[offload][lit] XFAIL all failing tests on the Level Zero plugin (#174804)

We finally got our buildbot added (to staging, at least) so we want to
start running L0 tests in CI.
We need `check-offload` to pass though, so XFAIL everything failing.
There's a couple `UNSUPPORTED` as well, those are for sporadic fails.

Also make set the `gpu` and `intelgpu` LIT variables when testing the
`spirv64-intel` triple.

We have no DeviceRTL yet so basically everything fails, but we manage to
get

```
Total Discovered Tests: 432
Unsupported      : 169 (39.12%)
Passed           :  67 (15.51%)
Expectedly Failed: 196 (45.37%)
```

We still don't build the level zero plugin by default and these tests
don't run unless the plugin was built, so this has no effect on most
builds.

---------

Signed-off-by: Nick Sarnie <nick.sarnie@intel.com>

offload/test/api/assert.c

@@ -1,5 +1,6 @@
 // RUN: %libomptarget-compile-run-and-check-generic
 // RUN: %libomptarget-compileopt-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <assert.h>
 #include <stdio.h>

offload/test/api/omp_device_alloc.c

@@ -2,6 +2,7 @@
 
 // UNSUPPORTED: nvidiagpu
 // UNSUPPORTED: amdgpu
+// XFAIL: intelgpu
 
 #include <assert.h>
 #include <omp.h>

offload/test/api/omp_device_managed_memory.c

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <stdio.h>

offload/test/api/omp_device_managed_memory_alloc.c

@@ -1,5 +1,6 @@
 // RUN: %libomptarget-compile-run-and-check-generic
 // RUN: %libomptarget-compileopt-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <stdio.h>

offload/test/api/omp_device_memory.c

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <stdio.h>

offload/test/api/omp_device_uid.c

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <stdio.h>

offload/test/api/omp_dynamic_shared_memory.c

@@ -7,6 +7,7 @@
 // RUN:   %libomptarget-run-generic | %fcheck-generic
 
 // REQUIRES: gpu
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <stdio.h>

offload/test/api/omp_env_vars.c

@@ -2,6 +2,7 @@
 // RUN: env OMP_NUM_TEAMS=1 OMP_TEAMS_THREAD_LIMIT=1 LIBOMPTARGET_INFO=16 \
 // RUN:   %libomptarget-run-generic 2>&1 | %fcheck-generic
 // REQUIRES: gpu
+// XFAIL: intelgpu
 
 #define N 256
 

offload/test/api/omp_get_device_num.c

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <stdio.h>

offload/test/api/omp_get_mapped_ptr.c

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-and-run-generic
+// XFAIL: intelgpu
 
 #include <assert.h>
 #include <omp.h>

offload/test/api/omp_get_num_devices.c

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <stdio.h>

offload/test/api/omp_get_num_procs.c

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <stdio.h>
 

offload/test/api/omp_host_call.c

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <assert.h>
 #include <omp.h>

offload/test/api/omp_host_pinned_memory.c

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <stdio.h>

offload/test/api/omp_host_pinned_memory_alloc.c

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <stdio.h>

offload/test/api/omp_indirect_call.c

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <assert.h>
 #include <stdio.h>

offload/test/api/omp_indirect_call_table_manual.c

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 #include <assert.h>
 #include <omp.h>
 #include <stdio.h>

offload/test/api/ompx_3d.c

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <ompx.h>

offload/test/api/ompx_3d.cpp

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compilexx-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <ompx.h>

offload/test/api/ompx_sync.c

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <ompx.h>

offload/test/api/ompx_sync.cpp

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compilexx-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <ompx.h>

offload/test/env/base_ptr_ref_count.c

@@ -1,5 +1,8 @@
+// clang-format off
 // RUN: %libomptarget-compile-generic && env LIBOMPTARGET_DEBUG=1 %libomptarget-run-generic 2>&1 | %fcheck-generic
 // REQUIRES: libomptarget-debug
+// XFAIL: intelgpu
+// clang-format on
 
