On 23 June 2014 13:49, Markus Armbruster <arm...@redhat.com> wrote:
> Peter Maydell <peter.mayd...@linaro.org> writes:
>
>> Convert the socket char backend to the new style QAPI framework;
>> this allows it to return an Error ** to callers who might not
>> want it to print directly about socket failures.
>>
>> Signed-off-by: Peter Maydell <peter.mayd...@linaro.org>
>> ---
>> I'm not 100% sure I have this correct -- review from somebody
>> who understands the char backends appreciated!
>
> Not sure anybody really understands this mess, but I can try.
Thanks.
>> + addr = g_new0(SocketAddress, 1);
>> + if (path) {
>> + addr->kind = SOCKET_ADDRESS_KIND_UNIX;
>
> Missing:
>
> addr->q_unix = g_new(UnixSocketAddress, 1);
>
> See also socket_parse().
Heh, yes. You can tell I didn't test this code path ;-)
>> + addr->q_unix->path = g_strdup(path);
>> + } else {
>> + addr->kind = SOCKET_ADDRESS_KIND_INET;
>> + addr->inet = g_new0(InetSocketAddress, 1);
>> + addr->inet->host = g_strdup(host);
>> + addr->inet->port = g_strdup(port);
>> + addr->inet->to = qemu_opt_get_number(opts, "to", 0);
>> + addr->inet->has_to = true;
>> + if (qemu_opt_get_bool(opts, "ipv4", 0)) {
>> + addr->inet->has_ipv4 = addr->inet->ipv4 = true;
>> + }
>> + if (qemu_opt_get_bool(opts, "ipv6", 0)) {
>> + addr->inet->has_ipv6 = addr->inet->ipv6 = true;
>> + }
>> + }
>> + backend->socket->addr = addr;
>> +}
>> +
>> typedef struct CharDriver {
>> const char *name;
>> /* old, pre qapi */
>> @@ -4064,7 +4064,8 @@ void qmp_chardev_remove(const char *id, Error **errp)
>> static void register_types(void)
>> {
>> register_char_driver_qapi("null", CHARDEV_BACKEND_KIND_NULL, NULL);
>> - register_char_driver("socket", qemu_chr_open_socket);
>> + register_char_driver_qapi("socket", CHARDEV_BACKEND_KIND_SOCKET,
>> + qemu_chr_parse_socket);
>> register_char_driver("udp", qemu_chr_open_udp);
>> register_char_driver_qapi("ringbuf", CHARDEV_BACKEND_KIND_RINGBUF,
>> qemu_chr_parse_ringbuf);
>
> Only one old-style driver left: "udp". It's closely related to
> "socket"... would you be willing to take care of that one, too?
Yeah, I noticed that too; I figured I'd get this patch
through review first, though.
thanks
-- PMM
