[ovs-dev] [PATCH v4] datapath-windows: Enable checksum offloads in STT

[Alin Serdean]

I added my comments inlined.

> -----Mesaj original-----
> De la: Nithin Raju [mailto:nithin at vmware.com]
> Trimis: Wednesday, September 23, 2015 9:05 AM
> Către: Jesse Gross <jesse at nicira.com>
> Cc: Alin Serdean <aserdean at cloudbasesolutions.com>; Ben Pfaff
> <blp at nicira.com>; dev at openvswitch.org
> Subiect: Re: [ovs-dev] [PATCH v4] datapath-windows: Enable checksum
> offloads in STT
> 
> hi Jesse,
> We are getting the Hyper-V solution to a state with the following goals:
> - Work “out of the box” ie. no need to make special settings such as disabling
> checksum offload, TSO, etc.
> - Reasonably stable
> 
> Most of the patches we have checked in so far into 2.4 are geared towards
> these two goals. Once all of the required changes go in, and we are
> reasonably confident about the stability, we can hopefully make an
> announcement about Hyper-V support.
[Alin Gabriel Serdean: ] +1
> 
> The 64-bit support is geared towards support Windows Nano which IIRC,
> supports only 64-bit apps (Alin, correct me if I’m wrong).
[Alin Gabriel Serdean: ] 64-bit is targeted mostly for NanoServer and to some
extent it increases performance.
It is important that we have a stable branch that supports NanoServers
especially for testing purposes because of the low disk requirements it offers.
(http://blogs.technet.com/b/windowsserver/archive/2015/04/08/microsoft-announces-nano-server-for-modern-apps-and-cloud.aspx)
> 
> STT and VXLAN checksumming patches are required to make sure that we
> don’t have to disable them at the VIF to make TCP/Ping traffic work.
[Alin Gabriel Serdean: ] I concur this is targeted directly for the users and it
will ease up a lot of configuration headaches when deploying.
> 
> The TCP flags patch is probably optional for 2.4, but maybe Alin has a good
> reason for it.
[Alin Gabriel Serdean: ] This is a full flat feature. I just thought it would be nice
to have it on both branches.
> 
> I agree that we should not destabilize 2.4 branch and we’ll take precautions
> for it.
[Alin Gabriel Serdean: ] +1
> 
> Pls. let us know if you have concerns.
> 
> thanks,
> -- Nithin
> 
> > On Sep 22, 2015, at 7:11 PM, Jesse Gross <jesse at nicira.com> wrote:
> >
> > On Tue, Sep 22, 2015 at 9:18 AM, Nithin Raju <nithin at vmware.com> wrote:
> >>
> >>> On Sep 22, 2015, at 9:07 AM, Ben Pfaff <blp at nicira.com> wrote:
> >>>
> >>> Thanks Sairam and Nithin, I applied this to master.
> >>
> >> Ben,
> >> Can you pls. apply this to 2.4 as well?
> >
> > I'm somewhat concerned about the number and size of Windows patches
> > that are targeted at 2.4 (it seems a lot went out today in
> > particular). Besides being large, many of them don't seem to meet the
> > criteria that I would normally expect for a backport. For example,
> > some look like features (64 bit support, TCP flags) while others
> > appear to be fixes for bugs so fundamental to the operation of things
> > that it seems unlikely that that part of the code can be deployed as
> > it currently exists (checksums in STT, VXLAN).
> >
> > Can you please explain the rationale?