suggestion from jameson's review: move wait for GC to before jl_exit_thread0 call

Author: d-netto

src/signals-unix.c

@@ -1033,10 +1033,6 @@ static void *signal_listener(void *arg)
         }
 #endif
         if (doexit) {
-            // Let's forbid threads from running GC while we're trying to exit,
-            // also let's make sure we're not in the middle of GC.
-            jl_atomic_fetch_add(&jl_gc_disable_counter, 1);
-            jl_safepoint_wait_gc(NULL);
             // The exit can get stuck if it happens at an unfortunate spot in thread 0
             // (unavoidable due to its async nature).
             // Try much harder to exit next time, if we get multiple exit requests.
@@ -1093,6 +1089,10 @@ static void *signal_listener(void *arg)
 //#if defined(_POSIX_C_SOURCE) && _POSIX_C_SOURCE >= 199309L && !HAVE_KEVENT
 //            si_code = info.si_code;
 //#endif
+            // Let's forbid threads from running GC while we're trying to exit,
+            // also let's make sure we're not in the middle of GC.
+            jl_atomic_fetch_add(&jl_gc_disable_counter, 1);
+            jl_safepoint_wait_gc(NULL);
             jl_exit_thread0(sig, signal_bt_data, signal_bt_size);
         }
         else if (critical) {