Re: [PATCH 7/7] Fix uninitialised variable warning inmei-amt-version.c

[Paul Bolle]

On Fri, 2014-01-03 at 16:08 +0000, David Howells wrote:
> Fix the following warning:
> 
> 	Documentation/misc-devices/mei/mei-amt-version.c: In function 'main':
> 	Documentation/misc-devices/mei/mei-amt-version.c:103:5: warning: 'acmd.fd' is used uninitialized in this function [-Wuninitialized]
> 	  if (cl->fd != -1)
> 	     ^
> 	Documentation/misc-devices/mei/mei-amt-version.c:443:21: note: 'acmd.fd' was declared here
> 	  struct amt_host_if acmd;
> 			     ^
> 
> The call chain:
> 
> 	-->main()
> 	   -->amt_host_if_init()
> 	      -->mei_init()
> 	         -->mei_deinit()
> 
> results in this on the first time through because main()::acmd has not yet
> been initialised.
> 
> To fix this, acmd.fd needs to be initialised to -1.
> 
> Signed-off-by: David Howells <dhowells@xxxxxxxxxx>
> cc: Tomas Winkler <tomas.winkler@xxxxxxxxx>
> cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>

I've been using a local patch now for quite some time. I currently use
it on v3.12.6. I haven't yet bothered to submit it. My (draft) commit
summary reads:

"Building mei-amt-version.o triggers a GCC warning:
    Documentation/misc-devices/mei/mei-amt-version.c: In function 'main':
    Documentation/misc-devices/mei/mei-amt-version.c:103:5: warning: 'acmd.fd' is used uninitialized in this function [-Wuninitialized]
      if (cl->fd != -1)
         ^
    Documentation/misc-devices/mei/mei-amt-version.c:443:21: note: 'acmd.fd' was declared here
      struct amt_host_if acmd;
                         ^

GCC is correct. See, the call chain that GCC detects must be
    main()
        amt_host_if_init()
            mei_init()
                mei_deinit()

But when we enter mei_deinit() struct amt_host_if acmd is still
unitialized. That makes the test for (effectively) amt_host_if->mei_cl->fd
bogus.

But it turns out that call of mei_deinit() isn't needed at all. All of
the members of mei_cl will be set later in mei_init() and none will be
used before they are set. So we can simply drop this call of
mei_deinit()."

Is that analysis (still) correct?


Paul Bolle

>  Documentation/misc-devices/mei/mei-amt-version.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/Documentation/misc-devices/mei/mei-amt-version.c b/Documentation/misc-devices/mei/mei-amt-version.c
> index 49e4f770864a..7027a8da0ba0 100644
> --- a/Documentation/misc-devices/mei/mei-amt-version.c
> +++ b/Documentation/misc-devices/mei/mei-amt-version.c
> @@ -440,7 +440,7 @@ out:
>  int main(int argc, char **argv)
>  {
>  	struct amt_code_versions ver;
> -	struct amt_host_if acmd;
> +	struct amt_host_if acmd = { .mei_cl.fd = -1 };
>  	unsigned int i;
>  	uint32_t status;
>  	int ret;
> 


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/