From: Rusty Russell <rusty@rustcorp.com.au>
To: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Christoph Hellwig <hch@infradead.org>,
linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
virtualization <virtualization@lists.linux-foundation.org>
Subject: Re: [PATCH 5 of 5] virtio: expose added descriptors immediately
Date: Tue, 22 Nov 2011 11:03:04 +1030 [thread overview]
Message-ID: <87lir9xagv.fsf@rustcorp.com.au> (raw)
In-Reply-To: <20111121115702.GA27580@redhat.com>
On Mon, 21 Nov 2011 13:57:04 +0200, "Michael S. Tsirkin" <mst@redhat.com> wrote:
> On Mon, Nov 21, 2011 at 12:18:45PM +1030, Rusty Russell wrote:
> > On Wed, 16 Nov 2011 09:18:38 +0200, "Michael S. Tsirkin" <mst@redhat.com> wrote:
> > > My unlocked kick patches will trip this warning: they make
> > > virtio-net do add + get without kick.
> >
> > Heh, it's a good sign if they do, since that means you're running really
> > well :)
>
> They don't in fact, in my testing :(. But I think they can with luck.
>
> > > I think block with unlocked kick can trip it too:
> > > add, lock is dropped and then an interrupt can get.
> > >
> > > We also don't need a kick each num - each 2^15 is enough.
> > > Why don't we do this at start of add_buf:
> > > if (vq->num_added >= 0x7fff)
> > > return -ENOSPC;
> >
> > The warning was there in case a driver is never doing a kick, and
> > getting away with it (mostly) because the device is polling. Let's not
> > penalize good drivers to catch bad ones.
> >
> > How about we do this properly, like so:
>
> Absolutely. But I think we also need to handle num_added
> overflow of a 15 bit counter, no? Otherwise the
> vring_need_event logic might give us false negatives ....
> I'm guessing we can just assume we need a kick in that case.
You're right. Thankyou. My immediate reaction of "make it an unsigned
long" doesn't work.
Here's the diff to what I posted before:
diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
--- a/drivers/virtio/virtio_ring.c
+++ b/drivers/virtio/virtio_ring.c
@@ -254,9 +254,10 @@ add_head:
vq->vring.avail->idx++;
vq->num_added++;
- /* If you haven't kicked in this long, you're probably doing something
- * wrong. */
- WARN_ON(vq->num_added > vq->vring.num);
+ /* This is very unlikely, but theoretically possible. Kick
+ * just in case. */
+ if (unlikely(vq->num_added == 65535))
+ virtqueue_kick(_vq);
pr_debug("Added buffer head %i to %p\n", head, vq);
END_USE(vq);
next prev parent reply other threads:[~2011-11-22 1:00 UTC|newest]
Thread overview: 34+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <patchbomb.1320306168@localhost6.localdomain6>
2011-11-03 7:42 ` [PATCH 1 of 5] virtio: document functions better Rusty Russell
2011-11-03 7:49 ` Christoph Hellwig
2011-11-03 7:42 ` [PATCH 2 of 5] virtio: rename virtqueue_add_buf_gfp to virtqueue_add_buf Rusty Russell
2011-11-03 7:50 ` Christoph Hellwig
2011-11-03 7:42 ` [PATCH 3 of 5] virtio: support unlocked queue kick Rusty Russell
2011-11-03 7:52 ` Christoph Hellwig
2011-11-04 10:09 ` Stefan Hajnoczi
2011-11-04 10:36 ` Rusty Russell
2011-11-03 7:42 ` [PATCH 4 of 5] virtio: avoid modulus operation Rusty Russell
2011-11-03 7:51 ` Pekka Enberg
2011-11-03 10:18 ` Rusty Russell
2011-11-03 7:42 ` [PATCH 5 of 5] virtio: expose added descriptors immediately Rusty Russell
2011-11-13 21:03 ` Michael S. Tsirkin
2011-11-14 0:43 ` Rusty Russell
2011-11-14 6:56 ` Michael S. Tsirkin
2011-11-16 0:21 ` Rusty Russell
2011-11-16 7:18 ` Michael S. Tsirkin
2011-11-21 1:48 ` Rusty Russell
2011-11-21 11:57 ` Michael S. Tsirkin
2011-11-22 0:33 ` Rusty Russell [this message]
2011-11-22 6:29 ` Michael S. Tsirkin
2011-11-23 1:19 ` Rusty Russell
2011-11-23 8:30 ` Michael S. Tsirkin
2012-07-01 9:20 ` RFD: virtio balloon API use (was Re: [PATCH 5 of 5] virtio: expose added descriptors immediately) Michael S. Tsirkin
2012-07-02 1:05 ` Rusty Russell
2012-07-02 7:25 ` Michael S. Tsirkin
2012-07-02 16:08 ` Rafael Aquini
2012-07-03 0:47 ` Rusty Russell
2012-07-03 16:26 ` Rafael Aquini
2012-07-04 10:55 ` Michael S. Tsirkin
2012-07-08 23:39 ` Rusty Russell
2012-07-04 10:55 ` Michael S. Tsirkin
2012-07-02 7:33 ` [PATCH RFC] virtio-balloon: fix add/get API use Michael S. Tsirkin
2012-07-04 3:27 ` Rusty Russell
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=87lir9xagv.fsf@rustcorp.com.au \
--to=rusty@rustcorp.com.au \
--cc=hch@infradead.org \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mst@redhat.com \
--cc=virtualization@lists.linux-foundation.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).