Re: [Qemu-devel] [PATCH v2 00/38] Delay destruction of memory regions to instance_finalize

[Peter Maydell]

On 18 September 2013 20:26, Paolo Bonzini <address@hidden> wrote:
> But does guest code actually care?  In many cases, I suspect that
> sticking a smp_rmb() in the read side of "unlocked" register accesses,
> and a smp_wmb() in the write side, will do just fine.  And add a
> compatibility property to place a device back under the BQL for guests
> that have problems.

Yuck. This sounds like a recipe for spending the next five years
debugging subtle race conditions. We need to continue to support
the semantics that the architecture and hardware specs define
for memory access orderings to emulated devices.

-- PMM