RE: [PATCH v3 2/4] usb: typec: ucsi: add a common function ucsi_unregister_connectors()

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



> From: Heikki Krogerus <heikki.krogerus@xxxxxxxxxxxxxxx>
> Sent: Wednesday, April 13, 2022 8:09 PM
> To: Linyu Yuan (QUIC) <quic_linyyuan@xxxxxxxxxxx>
> Cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>; linux-
> usb@xxxxxxxxxxxxxxx; Jack Pham (QUIC) <quic_jackp@xxxxxxxxxxx>
> Subject: Re: [PATCH v3 2/4] usb: typec: ucsi: add a common function
> ucsi_unregister_connectors()
> 
> On Wed, Apr 13, 2022 at 05:58:09PM +0800, Linyu Yuan wrote:
> > In error path of ucsi_init(), it will unregister all valid ucsi connector,
> > and samiliar operation also happen in ucsi_unregister(),
> 
> Sorry but I have to confirm this: with "samiliar" you mean "the same",
> right?

Only one small difference for original code which is no cancel_work_sync() of each connector in ucsi _init(),
But in ucsi_register_port(), we get role switch after connector work initialized,
So I think it is safe to call cancel_work_sync() to connector work if role switch return -EPROBE_DEFER.

> 
> > add a common function for two places.
> >
> > Signed-off-by: Linyu Yuan <quic_linyyuan@xxxxxxxxxxx>
> > ---
> > v2: improve ucsi_connector_clean(), check total number of connector.
> > v3: rename to ucsi_unregister_connectors(), suggest by maintainer
> >
> >  drivers/usb/typec/ucsi/ucsi.c | 51 ++++++++++++++++++++++++-----------
> --------
> >  1 file changed, 28 insertions(+), 23 deletions(-)
> >
> > diff --git a/drivers/usb/typec/ucsi/ucsi.c b/drivers/usb/typec/ucsi/ucsi.c
> > index 77ac0b7..af9a2a1 100644
> > --- a/drivers/usb/typec/ucsi/ucsi.c
> > +++ b/drivers/usb/typec/ucsi/ucsi.c
> > @@ -1187,6 +1187,32 @@ static int ucsi_register_port(struct ucsi *ucsi, int
> index)
> >  	return ret;
> >  }
> >
> > +static void ucsi_unregister_connectors(struct ucsi *ucsi)
> > +{
> > +	struct ucsi_connector *con;
> > +	int i;
> > +
> > +	if (!ucsi->connector)
> > +		return;
> 
> Can that actually ever happen?

Consider a case, ucsi_init() failed, we will call ucsi_unregister_connectors() to free all connectors,
After that the UCSI implementation like ucsi_acpi call ucsi_unregister() again,
It should not unregister connectors again.

> 
> > +	for (i = 0; i < ucsi->cap.num_connectors; i++) {
> > +		con = &ucsi->connector[i];
> > +		if (!con->port)
> > +			break;
> > +
> > +		cancel_work_sync(&con->work);
> > +		ucsi_unregister_partner(con);
> > +		ucsi_unregister_altmodes(con, UCSI_RECIPIENT_CON);
> > +		ucsi_unregister_port_psy(con);
> > +		if (con->wq)
> > +			destroy_workqueue(con->wq);
> > +		typec_unregister_port(con->port);
> > +	}
> > +
> > +	kfree(ucsi->connector);
> > +	ucsi->connector = NULL;
> > +}
> 
> Another way of doing this would be to just remove a single connector
> in the function, and leave the loops to the callers.
> 
> static void ucsi_unregister_connector(struct ucsi_connector *con)
> {
>         cancel_work_sync(&con->work);
>         ucsi_unregister_partner(con);
>         ucsi_unregister_altmodes(con, UCSI_RECIPIENT_CON);
>         ucsi_unregister_port_psy(con);
>         if (con->wq)
>                 destroy_workqueue(con->wq);
>         typec_unregister_port(con->port);
> }
> 
> I wonder would it actually be a bit more clearer to do it like that...
> I'll leave the decision to you.

I will keep it,
There will be no duplicate loop entry in two function, a little lazy.

> 
> >  /**
> >   * ucsi_init - Initialize UCSI interface
> >   * @ucsi: UCSI to be initialized
> > @@ -1195,7 +1221,6 @@ static int ucsi_register_port(struct ucsi *ucsi, int
> index)
> >   */
> >  static int ucsi_init(struct ucsi *ucsi)
> >  {
> > -	struct ucsi_connector *con;
> >  	u64 command;
> >  	int ret;
> >  	int i;
> > @@ -1250,15 +1275,7 @@ static int ucsi_init(struct ucsi *ucsi)
> >  	return 0;
> >
> >  err_unregister:
> > -	for (con = ucsi->connector; con->port; con++) {
> > -		ucsi_unregister_partner(con);
> > -		ucsi_unregister_altmodes(con, UCSI_RECIPIENT_CON);
> > -		ucsi_unregister_port_psy(con);
> > -		if (con->wq)
> > -			destroy_workqueue(con->wq);
> > -		typec_unregister_port(con->port);
> > -		con->port = NULL;
> > -	}
> > +	ucsi_unregister_connectors(ucsi);
> >
> >  err_reset:
> >  	memset(&ucsi->cap, 0, sizeof(ucsi->cap));
> > @@ -1364,7 +1381,6 @@ EXPORT_SYMBOL_GPL(ucsi_register);
> >  void ucsi_unregister(struct ucsi *ucsi)
> >  {
> >  	u64 cmd = UCSI_SET_NOTIFICATION_ENABLE;
> > -	int i;
> >
> >  	/* Make sure that we are not in the middle of driver initialization */
> >  	cancel_work_sync(&ucsi->work);
> > @@ -1372,18 +1388,7 @@ void ucsi_unregister(struct ucsi *ucsi)
> >  	/* Disable notifications */
> >  	ucsi->ops->async_write(ucsi, UCSI_CONTROL, &cmd, sizeof(cmd));
> >
> > -	for (i = 0; i < ucsi->cap.num_connectors; i++) {
> > -		cancel_work_sync(&ucsi->connector[i].work);
> > -		ucsi_unregister_partner(&ucsi->connector[i]);
> > -		ucsi_unregister_altmodes(&ucsi->connector[i],
> > -					 UCSI_RECIPIENT_CON);
> > -		ucsi_unregister_port_psy(&ucsi->connector[i]);
> > -		if (ucsi->connector[i].wq)
> > -			destroy_workqueue(ucsi->connector[i].wq);
> > -		typec_unregister_port(ucsi->connector[i].port);
> > -	}
> > -
> > -	kfree(ucsi->connector);
> > +	ucsi_unregister_connectors(ucsi);
> >  }
> >  EXPORT_SYMBOL_GPL(ucsi_unregister);
> 
> thanks,
> 
> --
> heikki




[Index of Archives]     [Linux Media]     [Linux Input]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [Old Linux USB Devel Archive]

  Powered by Linux