On Thu, Sep 29, 2016 at 11:21:48PM +0300, Michael S. Tsirkin wrote:
> On Thu, Sep 29, 2016 at 10:05:22PM +0200, Maxime Coquelin wrote:
> >
> >
> > On 09/29/2016 07:57 PM, Michael S. Tsirkin wrote:
> Yes but two points.
>
> 1. why is this memset expensive?
I don't have the exact answer, but just some rough thoughts:
It's an external clib function: there is a call stack and the
IP register will bounch back and forth. BTW, It's kind of an
overkill to use that for resetting 14 bytes structure.
Some trick like
*(struct virtio_net_hdr *)hdr = {0, };
Or even
hdr->xxx = 0;
hdr->yyy = 0;
should behaviour better.
There was an example: the vhost enqueue optmization patchset from
Zhihong [0] uses memset, and it introduces more than 15% drop (IIRC)
on my Ivybridge server: it has no such issue on his server though.
[0]: http://dpdk.org/ml/archives/dev/2016-August/045272.html
--yliu
> Is the test completely skipping looking
> at the packet otherwise?
>
> 2. As long as we are doing this, see
> Alignment vs. Networking
> ========================
> in Documentation/unaligned-memory-access.txt
>
>
> > From the micro-benchmarks results, we can expect +10% compared to
> > indirect descriptors, and + 5% compared to using 2 descs in the
> > virtqueue.
> > Also, it should have the same benefits as indirect descriptors for 0%
> > pkt loss (as we can fill 2x more packets in the virtqueue).
> >
> > What do you think?
> >
> > Thanks,
> > Maxime
