Re: [PATCH 2/2] i3c: master: svc: fix probe failure when no i3c device exist

[Miquel Raynal]

Hi Frank,

Frank.li@xxxxxxx wrote on Tue, 29 Aug 2023 11:26:42 -0400:

> On Tue, Aug 29, 2023 at 11:16:11AM +0200, Miquel Raynal wrote:
> > Hi Frank,
> > 
> > Frank.Li@xxxxxxx wrote on Mon, 28 Aug 2023 15:25:02 -0400:
> >   
> > > If there are not i3c device, all ccc command will get NACK. Set  
> > 
> > 		no					NACKed?
> >   
> > > i3c_ccc_cmd::err as I3C_ERROR_M2.  
> > 
> > This sentence should come last and be slightly more explicit.
> >   
> > > Return success when no i3c device found at svc_i3c_master_do_daa_locked().  
> > 
> > Please explain why this is important/useful.  
> 
> If return failure, driver will probe failure when there are no any i3c
> devices. I3C master supposed support hot join. The probe failure don't
> make sense if no i3c devices found when master driver probe.
> 
> How about rewrite commit log as

Much better, let me propose something slightly clearer.

> "I3C master supports device hot join, it doesn't make sense master driver
> probe failure when there are no I3C devices.
>
> When there are no I3C devices attached, all CCC commands are NACKed. So
> SVC I3C master fails to probe.

What about:

I3C masters are expected to support hot-join. This means at
initialization time we might not yet discover any device and this
should not be treated as a fatal error.

During the DAA procedure which happens at probe time, if no device has
joined, all CCC will be NACKed (from a bus perspective). This leads to
an early return with an error code which fails the probe of the master.

Let's avoid this by just telling the core through an I3C_ERROR_M2
return command code that no device was discovered, which is a valid
situation. This way the master will no longer bail out and fail to
probe for a wrong reason.

> Set i3c_cc_cmd:err as I3C_ERROR_M2. So I3C master framework
> i3c_master_rstdaa_locked() can return expected ERROR code and continue
> driver probe process.
> 
> Return success at svc_i3c_master_do_daa_locked() if no i3c devices found.
> So SVC master driver can probe successfully even if no I3C devices are
> attached"
> 
> >   
> > > Fixes: dd3c52846d59 ("i3c: master: svc: Add Silvaco I3C master driver")  
> > 
> > Shall we consider a backport into stable kernels?  
> 
> Yes, I can add stable tag at next version.

Yes, please.

> >   
> > > Signed-off-by: Frank Li <Frank.Li@xxxxxxx>
> > > ---
> > >  drivers/i3c/master/svc-i3c-master.c | 13 +++++++++++--
> > >  1 file changed, 11 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/drivers/i3c/master/svc-i3c-master.c b/drivers/i3c/master/svc-i3c-master.c
> > > index 770b40e28015e..a5620103acb73 100644
> > > --- a/drivers/i3c/master/svc-i3c-master.c
> > > +++ b/drivers/i3c/master/svc-i3c-master.c
> > > @@ -789,6 +789,9 @@ static int svc_i3c_master_do_daa_locked(struct svc_i3c_master *master,
> > >  				 */
> > >  				break;
> > >  			} else if (SVC_I3C_MSTATUS_NACKED(reg)) {
> > > +				/* No I3C devices attached */
> > > +				if (dev_nb == 0)
> > > +					break;  
> > 
> > \n ?
> >   
> > >  				/*
> > >  				 * A slave device nacked the address, this is
> > >  				 * allowed only once, DAA will be stopped and
> > > @@ -1263,11 +1266,17 @@ static int svc_i3c_master_send_ccc_cmd(struct i3c_master_controller *m,
> > >  {
> > >  	struct svc_i3c_master *master = to_svc_i3c_master(m);
> > >  	bool broadcast = cmd->id < 0x80;
> > > +	int ret;
> > >  
> > >  	if (broadcast)
> > > -		return svc_i3c_master_send_bdcast_ccc_cmd(master, cmd);
> > > +		ret = svc_i3c_master_send_bdcast_ccc_cmd(master, cmd);
> > >  	else
> > > -		return svc_i3c_master_send_direct_ccc_cmd(master, cmd);
> > > +		ret = svc_i3c_master_send_direct_ccc_cmd(master, cmd);
> > > +
> > > +	if (ret)
> > > +		cmd->err = I3C_ERROR_M2;
> > > +
> > > +	return ret;
> > >  }
> > >  
> > >  static int svc_i3c_master_priv_xfers(struct i3c_dev_desc *dev,  
> > 
> > 
> > Thanks,
> > Miquèl  


Thanks,
Miquèl