Re: [PATCH] selinux: add __randomize_layout to selinux_audit_data

Re: [PATCH] selinux: add __randomize_layout to selinux_audit_data

[Paul Moore]

On Tue, May 17, 2022 at 9:21 PM GONG, Ruiqi <gongruiqi1@huawei.com> wrote:
>
> Randomize the layout of struct selinux_audit_data as suggested in [1],
> since it contains a pointer to struct selinux_state, an already
> randomized strucure.
>
> [1]: https://github.com/KSPP/linux/issues/188
>
> Signed-off-by: GONG, Ruiqi <gongruiqi1@huawei.com>
> ---
>  security/selinux/include/avc.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/security/selinux/include/avc.h b/security/selinux/include/avc.h
> index 2b372f98f2d7..5525b94fd266 100644
> --- a/security/selinux/include/avc.h
> +++ b/security/selinux/include/avc.h
> @@ -53,7 +53,7 @@ struct selinux_audit_data {
>         u32 denied;
>         int result;
>         struct selinux_state *state;
> -};
> +} __randomize_layout;

I'll apologize in advance for the stupid question, but does
__randomize_layout result in any problems when the struct is used in a
trace event?  (see include/trace/events/avc.h)

-- 
paul-moore.com

Re: [PATCH] selinux: add __randomize_layout to selinux_audit_data

[Gong Ruiqi]

On 2022/05/18 9:39, Paul Moore wrote:
> On Tue, May 17, 2022 at 9:21 PM GONG, Ruiqi <gongruiqi1@huawei.com> wrote:
>>
>> Randomize the layout of struct selinux_audit_data as suggested in [1],
>> since it contains a pointer to struct selinux_state, an already
>> randomized strucure.
>>
>> [1]: https://github.com/KSPP/linux/issues/188
>>
>> Signed-off-by: GONG, Ruiqi <gongruiqi1@huawei.com>
>> ---
>>  security/selinux/include/avc.h | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/security/selinux/include/avc.h b/security/selinux/include/avc.h
>> index 2b372f98f2d7..5525b94fd266 100644
>> --- a/security/selinux/include/avc.h
>> +++ b/security/selinux/include/avc.h
>> @@ -53,7 +53,7 @@ struct selinux_audit_data {
>>         u32 denied;
>>         int result;
>>         struct selinux_state *state;
>> -};
>> +} __randomize_layout;
> 
> I'll apologize in advance for the stupid question, but does

Not at all :)

> __randomize_layout result in any problems when the struct is used in a
> trace event?  (see include/trace/events/avc.h)
> 

No, as least it doesn't in the testing I did. I believe we can use the
struct tagged with __randomize_layout as normal except that 1) it should
be initialized with a designated initializer, and 2) pointers to this
type can't be cast to/from pointers to another type. Other operations
like dereferencing members of the struct (as in
include/trace/events/avc.h) shouldn't be a problem.

I did a testing to the patch on a qemu vm by running the selinux
testsuite with tracing events "avc:selinux_audited" enabled. The
testsuite completed successfully and from the tracing log I saw nothing
abnormal with my bare eyes. You can do more testing if you want or you
have other ideas of how to do so ;)

[PATCH] selinux: add __randomize_layout to selinux_audit_data

[GONG, Ruiqi]

Randomize the layout of struct selinux_audit_data as suggested in [1],
since it contains a pointer to struct selinux_state, an already
randomized strucure.

[1]: https://github.com/KSPP/linux/issues/188

Signed-off-by: GONG, Ruiqi <gongruiqi1@huawei.com>
---
 security/selinux/include/avc.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/security/selinux/include/avc.h b/security/selinux/include/avc.h
index 2b372f98f2d7..5525b94fd266 100644
--- a/security/selinux/include/avc.h
+++ b/security/selinux/include/avc.h
@@ -53,7 +53,7 @@ struct selinux_audit_data {
 	u32 denied;
 	int result;
 	struct selinux_state *state;
-};
+} __randomize_layout;
 
 /*
  * AVC operations
-- 
2.17.1

Re: [PATCH] selinux: add __randomize_layout to selinux_audit_data

[Paul Moore]

