Re: [PATCH v2] linux: Signal datapath that unaligned Netlink message can be received

[Thomas Graf <tgraf@redhat.com>]

On 12/17/2013 02:23 AM, Jesse Gross wrote:
> On Sat, Nov 30, 2013 at 4:25 AM, Thomas Graf <tgraf@redhat.com> wrote:
>> diff --git a/lib/dpif-linux.c b/lib/dpif-linux.c
>> index 6c482d0..2d8a1aa 100644
>> --- a/lib/dpif-linux.c
>> +++ b/lib/dpif-linux.c
>> @@ -73,6 +73,7 @@ struct dpif_linux_dp {
>>       /* Attributes. */
>>       const char *name;                  /* OVS_DP_ATTR_NAME. */
>>       const uint32_t *upcall_pid;        /* OVS_DP_ATTR_UPCALL_PID. */
>> +    uint32_t user_features;            /* OVS_DP_ATTR_USER_FEATURES */
>>       struct ovs_dp_stats stats;         /* OVS_DP_ATTR_STATS. */
>>       struct ovs_dp_megaflow_stats megaflow_stats;
>>                                          /* OVS_DP_ATTR_MEGAFLOW_STATS.*/
>> @@ -228,6 +229,7 @@ dpif_linux_open(const struct dpif_class *class OVS_UNUSED, const char *name,
>>           dp_request.cmd = OVS_DP_CMD_NEW;
>>           upcall_pid = 0;
>>           dp_request.upcall_pid = &upcall_pid;
>> +        dp_request.user_features |= OVS_DP_F_UNALIGNED;
>>       } else {
>>           dp_request.cmd = OVS_DP_CMD_GET;
>>       }
>
> Does this handle the case where we are opening an existing datapath
> and need to update the features?

The patch (''openvswitch: Drop user features if old user space attempted
to create datapath'') takes care of resetting the features for the
downgrade case. The upgrade case with a persistent datapath object is
currently not implemented. The datapath object needs to be recreated.

Defining the NLM_F_REPLACE semantics is non trivial if we want to do
more than just update the settings. I will propose this in a follow up
patch.