[PATCH net 0/6] Fixes for NXP ENETC driver

[PATCH net 0/6] Fixes for NXP ENETC driver

[Vladimir Oltean]

From: Vladimir Oltean <vladimir.oltean@nxp.com>

This contains an assorted set of fixes collected over the past week on
the enetc driver. Some are related to VLAN processing, some to physical
link settings, some are fixups of previous hardware workarounds.

Vladimir Oltean (6):
  net: enetc: don't overwrite the RSS indirection table when
    initializing
  net: enetc: initialize RFS/RSS memories for unused ports too
  net: enetc: take the MDIO lock only once per NAPI poll cycle
  net: enetc: fix incorrect TPID when receiving 802.1ad tagged packets
  net: enetc: don't disable VLAN filtering in IFF_PROMISC mode
  net: enetc: force the RGMII speed and duplex instead of operating in
    inband mode

 drivers/net/ethernet/freescale/enetc/enetc.c  | 127 +++++-------------
 drivers/net/ethernet/freescale/enetc/enetc.h  |   5 +
 .../net/ethernet/freescale/enetc/enetc_cbdr.c |  54 ++++++++
 .../net/ethernet/freescale/enetc/enetc_hw.h   |  18 ++-
 .../net/ethernet/freescale/enetc/enetc_pf.c   | 103 +++++++++++---
 .../net/ethernet/freescale/enetc/enetc_vf.c   |   7 +
 6 files changed, 202 insertions(+), 112 deletions(-)

-- 
2.25.1

[PATCH net 1/6] net: enetc: don't overwrite the RSS indirection table when initializing

[Vladimir Oltean]

From: Vladimir Oltean <vladimir.oltean@nxp.com>

After the blamed patch, all RX traffic gets hashed to CPU 0 because the
hashing indirection table set up in:

enetc_pf_probe
-> enetc_alloc_si_resources
   -> enetc_configure_si
      -> enetc_setup_default_rss_table

is overwritten later in:

enetc_pf_probe
-> enetc_init_port_rss_memory

which zero-initializes the entire port RSS table in order to avoid ECC errors.

The trouble really is that enetc_init_port_rss_memory really neads
enetc_alloc_si_resources to be called, because it depends upon
enetc_alloc_cbdr and enetc_setup_cbdr. But that whole enetc_configure_si
thing could have been better thought out, it has nothing to do in a
function called "alloc_si_resources", especially since its counterpart,
"free_si_resources", does nothing to unwind the configuration of the SI.

The point is, we need to pull out enetc_configure_si out of
enetc_alloc_resources, and move it after enetc_init_port_rss_memory.
This allows us to set up the default RSS indirection table after
initializing the memory.

Fixes: 07bf34a50e32 ("net: enetc: initialize the RFS and RSS memories")
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
 drivers/net/ethernet/freescale/enetc/enetc.c    | 11 +++--------
 drivers/net/ethernet/freescale/enetc/enetc.h    |  1 +
 drivers/net/ethernet/freescale/enetc/enetc_pf.c |  7 +++++++
 drivers/net/ethernet/freescale/enetc/enetc_vf.c |  7 +++++++
 4 files changed, 18 insertions(+), 8 deletions(-)

diff --git a/drivers/net/ethernet/freescale/enetc/enetc.c b/drivers/net/ethernet/freescale/enetc/enetc.c
index c78d12229730..fdb6b9e8da78 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc.c
@@ -1058,13 +1058,12 @@ static int enetc_setup_default_rss_table(struct enetc_si *si, int num_groups)
 	return 0;
 }
 
-static int enetc_configure_si(struct enetc_ndev_priv *priv)
+int enetc_configure_si(struct enetc_ndev_priv *priv)
 {
 	struct enetc_si *si = priv->si;
 	struct enetc_hw *hw = &si->hw;
 	int err;
 
-	enetc_setup_cbdr(hw, &si->cbd_ring);
 	/* set SI cache attributes */
 	enetc_wr(hw, ENETC_SICAR0,
 		 ENETC_SICAR_RD_COHERENT | ENETC_SICAR_WR_COHERENT);
@@ -1112,6 +1111,8 @@ int enetc_alloc_si_resources(struct enetc_ndev_priv *priv)
 	if (err)
 		return err;
 
+	enetc_setup_cbdr(&si->hw, &si->cbd_ring);
+
 	priv->cls_rules = kcalloc(si->num_fs_entries, sizeof(*priv->cls_rules),
 				  GFP_KERNEL);
 	if (!priv->cls_rules) {
@@ -1119,14 +1120,8 @@ int enetc_alloc_si_resources(struct enetc_ndev_priv *priv)
 		goto err_alloc_cls;
 	}
 
-	err = enetc_configure_si(priv);
-	if (err)
-		goto err_config_si;
-
 	return 0;
 
-err_config_si:
-	kfree(priv->cls_rules);
 err_alloc_cls:
 	enetc_clear_cbdr(&si->hw);
 	enetc_free_cbdr(priv->dev, &si->cbd_ring);
diff --git a/drivers/net/ethernet/freescale/enetc/enetc.h b/drivers/net/ethernet/freescale/enetc/enetc.h
index 8532d23b54f5..f8275cef3b5c 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc.h
+++ b/drivers/net/ethernet/freescale/enetc/enetc.h
@@ -292,6 +292,7 @@ void enetc_get_si_caps(struct enetc_si *si);
 void enetc_init_si_rings_params(struct enetc_ndev_priv *priv);
 int enetc_alloc_si_resources(struct enetc_ndev_priv *priv);
 void enetc_free_si_resources(struct enetc_ndev_priv *priv);
+int enetc_configure_si(struct enetc_ndev_priv *priv);
 
 int enetc_open(struct net_device *ndev);
 int enetc_close(struct net_device *ndev);
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pf.c b/drivers/net/ethernet/freescale/enetc/enetc_pf.c
index 515c5b29d7aa..d02ecb2e46ae 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_pf.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_pf.c
@@ -1108,6 +1108,12 @@ static int enetc_pf_probe(struct pci_dev *pdev,
 		goto err_init_port_rss;
 	}
 
+	err = enetc_configure_si(priv);
+	if (err) {
+		dev_err(&pdev->dev, "Failed to configure SI\n");
+		goto err_config_si;
+	}
+
 	err = enetc_alloc_msix(priv);
 	if (err) {
 		dev_err(&pdev->dev, "MSIX alloc failed\n");
@@ -1136,6 +1142,7 @@ static int enetc_pf_probe(struct pci_dev *pdev,
 	enetc_mdiobus_destroy(pf);
 err_mdiobus_create:
 	enetc_free_msix(priv);
+err_config_si:
 err_init_port_rss:
 err_init_port_rfs:
 err_alloc_msix:
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_vf.c b/drivers/net/ethernet/freescale/enetc/enetc_vf.c
index 39c1a09e69a9..9b755a84c2d6 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_vf.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_vf.c
@@ -171,6 +171,12 @@ static int enetc_vf_probe(struct pci_dev *pdev,
 		goto err_alloc_si_res;
 	}
 
+	err = enetc_configure_si(priv);
+	if (err) {
+		dev_err(&pdev->dev, "Failed to configure SI\n");
+		goto err_config_si;
+	}
+
 	err = enetc_alloc_msix(priv);
 	if (err) {
 		dev_err(&pdev->dev, "MSIX alloc failed\n");
@@ -187,6 +193,7 @@ static int enetc_vf_probe(struct pci_dev *pdev,
 
 err_reg_netdev:
 	enetc_free_msix(priv);
+err_config_si:
 err_alloc_msix:
 	enetc_free_si_resources(priv);
 err_alloc_si_res:
-- 
2.25.1

[PATCH net 2/6] net: enetc: initialize RFS/RSS memories for unused ports too

[Vladimir Oltean]

From: Vladimir Oltean <vladimir.oltean@nxp.com>

Michael reports that since linux-next-20210211, the AER messages for ECC
errors have started reappearing, and this time they can be reliably
reproduced with the first ping on one of his LS1028A boards.

$ ping 1[   33.258069] pcieport 0000:00:1f.0: AER: Multiple Corrected error received: 0000:00:00.0
72.16.0.1
PING [   33.267050] pcieport 0000:00:1f.0: AER: can't find device of ID0000
172.16.0.1 (172.16.0.1): 56 data bytes
64 bytes from 172.16.0.1: seq=0 ttl=64 time=17.124 ms
64 bytes from 172.16.0.1: seq=1 ttl=64 time=0.273 ms

$ devmem 0x1f8010e10 32
0xC0000006

It isn't clear why this is necessary, but it seems that for the errors
to go away, we must clear the entire RFS and RSS memory, not just for
the ports in use.

Sadly the code is structured in such a way that we can't have unified
logic for the used and unused ports. For the minimal initialization of
an unused port, we need just to enable and ioremap the PF memory space,
and a control buffer descriptor ring. Unused ports must then free the
CBDR because the driver will exit, but used ports can not pick up from
where that code path left, since the CBDR API does not reinitialize a
ring when setting it up, so its producer and consumer indices are out of
sync between the software and hardware state. So a separate
enetc_init_unused_port function was created, and it gets called right
after the PF memory space is enabled.

Note that we need access from enetc_pf.c to the CBDR creation and
deletion methods, which were for some reason put in enetc.c. While
changing their definitions to be non-static, also move them to
enetc_cbdr.c which seems like a better place to hold these.

Fixes: 07bf34a50e32 ("net: enetc: initialize the RFS and RSS memories")
Reported-by: Michael Walle <michael@walle.cc>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
 drivers/net/ethernet/freescale/enetc/enetc.c  | 54 -------------------
 drivers/net/ethernet/freescale/enetc/enetc.h  |  4 ++
 .../net/ethernet/freescale/enetc/enetc_cbdr.c | 54 +++++++++++++++++++
 .../net/ethernet/freescale/enetc/enetc_pf.c   | 33 ++++++++++--
 4 files changed, 86 insertions(+), 59 deletions(-)

diff --git a/drivers/net/ethernet/freescale/enetc/enetc.c b/drivers/net/ethernet/freescale/enetc/enetc.c
index fdb6b9e8da78..43f0fae30080 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc.c
@@ -984,60 +984,6 @@ static void enetc_free_rxtx_rings(struct enetc_ndev_priv *priv)
 		enetc_free_tx_ring(priv->tx_ring[i]);
 }
 
-static int enetc_alloc_cbdr(struct device *dev, struct enetc_cbdr *cbdr)
-{
-	int size = cbdr->bd_count * sizeof(struct enetc_cbd);
-
-	cbdr->bd_base = dma_alloc_coherent(dev, size, &cbdr->bd_dma_base,
-					   GFP_KERNEL);
-	if (!cbdr->bd_base)
-		return -ENOMEM;
-
-	/* h/w requires 128B alignment */
-	if (!IS_ALIGNED(cbdr->bd_dma_base, 128)) {
-		dma_free_coherent(dev, size, cbdr->bd_base, cbdr->bd_dma_base);
-		return -EINVAL;
-	}
-
-	cbdr->next_to_clean = 0;
-	cbdr->next_to_use = 0;
-
-	return 0;
-}
-
-static void enetc_free_cbdr(struct device *dev, struct enetc_cbdr *cbdr)
-{
-	int size = cbdr->bd_count * sizeof(struct enetc_cbd);
-
-	dma_free_coherent(dev, size, cbdr->bd_base, cbdr->bd_dma_base);
-	cbdr->bd_base = NULL;
-}
-
-static void enetc_setup_cbdr(struct enetc_hw *hw, struct enetc_cbdr *cbdr)
-{
-	/* set CBDR cache attributes */
-	enetc_wr(hw, ENETC_SICAR2,
-		 ENETC_SICAR_RD_COHERENT | ENETC_SICAR_WR_COHERENT);
-
-	enetc_wr(hw, ENETC_SICBDRBAR0, lower_32_bits(cbdr->bd_dma_base));
-	enetc_wr(hw, ENETC_SICBDRBAR1, upper_32_bits(cbdr->bd_dma_base));
-	enetc_wr(hw, ENETC_SICBDRLENR, ENETC_RTBLENR_LEN(cbdr->bd_count));
-
-	enetc_wr(hw, ENETC_SICBDRPIR, 0);
-	enetc_wr(hw, ENETC_SICBDRCIR, 0);
-
-	/* enable ring */
-	enetc_wr(hw, ENETC_SICBDRMR, BIT(31));
-
-	cbdr->pir = hw->reg + ENETC_SICBDRPIR;
-	cbdr->cir = hw->reg + ENETC_SICBDRCIR;
-}
-
-static void enetc_clear_cbdr(struct enetc_hw *hw)
-{
-	enetc_wr(hw, ENETC_SICBDRMR, 0);
-}
-
 static int enetc_setup_default_rss_table(struct enetc_si *si, int num_groups)
 {
 	int *rss_table;
diff --git a/drivers/net/ethernet/freescale/enetc/enetc.h b/drivers/net/ethernet/freescale/enetc/enetc.h
index f8275cef3b5c..8b380fc13314 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc.h
+++ b/drivers/net/ethernet/freescale/enetc/enetc.h
@@ -310,6 +310,10 @@ int enetc_setup_tc(struct net_device *ndev, enum tc_setup_type type,
 void enetc_set_ethtool_ops(struct net_device *ndev);
 
 /* control buffer descriptor ring (CBDR) */
+int enetc_alloc_cbdr(struct device *dev, struct enetc_cbdr *cbdr);
+void enetc_free_cbdr(struct device *dev, struct enetc_cbdr *cbdr);
+void enetc_setup_cbdr(struct enetc_hw *hw, struct enetc_cbdr *cbdr);
+void enetc_clear_cbdr(struct enetc_hw *hw);
 int enetc_set_mac_flt_entry(struct enetc_si *si, int index,
 			    char *mac_addr, int si_map);
 int enetc_clear_mac_flt_entry(struct enetc_si *si, int index);
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_cbdr.c b/drivers/net/ethernet/freescale/enetc/enetc_cbdr.c
index 201cbc362e33..ad6aecda6b47 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_cbdr.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_cbdr.c
@@ -3,6 +3,60 @@
 
 #include "enetc.h"
 
+int enetc_alloc_cbdr(struct device *dev, struct enetc_cbdr *cbdr)
+{
+	int size = cbdr->bd_count * sizeof(struct enetc_cbd);
+
+	cbdr->bd_base = dma_alloc_coherent(dev, size, &cbdr->bd_dma_base,
+					   GFP_KERNEL);
+	if (!cbdr->bd_base)
+		return -ENOMEM;
+
+	/* h/w requires 128B alignment */
+	if (!IS_ALIGNED(cbdr->bd_dma_base, 128)) {
+		dma_free_coherent(dev, size, cbdr->bd_base, cbdr->bd_dma_base);
+		return -EINVAL;
+	}
+
+	cbdr->next_to_clean = 0;
+	cbdr->next_to_use = 0;
+
+	return 0;
+}
+
+void enetc_free_cbdr(struct device *dev, struct enetc_cbdr *cbdr)
+{
+	int size = cbdr->bd_count * sizeof(struct enetc_cbd);
+
+	dma_free_coherent(dev, size, cbdr->bd_base, cbdr->bd_dma_base);
+	cbdr->bd_base = NULL;
+}
+
+void enetc_setup_cbdr(struct enetc_hw *hw, struct enetc_cbdr *cbdr)
+{
+	/* set CBDR cache attributes */
+	enetc_wr(hw, ENETC_SICAR2,
+		 ENETC_SICAR_RD_COHERENT | ENETC_SICAR_WR_COHERENT);
+
+	enetc_wr(hw, ENETC_SICBDRBAR0, lower_32_bits(cbdr->bd_dma_base));
+	enetc_wr(hw, ENETC_SICBDRBAR1, upper_32_bits(cbdr->bd_dma_base));
+	enetc_wr(hw, ENETC_SICBDRLENR, ENETC_RTBLENR_LEN(cbdr->bd_count));
+
+	enetc_wr(hw, ENETC_SICBDRPIR, 0);
+	enetc_wr(hw, ENETC_SICBDRCIR, 0);
+
+	/* enable ring */
+	enetc_wr(hw, ENETC_SICBDRMR, BIT(31));
+
+	cbdr->pir = hw->reg + ENETC_SICBDRPIR;
+	cbdr->cir = hw->reg + ENETC_SICBDRCIR;
+}
+
+void enetc_clear_cbdr(struct enetc_hw *hw)
+{
+	enetc_wr(hw, ENETC_SICBDRMR, 0);
+}
+
 static void enetc_clean_cbdr(struct enetc_si *si)
 {
 	struct enetc_cbdr *ring = &si->cbd_ring;
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pf.c b/drivers/net/ethernet/freescale/enetc/enetc_pf.c
index d02ecb2e46ae..62ba4bf56f0d 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_pf.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_pf.c
@@ -1041,6 +1041,26 @@ static int enetc_init_port_rss_memory(struct enetc_si *si)
 	return err;
 }
 
+static void enetc_init_unused_port(struct enetc_si *si)
+{
+	struct device *dev = &si->pdev->dev;
+	struct enetc_hw *hw = &si->hw;
+	int err;
+
+	si->cbd_ring.bd_count = ENETC_CBDR_DEFAULT_SIZE;
+	err = enetc_alloc_cbdr(dev, &si->cbd_ring);
+	if (err)
+		return;
+
+	enetc_setup_cbdr(hw, &si->cbd_ring);
+
+	enetc_init_port_rfs_memory(si);
+	enetc_init_port_rss_memory(si);
+
+	enetc_clear_cbdr(hw);
+	enetc_free_cbdr(dev, &si->cbd_ring);
+}
+
 static int enetc_pf_probe(struct pci_dev *pdev,
 			  const struct pci_device_id *ent)
 {
@@ -1051,11 +1071,6 @@ static int enetc_pf_probe(struct pci_dev *pdev,
 	struct enetc_pf *pf;
 	int err;
 
-	if (node && !of_device_is_available(node)) {
-		dev_info(&pdev->dev, "device is disabled, skipping\n");
-		return -ENODEV;
-	}
-
 	err = enetc_pci_probe(pdev, KBUILD_MODNAME, sizeof(*pf));
 	if (err) {
 		dev_err(&pdev->dev, "PCI probing failed\n");
@@ -1069,6 +1084,13 @@ static int enetc_pf_probe(struct pci_dev *pdev,
 		goto err_map_pf_space;
 	}
 
+	if (node && !of_device_is_available(node)) {
+		enetc_init_unused_port(si);
+		dev_info(&pdev->dev, "device is disabled, skipping\n");
+		err = -ENODEV;
+		goto err_device_disabled;
+	}
+
 	pf = enetc_si_priv(si);
 	pf->si = si;
 	pf->total_vfs = pci_sriov_get_totalvfs(pdev);
@@ -1151,6 +1173,7 @@ static int enetc_pf_probe(struct pci_dev *pdev,
 	si->ndev = NULL;
 	free_netdev(ndev);
 err_alloc_netdev:
+err_device_disabled:
 err_map_pf_space:
 	enetc_pci_remove(pdev);
 
-- 
2.25.1

[PATCH net 3/6] net: enetc: take the MDIO lock only once per NAPI poll cycle

[Vladimir Oltean]

From: Vladimir Oltean <vladimir.oltean@nxp.com>

The workaround for the ENETC MDIO erratum caused a performance
degradation of 82 Kpps (seen with IP forwarding of two 1Gbps streams of
64B packets). This is due to excessive locking and unlocking in the fast
path, which can be avoided.

By taking the MDIO read-side lock only once per NAPI poll cycle, we are
able to regain 54 Kpps (65%) of the performance hit. The rest of the
performance degradation comes from the TX data path, but unfortunately
it doesn't look like we can optimize that away easily, even with
netdev_xmit_more(), there just isn't any skb batching done, to help with
taking the MDIO lock less often than once per packet.

