[coreboot] typo in amdk8/raminit_f.c

ron minnich rminnich at gmail.com
Mon Apr 13 18:25:21 CEST 2009

On Mon, Apr 13, 2009 at 9:15 AM, Peter Stuge <peter at stuge.se> wrote:
> ron minnich wrote:
>> I think this is one reason the Bell Labs crowd started using:
>> if (preprocessor_symbol == value)
>> That's the C compilers job, after all ...
>> The more code I see, the more I agree with them.
> I can see the point. But I also think excluding code from compilation
> is a big feature, especially in firmware.

ah yes, but ... just about any reasonable compiler will do something
smart with this: if (0) { ...} or if (1) { ... }.

If fact, romcc would not work at all without these types of
optimizations ... and this one is not even that aggressive.

So, since:
if (preprocessor_symbol == whatever) { ... }

is equivalent to either
if (0) {...}
if (1) {...}

I think it is safe to say there is no loss in code reduction.

> If we can do it in a smarter way, that could be good, but on the
> other hand things can be too smart as well.

What Pike et. al are really saying is "don't have cpp do something
that the C compiler can do better and with more error checking".

With which I agree.

The reason I care? Because we've had dead code for several years in
that fragment Ronald discovered, and that could have been avoided.


More information about the coreboot mailing list