openbmc.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
From: "Cédric Le Goater" <clg@kaod.org>
To: "Alex Bennée" <alex.bennee@linaro.org>,
	"Abhishek Singh Dagur" <abhishek@drut.io>
Cc: Peter Maydell <peter.maydell@linaro.org>,
	openbmc@lists.ozlabs.org,
	Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>,
	qemu-devel@nongnu.org
Subject: Re: Using QEMU how to redirect serial /dev/ttyS2 output of guest machine to host machine.
Date: Fri, 17 Mar 2023 18:24:26 +0100	[thread overview]
Message-ID: <b2c11d07-ee58-715d-2994-3add1c60b95d@kaod.org> (raw)
In-Reply-To: <87mt4b73sv.fsf@linaro.org>

On 3/17/23 17:06, Alex Bennée wrote:
> 
> Abhishek Singh Dagur <abhishek@drut.io> writes:
> 
> (cc aspeed maintainers)
> 
>> Hi all,
>>
>> We are using obmc-phosphor-image on an ast2500 board which is trying to communicate with other devices
>> over serial port /dev/ttyS2.
>> As we are trying to emulate the machine on qemu we need to redirect the request to the host machine so
>> that it can handle this request and return appropriately.
>> We tried using QEMU options like -serial ,-chardev but still not the
>> concrete way we get to do it.
> 
> Yeah I'm afraid its non-obvious, certainly for built in serial ports.
> Try something like:
> 
>    ./qemu-system-aarch64 -M ast2500-evb \
>      -serial null -serial null -serial chardev:myserial \
>      -chardev file,id=myserial,path=output.txt \
>      $MORE_OPTIONS
> 
> You have to add a -serial for each serial port up to the one you care
> about and then set the chardev for it.
> 
> If you where adding a device to the system then you can explicitly set
> the target chardev for it with something like:
> 
>    -device isa-serial,iobase=nnn,irq=nnn,chardev=ID
> 
>> It will be very helpful if you can provide us some guidance on this.
> 
> Another quirk for the aspeed boards seems to be the default uart can be
> an arbitrary one depending on the board model:
> 
> 334:    aspeed_soc_uart_set_chr(s, amc->uart_default, serial_hd(0));
> 336:        if (uart == amc->uart_default) {
> 1112:    amc->uart_default = ASPEED_DEV_UART5;
> 1407:    amc->uart_default = ASPEED_DEV_UART1;
> 
> as a result ASPEED_DEV_UART5 will always be the first serial port
> (serial_hd(0)). I don't know how Linux numbers them but worth being
> aware of.

Yes. UART5 is the general default but it depends on the board definition
and the fuji was the first to require an exception. See commit 5d63d0c76c
("hw/arm/aspeed: Allow machine to set UART default")

Then, it became more complex with commit d2b3eaefb4 ("aspeed: Refactor
UART init for multi-SoC machines"). That's another topic.

Abhishek,

I am afraid, you will need to add a new board to fit what's in the DT.

Or, here is a little patch adding a machine option to set the default uart.
It was never merged because it is a bit of hack, give it a try and we
will discuss. Use :

   -M ast2500-evb,uart-default=uart2


Thanks,

C.

 From 0d0700ae772fa5236914e96af1be5afcf0d4a994 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?C=C3=A9dric=20Le=20Goater?= <clg@kaod.org>
Date: Fri, 17 Mar 2023 18:21:54 +0100
Subject: [PATCH] aspeed: Add a "uart-default" machine option
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Cédric Le Goater <clg@kaod.org>
---
  hw/arm/aspeed.c | 31 +++++++++++++++++++++++++++++--
  1 file changed, 29 insertions(+), 2 deletions(-)

diff --git a/hw/arm/aspeed.c b/hw/arm/aspeed.c
index 806bb10707..e0335cf167 100644
--- a/hw/arm/aspeed.c
+++ b/hw/arm/aspeed.c
@@ -45,6 +45,7 @@ struct AspeedMachineState {
      bool mmio_exec;
      char *fmc_model;
      char *spi_model;
+    uint32_t uart_default;
      uint32_t hw_strap1;
  };
  
@@ -337,10 +338,11 @@ static void connect_serial_hds_to_uarts(AspeedMachineState *bmc)
      AspeedMachineClass *amc = ASPEED_MACHINE_GET_CLASS(bmc);
      AspeedSoCState *s = &bmc->soc;
      AspeedSoCClass *sc = ASPEED_SOC_GET_CLASS(s);
+    int uart_default = bmc->uart_default ? bmc->uart_default : amc->uart_default;
  
-    aspeed_soc_uart_set_chr(s, amc->uart_default, serial_hd(0));
+    aspeed_soc_uart_set_chr(s, uart_default, serial_hd(0));
      for (int i = 1, uart = ASPEED_DEV_UART1; i < sc->uarts_num; i++, uart++) {
-        if (uart == amc->uart_default) {
+        if (uart == uart_default) {
              continue;
          }
          aspeed_soc_uart_set_chr(s, uart, serial_hd(i));
@@ -1145,6 +1147,25 @@ static void aspeed_set_spi_model(Object *obj, const char *value, Error **errp)
      bmc->spi_model = g_strdup(value);
  }
  
+const QEnumLookup UartDefault_lookup = {
+    .array =
+        (const char *const[]) {
+            [0] = "none",
+            [1] = "uart1",
+            [2] = "uart2",
+            [3] = "uart3",
+            [4] = "uart4",
+            [5] = "uart5",
+        },
+    .size = 6
+};
+
+static void aspeed_set_uart_default(Object *obj, int val, Error **err)
+{
+    AspeedMachineState *bmc = ASPEED_MACHINE(obj);
+    bmc->uart_default = val + 1;
+}
+
  static void aspeed_machine_class_props_init(ObjectClass *oc)
  {
      object_class_property_add_bool(oc, "execute-in-place",
@@ -1153,6 +1174,12 @@ static void aspeed_machine_class_props_init(ObjectClass *oc)
      object_class_property_set_description(oc, "execute-in-place",
                             "boot directly from CE0 flash device");
  
+    object_class_property_add_enum(oc, "uart-default", "UartDefault",
+                                   &UartDefault_lookup, NULL,
+                                   aspeed_set_uart_default);
+    object_class_property_set_description(oc, "uart-default",
+                           "Change the default UART of the board");
+
      object_class_property_add_str(oc, "fmc-model", aspeed_get_fmc_model,
                                     aspeed_set_fmc_model);
      object_class_property_set_description(oc, "fmc-model",
-- 
2.39.2



  reply	other threads:[~2023-03-17 17:40 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-03-17 13:13 Using QEMU how to redirect serial /dev/ttyS2 output of guest machine to host machine Abhishek Singh Dagur
2023-03-17 16:06 ` Alex Bennée
2023-03-17 17:24   ` Cédric Le Goater [this message]
2023-03-17 17:40     ` Abhishek Singh Dagur

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=b2c11d07-ee58-715d-2994-3add1c60b95d@kaod.org \
    --to=clg@kaod.org \
    --cc=abhishek@drut.io \
    --cc=alex.bennee@linaro.org \
    --cc=mark.cave-ayland@ilande.co.uk \
    --cc=openbmc@lists.ozlabs.org \
    --cc=peter.maydell@linaro.org \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).