Re: [PATCH rtw-next v2 7/7] wifi: rtw88: 8812a, 8821a: respect peer A-MPDU size limits

[Jie Feng]

On Sat, 7 Nov 2026 at 00:35, Luka Gejak <luka.gejak@xxxxxxxxx> wrote:
>
> Hi Jie Feng,
>
> Two small comments below.
>
> October 31, 2026 at 8:09 AM, "Ho Jie Feng" <hjf3108@xxxxxxxxx> wrote:
>
> > diff --git a/drivers/net/wireless/realtek/rtw88/rtw8821a.c b/drivers/net/wireless/realtek/rtw88/rtw8821a.c
> > index f7391f6aa3039c7ba4d88249d968a5f1763ea866..f597edec290f64fda1c457812b240515360dc387 100644
> > --- a/drivers/net/wireless/realtek/rtw88/rtw8821a.c
> > +++ b/drivers/net/wireless/realtek/rtw88/rtw8821a.c
> > @@ -49,6 +49,13 @@ static void rtw8821a_cfg_ldo25(struct rtw_dev *rtwdev, bool enable)
> >  {
> >  }
> >
> > +static void rtw8821a_set_ampdu_factor(struct rtw_dev *rtwdev, u8 factor)
> > +{
> > +     factor = min_t(u8, factor, IEEE80211_HT_MAX_AMPDU_64K);
>
> For a VHT peer rtw_set_ampdu_factor() passes the VHT exponent here, so
> IEEE80211_VHT_MAX_AMPDU_64K would match the 8812a version. It is the
> same value, so this is only about readability.
>
> > +
> > +     rtw_write32(rtwdev, REG_AMPDU_MAX_LENGTH, (8192 << factor) - 1);
> > +}
>
> This is the same as rtw8812a_set_ampdu_factor() apart from the limit.
> Maybe a single rtw88xxa_set_ampdu_factor() that picks the limit by chip
> id, like rtw88xxa_set_crystal_cap() in patch 5 does for the mask?
I feel that this one would not deduplicate much code and defeats the point of a
per chip set_ampdu_factor callback anyway.

> Best regards,
> Luka Gejak