Skip to content

Commit fdea2cb

Browse files
omarahmed1111omarahmed1111
authored
Return the usage of -flto and use -lto-emit-asm instead (intel#16884)
This reverts this [PR](intel#16605) and add another solution to that problem mentioned in this [issue](intel#16413). This was the result of llvm upstream discussion [here](llvm/llvm-project#125243 (review))
1 parent 08b1747 commit fdea2cb

File tree

3 files changed

+15
-16
lines changed

3 files changed

+15
-16
lines changed

‎clang/test/Driver/linker-wrapper.c

Lines changed: 9 additions & 9 deletions
Original file line numberDiff line numberDiff line change
@@ -21,7 +21,7 @@ __attribute__((visibility("protected"), used)) int x;
2121
// RUN: clang-linker-wrapper --host-triple=x86_64-unknown-linux-gnu --dry-run \
2222
// RUN: --linker-path=/usr/bin/ld %t.o -o a.out 2>&1 | FileCheck %s --check-prefix=NVPTX-LINK
2323

24-
// NVPTX-LINK: clang{{.*}} -o {{.*}}.img --target=nvptx64-nvidia-cuda -march=sm_70 -O2 -foffload-lto {{.*}}.o {{.*}}.o
24+
// NVPTX-LINK: clang{{.*}} -o {{.*}}.img --target=nvptx64-nvidia-cuda -march=sm_70 -O2 -flto {{.*}}.o {{.*}}.o
2525

2626
// RUN: clang-offload-packager -o %t.out \
2727
// RUN: --image=file=%t.elf.o,kind=openmp,triple=nvptx64-nvidia-cuda,arch=sm_70 \
@@ -30,7 +30,7 @@ __attribute__((visibility("protected"), used)) int x;
3030
// RUN: clang-linker-wrapper --host-triple=x86_64-unknown-linux-gnu --dry-run --device-debug -O0 \
3131
// RUN: --linker-path=/usr/bin/ld %t.o -o a.out 2>&1 | FileCheck %s --check-prefix=NVPTX-LINK-DEBUG
3232

33-
// NVPTX-LINK-DEBUG: clang{{.*}} -o {{.*}}.img --target=nvptx64-nvidia-cuda -march=sm_70 -O2 -foffload-lto {{.*}}.o {{.*}}.o -g
33+
// NVPTX-LINK-DEBUG: clang{{.*}} -o {{.*}}.img --target=nvptx64-nvidia-cuda -march=sm_70 -O2 -flto {{.*}}.o {{.*}}.o -g
3434

3535
// RUN: clang-offload-packager -o %t.out \
3636
// RUN: --image=file=%t.elf.o,kind=openmp,triple=amdgcn-amd-amdhsa,arch=gfx908 \
@@ -39,7 +39,7 @@ __attribute__((visibility("protected"), used)) int x;
3939
// RUN: clang-linker-wrapper --host-triple=x86_64-unknown-linux-gnu --dry-run \
4040
// RUN: --linker-path=/usr/bin/ld %t.o -o a.out 2>&1 | FileCheck %s --check-prefix=AMDGPU-LINK
4141

42-
// AMDGPU-LINK: clang{{.*}} -o {{.*}}.img --target=amdgcn-amd-amdhsa -mcpu=gfx908 -O2 -foffload-lto -Wl,--no-undefined {{.*}}.o {{.*}}.o
42+
// AMDGPU-LINK: clang{{.*}} -o {{.*}}.img --target=amdgcn-amd-amdhsa -mcpu=gfx908 -O2 -flto -Wl,--no-undefined {{.*}}.o {{.*}}.o
4343

4444
// RUN: clang-offload-packager -o %t.out \
4545
// RUN: --image=file=%t.amdgpu.bc,kind=openmp,triple=amdgcn-amd-amdhsa,arch=gfx1030 \
@@ -48,7 +48,7 @@ __attribute__((visibility("protected"), used)) int x;
4848
// RUN: clang-linker-wrapper --host-triple=x86_64-unknown-linux-gnu --dry-run --save-temps -O2 \
4949
// RUN: --linker-path=/usr/bin/ld %t.o -o a.out 2>&1 | FileCheck %s --check-prefix=AMDGPU-LTO-TEMPS
5050

51-
// AMDGPU-LTO-TEMPS: clang{{.*}} -o {{.*}}.img --target=amdgcn-amd-amdhsa -mcpu=gfx1030 -O2 -foffload-lto -Wl,--no-undefined {{.*}} -save-temps
51+
// AMDGPU-LTO-TEMPS: clang{{.*}} -o {{.*}}.img --target=amdgcn-amd-amdhsa -mcpu=gfx1030 -O2 -flto -Wl,--no-undefined {{.*}} -save-temps
5252

5353
// RUN: clang-offload-packager -o %t.out \
5454
// RUN: --image=file=%t.elf.o,kind=openmp,triple=x86_64-unknown-linux-gnu \
@@ -148,7 +148,7 @@ __attribute__((visibility("protected"), used)) int x;
148148
// RUN: clang-linker-wrapper --host-triple=x86_64-unknown-linux-gnu --dry-run --clang-backend \
149149
// RUN: --linker-path=/usr/bin/ld %t.o -o a.out 2>&1 | FileCheck %s --check-prefix=CLANG-BACKEND
150150

151-
// CLANG-BACKEND: clang{{.*}} -o {{.*}}.img --target=amdgcn-amd-amdhsa -mcpu=gfx908 -O2 -foffload-lto -Wl,--no-undefined {{.*}}.o
151+
// CLANG-BACKEND: clang{{.*}} -o {{.*}}.img --target=amdgcn-amd-amdhsa -mcpu=gfx908 -O2 -flto -Wl,--no-undefined {{.*}}.o
152152

