This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug ada/19388] [4.0 Regression] MAX_OFILE_ALIGNMENT on PA HP-UX
- From: "dave at hiauly1 dot hia dot nrc dot ca" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 12 Jan 2005 02:53:55 -0000
- Subject: [Bug ada/19388] [4.0 Regression] MAX_OFILE_ALIGNMENT on PA HP-UX
- References: <20050112022518.19388.danglin@gcc.gnu.org>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
------- Additional Comments From dave at hiauly1 dot hia dot nrc dot ca 2005-01-12 02:53 -------
Subject: Re: [4.0 Regression] MAX_OFILE_ALIGNMENT on PA HP-UX
> ------- Additional Comments From pinskia at gcc dot gnu dot org 2005-01-12
> 02:32 -------
> Hmm, I think pa abose of the macro really is the cause, maybe we should
> change the macro to
> dependent on the decl and then force the explicat use in the pa back-end so
> we don't run into this
> again?
I certainly agree that I abused the macro. However, many people were
using the alignment attribute to request alignments that the linker
wasn't capable of providing without -fno-common. This is critical for
when the decl is used for a semaphore lock. They require 16-byte
alignment on the PA. Also, I believe -fno-common only applies to C.
Thus, I think the above suggestion is a good one but the ada routine
doesn't have access to the decl. The only other solution that I can
see is to give up using common for uninitialized data.
Dave
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19388