This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [libstdc++] atomicity.h PATCH
- From: Jason Merrill <jason at redhat dot com>
- To: Alan Modra <amodra at bigpond dot net dot au>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Thu, 27 Feb 2003 16:34:03 -0500
- Subject: Re: [libstdc++] atomicity.h PATCH
- References: <wvl8yy3maod.fsf@prospero.boston.redhat.com><20030107085824.GP943@bubble.sa.bigpond.net.au>
On Tue, 7 Jan 2003 19:28:24 +1030, Alan Modra <amodra at bigpond dot net dot au> wrote:
> On Thu, Jan 02, 2003 at 05:47:14PM -0500, Jason Merrill wrote:
>> The constraints in the i486 and m68k patterns were broken; they failed to
>> record that *__mem is also used for output. This caused problems with the
>> tree-ssa branch.
>
> Hmm, shouldn't you remove the "memory" clobber now that the output
> is correctly specified?
Makes sense to me. But I notice that some other atomicity.h files which
already use separate input and output operands for *__mem still have a
memory clobber. MIPS, for instance. Are these also wrong?
Jason