On Tue, 29 Mar 2011 13:17:55 +0800
Wen Congyang <wency@xxxxxxxxxxxxxx> wrote:
> At 03/29/2011 12:32 PM, KAMEZAWA Hiroyuki Write:
> > On Tue, 29 Mar 2011 11:24:23 +0800
> > Wen Congyang <wency@xxxxxxxxxxxxxx> wrote:
> > Hmm ? I'll add a spec. on qemuDomainAt(De)tachDevicePersistent() to
> > never update vmdef when return !0. Is it ok ?
>
> No.
> In patch 2, the function qemuDomainDeviceAddressFixup() may modify vmdef
> and return -1:
> =============
> +static int qemuDomainDeviceAddressFixup(virDomainDefPtr vmdef, bool pci)
> +{
> + if (!pci && virDomainDefAddImplicitControllers(vmdef))
> + return -1;
> + /* added controller requires PCI address */
> + return qemuDomainAssignPCIAddresses(vmdef);
> +}
> +
> =============
> The function virDomainDefAddImplicitControllers() may modify vmdef. If
> qemuDomainAssignPCIAddresses() failed, there is no way to rollback the
> operation that virDomainDefAddImplicitControllers() do.
>
Of course, I'll add a rollback. That's a bug in patch'2', not here.
Thanks,
-Kame
--
libvir-list mailing list
libvir-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/libvir-list