linux-stable.git/kernel/trace/ftrace.c, branch linux-2.6.29.y

tracing/function-graph-tracer: trace the idle tasks

2009-02-17T18:20:17+00:00

When the function graph tracer is activated, it iterates over the task_list
to allocate a stack to store the return addresses.

But the per cpu idle tasks are not iterated by using
do_each_thread / while_each_thread.

So we have to iterate on them manually.

This fixes somes weirdness in the traces and many losses of traces.
Examples on two cpus:

 0)   Xorg-4287    |   2.906 us    |              }
 0)   Xorg-4287    |   3.965 us    |            }
 0)   Xorg-4287    |   5.302 us    |          }
 ------------------------------------------
 0)   Xorg-4287    =>    -0
 ------------------------------------------

 0)    -0    |   2.861 us    |                        }
 0)    -0    |   0.526 us    |                        set_normalized_timespec();
 0)    -0    |   7.201 us    |                      }
 0)    -0    |   8.214 us    |                    }
 0)    -0    |               |                    clockevents_program_event() {
 0)    -0    |               |                      lapic_next_event() {
 0)    -0    |   0.510 us    |                        native_apic_mem_write();
 0)    -0    |   1.546 us    |                      }
 0)    -0    |   2.583 us    |                    }
 0)    -0    | + 12.435 us   |                  }
 0)    -0    | + 13.470 us   |                }
 0)    -0    |   0.608 us    |                _spin_unlock_irqrestore();
 0)    -0    | + 23.270 us   |              }
 0)    -0    | + 24.336 us   |            }
 0)    -0    | + 25.417 us   |          }
 0)    -0    |   0.593 us    |          _spin_unlock();
 0)    -0    | + 41.869 us   |        }
 0)    -0    | + 42.906 us   |      }
 0)    -0    | + 95.035 us   |    }
 0)    -0    |   0.540 us    |    menu_reflect();
 0)    -0    | ! 100.404 us  |  }
 0)    -0    |   0.564 us    |  mce_idle_callback();
 0)    -0    |               |  enter_idle() {
 0)    -0    |   0.526 us    |    mce_idle_callback();
 0)    -0    |   1.757 us    |  }
 0)    -0    |               |  cpuidle_idle_call() {
 0)    -0    |               |    menu_select() {
 0)    -0    |   0.525 us    |      pm_qos_requirement();
 0)    -0    |   0.518 us    |      tick_nohz_get_sleep_length();
 0)    -0    |   2.621 us    |    }
[...]
 1)    -0    |   0.518 us    |              touch_softlockup_watchdog();
 1)    -0    | + 14.355 us   |            }
 1)    -0    | + 22.840 us   |          }
 1)    -0    | + 25.949 us   |        }
 1)    -0    |               |        handle_irq() {
 1)    -0    |   0.511 us    |          irq_to_desc();
 1)    -0    |               |          handle_edge_irq() {
 1)    -0    |   0.638 us    |            _spin_lock();
 1)    -0    |               |            ack_apic_edge() {
 1)    -0    |   0.510 us    |              irq_to_desc();
 1)    -0    |               |              move_native_irq() {
 1)    -0    |   0.510 us    |                irq_to_desc();
 1)    -0    |   1.532 us    |              }
 1)    -0    |   0.511 us    |              native_apic_mem_write();
 ------------------------------------------
 1)    -0    =>    cat-5073
 ------------------------------------------

 1)    cat-5073    |   3.731 us    |                    }
 1)    cat-5073    |               |                    run_local_timers() {
 1)    cat-5073    |   0.533 us    |                      hrtimer_run_queues();
 1)    cat-5073    |               |                      raise_softirq() {
 1)    cat-5073    |               |                        __raise_softirq_irqoff() {
 1)    cat-5073    |               |                          /* nr: 1 */
 1)    cat-5073    |   2.718 us    |                        }
 1)    cat-5073    |   3.814 us    |                      }

Signed-off-by: Frederic Weisbecker 
Cc: Steven Rostedt 
Cc: Arnaldo Carvalho de Melo 
Signed-off-by: Ingo Molnar

ftrace: do_each_pid_task() needs rcu lock

2009-02-03T21:50:58+00:00

"ftrace: use struct pid" commit 978f3a45d9499c7a447ca7615455cefb63d44165
converted ftrace_pid_trace to "struct pid*".

But we can't use do_each_pid_task() without rcu_read_lock() even if
we know the pid itself can't go away (it was pinned in ftrace_pid_write).
The exiting task can detach itself from this pid at any moment.

Signed-off-by: Oleg Nesterov 
Signed-off-by: Ingo Molnar

tracing/function-graph-tracer: fix a regression while suspend to disk

2009-01-21T14:21:30+00:00

Impact: fix a crash while kernel image restore

When the function graph tracer is running and while suspend to disk, some racy
and dangerous things happen against this tracer.

