Nouveau - Apr 2017 - [PATCH 02/38] Annotate hardware config module parameters in arch/x86/mm/

When the kernel is running in secure boot mode, we lock down the kernel to
prevent userspace from modifying the running kernel image.  Whilst this
includes prohibiting access to things like /dev/mem, it must also prevent
access by means of configuring driver modules in such a way as to cause a
device to access or modify the kernel image.

To this end, annotate module_param* statements that refer to hardware
configuration and indicate for future reference what type of parameter they
specify.  The parameter parser in the core sees this information and can
skip such parameters with an error message if the kernel is locked down.
The module initialisation then runs as normal, but just sees whatever the
default values for those parameters is.

Note that we do still need to do the module initialisation because some
drivers have viable defaults set in case parameters aren't specified and
some drivers support automatic configuration (e.g. PNP or PCI) in addition
to manually coded parameters.

This patch annotates drivers in arch/x86/mm/.

Suggested-by: Alan Cox <gnomes at lxorguk.ukuu.org.uk>
Signed-off-by: David Howells <dhowells at redhat.com>
cc: Steven Rostedt <rostedt at goodmis.org>
cc: Ingo Molnar <mingo at kernel.org>
cc: Thomas Gleixner <tglx at linutronix.de>
cc: "H. Peter Anvin" <hpa at zytor.com>
cc: x86 at kernel.org
cc: linux-kernel at vger.kernel.org
cc: nouveau at lists.freedesktop.org
---

 arch/x86/mm/testmmiotrace.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/x86/mm/testmmiotrace.c b/arch/x86/mm/testmmiotrace.c
index 38868adf07ea..f6ae6830b341 100644
--- a/arch/x86/mm/testmmiotrace.c
+++ b/arch/x86/mm/testmmiotrace.c
@@ -9,7 +9,7 @@
 #include <linux/mmiotrace.h>
 
 static unsigned long mmio_address;
-module_param(mmio_address, ulong, 0);
+module_param_hw(mmio_address, ulong, iomem, 0);
 MODULE_PARM_DESC(mmio_address, " Start address of the mapping of 16 kB
"
 				"(or 8 MB if read_far is non-zero).");

On Wed, 05 Apr 2017 17:57:08 +0100
David Howells <dhowells at redhat.com> wrote:
> When the kernel is running in secure boot mode, we lock down the kernel to
> prevent userspace from modifying the running kernel image.  Whilst this
> includes prohibiting access to things like /dev/mem, it must also prevent
> access by means of configuring driver modules in such a way as to cause a
> device to access or modify the kernel image.
> 
> To this end, annotate module_param* statements that refer to hardware
> configuration and indicate for future reference what type of parameter they
> specify.  The parameter parser in the core sees this information and can
> skip such parameters with an error message if the kernel is locked down.
> The module initialisation then runs as normal, but just sees whatever the
> default values for those parameters is.
> 
> Note that we do still need to do the module initialisation because some
> drivers have viable defaults set in case parameters aren't specified
and
> some drivers support automatic configuration (e.g. PNP or PCI) in addition
> to manually coded parameters.
> 
> This patch annotates drivers in arch/x86/mm/.
> 
> Suggested-by: Alan Cox <gnomes at lxorguk.ukuu.org.uk>
> Signed-off-by: David Howells <dhowells at redhat.com>
Acked-by: Steven Rostedt (VMware) <rostedt at goodmis.org>

-- Steve
> cc: Ingo Molnar <mingo at kernel.org>
> cc: Thomas Gleixner <tglx at linutronix.de>
> cc: "H. Peter Anvin" <hpa at zytor.com>
> cc: x86 at kernel.org
> cc: linux-kernel at vger.kernel.org
> cc: nouveau at lists.freedesktop.org
> ---
> 
>  arch/x86/mm/testmmiotrace.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/x86/mm/testmmiotrace.c b/arch/x86/mm/testmmiotrace.c
> index 38868adf07ea..f6ae6830b341 100644
> --- a/arch/x86/mm/testmmiotrace.c
> +++ b/arch/x86/mm/testmmiotrace.c
> @@ -9,7 +9,7 @@
>  #include <linux/mmiotrace.h>
>  
>  static unsigned long mmio_address;
> -module_param(mmio_address, ulong, 0);
> +module_param_hw(mmio_address, ulong, iomem, 0);
>  MODULE_PARM_DESC(mmio_address, " Start address of the mapping of 16
kB "
>  				"(or 8 MB if read_far is non-zero).");
>

