2021-11-24 14:39:17

by Jari Ruusu

Subject: Re: [PATCH 5.10 130/154] block: Check ADMIN before NICE for IOPRIO_CLASS_RT

Greg Kroah-Hartman wrote:
> From: Alistair Delva <[email protected]>
>
> commit 94c4b4fd25e6c3763941bdec3ad54f2204afa992 upstream.
[SNIP]
> --- a/block/ioprio.c
> +++ b/block/ioprio.c
> @@ -69,7 +69,14 @@ int ioprio_check_cap(int ioprio)
>
> switch (class) {
> case IOPRIO_CLASS_RT:
> - if (!capable(CAP_SYS_NICE) && !capable(CAP_SYS_ADMIN))
> + /*
> + * Originally this only checked for CAP_SYS_ADMIN,
> + * which was implicitly allowed for pid 0 by security
> + * modules such as SELinux. Make sure we check
> + * CAP_SYS_ADMIN first to avoid a denial/avc for
> + * possibly missing CAP_SYS_NICE permission.
> + */
> + if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_NICE))
> return -EPERM;
> fallthrough;
> /* rt has prio field too */

What exactly is above patch trying to fix?
It does not change control flow at all, and added comment is misleading.

--
Jari Ruusu 4096R/8132F189 12D6 4C3A DCDA 0AA4 27BD ACDF F073 3C80 8132 F189

2021-11-24 15:31:29

by Greg KH

Subject: Re: [PATCH 5.10 130/154] block: Check ADMIN before NICE for IOPRIO_CLASS_RT

On Wed, Nov 24, 2021 at 04:22:50PM +0200, Jari Ruusu wrote:
> Greg Kroah-Hartman wrote:
> > From: Alistair Delva <[email protected]>
> >
> > commit 94c4b4fd25e6c3763941bdec3ad54f2204afa992 upstream.
> [SNIP]
> > --- a/block/ioprio.c
> > +++ b/block/ioprio.c
> > @@ -69,7 +69,14 @@ int ioprio_check_cap(int ioprio)
> >
> > switch (class) {
> > case IOPRIO_CLASS_RT:
> > - if (!capable(CAP_SYS_NICE) && !capable(CAP_SYS_ADMIN))
> > + /*
> > + * Originally this only checked for CAP_SYS_ADMIN,
> > + * which was implicitly allowed for pid 0 by security
> > + * modules such as SELinux. Make sure we check
> > + * CAP_SYS_ADMIN first to avoid a denial/avc for
> > + * possibly missing CAP_SYS_NICE permission.
> > + */
> > + if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_NICE))
> > return -EPERM;
> > fallthrough;
> > /* rt has prio field too */
>
> What exactly is above patch trying to fix?
> It does not change control flow at all, and added comment is misleading.

See the thread on the mailing list for what it does and why it is
needed.

It does change the result when selinux is enabled.

thanks,

greg k-h

2021-11-24 17:33:16

by Serge E. Hallyn

Subject: Re: [PATCH 5.10 130/154] block: Check ADMIN before NICE for IOPRIO_CLASS_RT

On Wed, Nov 24, 2021 at 04:31:22PM +0100, Greg Kroah-Hartman wrote:
> On Wed, Nov 24, 2021 at 04:22:50PM +0200, Jari Ruusu wrote:
> > Greg Kroah-Hartman wrote:
> > > From: Alistair Delva <[email protected]>
> > >
> > > commit 94c4b4fd25e6c3763941bdec3ad54f2204afa992 upstream.
> > [SNIP]
> > > --- a/block/ioprio.c
> > > +++ b/block/ioprio.c
> > > @@ -69,7 +69,14 @@ int ioprio_check_cap(int ioprio)
> > >
> > > switch (class) {
> > > case IOPRIO_CLASS_RT:
> > > - if (!capable(CAP_SYS_NICE) && !capable(CAP_SYS_ADMIN))
> > > + /*
> > > + * Originally this only checked for CAP_SYS_ADMIN,
> > > + * which was implicitly allowed for pid 0 by security
> > > + * modules such as SELinux. Make sure we check
> > > + * CAP_SYS_ADMIN first to avoid a denial/avc for
> > > + * possibly missing CAP_SYS_NICE permission.
> > > + */
> > > + if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_NICE))
> > > return -EPERM;
> > > fallthrough;
> > > /* rt has prio field too */
> >
> > What exactly is above patch trying to fix?
> > It does not change control flow at all, and added comment is misleading.
>
> See the thread on the mailing list for what it does and why it is
> needed.
>
> It does change the result when selinux is enabled.
>
> thanks,
>
> greg k-h