The current task will save its registers including the stack pointer which
contains the return address hooked by the tracer. But the current task will
continue to enter other functions after that to save the memory, and then
it will store other return addresses, and finally loose the old depth which
matches the return address saved in the old stack (during the registers saving).

So on image restore, the code will return to wrong addresses.
And there are other things: on restore, the task will have it's "current"
pointer overwritten during registers restoring....switching from one task to
another... That would be insane to try to trace function graphs at these
stages.

This patch makes the function graph tracer listening on power events, making
it's tracing disabled for the current task (the one that performs the
hibernation work) while suspend/resume to disk, making the tracing safe
during hibernation.

Signed-off-by: Frederic Weisbecker 
Signed-off-by: Steven Rostedt 
Signed-off-by: Ingo Molnar

ftrace: add not to regex on filtering functions

2008-12-18T11:57:09+00:00

Impact: enhancement

Ingo Molnar has asked about a way to remove items from the filter
lists. Currently, you can only add or replace items. The way
items are added to the list is through opening one of the list
files (set_ftrace_filter or set_ftrace_notrace) via append.
If the file is opened for truncate, the list is cleared.

  echo spin_lock > /debug/tracing/set_ftrace_filter

The above will replace the list with only spin_lock

  echo spin_lock >> /debug/tracing/set_ftrace_filter

The above will add spin_lock to the list.

Now this patch adds:

  echo '!spin_lock' >> /debug/tracing/set_ftrace_filter

This will remove spin_lock from the list.

The limited glob features of these lists also can be notted.

  echo '!spin_*' >> /debug/tracing/set_ftrace_filter

This will remove all functions that start with 'spin_'

Note:

  echo '!spin_*' > /debug/tracing/set_ftrace_filter

will simply clear out the list (notice the '>' instead of '>>')

Signed-off-by: Steven Rostedt 
Signed-off-by: Ingo Molnar

tracing/function-graph-tracer: append the tracing_graph_flag

2008-12-08T14:11:45+00:00

Impact: Provide a way to pause the function graph tracer

As suggested by Steven Rostedt, the previous patch that prevented from
spinlock function tracing shouldn't use the raw_spinlock to fix it.
It's much better to follow lockdep with normal spinlock, so this patch
adds a new flag for each task to make the function graph tracer able
to be paused. We also can send an ftrace_printk whithout worrying of
the irrelevant traced spinlock during insertion.

Signed-off-by: Frederic Weisbecker 
Signed-off-by: Ingo Molnar

ftrace: use init_struct_pid as swapper pid

2008-12-05T13:51:29+00:00

Impact: clean up

Using (struct pid *)-1 as the pointer for ftrace_swapper_pid is
a little confusing for others. This patch uses the address of the
actual init pid structure instead. This change is only for
clarity. It does not affect the code itself. Hopefully soon the
swapper tasks will all have their own pid structure and then
we can clean up the code a bit more.

Signed-off-by: Steven Rostedt 
Signed-off-by: Ingo Molnar

ftrace: avoid duplicated function when writing set_graph_function

2008-12-04T08:42:35+00:00

Impact: fix a bug in function filter setting

when writing function to set_graph_function, we should check whether it
has existed in set_graph_function to avoid duplicating.

Signed-off-by: Liming Wang 
Signed-off-by: Ingo Molnar

ftrace: add ability to only trace swapper tasks

2008-12-04T08:09:38+00:00

Impact: new feature

This patch lets the swapper tasks of all CPUS be filtered by the
set_ftrace_pid file.

If '0' is echoed into this file, then all the idle tasks (aka swapper)
is flagged to be traced.  This affects all CPU idle tasks.

Signed-off-by: Steven Rostedt 
Signed-off-by: Ingo Molnar

ftrace: use struct pid

2008-12-04T08:09:37+00:00

Impact: clean up, extend PID filtering to PID namespaces

Eric Biederman suggested using the struct pid for filtering on
pids in the kernel. This patch is based off of a demonstration
of an implementation that Eric sent me in an email.

Signed-off-by: Steven Rostedt 
Signed-off-by: Ingo Molnar

ftrace: trace single pid for function graph tracer

2008-12-04T08:09:36+00:00

Impact: New feature

This patch makes the changes to set_ftrace_pid apply to the function
graph tracer.

  # echo $$ > /debugfs/tracing/set_ftrace_pid
  # echo function_graph > /debugfs/tracing/current_tracer

Will cause only the current task to be traced. Note, the trace flags are
also inherited by child processes, so the children of the shell
will also be traced.

Signed-off-by: Steven Rostedt 
Signed-off-by: Ingo Molnar