linux-trace-devel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Steven Rostedt <rostedt@goodmis.org>
To: "Tzvetomir Stoyanov (VMware)" <tz.stoyanov@gmail.com>
Cc: y.karadz@gmail.com, linux-trace-devel@vger.kernel.org
Subject: Re: [PATCH 2/4] libtracefs: New API for getting dynamic event
Date: Wed, 1 Dec 2021 16:04:26 -0500	[thread overview]
Message-ID: <20211201160426.08d0c7d2@gandalf.local.home> (raw)
In-Reply-To: <20211130050057.336228-3-tz.stoyanov@gmail.com>

On Tue, 30 Nov 2021 07:00:55 +0200
"Tzvetomir Stoyanov (VMware)" <tz.stoyanov@gmail.com> wrote:

> A new API is proposed, to get tep event describing given dynamic event:
[blank line]
> 	tracefs_dynevent_get_event()
[blank line]
> The API detects any newly created dynamic events, but does not detect
> the deletion.
> 
> Signed-off-by: Tzvetomir Stoyanov (VMware) <tz.stoyanov@gmail.com>
> ---
>  Documentation/libtracefs-dynevents.txt | 14 +++++++++++--
>  include/tracefs.h                      |  2 ++
>  src/tracefs-dynevents.c                | 29 ++++++++++++++++++++++++++
>  3 files changed, 43 insertions(+), 2 deletions(-)
> 
> diff --git a/Documentation/libtracefs-dynevents.txt b/Documentation/libtracefs-dynevents.txt
> index a374651..7de06be 100644
> --- a/Documentation/libtracefs-dynevents.txt
> +++ b/Documentation/libtracefs-dynevents.txt
> @@ -4,8 +4,8 @@ libtracefs(3)
>  NAME
>  ----
>  tracefs_dynevent_create, tracefs_dynevent_destroy, tracefs_dynevent_destroy_all,
> -tracefs_dynevent_free, tracefs_dynevent_list_free, tracefs_dynevent_get_all, tracefs_dynevent_info -
> -Create, destroy, free and get dynamic events.
> +tracefs_dynevent_free, tracefs_dynevent_list_free, tracefs_dynevent_get_all, tracefs_dynevent_info,
> +tracefs_dynevent_get_event - Create, destroy, free and get dynamic events.
>  
>  SYNOPSIS
>  --------
> @@ -23,6 +23,7 @@ void *tracefs_dynevent_list_free*(struct tracefs_dynevent pass:[*]pass:[*]_event
>  struct tracefs_dynevent pass:[*]*tracefs_dynevent_get*(enum tracefs_dynevent_type _type_, const char pass:[*]_system_, const char pass:[*]_event_);
>  struct tracefs_dynevent pass:[*]pass:[*]*tracefs_dynevent_get_all*(unsigned int _types_, const char pass:[*]_system_);
>  enum tracefs_dynevent_type *tracefs_dynevent_info*(struct tracefs_dynevent pass:[*]_dynevent_, char pass:[*]pass:[*]_system_, char pass:[*]pass:[*]_event_, char pass:[*]pass:[*]_prefix_, char pass:[*]pass:[*]_addr_, char pass:[*]pass:[*]_format_);
> +struct tep_event pass:[*]*tracefs_dynevent_get_event*(struct tep_handle pass:[*]_tep_, struct tracefs_dynevent pass:[*]_dynevent_);
>  --
>  
>  DESCRIPTION
> @@ -68,6 +69,11 @@ if relevant for this event type. If _format_ is non NULL, it will hold the forma
>  dynamic event. Note, that the content in _group_, _event_, _prefix_, _addr_, and _format_ must be
>  freed with free(3) if they are set.
>  
> +The *tracefs_dynevent_get_event*() function returns tep event, describing the given dynamic event.

					  "returns a tep event"

> +If the dynamic event is newly created and not yet loaded in the @tep, the dynamic event system is

            "loaded in @tep"

> +rescanned for any new events. The returned pointer to tep event is controlled by @tep and must not
> +be freed.


