Re: [PATCH] riscv: add gdbserver support

["Maciej W. Rozycki" <macro@wdc.com>]

On Mon, 20 Jan 2020, Simon Marchi wrote:

> >  Offhand I can see the proposal fails to implement XML register 
> > descriptions, which I think every modern port is expected to do (we also 
> > need to disallow non-XML-enabled RISC-V stubs in GDB proper, as we 
> > discussed before; I fail to understand why it wasn't done right away with 
> > the initial implementation, as it's quite straightforward and would have 
> > set the policy for debug stubs right from the beginning).
> 
> I would also expect new ports to use XML target descriptions.  And I see
> that there is already code in arch/riscv.c to build target descriptions based
> on detected features... so should gdbserver use it?

 Yes, that's what I used with my implementation; a minor change was 
required for `riscv_create_target_description' not to return a `const' 
result (or `init_target_desc' couldn't be called on it).

 Unfortunately the result is not accepted by GDB proper as it does not 
know architecture names produced by this code (and native support bypasses 
this step as it uses a different code path for XML description setup).  I 
haven't got to fixing that yet and therefore I have concluded my 
`gdbserver' implementation is not ready for upstreaming at this point; 
test results, although clearly not catastrophic, are also affected by this 
problem I believe.

  Maciej