Re: linux-next: build failure after merge of the akpm-current tree

From	Paul Bolle <pebolle@tiscali.nl>
Newsgroups	linux.kernel
Subject	Re: linux-next: build failure after merge of the akpm-current tree
Date	2016-06-15 23:20 +0200
Message-ID	<rKqoa-7Kt-23@gated-at.bofh.it> (permalink)
References	<rvG1Q-6VA-3@gated-at.bofh.it> <rvGOd-7Es-1@gated-at.bofh.it> <rKfsK-Lj-19@gated-at.bofh.it> <rKjG1-3xJ-3@gated-at.bofh.it>
Organization	linux.* mail to news gateway

Show all headers | View raw

Hi Josh,

On wo, 2016-06-15 at 09:03 -0500, Josh Poimboeuf wrote:
> I agree it looks like a false positive, though the code is a bit
> convoluted, so I'm not surprised that gcc might get confused.

Agree.

> How about initializing opt to 0?

I'm unconvinced. Especially since this only shows up in 64 bits builds.

> --- a/net/netfilter/ipvs/ip_vs_sync.c
> +++ b/net/netfilter/ipvs/ip_vs_sync.c

>  static inline int ip_vs_proc_sync_conn(struct netns_ipvs *ipvs, __u8

> -	struct ip_vs_sync_conn_options opt;
> +	struct ip_vs_sync_conn_options opt = {0};

You forgot to append
                                                  /* I give up */

here. More common would be to use
                                                  /* silence GCC */
Thanks,


Paul Bolle

Back to linux.kernel | Previous | Next — Previous in thread | Find similar | Unroll thread

Thread

Re: linux-next: build failure after merge of the akpm-current tree Paul Bolle <pebolle@tiscali.nl> - 2016-06-15 11:40 +0200
  Re: linux-next: build failure after merge of the akpm-current tree Josh Poimboeuf <jpoimboe@redhat.com> - 2016-06-15 16:10 +0200
    Re: linux-next: build failure after merge of the akpm-current tree Paul Bolle <pebolle@tiscali.nl> - 2016-06-15 23:20 +0200

csiph-web