 #include <stdio.h>
 #include <stdlib.h>

offload/test/env/omp_target_debug.c

@@ -1,6 +1,9 @@
+// clang-format off
 // RUN: %libomptarget-compile-generic && env LIBOMPTARGET_DEBUG=1 %libomptarget-run-generic 2>&1 | %fcheck-generic -allow-empty -check-prefix=DEBUG
 // RUN: %libomptarget-compile-generic && env LIBOMPTARGET_DEBUG=0 %libomptarget-run-generic 2>&1 | %fcheck-generic -allow-empty -check-prefix=NDEBUG
 // REQUIRES: libomptarget-debug
+// XFAIL: intelgpu
+// clang-format on
 
 int main(void) {
 #pragma omp target

offload/test/jit/empty_kernel_lvl1.c

@@ -29,5 +29,6 @@
 // clang-format on
 
 // REQUIRES: gpu
+// XFAIL: intelgpu
 
 #include "empty_kernel.inc"

offload/test/jit/empty_kernel_lvl2.c

@@ -24,5 +24,6 @@
 // clang-format on
 
 // REQUIRES: gpu
+// XFAIL: intelgpu
 
 #include "empty_kernel.inc"

offload/test/jit/type_punning.c

@@ -9,6 +9,7 @@
 // clang-format on
 
 // REQUIRES: gpu
+// XFAIL: intelgpu
 
 // Ensure that there is only the kernel function left, not any outlined
 // parallel regions.

offload/test/libc/malloc_parallel.c

@@ -1,5 +1,6 @@
 // RUN: %libomptarget-compile-run-and-check-generic
 // RUN: %libomptarget-compileopt-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <stdio.h>

offload/test/lit.cfg

@@ -199,6 +199,9 @@ if config.libomptarget_current_target.startswith('nvptx'):
 if config.libomptarget_current_target.startswith('amdgcn'):
     config.available_features.add('gpu')
     config.available_features.add('amdgpu')
+if config.libomptarget_current_target.startswith('spirv64-intel'):
+    config.available_features.add('gpu')
+    config.available_features.add('intelgpu')
 if config.libomptarget_current_target in host_targets:
     config.available_features.add('host')
 
@@ -213,6 +216,9 @@ def remove_suffix_if_present(name):
 def add_libraries(source):
     if "gpu" not in config.available_features:
         return source
+    if "intelgpu" in config.available_features:
+        # There is no DeviceRTL for Intel yet and libc doesn't work.
+        return source
     if config.libomptarget_has_libc:
         return source + " -Xoffload-linker -lc " + \
                "-Xoffload-linker -lm " + \

offload/test/mapping/array_section_implicit_capture.c

@@ -1,6 +1,7 @@
 // RUN: %libomptarget-compile-generic
 // RUN: %libomptarget-run-generic 2>&1 \
 // RUN: | %fcheck-generic
+// XFAIL: intelgpu
 
 #include <stdio.h>
 #include <stdlib.h>

offload/test/mapping/chained_containing_structs_1.cc

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compilexx-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <cstdlib>
 #include <cstdio>

offload/test/mapping/chained_containing_structs_2.cc

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compilexx-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <cstdlib>
 #include <cstdio>

offload/test/mapping/chained_containing_structs_3.cc

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compilexx-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <cstdlib>
 #include <cstdio>

offload/test/mapping/data_member_ref.cpp

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compilexx-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <stdio.h>
 

offload/test/mapping/declare_mapper_nested_default_mappers.cpp

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compilexx-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <cstdio>
 #include <cstdlib>

offload/test/mapping/declare_mapper_nested_default_mappers_1.cpp

@@ -1,6 +1,7 @@
 // RUN: %libomptarget-compilexx-run-and-check-generic
 
 // UNSUPPORTED: amdgcn-amd-amdhsa
+// XFAIL: intelgpu
 
 extern "C" int printf(const char *, ...);
 

offload/test/mapping/declare_mapper_nested_mappers.cpp

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compilexx-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <cstdio>
 #include <cstdlib>

