lkml.org 
[lkml]   [2020]   [May]   [20]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
SubjectRe: [PATCH] tty: hvc: Fix data abort due to race in hvc_open
From
Date
On 15. 05. 20, 1:22, rananta@codeaurora.org wrote:
> On 2020-05-13 00:04, Greg KH wrote:
>> On Tue, May 12, 2020 at 02:39:50PM -0700, rananta@codeaurora.org wrote:
>>> On 2020-05-12 01:25, Greg KH wrote:
>>> > On Tue, May 12, 2020 at 09:22:15AM +0200, Jiri Slaby wrote:
>>> > > commit bdb498c20040616e94b05c31a0ceb3e134b7e829
>>> > > Author: Jiri Slaby <jslaby@suse.cz>
>>> > > Date:   Tue Aug 7 21:48:04 2012 +0200
>>> > >
>>> > >     TTY: hvc_console, add tty install
>>> > >
>>> > > added hvc_install but did not move 'tty->driver_data = NULL;' from
>>> > > hvc_open's fail path to hvc_cleanup.
>>> > >
>>> > > IOW hvc_open now NULLs tty->driver_data even for another task which
>>> > > opened the tty earlier. The same holds for
>>> > > "tty_port_tty_set(&hp->port,
>>> > > NULL);" there. And actually "tty_port_put(&hp->port);" is also
>>> > > incorrect
>>> > > for the 2nd task opening the tty.
>>> > >

...

> These are the traces you get when the issue happens:
> [  154.212291] hvc_install called for pid: 666
> [  154.216705] hvc_open called for pid: 666
> [  154.233657] hvc_open: request_irq failed with rc -22.
> [  154.238934] hvc_open called for pid: 678
> [  154.243012] Unable to handle kernel NULL pointer dereference at
> virtual address 00000000000000c4
> # hvc_install isn't called for pid: 678 as the file wasn't closed yet.

Nice. Does the attached help?

I wonder how comes the tty_port_put in hvc_open does not cause a UAF? I
would say hvc_open fails, tty_port_put is called. It decrements the
reference taken in hvc_install. So far so good.

Now, this should happen IMO:
tty_open
-> hvc_open (fails)
-> tty_port_put
-> tty_release
-> tty_release_struct
-> tty_kref_put
-> queue_release_one_tty
SCHEDULED WORKQUEUE
release_one_tty
-> hvc_cleanup
-> tty_port_put (should die terrible death now)

What am I missing?

thanks,
--
js
suse labs
From d891cdfcbd3b41eb23ddfc8d9e6cbe038ff8fb72 Mon Sep 17 00:00:00 2001
From: Jiri Slaby <jslaby@suse.cz>
Date: Wed, 20 May 2020 11:29:25 +0200
Subject: [PATCH] hvc_console: fix open

Signed-off-by: Jiri Slaby <jslaby@suse.cz>
---
drivers/tty/hvc/hvc_console.c | 23 ++++++++---------------
1 file changed, 8 insertions(+), 15 deletions(-)

diff --git a/drivers/tty/hvc/hvc_console.c b/drivers/tty/hvc/hvc_console.c
index 436cc51c92c3..cdcc64ea2554 100644
--- a/drivers/tty/hvc/hvc_console.c
+++ b/drivers/tty/hvc/hvc_console.c
@@ -371,15 +371,14 @@ static int hvc_open(struct tty_struct *tty, struct file * filp)
* tty fields and return the kref reference.
*/
if (rc) {
- tty_port_tty_set(&hp->port, NULL);
- tty->driver_data = NULL;
- tty_port_put(&hp->port);
printk(KERN_ERR "hvc_open: request_irq failed with rc %d.\n", rc);
- } else
+ } else {
/* We are ready... raise DTR/RTS */
if (C_BAUD(tty))
if (hp->ops->dtr_rts)
hp->ops->dtr_rts(hp, 1);
+ tty_port_set_initialized(&hp->port, true);
+ }

/* Force wakeup of the polling thread */
hvc_kick();
@@ -389,22 +388,12 @@ static int hvc_open(struct tty_struct *tty, struct file * filp)

static void hvc_close(struct tty_struct *tty, struct file * filp)
{
- struct hvc_struct *hp;
+ struct hvc_struct *hp = tty->driver_data;
unsigned long flags;

if (tty_hung_up_p(filp))
return;

- /*
- * No driver_data means that this close was issued after a failed
- * hvc_open by the tty layer's release_dev() function and we can just
- * exit cleanly because the kref reference wasn't made.
- */
- if (!tty->driver_data)
- return;
-
- hp = tty->driver_data;
-
spin_lock_irqsave(&hp->port.lock, flags);

if (--hp->port.count == 0) {
@@ -412,6 +401,9 @@ static void hvc_close(struct tty_struct *tty, struct file * filp)
/* We are done with the tty pointer now. */
tty_port_tty_set(&hp->port, NULL);

+ if (!tty_port_initialized(&hp->port))
+ return;
+
if (C_HUPCL(tty))
if (hp->ops->dtr_rts)
hp->ops->dtr_rts(hp, 0);
@@ -428,6 +420,7 @@ static void hvc_close(struct tty_struct *tty, struct file * filp)
* waking periodically to check chars_in_buffer().
*/
tty_wait_until_sent(tty, HVC_CLOSE_WAIT);
+ tty_port_set_initialized(&hp->port, false);
} else {
if (hp->port.count < 0)
printk(KERN_ERR "hvc_close %X: oops, count is %d\n",
--
2.26.2
\
 
 \ /
  Last update: 2020-05-20 11:39    [W:0.085 / U:1.620 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site