[Libvirt-cim] [PATCH] Remove has_vnc_passwd key from infostore

[Eduardo Lima (Etrunko)]

On 07/20/2011 05:36 PM, Chip Vincent wrote:
> One question below...
>
> On 07/20/2011 04:20 PM, Eduardo Lima (Etrunko) wrote:
>> # HG changeset patch
>> # User Eduardo Lima (Etrunko)<eblima at br.ibm.com>
>> # Date 1311192727 10800
>> # Node ID 9746544f39f508bb08b0c7ea71153dbbceda8dd9
>> # Parent 6056961c3c5347d3b8375039767d7bc78fa97eb5
>> Remove has_vnc_passwd key from infostore
>>
>> The logic is not necessary anymore since we return the vnc password
>> in every call to retrieve a domain XML. Also fix some other places
>> where this field was not handled properly.
>>
>> Signed-off-by: Eduardo Lima (Etrunko)<eblima at br.ibm.com>
>>
>> diff --git a/libxkutil/device_parsing.c b/libxkutil/device_parsing.c
>> --- a/libxkutil/device_parsing.c
>> +++ b/libxkutil/device_parsing.c
>> @@ -809,6 +809,7 @@
>> DUP_FIELD(dev, _dev, dev.graphics.port);
>> DUP_FIELD(dev, _dev, dev.graphics.host);
>> DUP_FIELD(dev, _dev, dev.graphics.keymap);
>> + DUP_FIELD(dev, _dev, dev.graphics.passwd);
>> } else if (dev->type == CIM_RES_TYPE_INPUT) {
>> DUP_FIELD(dev, _dev, dev.input.type);
>> DUP_FIELD(dev, _dev, dev.input.bus);
>> @@ -889,7 +890,7 @@
>> char *xml;
>> int ret;
>>
>> - xml = virDomainGetXMLDesc(dom, 0);
>> + xml = virDomainGetXMLDesc(dom, VIR_DOMAIN_XML_SECURE);
>> if (xml == NULL)
>> return 0;
>>
>> diff --git a/src/Virt_ComputerSystem.c b/src/Virt_ComputerSystem.c
>> --- a/src/Virt_ComputerSystem.c
>> +++ b/src/Virt_ComputerSystem.c
>> @@ -920,7 +920,7 @@
>> return s;
>> }
>>
>> - xml = virDomainGetXMLDesc(dom, 0);
>> + xml = virDomainGetXMLDesc(dom, VIR_DOMAIN_XML_SECURE);
>> if (xml == NULL) {
>> CU_DEBUG("Unable to retrieve domain XML");
>> virt_set_status(_BROKER,&s,
>> diff --git a/src/Virt_ComputerSystemIndication.c
>> b/src/Virt_ComputerSystemIndication.c
>> --- a/src/Virt_ComputerSystemIndication.c
>> +++ b/src/Virt_ComputerSystemIndication.c
>> @@ -151,7 +151,7 @@
>> }
>>
>> (*dom_xml_list)[i].xml = virDomainGetXMLDesc(dom_ptr_list[i],
>> - 0);
>> + VIR_DOMAIN_XML_SECURE);
>> if ((*dom_xml_list)[i].xml == NULL) {
>> cu_statusf(_BROKER,&s,
>> CMPI_RC_ERR_FAILED,
>> diff --git a/src/Virt_RASD.c b/src/Virt_RASD.c
>> --- a/src/Virt_RASD.c
>> +++ b/src/Virt_RASD.c
>> @@ -420,8 +420,6 @@
>> CMPIStatus s = {CMPI_RC_OK, NULL};
>> virConnectPtr conn = NULL;
>> virDomainPtr dom = NULL;
>> - struct infostore_ctx *infostore = NULL;
>> - bool has_passwd = false;
>>
>> CMSetProperty(inst, "ResourceSubType",
>> (CMPIValue *)dev->dev.graphics.type, CMPI_chars);
>> @@ -460,19 +458,12 @@
>> goto out;
>> }
>>
>> - infostore = infostore_open(dom);
>> - if (infostore != NULL)
>> - has_passwd = infostore_get_bool(infostore,
>> - "has_vnc_passwd");
>> -
>> - if (has_passwd) {
>> + if (dev->dev.graphics.passwd&& strlen(dev->dev.graphics.passwd)) {
>
> Is a NULL password different than a zero-length password? This check
> assumes that's possible, but one would expect '****' to be present in
> the instance for non-NULL, zero-length strings. Right?
>

In practical terms, they are not different at all. If you edit the XML 
for a domain adding an empty password attribute for vnc you will get the 
exact same behavior as when that attribute is not there.

You will only get prompted for a password when it's lenght is > 0. And 
that was the reason for this extra check.

Best regards,

-- 
Eduardo de Barros Lima
Software Engineer, Open Virtualization
Linux Technology Center - IBM/Brazil
eblima at br.ibm.com