[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <7b4d7b89-a69f-453a-bf9c-d0ccdd12d76e@tuxon.dev>
Date: Thu, 31 Jul 2025 12:07:28 +0300
From: Claudiu Beznea <claudiu.beznea@...on.dev>
To: "Karumanchi, Vineeth" <vineeth@....com>, vineeth.karumanchi@....com,
nicolas.ferre@...rochip.com, andrew+netdev@...n.ch, davem@...emloft.net,
edumazet@...gle.com, kuba@...nel.org, pabeni@...hat.com
Cc: git@....com, netdev@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH net-next 5/6] net: macb: Implement TAPRIO TC offload
command interface
On 29.07.2025 12:38, Karumanchi, Vineeth wrote:
> Hi Claudiu,
>
>
> On 7/26/2025 5:59 PM, claudiu beznea (tuxon) wrote:
> <...>
>>> + int err = 0;
>>> +
>>> + switch (taprio->cmd) {
>>> + case TAPRIO_CMD_REPLACE:
>>> + err = macb_taprio_setup_replace(ndev, taprio);
>>> + break;
>>> + case TAPRIO_CMD_DESTROY:
>>> + macb_taprio_destroy(ndev);
>>
>> macb_taprio_setup_replace() along with macb_taprio_destroy() touch HW
>> registers. Could macb_setup_taprio() be called when the interface is
>> runtime suspended?
>>
>>
>
> Nice catch!
>
> I will leverage pm_runtime_suspended(&pdev->dev) check before configuring.
>
>>> + break;
>>> + default:
>>> + err = -EOPNOTSUPP;
>>> + }
>>> +
>>> + return err;
>>> +}
>>> +
>>> +static int macb_setup_tc(struct net_device *dev, enum tc_setup_type
>>> type, void *type_data)
>>> +{
>>> + if (!dev || !type_data)
>>> + return -EINVAL;
>>> +
>>> + switch (type) {
>>> + case TC_SETUP_QDISC_TAPRIO:
>>> + return macb_setup_taprio(dev, type_data);
>>
>> Same here.
>>
>>> + default:
>>> + return -EOPNOTSUPP;
>>> + }
>>> +}
>>> +
>>> static const struct net_device_ops macb_netdev_ops = {
>>> .ndo_open = macb_open,
>>> .ndo_stop = macb_close,
>>> @@ -4284,6 +4316,7 @@ static const struct net_device_ops macb_netdev_ops
>>> = {
>>> .ndo_features_check = macb_features_check,
>>> .ndo_hwtstamp_set = macb_hwtstamp_set,
>>> .ndo_hwtstamp_get = macb_hwtstamp_get,
>>> + .ndo_setup_tc = macb_setup_tc,
>>
>> This patch (or parts of it) should be merged with the previous ones.
>> Otherwise you introduce patches with code that is unused.
>>
>
> Clubbing all comments on patch organization:
> I see that patch series gets merged into 2 set only.
>
> 1/6 + 2/6 + 3/6 + 4/6 + 5/6 ==> 1/2
> 6/6 ==> 2/2
That should be good.
Thank you,
Claudiu
>
> Please let me know your thoughts or suggestions.
>
>
> Thanks
Powered by blists - more mailing lists