This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: More on memory barriers
- From: Hans-Peter Nilsson <hp at bitrange dot com>
- To: Jason Merrill <jason at redhat dot com>
- Cc: gcc at gcc dot gnu dot org, Benjamin Kosnik <bkoz at redhat dot com>
- Date: Thu, 16 Sep 2004 21:52:45 -0400 (EDT)
- Subject: Re: More on memory barriers
- References: <xypllfcbl1s.fsf@miranda.boston.redhat.com>
On Tue, 14 Sep 2004, Jason Merrill wrote:
> It also means that my 2003-12-19 change to remove general memory
> clobbers from the atomic operations was misguided; I failed to consider the
> synchronization function of the memory clobber.
>
> Thoughts?
Your new insight is comforting. ;-)
<URL:http://gcc.gnu.org/ml/gcc-patches/2004-03/msg01799.html>.
(The ranting paragraph starting with "I also added back the
memory clobber removed last December".)
Wrt. your question later in the thread, no additional sync insns
is needed for CRIS.
I probably miss something, but I don't really see an OO arch as
effectively much different than GCC taking a quantum leap in
scheduler technology and moving a memory access across the
location where a memory barrier is needed (that is, moving it
across a needed but *absent* barrier).
brgds, H-P