linux-kernel - Re: [PATCH v5 2/2] usb: typec: tcpm: Support multiple capabilities

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <ZXnAiRa9bBacASQ5@kuha.fi.intel.com>
Date:   Wed, 13 Dec 2023 16:32:41 +0200
From:   Heikki Krogerus <heikki.krogerus@...ux.intel.com>
To:     Kyle Tso <kyletso@...gle.com>
Cc:     robh+dt@...nel.org, krzysztof.kozlowski+dt@...aro.org,
        conor+dt@...nel.org, linux@...ck-us.net,
        gregkh@...uxfoundation.org, badhri@...gle.com,
        devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-usb@...r.kernel.org
Subject: Re: [PATCH v5 2/2] usb: typec: tcpm: Support multiple capabilities

Hi Kyle,

On Tue, Dec 05, 2023 at 11:01:14AM +0800, Kyle Tso wrote:
> Refactor tcpm_fw_get_caps to support the multiple pd capabilities got
> from fwnode. For backward compatibility, the original single capability
> is still applicable. The fetched data is stored in the newly defined
> structure "pd_data" and there is an array "pd_list" to store the
> pointers to them. A dedicated array "pds" is used to store the handles
> of the registered usb_power_delivery instances.
> 
> Also implement the .pd_get and .pd_set ops which are introduced in
> commit a7cff92f0635 ("usb: typec: USB Power Delivery helpers for ports
> and partners"). Once the .pd_set is called, the current capability will
> be updated and state machine will re-negotiate the power contract if
> possible.
> 
> Signed-off-by: Kyle Tso <kyletso@...gle.com>
> ---
> v4 -> v5:
> - no change
> 
>  drivers/usb/typec/tcpm/tcpm.c | 389 ++++++++++++++++++++++++++--------
>  1 file changed, 303 insertions(+), 86 deletions(-)

<snip>

> @@ -6124,12 +6243,11 @@ static int tcpm_port_register_pd(struct tcpm_port *port)
>  	return ret;
>  }
>  
> -static int tcpm_fw_get_caps(struct tcpm_port *port,
> -			    struct fwnode_handle *fwnode)
> +static int tcpm_fw_get_properties(struct tcpm_port *port, struct fwnode_handle *fwnode)

The function names got me confused first :). Even though you now call
this "properties", I think the term "capablities" can in practice
still mean two things in this driver (you are for example still
calling typec_get_fw_cap() from this function).

So how about you don't change the name of this function, but
instead..

>  {
>  	const char *opmode_str;
> +	u32 frs_current;
>  	int ret;
> -	u32 mw, frs_current;
>  
>  	if (!fwnode)
>  		return -EINVAL;
> @@ -6149,28 +6267,10 @@ static int tcpm_fw_get_caps(struct tcpm_port *port,
>  
>  	port->port_type = port->typec_caps.type;
>  	port->pd_supported = !fwnode_property_read_bool(fwnode, "pd-disable");
> -
>  	port->slow_charger_loop = fwnode_property_read_bool(fwnode, "slow-charger-loop");
> -	if (port->port_type == TYPEC_PORT_SNK)
> -		goto sink;
> -
> -	/* Get Source PDOs for the PD port or Source Rp value for the non-PD port */
> -	if (port->pd_supported) {
> -		ret = fwnode_property_count_u32(fwnode, "source-pdos");
> -		if (ret == 0)
> -			return -EINVAL;
> -		else if (ret < 0)
> -			return ret;
> +	port->self_powered = fwnode_property_read_bool(fwnode, "self-powered");
>  
> -		port->nr_src_pdo = min(ret, PDO_MAX_OBJECTS);
> -		ret = fwnode_property_read_u32_array(fwnode, "source-pdos",
> -						     port->src_pdo, port->nr_src_pdo);
> -		if (ret)
> -			return ret;
> -		ret = tcpm_validate_caps(port, port->src_pdo, port->nr_src_pdo);
> -		if (ret)
> -			return ret;
> -	} else {
> +	if (!port->pd_supported) {
>  		ret = fwnode_property_read_string(fwnode, "typec-power-opmode", &opmode_str);
>  		if (ret)
>  			return ret;
> @@ -6180,43 +6280,156 @@ static int tcpm_fw_get_caps(struct tcpm_port *port,
>  		port->src_rp = tcpm_pwr_opmode_to_rp(ret);
>  	}
>  
> -	if (port->port_type == TYPEC_PORT_SRC)
> -		return 0;
> +	/* FRS can only be supported by DRP ports */
> +	if (port->port_type == TYPEC_PORT_DRP) {
> +		ret = fwnode_property_read_u32(fwnode, "new-source-frs-typec-current",
> +					       &frs_current);
> +		if (!ret && frs_current <= FRS_5V_3A)
> +			port->new_source_frs_current = frs_current;
> +	}
>  
> -sink:
> -	port->self_powered = fwnode_property_read_bool(fwnode, "self-powered");
> +	return 0;
> +}
> +
> +static int tcpm_fw_get_caps(struct tcpm_port *port, struct fwnode_handle *fwnode)

..make that tcpm_fw_get_pd_capabilities() (or something like that) to
be more clear?

Also, since you are only calling this ones, why not just call this
directly from tcpm_fw_get_cap() instead of right after it?

thanks,

-- 
heikki