tracing: Make __buffer_unlock_commit() always_inline
authorSteven Rostedt (Red Hat) <rostedt@goodmis.org>
Thu, 24 Nov 2016 01:28:38 +0000 (20:28 -0500)
committerSteven Rostedt <rostedt@goodmis.org>
Thu, 24 Nov 2016 01:30:51 +0000 (20:30 -0500)
commit52ffabe3848a1ebd944cdf7801a77247b1cb46d5
treee1c9162bdf75f41d279a1361f240ead9cce2850c
parent4239174570da080f3623724d97062bf55de7e36b
tracing: Make __buffer_unlock_commit() always_inline

The function __buffer_unlock_commit() is called in a few places outside of
trace.c. But for the most part, it should really be inlined, as it is in the
hot path of the trace_events. For the callers outside of trace.c, create a
new function trace_buffer_unlock_commit_nostack(), as the reason it was used
was to avoid the stack tracing that trace_buffer_unlock_commit() could do.

Link: http://lkml.kernel.org/r/20161121183700.GW26852@two.firstfloor.org
Reported-by: Andi Kleen <andi@firstfloor.org>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
kernel/trace/trace.c
kernel/trace/trace.h
kernel/trace/trace_branch.c
kernel/trace/trace_functions_graph.c
kernel/trace/trace_hwlat.c