Skip to content

Commit

Permalink
xen/smp/pvhvm: Don't initialize IRQ_WORKER as we are using the native…
Browse files Browse the repository at this point in the history
… one.

There is no need to use the PV version of the IRQ_WORKER mechanism
as under PVHVM we are using the native version. The native
version is using the SMP API.

They just sit around unused:

  69:          0          0  xen-percpu-ipi       irqwork0
  83:          0          0  xen-percpu-ipi       irqwork1

Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
  • Loading branch information
Konrad Rzeszutek Wilk committed Apr 16, 2013
1 parent 70dd499 commit 27d8b20
Showing 1 changed file with 12 additions and 1 deletion.
13 changes: 12 additions & 1 deletion arch/x86/xen/smp.c
Original file line number Diff line number Diff line change
Expand Up @@ -144,6 +144,13 @@ static int xen_smp_intr_init(unsigned int cpu)
goto fail;
per_cpu(xen_callfuncsingle_irq, cpu) = rc;

/*
* The IRQ worker on PVHVM goes through the native path and uses the
* IPI mechanism.
*/
if (xen_hvm_domain())
return 0;

callfunc_name = kasprintf(GFP_KERNEL, "irqwork%d", cpu);
rc = bind_ipi_to_irqhandler(XEN_IRQ_WORK_VECTOR,
cpu,
Expand All @@ -167,6 +174,9 @@ static int xen_smp_intr_init(unsigned int cpu)
if (per_cpu(xen_callfuncsingle_irq, cpu) >= 0)
unbind_from_irqhandler(per_cpu(xen_callfuncsingle_irq, cpu),
NULL);
if (xen_hvm_domain())
return rc;

if (per_cpu(xen_irq_work, cpu) >= 0)
unbind_from_irqhandler(per_cpu(xen_irq_work, cpu), NULL);

Expand Down Expand Up @@ -661,7 +671,8 @@ static void xen_hvm_cpu_die(unsigned int cpu)
unbind_from_irqhandler(per_cpu(xen_callfunc_irq, cpu), NULL);
unbind_from_irqhandler(per_cpu(xen_debug_irq, cpu), NULL);
unbind_from_irqhandler(per_cpu(xen_callfuncsingle_irq, cpu), NULL);
unbind_from_irqhandler(per_cpu(xen_irq_work, cpu), NULL);
if (!xen_hvm_domain())
unbind_from_irqhandler(per_cpu(xen_irq_work, cpu), NULL);
xen_uninit_lock_cpu(cpu);
xen_teardown_timer(cpu);
native_cpu_die(cpu);
Expand Down

0 comments on commit 27d8b20

Please sign in to comment.