linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Peter Hurley <peter@hurleysoftware.com>
To: Patrick Donnelly <batrick@batbytes.com>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Jiri Slaby <jslaby@suse.cz>, open list <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH v3] tty: add missing rcu_read_lock for task_pgrp
Date: Sat, 11 Jul 2015 22:05:45 -0400	[thread overview]
Message-ID: <55A1CB79.8040300@hurleysoftware.com> (raw)
In-Reply-To: <1435622379-3170-1-git-send-email-batrick@batbytes.com>

On 06/29/2015 07:59 PM, Patrick Donnelly wrote:
> task_pgrp requires an rcu or tasklist lock to be obtained if the returned pid
> is to be dereferenced, which kill_pgrp does. Obtain an RCU lock for the
> duration of use.
> 
> Signed-off-by: Patrick Donnelly <batrick@batbytes.com>
> ---
>  drivers/tty/n_tty.c  | 12 ++++++++++--
>  drivers/tty/tty_io.c | 17 ++++++++++++-----
>  2 files changed, 22 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/tty/n_tty.c b/drivers/tty/n_tty.c
> index c9c27f6..0d631f8 100644
> --- a/drivers/tty/n_tty.c
> +++ b/drivers/tty/n_tty.c
> @@ -2137,6 +2137,8 @@ extern ssize_t redirected_tty_write(struct file *, const char __user *,
>  
>  static int job_control(struct tty_struct *tty, struct file *file)
>  {
> +	struct pid *pgrp;
> +
>  	/* Job control check -- must be done at start and after
>  	   every sleep (POSIX.1 7.1.1.4). */
>  	/* NOTE: not yet done after every sleep pending a thorough
> @@ -2146,18 +2148,24 @@ static int job_control(struct tty_struct *tty, struct file *file)
>  	    current->signal->tty != tty)
>  		return 0;
>  
> +	rcu_read_lock();
> +	pgrp = task_pgrp(current);
> +
>  	spin_lock_irq(&tty->ctrl_lock);
> +
>  	if (!tty->pgrp)
>  		printk(KERN_ERR "n_tty_read: no tty->pgrp!\n");
> -	else if (task_pgrp(current) != tty->pgrp) {
> +	else if (pgrp != tty->pgrp) {
>  		spin_unlock_irq(&tty->ctrl_lock);
>  		if (is_ignored(SIGTTIN) || is_current_pgrp_orphaned())
>  			return -EIO;

I just realized there's a missing rcu_read_unlock() from this early return.

Regards,
Peter Hurley


> -		kill_pgrp(task_pgrp(current), SIGTTIN, 1);
> +		kill_pgrp(pgrp, SIGTTIN, 1);
> +		rcu_read_unlock();
>  		set_thread_flag(TIF_SIGPENDING);
>  		return -ERESTARTSYS;
>  	}
>  	spin_unlock_irq(&tty->ctrl_lock);
> +	rcu_read_unlock();
>  	return 0;
>  }
>  
> diff --git a/drivers/tty/tty_io.c b/drivers/tty/tty_io.c
> index 57fc6ee..6bdfb98 100644
> --- a/drivers/tty/tty_io.c
> +++ b/drivers/tty/tty_io.c
> @@ -388,33 +388,40 @@ EXPORT_SYMBOL_GPL(tty_find_polling_driver);
>  int tty_check_change(struct tty_struct *tty)
>  {
>  	unsigned long flags;
> +	struct pid *pgrp;
>  	int ret = 0;
>  
>  	if (current->signal->tty != tty)
>  		return 0;
>  
> +	rcu_read_lock();
> +	pgrp = task_pgrp(current);
> +
>  	spin_lock_irqsave(&tty->ctrl_lock, flags);
>  
>  	if (!tty->pgrp) {
>  		printk(KERN_WARNING "tty_check_change: tty->pgrp == NULL!\n");
>  		goto out_unlock;
>  	}
> -	if (task_pgrp(current) == tty->pgrp)
> +	if (pgrp == tty->pgrp)
>  		goto out_unlock;
>  	spin_unlock_irqrestore(&tty->ctrl_lock, flags);
> +
>  	if (is_ignored(SIGTTOU))
> -		goto out;
> +		goto out_rcuunlock;
>  	if (is_current_pgrp_orphaned()) {
>  		ret = -EIO;
> -		goto out;
> +		goto out_rcuunlock;
>  	}
> -	kill_pgrp(task_pgrp(current), SIGTTOU, 1);
> +	kill_pgrp(pgrp, SIGTTOU, 1);
> +	rcu_read_unlock();
>  	set_thread_flag(TIF_SIGPENDING);
>  	ret = -ERESTARTSYS;
> -out:
>  	return ret;
>  out_unlock:
>  	spin_unlock_irqrestore(&tty->ctrl_lock, flags);
> +out_rcuunlock:
> +	rcu_read_unlock();
>  	return ret;
>  }
>  
> 


  parent reply	other threads:[~2015-07-12  2:05 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-06-27 21:17 [PATCH 1/2] Add missing rcu_read_lock for task_pgrp Patrick Donnelly
2015-06-27 21:17 ` [PATCH 2/2] Check tcsetpgrp p is a process group Patrick Donnelly
2015-06-27 23:26   ` Greg Kroah-Hartman
2015-06-28  0:53     ` Patrick Donnelly
2015-06-27 23:27   ` Greg Kroah-Hartman
2015-06-27 23:25 ` [PATCH 1/2] Add missing rcu_read_lock for task_pgrp Greg Kroah-Hartman
2015-06-28  0:51 ` [PATCH v2 1/2] tty: add " Patrick Donnelly
2015-06-28  0:51   ` [PATCH v2 2/2] tty: check tcsetpgrp p is a process group Patrick Donnelly
2015-06-28 16:07     ` Peter Hurley
2015-06-29  1:27       ` Patrick Donnelly
2015-06-28 15:23   ` [PATCH v2 1/2] tty: add missing rcu_read_lock for task_pgrp Peter Hurley
2015-06-28 17:20     ` Patrick Donnelly
2015-06-28 19:21       ` Peter Hurley
2015-06-28 19:27   ` Peter Hurley
2015-06-29 23:38     ` Patrick Donnelly
2015-06-29 23:59 ` [PATCH v3] " Patrick Donnelly
2015-07-09  2:07   ` Peter Hurley
2015-07-12  2:05   ` Peter Hurley [this message]
2015-07-12 22:42     ` Patrick Donnelly
2015-07-12 22:51 ` [PATCH v4] " Patrick Donnelly
2015-07-13  0:35   ` Peter Hurley
2015-07-20  9:26   ` Jiri Slaby
2015-07-20 19:40     ` Oleg Nesterov

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=55A1CB79.8040300@hurleysoftware.com \
    --to=peter@hurleysoftware.com \
    --cc=batrick@batbytes.com \
    --cc=gregkh@linuxfoundation.org \
    --cc=jslaby@suse.cz \
    --cc=linux-kernel@vger.kernel.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).