Re: [Intel-wired-lan] [PATCH 0/4] Configuring traffic classes via new hardware offload mechanism in tc/mqprio

[John Fastabend <john.fastabend@gmail.com>]

On 05/19/2017 05:58 PM, Amritha Nambiar wrote:
> The following series introduces a new harware offload mode in
> tc/mqprio where the TCs, the queue configurations and bandwidth rate
> limits are offloaded to the hardware. The i40e driver enables the new
> mqprio hardware offload mechanism factoring the TCs, queue
> configuration and bandwidth rates by creating HW channel VSIs.
> 

nice work, fix your time stamp and line wrapping though.


> In this mode, the priority to traffic class mapping and the user
> specified queue ranges are used to configure the traffic class when
> the 'hw' option is set to 2. This is achieved by creating HW
> channels(VSI). A new channel is created for each of the traffic class
> configuration offloaded via mqprio framework except for the first TC
> (TC0) which is for the main VSI. TC0 for the main VSI is also
> reconfigured as per user provided queue parameters. Finally,
> bandwidth rate limits are set on these traffic classes through the
> mqprio offload framework by sending these rates in addition to the
> number of TCs and the queue configurations.
> Example:
> # tc qdisc add dev eth0 root mqprio num_tc 2  map 0 0 0 0 1 1 1 1\
>   queues 4@0 4@4 min_rate 0Mbit 0Mbit max_rate 55Mbit 60Mbit hw 2
> 
> To dump the bandwidth rates:
> 
> # tc qdisc show dev eth0
>   qdisc mqprio 804a: root  tc 2 map 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0
>                queues:(0:3) (4:7)
>                min rates:0bit 0bit
>                max rates:55Mbit 60Mbit
> 

Looks reasonable to me thanks. Previously, rate limits were being set
via dcbnl but I guess this interface is slightly nicer in that it puts
all configuration in one spot. IMO it would be nice to push dcbnl users
over to this.

Thanks,
.John