public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Alan Cox <alan@redhat.com>
To: linux-kernel@vger.kernel.org
Subject: [PATCH 36/76] tty: Make get_current_tty use a kref
Date: Sun, 05 Oct 2008 17:10:42 +0100	[thread overview]
Message-ID: <20081005161021.1997.38953.stgit@localhost.localdomain> (raw)
In-Reply-To: <20081005160231.1997.10462.stgit@localhost.localdomain>

We now return a kref covered tty reference. That ensures the tty structure
doesn't go away when you have a return from get_current_tty. This is not
enough to protect you from most of the resources being freed behind your
back - yet.

[Updated to include fixes for SELinux problems found by Andrew Morton and
 an s390 leak found while debugging the former]

Signed-off-by: Alan Cox <alan@redhat.com>
---

 drivers/char/tty_io.c      |   10 ++++++----
 drivers/s390/char/fs3270.c |    3 ++-
 fs/dquot.c                 |    6 +++---
 security/selinux/hooks.c   |    3 ++-
 4 files changed, 13 insertions(+), 9 deletions(-)

diff --git a/drivers/char/tty_io.c b/drivers/char/tty_io.c
index 9a76db3..4c0e4ed 100644
--- a/drivers/char/tty_io.c
+++ b/drivers/char/tty_io.c
@@ -786,12 +786,12 @@ void disassociate_ctty(int on_exit)
 	tty = get_current_tty();
 	if (tty) {
 		tty_pgrp = get_pid(tty->pgrp);
-		lock_kernel();
 		mutex_unlock(&tty_mutex);
-		/* XXX: here we race, there is nothing protecting tty */
+		lock_kernel();
 		if (on_exit && tty->driver->type != TTY_DRIVER_TYPE_PTY)
 			tty_vhangup(tty);
 		unlock_kernel();
+		tty_kref_put(tty);
 	} else if (on_exit) {
 		struct pid *old_pgrp;
 		spin_lock_irq(&current->sighand->siglock);
@@ -819,7 +819,6 @@ void disassociate_ctty(int on_exit)
 	spin_unlock_irq(&current->sighand->siglock);
 
 	mutex_lock(&tty_mutex);
-	/* It is possible that do_tty_hangup has free'd this tty */
 	tty = get_current_tty();
 	if (tty) {
 		unsigned long flags;
@@ -829,6 +828,7 @@ void disassociate_ctty(int on_exit)
 		tty->session = NULL;
 		tty->pgrp = NULL;
 		spin_unlock_irqrestore(&tty->ctrl_lock, flags);
+		tty_kref_put(tty);
 	} else {
 #ifdef TTY_DEBUG_HANGUP
 		printk(KERN_DEBUG "error attempted to write to tty [0x%p]"
@@ -1806,6 +1806,8 @@ retry_open:
 		index = tty->index;
 		filp->f_flags |= O_NONBLOCK; /* Don't let /dev/tty block */
 		/* noctty = 1; */
+		/* FIXME: Should we take a driver reference ? */
+		tty_kref_put(tty);
 		goto got_driver;
 	}
 #ifdef CONFIG_VT
@@ -3135,7 +3137,7 @@ struct tty_struct *get_current_tty(void)
 {
 	struct tty_struct *tty;
 	WARN_ON_ONCE(!mutex_is_locked(&tty_mutex));
-	tty = current->signal->tty;
+	tty = tty_kref_get(current->signal->tty);
 	/*
 	 * session->tty can be changed/cleared from under us, make sure we
 	 * issue the load. The obtained pointer, when not NULL, is valid as
diff --git a/drivers/s390/char/fs3270.c b/drivers/s390/char/fs3270.c
index d18e6d2..3ef5425 100644
--- a/drivers/s390/char/fs3270.c
+++ b/drivers/s390/char/fs3270.c
@@ -430,11 +430,12 @@ fs3270_open(struct inode *inode, struct file *filp)
 		mutex_lock(&tty_mutex);
 		tty = get_current_tty();
 		if (!tty || tty->driver->major != IBM_TTY3270_MAJOR) {
-			mutex_unlock(&tty_mutex);
+			tty_kref_put(tty);
 			rc = -ENODEV;
 			goto out;
 		}
 		minor = tty->index + RAW3270_FIRSTMINOR;
+		tty_kref_put(tty);
 		mutex_unlock(&tty_mutex);
 	}
 	/* Check if some other program is already using fullscreen mode. */
diff --git a/fs/dquot.c b/fs/dquot.c
index 8ec4d6c..7417a6c 100644
--- a/fs/dquot.c
+++ b/fs/dquot.c
@@ -897,8 +897,9 @@ static void print_warning(struct dquot *dquot, const int warntype)
 
