xen/spinlock: Don't enable them unconditionally.
authorKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Fri, 4 Apr 2014 18:48:04 +0000 (14:48 -0400)
committerDavid Vrabel <david.vrabel@citrix.com>
Tue, 15 Apr 2014 16:41:28 +0000 (17:41 +0100)
The git commit a945928ea2709bc0e8e8165d33aed855a0110279
('xen: Do not enable spinlocks before jump_label_init() has executed')
was added to deal with the jump machinery. Earlier the code
that turned on the jump label was only called by Xen specific
functions. But now that it had been moved to the initcall machinery
it gets called on Xen, KVM, and baremetal - ouch!. And the detection
machinery to only call it on Xen wasn't remembered in the heat
of merge window excitement.

This means that the slowpath is enabled on baremetal while it should
not be.

Reported-by: Waiman Long <waiman.long@hp.com>
Acked-by: Steven Rostedt <rostedt@goodmis.org>
CC: stable@vger.kernel.org
CC: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
arch/x86/xen/spinlock.c

index 4d3acc34a998e5aca97d58c9573e822a29905626..0ba5f3b967f0041c5ea6445f898a944008b96c58 100644 (file)
@@ -274,7 +274,7 @@ void __init xen_init_spinlocks(void)
                printk(KERN_DEBUG "xen: PV spinlocks disabled\n");
                return;
        }
-
+       printk(KERN_DEBUG "xen: PV spinlocks enabled\n");
        pv_lock_ops.lock_spinning = PV_CALLEE_SAVE(xen_lock_spinning);
        pv_lock_ops.unlock_kick = xen_unlock_kick;
 }
@@ -290,6 +290,9 @@ static __init int xen_init_spinlocks_jump(void)
        if (!xen_pvspin)
                return 0;
 
+       if (!xen_domain())
+               return 0;
+
        static_key_slow_inc(&paravirt_ticketlocks_enabled);
        return 0;
 }