dpif-netdev: Fix Auto Load Balance debug log.

In the case where there is a NUMA node that has a zero variance
improvement, the log will report it's variance improvement as value for
the previous NUMA node with a non-zero variance improvement.

For example in an artificial case:
|dpif_netdev|DBG|Numa node 1. Current variance 1000 Estimated variance 0.
Variance improvement 100%.
                     ^^^ correct value

|dpif_netdev|DBG|Numa node 0. Current variance 0 Estimated variance 0.
Variance improvement 100%.
                     ^^^ incorrect value for Numa 0, value from Numa 1

This is caused by not resetting the improvement variable between loops.

This is a debug log reporting issue only, non-zero variance improvement
will still trigger rebalance where appropriate.

Move improvement and other variables into the loop code block.

Fixes: 58fed7e ("dpif-netdev: Make PMD auto load balance use common rxq scheduling.")
Reported-at: https://issues.redhat.com/browse/FDP-1145
Signed-off-by: Kevin Traynor <ktraynor@redhat.com>

lib/dpif-netdev.c

@@ -6518,10 +6518,8 @@ pmd_rebalance_dry_run(struct dp_netdev *dp)
 {
     struct sched_numa_list numa_list_cur;
     struct sched_numa_list numa_list_est;
+    struct sched_numa *numa_cur;
     bool thresh_met = false;
-    uint64_t current_var, estimate_var;
-    struct sched_numa *numa_cur, *numa_est;
-    uint64_t improvement = 0;
 
     VLOG_DBG("PMD auto load balance performing dry run.");
 
@@ -6540,6 +6538,10 @@ pmd_rebalance_dry_run(struct dp_netdev *dp)
 
         /* Calculate variances. */
         HMAP_FOR_EACH (numa_cur, node, &numa_list_cur.numas) {
+            uint64_t current_var, estimate_var;
+            struct sched_numa *numa_est;
+            uint64_t improvement = 0;
+
             numa_est = sched_numa_list_lookup(&numa_list_est,
                                               numa_cur->numa_id);
             if (!numa_est) {