The idea is there are two type of errors:
- those returned by the system: i.e. the result code that you pointed below
which gets interpreted by
(https://github.com/openvswitch/ovs/blob/master/lib/netlink-socket.c#L567).
Which means that was a system error.
- netlink errors which will determine the userspace to make appropriate
decisions based on that error.
The rc is set to success so that we tell the userspace that there was no system
error but there was a netlink error.
Alin.
> done:
>+ if (nlError != NL_ERROR_SUCCESS) {
>+ POVS_MESSAGE_ERROR msgError = (POVS_MESSAGE_ERROR)
>+ usrParamsCtx->outputBuffer;
>+ NlBuildErrorMsg(msgIn, msgError, nlError);
>+ *replyLen = msgError->nlMsg.nlmsgLen;
Sai: Should rc be set to STATUS_SUCCESS? Isn¹t this a failure condition that
will affect the calling function?
>+ rc = STATUS_SUCCESS;
>+ }
>+
> return rc;
> }
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
