[dpdk-dev] [PATCH] virtio: fix virtio_net_hdr desc pointing to the same buffer

[Xie, Huawei]

> -----Original Message-----
> From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> Sent: Monday, December 14, 2015 9:45 PM
> To: Xie, Huawei
> Cc: Yuanhan Liu; dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH] virtio: fix virtio_net_hdr desc
> pointing to the same buffer
> 
> 2015-12-14 13:38, Xie, Huawei:
> > From: Yuanhan Liu [mailto:yuanhan.liu at linux.intel.com]
> > > On Mon, Dec 14, 2015 at 01:44:54PM +0100, Thomas Monjalon wrote:
> > > > 2015-12-14 19:47, Yuanhan Liu:
> > > > > On Mon, Dec 14, 2015 at 10:32:24AM +0100, Thomas Monjalon
> wrote:
> > > > > > 2015-12-14 11:01, Yuanhan Liu:
> > > > > > > On Fri, Dec 11, 2015 at 12:07:32AM +0800, Huawei at dpdk.org
> wrote:
> > > > > > > > The virtio_net_hdr desc all pointed to the same buffer.
> It
> > > doesn't cause
> > > > > > > > issue because in the simple TX mode we don't use the
> header. This
> > > patch
> > > > > > > > makes the header desc point to different buffer.
> > > > > > > >
> > > > > > > > Signed-off-by: Huawei Xie <huawei.xie at intel.com>
> > > > > > >
> > > > > > > Acked-by: Yuanhan Liu <yuanhan.liu at linux.intel.com>
> > > > > >
> > > > > > Does it fix something in the current behaviour?
> > > > >
> > > > > It's more like a logic fixing to me.
> > > > >
> > > > > > I have the feeling it may wait for 2.3.
> > > > >
> > > > > It's been introduced in v2.2, with Huawei's simple tx patchset.
> > > > > Therefore, I guess 2.2 is good to go?
> > > >
> > > > The vhost driver has been validated without with patch.
> > >
> > > Huawei stated in the commit log that "It doesn't cause issue
> because in
> > > the simple TX mode we don't use the header".
> > >
> > > > Merging it would be taking the risk of breaking something
> > > > (or just reduce performance) for no clear benefit.
> > > > Am I missing something?
> > >
> > Thomas, there is no risk at all with this patch, and it will not
> affect performance.
> > I prefer to integrate this patch, so that we have a good looking
> vhost library. :).
> 
> I'm not sure that "good looking" is enough.
> I'll wait for another opinion before merging, so we'll have 2
> responsibles
> in case of failure :)
Np. There is no issue either apply this patch or delay it to 2.3.
> 
> > > I know your concerns: we really should be cagy about making any
> changes
> > > when a release is close, especially when all stuff are validated.
> From
> > > this point of view, I agree with you we could delay it to v2.3.
> > >
> > > Maybe huawei have more inputs here?
> > >
> > > 	--yliu
>