[PATCH v5 2/2] serial: mxc: have putc use the TXFIFO

[Pali Rohár]

On Tuesday 25 October 2022 12:18:56 Tim Harvey wrote:
> On Tue, Sep 6, 2022 at 5:15 AM Johannes Schneider
> <johannes.schneider at leica-geosystems.com> wrote:
> >
> > only waiting for TXEMPTY leads to corrupted messages going over the
> > wire - which is fixed by making use of the FIFO
> >
> > this change is following the linux kernel uart driver
> > (drivers/tty/serial/imx.c), which also checks UTS_TXFULL
> > instead of UTS_TXEMPTY
> >
> > Signed-off-by: Johannes Schneider <johannes.schneider at leica-geosystems.com>
> > Reviewed-by: Peng Fan <peng.fan at nxp.com>
> > Reviewed-by: Fabio Estevam <festevam at denx.de>
> > ---
> >
> > (no changes since v1)
> >
> >  drivers/serial/serial_mxc.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/serial/serial_mxc.c b/drivers/serial/serial_mxc.c
> > index ee17a960d4..af1fd1ea9b 100644
> > --- a/drivers/serial/serial_mxc.c
> > +++ b/drivers/serial/serial_mxc.c
> > @@ -311,7 +311,7 @@ static int mxc_serial_putc(struct udevice *dev, const char ch)
> >         struct mxc_serial_plat *plat = dev_get_plat(dev);
> >         struct mxc_uart *const uart = plat->reg;
> >
> > -       if (!(readl(&uart->ts) & UTS_TXEMPTY))
> > +       if (readl(&uart->ts) & UTS_TXFULL)
> >                 return -EAGAIN;
> >
> >         writel(ch, &uart->txd);
> > --
> > 2.25.1
> >
> 
> Johannes,
> 
> Since this patch I find an issue with an IMX6 board of mine gwventana:
> 
> Prior to this patch the board boots with:
> DRAM:  1 GiB
> GSCv2   : v52 0x9981 RST:VIN WDT:disabled board_temp:43C
> RTC     : 1970-01-01   0:56:15 UTC
> Core:  67 devices, 22 uclasses, devicetree: separate
> WDT:   Started watchdog at 20bc000 with servicing every 1000ms (60s timeout)
> NAND:  2048 MiB
> ...
> 
> and following this patch I get:
> ...
> DRAM:  1 GiB
> GSCv2   : v52 0x9981 RST:VIN WDT:disabled board_temp:29C
> RTC     : 1970-01-01   ~�KW�'$H�$V�W��Y.KH�� uclasses, devicetree: separate
> WDT:   Started watchdog at 20bc000 with servicing every 1000ms (60s timeout)
> NAND:  2048 MiB
> ...
> 
> The RTC line is displayed from drivers/misc/gsc.c and the Core: comes
> from dm_announce. Somehow in between the FIFO does not get drained
> before dm_announce gets called.
> 
> Adding a delay after the RTC print or reverting this patch.
> 
> Any ideas?
> 
> Best Regards,
> 
> Tim

Hello! I do not have any MXC hardware but I see there one issue.
mxc_serial_putc() function probably should not return -EAGAIN when
device is busy. But instead it should wait until it is ready.

Could you try to change code to following?

    while (readl(&uart->ts) & UTS_TXFULL)
        ;

    writel(ch, &uart->txd);