Date:	Mon, 5 Jan 2009 13:07:22 -0600
From:	"Serge E. Hallyn" <serue@...ibm.com>
To:	David Howells <dhowells@...hat.com>
Cc:	James Morris <jmorris@...ei.org>, Christoph Hellwig <hch@....de>,
	linux-kernel@...r.kernel.org, linux-fsdevel@...r.kernel.org,
	linux-security-module@...r.kernel.org,
	Stephen Rothwell <sfr@...b.auug.org.au>,
	Andrew Morgan <morgan@...nel.org>
Subject: Re: [PATCH] CRED: Fix regression in cap_capable() as shown up by
	sys_faccessat() [ver #2]

Quoting David Howells (dhowells@...hat.com):
> Serge E. Hallyn <serue@...ibm.com> wrote:
> 
> > Yes, I'm sorry, I've been staring at it on and off all weekend...  From
> > a high level it looks right, but i think there's a problem with naming
> > here (which also could be said to have obfuscated the original bug).
> > The security_capable() should be security_capable_eff() or somesuch,
> > and security_task_capable() should be security_task_capable_real() or
> > something.
> 
> Yes.  We've got real and effective creds, each with effective caps.  It lends
> itself to confusion most readily:-).
> 
> > In fact the current-as-implicit optimization could be put off for a kernel
> > release or so while we just have security_capable_eff(tsk, cap) and
> > security_real_eff(tsk, cap), or just security_capable(tsk, cap, flag) where
> > flag is CAP_EFF or CAP_REAL.
> 
> I'd prefer not to add an extra argument like this, especially as CAP_EFF
> shouldn't be used if tsk != current.  Furthermore, security_capable_eff() may
> only be called with tsk == current.

Ok.

> OTOH, as I pointed out, the capable() op as I've modified it to be it is
> superfluous given the task_capable() op since the cred pointer is sufficient
> to distinguish which set of creds you want to observe.
> 
> Is the attached patch preferable, then?  I would prefer to go for the
> optimisation in the common case, especially as the common case is called
> rather a lot, but maybe the naming is more important...

You have the 'acting_as' name for subj/eff, which I like.  Is there
another name you could use in place of 'real' in the  name
task_real_capable()?

I do find this version much easier to read.  It seems easier to
track capable+current_cred() vs real_capable+get_task_cred().  Could
you do a few benchmarks to gauge whether the difference the
optimization makes?

> > I'm also not thrilled about security_task_capable() and
> > security_ops->task_capable() having different args and semantics, but
> > of course I see the reason for it and figure if there's a way to improve
> > on that we can do it later.
> 
> Well, security_ops->task_capable() should not be called, except by
> security_task_capable*() and other security_ops->task_capable(), so does that
> matter?

It's just something that could cause confusion 6 months down the road
when someone who wasn't involved in this thread is trying to do some
routine LSM maintenance...  But like I say I understand the reasons
and agree, so let's ignore it.

> The reason for the way I've done things is that the creds from the specified
> process need pinning in some way.  If I just pass the task pointer through,
> then each function that needs to examine those creds must pin them for
> itself.  With SELinux, that is both SELinux itself and commoncaps.
> 
> One thing I'm wondering is can I ditch the audit flag argument to the
> task_capable() sec op if I make it contingent on tsk != NULL instead?  The
> credentials are passed by cred pointer, so, currently, tsk is only needed for
> auditing.

I'm looking at a several-week-old linux-next, but only see one use of
capable on another task which audits, and that is in commoncap for
traceme, so it seems reasonable.

> > Anyway David the patch on its own doesn't look incorrect.  So far
> > the only code which manipulates subjective caps is in fact
> > sys_faccessat through cap_set_effective() right?  If so this at
> > least looks safe, looking through capable/has_capability callers.
> 
> fs/nfsd/auth.c also.

So yeah, I do like this version better.

