Re: [PATCH v2] selinux: log anon inode class name

From: Paul Moore <paul@paul-moore.com>
To: "Christian Göttsche" <cgzones@googlemail.com>
Cc: selinux@vger.kernel.org, James Morris <jmorris@namei.org>,
	"Serge E. Hallyn" <serge@hallyn.com>,
	Stephen Smalley <stephen.smalley.work@gmail.com>,
	Eric Paris <eparis@parisplace.org>,
	Richard Guy Briggs <rgb@redhat.com>,
	Ondrej Mosnacek <omosnace@redhat.com>,
	linux-kernel@vger.kernel.org,
	linux-security-module@vger.kernel.org
Subject: Re: [PATCH v2] selinux: log anon inode class name
Date: Wed, 27 Apr 2022 09:18:55 -0400	[thread overview]
Message-ID: <CAHC9VhR1d2aLKsZOxLb6b1uuTcWOpnJ22S5=mXygvjcv6Sm=xg@mail.gmail.com> (raw)
In-Reply-To: <CAHC9VhSiqvCbKQHYTGAj3vqECNto6eNm0MyzLd92kcJnvZSw1A@mail.gmail.com>

On Mon, Apr 4, 2022 at 4:18 PM Paul Moore <paul@paul-moore.com> wrote:
> On Tue, Mar 8, 2022 at 12:09 PM Christian Göttsche
> <cgzones@googlemail.com> wrote:
> >
> > Log the anonymous inode class name in the security hook
> > inode_init_security_anon.  This name is the key for name based type
> > transitions on the anon_inode security class on creation.  Example:
> >
> >     type=AVC msg=audit(02/16/22 22:02:50.585:216) : avc:  granted \
> >         { create } for  pid=2136 comm=mariadbd anonclass="[io_uring]" \
> >         scontext=system_u:system_r:mysqld_t:s0 \
> >         tcontext=system_u:system_r:mysqld_iouring_t:s0 tclass=anon_inode
> >
> > Add a new LSM audit data type holding the inode and the class name.
> >
> > Signed-off-by: Christian Göttsche <cgzones@googlemail.com>
> >
> > ---
> > v2:
> >   - drop dev= and name= output for anonymous inodes, and hence simplify
> >     the common_audit_data union member.
> >   - drop WARN_ON() on empty name passed to inode_init_security_anon hook
> > ---
> >  include/linux/lsm_audit.h | 2 ++
> >  security/lsm_audit.c      | 4 ++++
> >  security/selinux/hooks.c  | 4 ++--
> >  3 files changed, 8 insertions(+), 2 deletions(-)
> >
> > diff --git a/include/linux/lsm_audit.h b/include/linux/lsm_audit.h
> > index 17d02eda9538..97a8b21eb033 100644
> > --- a/include/linux/lsm_audit.h
> > +++ b/include/linux/lsm_audit.h
> > @@ -76,6 +76,7 @@ struct common_audit_data {
> >  #define LSM_AUDIT_DATA_IBENDPORT 14
> >  #define LSM_AUDIT_DATA_LOCKDOWN 15
> >  #define LSM_AUDIT_DATA_NOTIFICATION 16
> > +#define LSM_AUDIT_DATA_ANONINODE       17
> >         union   {
> >                 struct path path;
> >                 struct dentry *dentry;
> > @@ -96,6 +97,7 @@ struct common_audit_data {
> >                 struct lsm_ibpkey_audit *ibpkey;
> >                 struct lsm_ibendport_audit *ibendport;
> >                 int reason;
> > +               const char *anonclass;
> >         } u;
> >         /* this union contains LSM specific data */
> >         union {
> > diff --git a/security/lsm_audit.c b/security/lsm_audit.c
> > index 1897cbf6fc69..981f6a4e4590 100644
> > --- a/security/lsm_audit.c
> > +++ b/security/lsm_audit.c
> > @@ -433,6 +433,10 @@ static void dump_common_audit_data(struct audit_buffer *ab,
> >                 audit_log_format(ab, " lockdown_reason=\"%s\"",
> >                                  lockdown_reasons[a->u.reason]);
> >                 break;
> > +       case LSM_AUDIT_DATA_ANONINODE:
> > +               audit_log_format(ab, " anonclass=");
> > +               audit_log_untrustedstring(ab, a->u.anonclass);
>
> My apologies, I didn't notice this in the previous patch ... I don't
> think we need to log this as an untrusted string as the string value
> is coming from the kernel, not userspace, so we could rewrite the
> above as the following:
>
>   audit_log_format(ab, " anonclass=%s", a->u.anonclass);
>
> ... if you are okay with that, I can make the change when I merge the
> patch or you can submit another revision, let me know which you would
> prefer.
>
> The rest of the patch looks good, thanks!

Hi Christian,

I just wanted to follow up on this as we are at -rc4 this week and if
we want this to go during the next merge window this would need to be
merged soon ...

-- 
paul-moore.com