Currently if CUDA/HIP users use template class with virtual dtor
and std::string data member with C++20 and MSVC. When the template
class is explicitly instantiated, there is error about host
function called by host device function (used to be undefined
symbols in linking stage before member destructors were checked
by deferred diagnostics).
It was caused by clang inferring host/device attributes for
default dtors. Since all dtors of member and parent classes
have implicit host device attrs, clang infers the virtual dtor have
implicit host and device attrs. Since virtual dtor of
explicitly instantiated template class must be emitted,
this causes constexpr dtor of std::string emitted, which
calls a host function which was not emitted on device side.
This is a serious issue since it prevents users from
using std::string with C++20 on Windows.
When inferring host device attr of virtual dtor of explicit
template class instantiation, clang should be conservative
since it is sure to be emitted. Since an implicit host device
function may call a host function, clang cannot assume it is
always available on device. This guarantees dtors that
may call host functions not to have implicit device attr,
therefore will not be emitted on device side.
Fixes: https://github.com/llvm/llvm-project/issues/108548
Fixes: SWDEV-517435
Currently the deferred diag fails to diagnose calling of host function
in host device function in device compilation triggered by destructors.
This can be further divided into two issuse:
1. the deferred diag visitor does not visit dtor of member and parent
class when visiting dtor, which it should
2. the deferred diag visitor does not visit virtual dtor of explicit
template class instantiation, which it should
Due to these issues, some constexpr functions which call host functions
are emitted on device side, which causes undefind symbols in linking
stage, as revealed by
https://github.com/llvm/llvm-project/issues/108548
By fixing these issue, clang will diag the issues early during
compilation instead of linking.
This is take two of #70976. This iteration of the patch makes sure that
custom
diagnostics without any warning group don't get promoted by `-Werror` or
`-Wfatal-errors`.
This implements parts of the extension proposed in
https://discourse.llvm.org/t/exposing-the-diagnostic-engine-to-c/73092/7.
Specifically, this makes it possible to specify a diagnostic group in an
optional third argument.
This reverts commit e39205654dc11c50bd117e8ccac243a641ebd71f.
There are further discussions in
https://github.com/llvm/llvm-project/pull/70976, happening for past two
weeks. Since there were no responses for couple weeks now, reverting
until author is back.
This reverts commit e7f782e7481cea23ef452a75607d3d61f5bd0d22.
This had UBSan failures:
[----------] 1 test from ConfigCompileTests
[ RUN ] ConfigCompileTests.DiagnosticSuppression
Config fragment: compiling <unknown>:0 -> 0x00007B8366E2F7D8 (trusted=false)
/usr/local/google/home/fmayer/large/llvm-project/llvm/include/llvm/ADT/IntrusiveRefCntPtr.h:203:33: runtime error: reference binding to null pointer of type 'clang::DiagnosticIDs'
UndefinedBehaviorSanitizer: undefined-behavior /usr/local/google/home/fmayer/large/llvm-project/llvm/include/llvm/ADT/IntrusiveRefCntPtr.h:203:33
Pull Request: https://github.com/llvm/llvm-project/pull/108645
Through the new `-foffload-via-llvm` flag, CUDA kernels can now be
lowered to the LLVM/Offload API. On the Clang side, this is simply done
by using the OpenMP offload toolchain and emitting calls to `llvm*`
functions to orchestrate the kernel launch rather than `cuda*`
functions. These `llvm*` functions are implemented on top of the
existing LLVM/Offload API.
As we are about to redefine the Offload API, this wil help us in the
design process as a second offload language.
We do not support any CUDA APIs yet, however, we could:
https://www.osti.gov/servlets/purl/1892137
For proper host execution we need to resurrect/rebase
https://tianshilei.me/wp-content/uploads/2021/12/llpp-2021.pdf
(which was designed for debugging).
```
❯❯❯ cat test.cu
extern "C" {
void *llvm_omp_target_alloc_shared(size_t Size, int DeviceNum);
void llvm_omp_target_free_shared(void *DevicePtr, int DeviceNum);
}
__global__ void square(int *A) { *A = 42; }
int main(int argc, char **argv) {
int DevNo = 0;
int *Ptr = reinterpret_cast<int *>(llvm_omp_target_alloc_shared(4, DevNo));
*Ptr = 7;
printf("Ptr %p, *Ptr %i\n", Ptr, *Ptr);
square<<<1, 1>>>(Ptr);
printf("Ptr %p, *Ptr %i\n", Ptr, *Ptr);
llvm_omp_target_free_shared(Ptr, DevNo);
}
❯❯❯ clang++ test.cu -O3 -o test123 -foffload-via-llvm --offload-arch=native
❯❯❯ llvm-objdump --offloading test123
test123: file format elf64-x86-64
OFFLOADING IMAGE [0]:
kind elf
arch gfx90a
triple amdgcn-amd-amdhsa
producer openmp
❯❯❯ LIBOMPTARGET_INFO=16 ./test123
Ptr 0x155448ac8000, *Ptr 7
Ptr 0x155448ac8000, *Ptr 42
```
nvcc warns about the following code:
`void f();
__device__ void f() {}`
but clang does not since clang allows device function to overload host
function.
Users want clang to emit similar warning to help code to be compatible
with nvcc.
Since this may cause regression with existing code, the warning is off
by default and can be enabled by -Wnvcc-compat.
It won't cause warning in system headers, even with -Wnvcc-compat.
In -fgpu-rdc mode, when an external kernel is used by a host function
with weak_odr linkage (e.g. explicitly instantiated template function),
the kernel should not be marked as host-used external kernel, since the
host function may be dropped by the linker. Mark the external kernel as
host-used external kernel will force a reference to the external kernel,
which the user may not define in other TU.
Fixes: https://github.com/llvm/llvm-project/issues/83771
Treat ctor/dtor in device var init as host device function
so that they can be used to initialize file-scope
device variables to match nvcc behavior. If they are non-trivial
they will be diagnosed.
We cannot add implicit host device attrs to non-trivial
ctor/dtor since determining whether they are non-trivial
needs to know whether they have a trivial body and all their
member and base classes' ctor/dtor have trivial body, which
is affected by where their bodies are defined or instantiated.
Fixes: #72261
Fixes: SWDEV-432412
This reverts commit 27e6e4a4d0e3296cebad8db577ec0469a286795e.
This patch is reverted due to regression. A testcase is:
`template <class T>
struct ptr {
~ptr() { static int x = 1;}
};
template <class T>
struct Abc : ptr<T> {
public:
Abc();
~Abc() {}
};
template
class Abc<int>;
`
Make trivial ctor/dtor implicitly host device functions so that they can
be used to initialize file-scope
device variables to match nvcc behavior.
Fixes: https://github.com/llvm/llvm-project/issues/72261
Fixes: SWDEV-432412
Added option -foffload-implicit-host-device-templates which is off by
default.
When the option is on, template functions and specializations without
host/device attributes have implicit host device attributes.
They can be overridden by device template functions with the same
signagure.
They are emitted on device side only if they are used on device side.
This feature is added as an extension.
`__has_extension(cuda_implicit_host_device_templates)` can be used to
check whether it is enabled.
This is to facilitate using standard C++ headers for device.
Fixes: https://github.com/llvm/llvm-project/issues/69956
Fixes: SWDEV-428314
Currently clang diagnoses the following code:
(https://godbolt.org/z/s8zK3E5P5) but nvcc
does not.
`
struct A {
constexpr A(){}
};
struct B {
A a;
int b;
};
template<typename T>
__global__ void kernel( )
{
__shared__ B x;
}
`
Clang generates an implicit trivial ctor for struct B, which should be
allowed for initializing a shared variable.
However, the body of the ctor is defined only if the template kernel is
instantiated. Clang checks the initialization of variable in
non-instantiated templates, where it cannot find the body of the ctor,
therefore diagnoses it.
This patch skips the check for non-instantiated templates.