offload/test/mapping/declare_mapper_target.cpp

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <cstdio>
 #include <cstdlib>

offload/test/mapping/declare_mapper_target_data.cpp

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <cstdio>
 #include <cstdlib>

offload/test/mapping/declare_mapper_target_data_enter_exit.cpp

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <cstdio>
 #include <cstdlib>

offload/test/mapping/declare_mapper_target_update.cpp

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <cstdio>
 #include <cstdlib>

offload/test/mapping/delete_inf_refcount.c

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <stdio.h>

offload/test/mapping/device_ptr_update.c

@@ -2,6 +2,7 @@
 // RUN: env LIBOMPTARGET_DEBUG=1 %libomptarget-run-generic 2>&1 \
 // RUN: | %fcheck-generic -check-prefix=DEBUG -check-prefix=CHECK
 // REQUIRES: libomptarget-debug
+// XFAIL: intelgpu
 
 #include <stdio.h>
 

offload/test/mapping/duplicate_mappings_1.cpp

@@ -1,5 +1,6 @@
 // clang-format off
 // RUN: %libomptarget-compilexx-generic -Wno-openmp-mapping && %libomptarget-run-generic
+// XFAIL: intelgpu
 
 // clang-format on
 

offload/test/mapping/duplicate_mappings_2.cpp

@@ -1,5 +1,6 @@
 // clang-format off
 // RUN: %libomptarget-compilexx-generic -Wno-openmp-mapping && %libomptarget-run-generic
+// XFAIL: intelgpu
 
 #include <assert.h>
 

offload/test/mapping/firstprivate_aligned.cpp

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compilexx-generic -O3 && %libomptarget-run-generic
+// XFAIL: intelgpu
 
 #include <stdio.h>
 

offload/test/mapping/has_device_addr.cpp

@@ -1,6 +1,7 @@
 // RUN: %libomptarget-compilexx-generic -fopenmp-version=51
 // RUN: %libomptarget-run-generic 2>&1 \
 // RUN: | %fcheck-generic
+// XFAIL: intelgpu
 
 #include <assert.h>
 #include <iostream>

offload/test/mapping/implicit_device_ptr.c

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <stdio.h>

offload/test/mapping/is_accessible.cpp

@@ -8,6 +8,7 @@
 
 // REQUIRES: unified_shared_memory
 // XFAIL: nvptx
+// XFAIL: intelgpu
 
 // CHECK: SUCCESS
 // NO_USM: Not accessible

offload/test/mapping/is_device_ptr.cpp

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compilexx-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <assert.h>
 #include <iostream>

offload/test/mapping/lambda_by_value.cpp

@@ -1,5 +1,6 @@
 // RUN: %libomptarget-compileopt-generic -fno-exceptions
 // RUN: %libomptarget-run-generic 2>&1 | %fcheck-generic
+// XFAIL: intelgpu
 
 #include <stdint.h>
 #include <stdio.h>

offload/test/mapping/low_alignment.c

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compilexx-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <stdio.h>

offload/test/mapping/map_both_pointer_pointee.c

@@ -5,6 +5,7 @@
 //
 // FIXME: https://github.com/llvm/llvm-project/issues/161265
 // XFAIL: nvidiagpu
+// XFAIL: intelgpu
 
 #pragma omp declare target
 int *ptr1;

offload/test/mapping/ompx_hold/struct.c

@@ -1,5 +1,6 @@
 // RUN: %libomptarget-compile-generic -fopenmp-extensions
 // RUN: %libomptarget-run-generic | %fcheck-generic -strict-whitespace
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <stdio.h>

offload/test/mapping/ompx_hold/target.c

@@ -1,5 +1,6 @@
 // RUN: %libomptarget-compile-generic -fopenmp-extensions
 // RUN: %libomptarget-run-generic | %fcheck-generic -strict-whitespace
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <stdio.h>

offload/test/mapping/pr38704.c

@@ -3,6 +3,7 @@
 // Clang 6.0 doesn't use the new map interface, undefined behavior when
 // the compiler emits "old" interface code for structures.
 // UNSUPPORTED: clang-6
+// XFAIL: intelgpu
 
