diff mbox series

[v2] usb: usb251xb: Lock i2c-bus segment the hub resides

Message ID 20190425223256.20270-1-fancer.lancer@gmail.com (mailing list archive)
State Superseded
Headers show
Series [v2] usb: usb251xb: Lock i2c-bus segment the hub resides | expand

Commit Message

Serge Semin April 25, 2019, 10:32 p.m. UTC
SMBus slave configuration is activated by CFG_SEL[1:0]=0x1 pins
state. This is the mode the hub is supposed to be to let this driver
work correctly. But a race condition might happen right after reset
is cleared due to CFG_SEL[0] pin being multiplexed with SMBus SCL
function. In case if the reset pin is handled by a i2c GPIO expander,
which is also placed at the same i2c-bus segment as the usb251x
SMB-interface connected to, then the hub reset clearance might
cause the CFG_SEL[0] being latched in unpredictable state. So
sometimes the hub configuration mode might be 0x1 (as expected),
but sometimes being 0x0, which doesn't imply to have the hub SMBus-slave
interface activated and consequently causes this driver failure.

In order to fix the problem we must make sure the GPIO-reset chip doesn't
reside the same i2c-bus segment as the SMBus-interface of the hub. If
it doesn't, we can safely block the segment for the time the reset is
cleared to prevent anyone generating a traffic at the i2c-bus SCL lane
connected to the CFG_SEL[0] pin. But if it does, nothing we can do, so
just return an error. If we locked the i2c-bus segment and tried to
communicate with the GPIO-expander, it would cause a deadlock. If we didn't
lock the i2c-bus segment, it would randomly cause the CFG_SEL[0] bit flip.

Signed-off-by: Serge Semin <fancer.lancer@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
Changelog v2
- Execute the check code only if GPIOLIB is enabled (reported by
  a build robot).

---
 drivers/usb/misc/usb251xb.c | 62 +++++++++++++++++++++++++++++++++++++
 1 file changed, 62 insertions(+)

Comments

Serge Semin April 26, 2019, 10:45 a.m. UTC | #1
On Fri, Apr 26, 2019 at 01:32:56AM +0300, Serge Semin wrote:

Ahh, sorry. Forget about this email. I'll resend it with the valid In-Reply-To
message-id.

