[PATCH 1/3] ps3: vuart: fix error path locking

This stray down would cause a permanent sleep which doesn't seem correct. 
The other uses of this semaphore appear fairly mutex like it's even initialized
with init_MUTEX() ..  So here a patch for removing this one down().

Signed-off-by: Daniel Walker <redacted>

---
 drivers/ps3/ps3-vuart.c |    1 -
 1 file changed, 1 deletion(-)

Index: linux-2.6.23/drivers/ps3/ps3-vuart.c
===================================================================

--- linux-2.6.23.orig/drivers/ps3/ps3-vuart.c
+++ linux-2.6.23/drivers/ps3/ps3-vuart.c

@@ -1072,7 +1072,6 @@ static int ps3_vuart_probe(struct ps3_sy
 	if (result) {
 		dev_dbg(&dev->core, "%s:%d: drv->probe failed\n",
 			__func__, __LINE__);
-		down(&vuart_bus_priv.probe_mutex);
 		goto fail_probe;
 	}
 

-- 

This stray down would cause a permanent sleep which doesn't seem correct. 
The other uses of this semaphore appear fairly mutex like it's even initialized
with init_MUTEX() ..  So here a patch for removing this one down().

Signed-off-by: Daniel Walker <redacted>

---
 drivers/ps3/ps3-vuart.c |    1 -
 1 file changed, 1 deletion(-)

Signed-off-by: Geoff Levand <redacted>


Looks, good. 
Andrew, Please apply.

-Geoff 

On Wed, 12 Dec 2007 18:00:12 -0800
Geoff Levand [off-list ref] wrote:

quoted

This stray down would cause a permanent sleep which doesn't seem correct. 
The other uses of this semaphore appear fairly mutex like it's even initialized
with init_MUTEX() ..  So here a patch for removing this one down().

Signed-off-by: Daniel Walker <redacted>

---
 drivers/ps3/ps3-vuart.c |    1 -
 1 file changed, 1 deletion(-)

Signed-off-by: Geoff Levand <redacted>


Looks, good. 

Looks bad to me.

Andrew, Please apply.

The current code:

	if (result) {
		dev_dbg(&dev->core, "%s:%d: drv->probe failed\n",
			__func__, __LINE__);
		down(&vuart_bus_priv.probe_mutex);
		goto fail_probe;
	}

	up(&vuart_bus_priv.probe_mutex);

	return result;

fail_probe:
	ps3_vuart_set_interrupt_mask(dev, 0);
	kfree(dev->driver_priv);
	dev->driver_priv = NULL;
fail_dev_malloc:
	vuart_bus_priv.devices[dev->port_number] = NULL;
fail_busy:
	ps3_vuart_bus_interrupt_put();
fail_setup_interrupt:
	up(&vuart_bus_priv.probe_mutex);
	dev_dbg(&dev->core, "%s:%d: failed\n", __func__, __LINE__);
	return result;
}

is correct.  Although not exactly a thing of beauty.

On Tue, 2007-12-18 at 17:10 -0800, Andrew Morton wrote:

is correct.  Although not exactly a thing of beauty.

This isn't the worst I've seen ;( .. Do you think the ending should fall
through instead of having two returns?

Daniel

On 12/18/2007 05:10 PM, Andrew Morton wrote:

On Wed, 12 Dec 2007 18:00:12 -0800
Geoff Levand [off-list ref] wrote:

quoted

quoted

This stray down would cause a permanent sleep which doesn't seem correct. 
The other uses of this semaphore appear fairly mutex like it's even initialized
with init_MUTEX() ..  So here a patch for removing this one down().

Signed-off-by: Daniel Walker <redacted>

---
 drivers/ps3/ps3-vuart.c |    1 -
 1 file changed, 1 deletion(-)

Signed-off-by: Geoff Levand <redacted>


Looks, good. 

Looks bad to me.

