On Wed, Sep 12, 2012 at 04:32:52PM +0400, Dmitry Guryanov wrote:
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@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.
Okay
Postponed with the indentation fix waiting for the 4th patch to be resubmitted,
So i have 1-3 in the queue, waiting for the 4th one :-) Daniel -- Daniel Veillard | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ daniel@veillard.com | Rpmfind RPM search engine http://rpmfind.net/ http://veillard.com/ | virtualization library http://libvirt.org/