[clang] ba57828 - [CUDA][OpenMP] Fix the new driver crashing on multiple device-only outputs
Joseph Huber via cfe-commits
cfe-commits at lists.llvm.org
Wed Aug 24 06:48:06 PDT 2022
Author: Joseph Huber
Date: 2022-08-24T08:47:55-05:00
New Revision: ba57828e11c52f45420f2027dae0843cf83a57bf
URL: https://github.com/llvm/llvm-project/commit/ba57828e11c52f45420f2027dae0843cf83a57bf
DIFF: https://github.com/llvm/llvm-project/commit/ba57828e11c52f45420f2027dae0843cf83a57bf.diff
LOG: [CUDA][OpenMP] Fix the new driver crashing on multiple device-only outputs
The new driver supports device-only compilation for the offloading
device. The way this is handlded is a little different from the old
offloading driver. The old driver would put all the outputs in the final
action list akin to a linker job. The new driver however generated these
in the middle of the host's job so we instead put them all in a single
offloading action. However, we only handled these kinds of offloading
actions correctly when there was only a single input. When we had
multiple inputs we would instead attempt to get the host job, which
didn't exist, and crash.
This patch simply adds some extra logic to generate the jobs for all
dependencies if there is not host action.
Reviewed By: yaxunl
Differential Revision: https://reviews.llvm.org/D132248
Added:
Modified:
clang/lib/Driver/Driver.cpp
clang/test/Driver/cuda-bindings.cu
clang/test/Driver/hip-binding.hip
Removed:
################################################################################
diff --git a/clang/lib/Driver/Driver.cpp b/clang/lib/Driver/Driver.cpp
index d00f08d15ae5..f5b0cb85788b 100644
--- a/clang/lib/Driver/Driver.cpp
+++ b/clang/lib/Driver/Driver.cpp
@@ -4716,10 +4716,13 @@ void Driver::BuildJobs(Compilation &C) const {
// we are also generating .o files. So we allow more than one output file in
// this case as well.
//
+ // OffloadClass of type TY_Nothing: device-only output will place many outputs
+ // into a single offloading action. We should count all inputs to the action
+ // as outputs.
if (FinalOutput) {
unsigned NumOutputs = 0;
unsigned NumIfsOutputs = 0;
- for (const Action *A : C.getActions())
+ for (const Action *A : C.getActions()) {
if (A->getType() != types::TY_Nothing &&
!(A->getKind() == Action::IfsMergeJobClass ||
(A->getType() == clang::driver::types::TY_IFS_CPP &&
@@ -4728,6 +4731,10 @@ void Driver::BuildJobs(Compilation &C) const {
(A->getKind() == Action::BindArchClass && A->getInputs().size() &&
A->getInputs().front()->getKind() == Action::IfsMergeJobClass)))
++NumOutputs;
+ else if (A->getKind() == Action::OffloadClass &&
+ A->getType() == types::TY_Nothing)
+ NumOutputs += A->size();
+ }
if (NumOutputs > 1) {
Diag(clang::diag::err_drv_output_argument_with_multiple_files);
@@ -5265,20 +5272,21 @@ InputInfoList Driver::BuildJobsForActionNoCache(
// \
// Device Action 1 ---> OffloadAction -> Device Action 2
//
- // For a) and b), we just return the job generated for the dependence. For
+ // For a) and b), we just return the job generated for the dependences. For
// c) and d) we override the current action with the host/device dependence
// if the current toolchain is host/device and set the offload dependences
// info with the jobs obtained from the device/host dependence(s).
- // If there is a single device option, just generate the job for it.
- if (OA->hasSingleDeviceDependence()) {
+ // If there is a single device option or has no host action, just generate
+ // the job for it.
+ if (OA->hasSingleDeviceDependence() || !OA->hasHostDependence()) {
InputInfoList DevA;
OA->doOnEachDeviceDependence([&](Action *DepA, const ToolChain *DepTC,
const char *DepBoundArch) {
- DevA =
- BuildJobsForAction(C, DepA, DepTC, DepBoundArch, AtTopLevel,
- /*MultipleArchs*/ !!DepBoundArch, LinkingOutput,
- CachedResults, DepA->getOffloadingDeviceKind());
+ DevA.append(BuildJobsForAction(C, DepA, DepTC, DepBoundArch, AtTopLevel,
+ /*MultipleArchs*/ !!DepBoundArch,
+ LinkingOutput, CachedResults,
+ DepA->getOffloadingDeviceKind()));
});
return DevA;
}
diff --git a/clang/test/Driver/cuda-bindings.cu b/clang/test/Driver/cuda-bindings.cu
index 7115117e79ce..a7aa4c797885 100644
--- a/clang/test/Driver/cuda-bindings.cu
+++ b/clang/test/Driver/cuda-bindings.cu
@@ -146,3 +146,20 @@
// RUN: --cuda-gpu-arch=sm_52 --cuda-device-only -c -o foo.o %s 2>&1 \
// RUN: | FileCheck -check-prefix=D_ONLY %s
// D_ONLY: "foo.o"
+
+//
+// Check to make sure we can generate multiple outputs for device-only
+// compilation and fail with '-o'.
