x86/hotplug: Silence APIC only after all interrupts are migrated
authorAshok Raj <ashok.raj@intel.com>
Thu, 27 Aug 2020 04:12:10 +0000 (21:12 -0700)
committerThomas Gleixner <tglx@linutronix.de>
Thu, 27 Aug 2020 07:29:23 +0000 (09:29 +0200)
commit52d6b926aabc47643cd910c85edb262b7f44c168
treebb0fb5cfac29b8aa66984718cd6cb2b74ac63a9b
parentd4f07268d035721dd055ceb0de98ace6ac5f858b
x86/hotplug: Silence APIC only after all interrupts are migrated

There is a race when taking a CPU offline. Current code looks like this:

native_cpu_disable()
{
...
apic_soft_disable();
/*
 * Any existing set bits for pending interrupt to
 * this CPU are preserved and will be sent via IPI
 * to another CPU by fixup_irqs().
 */
cpu_disable_common();
{
....
/*
 * Race window happens here. Once local APIC has been
 * disabled any new interrupts from the device to
 * the old CPU are lost
 */
fixup_irqs(); // Too late to capture anything in IRR.
...
}
}

The fix is to disable the APIC *after* cpu_disable_common().

Testing was done with a USB NIC that provided a source of frequent
interrupts. A script migrated interrupts to a specific CPU and
then took that CPU offline.

Fixes: 60dcaad5736f ("x86/hotplug: Silence APIC and NMI when CPU is dead")
Reported-by: Evan Green <evgreen@chromium.org>
Signed-off-by: Ashok Raj <ashok.raj@intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Tested-by: Evan Green <evgreen@chromium.org>
Reviewed-by: Evan Green <evgreen@chromium.org>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/lkml/875zdarr4h.fsf@nanos.tec.linutronix.de/
Link: https://lore.kernel.org/r/1598501530-45821-1-git-send-email-ashok.raj@intel.com
arch/x86/kernel/smpboot.c