linux-kernel - Re: [PATCH 11/28] block: rdb: false-postive gcc-4.9 -Wmaybe-uninitialized

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:   Tue, 18 Oct 2016 11:57:33 +0200
From:   Ilya Dryomov <idryomov@...il.com>
To:     Arnd Bergmann <arnd@...db.de>
Cc:     Sage Weil <sage@...hat.com>,
        Linus Torvalds <torvalds@...ux-foundation.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        Alex Elder <elder@...nel.org>,
        Mike Christie <mchristi@...hat.com>,
        Ceph Development <ceph-devel@...r.kernel.org>
Subject: Re: [PATCH 11/28] block: rdb: false-postive gcc-4.9 -Wmaybe-uninitialized

On Tue, Oct 18, 2016 at 12:10 AM, Arnd Bergmann <arnd@...db.de> wrote:
> When building with gcc-4.9 -Wmaybe-uninitialized, we get a bogus
> warning in rbd_watch_cb, as the variable is not used at all
> in the one case in which it is not initialized first:
>
> drivers/block/rbd.c: In function ‘rbd_watch_cb’:
> drivers/block/rbd.c:3690:5: error: ‘struct_v’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
> drivers/block/rbd.c:3759:5: note: ‘struct_v’ was declared here
>
> Later compiler versions fix this, but adding another initialization
> here is harmless and lets us build cleanly with 4.9 as well.
>
> Signed-off-by: Arnd Bergmann <arnd@...db.de>
> ---
>  drivers/block/rbd.c | 1 +
>  1 file changed, 1 insertion(+)
>
> diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c
> index abb7162..4ab990b 100644
> --- a/drivers/block/rbd.c
> +++ b/drivers/block/rbd.c
> @@ -3776,6 +3776,7 @@ static void rbd_watch_cb(void *arg, u64 notify_id, u64 cookie,
>         } else {
>                 /* legacy notification for header updates */
>                 notify_op = RBD_NOTIFY_OP_HEADER_UPDATE;
> +               struct_v = 0;
>                 len = 0;
>         }

It already got silenced by initializing at declaration in one of the
downstream trees, so I'd rather we do

@@ -3756,7 +3819,7 @@ static void rbd_watch_cb(void *arg, u64
notify_id, u64 cookie,
        struct rbd_device *rbd_dev = arg;
        void *p = data;
        void *const end = p + data_len;
-       u8 struct_v;
+       u8 struct_v = 0;
        u32 len;
        u32 notify_op;
        int ret;

to reduce the churn.

The "block" prefix is redundant and "rdb" should be "rbd" in the subject.

Thanks,

                Ilya