On 2012年02月07日 03:13, Matthias Bolte wrote:
2012/2/6 Eric Blake<eblake@redhat.com>:
On 02/06/2012 02:11 AM, Osier Yang wrote:
The auto-generated WWN comply with the new addressing schema of WWN:
<quote> the first nibble is either hex 5 or 6 followed by a 3-byte vendor identifier and 36 bits for a vendor-specified serial number. </quote>
We choose hex 5 for the first nibble. And use Qumranet's OUI (00:1A:4A) as the 3-byte vendor indentifier. The last 36 bits are auto-generated. --- src/conf/node_device_conf.c | 36 +++++++++++++++++++----------------- src/libvirt_private.syms | 1 + src/util/virrandom.c | 18 ++++++++++++++++++ src/util/virrandom.h | 1 + 4 files changed, 39 insertions(+), 17 deletions(-)
+ +#define QUMRANET_OUI "001a4a" + +int +virRandomGenerateWWN(char **wwn) { + if (virAsprintf(wwn, "5" QUMRANET_OUI "%09" PRIx64, + virRandomBits(36))< 0) { + virReportOOMError(); + return -1; + } + + return 0;
ACK. It looks a lot shorter than v4, which is a good thing :)
Maybe I miss something here because I didn't follow the whole version history, but doesn't this result in using Qumranet's OUI for all generated WWN's in libvirt, also for ESX and Hyper-V (assume the drivers would support node devices yet) that might use different OUIs? Or do I misunderstand this?
Yeah, that's strange to see Qumranet's in ESX or other else environment, I created v6 to choose OUI according to the underlying hypervisor driver. Thanks for pointing it out. Regards, Osier