Mirror of the lttng-dev mailing list
 help / color / mirror / Atom feed
From: mathieu.desnoyers@polymtl.ca (Mathieu Desnoyers)
Subject: [ltt-dev] [UST PATCH] Fix return value handling for libustcmd
Date: Tue, 7 Sep 2010 16:56:53 -0400	[thread overview]
Message-ID: <20100907205653.GA13695@Krystal> (raw)
In-Reply-To: <1283891172-1695-1-git-send-email-david.goulet@polymtl.ca>

* David Goulet (david.goulet at polymtl.ca) wrote:
> Return value was not checked correctly so this was
> triggering a free() on an invalid pointer causing
> ustctl to fail badly.
> 
> Signed-off-by: David Goulet <david.goulet at polymtl.ca>

Acked-by: Mathieu Desnoyers <mathieu.desnoyers at efficios.com>

The fact that Pierre-Marc, Douglas and Philippe all got it wrong in the
function return value description/implementation supports my
recommandation to switch 

send_message_fd()

along with *all* its callers to the following returns values:

All negative values: Error returned by patient_send().
                     (includes -EPIPE: connection closed)
0: success.

Side-note: write_current_subbuffer() and on_read_partial_subbuffer() in
ustd/ustd.c use patient_write(), but the manpage states that the
returns values for the "0" case is different between files and
non-files (e.g. unix sockets). So we should probably look into this more
closely.

Thanks,

Mathieu

> ---
>  libustcmd/ustcmd.c |   11 ++++-------
>  1 files changed, 4 insertions(+), 7 deletions(-)
> 
> diff --git a/libustcmd/ustcmd.c b/libustcmd/ustcmd.c
> index 5b4fd02..f0a6ae0 100644
> --- a/libustcmd/ustcmd.c
> +++ b/libustcmd/ustcmd.c
> @@ -97,7 +97,7 @@ int ustcmd_set_marker_state(const char *mn, int state, pid_t pid)
>  	}
>  
>  	result = ustcmd_send_cmd(cmd, pid, NULL);
> -	if (result) {
> +	if (result != 1) {
>  		free(cmd);
>  		return USTCMD_ERR_GEN;
>  	}
> @@ -182,9 +182,8 @@ int ustcmd_get_subbuf_size(const char *channel, pid_t pid)
>  	}
>  
>  	result = ustcmd_send_cmd(cmd, pid, &reply);
> -	if (result) {
> +	if (result != 1) {
>  		free(cmd);
> -		free(reply);
>  		return -1;
>  	}
>  
> @@ -214,9 +213,8 @@ int ustcmd_get_subbuf_num(const char *channel, pid_t pid)
>  	}
>  
>  	result = ustcmd_send_cmd(cmd, pid, &reply);
> -	if (result) {
> +	if (result != 1) {
>  		free(cmd);
> -		free(reply);
>  		return -1;
>  	}
>  
> @@ -488,7 +486,6 @@ int ustcmd_get_sock_path(char **sock_path, pid_t pid)
>  	result = ustcmd_send_cmd(cmd, pid, &reply);
>  	if (result != 1) {
>  		free(cmd);
> -		free(reply);
>  		return USTCMD_ERR_GEN;
>  	}
>  
> @@ -516,7 +513,7 @@ int ustcmd_force_switch(pid_t pid)
>   * @param pid	Targeted PID
>   * @param reply	Pointer to string to be filled with a reply string (must
>   *		be NULL if no reply is needed for the given command).
> - * @return	-1 if successful, 0 on EOT, 1 on success
> + * @return	-1 if not successful, 0 on EOT, 1 on success
>   */
>  
>  int ustcmd_send_cmd(const char *cmd, const pid_t pid, char **reply)
> -- 
> 1.7.2.2
> 

-- 
Mathieu Desnoyers
Operating System Efficiency R&D Consultant
EfficiOS Inc.
http://www.efficios.com




  reply	other threads:[~2010-09-07 20:56 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-09-07 20:26 David Goulet
2010-09-07 20:56 ` Mathieu Desnoyers [this message]
2010-09-07 22:13   ` David Goulet

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=20100907205653.GA13695@Krystal \
    --to=mathieu.desnoyers@polymtl.ca \
    /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