perf/x86/intel: Make WARN()ings consistent
authorPeter Zijlstra <peterz@infradead.org>
Thu, 21 May 2015 08:57:28 +0000 (10:57 +0200)
committerIngo Molnar <mingo@kernel.org>
Wed, 27 May 2015 07:17:43 +0000 (09:17 +0200)
commit17186ccda374ae02ef231cbbc8f1825e7c19ddbd
treea19eac0a863534752861ade3892cc0f71f1ca1aa
parentaaf932e8161e45291cc85085b6d850f1bbdf53c8
perf/x86/intel: Make WARN()ings consistent

The intel_commit_scheduling() callback is pointlessly different from
the start and stop scheduling callback.

Furthermore, the constraint should never be NULL, so remove that test.

Even though we'll never get called (because we NULL the callbacks)
when !is_ht_workaround_enabled() put that test in.

Collapse the (pointless) WARN_ON_ONCE() and bail on !cpuc->excl_cntrs --
this is doubly pointless, because its the same condition as
is_ht_workaround_enabled() which was already pointless because the
whole method won't ever be called.

Furthremore, make all the !excl_cntrs test WARN_ON_ONCE(); they're all
pointless, because the above, either the function
({get,put}_excl_constraint) are already predicated on it existing or
the is_ht_workaround_enabled() thing is the same test.

Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
arch/x86/kernel/cpu/perf_event_intel.c