On Thu, 2021-11-18 at 08:40 -0800, greearb@xxxxxxxxxxxxxxx wrote:
> From: Ben Greear <greearb@xxxxxxxxxxxxxxx>
>
> Instead of every 500ms, this will do better job of catching wraps
> of 16-bit pkt counters.
>
> Signed-off-by: Ben Greear <greearb@xxxxxxxxxxxxxxx>
> ---
> drivers/net/wireless/mediatek/mt76/mt7915/mac.c | 6 +++++-
> 1 file changed, 5 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/net/wireless/mediatek/mt76/mt7915/mac.c
> b/drivers/net/wireless/mediatek/mt76/mt7915/mac.c
> index 3c89c7573015..7cfa6b2abda5 100644
> --- a/drivers/net/wireless/mediatek/mt76/mt7915/mac.c
> +++ b/drivers/net/wireless/mediatek/mt76/mt7915/mac.c
> @@ -2390,7 +2390,11 @@ void mt7915_mac_work(struct work_struct *work)
> mutex_lock(&mphy->dev->mutex);
>
> mt76_update_survey(mphy);
> - if (++mphy->mac_work_count == 5) {
> +
> + /* this method is called about every 100ms. Some pkt counters
> are 16-bit,
> + * so poll every 200ms to keep overflows at a minimum.
> + */
> + if (++mphy->mac_work_count == 2) {
> mphy->mac_work_count = 0;
>
> mt7915_mac_update_stats(phy);
I guess we need to do some testing on embedded platform to see it hurts
performance.
Ryder
