Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1480209 > unrolled thread

Re: [PATCH v5] i2c: imx: make bus recovery through pinctrl optional

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: [PATCH v5] i2c: imx: make bus recovery through pinctrl optional Stefan Agner <stefan@agner.ch> - 2016-09-09 19:00 +0200
    Re: [PATCH v5] i2c: imx: make bus recovery through pinctrl optional Leo Li <pku.leo@gmail.com> - 2016-09-09 21:40 +0200
      Re: [PATCH v5] i2c: imx: make bus recovery through pinctrl optional Stefan Agner <stefan@agner.ch> - 2016-09-09 22:50 +0200
        Re: [PATCH v5] i2c: imx: make bus recovery through pinctrl optional Uwe Kleine-König          <u.kleine-koenig@pengutronix.de> - 2016-09-09 23:00 +0200
          Re: [PATCH v5] i2c: imx: make bus recovery through pinctrl optional Leo Li <pku.leo@gmail.com> - 2016-09-12 18:50 +0200
      Re: [PATCH v5] i2c: imx: make bus recovery through pinctrl optional Lothar Waßmann <LW@KARO-electronics.de> - 2016-09-12 10:30 +0200
        Re: [PATCH v5] i2c: imx: make bus recovery through pinctrl optional Leo Li <pku.leo@gmail.com> - 2016-09-12 18:30 +0200