Acked-by: Serge Hallyn <serue@...ibm.com>

> > Finally, may i just say that i love the fact that a syscall is
> > checking the real user's access rights and so sets eff creds to
> > have the caps of the real user :)
> 
> Yes, it's quite mad.
> 
> > Hmm, did you at one time call the subjective creds 'working' or 'acting'
> > creds?  Might be a good name.  Subj/obj always makes me pause to think, and
> > real/eff while seemingly natural could be confusing in cases like this.
> > cap_set_acting() - i like it...
> 
> I was using acting and act_as.
> 
> David
> 
> ---
> diff --git a/include/linux/capability.h b/include/linux/capability.h
> index e22f48c..02bdb76 100644
> --- a/include/linux/capability.h
> +++ b/include/linux/capability.h
> @@ -529,8 +529,21 @@ extern const kernel_cap_t __cap_init_eff_set;
>   *
>   * Note that this does not set PF_SUPERPRIV on the task.
>   */
> -#define has_capability(t, cap) (security_capable((t), (cap)) == 0)
> -#define has_capability_noaudit(t, cap) (security_capable_noaudit((t), (cap)) == 0)
> +#define has_capability(t, cap) (security_real_capable((t), (cap)) == 0)
> +
> +/**
> + * has_capability_noaudit - Determine if a task has a superior capability available (unaudited)
> + * @t: The task in question
> + * @cap: The capability to be tested for
> + *
> + * Return true if the specified task has the given superior capability
> + * currently in effect, false if not, but don't write an audit message for the
> + * check.
> + *
> + * Note that this does not set PF_SUPERPRIV on the task.
> + */
> +#define has_capability_noaudit(t, cap) \
> +	(security_real_capable_noaudit((t), (cap)) == 0)
> 
>  extern int capable(int cap);
> 
> diff --git a/include/linux/security.h b/include/linux/security.h
> index b92b5e4..1f2ab63 100644
> --- a/include/linux/security.h
> +++ b/include/linux/security.h
> @@ -48,7 +48,8 @@ struct audit_krule;
>   * These functions are in security/capability.c and are used
>   * as the default capabilities functions
>   */
> -extern int cap_capable(struct task_struct *tsk, int cap, int audit);
> +extern int cap_capable(struct task_struct *tsk, const struct cred *cred,
> +		       int cap, int audit);
>  extern int cap_settime(struct timespec *ts, struct timezone *tz);
>  extern int cap_ptrace_may_access(struct task_struct *child, unsigned int mode);
>  extern int cap_ptrace_traceme(struct task_struct *parent);
> @@ -1251,9 +1252,12 @@ static inline void security_free_mnt_opts(struct security_mnt_opts *opts)
>   *	@permitted contains the permitted capability set.
>   *	Return 0 and update @new if permission is granted.
>   * @capable:
> - *	Check whether the @tsk process has the @cap capability.
> + *	Check whether the @tsk process has the @cap capability in the indicated
> + *	credentials.
>   *	@tsk contains the task_struct for the process.
> + *	@cred contains the credentials to use.
>   *	@cap contains the capability <include/linux/capability.h>.
> + *	@audit: Whether to write an audit message or not
>   *	Return 0 if the capability is granted for @tsk.
>   * @acct:
>   *	Check permission before enabling or disabling process accounting.  If
> @@ -1346,7 +1350,8 @@ struct security_operations {
>  		       const kernel_cap_t *effective,
>  		       const kernel_cap_t *inheritable,
>  		       const kernel_cap_t *permitted);
> -	int (*capable) (struct task_struct *tsk, int cap, int audit);
> +	int (*capable) (struct task_struct *tsk, const struct cred *cred,
> +			int cap, int audit);
>  	int (*acct) (struct file *file);
>  	int (*sysctl) (struct ctl_table *table, int op);
>  	int (*quotactl) (int cmds, int type, int id, struct super_block *sb);
> @@ -1628,8 +1633,9 @@ int security_capset(struct cred *new, const struct cred *old,
>  		    const kernel_cap_t *effective,
>  		    const kernel_cap_t *inheritable,
>  		    const kernel_cap_t *permitted);
> -int security_capable(struct task_struct *tsk, int cap);
> -int security_capable_noaudit(struct task_struct *tsk, int cap);
> +int security_capable(int cap);
> +int security_real_capable(struct task_struct *tsk, int cap);
