Merge pull request #12095 from wenduwan/cuda-doc-update-dlopen

docs: update static application build docs

Author: jsquyres

docs/building-apps/building-static-apps.rst

@@ -83,15 +83,14 @@ sub-optimal:
 
    In other words: *using dynamic linking saves memory.*
 
-#. If you disable Open MPI's ``dlopen`` functionality (which is
-   necessary to create *fully*-static MPI applications), you lose the
-   following:
-
-   * CUDA, because |mdash| among other reasons |mdash| the CUDA
-     library is dynamically loaded at run-time via ``dlopen(3)``.
-
-   * Memory manager functionality, which is important for OS-bypass
-     networks such as InfiniBand.
+#. *Fully*-static applications are not linked to the dynamic linking library,
+   e.g. ``libdl`` on Linux, which provides ``dlopen(3)``, ``dlsym(3)``, etc.
+   This will break Open MPI functionalities that depend on such interfaces.
+
+   .. warning:: Open MPI's memory management functionality, which provides
+                important performance optimizations on OS-bypass networks
+                such as InfiniBand, requires the ``dlsym(3)`` interface,
+                and therefore does not work with fully-static applications.
 
 Are you convinced yet?  *Please try to avoid building fully-static MPI
 applications if at all possible.*

docs/installing-open-mpi/configure-cli-options/misc.rst

@@ -13,8 +13,10 @@ above categories that can be used with ``configure``:
   is not *necessary* for OpenFabrics networks, but some performance
   loss may be observed without it).
 
-  However, it may be necessary to disable the memory manager in order
-  to build Open MPI statically.
+  .. warning:: Open MPI's memory management functionality, which provides
+               important performance optimizations on OS-bypass networks
+               such as InfiniBand, requires the ``dlsym(3)`` interface,
+               and therefore does not work with fully-static applications.
 
 * ``--with-ft=TYPE``:
   Specify the type of fault tolerance to enable.  The only allowed