usb: typec: tcpm/tcpci_maxim: better interrupt name

[PATCH] usb: typec: tcpm/tcpci_maxim: better interrupt name

Posted by André Draszik 10 months, 1 week ago

This changes the output of /proc/interrupts from the non-descriptive:
    1-0025
(or similar) to a more descriptive:
    1-0025-max33359

This makes it easier to find the device, in particular if there are
multiple i2c devices, as one doesn't have to remember (or lookup) where
it is located.

Signed-off-by: André Draszik <andre.draszik@linaro.org>
---
 drivers/usb/typec/tcpm/tcpci_maxim_core.c | 13 ++++++++++---
 1 file changed, 10 insertions(+), 3 deletions(-)

diff --git a/drivers/usb/typec/tcpm/tcpci_maxim_core.c b/drivers/usb/typec/tcpm/tcpci_maxim_core.c
index fd1b80593367..46fc626589db 100644
--- a/drivers/usb/typec/tcpm/tcpci_maxim_core.c
+++ b/drivers/usb/typec/tcpm/tcpci_maxim_core.c
@@ -420,12 +420,14 @@ static irqreturn_t max_tcpci_isr(int irq, void *dev_id)
 	return IRQ_WAKE_THREAD;
 }
 
-static int max_tcpci_init_alert(struct max_tcpci_chip *chip, struct i2c_client *client)
+static int max_tcpci_init_alert(struct max_tcpci_chip *chip,
+				struct i2c_client *client,
+				const char *name)
 {
 	int ret;
 
 	ret = devm_request_threaded_irq(chip->dev, client->irq, max_tcpci_isr, max_tcpci_irq,
-					(IRQF_TRIGGER_LOW | IRQF_ONESHOT), dev_name(chip->dev),
+					(IRQF_TRIGGER_LOW | IRQF_ONESHOT), name,
 					chip);
 
 	if (ret < 0)
@@ -485,6 +487,7 @@ static int max_tcpci_probe(struct i2c_client *client)
 	int ret;
 	struct max_tcpci_chip *chip;
 	u8 power_status;
+	const char *name;
 
 	chip = devm_kzalloc(&client->dev, sizeof(*chip), GFP_KERNEL);
 	if (!chip)
@@ -531,7 +534,11 @@ static int max_tcpci_probe(struct i2c_client *client)
 
 	chip->port = tcpci_get_tcpm_port(chip->tcpci);
 
-	ret = max_tcpci_init_alert(chip, client);
+	name = devm_kasprintf(&client->dev, GFP_KERNEL, "%s-%s",
+			      dev_name(&client->dev), client->name);
+	if (!name)
+		return -ENOMEM;
+	ret = max_tcpci_init_alert(chip, client, name);
 	if (ret < 0)
 		return dev_err_probe(&client->dev, ret,
 				     "IRQ initialization failed\n");

---
base-commit: 808eb958781e4ebb6e9c0962af2e856767e20f45
change-id: 20250213-max33359-irq-name-0665d6b69058

