Make CONFIG_DEBUG_RODATA boot-time configurable

Make CONFIG_DEBUG_RODATA boot-time configurable

[Dave Jones]

We had a request to disable DEBUG_RODATA in our kernel for debugging purposes.  
Turning it off completely seems a bit unfortunate for the minority case
of people wanting to do debugging, so I came up with this patch to make
it a boot-time 'disable_rodata' argument.

Signed-off-by: Dave Jones <davej@redhat.com>


diff --git a/init/main.c b/init/main.c
index a42fdf4..eb60fae 100644
--- a/init/main.c
+++ b/init/main.c
@@ -820,6 +820,15 @@ static void run_init_process(char *init_filename)
 	kernel_execve(init_filename, argv_init, envp_init);
 }
 
+static unsigned int disable_rodata;
+static int __init set_disable_rodata(char *str)
+{
+	disable_rodata = 1;
+	return 1;
+}
+__setup("disable_rodata", set_disable_rodata);
+
+
 /* This is a non __init function. Force it to be noinline otherwise gcc
  * makes it inline to init() and it becomes part of init.text section
  */
@@ -830,7 +839,8 @@ static noinline int init_post(void)
 	async_synchronize_full();
 	free_initmem();
 	unlock_kernel();
-	mark_rodata_ro();
+	if (disable_rodata == 0)
+		mark_rodata_ro();
 	system_state = SYSTEM_RUNNING;
 	numa_default_policy();
diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt
index 4ddb58d..1320d80 100644
--- a/Documentation/kernel-parameters.txt
+++ b/Documentation/kernel-parameters.txt
@@ -651,6 +651,10 @@ and is between 256 and 4096 characters. It is defined in the file
 			MTRR settings.  This parameter disables that behavior,
 			possibly causing your machine to run very slowly.
 
+	disable_rodata [X86]
+			Don't mark .rodata section as read-only.
+			May be useful for debugging.
+
 	disable_timer_pin_1 [X86]
 			Disable PIN 1 of APIC timer
 			Can be useful to work around chipset bugs.

Re: Make CONFIG_DEBUG_RODATA boot-time configurable

[Randy Dunlap]

On Thu, 15 Jul 2010 17:09:44 -0400 Dave Jones wrote:

> We had a request to disable DEBUG_RODATA in our kernel for debugging purposes.  
> Turning it off completely seems a bit unfortunate for the minority case
> of people wanting to do debugging, so I came up with this patch to make
> it a boot-time 'disable_rodata' argument.
> 
> Signed-off-by: Dave Jones <davej@redhat.com>
> 
> 
> diff --git a/init/main.c b/init/main.c
> index a42fdf4..eb60fae 100644
> --- a/init/main.c
> +++ b/init/main.c
> @@ -820,6 +820,15 @@ static void run_init_process(char *init_filename)
>  	kernel_execve(init_filename, argv_init, envp_init);
>  }
>  
> +static unsigned int disable_rodata;
> +static int __init set_disable_rodata(char *str)
> +{
> +	disable_rodata = 1;
> +	return 1;
> +}
> +__setup("disable_rodata", set_disable_rodata);
> +
> +
>  /* This is a non __init function. Force it to be noinline otherwise gcc
>   * makes it inline to init() and it becomes part of init.text section
>   */
> @@ -830,7 +839,8 @@ static noinline int init_post(void)
>  	async_synchronize_full();
>  	free_initmem();
>  	unlock_kernel();
> -	mark_rodata_ro();
> +	if (disable_rodata == 0)
> +		mark_rodata_ro();
>  	system_state = SYSTEM_RUNNING;
>  	numa_default_policy();
> diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt
> index 4ddb58d..1320d80 100644
> --- a/Documentation/kernel-parameters.txt
> +++ b/Documentation/kernel-parameters.txt
> @@ -651,6 +651,10 @@ and is between 256 and 4096 characters. It is defined in the file
>  			MTRR settings.  This parameter disables that behavior,
>  			possibly causing your machine to run very slowly.
>  
> +	disable_rodata [X86]

What is X86-specific about this boot option?

