Message ID | 20200409151125.32677-1-aishwaryarj100@gmail.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | iio: adc: at91-adc: Use devm_platform_ioremap_resource | expand |
On 09/04/2020 20:41:23+0530, Aishwarya R wrote: > Use the helper function that wraps the calls to > platform_get_resource() and devm_ioremap_resource() > together. > Please elaborate the actual value of doing that. > Signed-off-by: Aishwarya R <aishwaryarj100@gmail.com> > --- > drivers/iio/adc/at91_adc.c | 5 +---- > 1 file changed, 1 insertion(+), 4 deletions(-) > > diff --git a/drivers/iio/adc/at91_adc.c b/drivers/iio/adc/at91_adc.c > index abe99856c823..0368b6dc6d60 100644 > --- a/drivers/iio/adc/at91_adc.c > +++ b/drivers/iio/adc/at91_adc.c > @@ -1152,7 +1152,6 @@ static int at91_adc_probe(struct platform_device *pdev) > int ret; > struct iio_dev *idev; > struct at91_adc_state *st; > - struct resource *res; > u32 reg; > > idev = devm_iio_device_alloc(&pdev->dev, sizeof(struct at91_adc_state)); > @@ -1182,9 +1181,7 @@ static int at91_adc_probe(struct platform_device *pdev) > if (st->irq < 0) > return -ENODEV; > > - res = platform_get_resource(pdev, IORESOURCE_MEM, 0); > - > - st->reg_base = devm_ioremap_resource(&pdev->dev, res); > + st->reg_base = devm_platform_ioremap_resource(pdev, 0); > if (IS_ERR(st->reg_base)) > return PTR_ERR(st->reg_base); > > -- > 2.17.1 >
Use helper function to simply the code. The devm_platform_ioremap_resource() helper in the kernel which helps to make code cleaner a tad by replacing few "boilerplate" lines of code with a single line.
On Thu, Apr 9, 2020 at 7:00 PM Alexandre Belloni <alexandre.belloni@bootlin.com> wrote: > On 09/04/2020 20:41:23+0530, Aishwarya R wrote: > > Use the helper function that wraps the calls to > > platform_get_resource() and devm_ioremap_resource() > > together. > Please elaborate the actual value of doing that. Please, elaborate actual value of not doing that. Yes, I know that you are p* off of these changes, but why you not going further and forbid all clean ups we are doing in the code? To the point. Above change is reducing code base and showing the new comers modern APIs to use.
On 10/04/2020 13:55:26+0300, Andy Shevchenko wrote: > On Thu, Apr 9, 2020 at 7:00 PM Alexandre Belloni > <alexandre.belloni@bootlin.com> wrote: > > On 09/04/2020 20:41:23+0530, Aishwarya R wrote: > > > Use the helper function that wraps the calls to > > > platform_get_resource() and devm_ioremap_resource() > > > together. > > > Please elaborate the actual value of doing that. > > Please, elaborate actual value of not doing that. > > Yes, I know that you are p* off of these changes, but why you not > going further and forbid all clean ups we are doing in the code? > > To the point. Above change is reducing code base and showing the new > comers modern APIs to use. > The value of doing it is to reduce the code size by 16 bytes. The same people doing that will actively ruin that by adding error string for error that will never ever happen. Also, the commit message is definitively lacking. A good commit message would say that the patch has been generated using coccinelle, that no testing has been done and that no thought has been given. It would definitively make sense to send one patch per subsystem instead of having 475 different patches each changing only one location. The whole "let's let newcomers fix trivial bugs" thing is definitively not working and it is not leading to an increase of the number of useful reviewers and contributors
On Fri, Apr 10, 2020 at 2:22 PM Alexandre Belloni <alexandre.belloni@bootlin.com> wrote: > On 10/04/2020 13:55:26+0300, Andy Shevchenko wrote: > > On Thu, Apr 9, 2020 at 7:00 PM Alexandre Belloni > > <alexandre.belloni@bootlin.com> wrote: > > > On 09/04/2020 20:41:23+0530, Aishwarya R wrote: > > > > Use the helper function that wraps the calls to > > > > platform_get_resource() and devm_ioremap_resource() > > > > together. > > > > > Please elaborate the actual value of doing that. > > > > Please, elaborate actual value of not doing that. > > > > Yes, I know that you are p* off of these changes, but why you not > > going further and forbid all clean ups we are doing in the code? > > > > To the point. Above change is reducing code base and showing the new > > comers modern APIs to use. > The value of doing it is to reduce the code size by 16 bytes. The same > people doing that will actively ruin that by adding error string for > error that will never ever happen. I don't see it in the patch we are discussing, so, not an argument. > Also, the commit message is definitively lacking. A good commit message > would say that the patch has been generated using coccinelle, that no > testing has been done and that no thought has been given. That's I agree with. But somebody need to teach people how to do it better. > It would definitively make sense to send one patch per subsystem instead > of having 475 different patches each changing only one location. Depends on the maintainer and subsystem. This is arguable argument. In my subsystems (let's forget about PDx86, where one per subsystem in principle is not working by nature of the subsystem, but consider others I'm maintaining) I prefer to have a possibility to track changes independently on driver basis. > The whole "let's let newcomers fix trivial bugs" thing is definitively > not working and it is not leading to an increase of the number of useful > reviewers and contributors Semi-agree. That people can be self-organized into a reviewer gang and try to learn together. In any case, as for maintainers, the task has not only technical aspects, but mentoring as well.
