On Nov 8, 2011, at 4:06 PM, Ben Pfaff wrote:
> On Tue, Nov 08, 2011 at 03:57:28PM -0800, Justin Pettit wrote:
>> +/* The flow label in the IPv6 header.
>> + *
>> + * Prereqs: NXM_OF_ETH_TYPE must match 0x86dd exactly.
>> + *
>> + * Format: 20-bit IPv6 flow label.
>
> In the least-significant 20 bits of the field, right?
>
>> @@ -162,6 +163,7 @@ struct mf_field {
>> * - "dl_vlan" is 2 bytes but only 12 bits.
>> * - "dl_vlan_pcp" is 1 byte but only 3 bits.
>> * - "is_frag" is 1 byte but only 2 bits.
>
> s/ipv6_flabel/ipv6_label/ here:
>
>> + * - "ipv6_flabel" is 4 bytes but only 20 bits.
>> */
>> unsigned int n_bytes; /* Width of the field in bytes. */
>> unsigned int n_bits; /* Number of significant bits in field. */
>
>> @@ -2509,7 +2515,7 @@ ofputil_normalize_rule(struct cls_rule *rule, enum
>> nx_flow_format flow_format)
>> MAY_ARP_SHA = 1 << 4, /* arp_sha */
>> MAY_ARP_THA = 1 << 5, /* arp_tha */
>> MAY_IPV6_ADDR = 1 << 6, /* ipv6_src, ipv6_dst */
>
> This change is unnecessary:
>
>> - MAY_ND_TARGET = 1 << 7 /* nd_target */
>> + MAY_ND_TARGET = 1 << 7, /* nd_target */
>> } may_match;
>>
>> struct flow_wildcards wc;
>
> Otherwise this looks good, thank you.
Thanks. I've updated the commit with your suggestions.
--Justin
_______________________________________________
dev mailing list
[email protected]
http://openvswitch.org/mailman/listinfo/dev
