[llvm] [Dwarf][Transforms] Add dwarf support when func signature changed (PR #127855)

[Jose E. Marchesi via llvm-commits]

jemarch wrote:

> > > > Wouldn't it be possible to set the existing DW_AT_artificial attribute instead of introducing vendor-specific extensions?
> > > 
> > > 
> > > not sure if that'd connote either of the properties being expressed here (uncallable, and un-printable-returns) - for instance in C++, implicit ctors, assignment operators, and dtors are artificial but callable/printable returns based on ABI, so we couldn't have debuggers assuming artificial functions can't be called/have their return values printed.
> > 
> > 
> > The idea, which we were briefly discussing today in the #gcc IRC channel, would be to have the signature change of some given function represented in DWARF using a DW_TAG_artificial "container" with an inlined subroutine reflecting the original signature. Something like:
> > ```
> > foo.clone (int a, int b) { foo (1, { a, b }); } 
> > ```
> > 
> > 
> >     
> >       
> >     
> > 
> >       
> >     
> > 
> >     
> >   
> > The signature change would be described in the formal parameters of the DW_TAG_inlined_subroutine.
> 
> @jemarch Could you share a detailed dwarf info about how signature change is encoded?

I will try to synthesize one showing the typical constant propagation and scalar replacement of aggregates, which are the IPA optimizations that may result in changing the signature of the function. 

https://github.com/llvm/llvm-project/pull/127855