linux-kernel - Re: [RFC][PATCH 2/4] tracing: Use pid bitmap instead of a pid array for set_event

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] [day] [month] [year] [list]

Date:	Fri, 22 Apr 2016 11:30:19 -0400
From:	Steven Rostedt <rostedt@...dmis.org>
To:	Namhyung Kim <namhyung@...nel.org>
Cc:	<linux-kernel@...r.kernel.org>, Ingo Molnar <mingo@...nel.org>,
	Andrew Morton <akpm@...ux-foundation.org>,
	"H. Peter Anvin" <hpa@...or.com>,
	Thomas Gleixner <tglx@...utronix.de>,
	Jiri Olsa <jolsa@...nel.org>,
	Masami Hiramatsu <mhiramat@...nel.org>,
	<linux-trace-users@...r.kernel.org>
Subject: Re: [RFC][PATCH 2/4] tracing: Use pid bitmap instead of a pid array
 for set_event_pid

On Fri, 22 Apr 2016 11:45:30 +0900
Namhyung Kim <namhyung@...nel.org> wrote:

> > +	pid_list->pid_max = READ_ONCE(pid_max);
> > +	/* Only truncating will shrink pid_max */
> > +	if (filtered_pids && filtered_pids->pid_max > pid_list->pid_max)
> > +		pid_list->pid_max = filtered_pids->pid_max;
> > +	pid_list->pids = vzalloc((pid_list->pid_max + 7) >> 3);
> > +	if (!pid_list->pids) {
> > +		kfree(pid_list);
> > +		read = -ENOMEM;
> > +		goto out;
> > +	}
> > +	if (filtered_pids) {
> > +		/* copy the current bits to the new max */
> > +		pid = find_first_bit(filtered_pids->pids,
> > +				     filtered_pids->pid_max);
> > +		while (pid < filtered_pids->pid_max) {
> > +			set_bit(pid, pid_list->pids);
> > +			pid = find_next_bit(filtered_pids->pids,
> > +					    filtered_pids->pid_max,
> > +					    pid + 1);
> > +			nr_pids++;
> > +		}  
> 
> Why not just use memcpy and keep nr_pids in the pid_list?

This is the slow path (very slow ;-), and this was the first method
that came to my mind (while I programmed this during a conference). I
could use memcpy, or simply one of the bitmask copies, and then get the
nr_pids from bitmask_weight(). I would not keep nr_pids in pid_list
because that would mean that I would have to manage it in the fast path.

Maybe later I'll convert it to bitmap_copy(), but for now I'll just
keep it as is. I move this code in my queue for 4.8, and don't want to
deal with conflicts unless there's a real bug discovered.

Thanks for looking at this code!

-- Steve