[dpdk-dev] [PATCH v3 07/10] vmxnet3: support jumbo frames

Stephen Hemminger stephen at networkplumber.org
Tue Mar 10 05:18:49 CET 2015


On Mon, 9 Mar 2015 23:28:39 +0000
Yong Wang <yongwang at vmware.com> wrote:

> I didn’t see where mtu is used to calculate how many rx descriptors will
> be needed for each packet.  Furthermore, as pointed out by the following
> code comments, the device requires the first rx buffer of a packet be of
> type VMXNET3_RXD_BTYPE_HEAD with the remaining buffers of type
> VMXNET3_RXD_BTYPE_NODY.  This needs to be taken care of when populating rx
> rings in vmxnet3_post_rx_bufs(). Currently we don’t do this because no
> scatter-rx is supported and only one descriptor is needed for a packet
> (thus all types should be HEAD). Otherwise, the device will complain with
> error returned.  For the 2nd rx ring, type needs to be BODY for all
> descriptors still.

Yeah the reload logic needs work.

> Related to this, could you share what tests have been done to cover these
> new features?

Not really. We only test with our product which is built on top of DPDK,
not standalone.

Most of these came from figuring what features where missing from the DPDK
driver that Intel supplied. I had independently developed a much better
virtio and vmxnet3 drivers which were well tested, but DPDK maintainers
would not accept wholesale replacement of the driver. Therefore we have
been trying to selective add the bits from my driver back into the
Intel driver.


More information about the dev mailing list