Introduce crash_dump debugging helper.

Replace the global debuggerd with a per-process debugging helper that
gets exec'ed by the process that crashed.

Bug: http://b/30705528
Test: crasher/crasher64, `debuggerd <pid>`, `kill -ABRT <pid>`
Change-Id: Iad1b7478f7a4e2690720db4b066417d8b66834ed
diff --git a/private/system_server.te b/private/system_server.te
index 7e0c2fb..0e35660 100644
--- a/private/system_server.te
+++ b/private/system_server.te
@@ -2,7 +2,7 @@
 # public, but conceptually should go with this
 # Define a type for tmpfs-backed ashmem regions.
 tmpfs_domain(system_server)
-# Create a socket for connections from debuggerd.
+# Create a socket for connections from crash_dump.
 type_transition system_server system_data_file:sock_file system_ndebug_socket "ndebugsocket";
 
 allow system_server zygote_tmpfs:file read;