Received-SPF: pass (google.com: domain of linux-kernel+bounces-95093-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Message-ID: <85d71fd5ac7ffe9aee399b4754a8bb1ec44e7186.camel@huaweicloud.com>
Subject: Re: [PATCH v39 01/42] integrity: disassociate ima_filter_rule from
 security_audit_rule
From: Roberto Sassu <roberto.sassu@huaweicloud.com>
To: Casey Schaufler <casey@schaufler-ca.com>, paul@paul-moore.com, 
	linux-security-module@vger.kernel.org
Cc: jmorris@namei.org, serge@hallyn.com, keescook@chromium.org, 
 john.johansen@canonical.com, penguin-kernel@i-love.sakura.ne.jp, 
 stephen.smalley.work@gmail.com, linux-kernel@vger.kernel.org,
 mic@digikod.net,  linux-integrity@vger.kernel.org
Date: Thu, 07 Mar 2024 08:56:07 +0100
In-Reply-To: <5c99f987-d359-4366-984f-fe36fcff601b@schaufler-ca.com>
References: <20231215221636.105680-1-casey@schaufler-ca.com>
	 <20231215221636.105680-2-casey@schaufler-ca.com>
	 <d4cbe23822f7fdac900d1ebd5da9865d8bb96977.camel@huaweicloud.com>
	 <5c99f987-d359-4366-984f-fe36fcff601b@schaufler-ca.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.44.4-0ubuntu2 
Precedence: bulk
MIME-Version: 1.0

On Wed, 2024-03-06 at 08:56 -0800, Casey Schaufler wrote:
> On 3/6/2024 1:54 AM, Roberto Sassu wrote:
> > On Fri, 2023-12-15 at 14:15 -0800, Casey Schaufler wrote:
> > > Create real functions for the ima_filter_rule interfaces.
> > > These replace #defines that obscure the reuse of audit
> > > interfaces. The new functions are put in security.c because
> > > they use security module registered hooks that we don't
> > > want exported.
> > Beginner question: what makes IMA special, that the audit subsystem
> > does not need an AUDIT_LSM field to do the same that IMA would do?
> >=20
> > In other words, why can't we add the lsm_id parameter to
> > security_audit_*() functions, so that IMA can just call those?
>=20
> I have never liked the reuse of the audit filter functions, especially
> the way that they're hidden behind #define. The assumption that the
> two facilities (audit and IMA) are going to use them the same way, and
> that they will never diverge in their requirements, has always seemed
> questionable.
>=20
> In most cases audit needs an lsmblob rather than a secid+lsm_id pair,
> as there is information required about all the LSMs, not just the one
> actively involved.

Fair enough.

