Re: [PATCH] openvswitch: Orphan frags before sending to userspace via Netlink to avoid guest stall

[Zoltan Kiss]

On 07/03/14 04:46, Pravin Shelar wrote:
> On Thu, Mar 6, 2014 at 9:09 AM, Zoltan Kiss <zoltan.kiss@xxxxxxxxxx> wrote:
> > Do you have any feedback on this? I'm also adding KVM list as they might be
> > interested in this.
> >
> > Zoli
> >
> >
> > On 28/02/14 19:16, Zoltan Kiss wrote:
> > > The kernel datapath now switched to zerocopy Netlink messages, but that
> > > also
> > > means that the pages on frags array are sent straight to userspace. If
> > > those
> > > pages came outside the kernel, we have to swap them out with local copies.
> > >
> > > Signed-off-by: Zoltan Kiss <zoltan.kiss@xxxxxxxxxx>
>
> I do not think this is required, netlink zero copy only maps
> pre-allocated buffers to user-space.
How do you mean "pre-allocated"? By who?

As far as I've seen the skb in this function came straight from the 
device (vif in our case), and skb_zerocopy just copy the frags to 
user_skb, which is sent to the userspace. Those frags contain pages from 
guest, and it's a bad idea to pass them to userspace: e.g if userspace 
dies in the meantime, what happens with them? Also, in Xen's case they 
are actually not mapped to userspace, so accessing them can lead to garbage.

Zoli
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/