[PATCH v13 00/15]tracing: Hist trigger snapshot and onchange additions

* [PATCH v13 00/15]tracing: Hist trigger snapshot and onchange additions
@ 2019-02-04 21:44 Tom Zanussi
  2019-02-04 21:44 ` [PATCH v13 01/15] tracing: Refactor hist trigger action code Tom Zanussi
                   ` (14 more replies)
  0 siblings, 15 replies; 16+ messages in thread
From: Tom Zanussi @ 2019-02-04 21:44 UTC (permalink / raw)
  To: rostedt
  Cc: tglx, mhiramat, namhyung, vedang.patel, bigeasy, joel,
	mathieu.desnoyers, julia, linux-kernel, linux-rt-users

From: Tom Zanussi <tom.zanussi@linux.intel.com>

Hi,

This is v13 of the hist trigger snapshot and onchange additions
patchset.

It addresses a bug Namhyung found where taking a snapshot from the
command-line would cause an OOPs if there was a hist trigger snapshot
action in effect at the same time.  What should have happened in this
case was that the command-line snapshot return with -EBUSY.  -EBUSY
should also be returned if the user tries to start a tracer that takes
a snapshot, such as the wakeup or preemptoff tracers.  Conversely, if
any of those tracers is active or another snapshot() histogram trigger
is in effect, a new snapshot() histogram trigger should fail with
-EBUSY as well.

In addition, I did a couple of other cleanups - I realized that using
track_data to temporarily save the key and val was wrong and so added
a snapshot_context object and pass that around instead.  Also, since
with onchange 0 is now a valid value, introduced an 'updated' flag to
allow the trigger printing function to determine if the snapshot was
updated or not.  Finally, I simplified the overall ontrack processing
by getting rid of the get/set_track_val_fn function pointers along
with a couple other now unused fields.

Thanks,

Tom

v12->v13 changes:

  - Added -EBUSY checks to tracing_snapshot_cond_enable()
    tracing_snapshot_cond(), tracing_set_tracer() and
    tracing_snapshot_write().
  - Introduced snapshot_context for passing key/elt pair to
    tracing_snapshot_cond().
  - Removed key and cond_snapshot from action_data and cur_elt from
    track_data.
  - Added updated flag to track_data and test for it in
    hist_trigger_print_key().
  - Removed cond_snapshot_update_track_data() and simplified
    cond_snapshot_update().
  - Removed get/save_track_val_fn_t and corresponding
    get_val/save_val() members from action_data.
  - Moved code from get/save_val functions to save_data instead.

v11->v12 changes:

  - Removed parens from the param list specification in the trace()
    command description
  - Changed the Documentation for synthetic event generation to
    emphasize the trace() form over the function-call form.

v10->v11 changes:

  - Updated to use str_has_prefix() for static strings

v9->v10 changes:

  - Updated to use strncmp_prefix() for static strings

v8->v9 changes:

  - Moved v8 patches 17-22 to a separate series
  - Moved track_data.var_ref_idx removal into 'tracing: Generalize
    hist trigger onmax and save action'
  - Merged 'tracing: Add hist trigger handler.action documentation'
    into 'tracing: Refactor hist trigger action code'
  - Updated patches to change all strlen() to sizeof() for string
    constants
  - Moved test cases to the end of the series

v7->v8 changes:

  - Removed /dev/null redirection from test cases
  - Replaced event disable with tracing_on disable instead
  - New patch to remove hist_field.var_idx and track_data.var_ref_idx
  - New patch to use var_refs[] array for reference checking
  - New patch to remove open-coding of var_ref count manipulation
  - New patch to use var_refs[] array to destroy var_refs
  - New patch to remove synth_var_refs[] now that var_refs[] is used
  - New patch to add comments to variable-related fields
  - Added various new comments in the patches to clarify other things
  - Replaced the convoluted logic mentioned by Namhyung in
    update_track_val with wrappers and simpler check/save logic as
    suggested
  - Rebased to latest ftrace/core including Masami's dynamic event
    framework changes

v6->v7 changes:

  - Removed unnecessary HANDLER_ONMAX checks from onmax_print()/create()
  - Moved handler assignment to acion_parse()
  - Changed goto in ATION_TRACE case in action_create() to return
  - Changed EINVAL to EEXIST in action_create() ACTION_SAVE case
  - Made the return logic in create_actions() more consistent
  - Merged 'tracing: Move hist trigger key printing into a separate
    function' into 'tracing: Add hist trigger snapshot() action'
  - Updated the new snapshot, onchange, and trace test cases to match
    4.20 kselftest changes.
  - Added new xfail test case that make sure certain unsupported
    handler/action combinations fail as expected.
  - While updating the test cases, realized that the other testcases
    in the inter-event subdir needed SPDX license updates, so added
    them.

v5->v6 changes:

  - Added new Documentation patch explaining handler.action
  - Added new README patch explaining handler.action
  - Added separate snapshot() Documentation
  - Added new snapshot() test case
  - Updated README with snapshote()
  - Added separate onchange() Documentation
  - Added separate onchange() test case
  - Updated README with onchange()
  - Added separate trace() test case
  - Updated README with trace() and <synthetic_event>() syntax

v4->v5 changes:

  - added 'trace' keyword test case
  - added 'onchange' handler test case

v3->v4 changes:

  - added 'trace' keyword for generating synthetic events
  - fix elt_data leak
  - changed cond_update to cond_update_fn_t

v2->v3 changes:

  - fixed problem where trace actions were only being allowed for
    onmatch handlers - now trace actions can be used with any handler.
  - fixed problem where no action was being assigned to onmatch
    handlers if save or snapshot actions were specified.

v1->v2 changes:

  - added missing tracing_cond_snapshot_data() definition for when
    CONFIG_TRACER_SNAPSHOT not defined
  - removed an unnecessary WARN_ON() in track_data_snapshot_print()

Original text:

This patchset adds some useful new functions to the hist
trigger code: a snapshot action and an onchange handler.

In order to make it easier to add these and in the process make the
code more generic, I separated the code into explicit 'handlers' and
'actions', handlers being things like 'onmax' and 'onchange', and
'actions' being things like 'take a snapshot' or 'save some fields'.

The first few patches do that basic refactoring, which make it easier
to add the subsequent changes that arbitrarily combine actions and
handlers.

The fourth patch adds a 'conditional snapshot' capability that via a
new tracing_snaphot_cond() function extends the existing snapshot
code.  It allows the caller to associate some user data with the
snapshot that can be checked and saved in an update() callback whose
return value determines whether the snapshot should be taken or not.

The remaining patches finally add the new snapshot action and onchange
handler functionality - please see those patches for details and some
examples.

Thanks,

Tom

The following changes since commit 50c35ae1267d64eee975b8125e151e600071d4dc:

  tracing: Use strncpy instead of memcpy for string keys in hist triggers (2019-02-04 08:45:50 -0600)

are available in the git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/zanussi/linux-trace.git ftrace/hist-snapshot-onchange-v13

Tom Zanussi (15):
  tracing: Refactor hist trigger action code
  tracing: Make hist trigger Documentation better reflect
    actions/handlers
  tracing: Split up onmatch action data
  tracing: Generalize hist trigger onmax and save action
  tracing: Add conditional snapshot
  tracing: Add hist trigger snapshot() action
  tracing: Add hist trigger snapshot() action Documentation
  tracing: Add hist trigger onchange() handler
  tracing: Add hist trigger onchange() handler Documentation
  tracing: Add alternative synthetic event trace action syntax
  tracing: Add SPDX license GPL-2.0 license identifier to inter-event
    testcases
  tracing: Add hist trigger snapshot() action test case
  tracing: Add hist trigger onchange() handler test case
  tracing: Add alternative synthetic event trace action test case
  tracing: Add hist trigger action 'expected fail' test case

 Documentation/trace/histogram.rst                  |  316 +++++-
 kernel/trace/trace.c                               |  212 +++-
 kernel/trace/trace.h                               |   56 +-
 kernel/trace/trace_events_hist.c                   | 1036 ++++++++++++++------
 kernel/trace/trace_events_trigger.c                |    2 +-
 kernel/trace/trace_sched_wakeup.c                  |    2 +-
 .../inter-event/trigger-action-hist-xfail.tc       |   30 +
 .../inter-event/trigger-extended-error-support.tc  |    1 +
 .../inter-event/trigger-field-variable-support.tc  |    1 +
 .../trigger-inter-event-combined-hist.tc           |    1 +
 .../inter-event/trigger-multi-actions-accept.tc    |    1 +
 .../inter-event/trigger-onchange-action-hist.tc    |   28 +
 .../inter-event/trigger-onmatch-action-hist.tc     |    1 +
 .../trigger-onmatch-onmax-action-hist.tc           |    1 +
 .../inter-event/trigger-onmax-action-hist.tc       |    1 +
 .../inter-event/trigger-snapshot-action-hist.tc    |   43 +
 .../trigger-synthetic-event-createremove.tc        |    1 +
 .../inter-event/trigger-trace-action-hist.tc       |   42 +
 18 files changed, 1448 insertions(+), 327 deletions(-)
 create mode 100644 tools/testing/selftests/ftrace/test.d/trigger/inter-event/trigger-action-hist-xfail.tc
 create mode 100644 tools/testing/selftests/ftrace/test.d/trigger/inter-event/trigger-onchange-action-hist.tc
 create mode 100644 tools/testing/selftests/ftrace/test.d/trigger/inter-event/trigger-snapshot-action-hist.tc
 create mode 100644 tools/testing/selftests/ftrace/test.d/trigger/inter-event/trigger-trace-action-hist.tc

-- 
2.14.1

^ permalink raw reply	[flat|nested] 16+ messages in thread