 	mutex_lock(&tty_mutex);
 	tty = get_current_tty();
+	mutex_unlock(&tty_mutex);
 	if (!tty)
-		goto out_lock;
+		return;
 	tty_write_message(tty, dquot->dq_sb->s_id);
 	if (warntype == QUOTA_NL_ISOFTWARN || warntype == QUOTA_NL_BSOFTWARN)
 		tty_write_message(tty, ": warning, ");
@@ -926,8 +927,7 @@ static void print_warning(struct dquot *dquot, const int warntype)
 			break;
 	}
 	tty_write_message(tty, msg);
-out_lock:
-	mutex_unlock(&tty_mutex);
+	tty_kref_put(tty);
 }
 #endif
 
diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
index 03fc6a8..c856db8 100644
--- a/security/selinux/hooks.c
+++ b/security/selinux/hooks.c
@@ -2122,6 +2122,7 @@ static inline void flush_unauthorized_files(struct files_struct *files)
 
 	mutex_lock(&tty_mutex);
 	tty = get_current_tty();
+	mutex_unlock(&tty_mutex);
 	if (tty) {
 		file_list_lock();
 		file = list_entry(tty->tty_files.next, typeof(*file), f_u.fu_list);
@@ -2138,8 +2139,8 @@ static inline void flush_unauthorized_files(struct files_struct *files)
 			}
 		}
 		file_list_unlock();
+		tty_kref_put(tty);
 	}
-	mutex_unlock(&tty_mutex);
 	/* Reset controlling tty. */
 	if (drop_tty)
 		no_tty();


  parent reply	other threads:[~2008-10-05 16:17 UTC|newest]

