linux/kernel/trace
Steven Rostedt (VMware) 73a757e631 ring-buffer: Return reader page back into existing ring buffer
When reading the ring buffer for consuming, it is optimized for splice,
where a page is taken out of the ring buffer (zero copy) and sent to the
reading consumer. When the read is finished with the page, it calls
ring_buffer_free_read_page(), which simply frees the page. The next time the
reader needs to get a page from the ring buffer, it must call
ring_buffer_alloc_read_page() which allocates and initializes a reader page
for the ring buffer to be swapped into the ring buffer for a new filled page
for the reader.

The problem is that there's no reason to actually free the page when it is
passed back to the ring buffer. It can hold it off and reuse it for the next
iteration. This completely removes the interaction with the page_alloc
mechanism.

Using the trace-cmd utility to record all events (causing trace-cmd to
require reading lots of pages from the ring buffer, and calling
ring_buffer_alloc/free_read_page() several times), and also assigning a
stack trace trigger to the mm_page_alloc event, we can see how many times
the ring_buffer_alloc_read_page() needed to allocate a page for the ring
buffer.

Before this change:

  # trace-cmd record -e all -e mem_page_alloc -R stacktrace sleep 1
  # trace-cmd report |grep ring_buffer_alloc_read_page | wc -l
  9968

After this change:

  # trace-cmd record -e all -e mem_page_alloc -R stacktrace sleep 1
  # trace-cmd report |grep ring_buffer_alloc_read_page | wc -l
  4

Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
2017-05-01 10:26:40 -04:00
..
blktrace.c blktrace: use existing disk debugfs directory 2017-02-02 10:20:16 -07:00
bpf_trace.c bpf: mark all registered map/prog types as __ro_after_init 2017-02-17 13:40:04 -05:00
ftrace.c tracing/ftrace: Allow instances to have their own function probes 2017-04-20 22:06:47 -04:00
Kconfig ftrace: Add use of synchronize_rcu_tasks() with dynamic trampolines 2017-04-07 09:41:51 -04:00
Makefile perf/core: Rename CONFIG_[UK]PROBE_EVENT to CONFIG_[UK]PROBE_EVENTS 2017-03-01 10:26:39 +01:00
power-traces.c
ring_buffer_benchmark.c ring-buffer: Return reader page back into existing ring buffer 2017-05-01 10:26:40 -04:00
ring_buffer.c ring-buffer: Return reader page back into existing ring buffer 2017-05-01 10:26:40 -04:00
rpm-traces.c
trace_benchmark.c tracing: Have the trace_event benchmark thread call cond_resched_rcu_qs() 2017-04-17 15:21:19 -04:00
trace_benchmark.h tracing: Have the reg function allow to fail 2016-12-09 09:13:30 -05:00
trace_branch.c tracing: Add the constant count for branch tracer 2017-01-19 08:57:41 -05:00
trace_clock.c sched/headers: Prepare for new header dependencies before moving code to <linux/sched/clock.h> 2017-03-02 08:42:27 +01:00
trace_entries.h tracing: Add the constant count for branch tracer 2017-01-19 08:57:41 -05:00
trace_event_perf.c
trace_events_filter_test.h
trace_events_filter.c tracing/filter: Define op as the enum that it is 2016-11-14 16:42:59 -05:00
trace_events_hist.c sched/headers: Prepare to use <linux/rcuupdate.h> instead of <linux/rculist.h> in <linux/sched.h> 2017-03-02 08:42:38 +01:00
trace_events_trigger.c sched/headers: Prepare to use <linux/rcuupdate.h> instead of <linux/rculist.h> in <linux/sched.h> 2017-03-02 08:42:38 +01:00
trace_events.c tracing/ftrace: Add a better way to pass data via the probe functions 2017-04-20 22:06:46 -04:00
trace_export.c
trace_functions_graph.c tracing/fgraph: Have wakeup and irqsoff tracers ignore graph functions too 2016-12-09 09:21:35 -05:00
trace_functions.c tracing/ftrace: Allow for instances to trigger their own stacktrace probes 2017-04-20 22:06:49 -04:00
trace_hwlat.c sched/headers: Prepare for new header dependencies before moving code to <linux/sched/clock.h> 2017-03-02 08:42:27 +01:00
trace_irqsoff.c clocksource: Use a plain u64 instead of cycle_t 2016-12-25 11:04:12 +01:00
trace_kdb.c
trace_kprobe.c tracing/kprobes: expose maxactive for kretprobe in kprobe_events 2017-04-04 10:32:03 -04:00
trace_mmiotrace.c tracing: Expose CPU physical addresses (resource values) for PCI devices 2016-06-20 09:54:22 -04:00
trace_nop.c
trace_output.c sched/headers: Prepare for new header dependencies before moving code to <linux/sched/mm.h> 2017-03-02 08:42:28 +01:00
trace_output.h
trace_printk.c tracing: Handle NULL formats in hold_module_trace_bprintk_format() 2016-06-20 09:46:12 -04:00
trace_probe.c tracing/probes: Fix a warning message to show correct maximum length 2017-02-15 10:32:48 -05:00
trace_probe.h perf/core: Rename CONFIG_[UK]PROBE_EVENT to CONFIG_[UK]PROBE_EVENTS 2017-03-01 10:26:39 +01:00
trace_sched_switch.c
trace_sched_wakeup.c clocksource: Use a plain u64 instead of cycle_t 2016-12-25 11:04:12 +01:00
trace_selftest_dynamic.c
trace_selftest.c sched/headers: Prepare for new header dependencies before moving code to <uapi/linux/sched/types.h> 2017-03-02 08:42:27 +01:00
trace_seq.c
trace_stack.c rcu/tracing: Add rcu_disabled to denote when rcu_irq_enter() will not work 2017-04-10 15:22:03 -04:00
trace_stat.c
trace_stat.h
trace_syscalls.c tracing/syscalls: fix multiline in error message text 2016-09-29 10:25:23 +02:00
trace_uprobe.c sched/headers: Prepare to use <linux/rcuupdate.h> instead of <linux/rculist.h> in <linux/sched.h> 2017-03-02 08:42:38 +01:00
trace.c ring-buffer: Return reader page back into existing ring buffer 2017-05-01 10:26:40 -04:00
trace.h tracing/ftrace: Allow for the traceonoff probe be unique to instances 2017-04-20 22:06:48 -04:00
tracing_map.c
tracing_map.h