re PR go/59408 (Many Go tests FAIL with notesleep not on g0)
authorIan Lance Taylor <ian@gcc.gnu.org>
Fri, 6 Dec 2013 18:26:27 +0000 (18:26 +0000)
committerIan Lance Taylor <ian@gcc.gnu.org>
Fri, 6 Dec 2013 18:26:27 +0000 (18:26 +0000)
PR go/59408
runtime: Don't require g != m->g0 in sema notesleep.

From-SVN: r205756

libgo/runtime/lock_sema.c

index ce435119323fdcc9a58882eec3f9b1ecb706da33..000b9fcf7010349ec76e73a189f1cd46f78d8834 100644 (file)
@@ -152,8 +152,12 @@ runtime_notesleep(Note *n)
 
        m = runtime_m();
 
+  /* For gccgo it's OK to sleep in non-g0, and it happens in
+     stoptheworld because we have not implemented preemption.
+
        if(runtime_g() != m->g0)
                runtime_throw("notesleep not on g0");
+  */
 
        if(m->waitsema == 0)
                m->waitsema = runtime_semacreate();