Jeff Johnson <quic_jjohnson@xxxxxxxxxxx> writes:
>>> @@ -279,6 +334,23 @@ int ath12k_dp_tx(struct ath12k *ar, struct ath12k_vif *arvif,
>>> goto fail_remove_tx_buf;
>>> }
>>>
>>> + if (iova_mask &&
>>> + (unsigned long)skb->data & iova_mask) {
>>> + ret = ath12k_dp_tx_align_payload(ab, &skb);
>>> + if (ret) {
>>> + dev_warn_once(ab->dev, "failed to align TX buffer %d\n", ret);
>>
>> Why dev_warn_once()? I changed it to ath12k_warn() in the pending
>> branch.
>
> My concern was that if this is an ongoing issue that you'd end up spamming the
> kernel log. But I guess the rate limiting will reduce the spam to no more than
> 10 logs in a 5 second interval
Yeah, ratelimiting used by ath12k_warn() should be safe. It would be
consistent to have ath12k_warn_once() instead of using dev_warn_once()
but in this it's better to print the warning more than once so that
users don't miss it.
--
https://patchwork.kernel.org/project/linux-wireless/list/
https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
