From: Andrew Lunn <andrew@lunn.ch> Date: 2018-01-15 22:47:17
Further testing and core review found two sets of bugs.
Core review found a cut/paste error in the irq setup code.
A board which does not have an interrupt line from the switch to the
SoC, and experiancing an EPROBE_DEFER throw a splat when the ATU irq
was freed but never registered.
Andrew Lunn (2):
net: dsa: mv88e6xxx: Return error from irq_find_mapping()
net: dsa: mv88e6xxx: Free ATU/VTU irq only when there is chip irq
drivers/net/dsa/mv88e6xxx/chip.c | 6 ++++--
drivers/net/dsa/mv88e6xxx/global1_atu.c | 2 +-
drivers/net/dsa/mv88e6xxx/global1_vtu.c | 2 +-
3 files changed, 6 insertions(+), 4 deletions(-)
--
2.15.1
From: Andrew Lunn <andrew@lunn.ch> Date: 2018-01-15 22:46:47
We only register the ATU and VTU irq when we have a chip level IRQ.
In the error path, we should only attempt to remove the ATU and VTU
irq if we also have a chip level IRQ.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
---
drivers/net/dsa/mv88e6xxx/chip.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
From: Andrew Lunn <andrew@lunn.ch> Date: 2018-01-15 22:47:02
Fix a cut/paste error. When irq_find_mapping() returns an error for
the ATU or VTU interrupt, return that error, not the value of
chip->device_irq.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
---
drivers/net/dsa/mv88e6xxx/global1_atu.c | 2 +-
drivers/net/dsa/mv88e6xxx/global1_vtu.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
We only register the ATU and VTU irq when we have a chip level IRQ.
In the error path, we should only attempt to remove the ATU and VTU
irq if we also have a chip level IRQ.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
---
drivers/net/dsa/mv88e6xxx/chip.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
From: Andrew Lunn <andrew@lunn.ch> Date: 2018-01-15 23:02:26
On Mon, Jan 15, 2018 at 02:54:22PM -0800, Florian Fainelli wrote:
On 01/15/2018 02:45 PM, Andrew Lunn wrote:
quoted
We only register the ATU and VTU irq when we have a chip level IRQ.
In the error path, we should only attempt to remove the ATU and VTU
irq if we also have a chip level IRQ.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
---
drivers/net/dsa/mv88e6xxx/chip.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
@@ -3999,9 +3999,11 @@ static int mv88e6xxx_probe(struct mdio_device *mdiodev)out_mdio:mv88e6xxx_mdios_unregister(chip);out_g1_vtu_prob_irq:-mv88e6xxx_g1_vtu_prob_irq_free(chip);+if(chip->irq>0)+mv88e6xxx_g1_vtu_prob_irq_free(chip);
Why not move this check to mv88e6xxx_g1_vtu_prob_irq_free() and make it
a no-op if chip->irq <= 0?
Hi Florian
It keeps it symmetrical with the setup code:
if (chip->irq > 0) {
/* Has to be performed before the MDIO bus is created,
* because the PHYs will link there interrupts to these
* interrupt controllers
*/
mutex_lock(&chip->reg_lock);
err = mv88e6xxx_g1_irq_setup(chip);
mutex_unlock(&chip->reg_lock);
if (err)
goto out;
if (chip->info->g2_irqs > 0) {
err = mv88e6xxx_g2_irq_setup(chip);
if (err)
goto out_g1_irq;
}
err = mv88e6xxx_g1_atu_prob_irq_setup(chip);
if (err)
goto out_g2_irq;
err = mv88e6xxx_g1_vtu_prob_irq_setup(chip);
if (err)
goto out_g1_atu_prob_irq;
}
The general flow in probe() is to only call functions if they are
needed. So i would prefer to keep to that.
Andrew
From: David Miller <davem@davemloft.net> Date: 2018-01-16 20:18:35
From: Andrew Lunn <andrew@lunn.ch>
Date: Mon, 15 Jan 2018 23:45:56 +0100
Further testing and core review found two sets of bugs.
Core review found a cut/paste error in the irq setup code.
A board which does not have an interrupt line from the switch to the
SoC, and experiancing an EPROBE_DEFER throw a splat when the ATU irq
was freed but never registered.
From: David Miller <davem@davemloft.net> Date: 2018-01-16 20:23:07
From: David Miller <davem@davemloft.net>
Date: Tue, 16 Jan 2018 15:18:33 -0500 (EST)
From: Andrew Lunn <andrew@lunn.ch>
Date: Mon, 15 Jan 2018 23:45:56 +0100
quoted
Further testing and core review found two sets of bugs.
Core review found a cut/paste error in the irq setup code.
A board which does not have an interrupt line from the switch to the
SoC, and experiancing an EPROBE_DEFER throw a splat when the ATU irq
was freed but never registered.
Series applied, thanks Andrew.
Uhhh, this breaks the build. Reverting...
drivers/net/dsa/mv88e6xxx/global1_vtu.c: In function ‘mv88e6xxx_g1_vtu_prob_irq_setup’:
drivers/net/dsa/mv88e6xxx/global1_vtu.c:573:14: error: ‘struct mv88e6xxx_chip’ has no member named ‘chip’
return chip->chip->vtu_prob_irq;