[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20180124193217-mutt-send-email-mst@kernel.org>
Date: Wed, 24 Jan 2018 19:32:59 +0200
From: "Michael S. Tsirkin" <mst@...hat.com>
To: Wei Wang <wei.w.wang@...el.com>
Cc: virtio-dev@...ts.oasis-open.org, linux-kernel@...r.kernel.org,
virtualization@...ts.linux-foundation.org, kvm@...r.kernel.org,
linux-mm@...ck.org, mhocko@...nel.org, akpm@...ux-foundation.org,
pbonzini@...hat.com, liliang.opensource@...il.com,
yang.zhang.wz@...il.com, quan.xu0@...il.com, nilal@...hat.com,
riel@...hat.com
Subject: Re: [PATCH v23 2/2] virtio-balloon: VIRTIO_BALLOON_F_FREE_PAGE_VQ
On Wed, Jan 24, 2018 at 01:47:47PM +0800, Wei Wang wrote:
> On 01/24/2018 01:01 PM, Michael S. Tsirkin wrote:
> > On Wed, Jan 24, 2018 at 10:50:27AM +0800, Wei Wang wrote:
> > This will not DTRT in all cases. It's quite possible
> > that host does not need the kick when ring is half full but
> > does need it later when ring is full.
> > You can kick at ring half full as optimization but you absolutely
> > still must kick on ring full. Something like:
> >
> > if (vq->num_free == virtqueue_get_vring_size(vq) / 2 ||
> > vq->num_free <= 2)
>
> Right. Would "if (vq->num_free < virtqueue_get_vring_size(vq) / 2" be
> better?
>
>
> Best,
> Wei
It gives more kicks ... this reminds me, you need to validate
that vring size is at least 2, otherwise fail probe.
--
MST
Powered by blists - more mailing lists