Commit bc758133 authored Mar 26, 2012 by Peter Zijlstra Committed by Ingo Molnar Mar 27, 2012

sched/x86/smp: Do not enable IRQs over calibrate_delay()

We should not ever enable IRQs until we're fully set up. This opens up
a window where interrupts can hit the cpu and interrupts can do
wakeups, wakeups need state that isn't set-up yet, in particular this
cpu isn't elegible to run tasks, so if any cpu-affine task that got
created in CPU_UP_PREPARE manages to get a wakeup, its affinity mask
will get broken and we'll run into lots of 'interesting' problems.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/n/tip-yaezmlbriluh166tfkgni22m@git.kernel.org

Signed-off-by: Ingo Molnar <mingo@kernel.org>

parent 6c16a6dc

arch/x86/kernel/smpboot.c

+0 −5

Original line number	Diff line number	Diff line
		@@ -219,14 +219,9 @@ static void __cpuinit smp_callin(void)
		* Update loops_per_jiffy in cpu_data. Previous call to
		* smp_store_cpu_info() stored a value that is close but not as
		* accurate as the value just calculated.
		*
		* Need to enable IRQs because it can take longer and then
		* the NMI watchdog might kill us.
		*/
		local_irq_enable();
		calibrate_delay();
		cpu_data(cpuid).loops_per_jiffy = loops_per_jiffy;
		local_irq_disable();
		pr_debug("Stack at about %p\n", &cpuid);

		/*