#1480209 — Re: [PATCH v5] i2c: imx: make bus recovery through pinctrl optional

On 2016-09-08 16:57, Leo Li wrote:
> On Thu, Sep 8, 2016 at 5:39 PM, Stefan Agner <stefan@agner.ch> wrote:
>> On 2016-09-06 15:40, Leo Li wrote:
>>> On Tue, Sep 6, 2016 at 4:51 PM, Stefan Agner <stefan@agner.ch> wrote:
>>>> On 2016-09-06 13:06, Leo Li wrote:
>>>>> On Tue, Sep 6, 2016 at 1:58 PM, Uwe Kleine-König
>>>>> <u.kleine-koenig@pengutronix.de> wrote:
>>>>>> On Fri, Aug 19, 2016 at 05:05:22PM -0500, Li Yang wrote:
>> <snip>
>>>>>>> @@ -1081,8 +1090,11 @@ static int i2c_imx_probe(struct platform_device *pdev)
>>>>>>>               return ret;
>>>>>>>       }
>>>>>>>
>>>>>>> +     /* optional bus recovery feature through pinctrl */
>>>>>>>       i2c_imx->pinctrl = devm_pinctrl_get(&pdev->dev);
>>>>>>> -     if (IS_ERR(i2c_imx->pinctrl)) {
>>>>>>> +     /* bailout on -ENOMEM or -EPROBE_DEFER, continue for other errors */
>>>>>>> +     if (PTR_ERR(i2c_imx->pinctrl) == -ENOMEM ||
>>>>>>> +                     PTR_ERR(i2c_imx->pinctrl) == -EPROBE_DEFER) {
>>>>>>>               ret = PTR_ERR(i2c_imx->pinctrl);
>>>>>>>               goto clk_disable;
>>>>>>>       }
>>>>>>
>>>>>> devm_pinctrl_get might return the following error-valued pointers:
>>>>>>  - -EINVAL
>>>>>>  - -ENOMEM
>>>>>>  - -ENODEV
>>>>>>  - -EPROBE_DEFER
>>>>>>
>>>>>> There are several error paths returning -EINVAL, one is when an invalid
>>>>>> phandle is used. Do you really want to ignore that?
>>>>>>
>>>>>> IMO error handling is better done with inverse logic, that is continue
>>>>>> on some explicit error, bail out on all unknown stuff. This tends to be
>>>>>> more robust. Also the comment should be improved to not explain that for
>>>>>> -ENOMEM and -EPROBE_DEFER we bail out (which should be obvious for
>>>>>> anyone who can read C) but to explain why.
>>>>>
>>>>> What you said is true for normal error handling, but in this scenario
>>>>> it is intentional to ignore all pinctrl related errors except critical
>>>>> ones because failing to have pinctrl for an optional feature shouldn't
>>>>> impact the function of normal i2c.  We choose to catch -ENOMEM because
>>>>> the error could also cause problem for i2c probe, and -EPROBE_DEFER
>>>>> because it's possible that the pinctrl will be ready later and we want
>>>>> to give it a chance.  The i2c driver really don't care why the pinctrl
>>>>> was not usable.  I thought I added comment before the
>>>>
>>>> I don't agree. E.g. -EINVAL would appear if you pass devm_pinctrl_get an
>>>> invalid device. Currently you would silently ignore that, which is not
>>>> what you want.
>>>
>>> It is not silently ignored, there will be a message printed out saying
>>> pinctrl is not available and bus recovery is not supported.  On the
>>> contrary, without this change the entire i2c driver fails to work
>>> silently if pinctrl is somehow not working.  And if the system is so
>>> broken that the pointer to the i2c device is NULL, the probe of i2c
>>> would have already failed before this point.  We shouldn't count on an
>>> optional function of the driver to catch fundamental issues like this.
>>>
>>>>
>>>> You want to get the pinctrl in any case expect there isn't one. And that
>>>> is how you should formulate your if statement.
>>>>
>>>> /*
>>>>  * It is ok if no pinctrl device is available. We'll not be able to use
>>>> the
>>>>  * bus recovery feature, but otherwise the driver works fine...
>>>>  */
>>>> if (PTR_ERR(i2c_imx->pinctrl) != -ENODEV)
>>>
>>> I agree that there could be other possibilities that the pinctrl
>>> failed to work beside the reason I described in the commit
>>> message(platform doesn't support pinctrl at all).  But I don't think
>>> any of them other than the -ENOMEM and -EPROBE_DEFER deserves a bail
>>> out for the entire i2c driver.
>>
>> FWIW, I disagree. If there is pinctrl defined, you want be sure that it
>> gets applied properly, no matter what. E.g. when devm_pinctrl_get return
>> EINVAL (Uwe's example) the driver will continue and likely fail in
>> mysterious ways later on because the pins have not been muxed properly.
>> The driver should not load in that situation so that the developer is
>> forced to fix his mistakes. The only reason to bail out here is if there
>> is no pin controller (ENODEV). And it seems that Uwe also tends to that
>> solution.
> 
> With this patch the i2c bus recovery feature will be disabled if the
> devm_pinctrl_get() fails.  The pin mux setting will not be changed in
> either i2c probe stage or at runtime.  I don't think it can cause any
> trouble to normal I2C operation.  IMO, it is not good to *force*

If you have a pin controller, and you make a typo in your device tree
which leads to a wrong phandle and devm_pinctrl_get returning -EINVAL,
the system won't mux the pins... And that will certainly affect normal
I2C operation!

> people fix problem that they don't really care by deliberately enlarge
> the problem.  That's why we don't panic() on any error we found.  For
> those who do care about the bus recovery, they can get the information
> from the console.

IMHO, it is just stupid to ignore errors and then let the developer
later on trace back what the initial issue was. Error out early is a
common sense software design principle...

I am not asking for a panic(), I am just suggesting to only ignore
pinctrl if it returns -ENODEV, the case you care are about.

