Dan Smith wrote:
# HG changeset patch # User Dan Smith <danms@us.ibm.com> # Date 1208189730 25200 # Node ID abf9b4881cc061cca318fe39c3c849f01316cdc8 # Parent da267e967adbd316d30303cc064584ebd36d7123 Make the 'default' NetRASD allocation include a randomly-generated MAC
No changes from last time, except I think this can be put in by itself, to reduce the size of my queue for the VSMS stuff.
Signed-off-by: Dan Smith <danms@us.ibm.com>
static struct sdc_rasd_prop *net_def(const CMPIObjectPath *ref, CMPIStatus *s) { @@ -440,6 +482,7 @@ static struct sdc_rasd_prop *net_def(con struct sdc_rasd_prop tmp[] = { {"InstanceID", (CMPIValue *)"Default", CMPI_chars}, {"VirtualQuantity", (CMPIValue *)&num_nics, CMPI_uint16}, + {"Address", (CMPIValue *)_net_rand_mac(), CMPI_chars}, PROP_END };
Is this going to segfault if _net_rand_mac() hits an error and comes back NULL? Or is the code that actually turns the struct into an inst able to handle that gracefully? -- -Jay