Best regards,
-- 
André Draszik <andre.draszik@linaro.org>

Re: [PATCH] usb: typec: tcpm/tcpci_maxim: better interrupt name

Posted by William McVicker 10 months ago

On 02/13/2025, André Draszik wrote:
> This changes the output of /proc/interrupts from the non-descriptive:
>     1-0025
> (or similar) to a more descriptive:
>     1-0025-max33359
> 
> This makes it easier to find the device, in particular if there are
> multiple i2c devices, as one doesn't have to remember (or lookup) where
> it is located.
> 
> Signed-off-by: André Draszik <andre.draszik@linaro.org>
> ---
>  drivers/usb/typec/tcpm/tcpci_maxim_core.c | 13 ++++++++++---
>  1 file changed, 10 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/usb/typec/tcpm/tcpci_maxim_core.c b/drivers/usb/typec/tcpm/tcpci_maxim_core.c
> index fd1b80593367..46fc626589db 100644
> --- a/drivers/usb/typec/tcpm/tcpci_maxim_core.c
> +++ b/drivers/usb/typec/tcpm/tcpci_maxim_core.c
> @@ -420,12 +420,14 @@ static irqreturn_t max_tcpci_isr(int irq, void *dev_id)
>  	return IRQ_WAKE_THREAD;
>  }
>  
> -static int max_tcpci_init_alert(struct max_tcpci_chip *chip, struct i2c_client *client)
> +static int max_tcpci_init_alert(struct max_tcpci_chip *chip,
> +				struct i2c_client *client,
> +				const char *name)
>  {
>  	int ret;
>  
>  	ret = devm_request_threaded_irq(chip->dev, client->irq, max_tcpci_isr, max_tcpci_irq,
> -					(IRQF_TRIGGER_LOW | IRQF_ONESHOT), dev_name(chip->dev),
> +					(IRQF_TRIGGER_LOW | IRQF_ONESHOT), name,
>  					chip);
>  
>  	if (ret < 0)
> @@ -485,6 +487,7 @@ static int max_tcpci_probe(struct i2c_client *client)
>  	int ret;
>  	struct max_tcpci_chip *chip;
>  	u8 power_status;
> +	const char *name;
>  
>  	chip = devm_kzalloc(&client->dev, sizeof(*chip), GFP_KERNEL);
>  	if (!chip)
> @@ -531,7 +534,11 @@ static int max_tcpci_probe(struct i2c_client *client)
>  
>  	chip->port = tcpci_get_tcpm_port(chip->tcpci);
>  
> -	ret = max_tcpci_init_alert(chip, client);
> +	name = devm_kasprintf(&client->dev, GFP_KERNEL, "%s-%s",
> +			      dev_name(&client->dev), client->name);
> +	if (!name)
> +		return -ENOMEM;
> +	ret = max_tcpci_init_alert(chip, client, name);
>  	if (ret < 0)
>  		return dev_err_probe(&client->dev, ret,
>  				     "IRQ initialization failed\n");

Can you just change the `init_name` instead like this?

  chip->client->dev.init_name = name;

That way calling `dev_name()` will get you this more informative name as well?


Thanks,
Will

Re: [PATCH] usb: typec: tcpm/tcpci_maxim: better interrupt name

Posted by André Draszik 9 months, 3 weeks ago

Hi Will,

On Fri, 2025-02-14 at 13:57 -0800, William McVicker wrote:
> On 02/13/2025, André Draszik wrote:
> > This changes the output of /proc/interrupts from the non-descriptive:
> >     1-0025
> > (or similar) to a more descriptive:
> >     1-0025-max33359
> > 
> > This makes it easier to find the device, in particular if there are
> > multiple i2c devices, as one doesn't have to remember (or lookup) where
> > it is located.
> > 
> > Signed-off-by: André Draszik <andre.draszik@linaro.org>
> > ---
> >  drivers/usb/typec/tcpm/tcpci_maxim_core.c | 13 ++++++++++---
> >  1 file changed, 10 insertions(+), 3 deletions(-)
> > 
> > diff --git a/drivers/usb/typec/tcpm/tcpci_maxim_core.c b/drivers/usb/typec/tcpm/tcpci_maxim_core.c
> > index fd1b80593367..46fc626589db 100644
> > --- a/drivers/usb/typec/tcpm/tcpci_maxim_core.c
> > +++ b/drivers/usb/typec/tcpm/tcpci_maxim_core.c
> > @@ -420,12 +420,14 @@ static irqreturn_t max_tcpci_isr(int irq, void *dev_id)
> >  	return IRQ_WAKE_THREAD;
> >  }
> >  
> > -static int max_tcpci_init_alert(struct max_tcpci_chip *chip, struct i2c_client *client)
> > +static int max_tcpci_init_alert(struct max_tcpci_chip *chip,
> > +				struct i2c_client *client,
> > +				const char *name)
> >  {
> >  	int ret;
> >  
> >  	ret = devm_request_threaded_irq(chip->dev, client->irq, max_tcpci_isr, max_tcpci_irq,
> > -					(IRQF_TRIGGER_LOW | IRQF_ONESHOT), dev_name(chip->dev),
> > +					(IRQF_TRIGGER_LOW | IRQF_ONESHOT), name,
> >  					chip);
> >  
> >  	if (ret < 0)
> > @@ -485,6 +487,7 @@ static int max_tcpci_probe(struct i2c_client *client)
> >  	int ret;
> >  	struct max_tcpci_chip *chip;
> >  	u8 power_status;
> > +	const char *name;
> >  
> >  	chip = devm_kzalloc(&client->dev, sizeof(*chip), GFP_KERNEL);
> >  	if (!chip)
> > @@ -531,7 +534,11 @@ static int max_tcpci_probe(struct i2c_client *client)
> >  
> >  	chip->port = tcpci_get_tcpm_port(chip->tcpci);
> >  
> > -	ret = max_tcpci_init_alert(chip, client);
> > +	name = devm_kasprintf(&client->dev, GFP_KERNEL, "%s-%s",
> > +			      dev_name(&client->dev), client->name);
> > +	if (!name)
> > +		return -ENOMEM;
> > +	ret = max_tcpci_init_alert(chip, client, name);
> >  	if (ret < 0)
> >  		return dev_err_probe(&client->dev, ret,
> >  				     "IRQ initialization failed\n");
> 
> Can you just change the `init_name` instead like this?
> 
>   chip->client->dev.init_name = name;
> 
> That way calling `dev_name()` will get you this more informative name as well?

