Add caching to chunk_content's graph walking (vercel/turborepo#6242)

### Description

Extracts `chunk_content`'s processing of module references -> graph
nodes into a turbo function, allowing us to cache the result of
processing of every module.

### Testing Instructions

<!--
  Give a quick description of steps to test your changes.
-->


Closes WEB-1817

---------

Co-authored-by: Tobias Koppers <tobias.koppers@googlemail.com>

crates/turbopack-core/src/chunk/chunk_group.rs

@@ -30,7 +30,7 @@ pub async fn make_chunk_group(
         forward_edges_inherit_async,
         local_back_edges_inherit_async,
         available_async_modules_back_edges_inherit_async,
-    } = chunk_content(chunking_context, entries, Value::new(availability_info)).await?;
+    } = chunk_content(chunking_context, entries, availability_info).await?;
 
     // Find all local chunk items that are self async
     let self_async_children = chunk_items