On Wed, May 18, 2022 at 2:53 AM Gong Ruiqi <gongruiqi1@huawei.com> wrote:
> On 2022/05/18 9:39, Paul Moore wrote:
> > On Tue, May 17, 2022 at 9:21 PM GONG, Ruiqi <gongruiqi1@huawei.com> wrote:
> >>
> >> Randomize the layout of struct selinux_audit_data as suggested in [1],
> >> since it contains a pointer to struct selinux_state, an already
> >> randomized strucure.
> >>
> >> [1]: https://github.com/KSPP/linux/issues/188
> >>
> >> Signed-off-by: GONG, Ruiqi <gongruiqi1@huawei.com>
> >> ---
> >>  security/selinux/include/avc.h | 2 +-
> >>  1 file changed, 1 insertion(+), 1 deletion(-)
> >>
> >> diff --git a/security/selinux/include/avc.h b/security/selinux/include/avc.h
> >> index 2b372f98f2d7..5525b94fd266 100644
> >> --- a/security/selinux/include/avc.h
> >> +++ b/security/selinux/include/avc.h
> >> @@ -53,7 +53,7 @@ struct selinux_audit_data {
> >>         u32 denied;
> >>         int result;
> >>         struct selinux_state *state;
> >> -};
> >> +} __randomize_layout;
> >
> > I'll apologize in advance for the stupid question, but does
>
> Not at all :)
>
> > __randomize_layout result in any problems when the struct is used in a
> > trace event?  (see include/trace/events/avc.h)
>
> No, as least it doesn't in the testing I did. I believe we can use the
> struct tagged with __randomize_layout as normal except that 1) it should
> be initialized with a designated initializer, and 2) pointers to this
> type can't be cast to/from pointers to another type. Other operations
> like dereferencing members of the struct (as in
> include/trace/events/avc.h) shouldn't be a problem.
>
> I did a testing to the patch on a qemu vm by running the selinux
> testsuite with tracing events "avc:selinux_audited" enabled. The
> testsuite completed successfully and from the tracing log I saw nothing
> abnormal with my bare eyes. You can do more testing if you want or you
> have other ideas of how to do so ;)

That's great, thanks for verifying this.  I was aware of the other
restrictions but wasn't sure about tracing.  Now I know :)

It's too late to go into selinux/next for this dev cycle, but I'll
queue this up for selinux/next once the merge window closes.

Thanks again!

-- 
paul-moore.com

Re: [PATCH] selinux: add __randomize_layout to selinux_audit_data

[Paul Moore]

On Wed, May 18, 2022 at 12:03 PM Paul Moore <paul@paul-moore.com> wrote:
>
> On Wed, May 18, 2022 at 2:53 AM Gong Ruiqi <gongruiqi1@huawei.com> wrote:
> > On 2022/05/18 9:39, Paul Moore wrote:
> > > On Tue, May 17, 2022 at 9:21 PM GONG, Ruiqi <gongruiqi1@huawei.com> wrote:
> > >>
> > >> Randomize the layout of struct selinux_audit_data as suggested in [1],
> > >> since it contains a pointer to struct selinux_state, an already
> > >> randomized strucure.
> > >>
> > >> [1]: https://github.com/KSPP/linux/issues/188
> > >>
> > >> Signed-off-by: GONG, Ruiqi <gongruiqi1@huawei.com>
> > >> ---
> > >>  security/selinux/include/avc.h | 2 +-
> > >>  1 file changed, 1 insertion(+), 1 deletion(-)
> > >>
> > >> diff --git a/security/selinux/include/avc.h b/security/selinux/include/avc.h
> > >> index 2b372f98f2d7..5525b94fd266 100644
> > >> --- a/security/selinux/include/avc.h
> > >> +++ b/security/selinux/include/avc.h
> > >> @@ -53,7 +53,7 @@ struct selinux_audit_data {
> > >>         u32 denied;
> > >>         int result;
> > >>         struct selinux_state *state;
> > >> -};
> > >> +} __randomize_layout;
> > >
> > > I'll apologize in advance for the stupid question, but does
> >
> > Not at all :)
> >
> > > __randomize_layout result in any problems when the struct is used in a
> > > trace event?  (see include/trace/events/avc.h)
> >
> > No, as least it doesn't in the testing I did. I believe we can use the
> > struct tagged with __randomize_layout as normal except that 1) it should
> > be initialized with a designated initializer, and 2) pointers to this
> > type can't be cast to/from pointers to another type. Other operations
> > like dereferencing members of the struct (as in
> > include/trace/events/avc.h) shouldn't be a problem.
> >
> > I did a testing to the patch on a qemu vm by running the selinux
> > testsuite with tracing events "avc:selinux_audited" enabled. The
> > testsuite completed successfully and from the tracing log I saw nothing
> > abnormal with my bare eyes. You can do more testing if you want or you
> > have other ideas of how to do so ;)
>
> That's great, thanks for verifying this.  I was aware of the other
> restrictions but wasn't sure about tracing.  Now I know :)
>
> It's too late to go into selinux/next for this dev cycle, but I'll
> queue this up for selinux/next once the merge window closes.

I just merged this into my local selinux/next branch and I'll be
pushing it up to the mail selinux tree later today - thanks!

-- 
paul-moore.com