net: remove skb->vlan_present

skb->vlan_present seems redundant.

We can instead derive it from this boolean expression:

vlan_present = skb->vlan_proto != 0 || skb->vlan_tci != 0

Add a new union, to access both fields in a single load/store
when possible.

	union {
		u32	vlan_all;
		struct {
		__be16	vlan_proto;
		__u16	vlan_tci;
		};
	};

This allows following patch to remove a conditional test in GRO stack.

Note:
  We move remcsum_offload to keep TC_AT_INGRESS_MASK
  and SKB_MONO_DELIVERY_TIME_MASK unchanged.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Yonghong Song <yhs@fb.com>
Acked-by: Martin KaFai Lau <martin.lau@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

arch/sparc/net/bpf_jit_comp_32.c

@@ -555,11 +555,11 @@ void bpf_jit_compile(struct bpf_prog *fp)
 				emit_skb_load16(vlan_tci, r_A);
 				break;
 			case BPF_ANC | SKF_AD_VLAN_TAG_PRESENT:
-				__emit_skb_load8(__pkt_vlan_present_offset, r_A);
-				if (PKT_VLAN_PRESENT_BIT)
-					emit_alu_K(SRL, PKT_VLAN_PRESENT_BIT);
-				if (PKT_VLAN_PRESENT_BIT < 7)
-					emit_andi(r_A, 1, r_A);
+				emit_skb_load32(vlan_all, r_A);
+				emit_cmpi(r_A, 0);
+				emit_branch_off(BE, 12);
+				emit_nop();
+				emit_loadimm(1, r_A);
 				break;
 			case BPF_LD | BPF_W | BPF_LEN:
 				emit_skb_load32(len, r_A);

drivers/net/ethernet/marvell/octeontx2/nic/otx2_pf.c

@@ -1973,7 +1973,7 @@ static u16 otx2_select_queue(struct net_device *netdev, struct sk_buff *skb,
 #endif
 
 #ifdef CONFIG_DCB
-	if (!skb->vlan_present)
+	if (!skb_vlan_tag_present(skb))
 		goto pick_tx;
 
 	vlan_prio = skb->vlan_tci >> 13;

include/linux/if_vlan.h

@@ -76,7 +76,7 @@ static inline bool is_vlan_dev(const struct net_device *dev)
         return dev->priv_flags & IFF_802_1Q_VLAN;
 }
 
-#define skb_vlan_tag_present(__skb)	((__skb)->vlan_present)
+#define skb_vlan_tag_present(__skb)	(!!(__skb)->vlan_all)
 #define skb_vlan_tag_get(__skb)		((__skb)->vlan_tci)
 #define skb_vlan_tag_get_id(__skb)	((__skb)->vlan_tci & VLAN_VID_MASK)
 #define skb_vlan_tag_get_cfi(__skb)	(!!((__skb)->vlan_tci & VLAN_CFI_MASK))
@@ -471,7 +471,7 @@ static inline struct sk_buff *vlan_insert_tag_set_proto(struct sk_buff *skb,
  */
 static inline void __vlan_hwaccel_clear_tag(struct sk_buff *skb)
 {
-	skb->vlan_present = 0;
+	skb->vlan_all = 0;
 }
 
 /**
@@ -483,9 +483,7 @@ static inline void __vlan_hwaccel_clear_tag(struct sk_buff *skb)
  */
 static inline void __vlan_hwaccel_copy_tag(struct sk_buff *dst, const struct sk_buff *src)
 {
-	dst->vlan_present = src->vlan_present;
-	dst->vlan_proto = src->vlan_proto;
-	dst->vlan_tci = src->vlan_tci;
+	dst->vlan_all = src->vlan_all;
 }
 
 /*
@@ -519,7 +517,6 @@ static inline void __vlan_hwaccel_put_tag(struct sk_buff *skb,
 {
 	skb->vlan_proto = vlan_proto;
 	skb->vlan_tci = vlan_tci;
-	skb->vlan_present = 1;
 }
 
 /**

include/linux/skbuff.h

@@ -818,7 +818,7 @@ typedef unsigned char *sk_buff_data_t;
  *	@mark: Generic packet mark
  *	@reserved_tailroom: (aka @mark) number of bytes of free space available
  *		at the tail of an sk_buff
- *	@vlan_present: VLAN tag is present
+ *	@vlan_all: vlan fields (proto & tci)
  *	@vlan_proto: vlan encapsulation protocol
  *	@vlan_tci: vlan tag control information
  *	@inner_protocol: Protocol (encapsulation)
@@ -951,7 +951,7 @@ struct sk_buff {
 	/* private: */
 	__u8			__pkt_vlan_present_offset[0];
 	/* public: */
-	__u8			vlan_present:1;	/* See PKT_VLAN_PRESENT_BIT */
+	__u8			remcsum_offload:1;
 	__u8			csum_complete_sw:1;
 	__u8			csum_level:2;
 	__u8			dst_pending_confirm:1;
@@ -966,7 +966,6 @@ struct sk_buff {
 