--
Stefan

[toc] | [next] | [standalone]

#1480283

On Fri, Sep 9, 2016 at 11:51 AM, Stefan Agner <stefan@agner.ch> wrote:
> On 2016-09-08 16:57, Leo Li wrote:
>> On Thu, Sep 8, 2016 at 5:39 PM, Stefan Agner <stefan@agner.ch> wrote:
>>> On 2016-09-06 15:40, Leo Li wrote:
>>>> On Tue, Sep 6, 2016 at 4:51 PM, Stefan Agner <stefan@agner.ch> wrote:
>>>>> On 2016-09-06 13:06, Leo Li wrote:
>>>>>> On Tue, Sep 6, 2016 at 1:58 PM, Uwe Kleine-König
>>>>>> <u.kleine-koenig@pengutronix.de> wrote:
>>>>>>> On Fri, Aug 19, 2016 at 05:05:22PM -0500, Li Yang wrote:
>>> <snip>
>>>>>>>> @@ -1081,8 +1090,11 @@ static int i2c_imx_probe(struct platform_device *pdev)
>>>>>>>>               return ret;
>>>>>>>>       }
>>>>>>>>
>>>>>>>> +     /* optional bus recovery feature through pinctrl */
>>>>>>>>       i2c_imx->pinctrl = devm_pinctrl_get(&pdev->dev);
>>>>>>>> -     if (IS_ERR(i2c_imx->pinctrl)) {
>>>>>>>> +     /* bailout on -ENOMEM or -EPROBE_DEFER, continue for other errors */
>>>>>>>> +     if (PTR_ERR(i2c_imx->pinctrl) == -ENOMEM ||
>>>>>>>> +                     PTR_ERR(i2c_imx->pinctrl) == -EPROBE_DEFER) {
>>>>>>>>               ret = PTR_ERR(i2c_imx->pinctrl);
>>>>>>>>               goto clk_disable;
>>>>>>>>       }
>>>>>>>
>>>>>>> devm_pinctrl_get might return the following error-valued pointers:
>>>>>>>  - -EINVAL
>>>>>>>  - -ENOMEM
>>>>>>>  - -ENODEV
>>>>>>>  - -EPROBE_DEFER
>>>>>>>
>>>>>>> There are several error paths returning -EINVAL, one is when an invalid
>>>>>>> phandle is used. Do you really want to ignore that?
>>>>>>>
>>>>>>> IMO error handling is better done with inverse logic, that is continue
>>>>>>> on some explicit error, bail out on all unknown stuff. This tends to be
>>>>>>> more robust. Also the comment should be improved to not explain that for
>>>>>>> -ENOMEM and -EPROBE_DEFER we bail out (which should be obvious for
>>>>>>> anyone who can read C) but to explain why.
>>>>>>
>>>>>> What you said is true for normal error handling, but in this scenario
>>>>>> it is intentional to ignore all pinctrl related errors except critical
>>>>>> ones because failing to have pinctrl for an optional feature shouldn't
>>>>>> impact the function of normal i2c.  We choose to catch -ENOMEM because
>>>>>> the error could also cause problem for i2c probe, and -EPROBE_DEFER
>>>>>> because it's possible that the pinctrl will be ready later and we want
>>>>>> to give it a chance.  The i2c driver really don't care why the pinctrl
>>>>>> was not usable.  I thought I added comment before the
>>>>>
>>>>> I don't agree. E.g. -EINVAL would appear if you pass devm_pinctrl_get an
>>>>> invalid device. Currently you would silently ignore that, which is not
>>>>> what you want.
>>>>
>>>> It is not silently ignored, there will be a message printed out saying
>>>> pinctrl is not available and bus recovery is not supported.  On the
>>>> contrary, without this change the entire i2c driver fails to work
>>>> silently if pinctrl is somehow not working.  And if the system is so
>>>> broken that the pointer to the i2c device is NULL, the probe of i2c
>>>> would have already failed before this point.  We shouldn't count on an
>>>> optional function of the driver to catch fundamental issues like this.
>>>>
>>>>>
>>>>> You want to get the pinctrl in any case expect there isn't one. And that
>>>>> is how you should formulate your if statement.
>>>>>
>>>>> /*
>>>>>  * It is ok if no pinctrl device is available. We'll not be able to use
>>>>> the
>>>>>  * bus recovery feature, but otherwise the driver works fine...
>>>>>  */
>>>>> if (PTR_ERR(i2c_imx->pinctrl) != -ENODEV)
>>>>
>>>> I agree that there could be other possibilities that the pinctrl
>>>> failed to work beside the reason I described in the commit
>>>> message(platform doesn't support pinctrl at all).  But I don't think
>>>> any of them other than the -ENOMEM and -EPROBE_DEFER deserves a bail
>>>> out for the entire i2c driver.
>>>
>>> FWIW, I disagree. If there is pinctrl defined, you want be sure that it
>>> gets applied properly, no matter what. E.g. when devm_pinctrl_get return
>>> EINVAL (Uwe's example) the driver will continue and likely fail in
>>> mysterious ways later on because the pins have not been muxed properly.
>>> The driver should not load in that situation so that the developer is
>>> forced to fix his mistakes. The only reason to bail out here is if there
>>> is no pin controller (ENODEV). And it seems that Uwe also tends to that
>>> solution.
>>
>> With this patch the i2c bus recovery feature will be disabled if the
>> devm_pinctrl_get() fails.  The pin mux setting will not be changed in
>> either i2c probe stage or at runtime.  I don't think it can cause any
>> trouble to normal I2C operation.  IMO, it is not good to *force*
>
> If you have a pin controller, and you make a typo in your device tree
> which leads to a wrong phandle and devm_pinctrl_get returning -EINVAL,
> the system won't mux the pins... And that will certainly affect normal
> I2C operation!

