Re: [PATCH v2] iio: core: move 'indio_dev->info' null check first

Re: [PATCH v2] iio: core: move 'indio_dev->info' null check first

[Markus Elfring]

> Doesn't fix anything.

Would the change description be nicer without this sentence?


> Just moves this …

I suggest to choose an imperative wording.
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?id=7e63420847ae5f1036e4f7c42f0b3282e73efbc2#n151

Regards,
Markus

Re: [PATCH v2] iio: core: move 'indio_dev->info' null check first

[Ardelean, Alexandru]

On Tue, 2020-04-07 at 10:20 +0200, Markus Elfring wrote:
> [External]
> 
> > Doesn't fix anything.
> 
> Would the change description be nicer without this sentence?
> 

That part is slightly debate-able.
I've had my fair-share of comments [during reviews], where others were
asking/saying "is this fixing anything? please state if this is fixing
anything".

So, over time I got the habit of mentioning 'Doesn't fix anything.' if the patch
doesn't fix anything. 

So, I'd leave it as-is [if that's possible].

> 
> > Just moves this …
> 
> I suggest to choose an imperative wording.
> https://urldefense.com/v3/__https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?id=7e63420847ae5f1036e4f7c42f0b3282e73efbc2*n151__;Iw!!A3Ni8CS0y2Y!pMoqZILvWyu87fkNe16Sua0jMrZZK81oQOHENUhyTWZz9UrY4CisyKWSRCPFws0WTm50iQ$ 
> 

I admit, I didn't read that document too much in-depth.
And I probably won't, because I am lazy.

Personally, what I try to do, is guess the maintainer's/community's preferences
over time via interactions.
Things differ slightly from community to community.
Of course, there is some common ground across all communities, and I try to
apply that first.
Then, depending on the flexibility/past-comments from other sent-patches I try
to apply those on top [unless I forget; which I kinda do].

> Regards,
> Markus

Re: [v2] iio: core: move 'indio_dev->info' null check first

[Markus Elfring]

> So, over time I got the habit of mentioning 'Doesn't fix anything.'

I suggest to reconsider such information once more.
We come along different interpretations for remarkable software adjustments.


> if the patch doesn't fix anything.

Do you improve the input parameter validation a bit in principle here?

Regards,
Markus

Re: [v2] iio: core: move 'indio_dev->info' null check first

[Ardelean, Alexandru]

On Tue, 2020-04-07 at 12:42 +0200, Markus Elfring wrote:
> [External]
> 
> > So, over time I got the habit of mentioning 'Doesn't fix anything.'
> 
> I suggest to reconsider such information once more.
> We come along different interpretations for remarkable software adjustments.
> 
> 
> > if the patch doesn't fix anything.
> 
> Do you improve the input parameter validation a bit in principle here?

I thought that was obvious from the rest of the description + patch.
I guess it isn't.

I'll see.

> 
> Regards,
> Markus

Re: [v2] iio: core: move 'indio_dev->info' null check first

[Ardelean, Alexandru]

On Tue, 2020-04-07 at 11:01 +0000, Ardelean, Alexandru wrote:
> [External]
> 
> On Tue, 2020-04-07 at 12:42 +0200, Markus Elfring wrote:
> > [External]
> > 
> > > So, over time I got the habit of mentioning 'Doesn't fix anything.'
> > 
> > I suggest to reconsider such information once more.
> > We come along different interpretations for remarkable software adjustments.
> > 
> > 
> > > if the patch doesn't fix anything.
> > 
> > Do you improve the input parameter validation a bit in principle here?
> 
> I thought that was obvious from the rest of the description + patch.
> I guess it isn't.
> 
> I'll see.

So, how about this:
---------------------------------------------
    Doesn't fix anything. Just moves this to be the first check, as it's very
    simple and fails the registration earlier, instead of potentially
    initializing the 'indio_dev->label' and checking for duplicate indexes, and
    then failing with this simple-check.
    
    This is a minor optimization, since '__iio_device_register()' will waste
    fewer validation cycles in case 'indio_dev->info' is NULL.
---------------------------------------------
?    


> 
> > Regards,
> > Markus

Re: [v2] iio: core: move 'indio_dev->info' null check first

[Markus Elfring]

>     Doesn't fix anything.

You would like to change something for a specific reason.
Thus please omit this sentence.


>     Just moves this to be the first check, as it's very …

Wording alternative:
Move a null pointer check to the beginning of this function implementation.


>     This is a minor optimization, since '__iio_device_register()' will waste
>     fewer validation cycles in case 'indio_dev->info' is NULL.

How do you think about to use the text “iio: core: Improve input parameter validation
in __iio_device_register()” as a commit subject?

Regards,
Markus

[PATCH v2] iio: core: move 'indio_dev->info' null check first

[Alexandru Ardelean]

Doesn't fix anything. Just moves this to be the first check, as it's very
simple and fails the registration earlier, instead of potentially
initializing the 'indio_dev->label' and checking for duplicate indexes, and
then failing with this simple-check.

Signed-off-by: Alexandru Ardelean <alexandru.ardelean@analog.com>
---
Changelog v1 -> v2:
* fix typo 'regitration' 'registration'

 drivers/iio/industrialio-core.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/iio/industrialio-core.c b/drivers/iio/industrialio-core.c
index 157d95a24faa..56ff24d7a174 100644
--- a/drivers/iio/industrialio-core.c
+++ b/drivers/iio/industrialio-core.c
@@ -1711,6 +1711,9 @@ int __iio_device_register(struct iio_dev *indio_dev, struct module *this_mod)
 {
 	int ret;
 
+	if (!indio_dev->info)
+		return -EINVAL;
+
 	indio_dev->driver_module = this_mod;
 	/* If the calling driver did not initialize of_node, do it here */
 	if (!indio_dev->dev.of_node && indio_dev->dev.parent)
@@ -1723,9 +1726,6 @@ int __iio_device_register(struct iio_dev *indio_dev, struct module *this_mod)
 	if (ret < 0)
 		return ret;
 
-	if (!indio_dev->info)
-		return -EINVAL;
-
 	/* configure elements for the chrdev */
 	indio_dev->dev.devt = MKDEV(MAJOR(iio_devt), indio_dev->id);
 
-- 
2.17.1