Re: [PATCH] gcc-3.0 warnings

J . A . Magallon (jamagallon@able.es)
Fri, 23 Mar 2001 23:59:09 +0100


On 03.23 Linus Torvalds wrote:
>
> I agree. I'd much prefer that syntax also.
>
> Or just remove the "default:" altogether, when it doesn't make any
> difference.
>

Well, at last some sense. The same is with that ugly out: at the end
of the function. Just change all that 'goto out' for a return.
It does not matter, -O2 is going to do what it wants.

And the missing return 0 at the end of functions that call a 'noreturn'
function. gcc 2.96 still wants them. But it looks like a religious matter
to put ot not to put that stupid return just to shut up the compiler.
As I understand, the noreturn says that the function that is marked as
noreturn is allowed to have missing correct return paths, and the compiler
can build, for example <panic>, without worring about the global state
once it has entered <panic>. But <info gcc> says nothing about functions
that call a 'noreturn' function. So I see as INCORRECT to omit a return path
in a function that calls <panic>.

And if people is so worried about fast paths, begin to use 'const' or
'pure' functions. I think that can help the compiler to generate fast code
more than trying to do hancrafted fast paths that the compiler will reorganize.

-- 
J.A. Magallon                                          #  Let the source
mailto:jamagallon@able.es                              #  be with you, Luke... 

Linux werewolf 2.4.2-ac22 #3 SMP Fri Mar 23 02:06:00 CET 2001 i686

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/