Re: [PATCH v6 1/3] memory: drop guest writes to read-only ram device regions

[Yan Zhao]

On Mon, May 25, 2020 at 01:04:56PM +0200, Paolo Bonzini wrote:
> On 25/05/20 12:54, Philippe Mathieu-Daudé wrote:
> >> Not all of them, only those that need to return MEMTX_ERROR.  I would
> >> like some guidance from Peter as to whether (or when) reads from ROMs
> >> should return MEMTX_ERROR.  This way, we can use that information to
> >> device  what the read-only ram-device regions should do.
> > Is it only device-specific or might it be partly arch/machine-specific
> > (depending on the bus it is mapped)?
> 
> Good point, I think that could be handled by propagating the error up in
> the memory region hierarchy (i.e. the cached AddressSpaceDispatch radix
> tree is used in the common case, but when you have a failure you
> percolate it up through the whole hierarchy since that's not a fast path).
> 
>
but if we decide to propagate the error up by providing with
ops->write_with_attrs, then we have to remove ops->write correspondingly. 
as in

memory_region_dispatch_write()
{
    ...
    if (mr->ops->write) {
        return access_with_adjusted_size(addr, &data, size,
                                         mr->ops->impl.min_access_size,
                                         mr->ops->impl.max_access_size,
                                         memory_region_write_accessor, mr,
                                         attrs);
    } else {
        return
            access_with_adjusted_size(addr, &data, size,
                                      mr->ops->impl.min_access_size,
                                      mr->ops->impl.max_access_size,
                                      memory_region_write_with_attrs_accessor,
                                      mr, attrs);
    }
    ...
}

so which regions should keep ops->write and which regions should not?

Thanks
Yan