octeontx2-pf: fix FLOW_DIS_IS_FRAGMENT implementation

Upon reviewing the flower control flags handling in
this driver, I notice that the key wasn't being used,
only the mask.

Ie. `tc flower ... ip_flags nofrag` was hardware
offloaded as `... ip_flags frag`.

Only compile tested, no access to HW.

Fixes: c672e37 ("octeontx2-pf: Add support to filter packet based on IP fragment")
Signed-off-by: Asbjørn Sloth Tønnesen <ast@fiberby.net>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Author: Asbjørn Sloth Tønnesen

drivers/net/ethernet/marvell/octeontx2/nic/otx2_tc.c

@@ -689,6 +689,7 @@ static int otx2_tc_prepare_flow(struct otx2_nic *nic, struct otx2_tc_flow *node,
 
 	if (flow_rule_match_key(rule, FLOW_DISSECTOR_KEY_CONTROL)) {
 		struct flow_match_control match;
+		u32 val;
 
 		flow_rule_match_control(rule, &match);
 		if (match.mask->flags & FLOW_DIS_FIRST_FRAG) {
@@ -697,12 +698,14 @@ static int otx2_tc_prepare_flow(struct otx2_nic *nic, struct otx2_tc_flow *node,
 		}
 
 		if (match.mask->flags & FLOW_DIS_IS_FRAGMENT) {
+			val = match.key->flags & FLOW_DIS_IS_FRAGMENT;
 			if (ntohs(flow_spec->etype) == ETH_P_IP) {
-				flow_spec->ip_flag = IPV4_FLAG_MORE;
+				flow_spec->ip_flag = val ? IPV4_FLAG_MORE : 0;
 				flow_mask->ip_flag = IPV4_FLAG_MORE;
 				req->features |= BIT_ULL(NPC_IPFRAG_IPV4);
 			} else if (ntohs(flow_spec->etype) == ETH_P_IPV6) {
-				flow_spec->next_header = IPPROTO_FRAGMENT;
+				flow_spec->next_header = val ?
+							 IPPROTO_FRAGMENT : 0;
 				flow_mask->next_header = 0xff;
 				req->features |= BIT_ULL(NPC_IPFRAG_IPV6);
 			} else {