I see why we are having different understanding now.  The i2c-imx
driver doesn't rely on the pinctrl to get the correct pin-mux setting
for I2C operation.   The pinctrl code was just added for recently for
the bus recovery function.  The assumption is that the correct
pin-muxing has been taken care of by platform code or reset logic.
The pinctrl is only used to enter/exit bus recovery mode, nothing
else.  Hence optional.  So a typo in the pinctrl node only impacts bus
recovery function.

>
>> people fix problem that they don't really care by deliberately enlarge
>> the problem.  That's why we don't panic() on any error we found.  For
>> those who do care about the bus recovery, they can get the information
>> from the console.
>
> IMHO, it is just stupid to ignore errors and then let the developer
> later on trace back what the initial issue was. Error out early is a
> common sense software design principle...
>
> I am not asking for a panic(), I am just suggesting to only ignore
> pinctrl if it returns -ENODEV, the case you care are about.

It was just an analogy for enlarging the problem for getting
attention.  But you probably thought that it was not enlarging the
problem as you think pinctrl is required by the driver instead of an
optional thing.

Regards,
Leo

[toc] | [prev] | [next] | [standalone]

#1480346

On 2016-09-09 12:37, Leo Li wrote:
> On Fri, Sep 9, 2016 at 11:51 AM, Stefan Agner <stefan@agner.ch> wrote:
>> On 2016-09-08 16:57, Leo Li wrote:
>>> On Thu, Sep 8, 2016 at 5:39 PM, Stefan Agner <stefan@agner.ch> wrote:
>>>> On 2016-09-06 15:40, Leo Li wrote:
>>>>> On Tue, Sep 6, 2016 at 4:51 PM, Stefan Agner <stefan@agner.ch> wrote:
>>>>>> On 2016-09-06 13:06, Leo Li wrote:
>>>>>>> On Tue, Sep 6, 2016 at 1:58 PM, Uwe Kleine-König
>>>>>>> <u.kleine-koenig@pengutronix.de> wrote:
>>>>>>>> On Fri, Aug 19, 2016 at 05:05:22PM -0500, Li Yang wrote:
>>>> <snip>
>>>>>>>>> @@ -1081,8 +1090,11 @@ static int i2c_imx_probe(struct platform_device *pdev)
>>>>>>>>>               return ret;
>>>>>>>>>       }
>>>>>>>>>
>>>>>>>>> +     /* optional bus recovery feature through pinctrl */
>>>>>>>>>       i2c_imx->pinctrl = devm_pinctrl_get(&pdev->dev);
>>>>>>>>> -     if (IS_ERR(i2c_imx->pinctrl)) {
>>>>>>>>> +     /* bailout on -ENOMEM or -EPROBE_DEFER, continue for other errors */
>>>>>>>>> +     if (PTR_ERR(i2c_imx->pinctrl) == -ENOMEM ||
>>>>>>>>> +                     PTR_ERR(i2c_imx->pinctrl) == -EPROBE_DEFER) {
>>>>>>>>>               ret = PTR_ERR(i2c_imx->pinctrl);
>>>>>>>>>               goto clk_disable;
>>>>>>>>>       }
>>>>>>>>
>>>>>>>> devm_pinctrl_get might return the following error-valued pointers:
>>>>>>>>  - -EINVAL
>>>>>>>>  - -ENOMEM
>>>>>>>>  - -ENODEV
>>>>>>>>  - -EPROBE_DEFER
>>>>>>>>
>>>>>>>> There are several error paths returning -EINVAL, one is when an invalid
>>>>>>>> phandle is used. Do you really want to ignore that?
>>>>>>>>
>>>>>>>> IMO error handling is better done with inverse logic, that is continue
>>>>>>>> on some explicit error, bail out on all unknown stuff. This tends to be
>>>>>>>> more robust. Also the comment should be improved to not explain that for
>>>>>>>> -ENOMEM and -EPROBE_DEFER we bail out (which should be obvious for
>>>>>>>> anyone who can read C) but to explain why.
>>>>>>>
>>>>>>> What you said is true for normal error handling, but in this scenario
>>>>>>> it is intentional to ignore all pinctrl related errors except critical
