[RFC: 2.6 patch] kernel/audit.c: remove unused exports

[RFC: 2.6 patch] kernel/audit.c: remove unused exports

[Adrian Bunk]

This patch removes the following unused EXPORT_SYMBOL's:
- audit_log_start
- audit_log_end
- audit_log_format
- audit_log

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

--- linux-2.6.17-rc1-mm3-full/kernel/audit.c.old	2006-04-20 22:38:17.000000000 +0200
+++ linux-2.6.17-rc1-mm3-full/kernel/audit.c	2006-04-20 22:40:03.000000000 +0200
@@ -1092,7 +1092,3 @@
 	}
 }
 
-EXPORT_SYMBOL(audit_log_start);
-EXPORT_SYMBOL(audit_log_end);
-EXPORT_SYMBOL(audit_log_format);
-EXPORT_SYMBOL(audit_log);

Re: [RFC: 2.6 patch] kernel/audit.c: remove unused exports

[Tony Jones]

On Thu, Apr 20, 2006 at 11:05:21PM +0200, Adrian Bunk wrote:
> This patch removes the following unused EXPORT_SYMBOL's:
> - audit_log_start
> - audit_log_end
> - audit_log_format
> - audit_log
> 
> Signed-off-by: Adrian Bunk <bunk@stusta.de>
> 
> --- linux-2.6.17-rc1-mm3-full/kernel/audit.c.old	2006-04-20 22:38:17.000000000 +0200
> +++ linux-2.6.17-rc1-mm3-full/kernel/audit.c	2006-04-20 22:40:03.000000000 +0200
> @@ -1092,7 +1092,3 @@
>  	}
>  }
>  
> -EXPORT_SYMBOL(audit_log_start);
> -EXPORT_SYMBOL(audit_log_end);
> -EXPORT_SYMBOL(audit_log_format);
> -EXPORT_SYMBOL(audit_log);

It would seem useful to allow out of tree modules (even if they never have
a goal of submitting to become intree) to log audit data. No?  A reason to
prevent it seems utterly lost on me.

Tony

Re: [RFC: 2.6 patch] kernel/audit.c: remove unused exports

[Steve Grubb]

On Thursday 20 April 2006 17:05, Adrian Bunk wrote:
> This patch removes the following unused EXPORT_SYMBOL's:
> - audit_log_start
> - audit_log_end
> - audit_log_format
> - audit_log

These are to allow 3rd party security modules log to the audit system.

-Steve

Re: [RFC: 2.6 patch] kernel/audit.c: remove unused exports

[David Woodhouse]

On Thu, 2006-04-20 at 18:44 -0400, Steve Grubb wrote:
> These are to allow 3rd party security modules log to the audit system.

They should be EXPORT_SYMBOL_GPL though, at least.

-- 
dwmw2

Re: [RFC: 2.6 patch] kernel/audit.c: remove unused exports

[Adrian Bunk]

On Thu, Apr 20, 2006 at 02:25:19PM -0700, Tony Jones wrote:
> On Thu, Apr 20, 2006 at 11:05:21PM +0200, Adrian Bunk wrote:
> > This patch removes the following unused EXPORT_SYMBOL's:
> > - audit_log_start
> > - audit_log_end
> > - audit_log_format
> > - audit_log
> > 
> > Signed-off-by: Adrian Bunk <bunk@stusta.de>
> > 
> > --- linux-2.6.17-rc1-mm3-full/kernel/audit.c.old	2006-04-20 22:38:17.000000000 +0200
> > +++ linux-2.6.17-rc1-mm3-full/kernel/audit.c	2006-04-20 22:40:03.000000000 +0200
> > @@ -1092,7 +1092,3 @@
> >  	}
> >  }
> >  
> > -EXPORT_SYMBOL(audit_log_start);
> > -EXPORT_SYMBOL(audit_log_end);
> > -EXPORT_SYMBOL(audit_log_format);
> > -EXPORT_SYMBOL(audit_log);
> 
> It would seem useful to allow out of tree modules (even if they never have
> a goal of submitting to become intree) to log audit data. No?  A reason to
> prevent it seems utterly lost on me.

