re: Signed vs unsigned comparisons in sys/dev/pcio/if_wm.c

[Paul Goyette <paul%whooppee.com@localhost>]

On Sun, 13 Nov 2016, matthew green wrote:

> > While starting to investigate the possibility of modularizing the
> > if_wm(4) driver, I discovered some issues where signed expressions are
> > being compared to unsigned expressions.  When if_wm.c is being compiled
> > as a built-in driver, these errors are ignored.  However, when build as
> > a loadable module, they are fatal to the build!
> >
> > There are three instances within if_wm.c where this occurs, at source
> > lines 6164, 6688, and 9204 (based on revision 1.443).  The attached
> > diffs make the compiler happy in both cases (kernel and module).
> >
> > Is there any reason why these changes should not be committed?
>
> adding "(int)" casts to appease sign compare warnings considered
> not worth the churn or ugliness or casts.
>
> we shouldn't use a special warning for module builds, and while i
> tend to agree with using more warning messages than fewer, i'm not
> convinced this one is worth the casts and beyond churn required.

Yeah, no worry.

Turns out that there are a couple of rather complicated dependencies 
that would need to be modularized before if_wm.  Particularly, the mii 
stuff (complicated in that it would ideally require an ability to 
auto-load modules for whichever phy(s) are found), and if_ethersubr 
(which is complicated due to #ifdef/#endif maze).

I'm not going to move forward on any of this in the near future...



+------------------+--------------------------+------------------------+
| Paul Goyette     | PGP Key fingerprint:     | E-mail addresses:      |
| (Retired)        | FA29 0E3B 35AF E8AE 6651 | paul at whooppee.com   |
| Kernel Developer | 0786 F758 55DE 53BA 7731 | pgoyette at netbsd.org |
+------------------+--------------------------+------------------------+