Messages in this thread | | | Date | Mon, 9 Jun 2014 16:46:11 -0700 | Subject | Re: [Patch V4 01/42] x86, irq: update high address field when updating affinity for MSI IRQ | From | Yinghai Lu <> |
| |
On Mon, Jun 9, 2014 at 1:19 AM, Jiang Liu <jiang.liu@linux.intel.com> wrote: > If x2apic is enabled, the MSI high address field should also be aslo > updated when setting affinity for MSI IRQ, otherwise the MSI IRQ may > target wrong APIC IDs.
Do you have any test case to reveal the problem?
> > Signed-off-by: Jiang Liu <jiang.liu@linux.intel.com> > Cc: stable@vger.kernel.org > --- > arch/x86/kernel/apic/io_apic.c | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c > index 9d0a9795a0f8..2de992501a1b 100644 > --- a/arch/x86/kernel/apic/io_apic.c > +++ b/arch/x86/kernel/apic/io_apic.c > @@ -3007,6 +3007,10 @@ msi_set_affinity(struct irq_data *data, const struct cpumask *mask, bool force) > > __get_cached_msi_msg(data->msi_desc, &msg); > > + msg.address_hi = MSI_ADDR_BASE_HI; > + if (x2apic_enabled()) > + msg.address_hi |= MSI_ADDR_EXT_DEST_ID(dest); > + > msg.data &= ~MSI_DATA_VECTOR_MASK; > msg.data |= MSI_DATA_VECTOR(cfg->vector); > msg.address_lo &= ~MSI_ADDR_DEST_ID_MASK;
No. This change is not needed.
When x2apic is used, and apicid > 255, irq remapping is used.
msi_chip.irq_set_affinity will be changed from msi_set_affinity to x86_io_apic_ops.set_affinity (aka intel_setup_ioapic_entry), via irq_remap_modify_chip_defaults().
Thanks
Yinghai
| |