The case where we create a newer more fine grained capability which is a
sub-cap of a broader capability like CAP_SYS_ADMIN is analogous. See
check_syslog_permissions() for instance.

So I think a helper like

int capable_either_or(int cap1, int cap2) {
if (has_capability_noaudit(current, cap1))
return 0;
return capable(cap2);
}

might be worthwhile.

2021-11-24 18:16:12

by Greg KH

Subject: Re: [PATCH 5.10 130/154] block: Check ADMIN before NICE for IOPRIO_CLASS_RT

On Wed, Nov 24, 2021 at 11:33:11AM -0600, Serge E. Hallyn wrote:
> On Wed, Nov 24, 2021 at 04:31:22PM +0100, Greg Kroah-Hartman wrote:
> > On Wed, Nov 24, 2021 at 04:22:50PM +0200, Jari Ruusu wrote:
> > > Greg Kroah-Hartman wrote:
> > > > From: Alistair Delva <[email protected]>
> > > >
> > > > commit 94c4b4fd25e6c3763941bdec3ad54f2204afa992 upstream.
> > > [SNIP]
> > > > --- a/block/ioprio.c
> > > > +++ b/block/ioprio.c
> > > > @@ -69,7 +69,14 @@ int ioprio_check_cap(int ioprio)
> > > >
> > > > switch (class) {
> > > > case IOPRIO_CLASS_RT:
> > > > - if (!capable(CAP_SYS_NICE) && !capable(CAP_SYS_ADMIN))
> > > > + /*
> > > > + * Originally this only checked for CAP_SYS_ADMIN,
> > > > + * which was implicitly allowed for pid 0 by security
> > > > + * modules such as SELinux. Make sure we check
> > > > + * CAP_SYS_ADMIN first to avoid a denial/avc for
> > > > + * possibly missing CAP_SYS_NICE permission.
> > > > + */
> > > > + if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_NICE))
> > > > return -EPERM;
> > > > fallthrough;
> > > > /* rt has prio field too */
> > >
> > > What exactly is above patch trying to fix?
> > > It does not change control flow at all, and added comment is misleading.
> >
> > See the thread on the mailing list for what it does and why it is
> > needed.
> >
> > It does change the result when selinux is enabled.
> >
> > thanks,
> >
> > greg k-h
>
> The case where we create a newer more fine grained capability which is a
> sub-cap of a broader capability like CAP_SYS_ADMIN is analogous. See
> check_syslog_permissions() for instance.
>
> So I think a helper like
>
> int capable_either_or(int cap1, int cap2) {
> if (has_capability_noaudit(current, cap1))
> return 0;
> return capable(cap2);
> }
>
> might be worthwhile.

Sure, feel free to work on that and submit it, but for now, this change
is needed.

thanks,

greg k-h

2021-11-24 18:35:13

by Christian Göttsche

Subject: Re: [PATCH 5.10 130/154] block: Check ADMIN before NICE for IOPRIO_CLASS_RT