On Thu, 9 Apr 2020 20:41:23 +0530 Aishwarya R <aishwaryarj100@gmail.com> wrote: > Use the helper function that wraps the calls to > platform_get_resource() and devm_ioremap_resource() > together. > > Signed-off-by: Aishwarya R <aishwaryarj100@gmail.com> Hi Aishwarya R, Thanks for the patch. Good to add a little more detail to the description and note that I do prefer one patch per driver for IIO drivers. One other thing. Please use full name for the signed-off-by. I see you use a longer form in your email later in the thread. The DCO is a formal legal statement https://www.kernel.org/doc/html/v4.17/process/submitting-patches.html#sign-your-work-the-developer-s-certificate-of-origin so we tend to be fussy! Thanks, Jonathan > --- > drivers/iio/adc/at91_adc.c | 5 +---- > 1 file changed, 1 insertion(+), 4 deletions(-) > > diff --git a/drivers/iio/adc/at91_adc.c b/drivers/iio/adc/at91_adc.c > index abe99856c823..0368b6dc6d60 100644 > --- a/drivers/iio/adc/at91_adc.c > +++ b/drivers/iio/adc/at91_adc.c > @@ -1152,7 +1152,6 @@ static int at91_adc_probe(struct platform_device *pdev) > int ret; > struct iio_dev *idev; > struct at91_adc_state *st; > - struct resource *res; > u32 reg; > > idev = devm_iio_device_alloc(&pdev->dev, sizeof(struct at91_adc_state)); > @@ -1182,9 +1181,7 @@ static int at91_adc_probe(struct platform_device *pdev) > if (st->irq < 0) > return -ENODEV; > > - res = platform_get_resource(pdev, IORESOURCE_MEM, 0); > - > - st->reg_base = devm_ioremap_resource(&pdev->dev, res); > + st->reg_base = devm_platform_ioremap_resource(pdev, 0); > if (IS_ERR(st->reg_base)) > return PTR_ERR(st->reg_base); >
On Fri, 10 Apr 2020 19:37:25 +0300 Andy Shevchenko <andy.shevchenko@gmail.com> wrote: > On Fri, Apr 10, 2020 at 2:22 PM Alexandre Belloni > <alexandre.belloni@bootlin.com> wrote: > > On 10/04/2020 13:55:26+0300, Andy Shevchenko wrote: > > > On Thu, Apr 9, 2020 at 7:00 PM Alexandre Belloni > > > <alexandre.belloni@bootlin.com> wrote: > > > > On 09/04/2020 20:41:23+0530, Aishwarya R wrote: > > > > > Use the helper function that wraps the calls to > > > > > platform_get_resource() and devm_ioremap_resource() > > > > > together. > > > > > > > Please elaborate the actual value of doing that. > > > > > > Please, elaborate actual value of not doing that. > > > > > > Yes, I know that you are p* off of these changes, but why you not > > > going further and forbid all clean ups we are doing in the code? > > > > > > To the point. Above change is reducing code base and showing the new > > > comers modern APIs to use. > > > The value of doing it is to reduce the code size by 16 bytes. The same > > people doing that will actively ruin that by adding error string for > > error that will never ever happen. > > I don't see it in the patch we are discussing, so, not an argument. > > > Also, the commit message is definitively lacking. A good commit message > > would say that the patch has been generated using coccinelle, that no > > testing has been done and that no thought has been given. > > That's I agree with. > But somebody need to teach people how to do it better. > > > It would definitively make sense to send one patch per subsystem instead > > of having 475 different patches each changing only one location. > > Depends on the maintainer and subsystem. This is arguable argument. > In my subsystems (let's forget about PDx86, where one per subsystem in > principle is not working by nature of the subsystem, but consider > others I'm maintaining) I prefer to have a possibility to track > changes independently on driver basis. Likewise. I often request patches are broken out. > > > The whole "let's let newcomers fix trivial bugs" thing is definitively > > not working and it is not leading to an increase of the number of useful > > reviewers and contributors > > Semi-agree. That people can be self-organized into a reviewer gang and > try to learn together. > In any case, as for maintainers, the task has not only technical > aspects, but mentoring as well. > It would be interesting to see if we do have examples of people who have started with cleanup patches and moved on to more interesting things. Some of the outreach programs that start with simple patches as a means of getting familiar with the tools have led to more significant improvements but those are somewhat of a special case. I'd rather have hundreds of people send me these sorts of patches a year than loose one potential long term contributor. Jonathan p.s. Aishwarya - You've been a bit unlucky in a sense here running into the latest instance of a very long term debate!
diff --git a/drivers/iio/adc/at91_adc.c b/drivers/iio/adc/at91_adc.c index abe99856c823..0368b6dc6d60 100644 --- a/drivers/iio/adc/at91_adc.c +++ b/drivers/iio/adc/at91_adc.c @@ -1152,7 +1152,6 @@ static int at91_adc_probe(struct platform_device *pdev) int ret; struct iio_dev *idev; struct at91_adc_state *st; - struct resource *res; u32 reg; idev = devm_iio_device_alloc(&pdev->dev, sizeof(struct at91_adc_state)); @@ -1182,9 +1181,7 @@ static int at91_adc_probe(struct platform_device *pdev) if (st->irq < 0) return -ENODEV; - res = platform_get_resource(pdev, IORESOURCE_MEM, 0); - - st->reg_base = devm_ioremap_resource(&pdev->dev, res); + st->reg_base = devm_platform_ioremap_resource(pdev, 0); if (IS_ERR(st->reg_base)) return PTR_ERR(st->reg_base);
Use the helper function that wraps the calls to platform_get_resource() and devm_ioremap_resource() together. Signed-off-by: Aishwarya R <aishwaryarj100@gmail.com> --- drivers/iio/adc/at91_adc.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-)