Thread overview: 88+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-10-05 16:04 [PATCH 00/76] Queued TTY Patches Alan Cox
2008-10-05 16:04 ` [PATCH 01/76] drivers/serial/crisv10.c: add missing put_tty_driver Alan Cox
2008-10-05 16:04 ` [PATCH 02/76] drivers/char/hvc_console.c: adjust call to put_tty_driver Alan Cox
2008-10-05 16:04 ` [PATCH 03/76] coldfire: scheduled SERIAL_COLDFIRE removal Alan Cox
2008-10-05 16:05 ` [PATCH 04/76] epca: call tty_port_init Alan Cox
2008-10-05 16:05 ` [PATCH 05/76] 8250: Remove NR_IRQ usage Alan Cox
2008-10-05 16:05 ` [PATCH 06/76] Blackfin Serial Driver: use __initdata for data, not __init Alan Cox
2008-10-05 16:05 ` [PATCH 07/76] Blackfin Serial Driver: Fix bug - should suspend/resume/remove all uart ports Alan Cox
2008-10-05 16:05 ` [PATCH 08/76] Blackfin Serial Driver: trim trailing whitespace -- no functional changes Alan Cox
2008-10-05 16:05 ` [PATCH 09/76] Blackfin Serial Driver: move common variables out of serial headers and into the serial driver Alan Cox
2008-10-05 16:05 ` [PATCH 10/76] Blackfin Serial Driver: Remove useless stop Alan Cox
2008-10-05 16:06 ` [PATCH 11/76] Blackfin Serial Driver: Fix bug - Don't call tx_stop in tx_transfer Alan Cox
2008-10-05 16:06 ` [PATCH 12/76] Blackfin Serial Driver: Fix bug - ircp fails on sir over Blackfin UART Alan Cox
2008-10-05 16:06 ` [PATCH 13/76] Blackfin Serial Driver: Fix bug - request UART2/3 peripheral mapped interrupts in PIO mode Alan Cox
2008-10-05 16:06 ` [PATCH 14/76] tty: move tioclinux from a special case Alan Cox
2008-10-05 16:06 ` [PATCH 15/76] uml: small cleanups and note bugs to be dealt with by uml authors Alan Cox
2008-10-05 16:06 ` [PATCH 16/76] Char: merge ip2main and ip2base Alan Cox
2008-10-05 16:06 ` [PATCH 17/76] ip2, cleanup globals Alan Cox
2008-10-05 16:07 ` [PATCH 18/76] ip2, fix sparse warnings Alan Cox
2008-10-05 16:07 ` [PATCH 19/76] ip2, init/deinit cleanup Alan Cox
2008-10-05 16:07 ` [PATCH 20/76] ip2: avoid add_timer with pending timer Alan Cox
2008-10-05 16:07 ` [PATCH 21/76] audit: Handle embedded NUL in TTY input auditing Alan Cox
2008-10-05 16:07 ` [PATCH 22/76] serial: Make uart_port's ioport "unsigned long" Alan Cox
2008-10-06 12:53   ` Josh Boyer
2008-10-05 16:07 ` [PATCH 23/76] tty: split the buffering from tty_io Alan Cox
2008-10-05 16:07 ` [PATCH 24/76] tty: Split tty_port into its own file Alan Cox
2008-10-05 16:08 ` [PATCH 25/76] pps: Reserve a line discipline number for PPS Alan Cox
2008-10-05 16:08 ` [PATCH 26/76] tty: Add a kref count Alan Cox
2008-10-06 10:20   ` Louis Rilling
2008-10-06 10:52     ` Alan Cox
2008-10-05 16:08 ` [PATCH 27/76] tty: use krefs to protect driver module counts Alan Cox
2008-10-05 16:08 ` [PATCH 28/76] Char: cyclades. remove bogus iomap Alan Cox
2008-10-05 16:08 ` [PATCH 29/76] Char: sx, fix io unmapping Alan Cox
2008-10-05 16:08 ` [PATCH 30/76] 8250: remove a few inlines of dubious value Alan Cox
2008-10-05 16:09 ` [PATCH 31/76] tty: Cris has a nice RS485 ioctl so we should steal it Alan Cox
2008-10-05 16:09 ` [PATCH 32/76] tty: ipw need reworking Alan Cox
2008-10-05 16:09 ` [PATCH 33/76] tty: Add termiox Alan Cox
2008-10-05 16:10 ` [PATCH 34/76] tty: Termios locking - sort out real_tty confusions and lock reads Alan Cox
2008-10-05 16:10 ` [PATCH 35/76] tty: compare the tty winsize Alan Cox
2008-10-05 16:10 ` Alan Cox [this message]
2008-10-06 10:35   ` [PATCH 36/76] tty: Make get_current_tty use a kref Louis Rilling
2008-10-06 10:56     ` Alan Cox
2008-10-05 16:10 ` [PATCH 37/76] tty: Move tty_write_message out of kernel/printk Alan Cox