On Wed, 24 Nov 2021 at 19:16, Greg Kroah-Hartman
<[email protected]> wrote:
>
> On Wed, Nov 24, 2021 at 11:33:11AM -0600, Serge E. Hallyn wrote:
> > On Wed, Nov 24, 2021 at 04:31:22PM +0100, Greg Kroah-Hartman wrote:
> > > On Wed, Nov 24, 2021 at 04:22:50PM +0200, Jari Ruusu wrote:
> > > > Greg Kroah-Hartman wrote:
> > > > > From: Alistair Delva <[email protected]>
> > > > >
> > > > > commit 94c4b4fd25e6c3763941bdec3ad54f2204afa992 upstream.
> > > > [SNIP]
> > > > > --- a/block/ioprio.c
> > > > > +++ b/block/ioprio.c
> > > > > @@ -69,7 +69,14 @@ int ioprio_check_cap(int ioprio)
> > > > >
> > > > > switch (class) {
> > > > > case IOPRIO_CLASS_RT:
> > > > > - if (!capable(CAP_SYS_NICE) && !capable(CAP_SYS_ADMIN))
> > > > > + /*
> > > > > + * Originally this only checked for CAP_SYS_ADMIN,
> > > > > + * which was implicitly allowed for pid 0 by security
> > > > > + * modules such as SELinux. Make sure we check
> > > > > + * CAP_SYS_ADMIN first to avoid a denial/avc for
> > > > > + * possibly missing CAP_SYS_NICE permission.
> > > > > + */
> > > > > + if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_NICE))
> > > > > return -EPERM;
> > > > > fallthrough;
> > > > > /* rt has prio field too */
> > > >
> > > > What exactly is above patch trying to fix?
> > > > It does not change control flow at all, and added comment is misleading.
> > >
> > > See the thread on the mailing list for what it does and why it is
> > > needed.
> > >
> > > It does change the result when selinux is enabled.
> > >
> > > thanks,
> > >
> > > greg k-h
> >
> > The case where we create a newer more fine grained capability which is a
> > sub-cap of a broader capability like CAP_SYS_ADMIN is analogous. See
> > check_syslog_permissions() for instance.
> >
> > So I think a helper like
> >
> > int capable_either_or(int cap1, int cap2) {
> > if (has_capability_noaudit(current, cap1))
> > return 0;
> > return capable(cap2);
> > }
> >
> > might be worthwhile.
>

I proposed an early prototype at
https://patchwork.kernel.org/project/selinux/patch/[email protected]/

> Sure, feel free to work on that and submit it, but for now, this change
> is needed.
>

I would argue this change is not necessary since the actual syscall
still succeeds as this is only an informative avc denial message about
a failed capability check. But this ship has sailed...

> thanks,
>
> greg k-h

2021-11-24 23:29:13

by Serge E. Hallyn

Subject: Re: [PATCH 5.10 130/154] block: Check ADMIN before NICE for IOPRIO_CLASS_RT

On Wed, Nov 24, 2021 at 07:15:35PM +0100, Greg Kroah-Hartman wrote:
> On Wed, Nov 24, 2021 at 11:33:11AM -0600, Serge E. Hallyn wrote:
> > On Wed, Nov 24, 2021 at 04:31:22PM +0100, Greg Kroah-Hartman wrote:
> > > On Wed, Nov 24, 2021 at 04:22:50PM +0200, Jari Ruusu wrote:
> > > > Greg Kroah-Hartman wrote:
> > > > > From: Alistair Delva <[email protected]>
> > > > >
> > > > > commit 94c4b4fd25e6c3763941bdec3ad54f2204afa992 upstream.
> > > > [SNIP]
> > > > > --- a/block/ioprio.c
> > > > > +++ b/block/ioprio.c
> > > > > @@ -69,7 +69,14 @@ int ioprio_check_cap(int ioprio)
> > > > >
> > > > > switch (class) {
> > > > > case IOPRIO_CLASS_RT:
> > > > > - if (!capable(CAP_SYS_NICE) && !capable(CAP_SYS_ADMIN))
> > > > > + /*
> > > > > + * Originally this only checked for CAP_SYS_ADMIN,
> > > > > + * which was implicitly allowed for pid 0 by security
> > > > > + * modules such as SELinux. Make sure we check
> > > > > + * CAP_SYS_ADMIN first to avoid a denial/avc for
> > > > > + * possibly missing CAP_SYS_NICE permission.
> > > > > + */
> > > > > + if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_NICE))
> > > > > return -EPERM;
> > > > > fallthrough;
> > > > > /* rt has prio field too */
> > > >
> > > > What exactly is above patch trying to fix?
> > > > It does not change control flow at all, and added comment is misleading.
> > >
> > > See the thread on the mailing list for what it does and why it is
> > > needed.
> > >
> > > It does change the result when selinux is enabled.
> > >
> > > thanks,
> > >
> > > greg k-h
> >
> > The case where we create a newer more fine grained capability which is a
> > sub-cap of a broader capability like CAP_SYS_ADMIN is analogous. See
> > check_syslog_permissions() for instance.
> >
> > So I think a helper like
> >
> > int capable_either_or(int cap1, int cap2) {
> > if (has_capability_noaudit(current, cap1))
> > return 0;
> > return capable(cap2);
> > }
> >
> > might be worthwhile.
>
> Sure, feel free to work on that and submit it, but for now, this change
> is needed.