+//
+// RUN: %clang -### -target powerpc64le-ibm-linux-gnu --offload-new-driver -ccc-print-bindings \
+// RUN: --offload-arch=sm_70 --offload-arch=sm_52 --offload-device-only -c %s 2>&1 \
+// RUN: | FileCheck -check-prefix=MULTI-D-ONLY %s
+// MULTI-D-ONLY: # "nvptx64-nvidia-cuda" - "clang", inputs: ["[[INPUT:.+]]"], output: "[[PTX_70:.+]]"
+// MULTI-D-ONLY-NEXT: # "nvptx64-nvidia-cuda" - "NVPTX::Assembler", inputs: ["[[PTX_70]]"], output: "[[CUBIN_70:.+]]"
+// MULTI-D-ONLY-NEXT: # "nvptx64-nvidia-cuda" - "clang", inputs: ["[[INPUT]]"], output: "[[PTX_52:.+]]"
+// MULTI-D-ONLY-NEXT: # "nvptx64-nvidia-cuda" - "NVPTX::Assembler", inputs: ["[[PTX_52]]"], output: "[[CUBIN_52:.+]]"
+//
+// RUN: %clang -### -target powerpc64le-ibm-linux-gnu --offload-new-driver -ccc-print-bindings \
+// RUN: --offload-arch=sm_70 --offload-arch=sm_52 --offload-device-only -c -o %t %s 2>&1 \
+// RUN: | FileCheck -check-prefix=MULTI-D-ONLY-O %s
+// MULTI-D-ONLY-O: error: cannot specify -o when generating multiple output files
diff --git a/clang/test/Driver/hip-binding.hip b/clang/test/Driver/hip-binding.hip
index 6a27558dbf03..e9ced3004237 100644
--- a/clang/test/Driver/hip-binding.hip
+++ b/clang/test/Driver/hip-binding.hip
@@ -51,3 +51,20 @@
// NORDC-NOT: offload bundler
// NORDC: # "x86_64-unknown-linux-gnu" - "GNU::Linker", inputs: ["{{.*o}}"], output: "a.out"
+
+//
+// Check to make sure we can generate multiple outputs for device-only
+// compilation and fail with '-o'.
+//
+// RUN: %clang -### -target x86_64-linux-gnu --offload-new-driver -ccc-print-bindings \
+// RUN: --offload-arch=gfx90a --offload-arch=gfx908 --offload-device-only -c %s 2>&1 \
+// RUN: | FileCheck -check-prefix=MULTI-D-ONLY %s
+// MULTI-D-ONLY: # "amdgcn-amd-amdhsa" - "clang", inputs: ["[[INPUT:.+]]"], output: "[[GFX908:.+]]"
+// MULTI-D-ONLY-NEXT: # "amdgcn-amd-amdhsa" - "AMDGCN::Linker", inputs: ["[[GFX908]]"], output: "[[GFX908_OUT:.+]]"
+// MULTI-D-ONLY-NEXT: # "amdgcn-amd-amdhsa" - "clang", inputs: ["[[INPUT]]"], output: "[[GFX90a:.+]]"
+// MULTI-D-ONLY-NEXT: # "amdgcn-amd-amdhsa" - "AMDGCN::Linker", inputs: ["[[GFX90a]]"], output: "[[GFX90a_OUT:.+]]"
+//
+// RUN: %clang -### -target x86_64-linux-gnu --offload-new-driver -ccc-print-bindings \
+// RUN: --offload-arch=gfx90a --offload-arch=gfx908 --offload-device-only -c -o %t %s 2>&1 \
+// RUN: | FileCheck -check-prefix=MULTI-D-ONLY-O %s
+// MULTI-D-ONLY-O: error: cannot specify -o when generating multiple output files
More information about the cfe-commits
mailing list