[dpdk-dev] [PATCH v2] net/virtio-user: fix cannot get initialized

Tan, Jianfeng jianfeng.tan at intel.com
Fri Apr 14 07:55:17 CEST 2017



> -----Original Message-----
> From: Yuanhan Liu [mailto:yuanhan.liu at linux.intel.com]
> Sent: Friday, April 14, 2017 1:38 PM
> To: Tan, Jianfeng
> Cc: dev at dpdk.org; maxime.coquelin at redhat.com
> Subject: Re: [PATCH v2] net/virtio-user: fix cannot get initialized
> 
> On Fri, Apr 14, 2017 at 04:35:57AM +0000, Tan, Jianfeng wrote:
> >
> >
> > > -----Original Message-----
> > > From: Yuanhan Liu [mailto:yuanhan.liu at linux.intel.com]
> > > Sent: Friday, April 14, 2017 12:24 PM
> > > To: Tan, Jianfeng
> > > Cc: dev at dpdk.org; maxime.coquelin at redhat.com
> > > Subject: Re: [PATCH v2] net/virtio-user: fix cannot get initialized
> > >
> > > On Thu, Apr 13, 2017 at 10:11:27AM +0000, Jianfeng Tan wrote:
> > > > The feature negotiation in virtio-user is proven to be broken,
> > > > which results in device initialization failure.
> > > >
> > > > Originally, we get features from vhost backend, and remove those
> > > > that are not supported. But when new feature is added, for example,
> > > > VIRTIO_NET_F_MTU, we fail to remove this new feature. Then, this
> > > > new feature will be negotiated, as both frontend and backend claim
> > > > to support this feature.
> > > >
> > > > To fix it, we add a macro to record supported featues, as a filter
> > > > to remove newly added features.
> > >
> > > Yes, this is much better! You now don't have to worry that virtio-user
> > > will be broken every time we add a new feature.
> > >
> > > Applied to dpdk-next-virtio, with the title changed to "fix feature
> > > negotitation".
> > >
> >
> > Thanks. By the way, it's better backporting it to stable branches, however, I
> forget to CC stable at dpdk.org.
> 
> As long as we don't backport new features to a stable release, there
> should be no problem.
> 
> Besides, the virtio-user features supported in this patch may not match
> those supported by a stable release (say 16.11 LTS). For example, F_STATUS
> is just added in this feature. Thus, I don't think it's a good candidate
> for a stable release (though a backport could fix it).

Make sense. Thanks!

Jianfeng


More information about the dev mailing list