Unnecessary default-constructibility requirements for RPC types?

Edgard Schmidt schmidt at ...456...
Mon Oct 24 19:08:46 CEST 2016

as described in "Genode Foundations", all types which are used as  
argument or return types in RPC functions must have a default  
constructor. I do not think that such a limitation is really  
necessary. So I tried to write a proof of concept [1]. These  
modifications should allow RPC functions to pass most of the  
non-default-constructible objects which satisfy the rest of the  
requirements for RPC argument and return types. I extended the  
"hello_tutorial" with a little test code to show that it works [2].

Software patterns like RAII [3] are a possible reason for removing  
that limitation. Quite often classes cannot provide a useful  
default-constructor which does more than merely setting the object  
state to "invalid". Because often there are no useful default values.  
So if a class provides a default-constructor which just sets the  
object to "invalid", the code complexity increases since care has to  
be taken that the object is fed with data up to a specific point of  

Unfortunately, my implementation does not remove default-constructor  
calls completely:
1. Capabilities still need default-constructors. Msgbuf stores them  
directly in an array and I did not find the time to check changes of  
this storage behavior for possible side effects.
2. Output-only _arguments_ have to be default-constructible too.  
Please note, that non-default-constructible _return_ types are  
allowed. So this affects only RPC arguments which are explicitly  
declared as Output-only, for example by adding custom specializations  
of Rpc_direction. The reason for this limitation is that my Pod_tuple  
version still stores the RPC arguments directly. While input arguments  
are copy-constructed, it is hard to copy from an object that does not  
yet exist.

For the second point, I see no simple solutions: we could create a  
simple byte buffer instead of a default-constructed output-only  
argument. Since "normal" RPC arguments have to be trivially copyable  
[4], a RPC function could easily write to this buffer, which would be  
accessible by the correctly type-casted reference or pointer. A  
minimalist version of boost::optional [5] or the like as member of  
Server_args could make sense. However, it would require more template  
specializations and case differentiations because the special RPC  
types Capability and Native_capability are not trivially copyable.

I successfully tested the code with the repositories "hello_tutorial"  
and "demo" on base-linux. But if you decide to use my code  
contribution, please note that it is rather a proof of concept. There  
is room for improvements:
* I tried to minimize the code modifications. So some refactoring may  
be useful: I expanded Meta::Empty by constructors and a typedef, I  
inserted possibly confusing forward declarations, and Meta::Pod_tuples  
and Meta::Pod_args could be renamed because they do not only accept  
POD types [6].
* I completely ignored the trace-related code. I do not know if there  
is a need of modifications.

Thanks to Johannes Schlatow who gave me the idea of grappling with this topic.

[1] https://github.com/3dik/genode/tree/nodef
[3] https://en.wikipedia.org/wiki/Resource_Acquisition_Is_Initialization
[4] http://en.cppreference.com/w/cpp/concept/TriviallyCopyable
[6] http://en.cppreference.com/w/cpp/concept/PODType

More information about the users mailing list