>>>>>>> ones because failing to have pinctrl for an optional feature shouldn't
>>>>>>> impact the function of normal i2c.  We choose to catch -ENOMEM because
>>>>>>> the error could also cause problem for i2c probe, and -EPROBE_DEFER
>>>>>>> because it's possible that the pinctrl will be ready later and we want
>>>>>>> to give it a chance.  The i2c driver really don't care why the pinctrl
>>>>>>> was not usable.  I thought I added comment before the
>>>>>>
>>>>>> I don't agree. E.g. -EINVAL would appear if you pass devm_pinctrl_get an
>>>>>> invalid device. Currently you would silently ignore that, which is not
>>>>>> what you want.
>>>>>
>>>>> It is not silently ignored, there will be a message printed out saying
>>>>> pinctrl is not available and bus recovery is not supported.  On the
>>>>> contrary, without this change the entire i2c driver fails to work
>>>>> silently if pinctrl is somehow not working.  And if the system is so
>>>>> broken that the pointer to the i2c device is NULL, the probe of i2c
>>>>> would have already failed before this point.  We shouldn't count on an
>>>>> optional function of the driver to catch fundamental issues like this.
>>>>>
>>>>>>
>>>>>> You want to get the pinctrl in any case expect there isn't one. And that
>>>>>> is how you should formulate your if statement.
>>>>>>
>>>>>> /*
>>>>>>  * It is ok if no pinctrl device is available. We'll not be able to use
>>>>>> the
>>>>>>  * bus recovery feature, but otherwise the driver works fine...
>>>>>>  */
>>>>>> if (PTR_ERR(i2c_imx->pinctrl) != -ENODEV)
>>>>>
>>>>> I agree that there could be other possibilities that the pinctrl
>>>>> failed to work beside the reason I described in the commit
>>>>> message(platform doesn't support pinctrl at all).  But I don't think
>>>>> any of them other than the -ENOMEM and -EPROBE_DEFER deserves a bail
>>>>> out for the entire i2c driver.
>>>>
>>>> FWIW, I disagree. If there is pinctrl defined, you want be sure that it
>>>> gets applied properly, no matter what. E.g. when devm_pinctrl_get return
>>>> EINVAL (Uwe's example) the driver will continue and likely fail in
>>>> mysterious ways later on because the pins have not been muxed properly.
>>>> The driver should not load in that situation so that the developer is
>>>> forced to fix his mistakes. The only reason to bail out here is if there
>>>> is no pin controller (ENODEV). And it seems that Uwe also tends to that
>>>> solution.
>>>
>>> With this patch the i2c bus recovery feature will be disabled if the
>>> devm_pinctrl_get() fails.  The pin mux setting will not be changed in
>>> either i2c probe stage or at runtime.  I don't think it can cause any
>>> trouble to normal I2C operation.  IMO, it is not good to *force*
>>
>> If you have a pin controller, and you make a typo in your device tree
>> which leads to a wrong phandle and devm_pinctrl_get returning -EINVAL,
>> the system won't mux the pins... And that will certainly affect normal
>> I2C operation!
> 
> I see why we are having different understanding now.  The i2c-imx
> driver doesn't rely on the pinctrl to get the correct pin-mux setting
> for I2C operation.   The pinctrl code was just added for recently for
> the bus recovery function.  The assumption is that the correct
> pin-muxing has been taken care of by platform code or reset logic.
> The pinctrl is only used to enter/exit bus recovery mode, nothing
> else.  Hence optional.  So a typo in the pinctrl node only impacts bus
> recovery function.
> 

