On Thursday 17 February 2022 15:16:33 Marek Behún wrote:
> On Thu, 17 Feb 2022 10:26:18 +0100
> Pali Rohár <p...@kernel.org> wrote:
>
> > Allow to specify input parameters, define all available mbox commands
> > supported by CZ.NIC secure firmware + Marvell fuse.bin firmware and fix
> CZ.NIC's and also Marvell's
> > parsing response from Marvell OTP commands.
> >
> > Signed-off-by: Pali Rohár <p...@kernel.org>
>
> ...
>
> > #define MBOX_STS_ERROR(s) ((s) & (3 << 30))
> > #define MBOX_STS_VALUE(s) (((s) >> 10) & 0xfffff)
> > #define MBOX_STS_CMD(s) ((s) & 0x3ff)
> > +#define MBOX_STS_MARVELL_ERROR(s) ((s) == 0 ? 0 : (s) == 2 ? ETIMEDOUT :
> > (s) == 3 ? EINVAL : (s) == 4 ? ENOSYS : EIO)
>
> This is starting to become too complicated for a macro :-(
It is straightforward switch macro. I can reformat it to be more
readable, e.g.:
#define MBOX_STS_MARVELL_ERROR(s) ((s) == 0 ? 0 : \
(s) == 2 ? ETIMEDOUT : \
(s) == 3 ? EINVAL : \
(s) == 4 ? ENOSYS : \
EIO)
> What do you think about converting all these to a static function?
> Something like
>
> static int mbox_parse_status(u32 status, u16 *cmd, u32 *value,
> bool marvell)
> {
> ...
> }
>
> >
> > #ifndef _MVEBU_MBOX_H
> > @@ -16,8 +17,24 @@ enum mbox_cmd {
> >
> > MBOX_CMD_OTP_READ,
> > MBOX_CMD_OTP_WRITE,
> > +
> > + MBOX_CMD_REBOOT,
> > +
> > + /* OTP read commands supported by Marvell fuse.bin firmware */
> Marvell's
> > + MBOX_CMD_OTP_READ_1B = 257,
> > + MBOX_CMD_OTP_READ_8B,
> > + MBOX_CMD_OTP_READ_32B,
> > + MBOX_CMD_OTP_READ_64B,
> > + MBOX_CMD_OTP_READ_256B,
> > +
> > + /* OTP write commands supported by Marvell fuse.bin firmware */
> Marvell's
