On Mon, 2013-08-05 at 15:10 -0700, Andi Kleen wrote:
> From: Andi Kleen <a...@linux.intel.com>
>
> Check for const __devinitdata and non const __devinitconst
>
> People get this regularly wrong and it breaks the LTO builds,
> as it causes a section attribute conflict.
>
> This doesn't catch all mistakes -- spreading over multiple lines,
> getting const pointers wrong, but hopefully the common ones.
>
> Signed-off-by: Andi Kleen <a...@linux.intel.com>
> ---
> scripts/checkpatch.pl | 9 +++++++++
> 1 file changed, 9 insertions(+)
>
> diff --git a/scripts/checkpatch.pl b/scripts/checkpatch.pl
> index 2ee9eb7..5d68d9c 100755
> --- a/scripts/checkpatch.pl
> +++ b/scripts/checkpatch.pl
> @@ -2676,6 +2676,15 @@ sub process {
> }
> }
>
> +# check for __devinitdata with const or const without __devintconst
> +# XXX should scan multiple lines and handle misplaced consts for pointers
> + if ($line =~ /const/ && $line =~ /__(dev)?initdata/) {
> + ERROR("DEVINITCONST", "const init definition must use
> __devinitconst");
> + }
> + if ($line =~ /__(dev)?initconst/ && $line !~ /\Wconst\W/) {
There are no more uses of __devinitconst and
__devinitdata in the tree.
Shouldn't these be "\bconst\b", "\b__initdata\b"
and "\b__initconst\b"
I think also there'll be a few too many false
positives for function arguments.
It seems that every use of __initconst is of
the form
[static] const <type> [name] __initconst so
if ($line =~ /\b__initconst\b/ &&
$line !~
/^\+\s*(?:static\b)?\s*const\s+$Type\s*(?:$Ident)?\s*__initconst\b/) {
etc...
}
should work reasonably well.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