You are right, the framework usually makes sure that default pinctrl
gets applied. I got that wrong.

>>
>>> people fix problem that they don't really care by deliberately enlarge
>>> the problem.  That's why we don't panic() on any error we found.  For
>>> those who do care about the bus recovery, they can get the information
>>> from the console.
>>
>> IMHO, it is just stupid to ignore errors and then let the developer
>> later on trace back what the initial issue was. Error out early is a
>> common sense software design principle...
>>
>> I am not asking for a panic(), I am just suggesting to only ignore
>> pinctrl if it returns -ENODEV, the case you care are about.
> 
> It was just an analogy for enlarging the problem for getting
> attention.  But you probably thought that it was not enlarging the
> problem as you think pinctrl is required by the driver instead of an
> optional thing.

Yeah it is a bit a wording thing: In my understanding, pinctrl is
required on SoC's witch have a pin controller... It is just that the
driver does not need to get the pinctrl by itself because the stack is
taking care of it implicitly. And yes, that makes the particular example
not a real world example.

However, I still feel that only bailing out if -ENODEV is returned is
the cleaner solution, and expresses better what your intention is. But
then I am not the pinctrl maintainer and I don't feel that it is worth
debating more around it, so whatever...

--
Stefan

[toc] | [prev] | [next] | [standalone]

#1480354

Hello,

On Fri, Sep 09, 2016 at 01:34:31PM -0700, Stefan Agner wrote:
> Yeah it is a bit a wording thing: In my understanding, pinctrl is
> required on SoC's witch have a pin controller... It is just that the
> driver does not need to get the pinctrl by itself because the stack is
> taking care of it implicitly. And yes, that makes the particular example
> not a real world example.

At first I thought, too, that it's a fatal problem if getting the
pinctrl stuff fails. IMHO that shows that the comments (or the code) are
still not good enough.

Maybe we should do something like that:

/*
 * As the IP doesn't support bus recovery, we have to switch SCL and SDA
 * to their GPIO function and do some bitbanging. These alternative
 * pinmux settings can be described in the device tree by a separate
 * pinctrl state "gpio". If this is missing this is not a big problem,
 * the only implication is that we can't do bus recovery.
 */
static void i2c_imx_init_recovery_info(...)
{
	...

and then put

        i2c_imx->pinctrl = devm_pinctrl_get(&pdev->dev);
        if (IS_ERR(i2c_imx->pinctrl))
		return;

into this function (and remove it from i2c_imx_probe). This makes it
more obvious that .pinctrl is only ever used for recovery and as
i2c_imx_init_recovery_info is void there is no error to propagate.

