As of commit 18e8cf159177100e ("serial: sh-sci: increase RX FIFO trigger
defaults for (H)SCIF") in Linux v4.11-rc1, the serial console on the
QEMU SH4 target is broken: it delays serial input until enough data has
been received.

Since aformentioned commit, the Linux SCIF driver programs the Receive
FIFO Data Count Trigger bits in the FIFO Control Register, to postpone
generating a receive interrupt until:
  1. At least the receive trigger count of bytes of data are available
     in the receive FIFO, OR
  2. No further data has been received for at least 15 etu after the
     last received data.

While QEMU implements the former, it does not implement the latter.
Hence the receive interrupt is not generated until the former condition
is met.

Fix this by adding basic timeout handling.  As the QEMU SCIF emulation
ignores any serial speed programming, the timeout value used conforms to
a default speed of 9600 bps, which is fine for any interative console.

Reported-by: Rob Landley <rob@landley.net>
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
---
 hw/char/sh_serial.c | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/hw/char/sh_serial.c b/hw/char/sh_serial.c
index 373a40595fd975d1..12831561a6c8b137 100644
--- a/hw/char/sh_serial.c
+++ b/hw/char/sh_serial.c
@@ -29,6 +29,7 @@
 #include "hw/sh4/sh.h"
 #include "chardev/char-fe.h"
 #include "qapi/error.h"
+#include "qemu/timer.h"
 
 //#define DEBUG_SERIAL
 
@@ -63,6 +64,8 @@ typedef struct {
     int rtrg;
 
     CharBackend chr;
+    QEMUTimer *fifo_timeout_timer;
+    uint64_t etu; /* Elementary Time Unit (ns) */
 
     qemu_irq eri;
     qemu_irq rxi;
@@ -314,6 +317,16 @@ static int sh_serial_can_receive1(void *opaque)
     return sh_serial_can_receive(s);
 }
 
+static void sh_serial_timeout_int(void *opaque)
+{
+    sh_serial_state *s = opaque;
+
+    s->flags |= SH_SERIAL_FLAG_RDF;
+    if (s->scr & (1 << 6) && s->rxi) {
+        qemu_set_irq(s->rxi, 1);
+    }
+}
+
 static void sh_serial_receive1(void *opaque, const uint8_t *buf, int size)
 {
     sh_serial_state *s = opaque;
@@ -330,8 +343,12 @@ static void sh_serial_receive1(void *opaque, const uint8_t *buf, int size)
                 if (s->rx_cnt >= s->rtrg) {
                     s->flags |= SH_SERIAL_FLAG_RDF;
                     if (s->scr & (1 << 6) && s->rxi) {
+                        timer_del(s->fifo_timeout_timer);
                         qemu_set_irq(s->rxi, 1);
                     }
+                } else {
+                    timer_mod(s->fifo_timeout_timer,
+                        qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL) + 15 * s->etu);
                 }
             }
         }
@@ -402,6 +419,9 @@ void sh_serial_init(MemoryRegion *sysmem,
                                  sh_serial_event, NULL, s, NULL, true);
     }
 
+    s->fifo_timeout_timer = timer_new_ns(QEMU_CLOCK_VIRTUAL,
+                                         sh_serial_timeout_int, s);
+    s->etu = NANOSECONDS_PER_SECOND / 9600;
     s->eri = eri_source;
     s->rxi = rxi_source;
     s->txi = txi_source;
-- 
2.17.1

> On July 30, 2018 at 3:02 PM Geert Uytterhoeven <geert+renesas@glider.be> wrote:
> 
> 
> As of commit 18e8cf159177100e ("serial: sh-sci: increase RX FIFO trigger
> defaults for (H)SCIF") in Linux v4.11-rc1, the serial console on the
> QEMU SH4 target is broken: it delays serial input until enough data has
> been received.
> 
> Since aformentioned commit, the Linux SCIF driver programs the Receive
> FIFO Data Count Trigger bits in the FIFO Control Register, to postpone
> generating a receive interrupt until:
>   1. At least the receive trigger count of bytes of data are available
>      in the receive FIFO, OR
>   2. No further data has been received for at least 15 etu after the
>      last received data.
> 
> While QEMU implements the former, it does not implement the latter.
> Hence the receive interrupt is not generated until the former condition
> is met.
> 
> Fix this by adding basic timeout handling.  As the QEMU SCIF emulation
> ignores any serial speed programming, the timeout value used conforms to
> a default speed of 9600 bps, which is fine for any interative console.
> 
> Reported-by: Rob Landley <rob@landley.net>
> Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>

Works for me, kernel 4.18-rc7 for rts7751r2dplus.

Tested-by: Ulrich Hecht <uli@fpond.eu>

CU
Uli

On 07/30/2018 10:18 AM, Ulrich Hecht wrote:>> On July 30, 2018 at 3:02 PM Geert
Uytterhoeven <geert+renesas@glider.be> wrote:
>> Fix this by adding basic timeout handling.  As the QEMU SCIF emulation
>> ignores any serial speed programming, the timeout value used conforms to
>> a default speed of 9600 bps, which is fine for any interative console.
>>
>> Reported-by: Rob Landley <rob@landley.net>
>> Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
> 
> Works for me, kernel 4.18-rc7 for rts7751r2dplus.

Works for me too.

Tested-by: Rob Landley <rob@landley.net>

Rob

On Tue, Jul 31, 2018 at 11:36:46AM -0500, Rob Landley wrote:
> On 07/30/2018 10:18 AM, Ulrich Hecht wrote:>> On July 30, 2018 at 3:02 PM Geert
> Uytterhoeven <geert+renesas@glider.be> wrote:
> >> Fix this by adding basic timeout handling.  As the QEMU SCIF emulation
> >> ignores any serial speed programming, the timeout value used conforms to
> >> a default speed of 9600 bps, which is fine for any interative console.
> >>
> >> Reported-by: Rob Landley <rob@landley.net>
> >> Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
> > 
> > Works for me, kernel 4.18-rc7 for rts7751r2dplus.
> 
> Works for me too.
> 
> Tested-by: Rob Landley <rob@landley.net>

Me too. This would be a very welcome fix for a longstanding and
annoying problem.

Tested-by: Rich Felker <dalias@libc.org>

Rich