Re: [libvirt] [PATCH 3/4] parallels: fix parallelsDomainDefineXML for existing containers
On 120912 16:14:47, Daniel Veillard wrote:
On Mon, Sep 10, 2012 at 07:22:44PM +0400, Dmitry Guryanov wrote:
> Fix code, which checks what is changed in virDomainDef structure.
> It looks slightly different for containers and VMs: containers haven't
> boot devices, but have init path
>
> Signed-off-by: Dmitry Guryanov <dguryanov(a)parallels.com>
> ---
> src/parallels/parallels_driver.c | 42 ++++++++++++++++++++++++++++---------
> 1 files changed, 32 insertions(+), 10 deletions(-)
>
> diff --git a/src/parallels/parallels_driver.c b/src/parallels/parallels_driver.c
> index ace75a6..fd6ba88 100644
> --- a/src/parallels/parallels_driver.c
> +++ b/src/parallels/parallels_driver.c
> @@ -1484,24 +1484,46 @@ parallelsApplyChanges(virDomainObjPtr dom, virDomainDefPtr
new)
> return -1;
> }
>
> - /* we fill only type and arch fields in parallelsLoadDomain, so
> - * we can check that all other paramenters are null */
> + /* we fill only type and arch fields in parallelsLoadDomain for
> + * hvm type and also init for containers, so we can check that all
> + * other paramenters are null and boot devices config is default */
> +
> if (!STREQ_NULLABLE(old->os.type, new->os.type) ||
> !STREQ_NULLABLE(old->os.arch, new->os.arch) ||
So here you implicitely allow an update where the new os.type or
os.arch would be NULL, I assume that any definition in the system
should have those set, right ?
old->os.type and old->os.arch are always non-NULL, because we set them
in parallelsLoadDomain and don't change later. So if new->os.type/arch
is NULL or not equal to old - we report error.
> - new->os.machine != NULL || new->os.nBootDevs != 1 ||
> - new->os.bootDevs[0] != VIR_DOMAIN_BOOT_DISK ||
> - new->os.bootmenu != 0 || new->os.init != NULL ||
> - new->os.initargv != NULL || new->os.kernel != NULL ||
> - new->os.initrd != NULL || new->os.cmdline != NULL ||
> - new->os.root != NULL || new->os.loader != NULL ||
> - new->os.bootloader != NULL || new->os.bootloaderArgs != NULL ||
> - new->os.smbios_mode != 0 || new->os.bios.useserial != 0) {
> + new->os.machine != NULL || new->os.bootmenu != 0 ||
> + new->os.kernel != NULL || new->os.initrd != NULL ||
> + new->os.cmdline != NULL || new->os.root != NULL ||
> + new->os.loader != NULL || new->os.bootloader != NULL ||
> + new->os.bootloaderArgs != NULL || new->os.smbios_mode != 0 ||
> + new->os.bios.useserial != 0) {
>
> virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s",
> _("changing OS parameters is not supported "
> "by parallels driver"));
> return -1;
> }
> + if (STREQ(new->os.type, "hvm")) {
But here you cmpare to new->os.type, which could potentially be NULL
at this point. Shouldn't that be changed to old->os.type instead ?
Here we know, that new->os.type is equal to old one, and not NULL.
> + if (new->os.nBootDevs != 1 ||
> + new->os.bootDevs[0] != VIR_DOMAIN_BOOT_DISK ||
> + new->os.init != NULL || new->os.initargv != NULL) {
> +
> + virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s",
> + _("changing OS parameters is not supported "
> + "by parallels driver"));
> + return -1;
> + }
indentation error, here we are closing the second "if"
> + } else {
> + if (new->os.nBootDevs != 0 ||
> + !STREQ_NULLABLE(old->os.init, new->os.init) ||
> + (new->os.initargv != NULL && new->os.initargv[0] !=
NULL)) {
> +
> + virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s",
> + _("changing OS parameters is not supported "
> + "by parallels driver"));
> + return -1;
> + }
> + }
> +
>
> if (!STREQ_NULLABLE(old->emulator, new->emulator)) {
> virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s",
Postponed with the indentation fix waiting for the 4th patch to be
resubmitted,
Daniel
--
Daniel Veillard | libxml Gnome XML XSLT toolkit http://xmlsoft.org/
daniel(a)veillard.com | Rpmfind RPM search engine http://rpmfind.net/
http://veillard.com/ | virtualization library http://libvirt.org/