Re: [PATCH 2/2] virtio: use virtio accessor to access packed event

[Jason Wang]

On Thu, Nov 11, 2021 at 3:51 PM Philippe Mathieu-Daudé
<philmd@redhat.com> wrote:
>
> On 11/11/21 07:38, Jason Wang wrote:
> > We used to access packed descriptor event and off_wrap via
> > address_space_{write|read}_cached(). When we hit the cache, memcpy()
> > is used which is not atomic which may lead a wrong value to be read or
> > wrote.
> >
> > This patch fixes this by switching to use
> > virito_{stw|lduw}_phys_cached() to make sure the access is atomic.
> >
> > Fixes: 683f7665679c1 ("virtio: event suppression support for packed ring")
> > Cc: qemu-stable@nongnu.org
> > Signed-off-by: Jason Wang <jasowang@redhat.com>
> > ---
> >  hw/virtio/virtio.c | 13 ++++---------
> >  1 file changed, 4 insertions(+), 9 deletions(-)
>
> Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
>
> No cover so asking here, what about vring_packed_desc_read()?

In that function, the vring_packed_desc_read_flags() used for reading
the flags atomically. If the flags told us the buffer is available,
there's no need read the rest of descriptor in atomic operation since
the driver guarantee that the changes of flags are visible after the
rest of the descriptor is setup.

Thanks

>