The bounds of a c++ array is a _constant-expression_. And in C++ it is
also a constant expression.
But we also support VLAs, ie arrays with non-constant bounds.
We need to take care to handle the case of a consteval function (which
are specified to be only immediately called in non-constant contexts)
that appear in arrays bounds.
This introduces `Sema::isAlwayConstantEvaluatedContext`, and a flag in
ExpressionEvaluationContextRecord, such that immediate functions in
array bounds are always immediately invoked.
Sema had both `isConstantEvaluatedContext` and
`isConstantEvaluated`, so I took the opportunity to cleanup that.
The change in `TimeProfilerTest.cpp` is an unfortunate manifestation of
the problem that #66203 seeks to address.
Fixes#65520
This patch adds the Sema changes needed for enabling HIP parallel algorithm offload on AMDGPU targets. This change impacts the CUDA / HIP language specific checks, and only manifests if compiling in `hipstdpar` mode. In this case, we essentially do three things:
1. Allow device side callers to call host side callees - since the user visible HLL would be standard C++, with no annotations / restriction mechanisms, we cannot unambiguously establish that such a call is an error, so we conservatively allow all such calls, deferring actual cleanup to a subsequent pass over IR;
2. Allow host formed lambdas to capture by reference;
3. Allow device functions to use host global variables.
Reviewed by: yaxunl
Differential Revision: https://reviews.llvm.org/D155833
https://reviews.llvm.org/D158247 caused regressions for HIP on Windows
and was reverted.
A reduced test case is:
```
typedef void (__stdcall* funcTy)();
void invoke(funcTy f);
static void __stdcall callee() noexcept {
}
void foo() {
invoke(callee);
}
```
It is due to clang missing handling host/device attributes for calling
convention at a few places
This patch fixes that.
This reverts commit de0df639724b10001ea9a74539381ea494296be9.
It was reverted due to regression in HIP unit test on Windows:
In file included from C:\hip-tests\catch\unit\graph\hipGraphClone.cc:37:
In file included from C:\hip-tests\catch\.\include\hip_test_common.hh:24:
In file included from C:\hip-tests\catch\.\include/hip_test_context.hh:24:
In file included from C:/install/native/Release/x64/hip/include\hip/hip_runtime.h:54:
C:/dk/win\vc\14.31.31107\include\thread:76:70: error: cannot initialize a parameter of type '_beginthreadex_proc_type' (aka 'unsigned int (*)(void *) __attribute__((stdcall))') with an lvalue of type 'const unsigned int (*)(void *) noexcept __attribute__((stdcall))': different exception specifications
76 | reinterpret_cast<void*>(_CSTD _beginthreadex(nullptr, 0, _Invoker_proc, _Decay_copied.get(), 0, &_Thr._Id));
| ^~~~~~~~~~~~~
C:\hip-tests\catch\unit\graph\hipGraphClone.cc:290:21) &>' requested here
90 | _Start(_STD forward<_Fn>(_Fx), _STD forward<_Args>(_Ax)...);
| ^
C:\hip-tests\catch\unit\graph\hipGraphClone.cc:290:21) &, 0>' requested here
311 | std::thread t(lambdaFunc);
| ^
C:/dk/win\ms_wdk\e22621\Include\10.0.22621.0\ucrt\process.h:99:40: note: passing argument to parameter '_StartAddress' here
99 | _In_ _beginthreadex_proc_type _StartAddress,
| ^
1 error generated when compiling for gfx1030.
Currently, clang does not resolve certain overloaded functions correctly in the initializer
of global variables, e.g.
template<typename T1, typename U>
T1 mypow(T1, U);
__attribute__((device)) double mypow(double, int);
double t_extent = mypow(1.0, 2);
In the above example, mypow is supposed to resolve to the host version
but clang resolves it to the device version instead, and emits an error
(https://godbolt.org/z/17xxzaa67).
However, if the variable is assigned in a host function, there is no error.
The discrepancy in overloading resolution inside and outside of
a function is due to clang not accounting for the host/device target
when resolving functions called in the initializer of a global variable.