Sorry I misread the subject and thought this was just a resubmission.

FWIW I had acked an earlier version of this.

-serge

2021-11-24 23:30:57

by Serge E. Hallyn

Subject: Re: [PATCH 5.10 130/154] block: Check ADMIN before NICE for IOPRIO_CLASS_RT

On Wed, Nov 24, 2021 at 07:34:50PM +0100, Christian G?ttsche wrote:
> On Wed, 24 Nov 2021 at 19:16, Greg Kroah-Hartman
> <[email protected]> wrote:
> >
> > On Wed, Nov 24, 2021 at 11:33:11AM -0600, Serge E. Hallyn wrote:
> > > On Wed, Nov 24, 2021 at 04:31:22PM +0100, Greg Kroah-Hartman wrote:
> > > > On Wed, Nov 24, 2021 at 04:22:50PM +0200, Jari Ruusu wrote:
> > > > > Greg Kroah-Hartman wrote:
> > > > > > From: Alistair Delva <[email protected]>
> > > > > >
> > > > > > commit 94c4b4fd25e6c3763941bdec3ad54f2204afa992 upstream.
> > > > > [SNIP]
> > > > > > --- a/block/ioprio.c
> > > > > > +++ b/block/ioprio.c
> > > > > > @@ -69,7 +69,14 @@ int ioprio_check_cap(int ioprio)
> > > > > >
> > > > > > switch (class) {
> > > > > > case IOPRIO_CLASS_RT:
> > > > > > - if (!capable(CAP_SYS_NICE) && !capable(CAP_SYS_ADMIN))
> > > > > > + /*
> > > > > > + * Originally this only checked for CAP_SYS_ADMIN,
> > > > > > + * which was implicitly allowed for pid 0 by security
> > > > > > + * modules such as SELinux. Make sure we check
> > > > > > + * CAP_SYS_ADMIN first to avoid a denial/avc for
> > > > > > + * possibly missing CAP_SYS_NICE permission.
> > > > > > + */
> > > > > > + if (!capable(CAP_SYS_ADMIN) && !capable(CAP_SYS_NICE))
> > > > > > return -EPERM;
> > > > > > fallthrough;
> > > > > > /* rt has prio field too */
> > > > >
> > > > > What exactly is above patch trying to fix?
> > > > > It does not change control flow at all, and added comment is misleading.
> > > >
> > > > See the thread on the mailing list for what it does and why it is
> > > > needed.
> > > >
> > > > It does change the result when selinux is enabled.
> > > >
> > > > thanks,
> > > >
> > > > greg k-h
> > >
> > > The case where we create a newer more fine grained capability which is a
> > > sub-cap of a broader capability like CAP_SYS_ADMIN is analogous. See
> > > check_syslog_permissions() for instance.
> > >
> > > So I think a helper like
> > >
> > > int capable_either_or(int cap1, int cap2) {
> > > if (has_capability_noaudit(current, cap1))
> > > return 0;
> > > return capable(cap2);
> > > }
> > >
> > > might be worthwhile.
> >
>
> I proposed an early prototype at
> https://patchwork.kernel.org/project/selinux/patch/[email protected]/

I never saw this. Would you mind resending as a standalone patch?

(I do have comments, but this thread seems the wrong place)