-Sergey

> SMBus slave configuration is activated by CFG_SEL[1:0]=0x1 pins
> state. This is the mode the hub is supposed to be to let this driver
> work correctly. But a race condition might happen right after reset
> is cleared due to CFG_SEL[0] pin being multiplexed with SMBus SCL
> function. In case if the reset pin is handled by a i2c GPIO expander,
> which is also placed at the same i2c-bus segment as the usb251x
> SMB-interface connected to, then the hub reset clearance might
> cause the CFG_SEL[0] being latched in unpredictable state. So
> sometimes the hub configuration mode might be 0x1 (as expected),
> but sometimes being 0x0, which doesn't imply to have the hub SMBus-slave
> interface activated and consequently causes this driver failure.
> 
> In order to fix the problem we must make sure the GPIO-reset chip doesn't
> reside the same i2c-bus segment as the SMBus-interface of the hub. If
> it doesn't, we can safely block the segment for the time the reset is
> cleared to prevent anyone generating a traffic at the i2c-bus SCL lane
> connected to the CFG_SEL[0] pin. But if it does, nothing we can do, so
> just return an error. If we locked the i2c-bus segment and tried to
> communicate with the GPIO-expander, it would cause a deadlock. If we didn't
> lock the i2c-bus segment, it would randomly cause the CFG_SEL[0] bit flip.
> 
> Signed-off-by: Serge Semin <fancer.lancer@gmail.com>
> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> 
> ---
> Changelog v2
> - Execute the check code only if GPIOLIB is enabled (reported by
>   a build robot).
> 
> ---
>  drivers/usb/misc/usb251xb.c | 62 +++++++++++++++++++++++++++++++++++++
>  1 file changed, 62 insertions(+)
> 
> diff --git a/drivers/usb/misc/usb251xb.c b/drivers/usb/misc/usb251xb.c
> index 04684849d683..cdc80e8c2d8a 100644
> --- a/drivers/usb/misc/usb251xb.c
> +++ b/drivers/usb/misc/usb251xb.c
> @@ -12,6 +12,7 @@
>  
>  #include <linux/delay.h>
>  #include <linux/gpio/consumer.h>
> +#include <linux/gpio/driver.h>
>  #include <linux/i2c.h>
>  #include <linux/module.h>
>  #include <linux/nls.h>
> @@ -222,11 +223,51 @@ static const struct usb251xb_data usb2517i_data = {
>  	.product_str = "USB2517i",
>  };
>  
> +#ifdef CONFIG_GPIOLIB
> +static int usb251xb_check_dev_children(struct device *dev, void *child)
> +{
> +	if (dev->type == &i2c_adapter_type) {
> +		return device_for_each_child(dev, child,
> +					     usb251xb_check_dev_children);
> +	}
> +
> +	return (dev == child);
> +}
> +
> +static int usb251x_check_gpio_chip(struct usb251xb *hub)
> +{
> +	struct gpio_chip *gc = gpiod_to_chip(hub->gpio_reset);
> +	struct i2c_adapter *adap = hub->i2c->adapter;
> +	int ret;
> +
> +	if (!hub->gpio_reset)
> +		return 0;
> +
> +	if (!gc)
> +		return -EINVAL;
> +
> +	ret = usb251xb_check_dev_children(&adap->dev, gc->parent);
> +	if (ret) {
> +		dev_err(hub->dev, "Reset GPIO chip is at the same i2c-bus\n");
> +		return -EINVAL;
> +	}
> +
> +	return 0;
> +}
> +#else
> +static int usb251x_check_gpio_chip(struct usb251xb *hub)
> +{
> +	return 0;
> +}
> +#endif
> +
>  static void usb251xb_reset(struct usb251xb *hub, int state)
>  {
>  	if (!hub->gpio_reset)
>  		return;
>  
> +	i2c_lock_bus(hub->i2c->adapter, I2C_LOCK_SEGMENT);
> +
>  	gpiod_set_value_cansleep(hub->gpio_reset, state);
>  
>  	/* wait for hub recovery/stabilization */
> @@ -234,6 +275,8 @@ static void usb251xb_reset(struct usb251xb *hub, int state)
>  		usleep_range(500, 750);	/* >=500us at power on */
>  	else
>  		usleep_range(1, 10);	/* >=1us at power down */
> +
> +	i2c_unlock_bus(hub->i2c->adapter, I2C_LOCK_SEGMENT);
>  }
>  
>  static int usb251xb_connect(struct usb251xb *hub)
> @@ -621,6 +664,25 @@ static int usb251xb_probe(struct usb251xb *hub)
>  		}
>  	}
>  
> +	/*
> +	 * usb251x SMBus-slave SCL lane is muxed with CFG_SEL0 pin. So if anyone
> +	 * tries to work with the bus at the moment the hub reset is released,
> +	 * it may cause an invalid config being latched by usb251x. Particularly
> +	 * one of the config modes makes the hub loading a default registers
> +	 * value without SMBus-slave interface activation. If the hub
> +	 * accidentally gets this mode, this will cause the driver SMBus-
> +	 * functions failure. Normally we could just lock the SMBus-segment the
> +	 * hub i2c-interface resides for the device-specific reset timing. But
> +	 * the GPIO controller, which is used to handle the hub reset, might be
> +	 * placed at the same i2c-bus segment. In this case an error should be
> +	 * returned since we can't safely use the GPIO controller to clear the
> +	 * reset state (it may affect the hub configuration) and we can't lock
> +	 * the i2c-bus segment (it will cause a deadlock).
> +	 */
> +	err = usb251x_check_gpio_chip(hub);
> +	if (err)
> +		return err;
> +
>  	err = usb251xb_connect(hub);
>  	if (err) {
>  		dev_err(dev, "Failed to connect hub (%d)\n", err);
> -- 
> 2.21.0
>
diff mbox series

