Re: [PATCH 3/7] usb: typec: Separate the operations vector

[Heikki Krogerus <heikki.krogerus@linux.intel.com>]

On Tue, Oct 01, 2019 at 06:22:36AM -0700, Guenter Roeck wrote:
> On 10/1/19 2:48 AM, Heikki Krogerus wrote:
> > Introducing struct typec_operations which has the same
> > callbacks as struct typec_capability. The old callbacks are
> > kept for now, but after all users have been converted, they
> > will be removed.
> > 
> > Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
> > ---
> >   drivers/usb/typec/class.c | 90 +++++++++++++++++++++++++--------------
> >   include/linux/usb/typec.h | 19 +++++++++
> >   2 files changed, 76 insertions(+), 33 deletions(-)
> > 
> > diff --git a/drivers/usb/typec/class.c b/drivers/usb/typec/class.c
> > index 9fab0be8f08c..542be63795db 100644
> > --- a/drivers/usb/typec/class.c
> > +++ b/drivers/usb/typec/class.c
> > @@ -59,6 +59,7 @@ struct typec_port {
> >   	struct typec_mux		*mux;
> >   	const struct typec_capability	*cap;
> > +	const struct typec_operations	*ops;
> >   };
> >   #define to_typec_port(_dev_) container_of(_dev_, struct typec_port, dev)
> > @@ -961,11 +962,6 @@ preferred_role_store(struct device *dev, struct device_attribute *attr,
> >   		return -EOPNOTSUPP;
> >   	}
> > -	if (!port->cap->try_role) {
> > -		dev_dbg(dev, "Setting preferred role not supported\n");
> > -		return -EOPNOTSUPP;
> > -	}
> > -
> >   	role = sysfs_match_string(typec_roles, buf);
> >   	if (role < 0) {
> >   		if (sysfs_streq(buf, "none"))
> > @@ -974,9 +970,18 @@ preferred_role_store(struct device *dev, struct device_attribute *attr,
> >   			return -EINVAL;
> >   	}
> > -	ret = port->cap->try_role(port->cap, role);
> > -	if (ret)
> > -		return ret;
> > +	if (port->ops && port->ops->try_role) {
> > +		ret = port->ops->try_role(port, role);
> > +		if (ret)
> > +			return ret;
> > +	} else if (port->cap && port->cap->try_role) {
> > +		ret = port->cap->try_role(port->cap, role);
> > +		if (ret)
> > +			return ret;
> > +	} else {
> > +		dev_dbg(dev, "Setting preferred role not supported\n");
> > +		return -EOPNOTSUPP;
> > +	}
> 
> This is a semantic change: Support is now checked _after_ the string is evaluated.
> I understand the reason, but it should be noted in the patch description
> (not sure if it is worth it, though - it seems to me it makes the code more
> difficult to read).
> 
> >   	port->prefer_role = role;
> >   	return size;
> > @@ -1005,11 +1010,6 @@ static ssize_t data_role_store(struct device *dev,
> >   	struct typec_port *port = to_typec_port(dev);
> >   	int ret;
> > -	if (!port->cap->dr_set) {
> > -		dev_dbg(dev, "data role swapping not supported\n");
> > -		return -EOPNOTSUPP;
> > -	}
> > -
> >   	ret = sysfs_match_string(typec_data_roles, buf);
> >   	if (ret < 0)
> >   		return ret;
> > @@ -1020,9 +1020,19 @@ static ssize_t data_role_store(struct device *dev,
> >   		goto unlock_and_ret;
> >   	}
> > -	ret = port->cap->dr_set(port->cap, ret);
> > -	if (ret)
> > +	if (port->ops && port->ops->dr_set) {
> > +		ret = port->ops->dr_set(port, ret);
> > +		if (ret)
> > +			goto unlock_and_ret;
> > +	} else if (port->cap && port->cap->dr_set) {
> > +		ret = port->cap->dr_set(port->cap, ret);
> > +		if (ret)
> > +			goto unlock_and_ret;
> > +	} else {
> > +		dev_dbg(dev, "data role swapping not supported\n");
> > +		ret = -EOPNOTSUPP;
> >   		goto unlock_and_ret;
> 
> This really makes me wonder if the semantic change makes sense. Support
> is now evaluated _after_ the lock has been obtained. That seems like a
> waste.

OK, I'll re-think this.

> > +	}
> >   	ret = size;
> >   unlock_and_ret:
> > @@ -1055,11 +1065,6 @@ static ssize_t power_role_store(struct device *dev,
> >   		return -EOPNOTSUPP;
> >   	}
> > -	if (!port->cap->pr_set) {
> > -		dev_dbg(dev, "power role swapping not supported\n");
> > -		return -EOPNOTSUPP;
> > -	}
> > -
> >   	if (port->pwr_opmode != TYPEC_PWR_MODE_PD) {
> >   		dev_dbg(dev, "partner unable to swap power role\n");
> >   		return -EIO;
> > @@ -1077,11 +1082,21 @@ static ssize_t power_role_store(struct device *dev,
> >   		goto unlock_and_ret;
> >   	}
> > -	ret = port->cap->pr_set(port->cap, ret);
> > -	if (ret)
> > +	if (port->ops && port->ops->pr_set) {
> > +		ret = port->ops->pr_set(port, ret);
> > +		if (ret)
> > +			goto unlock_and_ret;
> > +	} else if (port->cap && port->cap->pr_set) {
> > +		ret = port->cap->pr_set(port->cap, ret);
> > +		if (ret)
> > +			goto unlock_and_ret;
> > +	} else {
> > +		dev_dbg(dev, "power role swapping not supported\n");
> > +		ret = -EOPNOTSUPP;
> >   		goto unlock_and_ret;
> > -
> > +	}
> >   	ret = size;
> > +
> >   unlock_and_ret:
> >   	mutex_unlock(&port->port_type_lock);
> >   	return ret;
> > @@ -1108,7 +1123,8 @@ port_type_store(struct device *dev, struct device_attribute *attr,
> >   	int ret;
> >   	enum typec_port_type type;
> > -	if (!port->cap->port_type_set || port->fixed_role != TYPEC_PORT_DRP) {
> > +	if ((!port->ops || !port->ops->port_type_set) ||
> > +	    !port->cap->port_type_set || port->fixed_role != TYPEC_PORT_DRP) {
> 
> The above now requires _all_ callbacks to exist, both ops and cap based ones.
> Is that on purpose ? Maybe this should be as follows ?
> 
> 	if (((!port->ops || !port->ops->port_type_set) &&
> 	     !port->cap->port_type_set) || port->fixed_role != TYPEC_PORT_DRP) {
> 
> or a bit better to read
> 	if (port->fixed_role != TYPEC_PORT_DRP ||
> 	    ((!port->ops || !port->ops->port_type_set) && !port->cap->port_type_set))

OK.


thanks,

-- 
heikki