linux-kernel - Re: [PATCH v2] usb: typec: tcpm: Relax disconnect threshold during power negotiation

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

Date:   Tue, 15 Jun 2021 11:05:50 -0700
From:   Badhri Jagan Sridharan <badhri@...gle.com>
To:     Kyle Tso <kyletso@...gle.com>
Cc:     Guenter Roeck <linux@...ck-us.net>,
        Heikki Krogerus <heikki.krogerus@...ux.intel.com>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        USB <linux-usb@...r.kernel.org>,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2] usb: typec: tcpm: Relax disconnect threshold during
 power negotiation

On Tue, Jun 15, 2021 at 1:08 AM Kyle Tso <kyletso@...gle.com> wrote:
>
> If the voltage is being decreased in power negotiation, the Source will
> set the power supply to operate at the new voltage level before sending
> PS_RDY. For non-PPS negotiation, relax the disconnect threshold on Sink
> after receiving Accept Message to ensure the relaxed setting is enabled
> before the voltage collapse. For PPS, relax the threshold before
> sending Request Message so that it will not race with Source which
> begins to adjust the voltage right after it sends Accept Message.
>
> The real threshold will be set after Sink receives PS_RDY Message.
>
> Fixes: f321a02caebd ("usb: typec: tcpm: Implement enabling Auto Discharge disconnect support")
> Cc: Badhri Jagan Sridharan <badhri@...gle.com>
> Signed-off-by: Kyle Tso <kyletso@...gle.com>
> ---
> Changes since v1:
> - move the timing of setting threshold up to "before sending Request"
>   for PPS power negotiation so that it won't race with the Source.
> - PPS: if it fails to send the Request, fallback to previous threshold
> - PPS: if the Source doesn't respond Accept, fallback to previous
>   threshold
> - update the commit message for above changes
>
>  drivers/usb/typec/tcpm/tcpm.c | 14 ++++++++++++++
>  1 file changed, 14 insertions(+)
>
> diff --git a/drivers/usb/typec/tcpm/tcpm.c b/drivers/usb/typec/tcpm/tcpm.c
> index 0db685d5d9c0..00f3fd7c05d6 100644
> --- a/drivers/usb/typec/tcpm/tcpm.c
> +++ b/drivers/usb/typec/tcpm/tcpm.c
> @@ -2599,6 +2599,11 @@ static void tcpm_pd_ctrl_request(struct tcpm_port *port,
>                             port->send_discover)
>                                 port->vdm_sm_running = true;
>
> +                       /* Threshold was relaxed before sending Request. Restore it back. */
> +                       tcpm_set_auto_vbus_discharge_threshold(port, TYPEC_PWR_MODE_PD,
> +                                                              port->pps_data.active,
> +                                                              port->supply_voltage);
> +
>                         tcpm_set_state(port, SNK_READY, 0);
>                         break;
>                 case DR_SWAP_SEND:
> @@ -2646,6 +2651,8 @@ static void tcpm_pd_ctrl_request(struct tcpm_port *port,
>                 switch (port->state) {
>                 case SNK_NEGOTIATE_CAPABILITIES:
>                         port->pps_data.active = false;
> +                       /* Voltage is going to be at new level. Relax the threshold here. */
> +                       tcpm_set_auto_vbus_discharge_threshold(port, TYPEC_PWR_MODE_USB, false, 0);
>                         tcpm_set_state(port, SNK_TRANSITION_SINK, 0);
>                         break;
>                 case SNK_NEGOTIATE_PPS_CAPABILITIES:
> @@ -3423,6 +3430,9 @@ static int tcpm_pd_send_pps_request(struct tcpm_port *port)
>         if (ret < 0)
>                 return ret;
>
> +       /* Relax the threshold as voltage will be adjusted right after Accept Message. */
This makes sense. Shouldn't we have the same approach for
tcpm_pd_send_request as it's equally applicable for fixed RDO as well
?

> +       tcpm_set_auto_vbus_discharge_threshold(port, TYPEC_PWR_MODE_USB, false, 0);
> +
>         memset(&msg, 0, sizeof(msg));
>         msg.header = PD_HEADER_LE(PD_DATA_REQUEST,
>                                   port->pwr_role,
> @@ -4196,6 +4206,10 @@ static void run_state_machine(struct tcpm_port *port)
>         case SNK_NEGOTIATE_PPS_CAPABILITIES:
>                 ret = tcpm_pd_send_pps_request(port);
>                 if (ret < 0) {
> +                       /* Restore back to the original state */
> +                       tcpm_set_auto_vbus_discharge_threshold(port, TYPEC_PWR_MODE_PD,
> +                                                              port->pps_data.active,
> +                                                              port->supply_voltage);
>                         port->pps_status = ret;
>                         /*
>                          * If this was called due to updates to sink
> --
> 2.32.0.272.g935e593368-goog
>