[PATCH] __ptrace_may_access() should not deny sub-threads

[Oleg Nesterov <oleg@redhat.com>]

From: Mark Grondona <mgrondona@llnl.gov>

__ptrace_may_access() checks get_dumpable/ptrace_has_cap/etc
if task != current, this can can lead to surprising results.

For example, a sub-thread can't readlink("/proc/self/exe") if
the executable is not readable. setup_new_exec()->would_dump()
notices that inode_permission(MAY_READ) fails and then it does
set_dumpable(suid_dumpable). After that get_dumpable() fails.

(It is not clear why proc_pid_readlink() checks get_dumpable(),
 perhaps we could add PTRACE_MODE_NODUMPABLE)

Change __ptrace_may_access() to use same_thread_group() instead
of "task == current". Any security check is pointless when the
tasks share the same ->mm.

Signed-off-by: Mark Grondona <mgrondona@llnl.gov>
Signed-off-by: Ben Woodard <woodard@redhat.com>
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
---
 kernel/ptrace.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/kernel/ptrace.c b/kernel/ptrace.c
index a146ee3..dd562e9 100644
--- a/kernel/ptrace.c
+++ b/kernel/ptrace.c
@@ -236,7 +236,7 @@ static int __ptrace_may_access(struct task_struct *task, unsigned int mode)
 	 */
 	int dumpable = 0;
 	/* Don't let security modules deny introspection */
-	if (task == current)
+	if (same_thread_group(task, current))
 		return 0;
 	rcu_read_lock();
 	tcred = __task_cred(task);
-- 
1.5.5.1