Hello,
On Fri, May 31, 2019 at 08:33:15AM +0300, Sergey Organov wrote:
> This is still wrong, as it turns off RTS/CTS handshake in hardware on
> TIOCM_RTS=0. Once again, set_mctrl() should not touch UCR2_CTSC, -- it's
> as simple as that.
>
> I still think it's rather what I did in the patch above is the right
> thing to do. It's simple and does the job, no surprises.
I agree that your patch implements the behaviour you intend as the right
one. I believe however that this behaviour isn't right.
We agree that with handshaking disabled (i.e CRTSCTS unset) the RTS
output should be set active iff TIOCM_RTS is requested.
With hardware handshaking enabled the two options we disagree about are:
a) RTS is not supposed to be changed, only the hardware controls the
level of RTS.
b) RTS should be set to inactive if !TIOCM_RTS. Otherwise the hardware
controls the RTS level.
An equivalent wording of b) is:
RTS should be active iff both the receiver and the software
request want it to be active.
Sergey thinks a) is the right one, I believe in b).
I think b) is sane, because with a) the kernel (or userspace) cannot
stop the other side sending data (unless they disable handshaking).
Documentation/serial/driver isn't clear here, we should improve that
once we found an agreement (probably by other people chiming in on what
they think is right).
Best regards
Uwe
--
Pengutronix e.K. | Uwe Kleine-König |
Industrial Linux Solutions | http://www.pengutronix.de/ |
