/proc/pid/status: add "Seccomp" field
authorKees Cook <keescook@chromium.org>
Tue, 18 Dec 2012 00:03:14 +0000 (16:03 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Tue, 18 Dec 2012 01:15:22 +0000 (17:15 -0800)
It is currently impossible to examine the state of seccomp for a given
process.  While attaching with gdb and attempting "call
prctl(PR_GET_SECCOMP,...)" will work with some situations, it is not
reliable.  If the process is in seccomp mode 1, this query will kill the
process (prctl not allowed), if the process is in mode 2 with prctl not
allowed, it will similarly be killed, and in weird cases, if prctl is
filtered to return errno 0, it can look like seccomp is disabled.

When reviewing the state of running processes, there should be a way to
externally examine the seccomp mode.  ("Did this build of Chrome end up
using seccomp?" "Did my distro ship ssh with seccomp enabled?")

This adds the "Seccomp" line to /proc/$pid/status.

Signed-off-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Cyrill Gorcunov <gorcunov@openvz.org>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: James Morris <jmorris@namei.org>
Acked-by: Serge E. Hallyn <serge.hallyn@ubuntu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Documentation/filesystems/proc.txt
fs/proc/array.c

index 74cb394..12665ee 100644 (file)
@@ -181,6 +181,7 @@ read the file /proc/PID/status:
   CapPrm: 0000000000000000
   CapEff: 0000000000000000
   CapBnd: ffffffffffffffff
+  Seccomp:        0
   voluntary_ctxt_switches:        0
   nonvoluntary_ctxt_switches:     1
 
@@ -237,6 +238,7 @@ Table 1-2: Contents of the status files (as of 2.6.30-rc7)
  CapPrm                      bitmap of permitted capabilities
  CapEff                      bitmap of effective capabilities
  CapBnd                      bitmap of capabilities bounding set
+ Seccomp                     seccomp mode, like prctl(PR_GET_SECCOMP, ...)
  Cpus_allowed                mask of CPUs on which this process may run
  Cpus_allowed_list           Same as previous, but in "list format"
  Mems_allowed                mask of memory nodes allowed to this process
index 377a373..077235f 100644 (file)
@@ -336,6 +336,13 @@ static inline void task_cap(struct seq_file *m, struct task_struct *p)
        render_cap_t(m, "CapBnd:\t", &cap_bset);
 }
 
+static inline void task_seccomp(struct seq_file *m, struct task_struct *p)
+{
+#ifdef CONFIG_SECCOMP
+       seq_printf(m, "Seccomp:\t%d\n", p->seccomp.mode);
+#endif
+}
+
 static inline void task_context_switch_counts(struct seq_file *m,
                                                struct task_struct *p)
 {
@@ -369,6 +376,7 @@ int proc_pid_status(struct seq_file *m, struct pid_namespace *ns,
        }
        task_sig(m, task);
        task_cap(m, task);
+       task_seccomp(m, task);
        task_cpus_allowed(m, task);
        cpuset_task_status_allowed(m, task);
        task_context_switch_counts(m, task);