> +int security_real_capable_noaudit(struct task_struct *tsk, int cap);
>  int security_acct(struct file *file);
>  int security_sysctl(struct ctl_table *table, int op);
>  int security_quotactl(int cmds, int type, int id, struct super_block *sb);
> @@ -1826,14 +1832,31 @@ static inline int security_capset(struct cred *new,
>  	return cap_capset(new, old, effective, inheritable, permitted);
>  }
> 
> -static inline int security_capable(struct task_struct *tsk, int cap)
> +static inline int security_capable(int cap)
>  {
> -	return cap_capable(tsk, cap, SECURITY_CAP_AUDIT);
> +	return cap_capable(current, current_cred(), cap, SECURITY_CAP_AUDIT);
>  }
> 
> -static inline int security_capable_noaudit(struct task_struct *tsk, int cap)
> +static inline int security_real_capable(struct task_struct *tsk, int cap)
>  {
> -	return cap_capable(tsk, cap, SECURITY_CAP_NOAUDIT);
> +	int ret;
> +
> +	rcu_read_lock();
> +	ret = cap_capable(tsk, __task_cred(tsk), cap, SECURITY_CAP_AUDIT);
> +	rcu_read_unlock();
> +	return ret;
> +}
> +
> +static inline
> +int security_real_capable_noaudit(struct task_struct *tsk, int cap)
> +{
> +	int ret;
> +
> +	rcu_read_lock();
> +	ret = cap_capable(tsk, __task_cred(tsk), cap,
> +			       SECURITY_CAP_NOAUDIT);
> +	rcu_read_unlock();
> +	return ret;
>  }
> 
>  static inline int security_acct(struct file *file)
> diff --git a/kernel/capability.c b/kernel/capability.c
> index c598d9d..688926e 100644
> --- a/kernel/capability.c
> +++ b/kernel/capability.c
> @@ -306,7 +306,7 @@ int capable(int cap)
>  		BUG();
>  	}
> 
> -	if (has_capability(current, cap)) {
> +	if (security_capable(cap) == 0) {
>  		current->flags |= PF_SUPERPRIV;
>  		return 1;
>  	}
> diff --git a/security/commoncap.c b/security/commoncap.c
> index 7971354..f0e671d 100644
> --- a/security/commoncap.c
> +++ b/security/commoncap.c
> @@ -45,26 +45,22 @@ EXPORT_SYMBOL(cap_netlink_recv);
>  /**
>   * cap_capable - Determine whether a task has a particular effective capability
>   * @tsk: The task to query
> + * @cred: The credentials to use
>   * @cap: The capability to check for
>   * @audit: Whether to write an audit message or not
>   *
>   * Determine whether the nominated task has the specified capability amongst
>   * its effective set, returning 0 if it does, -ve if it does not.
>   *
> - * NOTE WELL: cap_capable() cannot be used like the kernel's capable()
> - * function.  That is, it has the reverse semantics: cap_capable() returns 0
> - * when a task has a capability, but the kernel's capable() returns 1 for this
> - * case.
> + * NOTE WELL: cap_has_capability() cannot be used like the kernel's capable()
> + * and has_capability() functions.  That is, it has the reverse semantics:
> + * cap_has_capability() returns 0 when a task has a capability, but the
> + * kernel's capable() and has_capability() returns 1 for this case.
>   */
> -int cap_capable(struct task_struct *tsk, int cap, int audit)
> +int cap_capable(struct task_struct *tsk, const struct cred *cred, int cap,
> +		int audit)
>  {
> -	__u32 cap_raised;
> -
> -	/* Derived from include/linux/sched.h:capable. */
> -	rcu_read_lock();
> -	cap_raised = cap_raised(__task_cred(tsk)->cap_effective, cap);
> -	rcu_read_unlock();
> -	return cap_raised ? 0 : -EPERM;
> +	return cap_raised(cred->cap_effective, cap) ? 0 : -EPERM;
>  }
> 
>  /**
> @@ -160,7 +156,8 @@ static inline int cap_inh_is_capped(void)
>  	/* they are so limited unless the current task has the CAP_SETPCAP
>  	 * capability
>  	 */
> -	if (cap_capable(current, CAP_SETPCAP, SECURITY_CAP_AUDIT) == 0)
> +	if (cap_capable(current, current_cred(), CAP_SETPCAP,
> +			SECURITY_CAP_AUDIT) == 0)
>  		return 0;
>  #endif
>  	return 1;
> @@ -869,7 +866,8 @@ int cap_task_prctl(int option, unsigned long arg2, unsigned long arg3,
>  		     & (new->securebits ^ arg2))			/*[1]*/
>  		    || ((new->securebits & SECURE_ALL_LOCKS & ~arg2))	/*[2]*/
>  		    || (arg2 & ~(SECURE_ALL_LOCKS | SECURE_ALL_BITS))	/*[3]*/
> -		    || (cap_capable(current, CAP_SETPCAP, SECURITY_CAP_AUDIT) != 0) /*[4]*/
