Kumar Gala writes:

Can you hold off on the this.  The following patch is still under  
discussion and the problem exists on more than just PPC.

The patch that I put in fixes the problem in the situation where it
was found, right?  That the problem may exist on other architectures
isn't a reason not to fix it here.

quoted

Scott Wood:
      [POWERPC] IPIC: Fix spinlock recursion in set_irq_handler

Paul.