[SYCL] Add XFAIL trackers for E2E tests and remove hip_nvidia XFAILs entirely (#16307)

This is a cleanup of several tests marked as XFAIL on certain platforms.
It is a follow up of #16287 and it
follows the same principles in handling of XFAILs.

Author: lbushi25

sycl/test-e2e/Basic/aspects.cpp

@@ -1,8 +1,5 @@
 // RUN: %{build} -o %t.out -DSYCL_DISABLE_IMAGE_ASPECT_WARNING
 // RUN: %{run-unfiltered-devices} %t.out
-//
-// Hip is missing some of the parameters tested here so it fails with NVIDIA
-// XFAIL: hip_nvidia
 
 //==--------------- aspects.cpp - SYCL device test ------------------------==//
 //

sycl/test-e2e/Basic/device_event.cpp

@@ -1,9 +1,5 @@
 // RUN: %{build} -o %t.run
 // RUN: %{run} %t.run
-//
-// Returns error "Barrier is not supported on the host device
-// yet." with Nvidia.
-// XFAIL: hip_nvidia
 
 //==--------device_event.cpp - SYCL class device_event test ----------------==//
 //

sycl/test-e2e/Basic/diagnostics/handler.cpp

@@ -1,9 +1,6 @@
 // RUN: %{build} -o %t.out
 // RUN: %{run} %t.out | FileCheck %s
-//
-// Appears to fail on HIP Nvidia because 'no device of requested type available'
-// when constructing a queue with an exception_list.
-// XFAIL: hip_nvidia
+
 //==------------------- handler.cpp ----------------------------------------==//
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 // See https://llvm.org/LICENSE.txt for license information.

sycl/test-e2e/Basic/partition_supported.cpp

@@ -1,8 +1,5 @@
 // RUN: %{build} -o %t.out
 // RUN: %{run} %t.out
-//
-// Nvidia should not allow sub_devices but does not throw corresponding error.
-// XFAIL: hip_nvidia
 /* Check that:
 1) if partition_equally is supported, then we check that the correct
 invalid errc is returned if more than max_compute_units are requested

sycl/test-e2e/KernelAndProgram/kernel-bundle-merge-options.cpp

@@ -6,6 +6,7 @@
 // Debug option -g is not passed to device code compiler when CL-style driver
 // is used and /DEBUG options is passed.
 // XFAIL: cl_options
+// XFAIL-TRACKER: https://github.com/intel/llvm/issues/16309
 
 #include "kernel-bundle-merge-options.hpp"
 

sycl/test-e2e/Printf/int.cpp

@@ -7,7 +7,8 @@
 // UNSUPPORTED: hip_amd
 // FIXME: The 'short' type gets overflown with sporadic values on CUDA.
 // XFAIL: cuda
-//
+// XFAIL-TRACKER: https://github.com/intel/llvm/issues/14734
+
 // RUN: %{build} -o %t.out
 // RUN: %{run} %t.out | FileCheck %s
 // FIXME: Remove dedicated constant address space testing once generic AS

sycl/test-e2e/Printf/mixed-address-space.cpp

@@ -3,7 +3,7 @@
 //
 // UNSUPPORTED: hip_amd
 // XFAIL: cuda && windows
-//
+// XFAIL-TRACKER: https://github.com/intel/llvm/issues/14733
 // FIXME: Drop the test once generic AS support is considered stable and the
 //        dedicated constant AS overload of printf is removed from the library.
 //

sycl/test-e2e/Printf/percent-symbol.cpp

@@ -6,7 +6,7 @@
 //
 // UNSUPPORTED: hip_amd
 // XFAIL: cuda && windows
-//
+// XFAIL-TRACKER: https://github.com/intel/llvm/issues/14733
 // RUN: %{build} -o %t.out
 // RUN: %{run} %t.out | FileCheck %s
 // FIXME: Remove dedicated constant address space testing once generic AS

sycl/test-e2e/Regression/context_is_destroyed_after_exception.cpp

@@ -2,8 +2,6 @@
 
 // RUN: %{build} -o %t.out
 // RUN: env SYCL_UR_TRACE=2 %{run} %t.out %if !windows %{2>&1 | FileCheck %s %}
-//
-// XFAIL: hip_nvidia
 
 #include <sycl/detail/core.hpp>
 

sycl/test-e2e/Regression/multiple-targets.cpp

@@ -15,9 +15,6 @@
 //
 // RUN: %clangxx -fsycl -fsycl-targets=spir64,%{sycl_triple} -fsycl-device-code-split=per_kernel -o %t4.out %s
 // RUN: %{run} %t4.out
-//
-// XFAIL: hip_nvidia
-//
 
 #include <sycl/detail/core.hpp>