.
+
+%package -n python%{python3_pkgversion}-mlir
+%{?python_provide:%python_provide python%{python3_pkgversion}-mlir}
+Summary: MLIR python bindings
+
+Requires: python%{python3_pkgversion}
+Requires: python%{python3_pkgversion}-numpy
+
+%description -n python%{python3_pkgversion}-mlir
+MLIR python bindings.
+%endif
+#endregion MLIR packages
+
#endregion packages
#region prep
@@ -857,6 +928,9 @@ The package contains the LLDB Python module.
%if %{with lldb}
%global projects %{projects};lldb
%endif
+%if %{with mlir}
+%global projects %{projects};mlir
+%endif
# Copy CFLAGS into ASMFLAGS, so -fcf-protection is used when compiling assembly files.
export ASMFLAGS="%{build_cflags}"
@@ -990,6 +1064,18 @@ popd
-DLLVM_UTILS_INSTALL_DIR:PATH=bin
#endregion llvm options
+#region mlir options
+%if %{with mlir}
+%global cmake_config_args %{cmake_config_args} \\\
+ -DMLIR_INCLUDE_DOCS:BOOL=ON \\\
+ -DMLIR_INCLUDE_TESTS:BOOL=ON \\\
+ -DMLIR_INCLUDE_INTEGRATION_TESTS:BOOL=OFF \\\
+ -DMLIR_INSTALL_AGGREGATE_OBJECTS=OFF \\\
+ -DMLIR_BUILD_MLIR_C_DYLIB=ON \\\
+ -DMLIR_ENABLE_BINDINGS_PYTHON:BOOL=ON
+%endif
+#endregion mlir options
+
#region openmp options
%global cmake_config_args %{cmake_config_args} \\\
-DOPENMP_INSTALL_LIBDIR=%{unprefixed_libdir} \\\
@@ -1403,6 +1489,17 @@ ln -vsf "../../../${liblldb}" %{buildroot}%{python3_sitearch}/lldb/_lldb.so
%endif
#endregion LLDB installation
+#region mlir installation
+%if %{with mlir}
+mkdir -p %{buildroot}/%{python3_sitearch}
+mv %{buildroot}%{_prefix}/python_packages/mlir_core/mlir %{buildroot}/%{python3_sitearch}
+# These directories should be empty now.
+rmdir %{buildroot}%{_prefix}/python_packages/mlir_core %{buildroot}%{_prefix}/python_packages
+# Unneeded files.
+rm -rf %{buildroot}%{_prefix}/src/python
+%endif
+#endregion mlir installation
+
%if %{with compat_build}
# Add version suffix to binaries. Do this at the end so it includes any
# additional binaries that may be been added by other steps.
@@ -1494,6 +1591,9 @@ function reset_test_opts()
# Set for filtering out unit tests.
# See http://google.github.io/googletest/advanced.html#running-a-subset-of-the-tests
unset GTEST_FILTER
+
+ # Some test (e.g. mlir) require this to be set.
+ unset PYTHONPATH
}
# Convert array of test names into a regex.
@@ -1520,6 +1620,21 @@ function test_list_to_regex()
# Add enclosing parenthesis
echo "($arr)"
}
+
+# Similar to test_list_to_regex() except that this function exports
+# the LIT_FILTER_OUT if there are tests in the given list.
+# If there are no tests, the LIT_FILTER_OUT is unset in order to
+# avoid issues with the llvm test system.
+function adjust_lit_filter_out()
+{
+ local -n arr=$1
+ local res=$(test_list_to_regex test_list_filter_out)
+ if [[ "$res" != "()" ]]; then
+ export LIT_FILTER_OUT=$res
+ else
+ unset LIT_FILTER_OUT
+ fi
+}
#endregion Helper functions
#region Test LLVM lit
@@ -1606,7 +1721,7 @@ test_list_filter_out+=("libomp :: flush/omp_flush.c")
%endif
%ifarch aarch64 s390x
-# The following test has been failling intermittently on aarch64 and s390x.
+# The following test has been failing intermittently on aarch64 and s390x.
