Re: [PATCH rtw-next v3 7/9] wifi: rtw89: handle IEEE80211_TX_CTL_REQ_TX_STATUS frames for USB

[Fedor Pchelkin <pchelkin@ispras.ru>]

On Mon, 27. Oct 01:14, Ping-Ke Shih wrote:
> Fedor Pchelkin <pchelkin@ispras.ru> wrote:
> > On Wed, 22. Oct 07:16, Ping-Ke Shih wrote:
> > > Fedor Pchelkin <pchelkin@ispras.ru> wrote:
> > > > @@ -5849,6 +5852,7 @@ int rtw89_core_init(struct rtw89_dev *rtwdev)
> > > >         wiphy_work_init(&rtwdev->cancel_6ghz_probe_work, rtw89_cancel_6ghz_probe_work);
> > > >         INIT_WORK(&rtwdev->load_firmware_work, rtw89_load_firmware_work);
> > > >
> > > > +       skb_queue_head_init(&rtwdev->tx_rpt.queue);
> > >
> > > not sure if it's worth to initialize tx_rpt.sn to zero?
> > 
> > That shouldn't be needed because rtwdev is zero initialized in
> > rtw89_alloc_ieee80211_hw().  ieee80211_alloc_hw() fills the private
> > driver part with zeroes.
> 
> Ah. I mentioned this in wrong place. I meant that we can initialize tx_rpt.sn
> in rtw89_core_start() or do it right after downloading firmware in
> __rtw89_fw_download() like ' fw_info->h2c_seq = 0;'.

To my mind, it's not worth adding extra code to initialize tx_rpt.sn to
zero at some point as it's just a sequential number in [0x0, 0xF] range,
which is replayed to firmware and used to synchronize with it.  Actually
we can start counting from 0x1 or 0xA, it doesn't really matter to care
about counter initialization.