Out of tree modules without the goal of being submitted are not a good 
idea.

AppArmor/LSM is a good example what can happen if people don't submit 
their external modules for inclusion in the kernel.

> Tony

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: [RFC: 2.6 patch] kernel/audit.c: remove unused exports

[Steve Grubb]

On Friday 21 April 2006 05:27, Adrian Bunk wrote:
> Out of tree modules without the goal of being submitted are not a good
> idea.

And what are people to do while developing modules that are not ready for 
inclusion? This is a generic API for use by anyone. If virus scanning or IDS  
companies need to use the API, shouldn't they be allowed to?

-Steve

Re: [RFC: 2.6 patch] kernel/audit.c: remove unused exports

[David Woodhouse]

On Fri, 2006-04-21 at 09:10 -0400, Steve Grubb wrote:
> And what are people to do while developing modules that are not ready for 
> inclusion? This is a generic API for use by anyone. If virus scanning or IDS  
> companies need to use the API, shouldn't they be allowed to?

Yes, but only if their code is GPL'd.

-- 
dwmw2

Re: [RFC: 2.6 patch] kernel/audit.c: remove unused exports

[Tony Jones]

On Fri, Apr 21, 2006 at 11:27:55AM +0200, Adrian Bunk wrote:
> > > -EXPORT_SYMBOL(audit_log_start);
> > > -EXPORT_SYMBOL(audit_log_end);
> > > -EXPORT_SYMBOL(audit_log_format);
> > > -EXPORT_SYMBOL(audit_log);
> > 
> > It would seem useful to allow out of tree modules (even if they never have
> > a goal of submitting to become intree) to log audit data. No?  A reason to
> > prevent it seems utterly lost on me.
> 
> Out of tree modules without the goal of being submitted are not a good 
> idea.

What a stupid comment (in general and expecially given the nature of the audit 
subsystem ). Sorry. 

Also, there are probably at present no intree users of audit other than 
compiled built-in. I'm guessing by your patch you believe built-in to be the 
only valid configuration mode for said users :-)

EXPORT_SYMBOL_GPL, sure.

> AppArmor/LSM is a good example what can happen if people don't submit 
> their external modules for inclusion in the kernel.

LOL. Thanks for the daily laugh.

Tony

Re: [RFC: 2.6 patch] kernel/audit.c: remove unused exports

[Adrian Bunk]

On Fri, Apr 21, 2006 at 10:24:10AM -0700, Tony Jones wrote:
> On Fri, Apr 21, 2006 at 11:27:55AM +0200, Adrian Bunk wrote:
> > > > -EXPORT_SYMBOL(audit_log_start);
> > > > -EXPORT_SYMBOL(audit_log_end);
> > > > -EXPORT_SYMBOL(audit_log_format);
> > > > -EXPORT_SYMBOL(audit_log);
> > > 
> > > It would seem useful to allow out of tree modules (even if they never have
> > > a goal of submitting to become intree) to log audit data. No?  A reason to
> > > prevent it seems utterly lost on me.
> > 
> > Out of tree modules without the goal of being submitted are not a good 
> > idea.
> 
> What a stupid comment (in general and expecially given the nature of the audit 
> subsystem ). Sorry. 
> 
> Also, there are probably at present no intree users of audit other than 
> compiled built-in. I'm guessing by your patch you believe built-in to be the 
> only valid configuration mode for said users :-)
> 
> EXPORT_SYMBOL_GPL, sure.
>...

I don't have a fixed opinion about whether future in-kernel users of the 
audit subsystem could be modules.

Currently there are none, and if modular users get added at some point 
in the future re-exporting them will be trivial.

Until then, removing the exports simply makes the kernel image a bit 
smaller.

> Tony

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed