Re: [PATCH v4 2/2] virtio_net: Extend virtio to use VF datapath when available

["Samudrala, Sridhar" <sridhar.samudrala@intel.com>]

[-- Attachment #1: Type: text/plain, Size: 3880 bytes --]



On 3/2/2018 8:20 AM, Jiri Pirko wrote:
> Fri, Mar 02, 2018 at 04:26:25PM CET, alexander.duyck@gmail.com wrote:
>> On Fri, Mar 2, 2018 at 12:36 AM, Jiri Pirko <jiri@resnulli.us> wrote:
>>> Thu, Mar 01, 2018 at 09:08:43PM CET, sridhar.samudrala@intel.com wrote:
>>>> This patch enables virtio_net to switch over to a VF datapath when a VF
>>>> netdev is present with the same MAC address. It allows live migration
>>>> of a VM with a direct attached VF without the need to setup a bond/team
>>>> between a VF and virtio net device in the guest.
>>>>
>>>> The hypervisor needs to enable only one datapath at any time so that
>>>> packets don't get looped back to the VM over the other datapath. When a VF
>>>> is plugged, the virtio datapath link state can be marked as down. The
>>>> hypervisor needs to unplug the VF device from the guest on the source host
>>>> and reset the MAC filter of the VF to initiate failover of datapath to
>>>> virtio before starting the migration. After the migration is completed,
>>>> the destination hypervisor sets the MAC filter on the VF and plugs it back
>>>> to the guest to switch over to VF datapath.
>>>>
>>>> When BACKUP feature is enabled, an additional netdev(bypass netdev) is
>>>> created that acts as a master device and tracks the state of the 2 lower
>>>> netdevs. The original virtio_net netdev is marked as 'backup' netdev and a
>>>> passthru device with the same MAC is registered as 'active' netdev.
>>>>
>>>> This patch is based on the discussion initiated by Jesse on this thread.
>>>> https://marc.info/?l=linux-virtualization&m=151189725224231&w=2
>>>>
>>>> Signed-off-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
>>>> Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
>>>> Reviewed-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
>>>> ---
>>>> drivers/net/virtio_net.c | 683 ++++++++++++++++++++++++++++++++++++++++++++++-
>>> As I wrote to the discussion of the other version of this patchset,
>>> I strongly believe you need to have a common part in net/core/ that will
>>> be shared by both virtio_net and netvsc. There there should be explicit
>>> limits for this in-driver bonding solution, like max 1 vf slave, etc.
>> Yeah, this code essentially calls out the "shareable" code with a
>> comment at the start and end of the section what defines the
>> virtio_bypass functionality. It would just be a matter of mostly
>> cutting and pasting to put it into a separate driver module.
> Please put it there and unite the use of it with netvsc.

Based on Stephen's responses, it is not clear if netvsc would be OK to move to
the 3 netdev solution at this time. When another paravirtual driver would like
to use this solution, we can do refactoring at that time.

>
>> The design limits things to a 1:1 relationship since we just have the
>> child and backup pointers, but I don't think I am seeing exception
>> handling to prevent us from overwriting the child pointers so there
>> may be a leak there.
>
We only allow one active netdev registered as a child. There is a check to
make sure that if an active netdev is registered, another one with same MAC
is not allowed. I don't think there is a leak.

	vbi = netdev_priv(dev);
         backup = (child_netdev->dev.parent == dev->dev.parent);
         if (backup ? rtnl_dereference(vbi->backup_netdev) :
                         rtnl_dereference(vbi->active_netdev)) {
                 netdev_info(dev,
                             "%s attempting to join bypass dev when %s already present\n",
                             child_netdev->name, backup ? "backup" : "active");
                 return NOTIFY_DONE;
         }

         /* Avoid non pci devices as active netdev */
         if (!backup && (!child_netdev->dev.parent ||
                         !dev_is_pci(child_netdev->dev.parent)))
                 return NOTIFY_DONE;

Thanks
Sridhar


[-- Attachment #2: Type: text/html, Size: 40257 bytes --]