arm64, debug-monitors: Fix CPU hotplug callback registration
authorSrivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Mon, 10 Mar 2014 20:39:20 +0000 (02:09 +0530)
committerMark Brown <broonie@linaro.org>
Wed, 21 May 2014 18:03:33 +0000 (19:03 +0100)
commitb057e1d80f916922081c7c5bd2c08ff53ed9e4fc
tree93fb6f89d05a6d17426aee5caa9f1e148056388d
parent9da2fb2070585c138ccd041f0c4f9e72886312e6
arm64, debug-monitors: Fix CPU hotplug callback registration

Subsystems that want to register CPU hotplug callbacks, as well as perform
initialization for the CPUs that are already online, often do it as shown
below:

get_online_cpus();

for_each_online_cpu(cpu)
init_cpu(cpu);

register_cpu_notifier(&foobar_cpu_notifier);

put_online_cpus();

This is wrong, since it is prone to ABBA deadlocks involving the
cpu_add_remove_lock and the cpu_hotplug.lock (when running concurrently
with CPU hotplug operations).

Instead, the correct and race-free way of performing the callback
registration is:

cpu_notifier_register_begin();

for_each_online_cpu(cpu)
init_cpu(cpu);

/* Note the use of the double underscored version of the API */
__register_cpu_notifier(&foobar_cpu_notifier);

cpu_notifier_register_done();

Fix the debug-monitors code in arm64 by using this latter form of callback
registration.

Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Russell King <rmk+kernel@arm.linux.org.uk>
Cc: Ingo Molnar <mingo@kernel.org>
Acked-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
(cherry picked from commit 4b0b68af37b930403cf9074c0cf504fc2387c2fa)
Signed-off-by: Mark Brown <broonie@linaro.org>
arch/arm64/kernel/debug-monitors.c