Re: [PATCHSET 0/9] perf report: Support folded callchain output (v5)

[Brendan Gregg <brendan.d.gregg@gmail.com>]

On Sun, Nov 8, 2015 at 9:45 PM, Namhyung Kim <namhyung@kernel.org> wrote:
> Hello,
>
> This is what Brendan requested on the perf-users mailing list [1] to
> support FlameGraphs [2] more efficiently.  This patchset adds a few
> more callchain options to adjust the output for it.
>
>  * changes in v5)
>    - honor field separator from -t option
>    - add support for TUI and GTK
>
>  * changes in v4)
>    - add missing doc update
>      - cleanup/fix callchain value print code
>        - add Acked-by from Brendan and Jiri
>
>  * changes in v3)
>    - put the value before callchains
>      - fix compile error
>
>
> At first, 'folded' output mode was added.  The folded output puts the
> value, a space and all calchain nodes separated by semicolons.  Now it
> only supports --stdio as other UI provides some way of folding and/or
> expanding callchains dynamically.
>
> The value is now can be one of 'percent', 'period', or 'count'.  The
> percent is current default output and the period is the raw number of
> sample periods.  The count is the number of samples for each callchain.
>
> The proposed features of hiding hist lines with '-F none' and showing
> hist info with callchains can be added as later work.
>
> Here's an example:
>
>   $ perf report --no-children --show-nr-samples --stdio -g folded,count
>   ...
>     39.93%     80  swapper  [kernel.vmlinux]  [k] intel_idel
>   57 intel_idle;cpuidle_enter_state;cpuidle_enter;call_cpuidle;cpu_startup_entry;start_secondary
>   23 intel_idle;cpuidle_enter_state;cpuidle_enter;call_cpuidle;cpu_startup_entry;rest_init;...
>

Thanks, I tested it, it works!

It lets me do this:

# ./perf report --no-children -n --stdio -g folded,count -F pid
[...]
      0:swapper
1032 xen_hypercall_sched_op;default_idle;arch_cpu_idle;default_idle_call;cpu_startup_entry;cpu_bringup_and_idle
134 xen_hypercall_sched_op;default_idle;arch_cpu_idle;default_idle_call;cpu_startup_entry;rest_init;start_kernel;x86_64_start_reservations;xen_start_kernel
1 xen_hypercall_xen_version;check_events;__schedule;schedule;schedule_preempt_disabled;cpu_startup_entry;cpu_bringup_and_idle
1 xen_hypercall_xen_version;check_events;cpu_startup_entry;rest_init;start_kernel;x86_64_start_reservations;xen_start_kernel
   1248:bash
43 copy_page_range;copy_process;_do_fork;sys_clone;entry_SYSCALL_64_fastpath;__libc_fork;make_child
6 xen_hypercall_xen_version;check_events;xen_dup_mmap;copy_process;_do_fork;sys_clone;entry_SYSCALL_64_fastpath;__libc_fork;make_child
4 xen_hypercall_xen_version;check_events;copy_page_range;copy_process;_do_fork;sys_clone;entry_SYSCALL_64_fastpath;__libc_fork;make_child
[...]

This is a parsable call chain summary, and which can be flamegraph'd
after a touch of awk. Later on we can add a "-F none" and "-g pid"
etc, but this patch solves the number one issue of avoiding the
expense of needing to re-aggregate the call chain output (output of
perf script), so I'd be pretty happy to use this instead.

Brendan

>
>   $ perf report --no-children --stdio -g percent
>   ...
>     39.93%  swapper  [kernel.vmlinux]  [k] intel_idel
>             |
>             ---intel_idle
>                cpuidle_enter_state
>                cpuidle_enter
>                call_cpuidle
>                cpu_startup_entry
>                |
>                |--28.63%-- start_secondary
>                |
>                 --11.30%-- rest_init
>
>
>   $ perf report --no-children --stdio --show-total-period -g period
>   ...
>     39.93%   13018705  swapper  [kernel.vmlinux]  [k] intel_idel
>             |
>             ---intel_idle
>                cpuidle_enter_state
>                cpuidle_enter
>                call_cpuidle
>                cpu_startup_entry
>                |
>                |--9334403-- start_secondary
>                |
>                 --3684302-- rest_init
>
>
>   $ perf report --no-children --stdio --show-nr-samples -g count
>   ...
>     39.93%     80  swapper  [kernel.vmlinux]  [k] intel_idel
>             |
>             ---intel_idle
>                cpuidle_enter_state
>                cpuidle_enter
>                call_cpuidle
>                cpu_startup_entry
>                |
>                |--57-- start_secondary
>                |
>                 --23-- rest_init
>
>
> You can get it from 'perf/callchain-fold-v5' branch on my tree:
>
>   git://git.kernel.org/pub/scm/linux/kernel/git/namhyung/linux-perf.git
>
> Any comments are welcome, thanks
> Namhyung
>
>
> [1] http://www.spinics.net/lists/linux-perf-users/msg02498.html
> [2] http://www.brendangregg.com/FlameGraphs/cpuflamegraphs.html
>
>
> Namhyung Kim (9):
>   perf report: Support folded callchain mode on --stdio
>   perf callchain: Abstract callchain print function
>   perf callchain: Add count fields to struct callchain_node
>   perf report: Add callchain value option
>   perf hists browser: Factor out hist_browser__show_callchain_list()
>   perf hists browser: Support flat callchains
>   perf hists browser: Support folded callchains
>   perf ui/gtk: Support flat callchains
>   perf ui/gtk: Support folded callchains
>
>  tools/perf/Documentation/perf-report.txt |  14 +-
>  tools/perf/builtin-report.c              |   4 +-
>  tools/perf/ui/browsers/hists.c           | 316 ++++++++++++++++++++++++++++---
>  tools/perf/ui/gtk/hists.c                | 148 ++++++++++++++-
>  tools/perf/ui/stdio/hist.c               |  94 +++++++--
>  tools/perf/util/callchain.c              | 135 ++++++++++++-
>  tools/perf/util/callchain.h              |  28 ++-
>  tools/perf/util/util.c                   |   3 +-
>  8 files changed, 679 insertions(+), 63 deletions(-)
>
> --
> 2.6.2
>