On Tue, Sep 06, 2011 at 12:28:40PM +0300, Avi Kivity wrote: > On 09/06/2011 09:55 AM, Paolo Bonzini wrote: > >On 09/06/2011 05:12 AM, David Gibson wrote: > >>I'm not "fixing ppc". I'm fixing a fundamental flaw in the protocol > >>implementation._So far_ I've only observed the effects on ppc, but > >>that doesn't mean they don't exist. > > > >Actually Michael is right. The implementation is correct on x86, > >though wrong anywhere else (perhaps s390?). On those > >architectures you do not need rmb() and wmb(). > > Are we sure? Nothing prevents the guest from using weakly-ordered > writes, is there? For example, using MOVNTDQ.
Yes but the macros in question are here to order qemu accesses, not guest accesses. > > Although in that case the guest is probably required to issue an SFENCE. > -- > error compiling committee.c: too many arguments to function
