On Wed, Sep 06, 2017 at 10:02:29PM +0200, Maxime Coquelin wrote:
> On 09/06/2017 09:30 AM, Tiwei Bie wrote:
> > On Wed, Sep 06, 2017 at 09:15:47AM +0200, Maxime Coquelin wrote:
> > > Hi Tiwei,
> > >
> > > On 09/06/2017 03:15 AM, Tiwei Bie wrote:
> > > > On Tue, Sep 05, 2017 at 01:00:42PM +0200, Maxime Coquelin wrote:
> > > > > On 09/05/2017 12:07 PM, Tiwei Bie wrote:
> > > > > > On Tue, Sep 05, 2017 at 11:24:14AM +0200, Maxime Coquelin wrote:
> > > > > > > On 09/05/2017 06:45 AM, Tiwei Bie wrote:
> > > > > > > > On Thu, Aug 31, 2017 at 11:50:05AM +0200, Maxime Coquelin wrote:
> > > > > > > > > virtio_net device might be accessed while being reallocated
> > > > > > > > > in case of NUMA awareness. This case might be theoretical,
> > > > > > > > > but it will be needed anyway to protect vrings pages against
> > > > > > > > > invalidation.
> > > > > > > > >
> > > > > > > > > The virtio_net devs are now protected with a readers/writers
> > > > > > > > > lock, so that before reallocating the device, it is ensured
> > > > > > > > > that it is not being referenced by the processing threads.
> > > > > > > > >
> > > > > > > > [...]
> > > > > > > > > +struct virtio_net *
> > > > > > > > > +get_device(int vid)
> > > > > > > > > +{
> > > > > > > > > + struct virtio_net *dev;
> > > > > > > > > +
> > > > > > > > > + rte_rwlock_read_lock(&vhost_devices[vid].lock);
> > > > > > > > > +
> > > > > > > > > + dev = __get_device(vid);
> > > > > > > > > + if (unlikely(!dev))
> > > > > > > > > +
> > > > > > > > > rte_rwlock_read_unlock(&vhost_devices[vid].lock);
> > > > > > > > > +
> > > > > > > > > + return dev;
> > > > > > > > > +}
> > > > > > > > > +
> > > > > > > > > +void
> > > > > > > > > +put_device(int vid)
> > > > > > > > > +{
> > > > > > > > > + rte_rwlock_read_unlock(&vhost_devices[vid].lock);
> > > > > > > > > +}
> > > > > > > > > +
> > > > > > > >
> > > > > > > > This patch introduced a per-device rwlock which needs to be
> > > > > > > > acquired
> > > > > > > > unconditionally in the data path. So for each vhost device, the
> > > > > > > > IO
> > > > > > > > threads of different queues will need to acquire/release this
> > > > > > > > lock
> > > > > > > > during each enqueue and dequeue operation, which will cause
> > > > > > > > cache
> > > > > > > > contention when multiple queues are enabled and handled by
> > > > > > > > different
> > > > > > > > cores. With this patch alone, I saw ~7% performance drop when
> > > > > > > > enabling
> > > > > > > > 6 queues to do 64bytes iofwd loopback test. Is there any way to
> > > > > > > > avoid
> > > > > > > > introducing this lock to the data path?
> > > > > > >
> > > > > > > First, I'd like to thank you for running the MQ test.
> > > > > > > I agree it may have a performance impact in this case.
> > > > > > >
> > > > > > > This lock has currently two purposes:
> > > > > > > 1. Prevent referencing freed virtio_dev struct in case of
> > > > > > > numa_realloc.
> > > > > > > 2. Protect vring pages against invalidation.
> > > > > > >
> > > > > > > For 2., it can be fixed by using the per-vq IOTLB lock (it was
> > > > > > > not the
> > > > > > > case in my early prototypes that had per device IOTLB cache).
> > > > > > >
> > > > > > > For 1., this is an existing problem, so we might consider it is
> > > > > > > acceptable to keep current state. Maybe it could be improved by
> > > > > > > only
> > > > > > > reallocating in case VQ0 is not on the right NUMA node, the other
> > > > > > > VQs
> > > > > > > not being initialized at this point.
> > > > > > >
> > > > > > > If we do this we might be able to get rid of this lock, I need
> > > > > > > some more
> > > > > > > time though to ensure I'm not missing something.
> > > > > > >
> > > > > > > What do you think?
> > > > > > >
> > > > > >
> > > > > > Cool. So it's possible that the lock in the data path will be
> > > > > > acquired only when the IOMMU feature is enabled. It will be
> > > > > > great!
> > > > > >
> > > > > > Besides, I just did a very simple MQ test to verify my thoughts.
> > > > > > Lei (CC'ed in this mail) may do a thorough performance test for
> > > > > > this patch set to evaluate the performance impacts.
> > > > >
> > > > > I'll try to post v2 this week including the proposed change.
> > > > > Maybe it'll be better Lei waits for the v2.
> > > > >
> > > >
> > > > Cool. Sure. Thank you! :)
> > >
> > > I have done the changes, you can find the v2 on my gitlab repo:
> > > https://gitlab.com/mcoquelin/dpdk-next-virtio/commits/vhost_iotlb_v2
> > >
> > > I'm testing it right now, but if you'd like to run some early benchmark
> > > before I post the series, there it is!
> > >
> >
> > Got it. Thanks! :)
>
> Just to let you know that I have updated my branch to remove another
> regression with iommu=off by inlining the noiommu part of
> vhost_iova_to_vva call (See below for the patch, that is squashed into
> my branch).
>
> Without this, when running microbenchmarks (txonly, rxonly, ...) I
> noticed a 4% perf degradation.
>
Nice work!
Best regards,
Tiwei Bie
