[PATCH net-next] net: phy: improve stopping PHY

[PATCH net-next] net: phy: improve stopping PHY

[Heiner Kallweit]

phy_stop_interrupts() is called from phy_disconnect() only. Most of
what it does has been done by phy_stop() already which should have
been called before phy_disconnect(). Based on that we can do some
improvements:
- remove phy_stop_interrupts() and free interrupt in
  phy_disconnect() directly
- replace condition "phydev->irq > 0" with the appropriate helper
- make sure phy state machine is stopped after calling phy_stop()
- check in phy_disconnect() that PHY is in a stopped state. Else
  warn to detect misbehaving drivers and call phy_stop().

Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
---
 drivers/net/phy/phy.c        | 18 +-----------------
 drivers/net/phy/phy_device.c |  9 ++++++---
 include/linux/phy.h          |  1 -
 3 files changed, 7 insertions(+), 21 deletions(-)

diff --git a/drivers/net/phy/phy.c b/drivers/net/phy/phy.c
index 2ffe08537..745a705a5 100644
--- a/drivers/net/phy/phy.c
+++ b/drivers/net/phy/phy.c
@@ -813,23 +813,6 @@ int phy_start_interrupts(struct phy_device *phydev)
 }
 EXPORT_SYMBOL(phy_start_interrupts);
 
-/**
- * phy_stop_interrupts - disable interrupts from a PHY device
- * @phydev: target phy_device struct
- */
-int phy_stop_interrupts(struct phy_device *phydev)
-{
-	int err = phy_disable_interrupts(phydev);
-
-	if (err)
-		phy_error(phydev);
-
-	free_irq(phydev->irq, phydev);
-
-	return err;
-}
-EXPORT_SYMBOL(phy_stop_interrupts);
-
 /**
  * phy_stop - Bring down the PHY link, and stop checking the status
  * @phydev: target phy_device struct
@@ -853,6 +836,7 @@ void phy_stop(struct phy_device *phydev)
 	mutex_unlock(&phydev->lock);
 
 	phy_state_machine(&phydev->state_queue.work);
+	phy_stop_machine(phydev);
 
 	/* Cannot call flush_scheduled_work() here as desired because
 	 * of rtnl_lock(), but PHY_HALTED shall guarantee irq handler
diff --git a/drivers/net/phy/phy_device.c b/drivers/net/phy/phy_device.c
index b94095d74..4bd6c9787 100644
--- a/drivers/net/phy/phy_device.c
+++ b/drivers/net/phy/phy_device.c
@@ -981,10 +981,13 @@ EXPORT_SYMBOL(phy_connect);
  */
 void phy_disconnect(struct phy_device *phydev)
 {
-	if (phydev->irq > 0)
-		phy_stop_interrupts(phydev);
+	if (phy_is_started(phydev)) {
+		phydev_warn(phydev, "phy_stop should have been called before phy_disconnect!\n");
+		phy_stop(phydev);
+	}
 
-	phy_stop_machine(phydev);
+	if (phy_interrupt_is_valid(phydev))
+		free_irq(phydev->irq, phydev);
 
 	phydev->adjust_link = NULL;
 
diff --git a/include/linux/phy.h b/include/linux/phy.h
index 00b0533de..1fa7c367b 100644
--- a/include/linux/phy.h
+++ b/include/linux/phy.h
@@ -951,7 +951,6 @@ int phy_aneg_done(struct phy_device *phydev);
 int phy_speed_down(struct phy_device *phydev, bool sync);
 int phy_speed_up(struct phy_device *phydev);
 
-int phy_stop_interrupts(struct phy_device *phydev);
 int phy_restart_aneg(struct phy_device *phydev);
 int phy_reset_after_clk_enable(struct phy_device *phydev);
 
-- 
2.20.1

Re: [PATCH net-next] net: phy: improve stopping PHY

[Andrew Lunn]

On Wed, Jan 16, 2019 at 08:20:43PM +0100, Heiner Kallweit wrote:
> phy_stop_interrupts() is called from phy_disconnect() only. Most of
> what it does has been done by phy_stop() already which should have
> been called before phy_disconnect(). Based on that we can do some
> improvements:
> - remove phy_stop_interrupts() and free interrupt in
>   phy_disconnect() directly
> - replace condition "phydev->irq > 0" with the appropriate helper
> - make sure phy state machine is stopped after calling phy_stop()
> - check in phy_disconnect() that PHY is in a stopped state. Else
>   warn to detect misbehaving drivers and call phy_stop().

Hi Heiner

When i see a list like this, it makes me think there should be a
patchset, not a single patch. If something does break, we can bisect
it to just one change.

Please could you try to break this up into a few patches.

Thanks
	Andrew

Re: [PATCH net-next] net: phy: improve stopping PHY

[Heiner Kallweit]

On 16.01.2019 20:40, Andrew Lunn wrote:
> On Wed, Jan 16, 2019 at 08:20:43PM +0100, Heiner Kallweit wrote:
>> phy_stop_interrupts() is called from phy_disconnect() only. Most of
>> what it does has been done by phy_stop() already which should have
>> been called before phy_disconnect(). Based on that we can do some
>> improvements:
>> - remove phy_stop_interrupts() and free interrupt in
>>   phy_disconnect() directly
>> - replace condition "phydev->irq > 0" with the appropriate helper
>> - make sure phy state machine is stopped after calling phy_stop()
>> - check in phy_disconnect() that PHY is in a stopped state. Else
>>   warn to detect misbehaving drivers and call phy_stop().
> 
> Hi Heiner
> 
> When i see a list like this, it makes me think there should be a
> patchset, not a single patch. If something does break, we can bisect
> it to just one change.
> 
> Please could you try to break this up into a few patches.
> 
I thought the changes are simple enough to combine them and avoid the
overhead of a patchset. But I'll see to break this up, considering
that not all changes are fully independent.

> Thanks
> 	Andrew
> 
Heiner