Re: [PATCH wireless-next v4 11/14] wifi: cfg80211: add PD-specific preamble and bandwidth capabilities

[Johannes Berg <johannes@sipsolutions.net>]

On Fri, 2026-03-20 at 23:29 +0530, Peddolla Harshavardhan Reddy wrote:


> +++ b/net/wireless/nl80211.c
> @@ -2482,6 +2482,20 @@ nl80211_send_pmsr_ftm_capa(const struct cfg80211_pmsr_capabilities *cap,
>  				cap->pd_max_peer_rsta_role))
>  			return -ENOBUFS;
>  	}
> +
> +	if (nla_put_u32(msg, NL80211_PMSR_FTM_CAPA_ATTR_PD_EDCA_PREAMBLES,
> +			cap->ftm.pd_edca_preambles))
> +		return -ENOBUFS;
> +	if (nla_put_u32(msg, NL80211_PMSR_FTM_CAPA_ATTR_PD_NTB_PREAMBLES,
> +			cap->ftm.pd_ntb_preambles))
> +		return -ENOBUFS;
> +	if (nla_put_u32(msg, NL80211_PMSR_FTM_CAPA_ATTR_PD_EDCA_BANDWIDTHS,
> +			cap->ftm.pd_edca_bandwidths))
> +		return -ENOBUFS;
> +	if (nla_put_u32(msg, NL80211_PMSR_FTM_CAPA_ATTR_PD_NTB_BANDWIDTHS,
> +			cap->ftm.pd_ntb_bandwidths))
> +		return -ENOBUFS;

That probably should be contingent on pd_support or so, otherwise you
end up with a bunch of zero values in userspace, which is odd?

> +	if (out->ftm.pd_request &&
> +	    !(rdev->wiphy.pmsr_capa->ftm.pd_edca_bandwidths & BIT(out->chandef.width)) &&
> +	    !(rdev->wiphy.pmsr_capa->ftm.pd_ntb_bandwidths & BIT(out->chandef.width))) {
> +		NL_SET_ERR_MSG(info->extack, "FTM: unsupported bandwidth for pd request");
> +		return -EINVAL;
> +	}

This seems odd to me, why are you adding *two* fields when they're never
treated differently? Shouldn't you check if it's an NTB or EDCA request?

> +	if (out->ftm.pd_request &&
> +	    !(capa->ftm.pd_ntb_preambles & BIT(preamble)) &&
> +	    !(capa->ftm.pd_edca_preambles & BIT(preamble))) {
> +		NL_SET_ERR_MSG_ATTR(info->extack,
> +				    tb[NL80211_PMSR_FTM_REQ_ATTR_PREAMBLE],
> +				    "FTM: invalid preamble for PD request");
> +		return -EINVAL;
> +	}
> 

Same here.

johannes