Re: [PATCH 1/4] net-dhcp-leases: Implement the public APIs

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 09/11/2013 01:14 PM, Daniel P. Berrange wrote:
> On Wed, Sep 11, 2013 at 10:35:04PM +0530, Nehal J Wani wrote:
>> On Wed, Sep 11, 2013 at 9:25 PM, Daniel P. Berrange <berrange@xxxxxxxxxx> wrote:
>>> On Wed, Sep 11, 2013 at 09:00:09PM +0530, Nehal J Wani wrote:
>>>> diff --git a/include/libvirt/libvirt.h.in b/include/libvirt/libvirt.h.in
>>>> index a47e33c..3885baf 100644
>>>> --- a/include/libvirt/libvirt.h.in
>>>> +++ b/include/libvirt/libvirt.h.in
>>>> @@ -2800,6 +2800,28 @@ int                     virConnectNumOfDefinedInterfaces (virConnectPtr conn);
>>>>  int                     virConnectListDefinedInterfaces  (virConnectPtr conn,
>>>>                                                            char **const names,
>>>>                                                            int maxnames);
>>>> +
>>>> +typedef struct _virNetworkDHCPLeases virNetworkDHCPLeases;
>>>> +typedef virNetworkDHCPLeases *virNetworkDHCPLeasesPtr;
>>>> +struct _virNetworkDHCPLeases {
>>>> +    long long expirytime;
>>>> +    char *mac;
>>>> +    char *ipaddr;
>>>> +    char *hostname;
>>>> +    char *clientid;
>>>> +};
>>> This is missing
>>>
>>>     int type;                /* virIPAddrType */
>>>
>>> since it can be IPv4 or IPv6 in general.
>>>
>>> Also we have no 'prefix' field here. Although the 'prefix' is present in
>>> the network XML config, I think it would be useful to include it here
>>> too, for parity with the virDomainInterfaceIPAddress struct
>>>
>> The virNetworkDHCPLeases struct was made according to the lease
>> parameters returned by dnsmasq. Are you suggesting that libvirt might
>> shift to some other DHCP server in future? Also, in case the DHCP server
>> doesn't return the IP address type and/or prefix, in that case, should these
>> values be NULL, when returned to the user?
> When doing API design we must *not* make assumptions about the current
> impl that libvirt happens to have. We must always think of possible
> future changes.


I may be mistaken, but I thought I remembered someone talking about an
alternate implementation using ISC dhcpd instead of dnsmasq.

A concrete example of this is that in earlier versions of libvirt (and
still for earlier versions of dnsmasq even with current libvirt) we
use(d) radvd for ipv6 autoconf, but in newer libvirt with newer versions
of dnsmasq, we use the autoconf support that is now built into dnsmasq.
The libvirt network configuration remains the same; This is possible
because the configuration is agnostic to the underlying implementation.


>   The DHCP server doesn't need to return prefix - we have
> that available in virNetworkDefPtr. As for IP address type, we can currently
> hardcode that as IPV4, since we know dnsmasq only uses that. 

commit 2d5cd1d724084d9975b2514fb31776627acbe997
Author: Gene Czarcinski <gene@xxxxxxxxx>
Date:   Thu Dec 6 12:20:38 2012 -0500

    network: add support for DHCPv6


I haven't used it, but as far as I know it works (if it doesn't then
Gene is in a heap of trouble :-)


> If we later
> make use of DHCPv6  we can worry about distinguishing them at that point
>


--
libvir-list mailing list
libvir-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/libvir-list




[Index of Archives]     [Virt Tools]     [Libvirt Users]     [Lib OS Info]     [Fedora Users]     [Fedora Desktop]     [Fedora SELinux]     [Big List of Linux Books]     [Yosemite News]     [KDE Users]     [Fedora Tools]