# Re-enable it after https://github.com/llvm/llvm-project/issues/117773
# gets fixed.
test_list_filter_out+=("libarcher :: races/taskwait-depend.c")
@@ -1711,7 +1826,7 @@ export LIT_XFAIL="$LIT_XFAIL;offloading/thread_state_1.c"
export LIT_XFAIL="$LIT_XFAIL;offloading/thread_state_2.c"
%endif
-export LIT_FILTER_OUT=$(test_list_to_regex test_list_filter_out)
+adjust_lit_filter_out test_list_filter_out
%cmake_build --target check-openmp
#endregion Test OPENMP
@@ -1743,6 +1858,29 @@ reset_test_opts
%cmake_build --target check-lld
#endregion Test LLD
+#region Test MLIR
+%if %{with mlir}
+reset_test_opts
+%ifarch s390x
+# s390x does not support half-float
+test_list_filter_out+=("mlir/test/python/execution_engine.py")
+
+# https://discourse.llvm.org/t/mlir-s390x-linux-failure/76695/25
+test_list_filter_out+=("mlir/test/Target/LLVMIR/llvmir.mlir")
+test_list_filter_out+=("mlir/test/python/ir/array_attributes.py")
+%endif
+
+%ifarch ppc64le
+test_list_filter_out+=("test/python/execution_engine.py")
+%endif
+
+adjust_lit_filter_out test_list_filter_out
+
+export PYTHONPATH=%{buildroot}/%{python3_sitearch}
+%cmake_build --target check-mlir
+%endif
+#endregion Test MLIR
+
%endif
%if %{with snapshot_build}
@@ -2501,10 +2639,62 @@ fi
%{python3_sitearch}/lldb
%endif
#endregion LLDB files
+
+
+#region MLIR files
+%if %{with mlir}
+%files -n %{pkg_name_mlir}
+%license LICENSE.TXT
+%{_libdir}/libmlir_arm_runner_utils.so.%{maj_ver}*
+%{_libdir}/libmlir_arm_sme_abi_stubs.so.%{maj_ver}*
+%{_libdir}/libmlir_async_runtime.so.%{maj_ver}*
+%{_libdir}/libmlir_c_runner_utils.so.%{maj_ver}*
+%{_libdir}/libmlir_float16_utils.so.%{maj_ver}*
+%{_libdir}/libmlir_runner_utils.so.%{maj_ver}*
+%{_libdir}/libMLIR*.so.%{maj_ver}*
+
+%files -n %{pkg_name_mlir}-static
+%{_libdir}/libMLIR*.a
+
+%files -n %{pkg_name_mlir}-devel
+%{_bindir}/mlir-cat
+%{_bindir}/mlir-cpu-runner
+%{_bindir}/mlir-linalg-ods-yaml-gen
+%{_bindir}/mlir-lsp-server
+%{_bindir}/mlir-minimal-opt
+%{_bindir}/mlir-minimal-opt-canonicalize
+%{_bindir}/mlir-opt
+%{_bindir}/mlir-pdll
+%{_bindir}/mlir-pdll-lsp-server
+%{_bindir}/mlir-query
+%{_bindir}/mlir-reduce
+%{_bindir}/mlir-tblgen
+%{_bindir}/mlir-transform-opt
+%{_bindir}/mlir-translate
+%{_bindir}/tblgen-lsp-server
+%{_bindir}/tblgen-to-irdl
+%{_includedir}/mlir
+%{_includedir}/mlir-c
+%{_libdir}/cmake/mlir
+%{_libdir}/libmlir_arm_runner_utils.so
+%{_libdir}/libmlir_arm_sme_abi_stubs.so
+%{_libdir}/libmlir_async_runtime.so
+%{_libdir}/libmlir_c_runner_utils.so
+%{_libdir}/libmlir_float16_utils.so
+%{_libdir}/libmlir_runner_utils.so
+%{_libdir}/libMLIR*.so
+
+%files -n python%{python3_pkgversion}-%{pkg_name_mlir}
+%{python3_sitearch}/mlir/
+%endif
+#endregion MLIR files
#endregion files
#region changelog
%changelog
+* Wed Dec 04 2024 Konrad Kleine <kkleine@redhat.com> - 19.1.5-2
+- Add mlir
+
* Tue Dec 03 2024 Timm Bäder <tbaeder@redhat.com> - 19.1.5-1
- Update to 19.1.5

180
modular/llvm/Makefile Normal file
View File

@ -0,0 +1,180 @@
.DEFAULT_GOAL=help
# See ~/.config/mock/<CONFIG>.cfg or /etc/mock/<CONFIG>.cfg
# Tweak this to centos-stream-9-x86_64 to build for CentOS
MOCK_CHROOT?=fedora-rawhide-x86_64
MOCK_OPTS?=
MOCK_OPTS_RELEASE?=--no-clean --no-cleanup-after $(MOCK_OPTS)
MOCK_OPTS_SNAPSHOT?=$(MOCK_OPTS_RELEASE) --with snapshot_build $(MOCK_OPTS)
YYYYMMDD=$(shell date +%Y%m%d)
SOURCEDIR=$(shell pwd)
