On Thu, 23 Jun 2022, Jiri Slaby wrote:
> On 21. 06. 22, 14:49, Ilpo Järvinen wrote:
> > --- a/include/linux/serial.h
> > +++ b/include/linux/serial.h
> > @@ -9,7 +9,6 @@
> > #ifndef _LINUX_SERIAL_H
> > #define _LINUX_SERIAL_H
> > -#include <asm/page.h>
> > #include <uapi/linux/serial.h>
> > /* Helper for dealing with UART_LCR_WLEN* defines */
> > @@ -25,11 +24,6 @@ struct async_icount {
> > __u32 buf_overrun;
> > };
> > -/*
> > - * The size of the serial xmit buffer is 1 page, or 4096 bytes
> > - */
> > -#define SERIAL_XMIT_SIZE PAGE_SIZE
> > -
> > #include <linux/compiler.h>
> > #endif /* _LINUX_SERIAL_H */
> > diff --git a/include/linux/serial_core.h b/include/linux/serial_core.h
> > index 5518b70177b3..1767dee98021 100644
> > --- a/include/linux/serial_core.h
> > +++ b/include/linux/serial_core.h
> > @@ -296,6 +296,7 @@ struct uart_state {
> > struct uart_port *uart_port;
> > };
> > +/* The size of the serial xmit buffer is 1 page, or 4096 bytes */
>
> That is not completely true, I'd remove the "or" part. Hmm, but then the
> comment is somehow superfluous as it says what we can see below. So maybe
> don't add it at all?
>
> > #define UART_XMIT_SIZE PAGE_SIZE
Yeah, the comment is probably not that useful. I managed to get into
an "autopilot mode" with a sole goal to preserve the comment rather than
thinking/reading it through whether it's worth preserving.
--
i.
