Re: [PATCH] pl011: assert RTS signal in case the receiver uses flow control

[Tim Deegan <tim@xen.org>]

At 08:12 +0200 on 03 Sep (1378195944), Andre Przywara wrote:
> On 09/02/2013 06:21 PM, Tim Deegan wrote:
> >At 16:51 +0200 on 02 Sep (1378140705), Andre Przywara wrote:
> >>Although we do not support hardware flow control in the Xen driver
> >>for the PL011 UART, the other end may be configured to use it.
> >>In this case it waits in vain for the RTS signal to be asserted by
> >>the host and will never transmit any characters.
> >>This fixes the UART input on Calxeda Midway, which uses hardware
> >>flow control for the serial-over-LAN functionality.
> >>
> >>Signed-off-by: Andre Przywara <andre.przywara@linaro.org>
> >>---
> >>  xen/drivers/char/pl011.c         | 4 ++--
> >>  xen/include/asm-arm/pl011-uart.h | 4 ++++
> >>  2 files changed, 6 insertions(+), 2 deletions(-)
> >>
> >>diff --git a/xen/drivers/char/pl011.c b/xen/drivers/char/pl011.c
> >>index 3ec6e10..e340961 100644
> >>--- a/xen/drivers/char/pl011.c
> >>+++ b/xen/drivers/char/pl011.c
> >>@@ -120,8 +120,8 @@ static void __init pl011_init_preirq(struct 
> >>serial_port *port)
> >>      pl011_write(uart, IMSC, 0);
> >>      pl011_write(uart, ICR, ALLI);
> >>
> >>-    /* Enable the UART for RX and TX; no flow ctrl */
> >>-    pl011_write(uart, CR, RXE | TXE | UARTEN);
> >>+    /* Enable the UART for RX and TX; assert RTS in case the other end 
> >>cares */
> >>+    pl011_write(uart, CR, RTS | RXE | TXE | UARTEN);
> >>  }
> >>
> >>  static void __init pl011_init_postirq(struct serial_port *port)
> >>diff --git a/xen/include/asm-arm/pl011-uart.h 
> >>b/xen/include/asm-arm/pl011-uart.h
> >>index 3332c51..123f477 100644
> >>--- a/xen/include/asm-arm/pl011-uart.h
> >>+++ b/xen/include/asm-arm/pl011-uart.h
> >>@@ -38,6 +38,10 @@
> >>  #define DMACR  (0x48)
> >>
> >>  /* CR bits */
> >>+#define CTSEN  (1<<15) /* automatic CTS hardware flow control */
> >>+#define RTSEN  (1<<14) /* automatic RTS hardware flow control */
> >
> >Would this bit not be better than blindly setting RTS?
> 
> I don't think so. This sets and clears RTS according to the FIFO fill 
> level. Since we don't claim to support h/w flow control, we better leave 
> this disabled.
> Just asserting RTS all of the time should have no influence on 
> non-prepared devices

Sure, but frobbing RTS up and down won't have any effect on devices that
ignore it either, and if the hardware will actually DTRT when the other
end _wants_ flow control, why not?

But it's no big deal either way; asserting RTS is harmless and better
than what we have.

> [1]. So I will go with simply ORing in our requested 
> bits and leave the flow-control bits as they have been setup before.

That's fine too, at least in cases where we're alreay taking the baud
settings from the existing setup (which is all the time now?).

Tim.