153153
// RUN: clang-offload-packager -o %t.out \
154154
// RUN: --image=file=%t.elf.o,kind=openmp,triple=nvptx64-nvidia-cuda,arch=sm_70
@@ -171,8 +171,8 @@ __attribute__((visibility("protected"), used)) int x;
171171
// RUN: clang-linker-wrapper --host-triple=x86_64-unknown-linux-gnu --dry-run \
172172
// RUN: --linker-path=/usr/bin/ld %t-on.o %t-off.o %t.a -o a.out 2>&1 | FileCheck %s --check-prefix=AMD-TARGET-ID
173173

174-
// AMD-TARGET-ID: clang{{.*}} -o {{.*}}.img --target=amdgcn-amd-amdhsa -mcpu=gfx90a:xnack+ -O2 -foffload-lto -Wl,--no-undefined {{.*}}.o {{.*}}.o
175-
// AMD-TARGET-ID: clang{{.*}} -o {{.*}}.img --target=amdgcn-amd-amdhsa -mcpu=gfx90a:xnack- -O2 -foffload-lto -Wl,--no-undefined {{.*}}.o {{.*}}.o
174+
// AMD-TARGET-ID: clang{{.*}} -o {{.*}}.img --target=amdgcn-amd-amdhsa -mcpu=gfx90a:xnack+ -O2 -flto -Wl,--no-undefined {{.*}}.o {{.*}}.o
175+
// AMD-TARGET-ID: clang{{.*}} -o {{.*}}.img --target=amdgcn-amd-amdhsa -mcpu=gfx90a:xnack- -O2 -flto -Wl,--no-undefined {{.*}}.o {{.*}}.o
176176

177177
// RUN: clang-offload-packager -o %t-lib.out \
178178
// RUN: --image=file=%t.elf.o,kind=openmp,triple=amdgcn-amd-amdhsa,arch=generic
@@ -187,8 +187,8 @@ __attribute__((visibility("protected"), used)) int x;
187187
// RUN: clang-linker-wrapper --host-triple=x86_64-unknown-linux-gnu --dry-run \
188188
// RUN: --linker-path=/usr/bin/ld %t1.o %t2.o %t.a -o a.out 2>&1 | FileCheck %s --check-prefix=ARCH-ALL
189189

190-
// ARCH-ALL: clang{{.*}} -o {{.*}}.img --target=amdgcn-amd-amdhsa -mcpu=gfx90a -O2 -foffload-lto -Wl,--no-undefined {{.*}}.o {{.*}}.o
191-
// ARCH-ALL: clang{{.*}} -o {{.*}}.img --target=amdgcn-amd-amdhsa -mcpu=gfx908 -O2 -foffload-lto -Wl,--no-undefined {{.*}}.o {{.*}}.o
190+
// ARCH-ALL: clang{{.*}} -o {{.*}}.img --target=amdgcn-amd-amdhsa -mcpu=gfx90a -O2 -flto -Wl,--no-undefined {{.*}}.o {{.*}}.o
191+
// ARCH-ALL: clang{{.*}} -o {{.*}}.img --target=amdgcn-amd-amdhsa -mcpu=gfx908 -O2 -flto -Wl,--no-undefined {{.*}}.o {{.*}}.o
192192

193193
// RUN: clang-offload-packager -o %t.out \
194194
// RUN: --image=file=%t.elf.o,kind=openmp,triple=x86_64-unknown-linux-gnu \

‎clang/tools/clang-linker-wrapper/ClangLinkerWrapper.cpp

Lines changed: 2 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -1530,22 +1530,17 @@ Expected<StringRef> clang(ArrayRef<StringRef> InputFiles, const ArgList &Args,
15301530
};
15311531

15321532
// Forward all of the `--offload-opt` and similar options to the device.
1533+
CmdArgs.push_back("-flto");
15331534
for (auto &Arg : Args.filtered(OPT_offload_opt_eq_minus, OPT_mllvm))
15341535
CmdArgs.append(
15351536
{"-Xlinker",
15361537
Args.MakeArgString("--plugin-opt=" + StringRef(Arg->getValue()))});
15371538

1538-
if (Triple.isNVPTX() || Triple.isAMDGPU()) {
1539-
CmdArgs.push_back("-foffload-lto");
1540-
} else {
1541-
CmdArgs.push_back("-flto");
1542-
}
1543-
15441539
if (!Triple.isNVPTX() && !Triple.isSPIRV())
15451540
CmdArgs.push_back("-Wl,--no-undefined");
15461541

15471542
if (IsSYCLKind && Triple.isNVPTX())
1548-
CmdArgs.push_back("-S");
1543+
CmdArgs.push_back("-Wl,--lto-emit-asm");
15491544
for (StringRef InputFile : InputFiles)
15501545
CmdArgs.push_back(InputFile);
15511546

‎sycl/CMakeLists.txt

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -474,6 +474,10 @@ set( SYCL_TOOLCHAIN_DEPLOY_COMPONENTS
474474
${SYCL_TOOLCHAIN_DEPS}
475475
)
476476

477+
if("cuda" IN_LIST SYCL_ENABLE_BACKENDS)
478+
list(APPEND SYCL_TOOLCHAIN_DEPLOY_COMPONENTS clang-nvlink-wrapper)
479+
endif()
480+
477481
if (WIN32)
478482
list(APPEND SYCL_TOOLCHAIN_DEPLOY_COMPONENTS ur_win_proxy_loader)
479483
endif()

0 commit comments

Comments
 (0)