 #include <stdio.h>
 #include <stdlib.h>

offload/test/mapping/prelock.cpp

@@ -4,6 +4,7 @@
 // REQUIRES: gpu
 // UNSUPPORTED: nvidiagpu
 // UNSUPPORTED: amdgpu
+// XFAIL: intelgpu
 
 #include <cstdio>
 

offload/test/mapping/present/target.c

@@ -1,6 +1,7 @@
 // RUN: %libomptarget-compile-generic
 // RUN: %libomptarget-run-fail-generic 2>&1 \
 // RUN: | %fcheck-generic
+// XFAIL: intelgpu
 
 #include <stdio.h>
 

offload/test/mapping/present/target_array_extension.c

@@ -15,6 +15,7 @@
 // RUN:   -DEXTENDS=AFTER
 // RUN: %libomptarget-run-fail-generic 2>&1 \
 // RUN: | %fcheck-generic
+// XFAIL: intelgpu
 
 // END.
 

offload/test/mapping/present/unified_shared_memory.c

@@ -3,6 +3,7 @@
 // RUN: | %fcheck-generic
 
 // REQUIRES: unified_shared_memory
+// XFAIL: intelgpu
 
 #include <stdio.h>
 

offload/test/mapping/present/zero_length_array_section.c

@@ -1,6 +1,7 @@
 // RUN: %libomptarget-compile-generic
 // RUN: %libomptarget-run-fail-generic 2>&1 \
 // RUN: | %fcheck-generic
+// XFAIL: intelgpu
 
 #include <stdio.h>
 

offload/test/mapping/private_mapping.c

@@ -1,5 +1,6 @@
 // RUN: %libomptarget-compile-run-and-check-generic
 // UNSUPPORTED: amdgcn-amd-amdhsa
+// XFAIL: intelgpu
 
 #include <assert.h>
 #include <stdio.h>

offload/test/mapping/ptr_and_obj_motion.c

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compile-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <stdio.h>
 

offload/test/mapping/reduction_implicit_map.cpp

@@ -1,4 +1,5 @@
 // RUN: %libomptarget-compilexx-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <stdio.h>
 

offload/test/mapping/target_data_array_extension_at_exit.c

@@ -15,6 +15,7 @@
 // RUN:   -DEXTENDS=AFTER
 // RUN: %libomptarget-run-generic 2>&1 \
 // RUN: | %fcheck-generic
+// XFAIL: intelgpu
 
 // END.
 

offload/test/mapping/target_derefence_array_pointrs.cpp

@@ -6,6 +6,7 @@
 // UNSUPPORTED: amdgcn-amd-amdhsa
 // UNSUPPORTED: nvptx64-nvidia-cuda
 // UNSUPPORTED: nvptx64-nvidia-cuda-LTO
+// XFAIL: intelgpu
 
 #include <stdio.h>
 #include <stdlib.h>

offload/test/mapping/target_has_device_addr.c

@@ -3,6 +3,7 @@
 // RUN: | %fcheck-generic
 
 // UNSUPPORTED: amdgcn-amd-amdhsa
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <stdio.h>

offload/test/mapping/target_implicit_partial_map.c

@@ -1,6 +1,7 @@
 // RUN: %libomptarget-compile-generic
 // RUN: %libomptarget-run-generic 2>&1 \
 // RUN: | %fcheck-generic
+// XFAIL: intelgpu
 
 // END.
 

offload/test/mapping/target_map_for_member_data.cpp

@@ -1,6 +1,7 @@
 // RUN: %libomptarget-compile-generic -fopenmp-version=51
 // RUN: %libomptarget-run-generic 2>&1 \
 // RUN: | %fcheck-generic
+// XFAIL: intelgpu
 
 extern "C" int printf(const char *, ...);
 template <typename T> class A {

offload/test/mapping/target_uses_allocator.c

@@ -4,6 +4,7 @@
 // UNSUPPORTED: amdgcn-amd-amdhsa
 // UNSUPPORTED: nvptx64-nvidia-cuda
 // UNSUPPORTED: nvptx64-nvidia-cuda-LTO
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <stdio.h>

offload/test/mapping/use_device_addr/target_use_device_addr.c

@@ -1,6 +1,7 @@
 // RUN: %libomptarget-compile-generic -fopenmp-version=51
 // RUN: %libomptarget-run-generic 2>&1 \
 // RUN: | %fcheck-generic
+// XFAIL: intelgpu
 
