Re: [Qemu-devel] [PULL 06/18] COLO: Introduce checkpointing protocol

From: Hailiang Zhang <zhang.zhanghailiang@huawei.com>
To: Eduardo Habkost <ehabkost@redhat.com>, Amit Shah <amit.shah@redhat.com>
Cc: Peter Maydell <peter.maydell@linaro.org>,
	qemu list <qemu-devel@nongnu.org>,
	"Dr. David Alan Gilbert" <dgilbert@redhat.com>,
	Juan Quintela <quintela@redhat.com>
Subject: Re: [Qemu-devel] [PULL 06/18] COLO: Introduce checkpointing protocol
Date: Tue, 1 Nov 2016 09:48:09 +0800	[thread overview]
Message-ID: <5817F459.2060702@huawei.com> (raw)
In-Reply-To: <20161031182507.GB25480@thinpad.lan.raisama.net>

On 2016/11/1 2:25, Eduardo Habkost wrote:
> On Sun, Oct 30, 2016 at 04:16:58PM +0530, Amit Shah wrote:
> [...]
>> +static void colo_wait_handle_message(QEMUFile *f, int *checkpoint_request,
>> +                                     Error **errp)
>> +{
>> +    COLOMessage msg;
>> +    Error *local_err = NULL;
>> +
>> +    msg = colo_receive_message(f, &local_err);
>> +    if (local_err) {
>> +        error_propagate(errp, local_err);
>> +        return;
>> +    }
>> +
>> +    switch (msg) {
>> +    case COLO_MESSAGE_CHECKPOINT_REQUEST:
>> +        *checkpoint_request = 1;
>> +        break;
>> +    default:
>> +        *checkpoint_request = 0;
>> +        error_setg(errp, "Got unknown COLO message: %d", msg);
>> +        break;
>> +    }
>> +}
> [...]
>> +        colo_wait_handle_message(mis->from_src_file, &request, &local_err);
>> +        if (local_err) {
>> +            goto out;
>> +        }
>> +        assert(request);
>
> GCC 4.8.5 doesn't seem to be smart enough to notice that request
> will be always initialized:
>
>    /root/qemu/migration/colo.c: In function ‘colo_process_incoming_thread’:
>    /root/qemu/migration/colo.c:448:33: error: ‘request’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
>             assert(request);
>                                     ^
>    cc1: all warnings being treated as errors
>
>    $ gcc --version
>    gcc (GCC) 4.8.5 20150623 (Red Hat 4.8.5-4)
>

Yes, you are right, Jeff has sent a patch to fix this
'[PATCH 1/1] migration: fix compiler warning on uninitialized variable'

Thanks,
Hailiang