SPEC=llvm.spec
# When nothing is given, this will be determined based on
# release or snapshot builds.
SRPM_PATH?=
######### Get sources
.PHONY: get-sources-snapshot
## Downloads all sources we need for a snapshot build.
get-sources-snapshot:
YYYYMMDD=$(YYYYMMDD) ./.copr/snapshot-info.sh > $(SOURCEDIR)/version.spec.inc
spectool -g --define "_sourcedir $(SOURCEDIR)" --define "_with_snapshot_build 1" $(SPEC)
.PHONY: get-sources-release
## Downloads all sources we need for a release build.
get-sources-release:
spectool -g --define "_sourcedir $(SOURCEDIR)" $(SPEC)
######### Build SRPM
.PHONY: srpm-release
## Builds an SRPM that can be used for a release build.
srpm-release: get-sources-release
rpmbuild \
--define "_rpmdir $(SOURCEDIR)" \
--define "_sourcedir $(SOURCEDIR)" \
--define "_specdir $(SOURCEDIR)" \
--define "_srcrpmdir $(SOURCEDIR)" \
--define "_builddir $(SOURCEDIR)" \
-bs $(SPEC)
.PHONY: srpm-snapshot
## Builds an SRPM that can be used for a snapshot build.
srpm-snapshot: get-sources-snapshot
rpmbuild \
--with=snapshot_build \
--define "_rpmdir $(SOURCEDIR)" \
--define "_sourcedir $(SOURCEDIR)" \
--define "_specdir $(SOURCEDIR)" \
--define "_srcrpmdir $(SOURCEDIR)" \
--define "_builddir $(SOURCEDIR)" \
-bs $(SPEC)
######### Scrub mock chroot and cache
.PHONY: scrub-chroot
## Completely remove the fedora chroot and cache.
scrub-chroot:
mock -r $(MOCK_CHROOT) --scrub all
######### Do a mock build
.PHONY: mockbuild-release
## Start a mock build of the release SRPM.
mockbuild-release: srpm-release get-srpm-release
mock -r $(MOCK_CHROOT) $(MOCK_OPTS_RELEASE) $(srpm_path)
.PHONY: mockbuild-snapshot
## Start a mock build of the snapshot SRPM.
mockbuild-snapshot: srpm-snapshot get-srpm-snapshot
mock -r $(MOCK_CHROOT) $(MOCK_OPTS_SNAPSHOT) $(srpm_path)
######### Edit-last-failing-script
.PHONY: get-last-run-script
## Get the file that was last modified in /var/tmp/ within the chroot.
get-last-run-script:
$(eval last_run_script:=/var/tmp/$(shell ls -t1 /var/lib/mock/$(MOCK_CHROOT)/root/var/tmp | head -n1))
$(info last_run_script=$(last_run_script))
@echo > /dev/null
.PHONY: edit-last-failing-script
## Opens the last failing or running script from mock in your editor
## of choice for you to edit it and later re-run it in mock with:
## "make mockbuild-rerun-last-script".
edit-last-failing-script: get-last-run-script
$$EDITOR /var/lib/mock/$(MOCK_CHROOT)/root$(last_run_script)
######### Re-run the last failing script from mock
.PHONY: mockbuild-rerun-last-script
## Re-runs the last failing or running script of your release/snapshot mock mockbuild.
mockbuild-rerun-last-script: get-last-run-script
mock --root=$(MOCK_CHROOT) --shell 'sh -e $(last_run_script)'