2008-10-05 16:11 ` [PATCH 38/76] tty: usb-serial krefs Alan Cox
2008-10-06  5:36   ` Greg KH
2008-10-06  9:02     ` Alan Cox
2008-10-07  4:13       ` Greg KH
2008-10-05 16:12 ` [PATCH 39/76] tty: kref usage for isicom and moxa Alan Cox
2008-10-05 16:13 ` [PATCH 40/76] stallion: Use krefs Alan Cox
2008-10-05 16:14 ` [PATCH 41/76] mxser: Switch to kref tty Alan Cox
2008-10-05 16:15 ` [PATCH 42/76] tty: the vhangup syscall is racy Alan Cox
2008-10-05 16:15 ` [PATCH 43/76] tty: Redo current tty locking Alan Cox
2008-10-09 13:21   ` Derek Fawcus
2008-10-05 16:16 ` [PATCH 44/76] tty: Fix abusers of current->sighand->tty Alan Cox
2008-10-05 16:17 ` [PATCH 45/76] pty: If the administrator creates a device for a ptmx slave we should not error Alan Cox
2008-10-05 16:18 ` [PATCH 46/76] vt: remove bogus lock dropping Alan Cox
2008-10-05 16:18 ` [PATCH 47/76] tty: shutdown method Alan Cox
2008-10-05 16:18 ` [PATCH 48/76] tty: Remove more special casing and out of place code Alan Cox
2008-10-05 16:19 ` [PATCH 49/76] tty: Move parts of tty_init_dev into new functions Alan Cox
2008-10-05 16:19 ` [PATCH 50/76] tty: Clean up the tty_init_dev changes further Alan Cox
2008-10-05 16:20 ` [PATCH 51/76] tty: kref the tty driver object Alan Cox
2008-10-05 16:20 ` [PATCH 52/76] tty: More driver operations Alan Cox
2008-10-05 16:20 ` [PATCH 53/76] tty: Finish fixing up the init_dev interface to use ERR_PTR Alan Cox
2008-10-05 16:21 ` [PATCH 54/76] tty: extract the pty init time special cases Alan Cox
2008-10-05 16:21 ` [PATCH 55/76] Move tty lookup/reopen to caller Alan Cox
2008-10-05 16:21 ` [PATCH 56/76] Add an instance parameter devpts interfaces Alan Cox
2008-10-05 16:21 ` [PATCH 57/76] Simplify devpts_get_tty() Alan Cox
2008-10-05 16:21 ` [PATCH 58/76] Simplify devpts_pty_new() Alan Cox
2008-10-05 16:21 ` [PATCH 59/76] Simplify devpts_pty_kill Alan Cox
2008-10-05 16:22 ` [PATCH 60/76] pty: Coding style and polish Alan Cox
2008-10-05 16:22 ` [PATCH 61/76] pty: Fix allocation failure double free Alan Cox
2008-10-05 16:22 ` [PATCH 62/76] pty: simplify unix98 allocation Alan Cox
2008-10-05 16:22 ` [PATCH 63/76] tty: simplify ktermios allocation Alan Cox
2008-10-05 16:22 ` [PATCH 64/76] tty: some ICANON magic is in the wrong places Alan Cox
2008-10-05 16:23 ` [PATCH 65/76] tty: Fallout from tty-move-canon-specials Alan Cox
2008-10-05 16:23 ` [PATCH 66/76] tty: fix up gigaset a bit Alan Cox
2008-10-05 16:23 ` [PATCH 67/76] tty: Remove lots of NULL checks Alan Cox
2008-10-06  8:58   ` Geert Uytterhoeven
2008-10-06 10:50     ` Alan Cox
2008-10-05 16:23 ` [PATCH 68/76] tty: Minor tidyups and document fixes for n_tty Alan Cox
2008-10-05 16:24 ` [PATCH 69/76] hso: net driver using tty without locking Alan Cox
2008-10-05 16:24 ` [PATCH 70/76] applicom: Fix an unchecked user ioctl range and an error return Alan Cox
2008-10-05 16:24 ` [PATCH 71/76] serial-make-uart_ports-ioport-unsigned-long-fix Alan Cox
2008-10-05 16:24 ` [PATCH 72/76] serial: allow 8250 to be used on sparc Alan Cox
2008-10-05 16:24 ` [PATCH 73/76] serial: fix device name reporting when minor space is shared between drivers Alan Cox
2008-10-05 16:25 ` [PATCH 74/76] tty: tty_io.c shadows sparse fix Alan Cox
2008-10-05 16:25 ` [PATCH 75/76] usb: fix pl2303 initialization Alan Cox
2008-10-05 16:25 ` [PATCH 76/76] ftdi: A few errors are err() that should be debug which causes much spewage Alan Cox

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=20081005161021.1997.38953.stgit@localhost.localdomain \
    --to=alan@redhat.com \
    --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