Re: KASAN: use-after-free Read in device_release_driver_internal

From: Andrey Konovalov
Date: Wed Aug 07 2019 - 09:45:30 EST


On Wed, Aug 7, 2019 at 3:44 PM Andrey Konovalov <andreyknvl@xxxxxxxxxx> wrote:
>
> On Wed, Aug 7, 2019 at 3:44 PM Andrey Konovalov <andreyknvl@xxxxxxxxxx> wrote:
> >
> > On Wed, Aug 7, 2019 at 3:38 PM Oliver Neukum <oneukum@xxxxxxxx> wrote:
> > >
> > > Am Dienstag, den 06.08.2019, 14:50 +0200 schrieb Andrey Konovalov:
> > > > On Tue, Aug 6, 2019 at 2:36 PM Oliver Neukum <oneukum@xxxxxxxx> wrote:
> > > > >
> > > > > Am Donnerstag, den 01.08.2019, 14:47 -0400 schrieb Alan Stern:
> > > > > >
> > > > > > I think this must be caused by an unbalanced refcount. That is,
> > > > > > something must drop one more reference to the device than it takes.
> > > > > > That would explain why the invalid access occurs inside a single
> > > > > > bus_remove_device() call, between the klist_del() and
> > > > > > device_release_driver().
> > > > > >
> > > > > > The kernel log indicates that the device was probed by rndis_wlan,
> > > > > > rndis_host, and cdc_acm, all of which got errors because of the
> > > > > > device's bogus descriptors. Probably one of them is messing up the
> > > > > > refcount.
> > > > >
> > > > > Hi,
> > > > >
> > > > > you made me look at cdc-acm. I suspect
> > > > >
> > > > > cae2bc768d176bfbdad7035bbcc3cdc973eb7984 ("usb: cdc-acm: Decrement tty port's refcount if probe() fail")
> > > > >
> > > > > is buggy decrementing the refcount on the interface in destroy()
> > > > > even before the refcount is increased.
> > > > >
> > > > > Unfortunately I cannot tell from the bug report how many and which
> > > > > interfaces the emulated test device has. Hence it is unclear to me,
> > > > > when exactly probe() would fail cdc-acm.
> > > > >
> > > > > If you agree. I am attaching a putative fix.
> > > >
> > > > Let's see if it fixes the issue.
> > > >
> > > > #syz fix: https://github.com/google/kasan.git 6a3599ce
> > >
> > > Hi,
> > >
> > > did this ever produce a result? I saw none.
> >
> > Hm, that's weird, maybe that's caused by putting the bot into CC. Let
> > me try that again.
> >
> > #syz fix: https://github.com/google/kasan.git 6a3599ce
>
> Oh, wait, it should be syz test =)
>
> #syz test: https://github.com/google/kasan.git 6a3599ce

And now I forgot the patch :(

#syz test: https://github.com/google/kasan.git 6a3599ce

>
> >
> > >
> > > Regards
> > > Oliver
> > >
> > > --
> > > You received this message because you are subscribed to the Google Groups "syzkaller-bugs" group.
> > > To unsubscribe from this group and stop receiving emails from it, send an email to syzkaller-bugs+unsubscribe@xxxxxxxxxxxxxxxxx
> > > To view this discussion on the web visit https://groups.google.com/d/msgid/syzkaller-bugs/1565185131.15973.1.camel%40suse.com.
From 6b31904e6cf75f89441e308b9e428a1de7728fd8 Mon Sep 17 00:00:00 2001
From: Oliver Neukum <oneukum@xxxxxxxx>
Date: Tue, 6 Aug 2019 14:34:27 +0200
Subject: [PATCH] usb: cdc-acm: make sure a refcount is taken early enough

destroy() will decrement the refcount on the interface, so that
it needs to be taken so early that it never undercounts.

Signed-off-by: Oliver Neukum <oneukum@xxxxxxxx>
---
drivers/usb/class/cdc-acm.c | 12 +++++++-----
1 file changed, 7 insertions(+), 5 deletions(-)

diff --git a/drivers/usb/class/cdc-acm.c b/drivers/usb/class/cdc-acm.c
index 183b41753c98..28e3de775ada 100644
--- a/drivers/usb/class/cdc-acm.c
+++ b/drivers/usb/class/cdc-acm.c
@@ -1301,10 +1301,6 @@ static int acm_probe(struct usb_interface *intf,
tty_port_init(&acm->port);
acm->port.ops = &acm_port_ops;

- minor = acm_alloc_minor(acm);
- if (minor < 0)
- goto alloc_fail1;
-
ctrlsize = usb_endpoint_maxp(epctrl);
readsize = usb_endpoint_maxp(epread) *
(quirks == SINGLE_RX_URB ? 1 : 2);
@@ -1312,6 +1308,13 @@ static int acm_probe(struct usb_interface *intf,
acm->writesize = usb_endpoint_maxp(epwrite) * 20;
acm->control = control_interface;
acm->data = data_interface;
+
+ usb_get_intf(acm->control); /* undone in destroy() */
+
+ minor = acm_alloc_minor(acm);
+ if (minor < 0)
+ goto alloc_fail1;
+
acm->minor = minor;
acm->dev = usb_dev;
if (h.usb_cdc_acm_descriptor)
@@ -1458,7 +1461,6 @@ static int acm_probe(struct usb_interface *intf,
usb_driver_claim_interface(&acm_driver, data_interface, acm);
usb_set_intfdata(data_interface, acm);

- usb_get_intf(control_interface);
tty_dev = tty_port_register_device(&acm->port, acm_tty_driver, minor,
&control_interface->dev);
if (IS_ERR(tty_dev)) {
--
2.16.4