Re: [v10] i2c: imx: support slave mode for imx I2C driver

[Oleksij Rempel]

On Fri, Dec 11, 2020 at 09:57:19AM -0800, Kevin Herbert wrote:
> Thanks for the tip. I attempted to see if the register was implemented
> on the i.mx6, and it doesn't appear to be. I'll reach out to my FAE. I
> found the datasheet and the register definitely isn't documented
> there.
> 
> I was thinking of a patch that would keep track of state, and
> synthesize the I2C_SLAVE_STOP on the next interrupt if it was a new
> transaction. Does this seem too hacky to you? What are your thoughts?

I would say, send patches :)
Please provide in to the commit message enough description to understand and
reproduce your issue.
 
> 
> On Wed, Dec 9, 2020 at 6:16 PM Biwen Li <biwen.li@xxxxxxx> wrote:
> >
> > Hi Kevin,
> >
> >
> >
> > After enabling idle interrupts, the i2c register will generate an idle interrupts(whatever read or write) when i2c bus enter idle status. Then get I2C_SLAVE_STOP event.
> >
> > But don’t have the IBIC register(Maybe it’s a hidden register) in imx. You can query about the AE of imx about this.
> >
> > static void i2c_imx_enable_bus_idle(struct imx_i2c_struct *i2c_imx)
> >
> > {
> >
> >         if (is_vf610_i2c(i2c_imx)) {
> >
> >                 unsigned int temp;
> >
> >
> >
> >                 temp = imx_i2c_read_reg(i2c_imx, IMX_I2C_IBIC);
> >
> >                 temp |= IBIC_BIIE;
> >
> >                 imx_i2c_write_reg(temp, i2c_imx, IMX_I2C_IBIC);
> >
> >         }
> >
> > }
> >
> >
> >
> > Best Regards,
> >
> > Biwen Li
> >
> > From: Kevin Herbert <kph@xxxxxxxxxxxxxxxxxx>
> > Sent: 2020年12月10日 1:18
> > To: Wolfram Sang <wsa@xxxxxxxxxxxxx>; Kevin Herbert <kph@xxxxxxxxxxxxxxxxxx>; Biwen Li (OSS) <biwen.li@xxxxxxxxxxx>; Leo Li <leoyang.li@xxxxxxx>; linux@xxxxxxxxxxxxxxxx; kernel@xxxxxxxxxxxxxx; shawnguo@xxxxxxxxxx; s.hauer@xxxxxxxxxxxxxx; festevam@xxxxxxxxx; Aisheng Dong <aisheng.dong@xxxxxxx>; Clark Wang <xiaoning.wang@xxxxxxx>; o.rempel@xxxxxxxxxxxxxx; linux-i2c@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx; Jiafei Pan <jiafei.pan@xxxxxxx>; Xiaobo Xie <xiaobo.xie@xxxxxxx>; linux-arm-kernel@xxxxxxxxxxxxxxxxxxx; Biwen Li <biwen.li@xxxxxxx>
> > Subject: [EXT] Re: [v10] i2c: imx: support slave mode for imx I2C driver
> >
> >
> >
> > Caution: EXT Email
> >
> > Even on an operation like writing a byte, I get I2C_SLAVE_WRITE_REQUESTED followed by I2C_SLAVE_WRITE_RECEIVED, but no I2C_SLAVE_STOP. If I do a I2C write of multiple bytes, I get I2c_SLAVE_WRITE_REQUESTED followed by multiple I2C_SLAVE_WRITE_RECEIVED.
> >
> >
> >
> > Kevin
> >
> >
> >
> > On Wed, Dec 9, 2020 at 9:10 AM Wolfram Sang <wsa@xxxxxxxxxxxxx> wrote:
> >
> > On Wed, Dec 09, 2020 at 09:03:50AM -0800, Kevin Herbert wrote:
> > > What is the protocol for the I2C_SLAVE_STOP event? I am working on my own
> > > backend, and I've only tried it with this i.mx driver, and I do not receive
> > > I2C_SLAVE_STOP at the end of every I2C transaction. It was my expectation
> > > I'd receive this event at the end of every frame. In my testing, I've never
> > > received this event at all.
> > >
> > > Where are the I2C registers on the i.mx documented? My board is an i.mx6sx.
> >
> > Hmm, from a glimpse, it looks the STOP event is only sent after a write
> > and not after a read? Does this match your findings?
> 
> 

-- 
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |