RE: [PATCH v4 10/13] usb: typec: tcpm: set cc for drp toggling attach

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

 




> -----Original Message-----
> From: Guenter Roeck [mailto:groeck7@xxxxxxxxx] On Behalf Of Guenter Roeck
> Sent: 2018年3月30日 6:49
> To: Jun Li <jun.li@xxxxxxx>
> Cc: robh+dt@xxxxxxxxxx; gregkh@xxxxxxxxxxxxxxxxxxx;
> heikki.krogerus@xxxxxxxxxxxxxxx; a.hajda@xxxxxxxxxxx;
> shufan_lee@xxxxxxxxxxx; Peter Chen <peter.chen@xxxxxxx>;
> devicetree@xxxxxxxxxxxxxxx; linux-usb@xxxxxxxxxxxxxxx; dl-linux-imx
> <linux-imx@xxxxxxx>; devel@xxxxxxxxxxxxxxxxxxxx
> Subject: Re: [PATCH v4 10/13] usb: typec: tcpm: set cc for drp toggling attach
> 
> On Thu, Mar 29, 2018 at 12:06:15AM +0800, Li Jun wrote:
> > In case of drp toggling, we may need set correct cc value for role
> > control after attach as it may never been set.
> >
> 
> Isn't CC set by the lower level driver in this case ? In other words, is it ever
> necessary to call back into the low level driver to set CC again ? Doing that in
> attached state seems a bit late.

In question case(drp without try.src or try.snk), you can see tcpm never call
tcpm_set_cc() from drp toggling to attached state, start_drp_toggling
set cc at the beginning in this case, but the value is the *start* value, may not
the right value to match its final role.

Per tcpci spec
Figure 4-16. DRP Initialization and Connection Detection
After debounce, tcpm should set cc(again, but may different value) and polarity.

ShuFan encountered this case as I understood on his setup like below:
- Tcpc start drp toggling with Rp/Rp,
- Connect to adapter with also Rp/Rp
- After connection resolved, the tcpc becomes a sink(HW present Rd), but
role control register value is still Rp/Rp.
- This should be corrected to set cc for keep the un-contacted cc line open.

> 
> It may make more sense to update port->cc_req when the state machine leaves
> DRP_TOGGLING state, ie in _tcpm_cc_change(), and to do it without callback into
> the low level driver (it should not be necessary).

Currently there is no user of cc_req flag, so I can't catch the real intention
of this flag, as of now it's only set in tcpm_set_cc(), so I understood it's
for if tcpm already set tcpc a specific role.

Li Jun
> 
> Guenter
> 
> > Signed-off-by: Li Jun <jun.li@xxxxxxx>
> > ---
> >  drivers/usb/typec/tcpm.c | 5 +++++
> >  1 file changed, 5 insertions(+)
> >
> > diff --git a/drivers/usb/typec/tcpm.c b/drivers/usb/typec/tcpm.c index
> > 218c230..72d4232 100644
> > --- a/drivers/usb/typec/tcpm.c
> > +++ b/drivers/usb/typec/tcpm.c
> > @@ -2126,6 +2126,7 @@ static void tcpm_reset_port(struct tcpm_port *port)
> >  	tcpm_set_attached_state(port, false);
> >  	port->try_src_count = 0;
> >  	port->try_snk_count = 0;
> > +	port->cc_req = 0;
> >  }
> >
> >  static void tcpm_detach(struct tcpm_port *port) @@ -2361,6 +2362,8 @@
> > static void run_state_machine(struct tcpm_port *port)
> >  		break;
> >
> >  	case SRC_ATTACHED:
> > +		if (!port->cc_req)
> > +			tcpm_set_cc(port, tcpm_rp_cc(port));
> >  		ret = tcpm_src_attach(port);
> >  		tcpm_set_state(port, SRC_UNATTACHED,
> >  			       ret < 0 ? 0 : PD_T_PS_SOURCE_ON); @@ -2531,6 +2534,8
> @@
> > static void run_state_machine(struct tcpm_port *port)
> >  		tcpm_set_state(port, SNK_UNATTACHED, PD_T_PD_DEBOUNCE);
> >  		break;
> >  	case SNK_ATTACHED:
> > +		if (!port->cc_req)
> > +			tcpm_set_cc(port, TYPEC_CC_RD);
> >  		ret = tcpm_snk_attach(port);
> >  		if (ret < 0)
> >  			tcpm_set_state(port, SNK_UNATTACHED, 0);
> > --
> > 2.7.4
> >
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel




[Index of Archives]     [Linux Driver Backports]     [DMA Engine]     [Linux GPIO]     [Linux SPI]     [Video for Linux]     [Linux USB Devel]     [Linux Coverity]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [Yosemite Backpacking]
  Powered by Linux