> +		    || (cap_capable(current, current_cred(), CAP_SETPCAP,
> +				    SECURITY_CAP_AUDIT) != 0)		/*[4]*/
>  			/*
>  			 * [1] no changing of bits that are locked
>  			 * [2] no unlocking of locks
> @@ -950,7 +948,8 @@ int cap_vm_enough_memory(struct mm_struct *mm, long pages)
>  {
>  	int cap_sys_admin = 0;
> 
> -	if (cap_capable(current, CAP_SYS_ADMIN, SECURITY_CAP_NOAUDIT) == 0)
> +	if (cap_capable(current, current_cred(), CAP_SYS_ADMIN,
> +			SECURITY_CAP_NOAUDIT) == 0)
>  		cap_sys_admin = 1;
>  	return __vm_enough_memory(mm, pages, cap_sys_admin);
>  }
> diff --git a/security/security.c b/security/security.c
> index 678d4d0..c3586c0 100644
> --- a/security/security.c
> +++ b/security/security.c
> @@ -154,14 +154,32 @@ int security_capset(struct cred *new, const struct cred *old,
>  				    effective, inheritable, permitted);
>  }
> 
> -int security_capable(struct task_struct *tsk, int cap)
> +int security_capable(int cap)
>  {
> -	return security_ops->capable(tsk, cap, SECURITY_CAP_AUDIT);
> +	return security_ops->capable(current, current_cred(), cap,
> +				     SECURITY_CAP_AUDIT);
>  }
> 
> -int security_capable_noaudit(struct task_struct *tsk, int cap)
> +int security_real_capable(struct task_struct *tsk, int cap)
>  {
> -	return security_ops->capable(tsk, cap, SECURITY_CAP_NOAUDIT);
> +	const struct cred *cred;
> +	int ret;
> +
> +	cred = get_task_cred(tsk);
> +	ret = security_ops->capable(tsk, cred, cap, SECURITY_CAP_AUDIT);
> +	put_cred(cred);
> +	return ret;
> +}
> +
> +int security_real_capable_noaudit(struct task_struct *tsk, int cap)
> +{
> +	const struct cred *cred;
> +	int ret;
> +
> +	cred = get_task_cred(tsk);
> +	ret = security_ops->capable(tsk, cred, cap, SECURITY_CAP_NOAUDIT);
> +	put_cred(cred);
> +	return ret;
>  }
> 
>  int security_acct(struct file *file)
> diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
> index dbeaa78..e0cb106 100644
> --- a/security/selinux/hooks.c
> +++ b/security/selinux/hooks.c
> @@ -1433,12 +1433,13 @@ static int current_has_perm(const struct task_struct *tsk,
> 
>  /* Check whether a task is allowed to use a capability. */
>  static int task_has_capability(struct task_struct *tsk,
> +			       const struct cred *cred,
>  			       int cap, int audit)
>  {
>  	struct avc_audit_data ad;
>  	struct av_decision avd;
>  	u16 sclass;
> -	u32 sid = task_sid(tsk);
> +	u32 sid = cred_sid(cred);
>  	u32 av = CAP_TO_MASK(cap);
>  	int rc;
> 
> @@ -1865,15 +1866,16 @@ static int selinux_capset(struct cred *new, const struct cred *old,
>  	return cred_has_perm(old, new, PROCESS__SETCAP);
>  }
> 
> -static int selinux_capable(struct task_struct *tsk, int cap, int audit)
> +static int selinux_capable(struct task_struct *tsk, const struct cred *cred,
> +			   int cap, int audit)
>  {
>  	int rc;
> 
> -	rc = secondary_ops->capable(tsk, cap, audit);
> +	rc = secondary_ops->capable(tsk, cred, cap, audit);
>  	if (rc)
>  		return rc;
> 
> -	return task_has_capability(tsk, cap, audit);
> +	return task_has_capability(tsk, cred, cap, audit);
>  }
> 
>  static int selinux_sysctl_get_sid(ctl_table *table, u16 tclass, u32 *sid)
> @@ -2037,7 +2039,8 @@ static int selinux_vm_enough_memory(struct mm_struct *mm, long pages)
>  {
>  	int rc, cap_sys_admin = 0;
> 
> -	rc = selinux_capable(current, CAP_SYS_ADMIN, SECURITY_CAP_NOAUDIT);
> +	rc = selinux_capable(current, current_cred(), CAP_SYS_ADMIN,
> +			     SECURITY_CAP_NOAUDIT);
>  	if (rc == 0)
>  		cap_sys_admin = 1;
> 
> @@ -2880,7 +2883,8 @@ static int selinux_inode_getsecurity(const struct inode *inode, const char *name
>  	 * and lack of permission just means that we fall back to the
>  	 * in-core context value, not a denial.
>  	 */
> -	error = selinux_capable(current, CAP_MAC_ADMIN, SECURITY_CAP_NOAUDIT);
> +	error = selinux_capable(current, current_cred(), CAP_MAC_ADMIN,
> +				SECURITY_CAP_NOAUDIT);
>  	if (!error)
>  		error = security_sid_to_context_force(isec->sid, &context,
>  						      &size);
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Powered by blists - more mailing lists