linux/kernel/trace
Steven Rostedt (Red Hat) 03274a3ffb tracing/fgraph: Adjust fgraph depth before calling trace return callback
While debugging the virtual cputime with the function graph tracer
with a max_depth of 1 (most common use of the max_depth so far),
I found that I was missing kernel execution because of a race condition.

The code for the return side of the function has a slight race:

	ftrace_pop_return_trace(&trace, &ret, frame_pointer);
	trace.rettime = trace_clock_local();
	ftrace_graph_return(&trace);
	barrier();
	current->curr_ret_stack--;

The ftrace_pop_return_trace() initializes the trace structure for
the callback. The ftrace_graph_return() uses the trace structure
for its own use as that structure is on the stack and is local
to this function. Then the curr_ret_stack is decremented which
is what the trace.depth is set to.

If an interrupt comes in after the ftrace_graph_return() but
before the curr_ret_stack, then the called function will get
a depth of 2. If max_depth is set to 1 this function will be
ignored.

The problem is that the trace has already been called, and the
timestamp for that trace will not reflect the time the function
was about to re-enter userspace. Calls to the interrupt will not
be traced because the max_depth has prevented this.

To solve this issue, the ftrace_graph_return() can safely be
moved after the current->curr_ret_stack has been updated.
This way the timestamp for the return callback will reflect
the actual time.

If an interrupt comes in after the curr_ret_stack update and
ftrace_graph_return(), it will be traced. It may look a little
confusing to see it within the other function, but at least
it will not be lost.

Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2013-01-29 17:30:31 -05:00
..
blktrace.c tracing: Use this_cpu_ptr per-cpu helper 2013-01-21 13:22:30 -05:00
ftrace.c tracing: Avoid unnecessary multiple recursion checks 2013-01-22 23:38:01 -05:00
Kconfig ftrace: Move ARCH_SUPPORTS_FTRACE_SAVE_REGS in Kconfig 2013-01-21 13:22:35 -05:00
Makefile trace: Stop compiling in trace_clock unconditionally 2012-09-13 22:52:08 -04:00
power-traces.c perf: Clean up power events by introducing new, more generic ones 2011-01-04 08:16:54 +01:00
ring_buffer_benchmark.c tracing: Use NUMA allocation for per-cpu ring buffer pages 2011-06-14 22:04:39 -04:00
ring_buffer.c ring-buffer: Remove trace.h from ring_buffer.c 2013-01-22 23:38:03 -05:00
rpm-traces.c PM / Runtime: Introduce trace points for tracing rpm_* functions 2011-09-27 22:53:27 +02:00
trace_branch.c tracing: Cache comms only after an event occurred 2012-10-31 16:45:31 -04:00
trace_clock.c tracing: Remove trace.h header from trace_clock.c 2013-01-22 12:06:56 -05:00
trace_entries.h Merge branch 'tip/perf/urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace into perf/urgent 2012-03-24 08:19:09 +01:00
trace_event_perf.c perf/core improvements and fixes: 2012-08-21 11:27:00 +02:00
trace_events_filter_test.h tracing/filter: Add startup tests for events filter 2011-08-19 14:35:59 -04:00
trace_events_filter.c tracing: Replace strict_strto* with kstrto* 2012-10-31 16:45:23 -04:00
trace_events.c tracing: Remove the extra 4 bytes of padding in events 2013-01-21 21:05:41 -05:00
trace_export.c tracing: Do not enable function event with enable 2012-05-10 15:55:43 -04:00
trace_functions_graph.c tracing/fgraph: Adjust fgraph depth before calling trace return callback 2013-01-29 17:30:31 -05:00
trace_functions.c tracing: Fix unsigned int compare of zero in recursion check 2013-01-24 07:52:34 -05:00
trace_irqsoff.c Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial 2012-12-13 12:00:02 -08:00
trace_kdb.c
trace_kprobe.c tracing: Use irq_work for wake ups and remove *_nowake_*() functions 2012-11-02 10:21:52 -04:00
trace_mmiotrace.c atomic: use <linux/atomic.h> 2011-07-26 16:49:47 -07:00
trace_nop.c
trace_output.c tracing: Format non-nanosec times from tsc clock without a decimal point. 2012-11-13 15:48:40 -05:00
trace_output.h
trace_printk.c tracing: Add percpu buffers for trace_printk() 2012-04-23 21:15:55 -04:00
trace_probe.c tracing: Replace strict_strto* with kstrto* 2012-10-31 16:45:23 -04:00
trace_probe.h tracing: Provide trace events interface for uprobes 2012-05-07 14:30:17 +02:00
trace_sched_switch.c tracing: Use irq_work for wake ups and remove *_nowake_*() functions 2012-11-02 10:21:52 -04:00
trace_sched_wakeup.c Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial 2012-12-13 12:00:02 -08:00
trace_selftest_dynamic.c ftrace: Add self-tests for multiple function trace users 2011-05-18 19:24:51 -04:00
trace_selftest.c ftrace: Fix function tracing recursion self test 2013-01-22 23:37:58 -05:00
trace_stack.c tracing: Remove unneeded checks from the stack tracer 2012-11-19 15:07:13 -05:00
trace_stat.c
trace_stat.h
trace_syscalls.c tracing/syscalls: Make local functions static 2013-01-21 13:22:31 -05:00
trace_uprobe.c tracing: Verify target file before registering a uprobe event 2013-01-21 13:22:31 -05:00
trace.c tracing: Remove second iterator initializer 2013-01-29 09:57:49 -05:00
trace.h ring-buffer: User context bit recursion checking 2013-01-22 23:38:03 -05:00