e1000: Correct Rx descriptor threshold programming

Jumbo receive tuning on integrated controllers enabled PTHRESH without
a nonzero HTHRESH, contrary to the hardware programming requirements.
It also covered only the integrated MAC generations present when the
workaround was added.  Enumerate every jumbo-capable ICH and PCH type
and program PTHRESH=3 with HTHRESH=1.  Linux fixed the same HTHRESH
omission in b701cacdbcfb.

The 82574 path combined threshold values with the reset values using
bitwise OR.  Requesting WTHRESH=4 while the reset value was one thus
programmed five.  Clear the complete threshold fields before installing
the established PTHRESH=32, HTHRESH=4, WTHRESH=4 descriptor-granularity
policy.

MFC after:	2 weeks

(cherry picked from commit abe22383f1)

sys/dev/e1000/if_em.c

@@ -3899,6 +3899,27 @@ em_initialize_transmit_unit(if_ctx_t ctx)
  **********************************************************************/
 #define BSIZEPKT_ROUNDUP ((1<<E1000_SRRCTL_BSIZEPKT_SHIFT)-1)
 
+static bool
+em_integrated_jumbo_rx(struct e1000_hw *hw)
+{
+	switch (hw->mac.type) {
+	case e1000_ich9lan:
+	case e1000_ich10lan:
+	case e1000_pchlan:
+	case e1000_pch2lan:
+	case e1000_pch_lpt:
+	case e1000_pch_spt:
+	case e1000_pch_cnp:
+	case e1000_pch_tgp:
+	case e1000_pch_adp:
+	case e1000_pch_mtp:
+	case e1000_pch_ptp:
+		return (true);
+	default:
+		return (false);
+	}
+}
+
 static void
 em_initialize_receive_unit(if_ctx_t ctx)
 {
@@ -4048,24 +4069,25 @@ em_initialize_receive_unit(if_ctx_t ctx)
 		E1000_WRITE_REG(hw, E1000_RDT(i), 0);
 	}
 
-	/*
-	 * Set PTHRESH for improved jumbo performance
-	 * According to 10.2.5.11 of Intel 82574 Datasheet,
-	 * RXDCTL(1) is written whenever RXDCTL(0) is written.
-	 * Only write to RXDCTL(1) if there is a need for different
-	 * settings.
-	 */
-	if ((hw->mac.type == e1000_ich9lan || hw->mac.type == e1000_pch2lan ||
-	    hw->mac.type == e1000_ich10lan) && if_getmtu(ifp) > ETHERMTU) {
+	/* Increase receive-descriptor prefetching for integrated jumbo MACs. */
+	if (em_integrated_jumbo_rx(hw) && if_getmtu(ifp) > ETHERMTU) {
 		u32 rxdctl = E1000_READ_REG(hw, E1000_RXDCTL(0));
-		E1000_WRITE_REG(hw, E1000_RXDCTL(0), rxdctl | 3);
+
+		rxdctl &= ~(EM_RXDCTL_PTHRESH_MASK |
+		    EM_RXDCTL_HTHRESH_MASK);
+		rxdctl |= EM_JUMBO_RX_PTHRESH |
+		    (EM_JUMBO_RX_HTHRESH << 8);
+		E1000_WRITE_REG(hw, E1000_RXDCTL(0), rxdctl);
 	} else if (hw->mac.type == e1000_82574) {
+		/* RXDCTL(0) writes are mirrored to RXDCTL(1) on 82574. */
 		for (int i = 0; i < sc->rx_num_queues; i++) {
 			u32 rxdctl = E1000_READ_REG(hw, E1000_RXDCTL(i));
-			rxdctl |= 0x20; /* PTHRESH */
-			rxdctl |= 4 << 8; /* HTHRESH */
-			rxdctl |= 4 << 16;/* WTHRESH */
-			rxdctl |= 1 << 24; /* Switch to granularity */
+
+			rxdctl &= ~EM_RXDCTL_THRESH_MASK;
+			rxdctl |= EM_82574_RX_PTHRESH |
+			    (EM_82574_RX_HTHRESH << 8) |
+			    (EM_82574_RX_WTHRESH << 16) |
+			    E1000_RXDCTL_THRESH_UNIT_DESC;
 			E1000_WRITE_REG(hw, E1000_RXDCTL(i), rxdctl);
 		}
 	} else if (hw->mac.type >= igb_mac_min) {

sys/dev/e1000/if_em.h

@@ -297,6 +297,19 @@
 #define EM_TX_HTHRESH		1
 #define EM_TX_WTHRESH		1
 
+#define EM_RXDCTL_PTHRESH_MASK	0x0000003F
+#define EM_RXDCTL_HTHRESH_MASK	0x00003F00
+#define EM_RXDCTL_WTHRESH_MASK	0x003F0000
+#define EM_RXDCTL_THRESH_MASK	(EM_RXDCTL_PTHRESH_MASK | \
+				 EM_RXDCTL_HTHRESH_MASK | \
+				 EM_RXDCTL_WTHRESH_MASK)
+
+#define EM_JUMBO_RX_PTHRESH	3
+#define EM_JUMBO_RX_HTHRESH	1
+#define EM_82574_RX_PTHRESH	32
+#define EM_82574_RX_HTHRESH	4
+#define EM_82574_RX_WTHRESH	4
+
 #define IGB_RX_PTHRESH	((hw->mac.type == e1000_i354) ? 12 : \
 			    ((hw->mac.type <= e1000_82576) ? 16 : 8))
 #define IGB_RX_HTHRESH	8