On Sun, Dec 04, 2022 at 01:36:55PM +0100, Pali Rohár wrote:
U-Boot serial code already handles -EAGAIN value from getc and putc callbacks. So change drivers code to return -EAGAIN when HW is busy instead of doing its own busy loop and waiting until HW is ready.
Signed-off-by: Pali Rohár pali@kernel.org Reviewed-by: Simon Glass sjg@chromium.org Reviewed-by: Stefan Roese sr@denx.de
[snip]
diff --git a/drivers/serial/serial_lpuart.c b/drivers/serial/serial_lpuart.c index ff576da516d4..d7259d531b55 100644 --- a/drivers/serial/serial_lpuart.c +++ b/drivers/serial/serial_lpuart.c @@ -168,8 +168,8 @@ static void _lpuart_serial_setbrg(struct udevice *dev, static int _lpuart_serial_getc(struct lpuart_serial_plat *plat) { struct lpuart_fsl *base = plat->reg;
- while (!(__raw_readb(&base->us1) & (US1_RDRF | US1_OR)))
schedule();
if (!(__raw_readb(&base->us1) & (US1_RDRF | US1_OR)))
return -EAGAIN;barrier();
@@ -181,8 +181,8 @@ static void _lpuart_serial_putc(struct lpuart_serial_plat *plat, { struct lpuart_fsl *base = plat->reg;
- while (!(__raw_readb(&base->us1) & US1_TDRE))
schedule();
if (!(__raw_readb(&base->us1) & US1_TDRE))
return -EAGAIN;__raw_writeb(c, &base->ud);
}
This is non-trivially not right for this driver. ->putc here is set to lpuart_serial_putc which calls _lpuart_serial_putc OR _lpuart32_serial_putc, so there's the lpuart32 cases to address, and then lpuart_serial_putc needs to just return whatever one it called rather than always returning 0.