This is the mail archive of the
libc-alpha@sources.redhat.com
mailing list for the glibc project.
Re: libc_hidden_proto breaks things for PowerPC64
- From: "Steve Munroe" <sjmunroe at us dot ibm dot com>
- To: libc-alpha at sources dot redhat dot com
- Cc: Peter Bergner <bergner at us dot ibm dot com>
- Date: Thu, 15 Aug 2002 17:28:53 -0500
- Subject: Re: libc_hidden_proto breaks things for PowerPC64
Andreas Jaeger wrote:
> Yes, please change libc-symbols to get a libc_hidden_def that works
> for you.
The problem is that a single libc_hidden_def (etc.) macro can not be
written to handle both cases.
How would the macro know (at cpp time) if the symbol is a function (which
requires the extra DOT_SYMBOLS) or is static variable (which does need the
extra DOT_SYMBOLS)? I can make libc-symbols.h (and contained macros) do one
(extra DOT_SYMBOLS) or the other, but not both.
The macro processor does not know the intended type (function or data) of
the symbol. But the generated assembler must have DOTed (the actual
function entry) and nonDOTed (the function descriptor or odp entry) symbols
defined.
The only alternative that I see is to have two macros each (i.e
libc_hidden_def and libc_hidden_data_def). For most platforms they would
have the same definition, but for PowerPC64 (AIX???) they would have
slightly different definitions ... (with and without the extra DOTed
symbols).