linux-trace-devel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Yordan Karadzhov (VMware)" <y.karadz@gmail.com>
To: Steven Rostedt <rostedt@goodmis.org>, linux-trace-devel@vger.kernel.org
Subject: Re: [PATCH v4 5/7] libtracefs: Implement tracefs_kprobe_clear_all() to remove all kprobes
Date: Fri, 2 Jul 2021 14:15:44 +0300	[thread overview]
Message-ID: <e01fb2ef-a51f-f573-3869-829dd93f777d@gmail.com> (raw)
In-Reply-To: <20210702035443.154729-6-rostedt@goodmis.org>



On 2.07.21 г. 6:54, Steven Rostedt wrote:
> From: "Steven Rostedt (VMware)" <rostedt@goodmis.org>
> 
> A call to tracefs_kprobe_clear_all() will attempt to disable all kprobes. If
> any kprobe is set, and the @force parameter is set, it will fail with
> errno set to EBUSY. If @force is set, then it will attempt to disable all
> the defined kprobe events and then clear it.
> 
> Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
> ---
>   include/tracefs.h     |   1 +
>   src/tracefs-kprobes.c | 103 ++++++++++++++++++++++++++++++++++++++++++
>   2 files changed, 104 insertions(+)
> 
> diff --git a/include/tracefs.h b/include/tracefs.h
> index 5f59c480d572..3b57c596feab 100644
> --- a/include/tracefs.h
> +++ b/include/tracefs.h
> @@ -220,4 +220,5 @@ int tracefs_kprobe_raw(const char *system, const char *event,
>   int tracefs_kretprobe_raw(const char *system, const char *event,
>   			  const char *addr, const char *format);
>   char **tracefs_get_kprobes(void);
> +int tracefs_kprobe_clear_all(bool force);
>   #endif /* _TRACE_FS_H */
> diff --git a/src/tracefs-kprobes.c b/src/tracefs-kprobes.c
> index e875d6e8a65f..fe61a25baa10 100644
> --- a/src/tracefs-kprobes.c
> +++ b/src/tracefs-kprobes.c
> @@ -169,3 +169,106 @@ char **tracefs_get_kprobes(void)
>   	list = NULL;
>   	goto out;
>   }
> +
> +struct instance_list {
> +	struct instance_list	*next;
> +	struct tracefs_instance	*instance;
> +};

This is not used in this source file. I see that it was moved to 
tracefs-instance.c

Thanks!
Y.

> +
> +static void disable_events(const char *system, const char *event,
> +			   char **list)
> +{
> +	struct tracefs_instance *instance;
> +	int i;
> +
> +	/*
> +	 * Note, this will not fail even on error.
> +	 * That is because even if something fails, it may still
> +	 * work enough to clear the kprobes. If that's the case
> +	 * the clearing after the loop will succeed and the function
> +	 * is a success, even though other parts had failed. If
> +	 * one of the kprobe events is enabled in one of the
> +	 * instances that fail, then the clearing will fail too
> +	 * and the function will return an error.
> +	 */
> +
> +	tracefs_event_disable(NULL, system, event);
> +	/* No need to test results */
> +
> +	if (!list)
> +		return;
> +
> +	for (i = 0; list[i]; i++) {
> +		instance = tracefs_instance_alloc(NULL, list[i]);
> +		/* If this fails, try the next one */
> +		if (!instance)
> +			continue;
> +		tracefs_event_disable(instance, system, event);
> +		tracefs_instance_free(instance);
> +	}
> +	return;
> +}
> +
> +/**
> + * tracefs_kprobe_clear_all - clear kprobe events
> + * @force: Will attempt to disable all kprobe events and clear them
> + *
> + * Will remove all defined kprobe events. If any of them are enabled,
> + * and @force is not set, then it will error with -1 and errno to be
> + * EBUSY. If @force is set, then it will attempt to disable all the kprobe
> + * events in all instances, and try again.
> + *
> + * Returns zero on success, -1 otherwise.
> + */
> +int tracefs_kprobe_clear_all(bool force)
> +{
> +	char **instance_list;
> +	char **kprobe_list;
> +	char *saveptr;
> +	char *system;
> +	char *kprobe;
> +	char *event;
> +	int ret;
> +	int i;
> +
> +	ret = tracefs_instance_file_clear(NULL, KPROBE_EVENTS);
> +	if (!ret)
> +		return 0;
> +
> +	if (!force)
> +		return -1;
> +
> +	kprobe_list = tracefs_get_kprobes();
> +	if (!kprobe_list)
> +		return -1;
> +
> +	instance_list = tracefs_instances(NULL);
> +	/*
> +	 * Even if the above failed and instance_list is NULL,
> +	 * keep going, as the enabled event may simply be in the
> +	 * top level.
> +	 */
> +
> +	for (i = 0; kprobe_list[i]; i++) {
> +		kprobe = kprobe_list[i];
> +
> +		system = strtok_r(kprobe, "/", &saveptr);
> +		if (!system)
> +			goto out;
> +
> +		event = strtok_r(NULL," ", &saveptr);
> +		if (!event)
> +			goto out;
> +
> +		disable_events(system, event, instance_list);
> +
> +		ret = tracefs_instance_file_clear(NULL, KPROBE_EVENTS);
> +		/* On success stop the loop */
> +		if (!ret)
> +			goto out;
> +	}
> + out:
> +	tracefs_list_free(instance_list);
> +	tracefs_list_free(kprobe_list);
> +	return ret;
> +}
> 

  reply	other threads:[~2021-07-02 11:15 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-07-02  3:54 [PATCH v4 0/7] libtracefs: Facilitate adding and removing kprobes Steven Rostedt
2021-07-02  3:54 ` [PATCH v4 1/7] libtracefs: Implement tracefs_instances() Steven Rostedt
2021-07-02 11:10   ` Yordan Karadzhov (VMware)
2021-07-02 11:33     ` Steven Rostedt
2021-07-02 11:39       ` Yordan Karadzhov (VMware)
2021-07-02 11:50         ` Steven Rostedt
2021-07-02  3:54 ` [PATCH v4 2/7] libtracefs: Implement tracefs_kprobe_raw() Steven Rostedt
2021-07-02  3:54 ` [PATCH v4 3/7] libtracefs: Implement tracefs_kretprobe_raw() Steven Rostedt
2021-07-02  3:54 ` [PATCH v4 4/7] libtracefs: Implement tracefs_get_kprobes() Steven Rostedt
2021-07-02  3:54 ` [PATCH v4 5/7] libtracefs: Implement tracefs_kprobe_clear_all() to remove all kprobes Steven Rostedt
2021-07-02 11:15   ` Yordan Karadzhov (VMware) [this message]
2021-07-02 11:35     ` Steven Rostedt
2021-07-02  3:54 ` [PATCH v4 6/7] libtracefs: Implement tracefs_kprobe_clear_probe() Steven Rostedt
2021-07-02  3:54 ` [PATCH v4 7/7] libtracefs: Add man pages for kprobe functions 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=e01fb2ef-a51f-f573-3869-829dd93f777d@gmail.com \
    --to=y.karadz@gmail.com \
    --cc=linux-trace-devel@vger.kernel.org \
    --cc=rostedt@goodmis.org \
    /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).