[PATCH] docs: add documentation on printing kernel addresses

[PATCH] docs: add documentation on printing kernel addresses

[Tobin C. Harding]

Hashing addresses printed with printk specifier %p was implemented
recently. During development a number of issues were raised regarding
leaking kernel addresses to userspace. We should update the
documentation appropriately.

Add documentation regarding printing kernel addresses.

Signed-off-by: Tobin C. Harding <me@tobin.cc>
---

Is there a proffered method for subscripts in sphinx kernel docs? Here
we use '[*]' 

thanks,
Tobin.

 Documentation/security/self-protection.rst | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/Documentation/security/self-protection.rst b/Documentation/security/self-protection.rst
index 60c8bd8b77bf..e711280cfdd7 100644
--- a/Documentation/security/self-protection.rst
+++ b/Documentation/security/self-protection.rst
@@ -270,6 +270,20 @@ attacks, it is important to defend against exposure of both kernel memory
 addresses and kernel memory contents (since they may contain kernel
 addresses or other sensitive things like canary values).
 
+Kernel addresses
+----------------
+
+Printing kernel addresses to userspace leaks sensitive information about
+the kernel memory layout. Care should be exercised when using any printk
+specifier that prints the raw address, currently %px, %p[ad], (and %p[sSb]
+in certain circumstances [*]).  Any file written to using one of these
+specifiers should be readable only by privileged processes.
+
+Kernels 4.14 and older printed the raw address using %p. As of 4.15-rc1
+addresses printed with the specifier %p are hashed before printing.
+
+[*] If symbol lookup fails, the raw address is currently printed.
+
 Unique identifiers
 ------------------
 
-- 
2.7.4

Re: [PATCH] docs: add documentation on printing kernel addresses

[Kees Cook]

On Wed, Dec 6, 2017 at 4:26 PM, Tobin C. Harding <me@tobin.cc> wrote:
> Hashing addresses printed with printk specifier %p was implemented
> recently. During development a number of issues were raised regarding
> leaking kernel addresses to userspace. We should update the
> documentation appropriately.
>
> Add documentation regarding printing kernel addresses.
>
> Signed-off-by: Tobin C. Harding <me@tobin.cc>

Acked-by: Kees Cook <keescook@chromium.org>

> ---
>
> Is there a proffered method for subscripts in sphinx kernel docs? Here
> we use '[*]'

Great question... I can't find an answer to this. :P

>
> thanks,
> Tobin.
>
>  Documentation/security/self-protection.rst | 14 ++++++++++++++
>  1 file changed, 14 insertions(+)
>
> diff --git a/Documentation/security/self-protection.rst b/Documentation/security/self-protection.rst
> index 60c8bd8b77bf..e711280cfdd7 100644
> --- a/Documentation/security/self-protection.rst
> +++ b/Documentation/security/self-protection.rst
> @@ -270,6 +270,20 @@ attacks, it is important to defend against exposure of both kernel memory
>  addresses and kernel memory contents (since they may contain kernel
>  addresses or other sensitive things like canary values).
>
> +Kernel addresses
> +----------------
> +
> +Printing kernel addresses to userspace leaks sensitive information about
> +the kernel memory layout. Care should be exercised when using any printk
> +specifier that prints the raw address, currently %px, %p[ad], (and %p[sSb]
> +in certain circumstances [*]).  Any file written to using one of these
> +specifiers should be readable only by privileged processes.
> +
> +Kernels 4.14 and older printed the raw address using %p. As of 4.15-rc1
> +addresses printed with the specifier %p are hashed before printing.
> +
> +[*] If symbol lookup fails, the raw address is currently printed.

Is there a plan to adjust this case?

Thanks!

