2019-01-17 00:00:50

by Kees Cook

Subject: [PATCH] Yama: Check for pid death before checking ancestry

It's possible that a pid has died before we take the rcu lock, in which
case we can't walk the ancestry list as it may be detached. Instead, check
for death first before doing the walk.

Reported-by: [email protected]
Fixes: 2d514487faf1 ("security: Yama LSM")
Cc: [email protected]
Suggested-by: Oleg Nesterov <[email protected]>
Signed-off-by: Kees Cook <[email protected]>
---
James, can you please send this to Linus in your -fixes tree?
---
security/yama/yama_lsm.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/security/yama/yama_lsm.c b/security/yama/yama_lsm.c
index ffda91a4a1aa..02514fe558b4 100644
--- a/security/yama/yama_lsm.c
+++ b/security/yama/yama_lsm.c
@@ -368,7 +368,9 @@ static int yama_ptrace_access_check(struct task_struct *child,
break;
case YAMA_SCOPE_RELATIONAL:
rcu_read_lock();
- if (!task_is_descendant(current, child) &&
+ if (!pid_alive(child))
+ rc = -EPERM;
+ if (!rc && !task_is_descendant(current, child) &&
!ptracer_exception_found(current, child) &&
!ns_capable(__task_cred(child)->user_ns, CAP_SYS_PTRACE))
rc = -EPERM;
--
2.17.1


--
Kees Cook

2019-01-17 00:35:39

by James Morris

Subject: Re: [PATCH] Yama: Check for pid death before checking ancestry

On Wed, 16 Jan 2019, Kees Cook wrote:

> It's possible that a pid has died before we take the rcu lock, in which
> case we can't walk the ancestry list as it may be detached. Instead, check
> for death first before doing the walk.
>
> Reported-by: [email protected]
> Fixes: 2d514487faf1 ("security: Yama LSM")
> Cc: [email protected]
> Suggested-by: Oleg Nesterov <[email protected]>
> Signed-off-by: Kees Cook <[email protected]>
> ---
> James, can you please send this to Linus in your -fixes tree?

Done.

> ---
> security/yama/yama_lsm.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/security/yama/yama_lsm.c b/security/yama/yama_lsm.c
> index ffda91a4a1aa..02514fe558b4 100644
> --- a/security/yama/yama_lsm.c
> +++ b/security/yama/yama_lsm.c
> @@ -368,7 +368,9 @@ static int yama_ptrace_access_check(struct task_struct *child,
> break;
> case YAMA_SCOPE_RELATIONAL:
> rcu_read_lock();
> - if (!task_is_descendant(current, child) &&
> + if (!pid_alive(child))
> + rc = -EPERM;
> + if (!rc && !task_is_descendant(current, child) &&
> !ptracer_exception_found(current, child) &&
> !ns_capable(__task_cred(child)->user_ns, CAP_SYS_PTRACE))
> rc = -EPERM;
>

--
James Morris
<[email protected]>

2019-01-17 00:36:22

by James Morris

Subject: Re: [PATCH] Yama: Check for pid death before checking ancestry

On Thu, 17 Jan 2019, James Morris wrote:

> On Wed, 16 Jan 2019, Kees Cook wrote:
>
> > It's possible that a pid has died before we take the rcu lock, in which
> > case we can't walk the ancestry list as it may be detached. Instead, check
> > for death first before doing the walk.
> >
> > Reported-by: [email protected]
> > Fixes: 2d514487faf1 ("security: Yama LSM")
> > Cc: [email protected]
> > Suggested-by: Oleg Nesterov <[email protected]>
> > Signed-off-by: Kees Cook <[email protected]>
> > ---
> > James, can you please send this to Linus in your -fixes tree?
>
> Done.

I mean, queued in that tree.

>
> > ---
> > security/yama/yama_lsm.c | 4 +++-
> > 1 file changed, 3 insertions(+), 1 deletion(-)
> >
> > diff --git a/security/yama/yama_lsm.c b/security/yama/yama_lsm.c
> > index ffda91a4a1aa..02514fe558b4 100644
> > --- a/security/yama/yama_lsm.c
> > +++ b/security/yama/yama_lsm.c
> > @@ -368,7 +368,9 @@ static int yama_ptrace_access_check(struct task_struct *child,
> > break;
> > case YAMA_SCOPE_RELATIONAL:
> > rcu_read_lock();
> > - if (!task_is_descendant(current, child) &&
> > + if (!pid_alive(child))
> > + rc = -EPERM;
> > + if (!rc && !task_is_descendant(current, child) &&
> > !ptracer_exception_found(current, child) &&
> > !ns_capable(__task_cred(child)->user_ns, CAP_SYS_PTRACE))
> > rc = -EPERM;
> >
>
>

--
James Morris
<[email protected]>