[libvirt] [PATCH 3/5] libvirt/qemu - support persistent modification of devices.

[Eric Blake]

On 03/02/2011 06:11 PM, KAMEZAWA Hiroyuki wrote:
>>From f7a997e3cd58cfac81e131afdd20c3691267831d Mon Sep 17 00:00:00 2001
> From: KAMEZAWA Hiroyuki <kamezawa.hiroyu at jp.fujitsu.com>
> Date: Thu, 3 Mar 2011 09:43:07 +0900
> Subject: [PATCH 3/5] libivrt/qemu - support persistent modification of devices.
> 
> Now, virsh attach-***/detach-*** has --persistent option and
> it can update XML definition of inactive domains.

Hmm, looking at the recent additions of other virsh persistent support,
the single --persistent flag of virsh attach-{device,disk,interface} are
a bit limited; --persistent means the same thing as --config, but we may
need to add a counterpart --live option that specifies live only.  But
cleaning up virsh to better expose the underlying flags is a separate
patch; this patch only deals with implementing the flags internally.

> Now, disk and nic are supported in Xen, but none in qemu.

It would be nice to get all devices supported, but incremental is better
than nothing at all.

> 
> This patch adds a function for permanent modification for qemu.
> Following patches will add disk/network suport.
> 
> TODO:
>   - finally, need to do hotplug + XML modification for active domain.
> 
> Changelog v3->v4
>   - fixed trailing white spaces.
>   - splitted into 2 part and this patch only contains wrappers.
> Changelog v2->v3:
>   - clean ups.
>   - handle all VIR_DOMAIN_DEVICE_MODIFY_XXX flags.
> ---
>  src/qemu/qemu_driver.c |  143 ++++++++++++++++++++++++++++++++++++++++++++----
>  1 files changed, 131 insertions(+), 12 deletions(-)
> 
> diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c
> index c9095bb..3248cdc 100644
> --- a/src/qemu/qemu_driver.c
> +++ b/src/qemu/qemu_driver.c
> @@ -4099,16 +4099,131 @@ cleanup:
>      return ret;
>  }
>  
> -static int qemudDomainAttachDeviceFlags(virDomainPtr dom,
> +/*
> + * Attach a device given by XML, the change will be persistent
> + * and domain XML definition file is updated.
> + */
> +static int qemuDomainAttachDevicePersistent(virDomainDefPtr vmdef,
> +                                            virDomainDeviceDefPtr newdev)
> +{
> +
> +    /* At first, check device confliction */

s/device confliction/for device support/

> +    switch(newdev->type) {
> +    default:
> +        qemuReportError(VIR_ERR_INVALID_ARG, "%s",
> +                        _("Sorry, the device is not suppored for now"));

s/suppored/supported/ (twice)

> +
> +static int qemuDomainModifyDevicePersistent(virDomainPtr dom,
>                                          const char *xml,

Indentation is not consistent.

> -                                        unsigned int flags) {
> -    if (flags & VIR_DOMAIN_DEVICE_MODIFY_CONFIG) {
> -        qemuReportError(VIR_ERR_OPERATION_INVALID,
> -                        "%s", _("cannot modify the persistent configuration of a domain"));
> +                                        unsigned int attach, int flags)
> +{
> +    struct qemud_driver *driver;
> +    virDomainDeviceDefPtr device;
> +    virDomainDefPtr vmdef;
> +    virDomainObjPtr vm;
> +    int ret = -1;
> +
> +    if (!dom || !dom->conn || !dom->name || !xml) {
> +        qemuReportError(VIR_ERR_INTERNAL_ERROR,
> +                        _("internal error : %s"),  __FUNCTION__);
> +        return -1;
> +    }
> +    /*
> +     * When both of MODIFY_CONFIG and MODIFY_LIVE is specified at the same time,
> +     * return error for now. We should support this later.
> +     */
> +    if (flags & VIR_DOMAIN_DEVICE_MODIFY_LIVE) {
> +        qemuReportError(VIR_ERR_INVALID_ARG, "%s",
> +                        _("Now, cannot modify alive domain and its definition "
> +                          "at the same time."));

It shouldn't be too hard to attempt supporting this now - after all, we
already did just that with Taku Izumi's patch for virsh setmem.
qemudDomainSetMemoryFlags can be a good reference point for how to
implement a command that first validates all conditions required by the
flags, then modifies the live configuration if requested and still live,
then modifies the persistent configuration if requested.

I think there have been enough findings already (Wen's feedback has also
been useful), as well as quite a bit of upstream churn that will require
rebasing aspects of your patch, that I'll wait for v5 of this series; if
you use 'git send-email --subject-prefix=PATCHv5', that will make it
easier to recognize from the subject line that it is a resend.

-- 
Eric Blake   eblake at redhat.com    +1-801-349-2682
Libvirt virtualization library http://libvirt.org

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 619 bytes
Desc: OpenPGP digital signature
URL: <http://listman.redhat.com/archives/libvir-list/attachments/20110315/6ca6da04/attachment-0001.sig>