On Fri May 2, 2025 at 4:09 PM CEST, Andy Shevchenko wrote:

On Fri, May 02, 2025 at 03:58:04PM +0200, Mathieu Dubois-Briand wrote:

quoted

On Fri May 2, 2025 at 12:52 PM CEST, Andy Shevchenko wrote:

quoted

On Mon, Apr 28, 2025 at 01:57:28PM +0200, Mathieu Dubois-Briand wrote:

...

quoted

quoted

quoted

+				pos = max(0, (int)pos + steps);

Please, no castings for min()/max()/clamp(). It diminishes the use of those
macros.

Sorry, I'm not sure to get the point. Should I use MIN_T() instead?

Are the second argument is compile-time constant? I don't think so. Hence no
use for MIN*()/MAX*(). First of all, try to answer to the Q: Why is the explicit
casting being used? The second Q: How can it be easily fixed without using _t()
variants of the macros?

Err right, no MIN/MAX of course.

Explicit cast is here because the unsigned int + int sum would result in
an unsigned int.

I will use an intermediate signed value. Also the whole logic here can
be simplified a bit, so I will rework the whole block.

Best regards,
Mathieu

-- 
Mathieu Dubois-Briand, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com