The function machine__get_kernel_start_addr() was taking the first symbol
of kallsyms as the start address. This is incorrect in certain cases
where the first symbol is something at 0, while the actual kernel
functions begin at a later point (e.g. 0x80200000).
This patch fixes machine__get_kernel_start_addr() to search for the
symbol "_text" or "_stext", which marks the beginning of kernel mapping.
This was already being done in machine__create_kernel_maps(). Thus, this
patch is just a refactor, to move that code into
machine__get_kernel_start_addr().
Signed-off-by: Simon Que <sque@chromium.org>
Link: http://lkml.kernel.org/r/1402943529-13244-1-git-send-email-sque@chromium.org
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Caching dso data file descriptors to avoid expensive re-opens
especially during DWARF unwind.
We keep dsos data file descriptors open until their count reaches
the half of the current fd open limit (RLIMIT_NOFILE). In this case
we close file descriptor of the first opened dso object.
We've got overall speedup (~27% for my workload) of report:
'perf report --stdio -i perf-test.data' (3 runs)
(perf-test.data size was around 12GB)
current code:
545,640,944,228 cycles ( +- 0.53% )
785,255,798,320 instructions ( +- 0.03% )
366.340910010 seconds time elapsed ( +- 3.65% )
after change:
435,895,036,114 cycles ( +- 0.26% )
636,790,271,176 instructions ( +- 0.04% )
266.481463387 seconds time elapsed ( +- 0.13% )
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jean Pihet <jean.pihet@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1401892622-30848-7-git-send-email-jolsa@kernel.org
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
When configuring event perf checked a wrong condition that user
specified both of freq (-F) and period (-c) or the event has no
default value. This worked because most of events don't have default
value and only tracepoint events have default of 1 (and it's not
desirable to change it for those events).
However, Andi's downloadable event patch changes the situation so it
cannot change the value for those events. Fix it by allowing override
the default value if user gives one of the options.
$ perf record -a -e uops_retired.all -F 4000 sleep 1
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.325 MB perf.data (~14185 samples) ]
$ perf evlist -F
cpu/uops_retired.all/: sample_freq=4000
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Link: http://lkml.kernel.org/r/1402292617-26278-1-git-send-email-namhyung@kernel.org
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Pull perf/core improvements from Arnaldo Carvalho de Melo:
User visible:
* Improve 'perf probe' error messages, moving some diagnostic messages to
only appear in --verbose mode and fixing up some error reporting related
to variables and struct members. (Masami Hiramatsu)
* Reflow 'perf timechart' man page. (Stanislav Fomichev)
Developer stuff:
* Be more precise when reporting missing libraries in a static tool build.
(Arnaldo Carvalho de Melo)
* Show error messages from the multiple make invoked from 'make build-test'.
(Arnaldo Carvalho de Melo)
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Improve error messages of 'perf probe --line' mode.
Currently 'perf probe' shows the "Debuginfo analysis failed" message with
an error code when the given symbol is not found:
-----
# perf probe -L page_cgroup_init_flatmem
Debuginfo analysis failed. (-2)
Error: Failed to show lines.
-----
But -2 (-ENOENT) means that the given source line or function was not
found. With this patch, 'perf probe' shows the correct error message:
-----
# perf probe -L page_cgroup_init_flatmem
Specified source line is not found.
Error: Failed to show lines.
-----
There is also another debug error code is shown in the same function
after get_real_path(). This removes that too.
Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/20140606071406.6788.47850.stgit@kbuild-fedora.novalocal
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Fix an error message when failed to find given address in --vars
mode.
Without this fix, perf probe -V doesn't show the final "Error"
message if it fails to find given source line. Moreover, it
tells it fails to find "variables" instead of the source line.
