Re: [PATCH, CHKP] Restore transparent alias chains

[Jeff Law <law at redhat dot com>]

On 04/02/2015 12:48 PM, Jan Hubicka wrote:
> > On 03/20/2015 02:20 AM, Ilya Enkovich wrote:
> > > Hi,
> > >
> > > Identifiers read with input_identifier miss IDENTIFIER_TRANSPARENT_ALIAS bit.  We always expect it for instrumentation clones, thus restore it input_cgraph_1.  Bootstrapped and tested on x86_64-unknown-linux-gnu.  OK for trunk?
> > >
> > > Thanks,
> > > Ilya
> > > --
> > > 2015-03-20  Ilya Enkovich  <ilya.enkovich@intel.com>
> > >
> > > 	* lto-cgraph.c (input_cgraph_1): Always link instrumented
> > > 	assembler name with original one.
> > This appears to be a code path that only triggers when MPX is
> > enabled and is roughly analogous to the code in
> > chkp_build_instrumented_fndecl links things up.
> >
> > OK for the trunk.
>
> I think this will lead to wrong code. At this time, we may have multple
> declarations sharing single assembler name (and thus
> IDENTIFIER_TRANSPARENT_ALIAS link). Think of case where one unit defines static
> function and other global function of the same name. We may end up renaming the
> symbol but keeping bogus transparent alias link that will trigger on other
> symbol.
Then I think the code in ipa-chkp chkp_build_instrumented_fndecl (and 
more generally how we're using transparent aliases) may need some 
rethinking.

jeff