> +
>  RETURN VALUE
>  ------------
>  
> @@ -88,6 +94,10 @@ in case of an error or in case there are no events in the system. That array mus
>  on error. If _system_, _event_, _prefix_, _addr_, or _format_ are non NULL, they will contain
>  allocated strings that must be freed by free(3).
>  
> +The *tracefs_dynevent_get_event*() function returns pointer to tep event or NULL in case of an error

  "returns a pointer to a tep event'

> +or if the requested dynamic event is missing. The returned pointer to tep event is controlled by
> +@tep and must not be freed.
> +
>  ERRORS
>  ------
>  The following errors are for all the above calls:
> diff --git a/include/tracefs.h b/include/tracefs.h
> index 9662603..fbd7d31 100644
> --- a/include/tracefs.h
> +++ b/include/tracefs.h
> @@ -262,6 +262,8 @@ tracefs_dynevent_get(enum tracefs_dynevent_type type, const char *system, const
>  enum tracefs_dynevent_type
>  tracefs_dynevent_info(struct tracefs_dynevent *dynevent, char **system,
>  		      char **event, char **prefix, char **addr, char **format);
> +struct tep_event *
> +tracefs_dynevent_get_event(struct tep_handle *tep, struct tracefs_dynevent *dynevent);
>  
>  struct tracefs_dynevent *
>  tracefs_eprobe_alloc(const char *system, const char *event,
> diff --git a/src/tracefs-dynevents.c b/src/tracefs-dynevents.c
> index ea07d13..2ff3c87 100644
> --- a/src/tracefs-dynevents.c
> +++ b/src/tracefs-dynevents.c
> @@ -752,3 +752,32 @@ error:
>  
>  	return TRACEFS_DYNEVENT_UNKNOWN;
>  }
> +
> +/**
> + * tracefs_dynevent_get_event - return tep event, representing the given dynamic event

  remove the comma.

> + * @tep: a handle to the trace event parser context, that holds the events

  remove the comma.

> + * @dynevent: a dynamic event context, describing given dynamic event.

  remove the comma ;-)

> + *
> + * Returns pointer to tep event, describing the given dynamic event. The pointer

 "Returns a pointer to a tep event describing the given dynamic event".

> + * is managed by @tep handle and must not be freed. In case of an error, or in case

  "by the @tep handle"

> + * the requested dynamic event is missing in the @tep handler - NULL is returned.
> + */
> +struct tep_event *
> +tracefs_dynevent_get_event(struct tep_handle *tep, struct tracefs_dynevent *dynevent)
> +{
> +	struct tep_event *event;
> +
> +	if (!tep || !dynevent || !dynevent->event)
> +		return NULL;
> +
> +	/* If the dynamic event is already loaded in the tep, return it */
> +	event = tep_find_event_by_name(tep, dynevent->system, dynevent->event);
> +	if (event)
> +		return event;
> +
> +	/* Try to load any new events from the given system */
> +	if (trace_load_events(tep, NULL, dynevent->system))

Does this refresh, or does just append?

That is, if there are already dynamic events, wouldn't this re-add the
existing events that are already there?

Perhaps we need an interface in libtraceevent that rescans, and checks for
duplicates.

-- Steve

> +		return NULL;
> +
> +	return tep_find_event_by_name(tep, dynevent->system, dynevent->event);
> +}


  reply	other threads:[~2021-12-01 21:04 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-11-30  5:00 [PATCH 0/4] New tracefs APIs Tzvetomir Stoyanov (VMware)
2021-11-30  5:00 ` [PATCH 1/4] libtracefs: Reuse logic for loading events inside the library Tzvetomir Stoyanov (VMware)
2021-11-30  5:00 ` [PATCH 2/4] libtracefs: New API for getting dynamic event Tzvetomir Stoyanov (VMware)
2021-12-01 21:04   ` Steven Rostedt [this message]
2021-12-02  5:05     ` Tzvetomir Stoyanov
2021-11-30  5:00 ` [PATCH 3/4] libtracefs: Unit test for tracefs_dynevent_get_event() Tzvetomir Stoyanov (VMware)
2021-11-30  5:00 ` [PATCH 4/4] libtracefs: New API for applying filter on event Tzvetomir Stoyanov (VMware)
2021-12-01 19:55   ` Steven Rostedt

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20211201160426.08d0c7d2@gandalf.local.home \
    --to=rostedt@goodmis.org \
    --cc=linux-trace-devel@vger.kernel.org \
    --cc=tz.stoyanov@gmail.com \
    --cc=y.karadz@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).