 #include <stdio.h>
 int main() {

offload/test/offloading/CUDA/basic_launch.cu

@@ -9,6 +9,7 @@
 // UNSUPPORTED: aarch64-unknown-linux-gnu-LTO
 // UNSUPPORTED: x86_64-unknown-linux-gnu
 // UNSUPPORTED: x86_64-unknown-linux-gnu-LTO
+// XFAIL: intelgpu
 
 #include <stdio.h>
 

offload/test/offloading/CUDA/basic_launch_blocks_and_threads.cu

@@ -9,6 +9,7 @@
 // UNSUPPORTED: aarch64-unknown-linux-gnu-LTO
 // UNSUPPORTED: x86_64-unknown-linux-gnu
 // UNSUPPORTED: x86_64-unknown-linux-gnu-LTO
+// XFAIL: intelgpu
 
 #include <stdio.h>
 

offload/test/offloading/CUDA/basic_launch_multi_arg.cu

@@ -6,6 +6,7 @@
 // clang-format on
 
 // REQUIRES: gpu
+// XFAIL: intelgpu
 
 #include <stdio.h>
 

offload/test/offloading/CUDA/launch_tu.cu

@@ -9,6 +9,7 @@
 // UNSUPPORTED: aarch64-unknown-linux-gnu-LTO
 // UNSUPPORTED: x86_64-unknown-linux-gnu
 // UNSUPPORTED: x86_64-unknown-linux-gnu-LTO
+// XFAIL: intelgpu
 
 #include <stdio.h>
 

offload/test/offloading/assert.cpp

@@ -1,5 +1,6 @@
 // RUN: %libomptarget-compilexx-generic && %libomptarget-run-fail-generic
 // RUN: %libomptarget-compileoptxx-generic && %libomptarget-run-fail-generic
+// XFAIL: intelgpu
 
 int main(int argc, char *argv[]) {
 #pragma omp target

offload/test/offloading/atomic-compare-signedness.c

@@ -7,6 +7,7 @@
 // RUN: %libomptarget-run-generic | %fcheck-generic
 // RUN: %libomptarget-compileopt-generic -fopenmp-version=51
 // RUN: %libomptarget-run-generic | %fcheck-generic
+// XFAIL: intelgpu
 
 // High parallelism increases our chances of detecting a lack of atomicity.
 #define NUM_THREADS_TRY 256

offload/test/offloading/back2back_distribute.c

@@ -1,5 +1,7 @@
+// clang-format off
 // RUN: %libomptarget-compile-generic -O3 && %libomptarget-run-generic | %fcheck-generic
-
+// XFAIL: intelgpu
+// clang-format on
 #include <omp.h>
 #include <stdio.h>
 #include <stdlib.h>

offload/test/offloading/bug47654.cpp

@@ -1,5 +1,6 @@
 // RUN: %libomptarget-compilexx-run-and-check-generic
 // RUN: %libomptarget-compileoptxx-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <cassert>
 #include <iostream>

offload/test/offloading/bug49021.cpp

@@ -1,7 +1,10 @@
+// clang-format off
 // RUN: %libomptarget-compilexx-generic -O3 && %libomptarget-run-generic
 // RUN: %libomptarget-compilexx-generic -O3 -ffast-math && %libomptarget-run-generic
 // RUN: %libomptarget-compileoptxx-generic -O3 && %libomptarget-run-generic
 // RUN: %libomptarget-compileoptxx-generic -O3 -ffast-math && %libomptarget-run-generic
+// XFAIL: intelgpu
+// clang-format on
 
