[PATCH] InlineFunction doesn't update InlineFunctionInfo with allocas created for byval arguments

[Julien Lerouge]

On Tue, Apr 08, 2014 at 10:30:44AM -0700, Manman Ren wrote:
> On Fri, Apr 4, 2014 at 5:15 PM, Julien Lerouge <julien.lerouge at m4x.org>wrote:
> 
> > When llvm::InlineFunction inlines a function that has byval arguments,
> > it creates allocas in the caller. Those allocas aren't inserted in the
> > InlineFunctionInfo data structure. So after inlining, if the client code
> > wants to know where are the allocas that were created, it will miss
> > those.
> >
> > Should InlineFunctionInfo contain these allocas, or is the omission
> > deliberate ?
> >
> 
> Hi Julien,
> 
> I don't think the omission is deliberate, but I may be wrong.
> Do you have a testing case?
> 
> Thanks,
> Manman
> 

Yeah, sorry, I'll add one and resend the patch.

Thanks,
Julien

-- 
Julien Lerouge
PGP Key Id: 0xB1964A62
PGP Fingerprint: 392D 4BAD DB8B CE7F 4E5F FA3C 62DB 4AA7 B196 4A62
PGP Public Key from: keyserver.pgp.com