I don't think that member is supposed to be set by anybody. Looking
at device_add(), this seems to only be meant to be statically set
by statically allocated devices and is cleared upon device_add().
Essentially no driver sets this in the kernel.

So it doesn't appear to be appropriate.

Cheers,
Andre'

Re: [PATCH] usb: typec: tcpm/tcpci_maxim: better interrupt name

Posted by Greg Kroah-Hartman 10 months, 1 week ago

On Thu, Feb 13, 2025 at 09:37:54AM +0000, André Draszik wrote:
> This changes the output of /proc/interrupts from the non-descriptive:
>     1-0025
> (or similar) to a more descriptive:
>     1-0025-max33359
> 
> This makes it easier to find the device, in particular if there are
> multiple i2c devices, as one doesn't have to remember (or lookup) where
> it is located.
> 
> Signed-off-by: André Draszik <andre.draszik@linaro.org>
> ---
>  drivers/usb/typec/tcpm/tcpci_maxim_core.c | 13 ++++++++++---
>  1 file changed, 10 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/usb/typec/tcpm/tcpci_maxim_core.c b/drivers/usb/typec/tcpm/tcpci_maxim_core.c
> index fd1b80593367..46fc626589db 100644
> --- a/drivers/usb/typec/tcpm/tcpci_maxim_core.c
> +++ b/drivers/usb/typec/tcpm/tcpci_maxim_core.c
> @@ -420,12 +420,14 @@ static irqreturn_t max_tcpci_isr(int irq, void *dev_id)
>  	return IRQ_WAKE_THREAD;
>  }
>  
> -static int max_tcpci_init_alert(struct max_tcpci_chip *chip, struct i2c_client *client)
> +static int max_tcpci_init_alert(struct max_tcpci_chip *chip,
> +				struct i2c_client *client,
> +				const char *name)
>  {
>  	int ret;
>  
>  	ret = devm_request_threaded_irq(chip->dev, client->irq, max_tcpci_isr, max_tcpci_irq,
> -					(IRQF_TRIGGER_LOW | IRQF_ONESHOT), dev_name(chip->dev),
> +					(IRQF_TRIGGER_LOW | IRQF_ONESHOT), name,
>  					chip);
>  
>  	if (ret < 0)
> @@ -485,6 +487,7 @@ static int max_tcpci_probe(struct i2c_client *client)
>  	int ret;
>  	struct max_tcpci_chip *chip;
>  	u8 power_status;
> +	const char *name;
>  
>  	chip = devm_kzalloc(&client->dev, sizeof(*chip), GFP_KERNEL);
>  	if (!chip)
> @@ -531,7 +534,11 @@ static int max_tcpci_probe(struct i2c_client *client)
>  
>  	chip->port = tcpci_get_tcpm_port(chip->tcpci);
>  
> -	ret = max_tcpci_init_alert(chip, client);
> +	name = devm_kasprintf(&client->dev, GFP_KERNEL, "%s-%s",
> +			      dev_name(&client->dev), client->name);
> +	if (!name)
> +		return -ENOMEM;

