Re: [PATCH] Fix uninitialized variable warning in do_one_initcall

[Andrew Morton]

On Sat,  3 Jul 2010 20:26:28 -0300
Kevin Winchester <kjwinchester@xxxxxxxxx> wrote:

> The warning is corrected by extracting the debug path out into
> its own function.  This does en up duplicating one line of code
> between the debug and regular code paths (i.e. the actual call
> of the initcall function), but it seems worthwhile for the
> cleaner build.
> 

I assume the warning was for `calltime'?  Maybe other things?  Please,
remove all doubt and always quote the compiler output in the changelog.

Also please mention the compiler version - it looks like this is a new
warning.  It's not a false positive either - the compiler doesn't know
that initcall_debug's value never changes.

The patch doesn't apply to linux-next because someone went on a great
tromp through do_one_initcall() so could you please redo the patch
against linux-next?

I suggest that you not inline do_one_initcall_debug() - the compiler
will do that anwyay.

And if you're feeling keen, please do a separate patch which marks
do_one_initcall() and do_one_initcall_debug() with __init_or_module -
we don't need to leave that code in memory after bootup if
CONFIG_MODULES=n.

Thanks.

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