In addition to using refcounts for the struct thread lifetime
management, we need to protect access to machine->threads from
concurrent access.
That happens in 'perf top', where a thread processes events, inserting
and deleting entries from that rb_tree while another thread decays
hist_entries, that end up dropping references and ultimately deleting
threads from the rb_tree and releasing its resources when no further
hist_entry (or other data structures, like in 'perf sched') references
it.
So the rule is the same for refcounts + protected trees in the kernel,
get the tree lock, find object, bump the refcount, drop the tree lock,
return, use object, drop the refcount if no more use of it is needed,
keep it if storing it in some other data structure, drop when releasing
that data structure.
I.e. pair "t = machine__find(new)_thread()" with a "thread__put(t)", and
"perf_event__preprocess_sample(&al)" with "addr_location__put(&al)".
The addr_location__put() one is because as we return references to
several data structures, we may end up adding more reference counting
for the other data structures and then we'll drop it at
addr_location__put() time.
Acked-by: David Ahern <dsahern@gmail.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-bs9rt4n0jw3hi9f3zxyy3xln@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
The prior change fixes default output ordering with each column but it
breaks -o/--order option. This patch prepends a new hpp fmt struct to
sort list but not to output field list so that it can affect ordering
without adding a new output column.
The new hpp fmt uses its own compare functions which treats dummy
entries (which have no baseline) little differently - the delta field
can be computed without baseline but others (ratio and wdiff) are not.
The new output will look like below:
$ perf diff -o 2 perf.data.{old,cur,new}
...
# Baseline/0 Delta/1 Delta/2 Shared Object Symbol
# .......... ....... ....... ................. ..........................................
22.98% +0.51% +0.52% libc-2.20.so [.] _int_malloc
5.70% +0.28% +0.30% libc-2.20.so [.] free
4.38% -0.21% +0.25% a.out [.] main
1.32% -0.15% +0.05% a.out [.] free@plt
+0.01% [kernel.kallsyms] [k] intel_pstate_timer_func
+0.01% [kernel.kallsyms] [k] _raw_spin_lock_irqsave
+0.01% [kernel.kallsyms] [k] timekeeping_update.constprop.8
+0.01% +0.01% [kernel.kallsyms] [k] apic_timer_interrupt
0.01% -0.00% [kernel.kallsyms] [k] native_read_msr_safe
0.01% -0.01% -0.01% [kernel.kallsyms] [k] native_write_msr_safe
1.31% +0.03% -0.06% a.out [.] malloc@plt
31.50% -0.74% -0.23% libc-2.20.so [.] _int_free
32.75% +0.28% -0.83% libc-2.20.so [.] malloc
0.01% [kernel.kallsyms] [k] scheduler_tick
+0.01% [kernel.kallsyms] [k] read_tsc
+0.01% [kernel.kallsyms] [k] perf_adjust_freq_unthr_context.part.82
In above example, the output was sorted by 'Delta/2' column first, and
then 'Baseline/0' and finally 'Delta/1'.
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Kan Liang <kan.liang@intel.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1420677949-6719-8-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Current perf diff result is somewhat confusing since it sometimes hide
small result and sometimes there's no result. So do not hide small
result (less than 0.01%) and print "N/A" if baseline is not
recorded (for ratio and wdiff only). Blank means the baseline is
available but its pairs are not.
Before:
# Baseline Delta Shared Object Symbol
# ........ ....... ................. .........................
#
...
0.01% -0.01% [kernel.kallsyms] [k] native_write_msr_safe
0.01% [kernel.kallsyms] [k] scheduler_tick
0.01% [kernel.kallsyms] [k] native_read_msr_safe
0.00% [kernel.kallsyms] [k] __rcu_read_unlock
[kernel.kallsyms] [k] _raw_spin_lock
+0.01% [kernel.kallsyms] [k] apic_timer_interrupt
[kernel.kallsyms] [k] read_tsc
After:
# Baseline Delta Shared Object Symbol
# ........ ....... ................. .........................
#
...
0.01% -0.01% [kernel.kallsyms] [k] native_write_msr_safe
0.01% [kernel.kallsyms] [k] scheduler_tick
0.01% [kernel.kallsyms] [k] native_read_msr_safe
0.00% [kernel.kallsyms] [k] __rcu_read_unlock
+0.01% [kernel.kallsyms] [k] _raw_spin_lock
+0.01% [kernel.kallsyms] [k] apic_timer_interrupt
+0.01% [kernel.kallsyms] [k] read_tsc
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1419656793-32756-3-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
It also uses hists/hist_entries, hists__init() should be called before
creating any evsels.
Otherwise no extra space will be allocated per perf_evsel nor this space
will be initialized when allocating a new perf_evsel instance, resulting
in reads/writes to non allocated space, oops. Fix it.
Signed-off-by: Kan Liang <kan.liang@intel.com>
Link: http://lkml.kernel.org/r/1414004561-22096-1-git-send-email-kan.liang@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Currently vmlinux_path__init() only tries to find vmlinux file from
current directory, /boot and some canonical directories with version
number of the running kernel. This can be a problem when reporting old
data recorded on a kernel version not running currently.
We can use --symfs option for this but it's annoying for user to do it
always. As we already have the info in the perf.data file, it can be
changed to use it for the search automatically.
Before:
$ perf report
...
# Samples: 4K of event 'cpu-clock'
# Event count (approx.): 1067250000
#
# Overhead Command Shared Object Symbol
# ........ .......... ................. ..............................
71.87% swapper [kernel.kallsyms] [k] recover_probed_instruction
After:
# Overhead Command Shared Object Symbol
# ........ .......... ................. ....................
71.87% swapper [kernel.kallsyms] [k] native_safe_halt
This requires to change signature of symbol__init() to receive struct
perf_session_env *.
Reported-by: Minchan Kim <minchan@kernel.org>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1407825645-24586-14-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Currently, accounting each sample is done in multiple places - once
when adding them to the input tree, other when adding them to the
output tree. It's not only confusing but also can cause a subtle
problem since concurrent processing like in perf top might see the
updated stats before adding entries into the output tree - like seeing
more (blank) lines at the end and/or slight inaccurate percentage.
To fix this, only account the entries when it's moved into the output
tree so that they cannot be seen prematurely. There're some
exceptional cases here and there - they should be addressed separately
with comments.
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/r/1398327843-31845-7-git-send-email-namhyung@kernel.org
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
