[PATCH] tty: xilinx_uartps: move to arch_initcall for earlier console

[PATCH] tty: xilinx_uartps: move to arch_initcall for earlier console

[Shubhrajyoti Datta]

Signed-off-by: Shubhrajyoti Datta <shubhrajyoti.datta@xilinx.com>
---
 drivers/tty/serial/xilinx_uartps.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/tty/serial/xilinx_uartps.c b/drivers/tty/serial/xilinx_uartps.c
index ff1b115..a239343 100644
--- a/drivers/tty/serial/xilinx_uartps.c
+++ b/drivers/tty/serial/xilinx_uartps.c
@@ -1647,7 +1647,7 @@ static void __exit cdns_uart_exit(void)
 	uart_unregister_driver(&cdns_uart_uart_driver);
 }
 
-module_init(cdns_uart_init);
+arch_initcall(cdns_uart_init);
 module_exit(cdns_uart_exit);
 
 MODULE_DESCRIPTION("Driver for Cadence UART");
-- 
2.1.1

Re: [PATCH] tty: xilinx_uartps: move to arch_initcall for earlier console

[Greg KH]

On Wed, Aug 23, 2017 at 04:50:21PM +0530, Shubhrajyoti Datta wrote:
> Signed-off-by: Shubhrajyoti Datta <shubhrajyoti.datta@xilinx.com>

I can't take patches without any changelog text at all :(

Re: [PATCH] tty: xilinx_uartps: move to arch_initcall for earlier console

[Michal Simek]

Hi, +Linus

On 24.8.2017 01:24, Greg KH wrote:
> On Wed, Aug 23, 2017 at 04:50:21PM +0530, Shubhrajyoti Datta wrote:
>> Signed-off-by: Shubhrajyoti Datta <shubhrajyoti.datta@xilinx.com>
> 
> I can't take patches without any changelog text at all :(

ok. I see you have also commented this.

Anyway this is kind of old discussion about moving serial drivers to
arch_initcall from module_init.

There is one patch in the tree.

commit 4dd9e742df98f8f600b4302d3adbb087a68237f7
Author:     Alessandro Rubini <rubini@gnudd.com>
AuthorDate: Tue May 5 05:54:13 2009 +0100
Commit:     Russell King <rmk+kernel@arm.linux.org.uk>
CommitDate: Sun May 31 14:58:11 2009 +0100

    [ARM] 5505/1: serial amba-pl011: move to arch_initcall for earlier
console

    Signed-off-by: Alessandro Rubini <rubini@unipv.it>"
    Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>


and then there was one patch (also sent to linux-serial but don't have link)
https://patches.linaro.org/patch/14633/

where that discussion wasn't finished.


There is one more patch which does that without real description for
this change.
commit ce87122911f8db59d3c2bc355c694c7a38940804
Author:     Vladimir Murzin <vladimir.murzin@arm.com>
AuthorDate: Tue Jun 7 16:02:37 2016 +0100
Commit:     Greg Kroah-Hartman <gregkh@linuxfoundation.org>
CommitDate: Sat Jun 25 14:01:57 2016 -0700

    serial: mps2-uart: make driver explicitly non-modular


Most of drivers are using module_init and only some of them arch_initcall.

Thanks,
Michal

Re: [PATCH] tty: xilinx_uartps: move to arch_initcall for earlier console

[Linus Walleij]

On Thu, Aug 24, 2017 at 11:32 AM, Michal Simek <michal.simek@xilinx.com> wrote:
> On 24.8.2017 01:24, Greg KH wrote:
>> On Wed, Aug 23, 2017 at 04:50:21PM +0530, Shubhrajyoti Datta wrote:
>>> Signed-off-by: Shubhrajyoti Datta <shubhrajyoti.datta@xilinx.com>
>>
>> I can't take patches without any changelog text at all :(
>
> ok. I see you have also commented this.
>
> Anyway this is kind of old discussion about moving serial drivers to
> arch_initcall from module_init.
>
> There is one patch in the tree.
>
> commit 4dd9e742df98f8f600b4302d3adbb087a68237f7
> Author:     Alessandro Rubini <rubini@gnudd.com>
> AuthorDate: Tue May 5 05:54:13 2009 +0100
> Commit:     Russell King <rmk+kernel@arm.linux.org.uk>
> CommitDate: Sun May 31 14:58:11 2009 +0100
>
>     [ARM] 5505/1: serial amba-pl011: move to arch_initcall for earlier
> console
>
>     Signed-off-by: Alessandro Rubini <rubini@unipv.it>"
>     Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
>
>
> and then there was one patch (also sent to linux-serial but don't have link)
> https://patches.linaro.org/patch/14633/
>
> where that discussion wasn't finished.