-----
# perf probe -V foo@bar
Failed to find variables at foo@bar (0)
-----
The result also shows mysterious error code. Actually the error
returns 0 or -ENOENT means that it just fails to find the address
of given source line. (0 means there is no matching address,
and -ENOENT means there is an entry(DIE) but it has no instance,
e.g. an empty inlined function)
This fixes it to show what happened and the final error message
as below.
-----
# perf probe -V foo@bar
Failed to find the address of foo@bar
Error: Failed to show vars.
-----
Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/20140606071359.6788.84716.stgit@kbuild-fedora.novalocal
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Improve the error message if we can not find given member in the given
structure. Currently perf probe shows a wrong error message as below.
-----
# perf probe getname_flags:65 "result->BOGUS"
result(type:filename) has no member BOGUS.
Failed to find 'result' in this function.
Error: Failed to add events. (-22)
-----
The first message is correct, but the second one is not, since we didn't
fail to find a variable but fails to find the member of given variable.
-----
# perf probe getname_flags:65 "result->BOGUS"
result(type:filename) has no member BOGUS.
Error: Failed to add events. (-22)
-----
With this patch, the error message shows only the first one. And if we
really failed to find given variable, it tells us so.
-----
# perf probe getname_flags:65 "BOGUS"
Failed to find 'BOGUS' in this function.
Error: Failed to add events. (-2)
-----
Reported-by: Arnaldo Carvalho de Melo <acme@kernel.org>
Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/20140606071345.6788.23744.stgit@kbuild-fedora.novalocal
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
In perf's 'mem-mode', one can get access to a whole bunch of details specific to a
particular sample instruction. A bunch of those details relate to the data
address.
One interesting thing you can do with data addresses is to convert them into a unique
cacheline they belong too. Organizing these data cachelines into similar groups and sorting
them can reveal cache contention.
This patch creates an alogorithm based on various sample details that can help group
entries together into data cachelines and allows 'perf report' to sort on it.
The algorithm relies on having proper mmap2 support in the kernel to help determine
if the memory map the data address belongs to is private to a pid or globally shared.
The alogortithm is as follows:
o group cpumodes together
o group entries with discovered maps together
o sort on major, minor, inode and inode generation numbers
o if userspace anon, then sort on pid
o sort on cachelines based on data addresses
The 'dcacheline' sort option in 'perf report' only works in 'mem-mode'.
Sample output:
#
# Samples: 206 of event 'cpu/mem-loads/pp'
# Total weight : 2534
# Sort order : dcacheline,pid
#
# Overhead Samples Data Cacheline Command: Pid
# ........ ............ ...................................................................... ..................
#
13.22% 1 [k] 0xffff88042f08ebc0 swapper: 0
9.27% 1 [k] 0xffff88082e8cea80 swapper: 0
3.59% 2 [k] 0xffffffff819ba180 swapper: 0
0.32% 1 [k] arch_trigger_all_cpu_backtrace_handler_na.23901+0xffffffffffffffe0 swapper: 0
0.32% 1 [k] timekeeper_seq+0xfffffffffffffff8 swapper: 0
Note: Added a '+1' to symlen size in hists__calc_col_len to prevent the next column
from prematurely tabbing over and mis-aligning. Not sure what the problem is.
Signed-off-by: Don Zickus <dzickus@redhat.com>
Link: http://lkml.kernel.org/r/1401208087-181977-8-git-send-email-dzickus@redhat.com
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
With the Sebastian's change of handling num array argument (of raw
syscall enter), the script still failed to work like this:
$ perf record -e raw_syscalls:* sleep 1
$ perf script -g python
$ perf script -s perf-script.py
...
Traceback (most recent call last):
File "perf-script.py", line 42, in raw_syscalls__sys_enter
(id, args),
TypeError: %u format: a number is required, not list
Fatal Python error: problem in Python trace event handler
Aborted (core dumped)
This is because the generated script tries to print the array arg as
unsigned integer (%u). Since the python seems to convert arguments to
strings by default, just using %s solved the problem for me.
