Received-SPF: pass (google.com: domain of linux-kernel+bounces-145281-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:40f1:3f00::1 as permitted sender) client-ip=2604:1380:40f1:3f00::1;
Date: Mon, 15 Apr 2024 16:43:25 +0300
From: Dan Carpenter <dan.carpenter@linaro.org>
To: Robert Marko <robimarko@gmail.com>
Cc: Hanna Hawa <hhhawa@amazon.com>, andriy.shevchenko@linux.intel.com,
	wsa@kernel.org, linus.walleij@linaro.org, linux-i2c@vger.kernel.org,
	linux-kernel@vger.kernel.org, linux-gpio@vger.kernel.org,
	dwmw@amazon.co.uk, benh@amazon.com, ronenk@amazon.com,
	talel@amazon.com, jonnyc@amazon.com, hanochu@amazon.com,
	farbere@amazon.com, itamark@amazon.com
Subject: Re: [PATCH v5 2/2] i2c: Set i2c pinctrl recovery info from it's
 device pinctrl
Message-ID: <f7069e5b-0678-4f78-b06e-dba16c5e6088@moroto.mountain>
References: <20221228164813.67964-1-hhhawa@amazon.com>
 <20221228164813.67964-3-hhhawa@amazon.com>
 <416340b6-33a9-4b9e-bdc5-c5a9cffb3055@gmail.com>
 <ac51854b-09a6-4b79-b409-b950929655cb@moroto.mountain>
 <CAOX2RU6YOBBgEuwdp8P0GTJ5vB0M5Cbqf5SnVJ9Jbou9w5405g@mail.gmail.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CAOX2RU6YOBBgEuwdp8P0GTJ5vB0M5Cbqf5SnVJ9Jbou9w5405g@mail.gmail.com>

On Sun, Apr 14, 2024 at 07:47:50PM +0200, Robert Marko wrote:
> > > > @@ -282,7 +283,9 @@ static void i2c_gpio_init_pinctrl_recovery(struct i2c_adapter *adap)
> > > >   {
> > > >     struct i2c_bus_recovery_info *bri = adap->bus_recovery_info;
> > > >     struct device *dev = &adap->dev;
> > > > -   struct pinctrl *p = bri->pinctrl;
> > > > +   struct pinctrl *p = bri->pinctrl ?: dev_pinctrl(dev->parent);
> > > > +
> > > > +   bri->pinctrl = p;
> > >
> > > Hi Hanna,
> > > I know this has already been merged, but setting bri->pinctrl breaks PXA
> > > recovery.
> >
> > This is patch is a year and half old so it's a bit late to just revert
> > it...
> 
> Hi there,
> I know it's old but I just tried it on 6.6 in OpenWrt.
> 
> >
> > What does "breaks" mean in this context?  Is there a NULL dereference?
> > Do you have a stack trace?  It's really hard to get inspired to look at
> > the code when the bug report is so vague...
> 
> I admit that I did not explain this properly, but if bri->pinctrl is set then
> PXA I2C is completely broken as in it doesn't work at all, there are no errors
> other than trying to probe for I2C devices will time out.
> We had the same symptoms when PXA was converted to generic I2C recovery and that
> had to be reverted.
> 
> I think its probably some pinctrl issue but nobody has been able to
> track it down.

If you wanted you could try the following patch with the change to
i2c_gpio_init_pinctrl_recovery() and without it.  (It won't fix anything
it only prints information to dmesg).

regards,
dan carpenter

diff --git a/drivers/i2c/busses/i2c-pxa.c b/drivers/i2c/busses/i2c-pxa.c
index 888ca636f3f3..f9477089b980 100644
--- a/drivers/i2c/busses/i2c-pxa.c
+++ b/drivers/i2c/busses/i2c-pxa.c
@@ -34,6 +34,7 @@
 #include <linux/platform_data/i2c-pxa.h>
 #include <linux/property.h>
 #include <linux/slab.h>
+#include "../../pinctrl/core.h"
 
 /* I2C register field definitions */
 #define IBMR_SDAS	(1 << 0)
@@ -1345,6 +1346,12 @@ static int i2c_pxa_init_recovery(struct pxa_i2c *i2c)
 		return 0;
 
 	i2c->pinctrl = devm_pinctrl_get(dev);
+	if (IS_ERR(i2c->pinctrl))
+		dev_info(dev, "i2c->pinctrl: %pe\n", i2c->pinctrl);
+	else
+		dev_info(dev, "i2c->pinctrl: %s %s\n",
+			 dev_driver_string(i2c->pinctrl->dev),
+			 dev_name(i2c->pinctrl->dev));
 	if (PTR_ERR(i2c->pinctrl) == -ENODEV)
 		i2c->pinctrl = NULL;
 	if (IS_ERR(i2c->pinctrl))