On 28 Aug 2024, at 2:40, Ilya Maximets wrote:
> On 8/20/24 15:55, Mike Pattrick wrote:
>> The Clang analyzer will alert on the use of uninitialized variable local
>> despite the fact that this should be set by a syscall.
>>
>> To suppress the warning, this variable is now initialized.
>>
>> Signed-off-by: Mike Pattrick <[email protected]>
>> ---
>> lib/netlink-socket.c | 1 +
>> 1 file changed, 1 insertion(+)
>>
>> diff --git a/lib/netlink-socket.c b/lib/netlink-socket.c
>> index 5cb1fc89a..737e49cfc 100644
>> --- a/lib/netlink-socket.c
>> +++ b/lib/netlink-socket.c
>> @@ -233,6 +233,7 @@ nl_sock_create(int protocol, struct nl_sock **sockp)
>>
>> /* Obtain pid assigned by kernel. */
>> local_size = sizeof local;
>> + local.nl_family = AF_UNSPEC;
>
> It feels strange to do that. Just memset-ing the whole structure to 0 might
> be a better choice.
As Simon mentioned in the v1 patch, this is more a problem of clang than our
code, and wondering if we want to fix the code around it.
I just waved a couple of similar issues in Coverity.
However, if we do want to fix this, I would prefer the memset() as Ilya
suggested.
I have not looked at clang-analyze for a while, but maybe there finally is a
way to exclude results using an external file (rather than augment this inline).
>> if (getsockname(sock->fd, (struct sockaddr *) &local, &local_size) < 0)
>> {
>> VLOG_ERR("getsockname: %s", ovs_strerror(errno));
>> goto error;
>
> _______________________________________________
> dev mailing list
> [email protected]
> https://mail.openvswitch.org/mailman/listinfo/ovs-dev
_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