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Tom Zanussi <tzanussi@gmail.com>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Link: http://lkml.kernel.org/r/1401338695-18837-1-git-send-email-namhyung@kernel.org
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Coming in v3.16, trace events will be able to save bitmasks in raw
format in the ring buffer and output it with the __get_bitmask() macro.
In order for userspace tools to parse this, it must be able to handle
the __get_bitmask() call and be able to convert the data that's in
the ring buffer into a nice bitmask format. The output is similar to
what the kernel uses to print bitmasks, with a comma separator every
4 bytes (8 characters).
This allows for cpumasks to also be saved efficiently.
The first user is the thermal:thermal_power_limit event which has the
following output:
thermal_power_limit: cpus=0000000f freq=1900000 cdev_state=0 power=5252
Link: http://lkml.kernel.org/r/20140506132238.22e136d1@gandalf.local.home
Suggested-by: Javi Merino <javi.merino@arm.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Javi Merino <javi.merino@arm.com>
Link: http://lkml.kernel.org/r/20140603032224.229186537@goodmis.org
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Fix perf probe to find correct variable DIE which has location or
external instance by tracking down the lexical blocks.
Current die_find_variable() expects that the all variable DIEs
which has DW_TAG_variable have a location. However, since recent
dwarf information may have declaration variable DIEs at the
entry of function (subprogram), die_find_variable() returns it.
To solve this problem, it must track down the DIE tree to find
a DIE which has an actual location or a reference for external
instance.
e.g. finding a DIE which origin is <0xdc73>;
<1><11496>: Abbrev Number: 95 (DW_TAG_subprogram)
<11497> DW_AT_abstract_origin: <0xdc42>
<1149b> DW_AT_low_pc : 0x1850
[...]
<2><114cc>: Abbrev Number: 119 (DW_TAG_variable) <- this is a declaration
<114cd> DW_AT_abstract_origin: <0xdc73>
<2><114d1>: Abbrev Number: 119 (DW_TAG_variable)
[...]
<3><115a7>: Abbrev Number: 105 (DW_TAG_lexical_block)
<115a8> DW_AT_ranges : 0xaa0
<4><115ac>: Abbrev Number: 96 (DW_TAG_variable) <- this has a location
<115ad> DW_AT_abstract_origin: <0xdc73>
<115b1> DW_AT_location : 0x486c (location list)
Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Tested-by: Arnaldo Carvalho de Melo <acme@kernel.org>
Acked-by: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/r/20140529121930.30879.87092.stgit@ltc230.yrl.intra.hitachi.co.jp
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Fix a segfault bug by asking for variable it doesn't find.
Since the convert_variable() didn't handle error code returned
from convert_variable_location(), it just passed an incomplete
variable field and then a segfault was occurred when formatting
the field.
This fixes that bug by handling success code correctly in
convert_variable(). Other callers of convert_variable_location()
are correctly checking the return code.
This bug was introduced by following commit. But another hidden
erroneous error handling has been there previously (-ENOMEM case).
commit 3d918a12a1
Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Reported-by: Arnaldo Carvalho de Melo <acme@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/r/20140529105232.28251.30447.stgit@ltc230.yrl.intra.hitachi.co.jp
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
After output/sort fields refactoring, it's expensive
to check the elide bool in its current location inside
the 'struct sort_entry'.
The perf_hpp__should_skip function gets highly noticable in
workloads with high number of output/sort fields, like for:
$ perf report -i perf-test.data -F overhead,sample,period,comm,pid,dso,symbol,cpu --stdio
Performance report:
9.70% perf [.] perf_hpp__should_skip
Moving the elide bool into the 'struct perf_hpp_fmt', which
makes the perf_hpp__should_skip just single struct read.
Got speedup of around 22% for my test perf.data workload.
The change should not harm any other workload types.
