RE: [EXT] Re: [PATCH 1/2] i2c: imx-lpi2c: add bus recovery feature

[Carlos Song]

Hi, Andi

Sorry for the long time to reply. According to your advice, I found the patch is too redundant!
so I will send V2 patch. 
I find gpio and pinctrl assignement from the default i2c_init_recovery() have been defined very well.
Lpi2c have special initialization conditions for i2c recovery and I have added a comment in V2.

> -----Original Message-----
> From: Andi Shyti <andi.shyti@xxxxxxxxxx>
> Sent: Wednesday, June 14, 2023 7:20 AM
> To: Carlos Song <carlos.song@xxxxxxx>
> Cc: Aisheng Dong <aisheng.dong@xxxxxxx>; shawnguo@xxxxxxxxxx;
> s.hauer@xxxxxxxxxxxxxx; kernel@xxxxxxxxxxxxxx; festevam@xxxxxxxxx;
> robh+dt@xxxxxxxxxx; krzysztof.kozlowski+dt@xxxxxxxxxx; conor+dt@xxxxxxxxxx;
> Anson.Huang@xxxxxxx; Clark Wang <xiaoning.wang@xxxxxxx>; Bough Chen
> <haibo.chen@xxxxxxx>; dl-linux-imx <linux-imx@xxxxxxx>;
> linux-i2c@xxxxxxxxxxxxxxx; devicetree@xxxxxxxxxxxxxxx;
> linux-arm-kernel@xxxxxxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx
> Subject: [EXT] Re: [PATCH 1/2] i2c: imx-lpi2c: add bus recovery feature
> 
> Caution: This is an external email. Please take care when clicking links or
> opening attachments. When in doubt, report the message using the 'Report this
> email' button
> 
> 
> Hi,
> 
> On Mon, May 29, 2023 at 03:43:01PM +0800, carlos.song@xxxxxxx wrote:
> > From: Clark Wang <xiaoning.wang@xxxxxxx>
> >
> > Add bus recovery feature for LPI2C.
> > Need add gpio pinctrl, scl-gpios and sda-gpios configuration in dts.
> 
> please update the commit message according to the dts changes, as well.
> 
Carlos: There is still a need to add gpio pinctrl to set i2c sda/scl pin to gpio
> [...]
> 
> > +static void lpi2c_imx_prepare_recovery(struct i2c_adapter *adap) {
> > +     struct lpi2c_imx_struct *lpi2c_imx;
> > +
> > +     lpi2c_imx = container_of(adap, struct lpi2c_imx_struct,
> > + adapter);
> > +
> > +     pinctrl_select_state(lpi2c_imx->pinctrl,
> > +lpi2c_imx->pinctrl_pins_gpio); }
> > +
> > +static void lpi2c_imx_unprepare_recovery(struct i2c_adapter *adap) {
> > +     struct lpi2c_imx_struct *lpi2c_imx;
> > +
> > +     lpi2c_imx = container_of(adap, struct lpi2c_imx_struct,
> > + adapter);
> > +
> > +     pinctrl_select_state(lpi2c_imx->pinctrl,
> > +lpi2c_imx->pinctrl_pins_default); }
> > +
> > +/*
> > + * We switch SCL and SDA to their GPIO function and do some
> > +bitbanging
> > + * for bus recovery. These alternative pinmux settings can be
> > + * described in the device tree by a separate pinctrl state "gpio".
> > +If
> 
> is this still true?
> 

Carlos: Yes it is true.

> > + * this is missing this is not a big problem, the only implication is
> > + * that we can't do bus recovery.
> > + */
> > +static int lpi2c_imx_init_recovery_info(struct lpi2c_imx_struct *lpi2c_imx,
> > +             struct platform_device *pdev) {
> > +     struct i2c_bus_recovery_info *rinfo = &lpi2c_imx->rinfo;
> > +
> > +     lpi2c_imx->pinctrl = devm_pinctrl_get(&pdev->dev);
> > +     if (!lpi2c_imx->pinctrl || IS_ERR(lpi2c_imx->pinctrl)) {
> > +             dev_info(&pdev->dev, "can't get pinctrl, bus recovery not
> supported\n");
> > +             return PTR_ERR(lpi2c_imx->pinctrl);
> > +     }
> > +
> > +     lpi2c_imx->pinctrl_pins_default =
> pinctrl_lookup_state(lpi2c_imx->pinctrl,
> > +                     PINCTRL_STATE_DEFAULT);
> > +     lpi2c_imx->pinctrl_pins_gpio = pinctrl_lookup_state(lpi2c_imx->pinctrl,
> > +                     "gpio");
> > +     rinfo->sda_gpiod = devm_gpiod_get(&pdev->dev, "sda", GPIOD_IN);
> > +     rinfo->scl_gpiod = devm_gpiod_get(&pdev->dev, "scl",
> > + GPIOD_OUT_HIGH_OPEN_DRAIN);
> > +
> > +     if (PTR_ERR(rinfo->sda_gpiod) == -EPROBE_DEFER ||
> > +         PTR_ERR(rinfo->scl_gpiod) == -EPROBE_DEFER) {
> > +             return -EPROBE_DEFER;
> > +     } else if (IS_ERR(rinfo->sda_gpiod) ||
> > +                IS_ERR(rinfo->scl_gpiod) ||
> > +                IS_ERR(lpi2c_imx->pinctrl_pins_default) ||
> > +                IS_ERR(lpi2c_imx->pinctrl_pins_gpio)) {
> > +             dev_dbg(&pdev->dev, "recovery information incomplete\n");
> > +             return 0;
> > +     }
> 
> Why not use these assignement from the default i2c_init_recovery()? Is there
> anything you are doing I am not seeing?
> 

Carlos: these assignements are too redundant and I will fix it in V2 patch.

> > +
> > +     dev_info(&pdev->dev, "using scl%s for recovery\n",
> > +              rinfo->sda_gpiod ? ",sda" : "");
> 
> is there any case when sda_gpiod is NULL?
> 
Carlos: I will delete it in V2.
> > +
> > +     rinfo->prepare_recovery = lpi2c_imx_prepare_recovery;
> > +     rinfo->unprepare_recovery = lpi2c_imx_unprepare_recovery;
> > +     rinfo->recover_bus = i2c_generic_scl_recovery;
> > +     lpi2c_imx->adapter.bus_recovery_info = rinfo;
> 
> do you need also the set_scl() function? It should be mandatory.
> 
Carlos: I will use the default setting in V2 patch.
> > +
> > +     return 0;
> > +}
> 
> Besides, this is a copy/paste from i2c-imx.c, any chance to put the two things
> together?
> 
Carlos: I hope to apply the new recovery patch for lpi2c.
> Andi