lkml.org 
[lkml]   [2023]   [Mar]   [19]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH 4/4] usb: typec: tcpm: fix source caps may lost after soft reset
From
Hi Guenter,

On 2023/3/17 20:58, Guenter Roeck wrote:
> On 3/12/23 19:58, Frank Wang wrote:
>> Invoke set_pd_rx() may flush the RX FIFO of PD controller, so do
>> set_pd_rx() before sending Soft Reset in case Source caps may be flushed
>> at debounce time between SOFT_RESET_SEND and SNK_WAIT_CAPABILITIES
>> state.
>>
>
> Isn't that a problem of the fusb302 driver that it flushes its buffers
> unconditionally when its set_pd_rx() callback is called ?
>
> Guenter
>

The story goes like this,  the fusb302 notified SOFT_RESET completion to
TCPM and began to receive the Source Caps automatically,
TCPM got completion from fusb302 and changed state to
SNK_WAIT_CAPABILITIES and invoked set_pd_rx() callback. However, the
fusb302 or TCPM's worker may be not scheduled in time, set_pd_rx() would
be performed after the Source Caps packets had received
into fusb302's FIFO, even after the GoodCRC (Source Caps) had be replied.

So make forward set_pd_rx() process before PD_CTRL_SOFT_RESET sent at
SOFT_RESET_SEND state can cleanup the context in our side
and ensure the right PD commucation. I am not sure whether it is sensible?

BR.
Frank

>> Without this patch, in PD charger stress test, the FUSB302 driver may
>> occur the following exceptions in power negotiation stage.
>>
>> [ ...]
>> [ 4.512252] fusb302_irq_intn
>> [ 4.512260] AMS SOFT_RESET_AMS finished
>> [ 4.512269] state change SOFT_RESET_SEND ->SNK_WAIT_CAPABILITIES
>> [rev3 NONE_AMS]
>> [ 4.514511] pd := on
>> [ 4.514516] pending state change SNK_WAIT_CAPABILITIES
>> ->HARD_RESET_SEND @ 310 ms [rev3 NONE_AMS]
>> [ 4.515428] IRQ: 0x51, a: 0x00, b: 0x01, status0: 0x93
>> [ 4.515431] IRQ: BC_LVL, handler pending
>> [ 4.515435] IRQ: PD sent good CRC
>> [ 4.516434] PD message header: 0
>> [ 4.516437] PD message len: 0
>> [ 4.516444] PD RX, header: 0x0 [1]
>>
>> Signed-off-by: Frank Wang <frank.wang@rock-chips.com>
>> ---
>>   drivers/usb/typec/tcpm/tcpm.c | 11 +++++++----
>>   1 file changed, 7 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/usb/typec/tcpm/tcpm.c
>> b/drivers/usb/typec/tcpm/tcpm.c
>> index 9e583060e64fc..ba6bf71838eed 100644
>> --- a/drivers/usb/typec/tcpm/tcpm.c
>> +++ b/drivers/usb/typec/tcpm/tcpm.c
>> @@ -4321,10 +4321,12 @@ static void run_state_machine(struct
>> tcpm_port *port)
>>           tcpm_set_state(port, unattached_state(port), 0);
>>           break;
>>       case SNK_WAIT_CAPABILITIES:
>> -        ret = port->tcpc->set_pd_rx(port->tcpc, true);
>> -        if (ret < 0) {
>> -            tcpm_set_state(port, SNK_READY, 0);
>> -            break;
>> +        if (port->prev_state != SOFT_RESET_SEND) {
>> +            ret = port->tcpc->set_pd_rx(port->tcpc, true);
>> +            if (ret < 0) {
>> +                tcpm_set_state(port, SNK_READY, 0);
>> +                break;
>> +            }
>>           }
>>           /*
>>            * If VBUS has never been low, and we time out waiting
>> @@ -4603,6 +4605,7 @@ static void run_state_machine(struct tcpm_port
>> *port)
>>       case SOFT_RESET_SEND:
>>           port->message_id = 0;
>>           port->rx_msgid = -1;
>> +        port->tcpc->set_pd_rx(port->tcpc, true);
>>           if (tcpm_pd_send_control(port, PD_CTRL_SOFT_RESET))
>>               tcpm_set_state_cond(port, hard_reset_state(port), 0);
>>           else
>

\
 
 \ /
  Last update: 2023-03-27 01:07    [W:0.067 / U:0.804 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site