Please always test your code before sending it out.  You just leaked a
bunch of stuff here :(

thanks,

greg k-h

Re: [PATCH] usb: typec: tcpm/tcpci_maxim: better interrupt name

Posted by André Draszik 10 months, 1 week ago

Hi Greg,

On Thu, 2025-02-13 at 11:11 +0100, Greg Kroah-Hartman wrote:
> On Thu, Feb 13, 2025 at 09:37:54AM +0000, André Draszik wrote:
> > This changes the output of /proc/interrupts from the non-descriptive:
> >     1-0025
> > (or similar) to a more descriptive:
> >     1-0025-max33359
> > 
> > This makes it easier to find the device, in particular if there are
> > multiple i2c devices, as one doesn't have to remember (or lookup) where
> > it is located.
> > 
> > Signed-off-by: André Draszik <andre.draszik@linaro.org>
> > ---
> >  drivers/usb/typec/tcpm/tcpci_maxim_core.c | 13 ++++++++++---
> >  1 file changed, 10 insertions(+), 3 deletions(-)
> > 
> > diff --git a/drivers/usb/typec/tcpm/tcpci_maxim_core.c b/drivers/usb/typec/tcpm/tcpci_maxim_core.c
> > index fd1b80593367..46fc626589db 100644
> > --- a/drivers/usb/typec/tcpm/tcpci_maxim_core.c
> > +++ b/drivers/usb/typec/tcpm/tcpci_maxim_core.c
> > @@ -420,12 +420,14 @@ static irqreturn_t max_tcpci_isr(int irq, void *dev_id)
> >  	return IRQ_WAKE_THREAD;
> >  }
> >  
> > -static int max_tcpci_init_alert(struct max_tcpci_chip *chip, struct i2c_client *client)
> > +static int max_tcpci_init_alert(struct max_tcpci_chip *chip,
> > +				struct i2c_client *client,
> > +				const char *name)
> >  {
> >  	int ret;
> >  
> >  	ret = devm_request_threaded_irq(chip->dev, client->irq, max_tcpci_isr, max_tcpci_irq,
> > -					(IRQF_TRIGGER_LOW | IRQF_ONESHOT), dev_name(chip->dev),
> > +					(IRQF_TRIGGER_LOW | IRQF_ONESHOT), name,
> >  					chip);
> >  
> >  	if (ret < 0)
> > @@ -485,6 +487,7 @@ static int max_tcpci_probe(struct i2c_client *client)
> >  	int ret;
> >  	struct max_tcpci_chip *chip;
> >  	u8 power_status;
> > +	const char *name;
> >  
> >  	chip = devm_kzalloc(&client->dev, sizeof(*chip), GFP_KERNEL);
> >  	if (!chip)
> > @@ -531,7 +534,11 @@ static int max_tcpci_probe(struct i2c_client *client)
> >  
> >  	chip->port = tcpci_get_tcpm_port(chip->tcpci);
> >  
> > -	ret = max_tcpci_init_alert(chip, client);
> > +	name = devm_kasprintf(&client->dev, GFP_KERNEL, "%s-%s",
> > +			      dev_name(&client->dev), client->name);
> > +	if (!name)
> > +		return -ENOMEM;
> 
> Please always test your code before sending it out.  You just leaked a
> bunch of stuff here :(

Thanks for the review!

I can not see what leak you're referring to. Could you please clarify?

Thanks,
Andre'

