On Thu, 2023-05-25 at 19:07 +0300, Kalle Valo wrote:
>
> Ping-Ke Shih <pkshih@xxxxxxxxxxx> writes:
>
> > The register-based H2C/C2H are used to exchange commands and events with
> > firmware. The exchange data is limited, but it is relatively simple,
> > because it can work before HCI initialization. To make these code clean,
> > use struct to access them. This patch doesn't change logic at all.
> >
> > Signed-off-by: Ping-Ke Shih <pkshih@xxxxxxxxxxx>
>
> [...]
>
> > --- a/drivers/net/wireless/realtek/rtw89/fw.h
> > +++ b/drivers/net/wireless/realtek/rtw89/fw.h
> > @@ -18,15 +18,51 @@ enum rtw89_fw_dl_status {
> > RTW89_FWDL_WCPU_FW_INIT_RDY = 7
> > };
> >
> > -#define RTW89_GET_C2H_HDR_FUNC(info) \
> > - u32_get_bits(info, GENMASK(6, 0))
> > -#define RTW89_GET_C2H_HDR_LEN(info) \
> > - u32_get_bits(info, GENMASK(11, 8))
> > +struct rtw89_c2hreg_hdr {
> > + u32 w0;
> > +};
>
> Why this is u32? Shouldn't it be __le32?
>
> > +#define RTW89_C2HREG_HDR_FUNC_MASK GENMASK(6, 0)
> > +#define RTW89_C2HREG_HDR_ACK BIT(7)
> > +#define RTW89_C2HREG_HDR_LEN_MASK GENMASK(11, 8)
> > +#define RTW89_C2HREG_HDR_SEQ_MASK GENMASK(15, 12)
> > +
> > +struct rtw89_c2hreg_phycap {
> > + u32 w0;
> > + u32 w1;
> > + u32 w2;
> > + u32 w3;
> > +} __packed;
>
> Here as well? And I saw more in the patch.
>
> Of course these were already there so isn't a problem introduced by this
> patchset, but I started wondering if we are missing some little endian
> types?
>
I had the same question as yours when I did this conversion, but they
are correct because we access these H2C commands/C2H events via registers
which are CPU order.
A bug I found is that use struct bit-field to access these data. This
will cause big-endian machine wrong, and I fix them by this patch as well.
I hope we don't miss something.
Ping-Ke
