Bluetooth: hci_uart: check for missing tty operations in protocol handlers
Subsystems (2)
PATCH 1/1 Bluetooth: hci_uart: check for missing tty operations in protocol handlers
Certain ttys operations (pty_unix98_ops) lack tiocmget() and tiocmset() functions which are called by the certain HCI UART protocols (hci_ath, hci_bcm, hci_intel, hci_mrvl, hci_qca) via hci_uart_set_flow_control() or directly. This leads to an execution at NULL and can be triggered by an unprivileged user. Fix this by adding a check for the missing tty operations to the protocols which use them.
This fixes CVE-2019-10207.
--- drivers/bluetooth/hci_ath.c | 3 +++ drivers/bluetooth/hci_bcm.c | 5 +++++ drivers/bluetooth/hci_intel.c | 3 +++ drivers/bluetooth/hci_mrvl.c | 3 +++ drivers/bluetooth/hci_qca.c | 4 ++++ 5 files changed, 18 insertions(+)
drivers/bluetooth/hci_ath.c +3 -0
@@ -98,6 +98,9 @@ static int ath_open(struct hci_uart *hu)BT_DBG ("hu %p" , hu); +if (!hu->tty->driver->ops->tiocmget || !hu->tty->driver->ops->tiocmset) +return -ENOTSUPP ; + ath =kzalloc (sizeof (*ath),GFP_KERNEL );if (!ath)return -ENOMEM ;
drivers/bluetooth/hci_bcm.c +5 -0
@@ -406,6 +406,11 @@ static int bcm_open(struct hci_uart *hu)bt_dev_dbg (hu->hdev,"hu %p" , hu); +#ifdef CONFIG_PM +if (!hu->tty->driver->ops->tiocmget || !hu->tty->driver->ops->tiocmset) +return -ENOTSUPP ; +#endif + bcm =kzalloc (sizeof (*bcm),GFP_KERNEL );if (!bcm)return -ENOMEM ;
drivers/bluetooth/hci_intel.c +3 -0
@@ -391,6 +391,9 @@ static int intel_open(struct hci_uart *hu)BT_DBG ("hu %p" , hu); +if (!hu->tty->driver->ops->tiocmget || !hu->tty->driver->ops->tiocmset) +return -ENOTSUPP ; + intel =kzalloc (sizeof (*intel),GFP_KERNEL );if (!intel)return -ENOMEM ;
drivers/bluetooth/hci_mrvl.c +3 -0
@@ -59,6 +59,9 @@ static int mrvl_open(struct hci_uart *hu)BT_DBG ("hu %p" , hu); +if (!hu->tty->driver->ops->tiocmget || !hu->tty->driver->ops->tiocmset) +return -ENOTSUPP ; + mrvl =kzalloc (sizeof (*mrvl),GFP_KERNEL );if (!mrvl)return -ENOMEM ;
drivers/bluetooth/hci_qca.c +4 -0
@@ -28,6 +28,7 @@#include <linux/platform_device.h> #include <linux/regulator/consumer.h> #include <linux/serdev.h> +#include <linux/tty.h> #include <asm/unaligned.h> #include <net/bluetooth/bluetooth.h>
@@ -473,6 +474,9 @@ static int qca_open(struct hci_uart *hu)BT_DBG ("hu %p qca_open" , hu); +if (!hu->tty->driver->ops->tiocmget || !hu->tty->driver->ops->tiocmset) +return -ENOTSUPP ; + qca =kzalloc (sizeof (struct qca_data ),GFP_KERNEL );if (!qca)return -ENOMEM ;
Signature · 1 line
2.20.1
Marcel Holtmann
maintainerHi Vladis,
Quoted 45 lines
Certain ttys operations (
pty_unix98_ops) lacktiocmget()andtiocmset()functions which are called by the certain HCI UART protocols (hci_ath,hci_bcm,hci_intel,hci_mrvl,hci_qca) viahci_uart_set_flow_control()or directly. This leads to an execution at NULL and can be triggered by an unprivileged user. Fix this by adding a check for the missing tty operations to the protocols which use them.This fixes CVE-2019-10207.
