linux-kernel - Re: [PATCH 1/6] tracing: Turn event/id->i

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]

Date:	Thu, 25 Jul 2013 16:26:13 +0200
From:	Oleg Nesterov <oleg@...hat.com>
To:	Steven Rostedt <rostedt@...dmis.org>
Cc:	Masami Hiramatsu <masami.hiramatsu.pt@...achi.com>,
	Alexander Z Lam <azl@...gle.com>,
	Arnaldo Carvalho de Melo <acme@...stprotocols.net>,
	David Sharp <dhsharp@...gle.com>,
	Frederic Weisbecker <fweisbec@...il.com>,
	Ingo Molnar <mingo@...hat.com>,
	Peter Zijlstra <a.p.zijlstra@...llo.nl>,
	Srikar Dronamraju <srikar@...ux.vnet.ibm.com>,
	Vaibhav Nagarnaik <vnagarnaik@...gle.com>,
	"zhangwei(Jovi)" <jovi.zhangwei@...wei.com>,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/6] tracing: Turn event/id->i_private into
	call->event.type

On 07/24, Steven Rostedt wrote:
>
> On Tue, 2013-07-23 at 22:59 +0200, Oleg Nesterov wrote:
> > event_id_read() is racy, ftrace_event_call can be already freed
> > by trace_remove_event_call() callers.
> >
> > Change event_create_dir() to pass "data = call->event.type", this
> > is all event_id_read() needs. ftrace_event_id_fops no longer needs
> > tracing_open_generic().
> >
> > We add the new helper, event_file_data(), to read ->i_private, it
> > will have more users.
> >
> > Note: currently ACCESS_ONCE() and "id != 0" check are not needed,
> > but we are going to change event_remove/rmdir to clear ->i_mutex.
>
> We are changing i_mutex or i_private?

i_private, yes, will fix.

Steven, I agree with all your comments, thanks. Except this one.

> Anyway, this still looks too complex for the id. Just pass the id number
> to the filp->private_data, and use that. It should never be cleared, and
> as we are not using any pointers it will always return what the id
> is/was.
>
> Keep tracing_open_generic() and have:
>
> event_id_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t
> *ppos)
> {
> 	int id = (int)(unsigned long)filp->private_data;
> 	char buf[32];
> 	int len;
>
> 	if (*ppos)
> 		return 0;
>
> 	len = sprintf(buf, "%d\n", id);
> 	return simple_read_from_buffer(ubuf, cnt, ppos, buf, len);
> }
>
> No need for accessing the inode->i_private.

Unless you strongly object, I'd prefer to avoid filp->private_data
and use event_file_data(). Also, I'd like to keep "if (!id) ENODEV".

The main reason is the consistency with other f_op's, and this allows
us to change event_file_data/remove_event_file_dir without breaking
event_id_read().

Also. Even if we use filp->private_data, it can be NULL/0, ->open()
can race with rmdir. Of course we could update remove_event_file_dir()
to skip this file, but I am sure we should not do this.

Oleg.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/