Re: [PATCH net-next 02/37] rxrpc: Use umin() and umax() rather than min_t()/max_t() where possible

[David Howells]

David Laight <David.Laight@xxxxxxxxxx> wrote:

> > Use umin() and umax() rather than min_t()/max_t() where the type specified
> > is an unsigned type.
> 
> You are also changing some max() to umax().

Good point.  If I have to respin my patches again, I'll update that.

> Presumably they have always passed the type check so max() is fine.
> And max(foo, 1) would have required that 'foo' be 'signed int' and could
> potentially be negative when max(-1, 1) will be 1 but umax(-1, 1) is
> undefined.

There have been cases like this:

	unsigned long timeout;
	...
	timeout = max(timeout, 1);

where the macro would complain because it thought "timeout" and "1" were
different sizes, so "1UL" had to be used.  Using umax() deals with that issue.

David