On Sun, Mar 30, 2014 at 11:38 AM, Linus Torvalds
<torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
> The reason seems to be that our logic in check_declaration[] is
> broken: it only ties things together based on both symbols having
> 'extern'. So the actual declaration with an initializer, that lacks
> the extern (but is in global scope) is missed.
>
> The attached patch would seem to fix it. Chris, what is the reason for
> that MOD_INLINE check? You added it in commit 8cf99394ee4c ("move
> extern inline function to file scope") and I suspect it messes up the
> same_symbol logic. But I left it alone, and added a comment. If you
Possible. If I comment out that check, I will get this error in the
extern-inline.c
test case:
$ ./sparse validation/extern-inline.c validation/extern-inline.c
validation/extern-inline.c:12:1: warning: multiple definitions for function 'g'
validation/extern-inline.c:12:1: the previous one is here
$
Unfortunately I have to run now, I will take a closer look when I am back
tonight. I verify that applying your second patch did not trigger the error
in extern-inline.c.
I think it is possible the fix for the extern-inline.c is wrong regarding the
same_symbol chain, there should be better way to fix it. I need to spend
more time on it though.
> can resolve that comment, please apply this patch with my sign-off and
> some random commit message..
Sure
Chris
--
To unsubscribe from this list: send the line "unsubscribe linux-sparse" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