-Kees

> +
>  Unique identifiers
>  ------------------
>
> --
> 2.7.4
>



-- 
Kees Cook
Pixel Security

Re: [PATCH] docs: add documentation on printing kernel addresses

[Tobin C. Harding]

On Wed, Dec 06, 2017 at 04:43:40PM -0800, Kees Cook wrote:
> On Wed, Dec 6, 2017 at 4:26 PM, Tobin C. Harding <me@tobin.cc> wrote:
> > Hashing addresses printed with printk specifier %p was implemented
> > recently. During development a number of issues were raised regarding
> > leaking kernel addresses to userspace. We should update the
> > documentation appropriately.
> >
> > Add documentation regarding printing kernel addresses.
> >
> > Signed-off-by: Tobin C. Harding <me@tobin.cc>
> 
> Acked-by: Kees Cook <keescook@chromium.org>
> 
> > ---
> >
> > Is there a proffered method for subscripts in sphinx kernel docs? Here
> > we use '[*]'
> 
> Great question... I can't find an answer to this. :P
> 
> >
> > thanks,
> > Tobin.
> >
> >  Documentation/security/self-protection.rst | 14 ++++++++++++++
> >  1 file changed, 14 insertions(+)
> >
> > diff --git a/Documentation/security/self-protection.rst b/Documentation/security/self-protection.rst
> > index 60c8bd8b77bf..e711280cfdd7 100644
> > --- a/Documentation/security/self-protection.rst
> > +++ b/Documentation/security/self-protection.rst
> > @@ -270,6 +270,20 @@ attacks, it is important to defend against exposure of both kernel memory
> >  addresses and kernel memory contents (since they may contain kernel
> >  addresses or other sensitive things like canary values).
> >
> > +Kernel addresses
> > +----------------
> > +
> > +Printing kernel addresses to userspace leaks sensitive information about
> > +the kernel memory layout. Care should be exercised when using any printk
> > +specifier that prints the raw address, currently %px, %p[ad], (and %p[sSb]
> > +in certain circumstances [*]).  Any file written to using one of these
> > +specifiers should be readable only by privileged processes.
> > +
> > +Kernels 4.14 and older printed the raw address using %p. As of 4.15-rc1
> > +addresses printed with the specifier %p are hashed before printing.
> > +
> > +[*] If symbol lookup fails, the raw address is currently printed.
> 
> Is there a plan to adjust this case?

RFC is in flight at the moment

[RFC 0/3] kallsyms: don't leak address when printing symbol

You commented already that you liked it. Had no response from Steve, I
was intending to give him two weeks and then put in the patch for
real.

Or I could put it in without the ftrace stuff and just break tracing -
just kidding, I wouldn't do that :)

thanks,
Tobin.

Re: [PATCH] docs: add documentation on printing kernel addresses

[Tobin C. Harding]

Please drop this patch, needs amendment (commented inline).

On Thu, Dec 07, 2017 at 11:26:38AM +1100, Tobin C. Harding wrote:
> Hashing addresses printed with printk specifier %p was implemented
> recently. During development a number of issues were raised regarding
> leaking kernel addresses to userspace. We should update the
> documentation appropriately.
> 
> Add documentation regarding printing kernel addresses.
> 
> Signed-off-by: Tobin C. Harding <me@tobin.cc>
> ---
> 
> Is there a proffered method for subscripts in sphinx kernel docs? Here
> we use '[*]' 
> 
> thanks,
> Tobin.
> 
>  Documentation/security/self-protection.rst | 14 ++++++++++++++
>  1 file changed, 14 insertions(+)
> 
> diff --git a/Documentation/security/self-protection.rst b/Documentation/security/self-protection.rst
> index 60c8bd8b77bf..e711280cfdd7 100644
> --- a/Documentation/security/self-protection.rst
> +++ b/Documentation/security/self-protection.rst
> @@ -270,6 +270,20 @@ attacks, it is important to defend against exposure of both kernel memory
>  addresses and kernel memory contents (since they may contain kernel
>  addresses or other sensitive things like canary values).
>  
> +Kernel addresses
> +----------------
> +
> +Printing kernel addresses to userspace leaks sensitive information about
> +the kernel memory layout. Care should be exercised when using any printk
> +specifier that prints the raw address, currently %px, %p[ad], (and %p[sSb]
> +in certain circumstances [*]).  Any file written to using one of these
> +specifiers should be readable only by privileged processes.
> +
> +Kernels 4.14 and older printed the raw address using %p. As of 4.15-rc1
> +addresses printed with the specifier %p are hashed before printing.
> +
> +[*] If symbol lookup fails, the raw address is currently printed.

[*] If KALLSYMS is enabled and symbol lookup fails, the raw address is
currently printed. If KALLSYMS is not enabled the address is printed. 

thanks,
Tobin.