linux-kernel - Re: [PATCH 6/6] usb: typec: ucsi: pr_swap should check connector

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: <aOZiuxDfwYql6ZUu@kuha.fi.intel.com>
Date: Wed, 8 Oct 2025 16:10:19 +0300
From: Heikki Krogerus <heikki.krogerus@...ux.intel.com>
To: Jameson Thies <jthies@...gle.com>
Cc: linux-usb@...r.kernel.org, linux-kernel@...r.kernel.org,
	dmitry.baryshkov@....qualcomm.com, bleung@...omium.org,
	gregkh@...uxfoundation.org, akuchynski@...omium.org,
	abhishekpandit@...omium.org, sebastian.reichel@...labora.com,
	linux-pm@...r.kernel.org
Subject: Re: [PATCH 6/6] usb: typec: ucsi: pr_swap should check
 connector_status

Hi,

On Tue, Oct 07, 2025 at 12:00:07AM +0000, Jameson Thies wrote:
> From: Abhishek Pandit-Subedi <abhishekpandit@...omium.org>
> 
> Power role swaps initiated by the host system doesn't generate
> connection status change notifications.
> 
> >From UCSIv3.0 spec, section 6.5.10 Set Power Direction Role:
> 
> The execution of this command might require PPM to initiate a power
> role swap. If the power role swap fails for any reason, the command
> returns, and error and the power direction should remain unchanged.
> Note that if the execution of the command resulted in a successful
> power role swap, it should not result in a connector status change
> notification.
> 
> Signed-off-by: Abhishek Pandit-Subedi <abhishekpandit@...omium.org>
> Signed-off-by: Jameson Thies <jthies@...gle.com>
> ---
>  drivers/usb/typec/ucsi/ucsi.c | 30 +++++++++++++++++++++++++-----
>  1 file changed, 25 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/usb/typec/ucsi/ucsi.c b/drivers/usb/typec/ucsi/ucsi.c
> index 1a7d850b11ea..6e3797d7a144 100644
> --- a/drivers/usb/typec/ucsi/ucsi.c
> +++ b/drivers/usb/typec/ucsi/ucsi.c
> @@ -1526,20 +1526,40 @@ static int ucsi_pr_swap(struct typec_port *port, enum typec_role role)
>  	if (ret < 0)
>  		goto out_unlock;
>  
> -	mutex_unlock(&con->lock);
> +	command = UCSI_GET_CONNECTOR_STATUS | UCSI_CONNECTOR_NUMBER(con->num);
> +	ret = ucsi_send_command(con->ucsi, command, &con->status, sizeof(con->status));
> +	if (ret < 0)
> +		goto out_unlock;

Couldn't you use the helper ucsi_get_connector_status() ?

> -	if (!wait_for_completion_timeout(&con->complete,
> -					 msecs_to_jiffies(UCSI_SWAP_TIMEOUT_MS)))
> -		return -ETIMEDOUT;
> +	cur_role = !!UCSI_CONSTAT(con, PWR_DIR);
>  
> -	mutex_lock(&con->lock);
> +	/* Execution of SET_PDR should not result in connector status
> +	 * notifications. However, some legacy implementations may still defer
> +	 * the actual role swap and return immediately. Thus, check the
> +	 * connector status in case it immediately succeeded or wait for a later
> +	 * connector status change.
> +	 */
> +	if (cur_role != role) {
> +		mutex_unlock(&con->lock);
> +
> +		if (!wait_for_completion_timeout(
> +			    &con->complete,
> +			    msecs_to_jiffies(UCSI_SWAP_TIMEOUT_MS)))

Please align those properly.

> +			return -ETIMEDOUT;
> +
> +		mutex_lock(&con->lock);
> +	}
>  
>  	/* Something has gone wrong while swapping the role */
>  	if (UCSI_CONSTAT(con, PWR_OPMODE) != UCSI_CONSTAT_PWR_OPMODE_PD) {
>  		ucsi_reset_connector(con, true);
>  		ret = -EPROTO;
> +		goto out_unlock;
>  	}
>  
> +	/* Indicate successful power role swap */
> +	typec_set_pwr_role(con->port, role);
> +
>  out_unlock:
>  	mutex_unlock(&con->lock);
>  

Maybe this could be send separately? It does not seem to be directly
ucsi psy related.

thanks,

-- 
heikki