KernelAbstraction printing test fails

Maybe we need to have a rocprintf macro that accepts args... and automatically pick formats instead of calling it in a loop.

Got exception outside of a @test
InvalidIRError: compiling gpu_kernel_print(KernelAbstractions.CompilerMetadata{KernelAbstractions.NDIteration.DynamicSize, KernelAbstractions.NDIteration.DynamicCheck, Nothing, CartesianIndices{1, Tuple{Base.OneTo{Int64}}}, KernelAbstractions.NDIteration.NDRange{1, KernelAbstractions.NDIteration.DynamicSize, KernelAbstractions.NDIteration.StaticSize{(4,)}, CartesianIndices{1, Tuple{Base.OneTo{Int64}}}, Nothing}}) in world 34123 resulted in invalid LLVM IR
Reason: unsupported call to an unknown function (call to ijl_get_nth_field_checked)
Stacktrace:
 [1] getindex
   @ ./tuple.jl:29
 [2] iterate
   @ ./tuple.jl:68
 [3] #__print
   @ ~/.julia/dev/AMDGPU/src/ROCKernels.jl:182
 [4] macro expansion
   @ ~/.julia/packages/KernelAbstractions/XhtMv/src/KernelAbstractions.jl:251
 [5] macro expansion
   @ ~/.julia/packages/KernelAbstractions/XhtMv/test/print_test.jl:5
 [6] gpu_kernel_print
   @ ~/.julia/packages/KernelAbstractions/XhtMv/src/macros.jl:81
 [7] gpu_kernel_print
   @ ./none:0
Hint: catch this exception as `err` and call `code_typed(err; interactive = true)` to introspect the erronous code with Cthulhu.jl
Stacktrace:
  [1] check_ir(job::CompilerJob{GCNCompilerTarget, AMDGPU.Compiler.ROCCompilerParams}, args::LLVM.Module)
    @ GPUCompiler ~/.julia/packages/GPUCompiler/anMCs/src/validation.jl:154
  [2] macro expansion
    @ ~/.julia/packages/GPUCompiler/anMCs/src/driver.jl:442 [inlined]
  [3] macro expansion
    @ ~/.julia/packages/TimerOutputs/LHjFw/src/TimerOutput.jl:253 [inlined]
  [4] macro expansion
    @ ~/.julia/packages/GPUCompiler/anMCs/src/driver.jl:441 [inlined]
  [5] emit_llvm(job::CompilerJob, method_instance::Any; libraries::Bool, deferred_codegen::Bool, optimize::Bool, cleanup::Bool, only_entry::Bool, validate::Bool, ctx::ThreadSafeContext)
    @ GPUCompiler ~/.julia/packages/GPUCompiler/anMCs/src/utils.jl:83
  [6] emit_llvm
    @ ~/.julia/packages/GPUCompiler/anMCs/src/utils.jl:77 [inlined]
  [7] (::AMDGPU.Compiler.var"#61#64"{CompilerJob{GCNCompilerTarget, AMDGPU.Compiler.ROCCompilerParams}, Core.MethodInstance})(ctx::ThreadSafeContext)
    @ AMDGPU.Compiler ~/.julia/dev/AMDGPU/src/compiler/codegen.jl:128
  [8] ThreadSafeContext(f::AMDGPU.Compiler.var"#61#64"{CompilerJob{GCNCompilerTarget, AMDGPU.Compiler.ROCCompilerParams}, Core.MethodInstance})
    @ LLVM ~/.julia/packages/LLVM/HykgZ/src/executionengine/ts_module.jl:14
  [9] JuliaContext(f::AMDGPU.Compiler.var"#61#64"{CompilerJob{GCNCompilerTarget, AMDGPU.Compiler.ROCCompilerParams}, Core.MethodInstance})
    @ GPUCompiler ~/.julia/packages/GPUCompiler/anMCs/src/driver.jl:74
 [10] compile
