[dpdk-dev] [PATCH 35/38] net/dpaa: add support for packet type parsing

[Shreyansh Jain]

On Friday 30 June 2017 05:10 PM, Shreyansh Jain wrote:
> On Wednesday 28 June 2017 09:20 PM, Ferruh Yigit wrote:
>> On 6/16/2017 6:41 AM, Shreyansh Jain wrote:
>>> Signed-off-by: Hemant Agrawal <hemant.agrawal at nxp.com>
>>> Signed-off-by: Shreyansh Jain <shreyansh.jain at nxp.com>
>>
[...]
>>
>>> +
>>> +	m->tx_offload = annot->parse.ip_off[0];
>>> +	m->tx_offload |= (annot->parse.l4_off - annot->parse.ip_off[0])
>>> +					<< DPAA_PKT_L3_LEN_SHIFT;
>>
>> This is a received mbuf right? Is it required to set tx_offload flag?
>>
[...]

I had not replied to this in my previous response.
DPAA hardware fills parsed information into the annotation (annot) area.
When a packet is Rx'd, it would contain information about where the IP
offset field it. Once we read the packet, the 'annot' area is
overwritten in subsequent cycles.

This packet received may be forwarded, in which case, this information
(preserved in m->tx_offload) would be useful for optimized performance.

Indeed, this is one of the cases, but at least some optimization is
achieved using this.

-
Shreyansh