[flang-commits] [flang] [Flang][MLIR] - Access the LEN for a `fir.boxchar<k>` and use it to set the bounds `omp.map.info` ops. (PR #134967)

[Slava Zakharin via flang-commits]

================
@@ -268,8 +274,54 @@ static void bindEntryBlockArgs(lower::AbstractConverter &converter,
                                             cloneBounds(v.getLBounds())));
           },
           [&](const fir::CharBoxValue &v) {
-            converter.bindSymbol(
-                sym, fir::CharBoxValue(arg, cloneBound(v.getLen())));
+            // In some cases, v.len could reference the input to the
+            // hlfir.declare which is the corresponding v.addr. While this isn't
+            // a big problem by itself, it is desirable to extract this out of
+            // v.addr itself since it's first result will be of type
+            // fir.boxchar<>. For example, consider the following
+            //
+            // func.func private @_QFPrealtest(%arg0: !fir.boxchar<1>)
+            //  %2 = fir.dummy_scope : !fir.dscope
+            //  %3:2 = fir.unboxchar %arg0 : (!fir.boxchar<1>) ->
+            //         (!fir.ref<!fir.char<1,?>>, index)
+            //  %4:2 = hlfir.declare (%3#0, %3#1, %2):(!fir.ref<!fir.char<1,?>>,
+            //                        index,!fir.dscope) ->
+            //                       (!fir.boxchar<1>, !fir.ref<!fir.char<1,?>>)
+
+            // In the case above,
+            // v.addr is
+            //  %4:2 = hlfir.declare (%3#0, %3#1, %2):(!fir.ref<!fir.char<1,?>>,
+            //                        index,!fir.dscope) ->
+            //                       (!fir.boxchar<1>, !fir.ref<!fir.char<1,?>>)
+            // v.len is
+            //  %3:2 = fir.unboxchar %arg0 : (!fir.boxchar<1>) ->
+            //         (!fir.ref<!fir.char<1,?>>, index)
+
+            // Mapping this to the target will create a use of %arg0 on the
+            // target. Since omp.target is IsolatedFromAbove, %arg0 will have to
----------------
vzakhari wrote:

Isn't the length of the character (i.e. the `index` value itself, not its defining op) also needed in the `hlfir.declare` of the character var?  Or is it different because in the array lower bound there is `!fir.ref<i32>` and in the length case it is `i32`?

> The problem in the case that this PR is solving is that this process of creating a temporary for an implicitly captured value is contingent upon such a value having a definingOp that is non-null. In the case of block arguments, the definingOp is null.

Can you please point me to where this is done?


Again, my general concern here is that we should probably just pass the chracter length value as a literal firstprivate, instead of cloning the defining `fir.unboxchar`.

https://github.com/llvm/llvm-project/pull/134967