x86/ioapic: Force affinity setting in setup_ioapic_dest()
authorThomas Gleixner <tglx@linutronix.de>
Mon, 14 Sep 2015 10:00:55 +0000 (12:00 +0200)
committerThomas Gleixner <tglx@linutronix.de>
Mon, 14 Sep 2015 16:28:15 +0000 (18:28 +0200)
commit4857c91f0d195f05908fff296ba1ec5fca87066c
treeb82615a3360e3003cce9fdf11d5a2fa0e3f5f6de
parentcda34fc774d114afe98515a21c2063a803f922bc
x86/ioapic: Force affinity setting in setup_ioapic_dest()

The recent ioapic cleanups changed the affinity setting in
setup_ioapic_dest() from a direct write to the hardware to the delayed
affinity setup via irq_set_affinity().

That results in a warning from chained_irq_exit():
WARNING: CPU: 0 PID: 5 at kernel/irq/migration.c:32 irq_move_masked_irq
[<ffffffff810a0a88>] irq_move_masked_irq+0xb8/0xc0
[<ffffffff8103c161>] ioapic_ack_level+0x111/0x130
[<ffffffff812bbfe8>] intel_gpio_irq_handler+0x148/0x1c0

The reason is that irq_set_affinity() does not write directly to the
hardware. It marks the affinity setting as pending and executes it
from the next interrupt. The chained handler infrastructure does not
take the irq descriptor lock for performance reasons because such a
chained interrupt is not visible to any interfaces. So the delayed
affinity setting triggers the warning in irq_move_masked_irq().

Restore the old behaviour by calling the set_affinity function of the
ioapic chip in setup_ioapic_dest(). This is safe as none of the
interrupts can be on the fly at this point.

Fixes: aa5cb97f14a2 'x86/irq: Remove x86_io_apic_ops.set_affinity and related interfaces'
Reported-and-tested-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Jiang Liu <jiang.liu@linux.intel.com>
Cc: jarkko.nikula@linux.intel.com
arch/x86/kernel/apic/io_apic.c