This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: -flto / -flto-partition=none question on ia64-hp-hpux11.23
- From: Steve Ellcey <sje at cup dot hp dot com>
- To: Ian Lance Taylor <iant at google dot com>
- Cc: Richard Guenther <richard dot guenther at gmail dot com>, gcc at gcc dot gnu dot org
- Date: Tue, 01 Mar 2011 15:05:15 -0800
- Subject: Re: -flto / -flto-partition=none question on ia64-hp-hpux11.23
- References: <201102252322.p1PNM2W08338@lucas.cup.hp.com> <AANLkTikhtvStR_Lhs4+2CbGh6UdPYiktgWAWaX7x-EV5@mail.gmail.com> <1298917338.2267.235.camel@hpsje.cup.hp.com> <AANLkTik3n1oAQdSg6ytvLpC4uJ0n_dvzrLxgu8p-wnzL@mail.gmail.com> <1298935631.2267.282.camel@hpsje.cup.hp.com> <AANLkTik29_qN-BrSeB3Yqs18B_8qr8GumWJRTPU-jQ+A@mail.gmail.com> <1299000877.30497.7.camel@hpsje.cup.hp.com> <mcr4o7mbjm6.fsf@google.com>
- Reply-to: sje at cup dot hp dot com
On Tue, 2011-03-01 at 13:45 -0800, Ian Lance Taylor wrote:
> Steve Ellcey <sje@cup.hp.com> writes:
>
> > According to PR 20095 the reason for not implementing
> > __gcc_personality_v0 is that the HP unwind library does not implement
> > _UA_END_OF_STACK which is a GCC extension to the standard IA64 unwind
> > ABI and the HP-UX system unwind does not support it so the only way
> > to implement __gcc_personality_v0 is to use the GCC libunwind instead of
> > the HP-UX system one.
>
> I don't understand this point. It's true that gcc's unwind library will
> pass _UA_END_OF_STACK to the stop function passed to
> _Unwind_ForcedUnwind if the unwind library returns _URC_END_OF_STACK.
> But I don't see what that has to do with __gcc_personality_v0.
> __gcc_personality_v0 does not generate or look for _UA_END_OF_STACK, nor
> should it, as it is not a stop function passed to _Unwind_ForcedUnwind.
>
> Ian
I guess it isn't related directly to __gcc_personality_v0. If I include
unwind-c.c in the ia64-hp-hpux11.23 build then I get the C personality
function and the pr33572.C test passes with -flto. The reason we didn't
include this file before (in response to PR 20095) was that even if we
added unwind-c.c to the build, it didn't fix gcc.dg/cleanup-5.c. That
test still fails with a segfault in the HP-UX unwind library (In
_Unwind_ForcedUnwind).
So maybe we should include unwind-c.c to fix the -flto failures.
Steve Ellcey
sje@cup.hp.com