 	__u8			ipvs_property:1;
 	__u8			inner_protocol_type:1;
-	__u8			remcsum_offload:1;
 #ifdef CONFIG_NET_SWITCHDEV
 	__u8			offload_fwd_mark:1;
 	__u8			offload_l3_fwd_mark:1;
@@ -999,8 +998,13 @@ struct sk_buff {
 	__u32			priority;
 	int			skb_iif;
 	__u32			hash;
-	__be16			vlan_proto;
-	__u16			vlan_tci;
+	union {
+		u32		vlan_all;
+		struct {
+			__be16	vlan_proto;
+			__u16	vlan_tci;
+		};
+	};
 #if defined(CONFIG_NET_RX_BUSY_POLL) || defined(CONFIG_XPS)
 	union {
 		unsigned int	napi_id;
@@ -1059,15 +1063,13 @@ struct sk_buff {
 #endif
 #define PKT_TYPE_OFFSET		offsetof(struct sk_buff, __pkt_type_offset)
 
-/* if you move pkt_vlan_present, tc_at_ingress, or mono_delivery_time
+/* if you move tc_at_ingress or mono_delivery_time
  * around, you also must adapt these constants.
  */
 #ifdef __BIG_ENDIAN_BITFIELD
-#define PKT_VLAN_PRESENT_BIT	7
 #define TC_AT_INGRESS_MASK		(1 << 0)
 #define SKB_MONO_DELIVERY_TIME_MASK	(1 << 2)
 #else
-#define PKT_VLAN_PRESENT_BIT	0
 #define TC_AT_INGRESS_MASK		(1 << 7)
 #define SKB_MONO_DELIVERY_TIME_MASK	(1 << 5)
 #endif

lib/test_bpf.c

@@ -14346,7 +14346,6 @@ static struct sk_buff *populate_skb(char *buf, int size)
 	skb->hash = SKB_HASH;
 	skb->queue_mapping = SKB_QUEUE_MAP;
 	skb->vlan_tci = SKB_VLAN_TCI;
-	skb->vlan_present = SKB_VLAN_PRESENT;
 	skb->vlan_proto = htons(ETH_P_IP);
 	dev_net_set(&dev, &init_net);
 	skb->dev = &dev;

net/core/filter.c

@@ -325,11 +325,11 @@ static u32 convert_skb_access(int skb_field, int dst_reg, int src_reg,
 				      offsetof(struct sk_buff, vlan_tci));
 		break;
 	case SKF_AD_VLAN_TAG_PRESENT:
-		*insn++ = BPF_LDX_MEM(BPF_B, dst_reg, src_reg, PKT_VLAN_PRESENT_OFFSET);
-		if (PKT_VLAN_PRESENT_BIT)
-			*insn++ = BPF_ALU32_IMM(BPF_RSH, dst_reg, PKT_VLAN_PRESENT_BIT);
-		if (PKT_VLAN_PRESENT_BIT < 7)
-			*insn++ = BPF_ALU32_IMM(BPF_AND, dst_reg, 1);
+		BUILD_BUG_ON(sizeof_field(struct sk_buff, vlan_all) != 4);
+		*insn++ = BPF_LDX_MEM(BPF_W, dst_reg, src_reg,
+				      offsetof(struct sk_buff, vlan_all));
+		*insn++ = BPF_JMP_IMM(BPF_JEQ, dst_reg, 0, 1);
+		*insn++ = BPF_ALU32_IMM(BPF_MOV, dst_reg, 1);
 		break;
 	}
 
@@ -9290,13 +9290,11 @@ static u32 bpf_convert_ctx_access(enum bpf_access_type type,
 		break;
 
 	case offsetof(struct __sk_buff, vlan_present):
-		*target_size = 1;
-		*insn++ = BPF_LDX_MEM(BPF_B, si->dst_reg, si->src_reg,
-				      PKT_VLAN_PRESENT_OFFSET);
-		if (PKT_VLAN_PRESENT_BIT)
-			*insn++ = BPF_ALU32_IMM(BPF_RSH, si->dst_reg, PKT_VLAN_PRESENT_BIT);
-		if (PKT_VLAN_PRESENT_BIT < 7)
-			*insn++ = BPF_ALU32_IMM(BPF_AND, si->dst_reg, 1);
+		*insn++ = BPF_LDX_MEM(BPF_W, si->dst_reg, si->src_reg,
+				      bpf_target_off(struct sk_buff,
+						     vlan_all, 4, target_size));
+		*insn++ = BPF_JMP_IMM(BPF_JEQ, si->dst_reg, 0, 1);
+		*insn++ = BPF_ALU32_IMM(BPF_MOV, si->dst_reg, 1);
 		break;
 
 	case offsetof(struct __sk_buff, vlan_tci):