On Mon, 29 Nov 2021 at 10:52, Krzysztof Kozlowski
[off-list ref] wrote:

On 27/11/2021 23:32, Sam Protsenko wrote:

quoted

Enable serial driver to be built as a module. To do so, init the console
support on driver/module load instead of using console_initcall().

This is needed for proper support of USIv2 driver (which can be built as
a module, which in turn makes SERIAL_SAMSUNG be a module too). It also
might be useful for Android GKI modularization efforts.

Inspired by commit 87a0b9f98ac5 ("tty: serial: meson: enable console as
module").

Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
---
 drivers/tty/serial/Kconfig       |  2 +-
 drivers/tty/serial/samsung_tty.c | 21 +++++++++++++++++++--
 2 files changed, 20 insertions(+), 3 deletions(-)

diff --git a/drivers/tty/serial/Kconfig b/drivers/tty/serial/Kconfig
index fc543ac97c13..0e5ccb25bdb1 100644
--- a/drivers/tty/serial/Kconfig
+++ b/drivers/tty/serial/Kconfig

@@ -263,7 +263,7 @@ config SERIAL_SAMSUNG_UARTS

 config SERIAL_SAMSUNG_CONSOLE
      bool "Support for console on Samsung SoC serial port"
-     depends on SERIAL_SAMSUNG=y
+     depends on SERIAL_SAMSUNG
      select SERIAL_CORE_CONSOLE
      select SERIAL_EARLYCON
      help

diff --git a/drivers/tty/serial/samsung_tty.c b/drivers/tty/serial/samsung_tty.c
index f986a9253dc8..92a63e9392ed 100644
--- a/drivers/tty/serial/samsung_tty.c
+++ b/drivers/tty/serial/samsung_tty.c

@@ -1720,10 +1720,10 @@ static int __init s3c24xx_serial_console_init(void)
      register_console(&s3c24xx_serial_console);
      return 0;
 }
-console_initcall(s3c24xx_serial_console_init);

 #define S3C24XX_SERIAL_CONSOLE &s3c24xx_serial_console
 #else
+static inline int s3c24xx_serial_console_init(void) { return 0; }
 #define S3C24XX_SERIAL_CONSOLE NULL
 #endif

@@ -2898,7 +2898,24 @@ static struct platform_driver samsung_serial_driver = {
      },
 };

-module_platform_driver(samsung_serial_driver);
+static int __init samsung_serial_init(void)
+{
+     int ret;
+
+     ret = s3c24xx_serial_console_init();
+     if (ret)
+             return ret;

This will trigger warns on module re-loading, won't it? Either suppress
unbind or cleanup in module exit.

I guess that's already taken care of in  samsung_serial_remove(): it's
doing uart_remove_one_port(), which in turn does unregister_console().
So I don't think anything extra should be done on module exit. Or I'm
missing something?

That case (unload/load) actually doesn't work well in my case: serial
console doesn't work after doing "modprobe -r samsung_tty; modprobe
samsung_tty" (but it works fine e.g. in case of i2c_exynos5 driver).
Not sure what is wrong, but I can see that my board keeps running
(heartbeat LED is still blinking). Not even sure if that use case
(unload/load) was ever functional before.

Anyway, please let me know if you think something should be done about
this particular patch. Right now I don't see anything missing.

quoted

+
+     return platform_driver_register(&samsung_serial_driver);
+}
+
+static void __exit samsung_serial_exit(void)
+{
+     platform_driver_unregister(&samsung_serial_driver);
+}
+
+module_init(samsung_serial_init);
+module_exit(samsung_serial_exit);

 #ifdef CONFIG_SERIAL_SAMSUNG_CONSOLE
 /*

Best regards,
Krzysztof