RE: __libc_stack_end and friends on IA64

["Boehm, Hans" <hans_boehm at hp dot com>]

Jakub -

Thanks for the information.

Presumably it should be set correctly on IA64, even if it is libc internal?

I would like to see it exported again.  Hiding it probably won't break
anything on Linux systems.  But it will force some applications (notably
anything generated by gcj) to go out and read and parse /proc/stat to
rediscover what the startup code already knew.  (Exporting it and
initializing it to a bad nonzero value will break things.)

Hans

> -----Original Message-----
> From: Jakub Jelinek [mailto:jakub@redhat.com]
> Sent: Friday, March 15, 2002 2:41 PM
> To: Boehm, Hans
> Cc: libc-alpha@sources.redhat.com
> Subject: Re: __libc_stack_end and friends on IA64
> 
> 
> On Fri, Mar 15, 2002 at 02:32:20PM -0800, Boehm, Hans wrote:
> > I see __libc_stack_end exported by ld-2.2.4.so on both my 
> 7.2 RH laptop and
> > a local Itanium machine.  It's referenced by libc.so.  On 
> the Itanium
> > machine __libc_ia64_register_backing_store_base is exported 
> from crt1.o.  Am
> > I misunderstanding the issue?
> 
> Sorry for __libc_ia64_register_backing_store_base, it is a 
> non-issue then.
> But __libc_stack_end is @GLIBC_PRIVATE in current glibc CVS 
> trunk, which
> means it is considered glibc private symbol which apps or 
> libraries outside
> of glibc shouldn't reference.
> 
> 	Jakub
>