Reported-by [email protected]Cc [email protected] # v2.6.36+Fixes b3190df62861 ("Bluetooth: Support for Atheros AR300x serial chip")Fixes 118612fb9165 ("Bluetooth: hci_bcm: Add suspend/resume PM functions")Fixes ff2895592f0f ("Bluetooth: hci_intel: Add Intel baudrate configuration support")Fixes 162f812f23ba ("Bluetooth: hci_uart: Add Marvell support")Fixes fa9ad876b8e0 ("Bluetooth: hci_qca: Add support for Qualcomm Bluetooth chip wcn3990")Signed-off-by Vladis Dronov <[email protected]>---
drivers/bluetooth/hci_ath.c | 3 +++ drivers/bluetooth/hci_bcm.c | 5 +++++ drivers/bluetooth/hci_intel.c | 3 +++ drivers/bluetooth/hci_mrvl.c | 3 +++ drivers/bluetooth/hci_qca.c | 4 ++++ 5 files changed, 18 insertions(+)drivers/bluetooth/hci_ath.c +3
-0@@ -98,6 +98,9 @@ static int ath_open(struct hci_uart *hu)BT_DBG ("hu %p" , hu); +if (!hu->tty->driver->ops->tiocmget || !hu->tty->driver->ops->tiocmset) +return -ENOTSUPP ; + ath =kzalloc (sizeof (*ath),GFP_KERNEL );if (!ath)return -ENOMEM ;drivers/bluetooth/hci_bcm.c +0
-0
drivers/bluetooth/hci_bcm.c static int bcm_open(struct hci_uart *hu) · line 413
Quoted 8 lines
drivers/bluetooth/hci_bcm.c +5
-0@@ -406,6 +406,11 @@ static int bcm_open(struct hci_uart *hu)bt_dev_dbg (hu->hdev,"hu %p" , hu); +#ifdef CONFIG_PM +if (!hu->tty->driver->ops->tiocmget || !hu->tty->driver->ops->tiocmset) +return -ENOTSUPP ; +#endif +
why is this one hidden behind CONFIG_PM? The general baud rate changes are independent of runtime power management support.
And I would introduce a bool hci_uart_has_tiocm_support(struct hci_uart *) helper.
Regards
Marcel
Hello, Marcel,
Quoted 2 lines
why is this one hidden behind
CONFIG_PM? The general baud rate changes are independent of runtime power management support.
hci_bcm calls hci_uart_set_flow_control() only from functions hidden behind #ifdef-CONFIG_PM (surely this can change in the future), and so without CONFIG_PM set it cannot hit the bug (as of now). So I've hidden the check for tiocm[gs]et() behind #ifdef-CONFIG_PM too.
If you tell me it is better to remove this #ifdef, I'll remove it.
Quoted 2 lines
And I would introduce a bool
hci_uart_has_tiocm_support(struct hci_uart *)helper.
Great, I will add it to the v2 fix. I guess a good place for it is hci_ldisc.c, near hci_uart_set_flow_control(), isn't it?
Best regards,
Vladis Dronov | Red Hat, Inc. | The Core Kernel | Senior Software Engineer
----- Original Message -----
Quoted 76 lines
From: "Marcel Holtmann" <[email protected]>
To: "Vladis Dronov" <[email protected]>
Cc: "Johan Hedberg" <[email protected]>, [email protected], [email protected], "Suraj Sumangala" <[email protected]>, "Frederic Danis" <[email protected]>, "Loic Poulain" <[email protected]>, "Balakrishna Godavarthi" <[email protected]>, [email protected]
Sent: Thursday, July 25, 2019 2:59:42 PM
Subject: Re: [PATCH] Bluetooth:hci_uart: check for missing tty operations in protocol handlersHi Vladis,
Quoted 48 lines
Certain ttys operations (
pty_unix98_ops) lacktiocmget()andtiocmset()functions which are called by the certain HCI UART protocols (hci_ath,hci_bcm,hci_intel,hci_mrvl,hci_qca) viahci_uart_set_flow_control()or directly. This leads to an execution at NULL and can be triggered by an unprivileged user. Fix this by adding a check for the missing tty operations to the protocols which use them.This fixes CVE-2019-10207.
