[PATCH ath-next 1/2] wifi: ath12k: Add support for handling incumbent signal interference in 6 GHz
Rameshkumar Sundaram
rameshkumar.sundaram at oss.qualcomm.com
Thu May 7 23:13:45 PDT 2026
On 5/5/2026 8:08 PM, Amith A wrote:
> From: Aishwarya R <aishwarya.r at oss.qualcomm.com>
>
> When incumbent signal interference is detected by an AP/mesh interface
> operating in the 6 GHz band, as mandated by the FCC, it is expected to
> vacate the affected channels. The firmware indicates the interference to
> the host using the WMI_DCS_INTERFERENCE_EVENT.
>
> To handle the new WMI event, first parse it to retrieve the interference
> information. Next, validate the interference-detected channel and
> the interference bitmap. The interference bitmap received from the
> firmware uses a mapping where bit 0 corresponds to the primary
> 20 MHz segment, regardless of its position within the operating
> bandwidth. Bit 1 represents the next adjacent 20 MHz segment, bit 2
> the lower 20 MHz segment of the adjacent 40 MHz segment, and so
> on, progressing sequentially across the bandwidth. However, for userspace
> consumption via mac80211, this bitmap must be transformed into a
> standardized format such that each bit position directly maps to the
> corresponding sub-channel index within the operating bandwidth.
> Finally, indicate the transformed interference bitmap to mac80211, which
> then notifies userspace of the interference. Once the incumbent signal
> interference is detected, firmware suspends TX internally on the affected
> operating channel while userspace decides the mitigation action. Userspace
> is expected to trigger a channel switch or bandwidth reduction to mitigate
> the interference. Also, add a flag handling_in_progress to indicate that
> handling of interference is in progress. Set it to true after
> indicating to mac80211 about the interference. Reset the flag to false
> after the operating channel is switched by userspace. This prevents
> processing any further interference events when there is already a
> previous event being handled. Hence, further events are processed only
> after a channel switch request is received from userspace for the
> previous event.
>
> Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.5-01651-QCAHKSWPL_SILICONZ-1
>
> Signed-off-by: Aishwarya R <aishwarya.r at oss.qualcomm.com>
> Co-developed-by: Hari Chandrakanthan <quic_haric at quicinc.com>
> Signed-off-by: Hari Chandrakanthan <quic_haric at quicinc.com>
> Signed-off-by: Amith A <amith.a at oss.qualcomm.com>
> ---
> drivers/net/wireless/ath/ath12k/core.h | 8 +
> drivers/net/wireless/ath/ath12k/mac.c | 46 +++
> drivers/net/wireless/ath/ath12k/wmi.c | 382 +++++++++++++++++++++++++
> drivers/net/wireless/ath/ath12k/wmi.h | 58 +++-
> 4 files changed, 493 insertions(+), 1 deletion(-)
>
{ ... }
> +static int
> +ath12k_wmi_incumbent_signal_interference_subtlv_parser(struct ath12k_base *ab,
> + u16 tag, u16 len,
> + const void *ptr,
> + void *data)
> +{
> + const struct ath12k_wmi_incumbent_signal_interference_params *info;
> + struct ath12k_wmi_incumbent_signal_interference_arg *arg = data;
> +
> + switch (tag) {
> + case WMI_TAG_DCS_INCUMBENT_SIGNAL_INTERFERENCE_TYPE:
> + info = ptr;
> +
should we validate len before accessing info ? or may be add an entry
for WMI_TAG_DCS_INCUMBENT_SIGNAL_INTERFERENCE_TYPE in
ath12k_wmi_tlv_policies so that ath12k_wmi_tlv_iter() can take care of
the validation.
> + arg->chan_width = le32_to_cpu(info->chan_width);
> + arg->chan_freq = le32_to_cpu(info->chan_freq);
> + arg->center_freq0 = le32_to_cpu(info->center_freq0);
> + arg->center_freq1 = le32_to_cpu(info->center_freq1);
> + arg->chan_bw_interference_bitmap =
> + le32_to_cpu(info->chan_bw_interference_bitmap);
> +
> + ath12k_dbg(ab, ATH12K_DBG_WMI,
> + "incumbent signal interference chan width %u freq %u center_freq0 %u center_freq1 %u bitmap 0x%x\n",
> + arg->chan_width, arg->chan_freq,
> + arg->center_freq0, arg->center_freq1,
> + arg->chan_bw_interference_bitmap);
> + break;
> + default:
> + ath12k_warn(ab, "Received invalid tag 0x%x for WMI DCS interference in subtlvs\n",
> + tag);
> + return -EINVAL;
> + }
> +
> + return 0;
> +}
--
Ramesh
More information about the ath12k
mailing list