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]
Date:   Tue, 14 Apr 2020 15:44:19 -0700
From:   Prashant Malani <pmalani@...omium.org>
To:     Enric Balletbo i Serra <enric.balletbo@...labora.com>
Cc:     Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Heikki Krogerus <heikki.krogerus@...ux.intel.com>,
        Jon Flatley <jflat@...omium.org>,
        Benson Leung <bleung@...omium.org>,
        Guenter Roeck <groeck@...omium.org>
Subject: Re: [PATCH v3 3/3] platform/chrome: typec: Register port partner

Thanks for the review Enric,

On Tue, Apr 14, 2020 at 7:08 AM Enric Balletbo i Serra
<enric.balletbo@...labora.com> wrote:
>
> Hi Prashant,
>
> Thank you for your patch.
>
> On 10/4/20 2:23, Prashant Malani wrote:
> > Register (and unregister) the port partner when a connect (and
> > disconnect) is detected.
> >
> > Co-developed-by: Jon Flatley <jflat@...omium.org>
> > Signed-off-by: Prashant Malani <pmalani@...omium.org>
> > ---
> >
> > Changes in v3:
> > - No changes.
> >
> > Changes in v2:
> > - Fixed error pointer return value.
> >
> >  drivers/platform/chrome/cros_ec_typec.c | 48 +++++++++++++++++++++++++
> >  1 file changed, 48 insertions(+)
> >
> > diff --git a/drivers/platform/chrome/cros_ec_typec.c b/drivers/platform/chrome/cros_ec_typec.c
> > index 56ded09a60ffb..304e0b20f279b 100644
> > --- a/drivers/platform/chrome/cros_ec_typec.c
> > +++ b/drivers/platform/chrome/cros_ec_typec.c
> > @@ -22,6 +22,9 @@ struct cros_typec_port {
> >       struct typec_port *port;
> >       /* Initial capabilities for the port. */
> >       struct typec_capability caps;
> > +     struct typec_partner *partner;
> > +     /* Port partner PD identity info. */
> > +     struct usb_pd_identity p_identity;
> >  };
> >
> >  /* Platform-specific data for the Chrome OS EC Type C controller. */
> > @@ -190,6 +193,30 @@ static int cros_typec_ec_command(struct cros_typec_data *typec,
> >       return ret;
> >  }
> >
> > +static int cros_typec_add_partner(struct cros_typec_data *typec, int port_num,
> > +                               bool pd_en)
> > +{
> > +     struct cros_typec_port *port = typec->ports[port_num];
> > +     struct typec_partner_desc p_desc = {
> > +             .usb_pd = pd_en,
> > +     };
> > +     int ret = 0;
> > +
> > +     /*
> > +      * Fill an initial PD identity, which will then be updated with info
> > +      * from the EC.
> > +      */
> > +     p_desc.identity = &port->p_identity;
> > +
> > +     port->partner = typec_register_partner(port->port, &p_desc);
> > +     if (IS_ERR_OR_NULL(port->partner)) {
> > +             ret = PTR_ERR(port->partner);
>
> If you're checking for IS_ERR_OR_NULL that means that port->partner can be NULL,
> so PTR_ERR(NULL) is 0 returning something that you don't really want.
>
> But looking at the typec_register_partner, NULL is not an option as returns a
> handle to the partner on success or ERR_PTR on failure. So, the code should just do:
>
> if (IS_ERR(port->partner))
>     return PTR_ERR(port->partner);
Will update the patch to do this on the next version, thanks.
>
> And AFAICS you don't need to set port->partner to NULL.
>
> Can you double check this?
I think we will need it, since we check for it in
cros_typec_set_port_params_v1() to avoid double registration of a
partner:

+       /* Register/remove partners when a connect/disconnect occurs. */
+       if (resp->enabled & PD_CTRL_RESP_ENABLED_CONNECTED) {
+               if (typec->ports[port_num]->partner)
+                       return;

So if it is not set to NULL on error earlier, the registration won't
take place here (in case another PD notification comes in while the
PD_CTRL_RESP_ENABLED_CONNECTED bit is still set). Kindly correct me if
this understanding is incorrect.

Best regards,

-Prashant

>
> Thanks,
>  Enric
>
> > +             port->partner = NULL;
> > +     }
> > +
> > +     return ret;
> > +}
> > +
> >  static void cros_typec_set_port_params_v0(struct cros_typec_data *typec,
> >               int port_num, struct ec_response_usb_pd_control *resp)
> >  {
> > @@ -212,6 +239,8 @@ static void cros_typec_set_port_params_v1(struct cros_typec_data *typec,
> >  {
> >       struct typec_port *port = typec->ports[port_num]->port;
> >       enum typec_orientation polarity;
> > +     bool pd_en;
> > +     int ret;
> >
> >       if (!(resp->enabled & PD_CTRL_RESP_ENABLED_CONNECTED))
> >               polarity = TYPEC_ORIENTATION_NONE;
> > @@ -226,6 +255,25 @@ static void cros_typec_set_port_params_v1(struct cros_typec_data *typec,
> >                       TYPEC_SOURCE : TYPEC_SINK);
> >       typec_set_vconn_role(port, resp->role & PD_CTRL_RESP_ROLE_VCONN ?
> >                       TYPEC_SOURCE : TYPEC_SINK);
> > +
> > +     /* Register/remove partners when a connect/disconnect occurs. */
> > +     if (resp->enabled & PD_CTRL_RESP_ENABLED_CONNECTED) {
> > +             if (typec->ports[port_num]->partner)
> > +                     return;
> > +
> > +             pd_en = resp->enabled & PD_CTRL_RESP_ENABLED_PD_CAPABLE;
> > +             ret = cros_typec_add_partner(typec, port_num, pd_en);
> > +             if (!ret)
> > +                     dev_warn(typec->dev,
> > +                              "Failed to register partner on port: %d\n",
> > +                              port_num);
> > +     } else {
> > +             if (!typec->ports[port_num]->partner)
> > +                     return;
> > +
> > +             typec_unregister_partner(typec->ports[port_num]->partner);
> > +             typec->ports[port_num]->partner = NULL;
> > +     }
> >  }
> >
> >  static int cros_typec_port_update(struct cros_typec_data *typec, int port_num)
> >

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