Re: [PATCH] iio: imu: adis16550: rework clock range test
From: David Lechner
Date: Wed Jul 02 2025 - 11:08:06 EST
On 7/2/25 9:59 AM, Andy Shevchenko wrote:
> On Wed, Jul 02, 2025 at 05:53:57PM +0300, Andy Shevchenko wrote:
>> On Wed, Jul 02, 2025 at 09:27:45AM -0500, David Lechner wrote:
>>> Rework the clock rate range test to test if sync_mode_data != NULL
>>> instead of testing if the for loop index variable. This makes it easier
>>> for static analyzers to see that we aren't using an uninitialized
>>> sync_mode_data [1].
>>
>> But at the same time it makes it not to be the usual pattern.,,
>
> Reading the static analyser output I think the first hunk is only what we need,
> but this is still false positive and it's problem of that static
> analyser. Have you filed a bug there? (My point is that modifying the code for
> the advantage of false positives of some static analyser is wrong road to go
> in my opinion.)
>
I agree that we shouldn't fix this _only_ to make the static analyzer
happy. But I had to think quite a bit harder to see that the existing
code was correct compared to what I have proposed here.
But if this is a common pattern that I just haven't learned to identify
at a glance yet and everybody else can easily see that the existing code
is correct, then perhaps it isn't worth the change.