x86/apic: Optimize cpu traversal in __assign_irq_vector() using domain membership
authorSuresh Siddha <suresh.b.siddha@intel.com>
Mon, 25 Jun 2012 20:38:27 +0000 (13:38 -0700)
committerIngo Molnar <mingo@kernel.org>
Fri, 6 Jul 2012 09:00:21 +0000 (11:00 +0200)
commitb39f25a849d7677a7dbf183f2483fd41c201a5ce
tree87f06060976461b91d5cb196f0ed04fbcea0bb26
parentabf71f3066740f3b59c3f731b4b68ed335f7b24d
x86/apic: Optimize cpu traversal in __assign_irq_vector() using domain membership

Currently __assign_irq_vector() goes through each cpu in the
specified mask until it finds a free vector in all the cpu's
that are part of the same interrupt domain. We visit all the
interrupt domain sibling cpus to reserve the free vector. So,
when we fail to find a free vector in an interrupt domain, it is
safe to continue our search with a cpu belonging to a new
interrupt domain. No need to go through each cpu, if the domain
containing that cpu is already visited.

Use the irq_cfg's old_domain to track the visited domains and
optimize the cpu traversal while finding a free vector in the
given cpumask.

NOTE: We can also optimize the search by using for_each_cpu() and
skip the current cpu, if it is not the first cpu in the mask
returned by the vector_allocation_domain(). But re-using the
cfg->old_domain to track the visited domains will be slightly
faster.

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Acked-by: Yinghai Lu <yinghai@kernel.org>
Acked-by: Alexander Gordeev <agordeev@redhat.com>
Acked-by: Cyrill Gorcunov <gorcunov@openvz.org>
Link: http://lkml.kernel.org/r/1340656709-11423-2-git-send-email-suresh.b.siddha@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
arch/x86/include/asm/apic.h
arch/x86/kernel/apic/apic_noop.c
arch/x86/kernel/apic/io_apic.c
arch/x86/kernel/apic/x2apic_cluster.c
arch/x86/kernel/vsmp_64.c