[PATCH] perf: Require CAP_KILL if sigtrap is requested

[Ondrej Mosnacek]

On Wed, Jun 30, 2021 at 11:38 AM Marco Elver <elver at google.com> wrote:
> If perf_event_open() is called with another task as target and
> perf_event_attr::sigtrap is set, and the target task's user does not
> match the calling user, also require the CAP_KILL capability.
>
> Otherwise, with the CAP_PERFMON capability alone it would be possible
> for a user to send SIGTRAP signals via perf events to another user's
> tasks. This could potentially result in those tasks being terminated if
> they cannot handle SIGTRAP signals.
>
> Fixes: 97ba62b27867 ("perf: Add support for SIGTRAP on perf events")
> Reported-by: Dmitry Vyukov <dvyukov at google.com>
> Signed-off-by: Marco Elver <elver at google.com>
> ---
>  include/linux/capability.h |  5 +++++
>  kernel/events/core.c       | 13 ++++++++++++-
>  2 files changed, 17 insertions(+), 1 deletion(-)
>
> diff --git a/include/linux/capability.h b/include/linux/capability.h
> index 65efb74c3585..1c6be4743dbe 100644
> --- a/include/linux/capability.h
> +++ b/include/linux/capability.h
> @@ -264,6 +264,11 @@ static inline bool bpf_capable(void)
>         return capable(CAP_BPF) || capable(CAP_SYS_ADMIN);
>  }
>
> +static inline bool kill_capable(void)
> +{
> +       return capable(CAP_KILL) || capable(CAP_SYS_ADMIN);

Is it really necessary to fall back to CAP_SYS_ADMIN here? CAP_PERFMON
and CAP_BPF have been split off from CAP_SYS_ADMIN recently, so they
have it for backwards compatibility. You are adding a new restriction
for a very specific action, so I don't think the fallback is needed.

> +}
> +
>  static inline bool checkpoint_restore_ns_capable(struct user_namespace *ns)
>  {
>         return ns_capable(ns, CAP_CHECKPOINT_RESTORE) ||
> diff --git a/kernel/events/core.c b/kernel/events/core.c
> index fe88d6eea3c2..1ab4bc867531 100644
> --- a/kernel/events/core.c
> +++ b/kernel/events/core.c
> @@ -12152,10 +12152,21 @@ SYSCALL_DEFINE5(perf_event_open,
>         }
>
>         if (task) {
> +               bool is_capable;
> +
>                 err = down_read_interruptible(&task->signal->exec_update_lock);
>                 if (err)
>                         goto err_file;
>
> +               is_capable = perfmon_capable();
> +               if (attr.sigtrap) {
> +                       /*
> +                        * perf_event_attr::sigtrap sends signals to the other
> +                        * task. Require the current task to have CAP_KILL.
> +                        */
> +                       is_capable &= kill_capable();

Is it necessary to do all this dance just to call perfmon_capable()
first? Couldn't this be simply:

err = -EPERM;
if (attr.sigtrap && !capable(CAP_KILL))
        goto err_cred;

Also, looking at kill_ok_by_cred() in kernel/signal.c, would it
perhaps be more appropriate to do
ns_capable(__task_cred(task)->user_ns, CAP_KILL) instead? (There might
also need to be some careful locking around getting the target task's
creds - I'm not sure...)

> +               }
> +
>                 /*
>                  * Preserve ptrace permission check for backwards compatibility.
>                  *
> @@ -12165,7 +12176,7 @@ SYSCALL_DEFINE5(perf_event_open,
>                  * perf_event_exit_task() that could imply).
>                  */
>                 err = -EACCES;

BTW, shouldn't this (and several other such cases in this file...)
actually be EPERM, as is the norm for capability checks?

> -               if (!perfmon_capable() && !ptrace_may_access(task, PTRACE_MODE_READ_REALCREDS))
> +               if (!is_capable && !ptrace_may_access(task, PTRACE_MODE_READ_REALCREDS))
>                         goto err_cred;
>         }
>
> --
> 2.32.0.93.g670b81a890-goog
>

-- 
Ondrej Mosnacek
Software Engineer, Linux Security - SELinux kernel
Red Hat, Inc.