[CIR][Lowering] Lower delete array to LLVM

[ahmedshakill]

Initial implementation of lowering cir.delete.array #1285

lowered to a call to _ZdaPv keeping similarity with how Clang AST is lowered.

before

module @"/opt/tmp/input.cpp" attributes {cir.lang = #cir.lang<cxx>, cir.sob = #cir.signed_overflow_behavior<undefined>, cir.triple = "x86_64-unknown-linux-gnu", cir.type_size_info = #cir.type_size_info<char = 8, int = 32, size_t = 64>, dlti.dl_spec = #dlti.dl_spec<!llvm.ptr<271> = dense<32> : vector<4xi64>, !llvm.ptr<272> = dense<64> : vector<4xi64>, i128 = dense<128> : vector<2xi64>, !llvm.ptr<270> = dense<32> : vector<4xi64>, f80 = dense<128> : vector<2xi64>, i1 = dense<8> : vector<2xi64>, !llvm.ptr = dense<64> : vector<4xi64>, i8 = dense<8> : vector<2xi64>, i16 = dense<16> : vector<2xi64>, i32 = dense<32> : vector<2xi64>, f64 = dense<64> : vector<2xi64>, f128 = dense<128> : vector<2xi64>, f16 = dense<16> : vector<2xi64>, i64 = dense<64> : vector<2xi64>, "dlti.stack_alignment" = 128 : i64, "dlti.endianness" = "little", "dlti.mangling_mode" = "e">} {
  cir.func @_Z17test_delete_arrayPi(%arg0: !cir.ptr<!s32i> loc(fused[#loc3, #loc4])) extra(#fn_attr) {
    %0 = cir.alloca !cir.ptr<!s32i>, !cir.ptr<!cir.ptr<!s32i>>, ["ptr", init] {alignment = 8 : i64} loc(#loc8)
    cir.store %arg0, %0 : !cir.ptr<!s32i>, !cir.ptr<!cir.ptr<!s32i>> loc(#loc5)
    %1 = cir.load %0 : !cir.ptr<!cir.ptr<!s32i>>, !cir.ptr<!s32i> loc(#loc6)
    cir.delete.array %1 : <!s32i> loc(#loc6)
    cir.return loc(#loc2)
  } loc(#loc7)
} loc(#loc)

after

define dso_local void @_Z17test_delete_arrayPi(ptr %0) #0 {
  %2 = alloca ptr, i64 1, align 8
  store ptr %0, ptr %2, align 8
  %3 = load ptr, ptr %2, align 8
  call void @_ZdaPv(ptr %3)
  ret void
}

[github-actions]

✅ With the latest revision this PR passed the C/C++ code formatter.

[bcardosolopes]

Thanks for the first contribution and for adding support for this. Few nits and this should be ready to go.

[bcardosolopes]

LGTM once it passes tests!

[ahmedshakill]

@bcardosolopes Sorry about earlier, it seems I didn't ran git-clang-format properly. Hopefully this should be okay now.

[bcardosolopes]

Dispatched another test attempt!