Hi Andrew,

Unfortunately there wasn't enough context in the patch to see
that there is a down() earlier in the routine, and that the patch
does indeed remove an incorrectly placed down().  Here is the
entire routine, marked with what the patch removes.

static int ps3_vuart_probe(struct ps3_system_bus_device *dev)
{
	int result;
	struct ps3_vuart_port_driver *drv;
	struct ps3_vuart_port_priv *priv = NULL;

	dev_dbg(&dev->core, "%s:%d\n", __func__, __LINE__);

	drv = ps3_system_bus_dev_to_vuart_drv(dev);

	dev_dbg(&dev->core, "%s:%d: (%s)\n", __func__, __LINE__,
		drv->core.core.name);

	BUG_ON(!drv);

	if (dev->port_number >= PORT_COUNT) {
		BUG();
		return -EINVAL;
	}

	down(&vuart_bus_priv.probe_mutex);

	result = ps3_vuart_bus_interrupt_get();

	if (result)
		goto fail_setup_interrupt;

	if (vuart_bus_priv.devices[dev->port_number]) {
		dev_dbg(&dev->core, "%s:%d: port busy (%d)\n", __func__,
			__LINE__, dev->port_number);
		result = -EBUSY;
		goto fail_busy;
	}

	vuart_bus_priv.devices[dev->port_number] = dev;

	/* Setup dev->driver_priv. */

	dev->driver_priv = kzalloc(sizeof(struct ps3_vuart_port_priv),
		GFP_KERNEL);

	if (!dev->driver_priv) {
		result = -ENOMEM;
		goto fail_dev_malloc;
	}

	priv = to_port_priv(dev);

	INIT_LIST_HEAD(&priv->tx_list.head);
	spin_lock_init(&priv->tx_list.lock);

	INIT_LIST_HEAD(&priv->rx_list.head);
	spin_lock_init(&priv->rx_list.lock);

	INIT_WORK(&priv->rx_list.work.work, NULL);
	priv->rx_list.work.trigger = 0;
	priv->rx_list.work.dev = dev;

	/* clear stale pending interrupts */

	ps3_vuart_clear_rx_bytes(dev, 0);

	ps3_vuart_set_interrupt_mask(dev, INTERRUPT_MASK_RX);

	ps3_vuart_set_triggers(dev, 1, 1);

	if (drv->probe)
		result = drv->probe(dev);
	else {
		result = 0;
		dev_info(&dev->core, "%s:%d: no probe method\n", __func__,
			__LINE__);
	}

	if (result) {
		dev_dbg(&dev->core, "%s:%d: drv->probe failed\n",
			__func__, __LINE__);
removed >>>>>>	down(&vuart_bus_priv.probe_mutex); <<<<<<<<<<<
		goto fail_probe;
	}

	up(&vuart_bus_priv.probe_mutex);

	return result;

fail_probe:
	ps3_vuart_set_interrupt_mask(dev, 0);
	kfree(dev->driver_priv);
	dev->driver_priv = NULL;
fail_dev_malloc:
	vuart_bus_priv.devices[dev->port_number] = NULL;
fail_busy:
	ps3_vuart_bus_interrupt_put();
fail_setup_interrupt:
	up(&vuart_bus_priv.probe_mutex);
	dev_dbg(&dev->core, "%s:%d: failed\n", __func__, __LINE__);
	return result;
}

Thanks for taking the time to scrutinize.

-Geoff

On Tue, 2007-12-18 at 19:04 -0800, Geoff Levand wrote:

Unfortunately there wasn't enough context in the patch to see
that there is a down() earlier in the routine, and that the patch
does indeed remove an incorrectly placed down().  Here is the
entire routine, marked with what the patch removes.

Andrew have you had a chance to review this?

Daniel

On Thu, 20 Dec 2007 11:32:25 -0800 Daniel Walker [off-list ref] wrote:

On Tue, 2007-12-18 at 19:04 -0800, Geoff Levand wrote:

quoted

Unfortunately there wasn't enough context in the patch to see
that there is a down() earlier in the routine, and that the patch
does indeed remove an incorrectly placed down().  Here is the
entire routine, marked with what the patch removes.

Andrew have you had a chance to review this?

Confused.  I did review it: http://lkml.org/lkml/2007/12/18/384

On Thu, 2007-12-20 at 12:06 -0800, Andrew Morton wrote:

On Thu, 20 Dec 2007 11:32:25 -0800 Daniel Walker [off-list ref] wrote:

quoted

On Tue, 2007-12-18 at 19:04 -0800, Geoff Levand wrote:

quoted

Unfortunately there wasn't enough context in the patch to see
that there is a down() earlier in the routine, and that the patch
does indeed remove an incorrectly placed down().  Here is the
entire routine, marked with what the patch removes.

Andrew have you had a chance to review this?

Confused.  I did review it: http://lkml.org/lkml/2007/12/18/384

Yeah, but Geoff countered http://lkml.org/lkml/2007/12/18/409

Do you still think the patch is wrong, given the whole function?

Daniel