[Bug 225450] 11.1-* panics on AMD Opteron 2k due to EARLY_AP_STARTUP

[bugzilla-noreply at freebsd.org]

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=225450

--- Comment #16 from John Baldwin <jhb at FreeBSD.org> ---
(In reply to Pablo Ruiz from comment #14)
Thanks, this gives me what I was looking for.  We have infinite recursion
because the spin lock code is calling DELAY() which is trying to grab a spin
lock.

One question is why isn't DELAY using the tsc?

Hmm, it seems 'tsc_is_invariant' isn't set.  Are these older AMD CPUs?

You can try this as a hack-workaround to verify it fixes the issue, but I need
to think a bit more about what the right fix might be:

Index: x86/x86/delay.c
===================================================================
--- delay.c     (revision 329004)
+++ delay.c     (working copy)
@@ -72,7 +72,7 @@
                func = get_tsc;
                mask = ~0u;
        } else {
-               if (tc->tc_quality <= 0)
+               if (tc->tc_quality <= 0 || n == 1)
                        return (0);
                func = tc->tc_get_timecount;
                mask = tc->tc_counter_mask;

-- 
You are receiving this mail because:
You are the assignee for the bug.