Re: [PATCH] usb: typec: tcpm/tcpci_maxim: better interrupt name

Posted by Greg Kroah-Hartman 10 months, 1 week ago

On Thu, Feb 13, 2025 at 11:44:59AM +0000, André Draszik wrote:
> Hi Greg,
> 
> On Thu, 2025-02-13 at 11:11 +0100, Greg Kroah-Hartman wrote:
> > On Thu, Feb 13, 2025 at 09:37:54AM +0000, André Draszik wrote:
> > > This changes the output of /proc/interrupts from the non-descriptive:
> > >     1-0025
> > > (or similar) to a more descriptive:
> > >     1-0025-max33359
> > > 
> > > This makes it easier to find the device, in particular if there are
> > > multiple i2c devices, as one doesn't have to remember (or lookup) where
> > > it is located.
> > > 
> > > Signed-off-by: André Draszik <andre.draszik@linaro.org>
> > > ---
> > >  drivers/usb/typec/tcpm/tcpci_maxim_core.c | 13 ++++++++++---
> > >  1 file changed, 10 insertions(+), 3 deletions(-)
> > > 
> > > diff --git a/drivers/usb/typec/tcpm/tcpci_maxim_core.c b/drivers/usb/typec/tcpm/tcpci_maxim_core.c
> > > index fd1b80593367..46fc626589db 100644
> > > --- a/drivers/usb/typec/tcpm/tcpci_maxim_core.c
> > > +++ b/drivers/usb/typec/tcpm/tcpci_maxim_core.c
> > > @@ -420,12 +420,14 @@ static irqreturn_t max_tcpci_isr(int irq, void *dev_id)
> > >  	return IRQ_WAKE_THREAD;
> > >  }
> > >  
> > > -static int max_tcpci_init_alert(struct max_tcpci_chip *chip, struct i2c_client *client)
> > > +static int max_tcpci_init_alert(struct max_tcpci_chip *chip,
> > > +				struct i2c_client *client,
> > > +				const char *name)
> > >  {
> > >  	int ret;
> > >  
> > >  	ret = devm_request_threaded_irq(chip->dev, client->irq, max_tcpci_isr, max_tcpci_irq,
> > > -					(IRQF_TRIGGER_LOW | IRQF_ONESHOT), dev_name(chip->dev),
> > > +					(IRQF_TRIGGER_LOW | IRQF_ONESHOT), name,
> > >  					chip);
> > >  
> > >  	if (ret < 0)
> > > @@ -485,6 +487,7 @@ static int max_tcpci_probe(struct i2c_client *client)
> > >  	int ret;
> > >  	struct max_tcpci_chip *chip;
> > >  	u8 power_status;
> > > +	const char *name;
> > >  
> > >  	chip = devm_kzalloc(&client->dev, sizeof(*chip), GFP_KERNEL);
> > >  	if (!chip)
> > > @@ -531,7 +534,11 @@ static int max_tcpci_probe(struct i2c_client *client)
> > >  
> > >  	chip->port = tcpci_get_tcpm_port(chip->tcpci);
> > >  
> > > -	ret = max_tcpci_init_alert(chip, client);
> > > +	name = devm_kasprintf(&client->dev, GFP_KERNEL, "%s-%s",
> > > +			      dev_name(&client->dev), client->name);
> > > +	if (!name)
> > > +		return -ENOMEM;
> > 
> > Please always test your code before sending it out.  You just leaked a
> > bunch of stuff here :(
> 
> Thanks for the review!
> 
> I can not see what leak you're referring to. Could you please clarify?

At a quick glance, tcpci_register_port() is called earlier in the
function, but when you error out here you did not call
tcpci_unregister_port().  What else needs to also be unwound?

thanks,

greg k-h

Re: [PATCH] usb: typec: tcpm/tcpci_maxim: better interrupt name

Posted by André Draszik 10 months, 1 week ago

Thanks Greg,

On Thu, 2025-02-13 at 12:56 +0100, Greg Kroah-Hartman wrote:
> At a quick glance, tcpci_register_port() is called earlier in the
> function, but when you error out here you did not call
> tcpci_unregister_port().  What else needs to also be unwound?

This driver manages everything using devres, including calling of
tcpci_unregister_port() via devres:

        ret = devm_add_action_or_reset(&client->dev,
				       max_tcpci_unregister_tcpci_port,
				       chip->tcpci);

is done just after tcpci_register_port(). As far as I can see nothing
needs to be unwound explicitly.

Cheers,
Andre'

Re: [PATCH] usb: typec: tcpm/tcpci_maxim: better interrupt name

Posted by Greg Kroah-Hartman 10 months, 1 week ago

On Thu, Feb 13, 2025 at 01:10:49PM +0000, André Draszik wrote:
> Thanks Greg,
> 
> On Thu, 2025-02-13 at 12:56 +0100, Greg Kroah-Hartman wrote:
> > At a quick glance, tcpci_register_port() is called earlier in the
> > function, but when you error out here you did not call
> > tcpci_unregister_port().  What else needs to also be unwound?
> 
> This driver manages everything using devres, including calling of
> tcpci_unregister_port() via devres:
> 
>         ret = devm_add_action_or_reset(&client->dev,
> 				       max_tcpci_unregister_tcpci_port,
> 				       chip->tcpci);
> 
> is done just after tcpci_register_port(). As far as I can see nothing
> needs to be unwound explicitly.

Ugh, that wasn't obvious at all, sorry about that.  Yet another reason
to hate devm apis :)

thanks,

greg k-h