 #include <iostream>
 

offload/test/offloading/bug49334.cpp

@@ -8,6 +8,7 @@
 // REQUIRES: gpu
 // UNSUPPORTED: nvidiagpu
 // UNSUPPORTED: amdgpu
+// XFAIL: intelgpu
 
 #include <cassert>
 #include <cmath>

offload/test/offloading/bug49779.cpp

@@ -5,6 +5,7 @@
 
 // We need malloc/global_alloc support
 // UNSUPPORTED: amdgcn-amd-amdhsa
+// XFAIL: intelgpu
 
 #include <cassert>
 #include <iostream>

offload/test/offloading/bug50022.cpp

@@ -1,5 +1,6 @@
 // RUN: %libomptarget-compilexx-and-run-generic
 // RUN: %libomptarget-compileoptxx-and-run-generic
+// XFAIL: intelgpu
 
 #include <cassert>
 #include <iostream>

offload/test/offloading/bug51781.c

@@ -33,6 +33,7 @@
 // RUN: %fcheck-nvptx64-nvidia-cuda -check-prefix=CUSTOM -input-file=%t.custom
 // RUN: %fcheck-amdgcn-amd-amdhsa -check-prefix=CUSTOM -input-file=%t.custom
 // RUN: %libomptarget-run-generic 2>&1 | %fcheck-generic
+// XFAIL: intelgpu
 //
 // CUSTOM: Rewriting generic-mode kernel with a customized state machine.
 

offload/test/offloading/bug51982.c

@@ -1,6 +1,7 @@
 // RUN: %libomptarget-compile-generic -O2 && %libomptarget-run-generic
 // -O2 to run openmp-opt
 // RUN: %libomptarget-compileopt-generic -O2 && %libomptarget-run-generic
+// XFAIL: intelgpu
 
 int main(void) {
   long int aa = 0;

offload/test/offloading/bug53727.cpp

@@ -1,5 +1,6 @@
 // RUN: %libomptarget-compilexx-and-run-generic
 // RUN: %libomptarget-compileoptxx-and-run-generic
+// XFAIL: intelgpu
 
 #include <cassert>
 #include <iostream>

offload/test/offloading/bug64959.c

@@ -6,6 +6,7 @@
 // UNSUPPORTED: amdgcn-amd-amdhsa
 // UNSUPPORTED: nvptx64-nvidia-cuda
 // UNSUPPORTED: nvptx64-nvidia-cuda-LTO
+// XFAIL: intelgpu
 
 #include <omp.h>
 #include <stdio.h>

offload/test/offloading/bug64959_compile_only.c

@@ -1,5 +1,6 @@
 // RUN: %libomptarget-compile-generic
 // RUN: %libomptarget-compileopt-generic
+// XFAIL: intelgpu
 
 #include <stdio.h>
 #define N 10

offload/test/offloading/bug74582.c

@@ -1,5 +1,6 @@
 // RUN: %libomptarget-compile-generic && %libomptarget-run-generic
 // RUN: %libomptarget-compileopt-generic && %libomptarget-run-generic
+// XFAIL: intelgpu
 
 // Verify we do not read bits in the image that are not there (nobits section).
 

offload/test/offloading/complex_reduction.cpp

@@ -1,5 +1,8 @@
+// clang-format off
 // RUN: %libomptarget-compilexx-generic -O3 && %libomptarget-run-generic
 // RUN: %libomptarget-compilexx-generic -O3 -ffast-math && %libomptarget-run-generic
+// XFAIL: intelgpu
+// clang-format on
 
 #include <complex>
 #include <iostream>

offload/test/offloading/ctor_dtor.cpp

@@ -1,5 +1,6 @@
 // RUN: %libomptarget-compilexx-run-and-check-generic
 // RUN: %libomptarget-compileoptxx-run-and-check-generic
+// XFAIL: intelgpu
 
 #include <cstdio>
 struct S {

offload/test/offloading/ctor_dtor_api.cpp

@@ -1,5 +1,6 @@
 // RUN: %libomptarget-compilexx-run-and-check-generic
 // RUN: %libomptarget-compileoptxx-run-and-check-generic
+// UNSUPPORTED: intelgpu
 
