linux-kernel - Re: [PATCH] rbd: avoid clang -Wuninitialized warning

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 22 Mar 2019 17:34:26 +0100
From:   Ilya Dryomov <idryomov@...il.com>
To:     Arnd Bergmann <arnd@...db.de>
Cc:     Sage Weil <sage@...hat.com>, Alex Elder <elder@...nel.org>,
        Jens Axboe <axboe@...nel.dk>,
        clang-built-linux@...glegroups.com,
        Nick Desaulniers <ndesaulniers@...gle.com>,
        Nathan Chancellor <natechancellor@...il.com>,
        Jason Dillaman <dillaman@...hat.com>,
        Ceph Development <ceph-devel@...r.kernel.org>,
        linux-block <linux-block@...r.kernel.org>,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] rbd: avoid clang -Wuninitialized warning

On Fri, Mar 22, 2019 at 3:36 PM Arnd Bergmann <arnd@...db.de> wrote:
>
> clang fails to see that rbd_assert(0) ends in an unreachable code
> path and warns about a subsequent use of an uninitialized variable
> when CONFIG_PROFILE_ANNOTATED_BRANCHES is set:
>
> drivers/block/rbd.c:2402:4: error: variable 'ret' is used uninitialized whenever 'if' condition is false
>       [-Werror,-Wsometimes-uninitialized]
>                         rbd_assert(0);
>                         ^~~~~~~~~~~~~
> drivers/block/rbd.c:563:7: note: expanded from macro 'rbd_assert'
>                 if (unlikely(!(expr))) {                                \
>                     ^~~~~~~~~~~~~~~~~
> include/linux/compiler.h:48:23: note: expanded from macro 'unlikely'
>  #  define unlikely(x)   (__branch_check__(x, 0, __builtin_constant_p(x)))
>                         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> drivers/block/rbd.c:2410:6: note: uninitialized use occurs here
>         if (ret) {
>             ^~~
> drivers/block/rbd.c:2402:4: note: remove the 'if' if its condition is always true
>                         rbd_assert(0);
>                         ^
> drivers/block/rbd.c:563:3: note: expanded from macro 'rbd_assert'
>                 if (unlikely(!(expr))) {                                \
>                 ^
> drivers/block/rbd.c:2376:9: note: initialize the variable 'ret' to silence this warning
>         int ret;
>                ^
>                 = 0
> 1 error generated.
>
> This seems to be a bug in clang, but is easy to work around by using
> an unconditional BUG().
>
> Signed-off-by: Arnd Bergmann <arnd@...db.de>
> ---
>  drivers/block/rbd.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c
> index 4ba967d65cf9..cbcc3baf3807 100644
> --- a/drivers/block/rbd.c
> +++ b/drivers/block/rbd.c
> @@ -2399,7 +2399,7 @@ static int rbd_obj_read_from_parent(struct rbd_obj_request *obj_req)
>                                                       &obj_req->bvec_pos);
>                         break;
>                 default:
> -                       rbd_assert(0);
> +                       BUG();
>                 }
>         } else {
>                 ret = rbd_img_fill_from_bvecs(child_img_req,

Hi Arnd,

You did a couple of these last year in commit c6244b3b2377 ("rbd: avoid
Wreturn-type warnings").  Let's change all of those default cases to BUG
in one go.  Do you want to do that or should I?

Thanks,

                Ilya