Message ID | 20200730194858.28523-4-paul@xen.org (mailing list archive) |
---|---|
State | Superseded |
Headers | show |
Series | tools: propagate bridge MTU to vif frontends | expand |
On 30.07.20 21:48, Paul Durrant wrote: > From: Paul Durrant <pdurrant@amazon.com> > > There is currently no documentation to state what MTU a frontend should > adertise to its network stack. It has however long been assumed that the > default value of 1500 is correct. > > This patch specifies a mechanism to allow the tools to set the MTU via a > xenstore node in the frontend area and states that the absence of that node > means the frontend should assume an MTU of 1500 octets. > > NOTE: The Windows PV frontend has used an MTU sampled from the xenstore > node specified in this patch. > > Signed-off-by: Paul Durrant <pdurrant@amazon.com> Can you please update docs/misc/xenstore-paths.pandoc accordingly? With that you can have my: Reviewed-by: Juergen Gross <jgross@suse.com> Juergen
> -----Original Message----- > From: Jürgen Groß <jgross@suse.com> > Sent: 03 August 2020 06:09 > To: Paul Durrant <paul@xen.org>; xen-devel@lists.xenproject.org > Cc: Durrant, Paul <pdurrant@amazon.co.uk> > Subject: RE: [EXTERNAL] [PATCH 3/4] public/io/netif: specify MTU override node > > On 30.07.20 21:48, Paul Durrant wrote: > > From: Paul Durrant <pdurrant@amazon.com> > > > > There is currently no documentation to state what MTU a frontend should > > adertise to its network stack. It has however long been assumed that the > > default value of 1500 is correct. > > > > This patch specifies a mechanism to allow the tools to set the MTU via a > > xenstore node in the frontend area and states that the absence of that node > > means the frontend should assume an MTU of 1500 octets. > > > > NOTE: The Windows PV frontend has used an MTU sampled from the xenstore > > node specified in this patch. > > > > Signed-off-by: Paul Durrant <pdurrant@amazon.com> > > Can you please update docs/misc/xenstore-paths.pandoc accordingly? Sure. Given the path is for use by tools then it should indeed be documented there as well. > With that you can have my: > > Reviewed-by: Juergen Gross <jgross@suse.com> > Thanks, Paul > > Juergen
diff --git a/xen/include/public/io/netif.h b/xen/include/public/io/netif.h index 9fcf91a2fe..00dd258712 100644 --- a/xen/include/public/io/netif.h +++ b/xen/include/public/io/netif.h @@ -204,6 +204,18 @@ * present). */ +/* + * MTU + * === + * + * The toolstack may set a value of MTU for the frontend by setting the + * /local/domain/<domid>/device/vif/<vif>/mtu node with the MTU value in + * octets. If this node is absent the frontend should assume an MTU value + * of 1500 octets. A frontend is also at liberty to ignore this value so + * it is only suitable for informing the frontend that a packet payload + * >1500 octets is permitted. + */ + /* * Hash types * ==========