> +			Don't mark .rodata section as read-only.
> +			May be useful for debugging.
> +
>  	disable_timer_pin_1 [X86]
>  			Disable PIN 1 of APIC timer
>  			Can be useful to work around chipset bugs.
> --


---
~Randy
*** Remember to use Documentation/SubmitChecklist when testing your code ***

Re: Make CONFIG_DEBUG_RODATA boot-time configurable

[Dave Jones]

On Thu, Jul 15, 2010 at 02:35:25PM -0700, Randy Dunlap wrote:
 
 > > diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt
 > > index 4ddb58d..1320d80 100644
 > > --- a/Documentation/kernel-parameters.txt
 > > +++ b/Documentation/kernel-parameters.txt
 > > @@ -651,6 +651,10 @@ and is between 256 and 4096 characters. It is defined in the file
 > >  			MTRR settings.  This parameter disables that behavior,
 > >  			possibly causing your machine to run very slowly.
 > >  
 > > +	disable_rodata [X86]
 > 
 > What is X86-specific about this boot option?

Just hadn't realized that other archs had implemented it.

	Dave

Re: Make CONFIG_DEBUG_RODATA boot-time configurable

[Thomas Gleixner]

On Thu, 15 Jul 2010, Dave Jones wrote:

> We had a request to disable DEBUG_RODATA in our kernel for debugging purposes.  
> Turning it off completely seems a bit unfortunate for the minority case
> of people wanting to do debugging, so I came up with this patch to make
> it a boot-time 'disable_rodata' argument.
> 
> Signed-off-by: Dave Jones <davej@redhat.com>
 
Nice one, shouldn't we get rid of the DEBUG_RODATA config and enable
it unconditionally ?

Thanks,

	tglx

Re: Make CONFIG_DEBUG_RODATA boot-time configurable

[Dave Jones]

On Fri, Jul 16, 2010 at 12:06:16AM +0200, Thomas Gleixner wrote:
 > On Thu, 15 Jul 2010, Dave Jones wrote:
 > 
 > > We had a request to disable DEBUG_RODATA in our kernel for debugging purposes.  
 > > Turning it off completely seems a bit unfortunate for the minority case
 > > of people wanting to do debugging, so I came up with this patch to make
 > > it a boot-time 'disable_rodata' argument.
 > > 
 > > Signed-off-by: Dave Jones <davej@redhat.com>
 >  
 > Nice one, shouldn't we get rid of the DEBUG_RODATA config and enable
 > it unconditionally ?

I personally have no objection, but others might.
follow-up patch rather than including it in this maybe?

	Dave

Re: Make CONFIG_DEBUG_RODATA boot-time configurable

[Arjan van de Ven]

On Fri, 16 Jul 2010 00:06:16 +0200 (CEST)
Thomas Gleixner <tglx@linutronix.de> wrote:

> On Thu, 15 Jul 2010, Dave Jones wrote:
> 
> > We had a request to disable DEBUG_RODATA in our kernel for
> > debugging purposes. Turning it off completely seems a bit
> > unfortunate for the minority case of people wanting to do
> > debugging, so I came up with this patch to make it a boot-time
> > 'disable_rodata' argument.
> > 
> > Signed-off-by: Dave Jones <davej@redhat.com>
>  
> Nice one, shouldn't we get rid of the DEBUG_RODATA config and enable
> it unconditionally ?
> 

I'd go one step further; enable unconditional, and have the config
option control if you can turn it off or not...

Re: Make CONFIG_DEBUG_RODATA boot-time configurable

[Arjan van de Ven]

On Thu, 15 Jul 2010 17:09:44 -0400
Dave Jones <davej@redhat.com> wrote:

> We had a request to disable DEBUG_RODATA in our kernel for debugging
> purposes. Turning it off completely seems a bit unfortunate for the
> minority case of people wanting to do debugging, so I came up with
> this patch to make it a boot-time 'disable_rodata' argument.


I'd be very curious what kind of debugging this is...
is this the rootkit type of debugging?


-- 
Arjan van de Ven 	Intel Open Source Technology Centre
For development, discussion and tips for power savings, 
visit http://www.lesswatts.org