Link:
https://syzkaller.appspot.com/bug?id=1b42faa2848963564a5b1b7f8c837ea7b55ffa50Reported-by [email protected]Cc [email protected] # v2.6.36+Fixes b3190df62861 ("Bluetooth: Support for Atheros AR300x serial chip")Fixes 118612fb9165 ("Bluetooth: hci_bcm: Add suspend/resume PM functions")Fixes ff2895592f0f ("Bluetooth: hci_intel: Add Intel baudrateconfiguration support")
Fixes 162f812f23ba ("Bluetooth: hci_uart: Add Marvell support")Fixes fa9ad876b8e0 ("Bluetooth: hci_qca: Add support for QualcommBluetooth chip wcn3990")
Signed-off-by Vladis Dronov <[email protected]>---
drivers/bluetooth/hci_ath.c | 3 +++ drivers/bluetooth/hci_bcm.c | 5 +++++ drivers/bluetooth/hci_intel.c | 3 +++ drivers/bluetooth/hci_mrvl.c | 3 +++ drivers/bluetooth/hci_qca.c | 4 ++++ 5 files changed, 18 insertions(+)drivers/bluetooth/hci_ath.c +3
-0@@ -98,6 +98,9 @@ static int ath_open(struct hci_uart *hu)BT_DBG ("hu %p" , hu); +if (!hu->tty->driver->ops->tiocmget || !hu->tty->driver->ops->tiocmset) +return -ENOTSUPP ; + ath =kzalloc (sizeof (*ath),GFP_KERNEL );if (!ath)return -ENOMEM ;drivers/bluetooth/hci_bcm.c +0
-0drivers/bluetooth/hci_bcm.c static int bcm_open(struct hci_uart *hu) · line 413
Quoted 8 lines
drivers/bluetooth/hci_bcm.c +5
-0@@ -406,6 +406,11 @@ static int bcm_open(struct hci_uart *hu)bt_dev_dbg (hu->hdev,"hu %p" , hu); +#ifdef CONFIG_PM +if (!hu->tty->driver->ops->tiocmget || !hu->tty->driver->ops->tiocmset) +return -ENOTSUPP ; +#endif +why is this one hidden behind
CONFIG_PM? The general baud rate changes are independent of runtime power management support.And I would introduce a bool
hci_uart_has_tiocm_support(struct hci_uart *)helper.Regards
Marcel
Marcel Holtmann
maintainerHi Vladis,
Quoted 9 lines
Quoted 2 lines
why is this one hidden behind
CONFIG_PM? The general baud rate changes are independent of runtime power management support.
hci_bcmcallshci_uart_set_flow_control()only from functions hidden behind #ifdef-CONFIG_PM (surely this can change in the future), and so withoutCONFIG_PMset it cannot hit the bug (as of now). So I've hidden the check for tiocm[gs]et() behind #ifdef-CONFIG_PM too.If you tell me it is better to remove this #ifdef, I'll remove it.
just remove it since I think that is a coincidence actually. And I think you also need to add a check in hci_h5.c since that got Realtek support, it should also use flow control or baud rate settings.
Quoted 5 lines
Quoted 2 lines
And I would introduce a bool
hci_uart_has_tiocm_support(struct hci_uart *)helper.Great, I will add it to the v2 fix. I guess a good place for it is
hci_ldisc.c, nearhci_uart_set_flow_control(), isn't it?
Sounds good.
Regards
Marcel