On Wed, 5 Apr 2017, David Howells wrote:

The subject line hardly qualifies as a valid one.

    arch/subsys: Short description

Do I really have to explain that to you?
> When the kernel is running in secure boot mode, we lock down the kernel to
> prevent userspace from modifying the running kernel image.  Whilst this
> includes prohibiting access to things like /dev/mem, it must also prevent
> access by means of configuring driver modules in such a way as to cause a
> device to access or modify the kernel image.
> 
> To this end, annotate module_param* statements that refer to hardware
> configuration and indicate for future reference what type of parameter they
> specify.  The parameter parser in the core sees this information and can
> skip such parameters with an error message if the kernel is locked down.
> The module initialisation then runs as normal, but just sees whatever the
> default values for those parameters is.
> 
> Note that we do still need to do the module initialisation because some
> drivers have viable defaults set in case parameters aren't specified
and
> some drivers support automatic configuration (e.g. PNP or PCI) in addition
> to manually coded parameters.
> 
> This patch annotates drivers in arch/x86/mm/.
We already know that this is a patch. There is a chapter in
Documentation/process/SubmittingPatches which explains that. That file is
not only for newbies.
> diff --git a/arch/x86/mm/testmmiotrace.c b/arch/x86/mm/testmmiotrace.c
> index 38868adf07ea..f6ae6830b341 100644
> --- a/arch/x86/mm/testmmiotrace.c
> +++ b/arch/x86/mm/testmmiotrace.c
> @@ -9,7 +9,7 @@
>  #include <linux/mmiotrace.h>
>  
>  static unsigned long mmio_address;
> -module_param(mmio_address, ulong, 0);
> +module_param_hw(mmio_address, ulong, iomem, 0);
>  MODULE_PARM_DESC(mmio_address, " Start address of the mapping of 16
kB "
>  				"(or 8 MB if read_far is non-zero).");
The copied boilerplate above is really nonsensical here. The default
address is 0, so the init function will emit:

     pr_err("you have to use the module argument mmio_address.\n");
     pr_err("DO NOT LOAD THIS MODULE UNLESS YOU REALLY KNOW WHAT YOU ARE
DOING!\n");

Pretty useless when you can't supply a valid address.

       if (kernel_locked_down()) {
       		pr_info("This is not allowed because ...");
		return -EPERM;
       }

would make too much sense for the user, right?

Thanks,

	tglx

Thomas Gleixner <tglx at linutronix.de> wrote:
> > -module_param(mmio_address, ulong, 0);
> > +module_param_hw(mmio_address, ulong, iomem, 0);
> >  MODULE_PARM_DESC(mmio_address, " Start address of the mapping of
16 kB "
> >  				"(or 8 MB if read_far is non-zero).");
> 
> The copied boilerplate above is really nonsensical here. The default
> address is 0, so the init function will emit:
> 
>      pr_err("you have to use the module argument
mmio_address.\n");
>      pr_err("DO NOT LOAD THIS MODULE UNLESS YOU REALLY KNOW WHAT YOU
ARE DOING!\n");
> 
> Pretty useless when you can't supply a valid address.
> 
>        if (kernel_locked_down()) {
>        		pr_info("This is not allowed because ...");
> 		return -EPERM;
>        }
> 
> would make too much sense for the user, right?
In some drivers, this would be wrong - ipmi, for example - and we've already
been through this.  The hwparam series of patches annotates *all*
ioport/iomem/irq/dma specifiers unconditionally.  The hwparam series is the
way it is is because this has no overhead if it's not used - and also has
the
potentially useful side effect of making such parameters greppable.

It may well make sense to add your above suggestion also - but in the other
patch series.

David