    @ ~/.julia/dev/AMDGPU/src/compiler/codegen.jl:127 [inlined]
 [11] actual_compilation(cache::Dict{UInt64, Any}, key::UInt64, cfg::CompilerConfig{GCNCompilerTarget, AMDGPU.Compiler.ROCCompilerParams}, ft::Type, tt::Type, world::UInt64, compiler::typeof(AMDGPU.Compiler.compile), linker::typeof(AMDGPU.Compiler.link))
    @ GPUCompiler ~/.julia/packages/GPUCompiler/anMCs/src/cache.jl:184
 [12] cached_compilation(cache::Dict{UInt64, Any}, cfg::CompilerConfig{GCNCompilerTarget, AMDGPU.Compiler.ROCCompilerParams}, ft::Type, tt::Type, compiler::Function, linker::Function)
    @ GPUCompiler ~/.julia/packages/GPUCompiler/anMCs/src/cache.jl:163
 [13] macro expansion
    @ ~/.julia/dev/AMDGPU/src/compiler/codegen.jl:111 [inlined]
 [14] macro expansion
    @ ./lock.jl:267 [inlined]
 [15] rocfunction(f::typeof(Main.Testsuite.gpu_kernel_print), tt::Type; device::ROCDevice, global_hooks::NamedTuple{(), Tuple{}}, kwargs::Base.Pairs{Symbol, Union{}, Tuple{}, NamedTuple{(), Tuple{}}})
    @ AMDGPU.Compiler ~/.julia/dev/AMDGPU/src/compiler/codegen.jl:105
 [16] rocfunction
    @ ~/.julia/dev/AMDGPU/src/compiler/codegen.jl:101 [inlined]
 [17] macro expansion
    @ ~/.julia/dev/AMDGPU/src/highlevel.jl:599 [inlined]
 [18] (::KernelAbstractions.Kernel{ROCBackend, KernelAbstractions.NDIteration.StaticSize{(4,)}, KernelAbstractions.NDIteration.DynamicSize, typeof(Main.Testsuite.gpu_kernel_print)})(; ndrange::Tuple{Int64}, workgroupsize::Nothing)
    @ AMDGPU.ROCKernels ~/.julia/dev/AMDGPU/src/ROCKernels.jl:85
 [19] Kernel
    @ ~/.julia/dev/AMDGPU/src/ROCKernels.jl:82 [inlined]
 [20] macro expansion
    @ ~/.julia/packages/KernelAbstractions/XhtMv/test/print_test.jl:11 [inlined]
 [21] macro expansion
    @ ~/bin/julia-1.9.0-rc1/share/julia/stdlib/v1.9/Test/src/Test.jl:1498 [inlined]
 [22] printing_testsuite(backend::Type{ROCBackend})
    @ Main.Testsuite ~/.julia/packages/KernelAbstractions/XhtMv/test/print_test.jl:10
 [23] macro expansion
    @ ~/.julia/packages/KernelAbstractions/XhtMv/test/testsuite.jl:61 [inlined]
 [24] macro expansion
    @ ~/bin/julia-1.9.0-rc1/share/julia/stdlib/v1.9/Test/src/Test.jl:1498 [inlined]
 [25] testsuite(backend::Type, backend_str::String, backend_mod::Module, AT::Type, DAT::Type; skip_tests::Set{String})
    @ Main.Testsuite ~/.julia/packages/KernelAbstractions/XhtMv/test/testsuite.jl:9
 [26] testsuite
    @ ~/.julia/packages/KernelAbstractions/XhtMv/test/testsuite.jl:31 [inlined]
 [27] #5
    @ ~/.julia/dev/AMDGPU/test/runtests.jl:50 [inlined]

Start by reproducing the failure in test/print_test.jl and trace the printing path through src/ROCKernels.jl and src/macros.jl. Inspect the reported rocprintf and KernelAbstractions call sites to determine the supported argument shape. Done means the KernelAbstractions printing test compiles without invalid LLVM IR.