Re: [PATCH] serial/liteuart; Add support for earlycon

[Stafford Horne]

On Mon, May 17, 2021 at 10:25:46AM +0200, Jiri Slaby wrote:
> On 17. 05. 21, 10:23, Stafford Horne wrote:
> > On Mon, May 17, 2021 at 10:16:43AM +0200, Jiri Slaby wrote:
> > > On 15. 05. 21, 10:45, Stafford Horne wrote:
> > > > Most litex boards using RISC-V soft cores us the sbi earlycon, however
> > > > this is not available for non RISC-V litex SoC.  This patch enables
> > > > earlycon for liteuart which is available on all Litex SoC's making
> > > > support for earycon debugging more widely available.
> > > > 
> > > > Signed-off-by: Stafford Horne <shorne@xxxxxxxxx>
> > > > Cc: Florent Kermarrec <florent@xxxxxxxxxxxxxxxx>
> > > > Cc: Mateusz Holenko <mholenko@xxxxxxxxxxxx>
> > > > Cc: Joel Stanley <joel@xxxxxxxxx>
> > > > Cc: Gabriel L. Somlo <gsomlo@xxxxxxxxx>
> > > > ---
> > > >    .../admin-guide/kernel-parameters.txt         |  5 +++
> > > >    drivers/tty/serial/Kconfig                    |  1 +
> > > >    drivers/tty/serial/liteuart.c                 | 31 +++++++++++++++++++
> > > >    3 files changed, 37 insertions(+)
> > > > 
> > > > diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt
> > > > index 04545725f187..2d4a43af8de2 100644
> > > > --- a/Documentation/admin-guide/kernel-parameters.txt
> > > > +++ b/Documentation/admin-guide/kernel-parameters.txt
> > > > @@ -1084,6 +1084,11 @@
> > > >    			the driver will use only 32-bit accessors to read/write
> > > >    			the device registers.
> > > > +		liteuart,<addr>
> > > > +			Start an early console on a litex serial port at the
> > > > +			specified address. The serial port must already be
> > > > +			setup and configured. Options are not yet supported.
> > > > +
> > > >    		meson,<addr>
> > > >    			Start an early, polled-mode console on a meson serial
> > > >    			port at the specified address. The serial port must
> > > > diff --git a/drivers/tty/serial/Kconfig b/drivers/tty/serial/Kconfig
> > > > index 0c4cd4a348f4..9ceffe6ab6fd 100644
> > > > --- a/drivers/tty/serial/Kconfig
> > > > +++ b/drivers/tty/serial/Kconfig
> > > > @@ -1531,6 +1531,7 @@ config SERIAL_LITEUART
> > > >    	depends on OF || COMPILE_TEST
> > > >    	depends on LITEX
> > > >    	select SERIAL_CORE
> > > > +	select SERIAL_EARLYCON
> > > >    	help
> > > >    	  This driver is for the FPGA-based LiteUART serial controller from LiteX
> > > >    	  SoC builder.
> > > > diff --git a/drivers/tty/serial/liteuart.c b/drivers/tty/serial/liteuart.c
> > > > index 64842f3539e1..38c472487e68 100644
> > > > --- a/drivers/tty/serial/liteuart.c
> > > > +++ b/drivers/tty/serial/liteuart.c
> > > > @@ -372,6 +372,37 @@ static int __init liteuart_console_init(void)
> > > >    console_initcall(liteuart_console_init);
> > > >    #endif /* CONFIG_SERIAL_LITEUART_CONSOLE */
> > > > +#ifdef CONFIG_SERIAL_EARLYCON
> > > > +static void early_liteuart_putc(struct uart_port *port, int c)
> > > > +{
> > > > +	while (litex_read8(port->membase + OFF_TXFULL))
> > > > +		cpu_relax();
> > > > +
> > > > +	litex_write8(port->membase + OFF_RXTX, c);
> > > 
> > > Hi,
> > > 
> > > am I missing something or this doesn't differ to liteuart_putchar?
> > 
> > Hi, No you are right, I missed that thanks for catching it.
> > 
> > I should be able to remove this function and use putchar below.
> 
> OK.
> 
> I've just noticed you add an ifdef CONFIG_SERIAL_EARLYCON here, but you
> select SERIAL_EARLYCON in Kconfig above. So the ifdef is sort of pointless?
> (Or you shouldn't select SERIAL_EARLYCON.)

Ah yes, this I thought might have been wrong when I wrote it.  I think my
problem here is I was just copying how some other patches implemented earlycon.

 omap-serial - has ifdefs, doesn't select SERIAL_EARLYCON - looks ok
 sifive - has ifdefs + select SERIAL_EARLYCON - maybe needs cleanup too

-Stafford