.PHONY: help
# Based on https://gist.github.com/rcmachado/af3db315e31383502660
## Display this help text.
help:/
$(info Available targets)
$(info -----------------)
@awk '/^[a-zA-Z\-0-9]+:/ { \
helpMessage = match(lastLine, /^## (.*)/); \
helpCommand = substr($$1, 0, index($$1, ":")-1); \
if (helpMessage) { \
helpMessage = substr(lastLine, RSTART + 3, RLENGTH); \
gsub(/##/, "\n ", helpMessage); \
} else { \
helpMessage = "(No documentation)"; \
} \
printf "%-37s - %s\n", helpCommand, helpMessage; \
lastLine = "" \
} \
{ hasComment = match(lastLine, /^## (.*)/); \
if(hasComment) { \
lastLine=lastLine$$0; \
} \
else { \
lastLine = $$0 \
} \
}' $(MAKEFILE_LIST)
######### Deprecated targets
# Map deprecated targets to new targets
.PHONY: snapshot-srpm release-srpm
snapshot-srpm release-srpm:
$(eval mapped_target:=$(subst snapshot-srpm,srpm-snapshot,$(MAKECMDGOALS)))
$(eval mapped_target:=$(subst release-srpm,srpm-release,$(mapped_target)))
$(info WARNING: "$(MAKECMDGOALS)" is deprecated. Instead running "$(mapped_target)")
$(MAKE) $(mapped_target)
######### Version/Release helper targets to build name of SRPM
.PHONY: get-llvm-version-release
## Determines the LLVM version given in the llvm.spec file.
get-llvm-version-release:
$(eval llvm_version_release:=$(shell grep -ioP "%global\s+(maj|min|patch)_ver[^0-9]\K[0-9]+" $(SPEC) | paste -sd'.'))
$(info LLVM Release Version: $(llvm_version_release))
@echo > /dev/null
.PHONY: get-llvm-version-snapshot
## Determines the LLVM version given in the version.spec.inc file.
get-llvm-version-snapshot:
$(eval llvm_version_snapshot:=$(shell grep -ioP "%global\s+(maj|min|patch)_ver[^0-9]\K[0-9]+" version.spec.inc | paste -sd'.'))
$(info LLVM Snapshot Version: $(llvm_version_snapshot))
@echo > /dev/null
.PHONY: get-spec-file-release
## Parses the spec file for the Release: tag
get-spec-file-release:
$(eval spec_file_release:=$(shell grep -ioP '^Release:\s*\K[0-9]+' $(SPEC)))
$(info LLVM Spec file Release: $(spec_file_release))
@echo > /dev/null
.PHONY: get-srpm-release
## Determines the name of the SRPM used for release builds
## Can be overriden by giving "make ... SRPM_PATH=foo.src.rpm".
get-srpm-release: get-llvm-version-release get-spec-file-release
ifeq ($(SRPM_PATH),)
$(eval srpm_path:=llvm-$(llvm_version_release)-$(spec_file_release).*.src.rpm)
else
$(eval srpm_path:=$(SRPM_PATH))
endif
$(info LLVM SRPM Release: $(srpm_path))
@echo > /dev/null
.PHONY: get-srpm-snapshot
## Determines the name of the SRPM used for snapshot builds
## Can be overriden by giving "make ... SRPM_PATH=foo.src.rpm".
get-srpm-snapshot: get-llvm-version-snapshot get-spec-file-release
ifeq ($(SRPM_PATH),)
$(eval yyyymmdd:=$(shell grep -ioP "%global\s+llvm_snapshot_yyyymmdd\s+\K[0-9]+" version.spec.inc))
$(eval git_short:=$(shell grep -ioP "%global\s+llvm_snapshot_git_revision_short\s+\K[a-zA-Z0-9]+" version.spec.inc))
$(eval srpm_path:=llvm-$(llvm_version_snapshot)~pre$(yyyymmdd).g$(git_short)-$(spec_file_release).*.src.rpm)
else
$(eval srpm_path:=$(SRPM_PATH))
endif
$(info LLVM SRPM Snapshot: $(srpm_path))
@echo > /dev/null

849
modular/llvm/llvm.spec
View File

File diff suppressed because it is too large Load Diff

3203
modular/llvm/llvm.spec.orig Normal file
View File

File diff suppressed because it is too large Load Diff