 #include <cstdio>
 #include <omp.h>

offload/test/offloading/ctor_dtor_lazy.cpp

@@ -4,6 +4,7 @@
 // RUN: %not --crash %libomptarget-run-generic
 // RUN: %libomptarget-compilexx-generic -DCTOR_API
 // RUN: %not --crash %libomptarget-run-generic
+// XFAIL: intelgpu
 
 #include <cstdio>
 #include <omp.h>

offload/test/offloading/cuda_no_devices.c

@@ -5,6 +5,7 @@
 // RUN: %libomptarget-compile-generic
 // RUN: env CUDA_VISIBLE_DEVICES= \
 // RUN:   %libomptarget-run-generic 2>&1 | %fcheck-generic
+// XFAIL: intelgpu
 
 #include <stdio.h>
 

offload/test/offloading/d2d_memcpy.c

@@ -4,6 +4,7 @@
 // RUN: %libomptarget-compileopt-generic && \
 // RUN: env OMP_MAX_ACTIVE_LEVELS=2 %libomptarget-run-generic | \
 // RUN: %fcheck-generic -allow-empty
+// XFAIL: intelgpu
 
 #include <assert.h>
 #include <omp.h>

offload/test/offloading/default_thread_limit.c

@@ -7,6 +7,7 @@
 // RUN:   %libomptarget-run-generic 2>&1 | %fcheck-generic --check-prefix=DEFAULT
 
 // REQUIRES: amdgpu
+// XFAIL: intelgpu
 
 __attribute__((optnone)) int optnone() { return 1; }
 

offload/test/offloading/dynamic_module.c

@@ -1,9 +1,12 @@
+// clang-format off
 // RUN: %libomptarget-compile-generic -DSHARED -fPIC -shared -o %t.so && \
 // RUN: %libomptarget-compile-generic %t.so && %libomptarget-run-generic 2>&1 | %fcheck-generic
 // RUN: %libomptarget-compileopt-generic -DSHARED -fPIC -shared -o %t.so && \
 // RUN: %libomptarget-compileopt-generic %t.so && %libomptarget-run-generic 2>&1 | %fcheck-generic
 //
 // REQUIRES: gpu
+// XFAIL: intelgpu
+// clang-format on
 
 #ifdef SHARED
 void foo() {}

offload/test/offloading/dynamic_module_load.c

@@ -1,5 +1,7 @@
+// clang-format off
 // RUN: %libomptarget-compile-generic -DSHARED -fPIC -shared -o %t.so && %clang %flags %s -o %t -ldl && %libomptarget-run-generic %t.so 2>&1 | %fcheck-generic
-
+// XFAIL: intelgpu
+// clang-format on
 #ifdef SHARED
 #include <stdio.h>
 int foo() {

offload/test/offloading/extern.c

@@ -1,6 +1,8 @@
+// clang-format off
 // RUN: %libomptarget-compile-generic -DVAR -c -o %t.o
 // RUN: %libomptarget-compile-generic %t.o && %libomptarget-run-generic | %fcheck-generic
-
+// XFAIL: intelgpu
+// clang-format on
 #ifdef VAR
 int x = 1;
 #else

offload/test/offloading/force-usm.cpp

@@ -10,6 +10,7 @@
 //
 // UNSUPPORTED: nvptx64-nvidia-cuda
 // UNSUPPORTED: nvptx64-nvidia-cuda-LTO
+// XFAIL: intelgpu
 // clang-format on
 
 #include <cassert>

offload/test/offloading/fortran/basic-target-parallel-do.f90

@@ -3,6 +3,7 @@
 
 ! RUN: %libomptarget-compile-fortran-generic
 ! RUN: env LIBOMPTARGET_INFO=16 %libomptarget-run-generic 2>&1 | %fcheck-generic
+! XFAIL: intelgpu
 program main
    use omp_lib
    integer :: x(100)