Re: Quo vadis, thunk? (Was: Re: [PATCH/RFA] PR target/16665: SH thunk

[Mark Mitchell <mark at codesourcery dot com>]

Joern Rennecke wrote:

> > Thus, I would consider the fact that G++ presently makes the thunks weak 
> > even when the underlying function is not weak to be an odd quirk, if not 
> > a bug.  I think that could safely be changed without endangering 
> > anything.  It may be that it is the way it is for historical reasons; it 
> > could also be that until my recent linkage cleanup, one could not be 
> > sure whether the associated function would be COMDAT or not.  In any 
> > case, I think you could change it now.
> >    
>
> I see that use_thunk has now another call to comdat_linkage, so changing
> things beforehand would be no good.
> I changed the code in use_thunk to be more thorough in making the thunk
> agree with the function.  (If you think all this adjusting of the thunk
> can migrate to make_thunk, that is really a separate issue.)
>  
In fact, I think that all of this ought to migrate to 
import_export_decl; that's the function that's responsible for 
determining whether things are COMDAT or not.  But, as you say, this is 
a separate issue.

> What kind of testing would this patch require?
>  
First, show that it works: have tests with virtual functions (a) in 
COMDAT, and (b) in a named section, and show that in both cases you get 
the thunk in the same section.  Second, test on a major platform 
(GNU/Linux, OS X, Solaris, ...) and show that you get no regressions.

I think it should be OK if those tests pass.

--
Mark Mitchell
CodeSourcery, LLC
(916) 791-8304
mark@codesourcery.com