sched,preempt: Fix preempt_count manipulations

Vikram reported that his ARM64 compiler managed to 'optimize' away the
preempt_count manipulations in code like:

        preempt_enable_no_resched();
        put_user();
        preempt_disable();

Irrespective of that fact that that is horrible code that should be
fixed for many reasons, it does highlight a deficiency in the generic
preempt_count manipulators. As it is never right to combine/elide
preempt_count manipulations like this.

Therefore sprinkle some volatile in the two generic accessors to
ensure the compiler is aware of the fact that the preempt_count is
observed outside of the regular program-order view and thus cannot be
optimized away like this.

x86; the only arch not using the generic code is not affected as we
do all this in asm in order to use the segment base per-cpu stuff.

Change-Id: I781dc34fdf52823fd34f4bb93872f85847076c66
Cc: stable@vger.kernel.org
Cc: Thomas Gleixner <tglx@linutronix.de>
Fixes: a7878709 ("sched, arch: Create asm/preempt.h")
Reported-by: Vikram Mulukutla <markivx@codeaurora.org>
Tested-by: Vikram Mulukutla <markivx@codeaurora.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Git-commit: 2e636d5e66c35dfcbaf617aa8fa963f6847478fe
Git-repo: git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
Link: https://lkml.org/lkml/2016/5/16/190


[vmulukut@codeaurora.org: merge conflict fixups]
Signed-off-by: Vikram Mulukutla <markivx@codeaurora.org>