x86/stacktrace: Make zombie stack traces reliable

Commit:

  1959a60182f4 ("x86/dumpstack: Pin the target stack when dumping it")

changed the behavior of stack traces for zombies.  Before that commit,
/proc/<pid>/stack reported the last execution path of the zombie before
it died:

  [<ffffffff8105b877>] do_exit+0x6f7/0xa80
  [<ffffffff8105bc79>] do_group_exit+0x39/0xa0
  [<ffffffff8105bcf0>] __wake_up_parent+0x0/0x30
  [<ffffffff8152dd09>] system_call_fastpath+0x16/0x1b
  [<00007fd128f9c4f9>] 0x7fd128f9c4f9
  [<ffffffffffffffff>] 0xffffffffffffffff

After the commit, it just reports an empty stack trace.

The new behavior is actually probably more correct.  If the stack
refcount has gone down to zero, then the task has already gone through
do_exit() and isn't going to run anymore.  The stack could be freed at
any time and is basically gone, so reporting an empty stack makes sense.

However, save_stack_trace_tsk_reliable() treats such a missing stack
condition as an error.  That can cause livepatch transition stalls if
there are any unreaped zombies.  Instead, just treat it as a reliable,
empty stack.

Reported-and-tested-by: Miroslav Benes <mbenes@suse.cz>
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: live-patching@vger.kernel.org
Fixes: af085d9084b4 ("stacktrace/x86: add function for detecting reliable stack traces")
Link: http://lkml.kernel.org/r/e4b09e630e99d0c1080528f0821fc9d9dbaeea82.1513631620.git.jpoimboe@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>

1 files changed, 5 insertions, 1 deletions

diff --git a/arch/x86/kernel/stacktrace.c b/arch/x86/kernel/stacktrace.c
index 77835bc..20161ef 100644
--- a/arch/x86/kernel/stacktrace.c
+++ b/arch/x86/kernel/stacktrace.c
@@ -164,8 +164,12 @@ int save_stack_trace_tsk_reliable(struct task_struct *tsk,
 {
 	int ret;
 
+	/*
+	 * If the task doesn't have a stack (e.g., a zombie), the stack is
+	 * "reliably" empty.
+	 */
 	if (!try_get_task_stack(tsk))
-		return -EINVAL;
+		return 0;
 
 	ret = __save_stack_trace_reliable(trace, tsk);