[2.6 patch] unexport is_console_locked

[2.6 patch] unexport is_console_locked

[Adrian Bunk]

I didn't find any possible modular usage in the kernel.

Signed-off-by: Adrian Bunk <bunk@stusta.de>

--- linux-2.6.11-rc5-mm1-full/kernel/printk.c.old	2005-03-04 00:58:16.000000000 +0100
+++ linux-2.6.11-rc5-mm1-full/kernel/printk.c	2005-03-04 00:58:22.000000000 +0100
@@ -675,7 +675,6 @@
 {
 	return console_locked;
 }
-EXPORT_SYMBOL(is_console_locked);
 
 /**
  * release_console_sem - unlock the console system

[2.6 patch] unexport is_console_locked

[Adrian Bunk]

I didn't find any possible modular usage in the kernel.

Signed-off-by: Adrian Bunk <bunk@stusta.de>

---

This patch was already sent on:
- 4 Mar 2005

--- linux-2.6.11-rc5-mm1-full/kernel/printk.c.old	2005-03-04 00:58:16.000000000 +0100
+++ linux-2.6.11-rc5-mm1-full/kernel/printk.c	2005-03-04 00:58:22.000000000 +0100
@@ -675,7 +675,6 @@
 {
 	return console_locked;
 }
-EXPORT_SYMBOL(is_console_locked);
 
 /**
  * release_console_sem - unlock the console system

Re: [2.6 patch] unexport is_console_locked

[Andrew Morton]

Adrian Bunk <bunk@stusta.de> wrote:
>
> I didn't find any possible modular usage in the kernel.
> 

Is needed for WARN_CONSOLE_UNLOCKED().

It just happens that WARN_CONSOLE_UNLOCKED() is currently only using in
vt.c, which just happens to not support modular usage.

But disabling this export would make WARN_CONSOLE_UNLOCKED() pretty useless
for driver development.

So I'd suggest that we either leave this symbol exported for driver
developers or we nuke WARN_CONSOLE_UNLOCKED() altogether.

Ben, do you still want it?


> --- linux-2.6.11-rc5-mm1-full/kernel/printk.c.old	2005-03-04 00:58:16.000000000 +0100
> +++ linux-2.6.11-rc5-mm1-full/kernel/printk.c	2005-03-04 00:58:22.000000000 +0100
> @@ -675,7 +675,6 @@
>  {
>  	return console_locked;
>  }
> -EXPORT_SYMBOL(is_console_locked);
>  
>  /**
>   * release_console_sem - unlock the console system

Re: [2.6 patch] unexport is_console_locked

[Benjamin Herrenschmidt]

On Sat, 2005-04-23 at 16:49 -0700, Andrew Morton wrote:
> Adrian Bunk <bunk@stusta.de> wrote:
> >
> > I didn't find any possible modular usage in the kernel.
> > 
> 
> Is needed for WARN_CONSOLE_UNLOCKED().
> 
> It just happens that WARN_CONSOLE_UNLOCKED() is currently only using in
> vt.c, which just happens to not support modular usage.
> 
> But disabling this export would make WARN_CONSOLE_UNLOCKED() pretty useless
> for driver development.
> 
> So I'd suggest that we either leave this symbol exported for driver
> developers or we nuke WARN_CONSOLE_UNLOCKED() altogether.
> 
> Ben, do you still want it?

Yes, I do. Please, keep it exported. On that note, I don't understand
that frenzy of unexporting everything you can, I find this useless
burden in general.

Ben.

Re: [2.6 patch] unexport is_console_locked

[Christoph Hellwig]

On Sun, Apr 24, 2005 at 10:51:30AM +1000, Benjamin Herrenschmidt wrote:
> Yes, I do. Please, keep it exported. On that note, I don't understand
> that frenzy of unexporting everything you can, I find this useless
> burden in general.

If we ever want to a have a semi-stable driver API (1) we need to stop
exporting random symbols and move towards APIs that make sense and
operate at the right level instead of poking into implementation details.
Note that even without that goal only having the right primitives avaiable
to drivers makes sure they use the right tool instead of copy & pasting
random crap.

(1) that is changing only when we have a real need instead of all the time.