Re: [PATCH] make vfio and DAX cache work together

["Dr. David Alan Gilbert" <dgilbert@redhat.com>]

* Edge NFV (edgenfv@gmail.com) wrote:
>  Signed-off-by: Edge NFV <edgenfv@gmail.com>

Hi,
  I take it that 'Edge NFV' isn't your real name; apologies if it is.
It's unusual not to use a real name; I would be interested to know
why you feel uncomfortable not doing.

> ---
>  hw/vfio/common.c | 9 +++++++++
>  1 file changed, 9 insertions(+)
> 
> diff --git a/hw/vfio/common.c b/hw/vfio/common.c
> index ae5654fcdb..83e15bf7a3 100644
> --- a/hw/vfio/common.c
> +++ b/hw/vfio/common.c
> @@ -668,6 +668,15 @@ static void vfio_listener_region_add(MemoryListener
> *listener,
>                  int128_get64(int128_sub(section->size, int128_one())));
>          return;
>      }
> +
> +    /* Do not add virtio fs cache section */
> +    if (!strcmp(memory_region_name(section->mr), "virtio-fs-cache")) {

So first, this is a patch that fixes something that isn't yet in qemu;
the DAX mode of virtiofs.
Secondly, hard coding the name like this is probably the wrong thing to
do; we need a way for the cache to declare it wants to be omitted.
Thirdly, shouldn't this actually be a change to
vfio_listener_skip_section to add this test?

Dave

> +        trace_vfio_listener_region_add_skip(
> +                section->offset_within_address_space,
> +                section->offset_within_address_space +
> +                int128_get64(int128_sub(section->size, int128_one())));
> +        return;
> +    }
> 
>      if (unlikely((section->offset_within_address_space &
>                    ~qemu_real_host_page_mask) !=
> -- 
> 2.25.1
-- 
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK