[Offload] XFAIL four tests while working on fix

[jplehr]

omp_dynamic_shared_memory_mixed_amdgpu.c
omp_dynamic_shared_memory_amdgpu.c
amdgcn-amd-amdhsa::bug51982.c
amdgcn-amd-amdhsa::bug51781.c

[llvmbot]

@llvm/pr-subscribers-offload

@llvm/pr-subscribers-backend-amdgpu

Author: Jan Patrick Lehr (jplehr)

Changes

omp_dynamic_shared_memory_mixed_amdgpu.c
omp_dynamic_shared_memory_amdgpu.c
amdgcn-amd-amdhsa::bug51982.c
amdgcn-amd-amdhsa::bug51781.c

---

Full diff: https://github.com/llvm/llvm-project/pull/98899.diff

4 Files Affected:

• (modified) offload/test/api/omp_dynamic_shared_memory_amdgpu.c (+1)
• (modified) offload/test/api/omp_dynamic_shared_memory_mixed_amdgpu.c (+1)
• (modified) offload/test/offloading/bug51781.c (+1)
• (modified) offload/test/offloading/bug51982.c (+1)

diff --git a/offload/test/api/omp_dynamic_shared_memory_amdgpu.c b/offload/test/api/omp_dynamic_shared_memory_amdgpu.c
index 0b4d9d6ea9d46..dd4d59257714b 100644
--- a/offload/test/api/omp_dynamic_shared_memory_amdgpu.c
+++ b/offload/test/api/omp_dynamic_shared_memory_amdgpu.c
@@ -2,6 +2,7 @@
 // RUN: env LIBOMPTARGET_SHARED_MEMORY_SIZE=256 \
 // RUN:   %libomptarget-run-amdgcn-amd-amdhsa | %fcheck-amdgcn-amd-amdhsa
 // REQUIRES: amdgcn-amd-amdhsa
+// XFAIL: amdgcn-amd-amdhsa
 
 #include <omp.h>
 #include <stdio.h>
diff --git a/offload/test/api/omp_dynamic_shared_memory_mixed_amdgpu.c b/offload/test/api/omp_dynamic_shared_memory_mixed_amdgpu.c
index 656c3a20aaf82..844c25dc9e025 100644
--- a/offload/test/api/omp_dynamic_shared_memory_mixed_amdgpu.c
+++ b/offload/test/api/omp_dynamic_shared_memory_mixed_amdgpu.c
@@ -2,6 +2,7 @@
 // RUN: env LIBOMPTARGET_NEXTGEN_PLUGINS=1 \
 // RUN:   %libomptarget-run-amdgcn-amd-amdhsa | %fcheck-amdgcn-amd-amdhsa
 // REQUIRES: amdgcn-amd-amdhsa
+// XFAIL: amdgcn-amd-amdhsa
 
 #include "omp_dynamic_shared_memory_mixed.inc"
 // CHECK: PASS
diff --git a/offload/test/offloading/bug51781.c b/offload/test/offloading/bug51781.c
index 35ecf55aa8c53..237e1585455c5 100644
--- a/offload/test/offloading/bug51781.c
+++ b/offload/test/offloading/bug51781.c
@@ -31,6 +31,7 @@
 // RUN: %libomptarget-run-generic 2>&1 | %fcheck-generic
 //
 // CUSTOM: Rewriting generic-mode kernel with a customized state machine.
+// XFAIL: amdgcn-amd-amdhsa
 
 #if ADD_REDUCTION
 #define REDUCTION(...) reduction(__VA_ARGS__)
diff --git a/offload/test/offloading/bug51982.c b/offload/test/offloading/bug51982.c
index 91ce4a264e238..c037d215a8a79 100644
--- a/offload/test/offloading/bug51982.c
+++ b/offload/test/offloading/bug51982.c
@@ -1,6 +1,7 @@
 // RUN: %libomptarget-compile-generic -O1 && %libomptarget-run-generic
 // -O1 to run openmp-opt
 // RUN: %libomptarget-compileopt-generic -O1 && %libomptarget-run-generic
+// XFAIL: amdgcn-amd-amdhsa
 
 int main(void) {
   long int aa = 0;

[jhuber6]

What are the fails? Can we make an issue to track this before we land anything?

[github-actions]

⚠️ C/C++ code formatter, clang-format found issues in your code. ⚠️

You can test this locally with the following command:

git-clang-format --diff 861a8ed68be7c6d5e2605b1ab3810fde72b5f66a 6cb954bd051e87eef91301875574a71b53ca1941 --extensions c -- offload/test/api/omp_dynamic_shared_memory_amdgpu.c offload/test/api/omp_dynamic_shared_memory_mixed_amdgpu.c offload/test/offloading/bug51781.c offload/test/offloading/bug51982.c

View the diff from clang-format here.

diff --git a/offload/test/api/omp_dynamic_shared_memory_amdgpu.c b/offload/test/api/omp_dynamic_shared_memory_amdgpu.c
index dd4d592577..3c140aa7ee 100644
--- a/offload/test/api/omp_dynamic_shared_memory_amdgpu.c
+++ b/offload/test/api/omp_dynamic_shared_memory_amdgpu.c
@@ -1,4 +1,5 @@
-// RUN: %libomptarget-compile-amdgcn-amd-amdhsa -O1 -mllvm -openmp-opt-inline-device
+// RUN: %libomptarget-compile-amdgcn-amd-amdhsa -O1 -mllvm
+// -openmp-opt-inline-device
 // RUN: env LIBOMPTARGET_SHARED_MEMORY_SIZE=256 \
 // RUN:   %libomptarget-run-amdgcn-amd-amdhsa | %fcheck-amdgcn-amd-amdhsa
 // REQUIRES: amdgcn-amd-amdhsa
diff --git a/offload/test/api/omp_dynamic_shared_memory_mixed_amdgpu.c b/offload/test/api/omp_dynamic_shared_memory_mixed_amdgpu.c
index 844c25dc9e..08412dcacf 100644
--- a/offload/test/api/omp_dynamic_shared_memory_mixed_amdgpu.c
+++ b/offload/test/api/omp_dynamic_shared_memory_mixed_amdgpu.c
@@ -1,4 +1,5 @@
-// RUN: %libomptarget-compile-amdgcn-amd-amdhsa -O1 -mllvm -openmp-opt-inline-device -I %S
+// RUN: %libomptarget-compile-amdgcn-amd-amdhsa -O1 -mllvm
+// -openmp-opt-inline-device -I %S
 // RUN: env LIBOMPTARGET_NEXTGEN_PLUGINS=1 \
 // RUN:   %libomptarget-run-amdgcn-amd-amdhsa | %fcheck-amdgcn-amd-amdhsa
 // REQUIRES: amdgcn-amd-amdhsa

[jplehr]

Yeah, my bad, I forgot to link to the PR that caused the fails in the commit message.
#83131

There is an assertion error that we hit. Matt asked to produce IR before/after the patch, so I'm going to llvm-reduce these tests and provide it to him.

[jhuber6]

LG for now as long as Matt fixes it and there's an issue.

[jplehr]

I have created #98903 to track this.