Thanks

Roberto

> >=20
> > Thanks
> >=20
> > Roberto
> >=20
> > > Acked-by: Paul Moore <paul@paul-moore.com>
> > > Reviewed-by: John Johansen <john.johansen@canonical.com>
> > > Signed-off-by: Casey Schaufler <casey@schaufler-ca.com>
> > > To: Mimi Zohar <zohar@linux.ibm.com>
> > > Cc: linux-integrity@vger.kernel.org
> > > ---
> > >  include/linux/security.h     | 24 ++++++++++++++++++++++++
> > >  security/integrity/ima/ima.h | 26 --------------------------
> > >  security/security.c          | 21 +++++++++++++++++++++
> > >  3 files changed, 45 insertions(+), 26 deletions(-)
> > >=20
> > > diff --git a/include/linux/security.h b/include/linux/security.h
> > > index 750130a7b9dd..4790508818ee 100644
> > > --- a/include/linux/security.h
> > > +++ b/include/linux/security.h
> > > @@ -2009,6 +2009,30 @@ static inline void security_audit_rule_free(vo=
id *lsmrule)
> > >  #endif /* CONFIG_SECURITY */
> > >  #endif /* CONFIG_AUDIT */
> > > =20
> > > +#if defined(CONFIG_IMA_LSM_RULES) && defined(CONFIG_SECURITY)
> > > +int ima_filter_rule_init(u32 field, u32 op, char *rulestr, void **ls=
mrule);
> > > +int ima_filter_rule_match(u32 secid, u32 field, u32 op, void *lsmrul=
e);
> > > +void ima_filter_rule_free(void *lsmrule);
> > > +
> > > +#else
> > > +
> > > +static inline int ima_filter_rule_init(u32 field, u32 op, char *rule=
str,
> > > +					   void **lsmrule)
> > > +{
> > > +	return 0;
> > > +}
> > > +
> > > +static inline int ima_filter_rule_match(u32 secid, u32 field, u32 op=
,
> > > +					    void *lsmrule)
> > > +{
> > > +	return 0;
> > > +}
> > > +
> > > +static inline void ima_filter_rule_free(void *lsmrule)
> > > +{ }
> > > +
> > > +#endif /* defined(CONFIG_IMA_LSM_RULES) && defined(CONFIG_SECURITY) =
*/
> > > +
> > >  #ifdef CONFIG_SECURITYFS
> > > =20
> > >  extern struct dentry *securityfs_create_file(const char *name, umode=
_t mode,
> > > diff --git a/security/integrity/ima/ima.h b/security/integrity/ima/im=
a.h
> > > index c29db699c996..560d6104de72 100644
> > > --- a/security/integrity/ima/ima.h
> > > +++ b/security/integrity/ima/ima.h
> > > @@ -420,32 +420,6 @@ static inline void ima_free_modsig(struct modsig=
 *modsig)
> > >  }
> > >  #endif /* CONFIG_IMA_APPRAISE_MODSIG */
> > > =20
> > > -/* LSM based policy rules require audit */
> > > -#ifdef CONFIG_IMA_LSM_RULES
> > > -
> > > -#define ima_filter_rule_init security_audit_rule_init
> > > -#define ima_filter_rule_free security_audit_rule_free
> > > -#define ima_filter_rule_match security_audit_rule_match
> > > -
> > > -#else
> > > -
> > > -static inline int ima_filter_rule_init(u32 field, u32 op, char *rule=
str,
> > > -				       void **lsmrule)
> > > -{
> > > -	return -EINVAL;
> > > -}
> > > -
> > > -static inline void ima_filter_rule_free(void *lsmrule)
> > > -{
> > > -}
> > > -
> > > -static inline int ima_filter_rule_match(u32 secid, u32 field, u32 op=
,
> > > -					void *lsmrule)
> > > -{
> > > -	return -EINVAL;
> > > -}
> > > -#endif /* CONFIG_IMA_LSM_RULES */
> > > -
> > >  #ifdef	CONFIG_IMA_READ_POLICY
> > >  #define	POLICY_FILE_FLAGS	(S_IWUSR | S_IRUSR)
> > >  #else
> > > diff --git a/security/security.c b/security/security.c
> > > index d7b15ea67c3f..8e5379a76369 100644
> > > --- a/security/security.c
> > > +++ b/security/security.c
> > > @@ -5350,6 +5350,27 @@ int security_audit_rule_match(u32 secid, u32 f=
ield, u32 op, void *lsmrule)
> > >  }
> > >  #endif /* CONFIG_AUDIT */
> > > =20
> > > +#ifdef CONFIG_IMA_LSM_RULES
> > > +/*
> > > + * The integrity subsystem uses the same hooks as
> > > + * the audit subsystem.
> > > + */
> > > +int ima_filter_rule_init(u32 field, u32 op, char *rulestr, void **ls=
mrule)
> > > +{
> > > +	return call_int_hook(audit_rule_init, 0, field, op, rulestr, lsmrul=
e);
> > > +}
> > > +
> > > +void ima_filter_rule_free(void *lsmrule)
> > > +{
> > > +	call_void_hook(audit_rule_free, lsmrule);
> > > +}
> > > +
> > > +int ima_filter_rule_match(u32 secid, u32 field, u32 op, void *lsmrul=
e)
> > > +{
> > > +	return call_int_hook(audit_rule_match, 0, secid, field, op, lsmrule=
);
> > > +}
> > > +#endif /* CONFIG_IMA_LSM_RULES */
> > > +
> > >  #ifdef CONFIG_BPF_SYSCALL
> > >  /**
> > >   * security_bpf() - Check if the bpf syscall operation is allowed