Patch

diff --git a/drivers/usb/misc/usb251xb.c b/drivers/usb/misc/usb251xb.c
index 04684849d683..cdc80e8c2d8a 100644
--- a/drivers/usb/misc/usb251xb.c
+++ b/drivers/usb/misc/usb251xb.c
@@ -12,6 +12,7 @@ 
 
 #include <linux/delay.h>
 #include <linux/gpio/consumer.h>
+#include <linux/gpio/driver.h>
 #include <linux/i2c.h>
 #include <linux/module.h>
 #include <linux/nls.h>
@@ -222,11 +223,51 @@  static const struct usb251xb_data usb2517i_data = {
 	.product_str = "USB2517i",
 };
 
+#ifdef CONFIG_GPIOLIB
+static int usb251xb_check_dev_children(struct device *dev, void *child)
+{
+	if (dev->type == &i2c_adapter_type) {
+		return device_for_each_child(dev, child,
+					     usb251xb_check_dev_children);
+	}
+
+	return (dev == child);
+}
+
+static int usb251x_check_gpio_chip(struct usb251xb *hub)
+{
+	struct gpio_chip *gc = gpiod_to_chip(hub->gpio_reset);
+	struct i2c_adapter *adap = hub->i2c->adapter;
+	int ret;
+
+	if (!hub->gpio_reset)
+		return 0;
+
+	if (!gc)
+		return -EINVAL;
+
+	ret = usb251xb_check_dev_children(&adap->dev, gc->parent);
+	if (ret) {
+		dev_err(hub->dev, "Reset GPIO chip is at the same i2c-bus\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+#else
+static int usb251x_check_gpio_chip(struct usb251xb *hub)
+{
+	return 0;
+}
+#endif
+
 static void usb251xb_reset(struct usb251xb *hub, int state)
 {
 	if (!hub->gpio_reset)
 		return;
 
+	i2c_lock_bus(hub->i2c->adapter, I2C_LOCK_SEGMENT);
+
 	gpiod_set_value_cansleep(hub->gpio_reset, state);
 
 	/* wait for hub recovery/stabilization */
@@ -234,6 +275,8 @@  static void usb251xb_reset(struct usb251xb *hub, int state)
 		usleep_range(500, 750);	/* >=500us at power on */
 	else
 		usleep_range(1, 10);	/* >=1us at power down */
+
+	i2c_unlock_bus(hub->i2c->adapter, I2C_LOCK_SEGMENT);
 }
 
 static int usb251xb_connect(struct usb251xb *hub)
@@ -621,6 +664,25 @@  static int usb251xb_probe(struct usb251xb *hub)
 		}
 	}
 
+	/*
+	 * usb251x SMBus-slave SCL lane is muxed with CFG_SEL0 pin. So if anyone
+	 * tries to work with the bus at the moment the hub reset is released,
+	 * it may cause an invalid config being latched by usb251x. Particularly
+	 * one of the config modes makes the hub loading a default registers
+	 * value without SMBus-slave interface activation. If the hub
+	 * accidentally gets this mode, this will cause the driver SMBus-
+	 * functions failure. Normally we could just lock the SMBus-segment the
+	 * hub i2c-interface resides for the device-specific reset timing. But
+	 * the GPIO controller, which is used to handle the hub reset, might be
+	 * placed at the same i2c-bus segment. In this case an error should be
+	 * returned since we can't safely use the GPIO controller to clear the
+	 * reset state (it may affect the hub configuration) and we can't lock
+	 * the i2c-bus segment (it will cause a deadlock).
+	 */
+	err = usb251x_check_gpio_chip(hub);
+	if (err)
+		return err;
+
 	err = usb251xb_connect(hub);
 	if (err) {
 		dev_err(dev, "Failed to connect hub (%d)\n", err);