This patch introduces a global host/device target context for CUDA/HIP
for functions called outside of functions. For global variable initialization,
it is determined by the host/device attribute of the variable. For other
situations, a default value of host_device is sufficient.
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D158247
Fixes: SWDEV-416731
std::optional::value() has undesired exception checking semantics and is
unavailable in older Xcode (see _LIBCPP_AVAILABILITY_BAD_OPTIONAL_ACCESS). The
call sites block std::optional migration.
This makes `ninja clang` work in the absence of llvm::Optional::value.
For -fgpu-rdc, a host function may call an external kernel
which is defined in an archive of bitcode. Since this external
kernel is only referenced in host function, the device
bitcode does not contain reference to this external
kernel, then the linker will not try to resolve this external
kernel in the archive.
To fix this issue, host-used external kernels and device
variables are tracked. A global array containing pointers
to these external kernels and variables is emitted which
serves as an artificial references to the external kernels
and variables used by host.
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D123441
CUDA/HIP determines whether a function can be called based on
the device/host attributes of callee and caller. Clang assumes the
caller is CurContext. This is correct in most cases, however, it is
not correct in OpenMP parallel region when CUDA/HIP program
is compiled with -fopenmp. This causes incorrect overloading
resolution and missed diagnostics.
To get the correct caller, clang needs to chase the parent chain
of DeclContext starting from CurContext until a function decl
or a lambda decl is reached. Sema API is adapted to achieve that
and used to determine the caller in hostness check.
Reviewed by: Artem Belevich, Richard Smith
Differential Revision: https://reviews.llvm.org/D121765
constexpr var may be initialized with address of non-const variable.
In this case the initializer is not constant in device compilation.
This has been handled for const vars but not for constexpr vars.
This patch makes handling of const var and constexpr var
consistent.
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D119615
Fixes: https://github.com/llvm/llvm-project/issues/53780
Currently clang treats host var address as constant in device compilation,
which causes const vars initialized with host var address promoted to
device variables incorrectly and results in undefined symbols.
This patch fixes that.
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D118153
Fixes: SWDEV-309881
Change-Id: I0a69357063c6f8539ef259c96c250d04615f4473
HIP currently diagnose capture of this pointer in device lambda in
host member functions. If this pointer points to managed memory,
it can be used in both device and host functions. Under this
situation, capturing this pointer in device lambda functions
in host member functions is valid usage. Change the diagnostic
about capturing this pointer to warning.
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D108493
Recently we added diagnosing ODR-use of host variables
in device functions, which includes ODR-use of const
host variables since they are not really emitted on
device side. This caused regressions since we used
to allow ODR-use of const host variables in device
functions.
This patch allows ODR-use of const variables in device
functions if the const variables can be statically initialized
and have an empty dtor. Such variables are marked with
implicit constant attrs and emitted on device side. This is
in line with what clang does for constexpr variables.
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D103108
variables emitted on both host and device side with different addresses
when ODR-used by host function should not cause device side counter-part
to be force emitted.
This fixes the regression caused by https://reviews.llvm.org/D102237
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D102801
Defer constant checking of dependent initializer to template instantiation
since it cannot be done for dependent values.
Reviewed by: Artem Belevich
Differential Revision: https://reviews.llvm.org/D95840
Defaulted destructor was treated inconsistently, compared to other
compiler-generated functions.
When Sema::IdentifyCUDATarget() got called on just-created dtor which didn't
have implicit __host__ __device__ attributes applied yet, it would treat it as a
host function. That happened to (sometimes) hide the error when dtor referred
to a host-only functions.
Even when we had identified defaulted dtor as a HD function, we still treated it
inconsistently during selection of usual deallocators, where we did not allow
referring to wrong-side functions, while it is allowed for other HD functions.
This change brings handling of defaulted dtors in line with other HD functions.
Differential Revision: https://reviews.llvm.org/D94732
This patch diagnoses invalid references of global host variables in device,
global, or host device functions.
Differential Revision: https://reviews.llvm.org/D91281
