Xen devel - Aug 2012 - [PATCH] arm: fix gic_init_secondary_cpu.

Using spin_lock_irq here is unnecessary (interrupts are not yet enabled) and
wrong (since they will get unexpectedly renabled by spin_unlock_irq).

We can just use spin_lock/spin_unlock.

Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Cc: tim@xen.org
---

Now an SMP model gets as far as hanging at dom0 "Calibrating delay
loop...".

Tim, didn''t you diagnose that a while ago?
---
 xen/arch/arm/gic.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/xen/arch/arm/gic.c b/xen/arch/arm/gic.c
index 378158b..6f5b0e1 100644
--- a/xen/arch/arm/gic.c
+++ b/xen/arch/arm/gic.c
@@ -323,10 +323,10 @@ int __init gic_init(void)
 /* Set up the per-CPU parts of the GIC for a secondary CPU */
 void __cpuinit gic_init_secondary_cpu(void)
 {
-    spin_lock_irq(&gic.lock);
+    spin_lock(&gic.lock);
     gic_cpu_init();
     gic_hyp_init();
-    spin_unlock_irq(&gic.lock);
+    spin_unlock(&gic.lock);
 }
 
 /* Shut down the per-CPU GIC interface */
-- 
1.7.9.1

At 16:43 +0000 on 01 Aug (1343839437), Ian Campbell
wrote:
> Using spin_lock_irq here is unnecessary (interrupts are not yet enabled)
and
> wrong (since they will get unexpectedly renabled by spin_unlock_irq).
> 
> We can just use spin_lock/spin_unlock.
> 
> Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Acked-by: Tim Deegan <tim@xen.org>
> Now an SMP model gets as far as hanging at dom0 "Calibrating delay
loop...".
> 
> Tim, didn''t you diagnose that a while ago?
I only got as far as showing that dom0 is spinning waiting for jiffies
to increase, but never taking the timer interrupts that would cause that
to happen.

Tim.

On Thu, 2012-08-02 at 12:09 +0100, Tim Deegan wrote:
> Acked-by: Tim Deegan <tim@xen.org>
Applied, thanks.

Ian.