linux-kernel - Re: [PATCH] tracing: Add saved_tgids file to show cached pid to tgid mappings

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20170630175057.71de9d39@gandalf.local.home>
Date:   Fri, 30 Jun 2017 17:50:57 -0400
From:   Steven Rostedt <rostedt@...dmis.org>
To:     Michael Sartain <mikesart@...tmail.com>
Cc:     linux-kernel@...r.kernel.org, Joel Fernandes <joelaf@...gle.com>,
        Ingo Molnar <mingo@...hat.com>
Subject: Re: [PATCH] tracing: Add saved_tgids file to show cached pid to
 tgid mappings

On Fri, 30 Jun 2017 11:17:50 -0600
Michael Sartain <mikesart@...tmail.com> wrote:

> Export the cached pid / tgid mappings to userspace. This allows user
> apps to translate the pids from a trace to their respective thread
> group.
> 
> Example saved_tgids file with pid / tgid values separated by ' ':
> 
>   # cat saved_tgids
>   1048 1048
>   1047 1047
>   7 7
>   1049 1047
>   1054 1047
>   1053 1047
> 
> [ Impact: let userspace apps reading binary buffer know tgid's ]

Impact line? Linus put a kibosh on this a while ago...

  https://lkml.org/lkml/2009/4/15/296

Although, I will admit that this line is actually useful. Just remove
the brackets.

> 
> Signed-off-by: Michael Sartain <mikesart@...tmail.com>
> ---
>  kernel/trace/trace.c | 60 +++++++++++++++++++++++++++++++++++++++++++++++++++-
>  1 file changed, 59 insertions(+), 1 deletion(-)
> 
> diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
> index 68c214b..ca84c97 100644
> --- a/kernel/trace/trace.c
> +++ b/kernel/trace/trace.c
> @@ -4692,6 +4692,7 @@ static const struct file_operations tracing_readme_fops = {
>  static void *saved_cmdlines_next(struct seq_file *m, void *v, loff_t *pos)
>  {
>  	unsigned int *ptr = v;
> +	long tgid_check = (long) m->private;

I really don't like the subtle use of having m->private != NULL mean
this is for tgid listing.

In fact, I don't see the purpose of reusing the seq code. The cmdlines
and tgid map are quite different. Just create its own functions. I
don't see the benefit of trying to reuse this except for making the
code more complex.

-- Steve

>  
>  	if (*pos || m->count)
>  		ptr++;
> @@ -4702,6 +4703,8 @@ static void *saved_cmdlines_next(struct seq_file *m, void *v, loff_t *pos)
>  	     ptr++) {
>  		if (*ptr == -1 || *ptr == NO_CMDLINE_MAP)
>  			continue;
> +		if (tgid_check && !trace_find_tgid(*ptr))
> +			continue;
>  
>  		return ptr;
>  	}
> @@ -4713,6 +4716,10 @@ static void *saved_cmdlines_start(struct seq_file *m, loff_t *pos)
>  {
>  	void *v;
>  	loff_t l = 0;
> +	long tgid_check = (long) m->private;
> +
> +	if (tgid_check && !tgid_map)
> +		return NULL;
>  
>  	preempt_disable();
>  	arch_spin_lock(&trace_cmdline_lock);
> @@ -4743,6 +4750,14 @@ static int saved_cmdlines_show(struct seq_file *m, void *v)
>  	return 0;
>  }
>