Best regards
Uwe

-- 
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |

[toc] | [prev] | [next] | [standalone]

#1481543

On Fri, Sep 9, 2016 at 3:59 PM, Uwe Kleine-König
<u.kleine-koenig@pengutronix.de> wrote:
> Hello,
>
> On Fri, Sep 09, 2016 at 01:34:31PM -0700, Stefan Agner wrote:
>> Yeah it is a bit a wording thing: In my understanding, pinctrl is
>> required on SoC's witch have a pin controller... It is just that the
>> driver does not need to get the pinctrl by itself because the stack is
>> taking care of it implicitly. And yes, that makes the particular example
>> not a real world example.
>
> At first I thought, too, that it's a fatal problem if getting the
> pinctrl stuff fails. IMHO that shows that the comments (or the code) are
> still not good enough.

Ya.  If it has confused more than one people, it is likely to confuse
more.  I agree with you that we should make it more clear.

Thanks,
Leo

[toc] | [prev] | [next] | [standalone]

#1481033

Hi,

On Fri, 9 Sep 2016 14:37:12 -0500 Leo Li wrote:
> On Fri, Sep 9, 2016 at 11:51 AM, Stefan Agner <stefan@agner.ch> wrote:
> > On 2016-09-08 16:57, Leo Li wrote:
[...]
> >> people fix problem that they don't really care by deliberately enlarge
> >> the problem.  That's why we don't panic() on any error we found.  For
> >> those who do care about the bus recovery, they can get the information
> >> from the console.
> >
> > IMHO, it is just stupid to ignore errors and then let the developer
> > later on trace back what the initial issue was. Error out early is a
> > common sense software design principle...
> >
> > I am not asking for a panic(), I am just suggesting to only ignore
> > pinctrl if it returns -ENODEV, the case you care are about.
> 
> It was just an analogy for enlarging the problem for getting
> attention.  But you probably thought that it was not enlarging the
> problem as you think pinctrl is required by the driver instead of an
> optional thing.
> 
The I2C bus recovery is a feature which is not used during normal
operation and you won't find that it's not working unless something
unusual happens (probably only in a productive environment). It may even
be hard to trigger a condition to test the feature.
Thus IMO it is vital that the driver complains loudly if something is
missing to make this feature work if requested.


Lothar Waßmann

[toc] | [prev] | [next] | [standalone]

#1481526

On Mon, Sep 12, 2016 at 3:21 AM, Lothar Waßmann <LW@karo-electronics.de> wrote:
> Hi,
>
> On Fri, 9 Sep 2016 14:37:12 -0500 Leo Li wrote:
>> On Fri, Sep 9, 2016 at 11:51 AM, Stefan Agner <stefan@agner.ch> wrote:
>> > On 2016-09-08 16:57, Leo Li wrote:
> [...]
>> >> people fix problem that they don't really care by deliberately enlarge
>> >> the problem.  That's why we don't panic() on any error we found.  For
>> >> those who do care about the bus recovery, they can get the information
>> >> from the console.
>> >
>> > IMHO, it is just stupid to ignore errors and then let the developer
>> > later on trace back what the initial issue was. Error out early is a
>> > common sense software design principle...
>> >
>> > I am not asking for a panic(), I am just suggesting to only ignore
>> > pinctrl if it returns -ENODEV, the case you care are about.
>>
>> It was just an analogy for enlarging the problem for getting
>> attention.  But you probably thought that it was not enlarging the
>> problem as you think pinctrl is required by the driver instead of an
>> optional thing.
>>
> The I2C bus recovery is a feature which is not used during normal
> operation and you won't find that it's not working unless something
> unusual happens (probably only in a productive environment). It may even
> be hard to trigger a condition to test the feature.
> Thus IMO it is vital that the driver complains loudly if something is
> missing to make this feature work if requested.

I agree.  A message is already added in this patch to mention that bus
recovery is not supported due to pinctrl when there is a problem.

Regards,
Leo

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web