diff mbox

selinux: check CAP_SETFCAP for a particular inode & mapped user

Message ID 20171005061611.6649-1-lkundrak@v3.sk (mailing list archive)
State New, archived
Headers show

Commit Message

Lubomir Rintel Oct. 5, 2017, 6:16 a.m. UTC
This allows setting "security.capability" xattr by a user that has
CAP_SETFCAP in an userns with SELinux. Namespaced capabilities are
supported, as of commit 8db6c34f1dbc ("Introduce v3 namespaced file
capabilities").

Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
---
A casual warning: Please not assume I know what I'm really doing. I can now 
install iputils and /bin/ping in a LXC container, which is great, but I 
typically wouldn't be allowed anywhere near anything security related.

Thank you!

 security/selinux/hooks.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

Comments

Serge Hallyn Oct. 5, 2017, 6:48 a.m. UTC | #1
On Thu, Oct 05, 2017 at 08:16:11AM +0200, Lubomir Rintel wrote:
> This allows setting "security.capability" xattr by a user that has
> CAP_SETFCAP in an userns with SELinux. Namespaced capabilities are
> supported, as of commit 8db6c34f1dbc ("Introduce v3 namespaced file
> capabilities").
> 
> Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>

The fix is already on its way into the kernel - see the thread at
https://marc.info/?l=selinux&m=150713903925728&w=2

Thanks though :)

-serge

> ---
> A casual warning: Please not assume I know what I'm really doing. I can now 
> install iputils and /bin/ping in a LXC container, which is great, but I 
> typically wouldn't be allowed anywhere near anything security related.
> 
> Thank you!
> 
>  security/selinux/hooks.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
> index f5d304736852..7bc5f4ffb4e0 100644
> --- a/security/selinux/hooks.c
> +++ b/security/selinux/hooks.c
> @@ -3127,11 +3127,12 @@ static int selinux_inode_getattr(const struct path *path)
>  static int selinux_inode_setotherxattr(struct dentry *dentry, const char *name)
>  {
>  	const struct cred *cred = current_cred();
> +	struct inode *inode = d_backing_inode(dentry);
>  
>  	if (!strncmp(name, XATTR_SECURITY_PREFIX,
>  		     sizeof XATTR_SECURITY_PREFIX - 1)) {
>  		if (!strcmp(name, XATTR_NAME_CAPS)) {
> -			if (!capable(CAP_SETFCAP))
> +			if (!capable_wrt_inode_uidgid(inode, CAP_SETFCAP))
>  				return -EPERM;
>  		} else if (!capable(CAP_SYS_ADMIN)) {
>  			/* A different attribute in the security namespace.
> -- 
> 2.13.6
--
To unsubscribe from this list: send the line "unsubscribe linux-security-module" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Lubomir Rintel Oct. 9, 2017, 6:14 p.m. UTC | #2
On Thu, 2017-10-05 at 01:48 -0500, Serge E. Hallyn wrote:
> On Thu, Oct 05, 2017 at 08:16:11AM +0200, Lubomir Rintel wrote:
> > This allows setting "security.capability" xattr by a user that has
> > CAP_SETFCAP in an userns with SELinux. Namespaced capabilities are
> > supported, as of commit 8db6c34f1dbc ("Introduce v3 namespaced file
> > capabilities").
> > 
> > Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
> 
> The fix is already on its way into the kernel - see the thread at
> https://marc.info/?l=selinux&m=150713903925728&w=2

Thanks for the response. I'm wondering if there are chances this could
go to 4.14, given the patch that got merged seems reasonably self-
contained, quite useful, and the release is going to be a long-term
one.

If rc5 is too late for such changes, are there changes it could perhaps
go to a subsequent -stable release?

Thanks,
Lubo

> 
> Thanks though :)
> 
> -serge
> 
> > ---
> > A casual warning: Please not assume I know what I'm really doing. I
> > can now 
> > install iputils and /bin/ping in a LXC container, which is great,
> > but I 
> > typically wouldn't be allowed anywhere near anything security
> > related.
> > 
> > Thank you!
> > 
> >  security/selinux/hooks.c | 3 ++-
> >  1 file changed, 2 insertions(+), 1 deletion(-)
> > 
> > diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
> > index f5d304736852..7bc5f4ffb4e0 100644
> > --- a/security/selinux/hooks.c
> > +++ b/security/selinux/hooks.c
> > @@ -3127,11 +3127,12 @@ static int selinux_inode_getattr(const
> > struct path *path)
> >  static int selinux_inode_setotherxattr(struct dentry *dentry,
> > const char *name)
> >  {
> >  	const struct cred *cred = current_cred();
> > +	struct inode *inode = d_backing_inode(dentry);
> >  
> >  	if (!strncmp(name, XATTR_SECURITY_PREFIX,
> >  		     sizeof XATTR_SECURITY_PREFIX - 1)) {
> >  		if (!strcmp(name, XATTR_NAME_CAPS)) {
> > -			if (!capable(CAP_SETFCAP))
> > +			if (!capable_wrt_inode_uidgid(inode,
> > CAP_SETFCAP))
> >  				return -EPERM;
> >  		} else if (!capable(CAP_SYS_ADMIN)) {
> >  			/* A different attribute in the security
> > namespace.
> > -- 
> > 2.13.6
--
To unsubscribe from this list: send the line "unsubscribe linux-security-module" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Paul Moore Oct. 9, 2017, 6:21 p.m. UTC | #3
On Mon, Oct 9, 2017 at 2:14 PM, Lubomir Rintel <lkundrak@v3.sk> wrote:
> On Thu, 2017-10-05 at 01:48 -0500, Serge E. Hallyn wrote:
>> On Thu, Oct 05, 2017 at 08:16:11AM +0200, Lubomir Rintel wrote:
>> > This allows setting "security.capability" xattr by a user that has
>> > CAP_SETFCAP in an userns with SELinux. Namespaced capabilities are
>> > supported, as of commit 8db6c34f1dbc ("Introduce v3 namespaced file
>> > capabilities").
>> >
>> > Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
>>
>> The fix is already on its way into the kernel - see the thread at
>> https://marc.info/?l=selinux&m=150713903925728&w=2
>
> Thanks for the response. I'm wondering if there are chances this could
> go to 4.14, given the patch that got merged seems reasonably self-
> contained, quite useful, and the release is going to be a long-term
> one.
>
> If rc5 is too late for such changes, are there changes it could perhaps
> go to a subsequent -stable release?

As a general rule I only send fixes, not new functionality, upwards
during the -rcX releases and this falls into the "functionality" not
"fix" category of patches.  I apply a similar approach to -stable
kernels.
diff mbox

Patch

diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
index f5d304736852..7bc5f4ffb4e0 100644
--- a/security/selinux/hooks.c
+++ b/security/selinux/hooks.c
@@ -3127,11 +3127,12 @@  static int selinux_inode_getattr(const struct path *path)
 static int selinux_inode_setotherxattr(struct dentry *dentry, const char *name)
 {
 	const struct cred *cred = current_cred();
+	struct inode *inode = d_backing_inode(dentry);
 
 	if (!strncmp(name, XATTR_SECURITY_PREFIX,
 		     sizeof XATTR_SECURITY_PREFIX - 1)) {
 		if (!strcmp(name, XATTR_NAME_CAPS)) {
-			if (!capable(CAP_SETFCAP))
+			if (!capable_wrt_inode_uidgid(inode, CAP_SETFCAP))
 				return -EPERM;
 		} else if (!capable(CAP_SYS_ADMIN)) {
 			/* A different attribute in the security namespace.