Enable InvokeSimd/call_vadd_1d_loop for PVC, disable InvokeSimd popcnt tests (#8970)

Driver issue fixed some time ago. Recently it started to pass
'unexpectedly' in some CIs.

Author: fveselov

sycl/test-e2e/InvokeSimd/Feature/ImplicitSubgroup/popcnt.cpp

@@ -2,8 +2,9 @@
 // REQUIRES: gpu && linux
 // UNSUPPORTED: cuda || hip
 //
-// TODO: enable after simd_mask supported
-// XFAIL: gpu
+// NOTE: The feature is not yet supported, there is a discussion on the
+// feasibility of tests
+// REQUIRES: TEMPORARY_DISABLED
 //
 // Check that full compilation works:
 // RUN: %clangxx -DIMPL_SUBGROUP -fsycl -fno-sycl-device-code-split-esimd -Xclang -fsycl-allow-func-ptr %S/../popcnt.cpp -o %t.out

sycl/test-e2e/InvokeSimd/Feature/popcnt.cpp

@@ -2,8 +2,9 @@
 // REQUIRES: gpu && linux
 // UNSUPPORTED: cuda || hip
 //
-// TODO: enable after simd_mask supported
-// XFAIL: gpu
+// NOTE: The feature is not yet supported, there is a discussion on the
+// feasibility of tests
+// REQUIRES: TEMPORARY_DISABLED
 //
 // Check that full compilation works:
 // RUN: %clangxx -fsycl -fno-sycl-device-code-split-esimd -Xclang -fsycl-allow-func-ptr %s -o %t.out

sycl/test-e2e/InvokeSimd/Regression/call_vadd_1d_loop.cpp

@@ -2,9 +2,6 @@
 // REQUIRES: gpu && linux
 // UNSUPPORTED: cuda || hip
 //
-// TODO: enable when Jira ticket resolved
-// XFAIL: gpu-intel-pvc
-//
 // Check that full compilation works:
 // RUN: %clangxx -fsycl -fno-sycl-device-code-split-esimd -Xclang -fsycl-allow-func-ptr %s -o %t.out
 // RUN: env IGC_VCSaveStackCallLinkage=1 IGC_VCDirectCallsOnly=1 %GPU_RUN_PLACEHOLDER %t.out

sycl/test-e2e/InvokeSimd/Regression/nbarrier_exec_in_order.cpp

@@ -57,9 +57,9 @@ ESIMD_INLINE void ESIMD_CALLEE_nbarrier(int localID,
   esimd::simd<int, VL> val(localID);
 
   if constexpr (UseSLM) {
-    /* TODO: SLM has to be allocated of outside invoke_simd, but propper
-     * intarface is not yet ready. Current test implementation in this regard
-     * is a subject to future changes.
+    /* TODO: SLM needs to be allocated from outside invoke_simd, but the proper
+     * intarface is not yet ready. Current test implementation in this regard is
+     * subject to future changes.
      */
     esimd::slm_init(Size * sizeof(int));
     esimd::simd<int, VL> zero(0);

sycl/test-e2e/InvokeSimd/Regression/slm_load_store.cpp

@@ -45,9 +45,9 @@ constexpr int slm_size = 32 * 1024;
 
 ESIMD_INLINE void slm_load_store_test(nd_item<1> *ndi, dtype *A,
                                       dtype *C) SYCL_ESIMD_FUNCTION {
-  /* TODO: SLM has to be allocated of outside invoke_simd, but propper
-   * intarface is not yet ready. Current test implementation in this regard
-   * is a subject to future changes.
+  /* TODO: SLM needs to be allocated from outside invoke_simd, but the proper
+   * intarface is not yet ready. Current test implementation in this regard is
+   * subject to future changes.
    */
   esimd::slm_init<slm_size>();