Performance counter stats for (10 runs):
before:
358,319,732,626 cycles ( +- 0.55% )
467,129,581,515 instructions # 1.30 insns per cycle ( +- 0.00% )
150.943975206 seconds time elapsed ( +- 0.62% )
now:
278,785,972,990 cycles ( +- 0.12% )
370,146,797,640 instructions # 1.33 insns per cycle ( +- 0.00% )
116.416670507 seconds time elapsed ( +- 0.31% )
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
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/20140601142622.GA9131@krava.brq.redhat.com
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
The new ->add_entry_cb() will be called after an entry was added to
the histogram. It's used for code sharing between perf report and
perf top. Note that ops->add_*_entry() should set iter->he properly
in order to call the ->add_entry_cb.
Also pass @arg to the callback function. It'll be used by perf top
later.
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Link: http://lkml.kernel.org/r/87k393g999.fsf@sejong.aot.lge.com
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Print accumulated stat of a hist entry if requested.
To do that, add new HPP_PERCENT_ACC_FNS macro and generate a
perf_hpp_fmt using it. The __hpp__sort_acc() function sorts entries
by accumulated period value. When accumulated periods of two entries
are same (i.e. single path callchain) put the caller above since
accumulation tends to put callers on higher position for obvious
reason.
Also add "overhead_children" output field to be selected by user.
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Link: http://lkml.kernel.org/r/1401335910-16832-11-git-send-email-namhyung@kernel.org
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
There're some duplicate code when adding hist entries. They are
different in that some have branch info or mem info but generally do
same thing. So introduce new struct hist_entry_iter and add callbacks
to customize each case in general way.
The new perf_evsel__add_entry() function will look like:
iter->prepare_entry();
iter->add_single_entry();
while (iter->next_entry())
iter->add_next_entry();
iter->finish_entry();
This will help further work like the cumulative callchain patchset.
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1401335910-16832-3-git-send-email-namhyung@kernel.org
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
This patch automatically adjusts the path of MMAP records
associated with Android system libraries.
The Android system is organized with system libraries found in
/system/lib and user libraries in /data/app-lib. On the host system
(not running Android), system libraries can be found in the downloaded
NDK directory under ${NDK_ROOT}/platforms/${APP_PLATFORM}/arch-${ARCH}/usr/lib
and the user libraries are installed under libs/${APP_ABI} within
the apk build directory. This patch makes running the reporting
tools possible on the host system using the libraries from the NDK.
Signed-off-by: Michael Lentine <mlentine@google.com>
Reviewed-by: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1400579330-5043-3-git-send-email-eranian@google.com
[ fixed 'space required before the open parenthesis' checkpatch.pl errors ]
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
The -F/--fields option is to allow user setup output field in any
order. It can receive any sort keys and following (hpp) fields:
overhead, overhead_sys, overhead_us, sample and period
If guest profiling is enabled, overhead_guest_{sys,us} will be
available too.
The output fields also affect sort order unless you give -s/--sort
option. And any keys specified on -s option, will also be added to
the output field list automatically.
$ perf report -F sym,sample,overhead
...
# Symbol Samples Overhead
# .......................... ............ ........
#
[.] __cxa_atexit 2 2.50%
[.] __libc_csu_init 4 5.00%
[.] __new_exitfn 3 3.75%
[.] _dl_check_map_versions 1 1.25%
[.] _dl_name_match_p 4 5.00%
[.] _dl_setup_hash 1 1.25%
[.] _dl_sysdep_start 1 1.25%
[.] _init 5 6.25%
[.] _setjmp 6 7.50%
[.] a 8 10.00%
[.] b 8 10.00%
[.] brk 1 1.25%
[.] c 8 10.00%
Note that, the example output above is captured after applying next
patch which fixes sort/comparing behavior.
Requested-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Ingo Molnar <mingo@kernel.org>
Link: http://lkml.kernel.org/r/1400480762-22852-12-git-send-email-namhyung@kernel.org
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
