This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug middle-end/14311] builtins for atomic operations needed
- From: "bkoz at redhat dot com" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 11 Apr 2005 17:14:25 -0000
- Subject: [Bug middle-end/14311] builtins for atomic operations needed
- References: <20040227003937.14311.bkoz@gcc.gnu.org>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
------- Additional Comments From bkoz at redhat dot com 2005-04-11 17:14 -------
Subject: Re: builtins for atomic operations needed
> I'm working on atomic builtins, but this will *not* resolve the problem of
> compiling for i386 and i486+. Indeed, it could easily make it worse because
> you won't have the kind of control you did before wrt hiding these operations
> in out-of-line functions.
Hurray. I really think that the builtin solution is the right way
forward for everybody, design-wise. Thanks for doing this Richard.
Pre gcc-3.4, we had the same problem with i386/i486+ separation.
I think, realistically, that i486 compilation should be a minimum
requirement for linux distros circa 2005. (If not i586.). Thus, the
solution I will be in favor of is eliminating i386 compatibility with
the version of the toolchain that first implements the atomic builtins.
For people that really want to support i386, they should use gcc-2.7 to gcc-3.4.
-benjamin
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14311