From: Jesper Dangaard Brouer <redacted>
Date: Fri, 10 Feb 2023 16:07:59 +0100

When function igc_rx_hash() was introduced in v4.20 via commit 0507ef8a0372
("igc: Add transmit and receive fastpath and interrupt handlers"), the
hardware wasn't configured to provide RSS hash, thus it made sense to not
enable net_device NETIF_F_RXHASH feature bit.

[...]

GENMASK()?

+
+/* igc_rss_type - Rx descriptor RSS type field */
+static inline u8 igc_rss_type(union igc_adv_rx_desc *rx_desc)

Why use types shorter than u32 on the stack?
Why this union is not const here, since there are no modifications?

+{
+	/* RSS Type 4-bit number: 0-9 (above 9 is reserved) */
+	return rx_desc->wb.lower.lo_dword.hs_rss.pkt_info & IGC_RSS_TYPE_MASK;

The most important I wanted to mention: doesn't this function make the
CPU read the uncached field again, while you could just read it once
onto the stack and then extract all such data from there?

+}
+
+/* Packet header type identified by hardware (when BIT(11) is zero).
+ * Even when UDP ports are not part of RSS hash HW still parse and mark UDP bits
+ */
+enum igc_pkt_type_bits {
+	IGC_PKT_TYPE_HDR_IPV4	=	BIT(0),
+	IGC_PKT_TYPE_HDR_IPV4_WITH_OPT=	BIT(1), /* IPv4 Hdr includes IP options */
+	IGC_PKT_TYPE_HDR_IPV6	=	BIT(2),
+	IGC_PKT_TYPE_HDR_IPV6_WITH_EXT=	BIT(3), /* IPv6 Hdr includes extensions */
+	IGC_PKT_TYPE_HDR_L4_TCP	=	BIT(4),
+	IGC_PKT_TYPE_HDR_L4_UDP	=	BIT(5),
+	IGC_PKT_TYPE_HDR_L4_SCTP=	BIT(6),
+	IGC_PKT_TYPE_HDR_NFS	=	BIT(7),
+	/* Above only valid when BIT(11) is zero */
+	IGC_PKT_TYPE_L2		=	BIT(11),
+	IGC_PKT_TYPE_VLAN	=	BIT(12),
+	IGC_PKT_TYPE_MASK	=	0x1FFF, /* 13-bits */

Also GENMASK().

+};
+
+/* igc_pkt_type - Rx descriptor Packet type field */
+static inline u16 igc_pkt_type(union igc_adv_rx_desc *rx_desc)

Also short types and consts.

+{
+	u32 data = le32_to_cpu(rx_desc->wb.lower.lo_dword.data);
+	/* Packet type is 13-bits - as bits (16:4) in lower.lo_dword*/
+	u16 pkt_type = (data >> 4) & IGC_PKT_TYPE_MASK;

Perfect candidate for FIELD_GET(). No, even for le32_get_bits().

Also my note above about excessive expensive reads.

Why make a struct for one field? + short type note

+} igc_rss_type_table[IGC_RSS_TYPE_MAX_TABLE] = {
+	[IGC_RSS_TYPE_NO_HASH].hash_type	  = PKT_HASH_TYPE_L2,
+	[IGC_RSS_TYPE_HASH_TCP_IPV4].hash_type	  = PKT_HASH_TYPE_L4,
+	[IGC_RSS_TYPE_HASH_IPV4].hash_type	  = PKT_HASH_TYPE_L3,
+	[IGC_RSS_TYPE_HASH_TCP_IPV6].hash_type	  = PKT_HASH_TYPE_L4,
+	[IGC_RSS_TYPE_HASH_IPV6_EX].hash_type	  = PKT_HASH_TYPE_L3,
+	[IGC_RSS_TYPE_HASH_IPV6].hash_type	  = PKT_HASH_TYPE_L3,
+	[IGC_RSS_TYPE_HASH_TCP_IPV6_EX].hash_type = PKT_HASH_TYPE_L4,
+	[IGC_RSS_TYPE_HASH_UDP_IPV4].hash_type	  = PKT_HASH_TYPE_L4,
+	[IGC_RSS_TYPE_HASH_UDP_IPV6].hash_type	  = PKT_HASH_TYPE_L4,
+	[IGC_RSS_TYPE_HASH_UDP_IPV6_EX].hash_type = PKT_HASH_TYPE_L4,
+	[10].hash_type = PKT_HASH_TYPE_L2, /* RSS Type above 9 "Reserved" by HW */
+	[11].hash_type = PKT_HASH_TYPE_L2,
+	[12].hash_type = PKT_HASH_TYPE_L2,
+	[13].hash_type = PKT_HASH_TYPE_L2,
+	[14].hash_type = PKT_HASH_TYPE_L2,
+	[15].hash_type = PKT_HASH_TYPE_L2,

Why define those empty if you could do a bound check in the code
instead? E.g. `if (unlikely(bigger_than_9)) return PKT_HASH_TYPE_L2`.

+};
+
 static inline void igc_rx_hash(struct igc_ring *ring,
 			       union igc_adv_rx_desc *rx_desc,
 			       struct sk_buff *skb)
 {
-	if (ring->netdev->features & NETIF_F_RXHASH)
-		skb_set_hash(skb,
-			     le32_to_cpu(rx_desc->wb.lower.hi_dword.rss),
-			     PKT_HASH_TYPE_L3);
+	if (ring->netdev->features & NETIF_F_RXHASH) {

	if (!(feature & HASH))
		return;

and -1 indent level?

+		u32 rss_hash = le32_to_cpu(rx_desc->wb.lower.hi_dword.rss);
+		u8  rss_type = igc_rss_type(rx_desc);
+		enum pkt_hash_types hash_type;
+
+		hash_type = igc_rss_type_table[rss_type].hash_type;
+		skb_set_hash(skb, rss_hash, hash_type);
+	}
 }

[...]

Thanks,
Olek