Yeah Alessandro never came back on that.

I *guess* it is to get the console up really early, simply becaus it is
quite helpful for spotting early boot problems.

It is not for earlydebug, because for that we use another routine
that hammers out the characters on the console. That is what we use
before even going to console, with printascii(), I usually use a patch
like this:

--- a/init/main.c
+++ b/init/main.c
@@ -515,6 +515,12 @@ asmlinkage __visible void __init start_kernel(void)
        smp_setup_processor_id();
        debug_objects_early_init();

+#ifdef CONFIG_ARM
+       {
+         extern void printascii(char *);
+         printascii("start_kernel\n");
+       }
+#endif
        /*
         * Set up the initial canary ASAP:
         */
diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c
index fc47863f629c..85edd1401406 100644
--- a/kernel/printk/printk.c
+++ b/kernel/printk/printk.c
@@ -1726,6 +1726,13 @@ asmlinkage int vprintk_emit(int facility, int level,
                lflags |= LOG_NEWLINE;
        }

+#ifdef CONFIG_ARM
+       {
+               extern void printascii(char *);
+               printascii(textbuf);
+       }
+#endif
+
        /* strip kernel syslog prefix and extract log level or control flags */
        if (facility == 0) {
                int kern_level;


Yours,
Linus Walleij

Re: [PATCH] tty: xilinx_uartps: move to arch_initcall for earlier console

[Michal Simek]

On 31.8.2017 14:59, Linus Walleij wrote:
> On Thu, Aug 24, 2017 at 11:32 AM, Michal Simek <michal.simek@xilinx.com> wrote:
>> On 24.8.2017 01:24, Greg KH wrote:
>>> On Wed, Aug 23, 2017 at 04:50:21PM +0530, Shubhrajyoti Datta wrote:
>>>> Signed-off-by: Shubhrajyoti Datta <shubhrajyoti.datta@xilinx.com>
>>>
>>> I can't take patches without any changelog text at all :(
>>
>> ok. I see you have also commented this.
>>
>> Anyway this is kind of old discussion about moving serial drivers to
>> arch_initcall from module_init.
>>
>> There is one patch in the tree.
>>
>> commit 4dd9e742df98f8f600b4302d3adbb087a68237f7
>> Author:     Alessandro Rubini <rubini@gnudd.com>
>> AuthorDate: Tue May 5 05:54:13 2009 +0100
>> Commit:     Russell King <rmk+kernel@arm.linux.org.uk>
>> CommitDate: Sun May 31 14:58:11 2009 +0100
>>
>>     [ARM] 5505/1: serial amba-pl011: move to arch_initcall for earlier
>> console
>>
>>     Signed-off-by: Alessandro Rubini <rubini@unipv.it>"
>>     Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
>>
>>
>> and then there was one patch (also sent to linux-serial but don't have link)
>> https://patches.linaro.org/patch/14633/
>>
>> where that discussion wasn't finished.
> 
> Yeah Alessandro never came back on that.
> 
> I *guess* it is to get the console up really early, simply becaus it is
> quite helpful for spotting early boot problems.


ok. It means no concern about the move to arch_initcall.
Shubhrajyoti: Please send v2 with changelog.

Thanks,
Michal

Re: [PATCH] tty: xilinx_uartps: move to arch_initcall for earlier console

[Shubhrajyoti Datta]

...
>>
>> I *guess* it is to get the console up really early, simply becaus it is
>> quite helpful for spotting early boot problems.
>
>
> ok. It means no concern about the move to arch_initcall.
> Shubhrajyoti: Please send v2 with changelog.

Will do thanks for the review.

>
> Thanks,
> Michal

Re: [PATCH] tty: xilinx_uartps: move to arch_initcall for earlier console

[Michal Simek]

On 23.8.2017 12:55, Shubhrajyoti Datta wrote:
> Signed-off-by: Shubhrajyoti Datta <shubhrajyoti.datta@xilinx.com>

Empty commit message?
What's the reason for this change?

M

> ---
>  drivers/tty/serial/xilinx_uartps.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/tty/serial/xilinx_uartps.c b/drivers/tty/serial/xilinx_uartps.c
> index ff1b115..a239343 100644
> --- a/drivers/tty/serial/xilinx_uartps.c
> +++ b/drivers/tty/serial/xilinx_uartps.c
> @@ -1647,7 +1647,7 @@ static void __exit cdns_uart_exit(void)
>  	uart_unregister_driver(&cdns_uart_uart_driver);
>  }
>  
> -module_init(cdns_uart_init);
> +arch_initcall(cdns_uart_init);
>  module_exit(cdns_uart_exit);
>  
>  MODULE_DESCRIPTION("Driver for Cadence UART");
>