On Fri, 17 Feb 2023, Biju Das wrote:
> Hi Ilpo,
>
> Thanks for the feedback.
>
> > Subject: RE: [PATCH v4 5/6] serial: 8250_em: Use pseudo offset for UART_FCR
> >
> > On Fri, 17 Feb 2023, Biju Das wrote:
> >
> > > HI Ilpo,
> > >
> > > > Subject: RE: [PATCH v4 5/6] serial: 8250_em: Use pseudo offset for
> > > > UART_FCR
> > > >
> > > > On Fri, 17 Feb 2023, Biju Das wrote:
> > > >
> > > > > Hi Ilpo,
> > > > >
> > > > > Thanks for the feedback.
> > > > >
> > > > > > Subject: Re: [PATCH v4 5/6] serial: 8250_em: Use pseudo offset
> > > > > > for UART_FCR
> > > > > >
> > > > > > On Fri, 17 Feb 2023, Biju Das wrote:
> > > > > >
> > > > > > > UART_FCR shares the same offset with UART_IIR. We cannot use
> > > > > > > UART_FCR in serial8250_em_serial_in() as it overlaps with
> > UART_IIR.
> > > > > > >
> > > > > > > This patch introduces a macro UART_FCR_EM with a high value to
> > > > > > > avoid overlapping with existing UART_* register defines.
> > > > > > >
> > > > > > > This will help us to use UART_FCR_EM consistently in
> > > > > > > serial8250_em_ serial{_in/_out} function to read/write FCR
> > register.
> > > > > > >
> > > > > > > Signed-off-by: Biju Das <biju.das.jz@xxxxxxxxxxxxxx>
> > > > > > > Suggested-by: Ilpo Järvinen <ilpo.jarvinen@xxxxxxxxxxxxxxx>
> > > > > > > ---
> > > > > > > v4:
> > > > > > > * New patch. Used UART_FCR_EM for read/write of FCR register.
> > > > > > > ---
> > > > > > > drivers/tty/serial/8250/8250_em.c | 13 ++++++++++++-
> > > > > > > 1 file changed, 12 insertions(+), 1 deletion(-)
> > > > > > >
> > > > > > > diff --git a/drivers/tty/serial/8250/8250_em.c
> > > > > > > b/drivers/tty/serial/8250/8250_em.c
> > > > > > > index 499d7a8847ec..4165fd3bb17a 100644
> > > > > > > --- a/drivers/tty/serial/8250/8250_em.c
> > > > > > > +++ b/drivers/tty/serial/8250/8250_em.c
> > > > > > > @@ -19,6 +19,13 @@
> > > > > > > #define UART_DLL_EM 9
> > > > > > > #define UART_DLM_EM 10
> > > > > > >
> > > > > > > +/*
> > > > > > > + * A high value for UART_FCR_EM avoids overlapping with
> > > > > > > +existing
> > > > > > > +UART_*
> > > > > > > + * register defines. UART_FCR_EM_HW is the real HW register
> > offset.
> > > > > > > + */
> > > > > > > +#define UART_FCR_EM 0x10003
> > > > > > > +#define UART_FCR_EM_HW 3
> > > > > > > +
> > > > > > > struct serial8250_em_priv {
> > > > > > > int line;
> > > > > > > };
> > > > > > > @@ -29,12 +36,14 @@ static void
> > > > > > > serial8250_em_serial_out(struct uart_port
> > > > > > *p, int offset, int value)
> > > > > > > case UART_TX: /* TX @ 0x00 */
> > > > > > > writeb(value, p->membase);
> > > > > > > break;
> > > > > > > - case UART_FCR: /* FCR @ 0x0c (+1) */
> > > > > >
> > > > > > 8250_port wants this to remain in place, I think. Otherwise it's
> > > > > > attempts to set UART_FCR will end up into wrong destination.
> > > > >
> > > > > Oops, next patch has this change.
> > > > >
> > > > > + case UART_FCR:
> > > > > + serial8250_em_reg_update(p, UART_FCR_EM, value);
> > > > >
> > > > > I just need to keep UART_FCR for this patch and remove it from
> > > > > "serial8250_em_serial_out_helper" on the next patch
> > > >
> > > > IMHO, the extra layer + _helper seems pretty unnecessary. I don't
> > > > see any useful thing it achieves over just having to serial_out
> > functions.
> > >
> > > Without helper, it will become cyclic right??
> > >
> > > serial8250_em_serial_out() needs to call serial8250_em_reg_update()
> > >
> > > serial8250_em_reg_update() will call serial8250_em_serial_out() for
> > writes??
> >
> > With your most recent code, yes it seems to happen. But why did you remove
> > the separate serial_out for RZ. There wasn't any cyclicness when it called
> > serial8250_em_serial_out().
> >
> > I'm a bit lost now, are you saying that this change is needed for all
> > devices 8250_em supports (which is different from what you initially
> > presented in the early versions of this patchset)?
>
> Yes, That is correct. Please see the discussion thread related to this[1].
>
> Geert pointed out that UART register sets between RZ/V2M and
> EMMA mobile SoC are almost identical.
>
> "According to R19UH0040EJ0400 Rev.4.00 it is available on EMEV2, and the
> layout looks identical to RZ/V2M."
>
> Niklas tested previous patch series on EMEV2 board and
> It detected port type as 16750 and done read/write test
> with 64-bytes fifo enabled.
>
> EMMA mobile SoC is very old and hardware manual
> is not updated for long time related.
>
> So we guess it is safe to apply this restriction for this SoC
> aswell, since there is no regression.
Okay, fair enough.
--
i.