Fixes: fd5736bf9f23 ("enetc: Workaround for MDIO register access issue")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
 drivers/net/ethernet/freescale/enetc/enetc.c  | 31 ++++++-------------
 .../net/ethernet/freescale/enetc/enetc_hw.h   |  2 ++
 2 files changed, 11 insertions(+), 22 deletions(-)

diff --git a/drivers/net/ethernet/freescale/enetc/enetc.c b/drivers/net/ethernet/freescale/enetc/enetc.c
index 43f0fae30080..eebe08a99270 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc.c
@@ -281,6 +281,8 @@ static int enetc_poll(struct napi_struct *napi, int budget)
 	int work_done;
 	int i;
 
+	enetc_lock_mdio();
+
 	for (i = 0; i < v->count_tx_rings; i++)
 		if (!enetc_clean_tx_ring(&v->tx_ring[i], budget))
 			complete = false;
@@ -291,8 +293,10 @@ static int enetc_poll(struct napi_struct *napi, int budget)
 	if (work_done)
 		v->rx_napi_work = true;
 
-	if (!complete)
+	if (!complete) {
+		enetc_unlock_mdio();
 		return budget;
+	}
 
 	napi_complete_done(napi, work_done);
 
@@ -301,8 +305,6 @@ static int enetc_poll(struct napi_struct *napi, int budget)
 
 	v->rx_napi_work = false;
 
-	enetc_lock_mdio();
-
 	/* enable interrupts */
 	enetc_wr_reg_hot(v->rbier, ENETC_RBIER_RXTIE);
 
@@ -327,8 +329,8 @@ static void enetc_get_tx_tstamp(struct enetc_hw *hw, union enetc_tx_bd *txbd,
 {
 	u32 lo, hi, tstamp_lo;
 
-	lo = enetc_rd(hw, ENETC_SICTR0);
-	hi = enetc_rd(hw, ENETC_SICTR1);
+	lo = enetc_rd_hot(hw, ENETC_SICTR0);
+	hi = enetc_rd_hot(hw, ENETC_SICTR1);
 	tstamp_lo = le32_to_cpu(txbd->wb.tstamp);
 	if (lo <= tstamp_lo)
 		hi -= 1;
@@ -358,9 +360,7 @@ static bool enetc_clean_tx_ring(struct enetc_bdr *tx_ring, int napi_budget)
 	i = tx_ring->next_to_clean;
 	tx_swbd = &tx_ring->tx_swbd[i];
 
-	enetc_lock_mdio();
 	bds_to_clean = enetc_bd_ready_count(tx_ring, i);
-	enetc_unlock_mdio();
 
 	do_tstamp = false;
 
@@ -403,8 +403,6 @@ static bool enetc_clean_tx_ring(struct enetc_bdr *tx_ring, int napi_budget)
 			tx_swbd = tx_ring->tx_swbd;
 		}
 
-		enetc_lock_mdio();
-
 		/* BD iteration loop end */
 		if (is_eof) {
 			tx_frm_cnt++;
@@ -415,8 +413,6 @@ static bool enetc_clean_tx_ring(struct enetc_bdr *tx_ring, int napi_budget)
 
 		if (unlikely(!bds_to_clean))
 			bds_to_clean = enetc_bd_ready_count(tx_ring, i);
-
-		enetc_unlock_mdio();
 	}
 
 	tx_ring->next_to_clean = i;
@@ -660,8 +656,6 @@ static int enetc_clean_rx_ring(struct enetc_bdr *rx_ring,
 		u32 bd_status;
 		u16 size;
 
-		enetc_lock_mdio();
-
 		if (cleaned_cnt >= ENETC_RXBD_BUNDLE) {
 			int count = enetc_refill_rx_ring(rx_ring, cleaned_cnt);
 
@@ -672,19 +666,15 @@ static int enetc_clean_rx_ring(struct enetc_bdr *rx_ring,
 
 		rxbd = enetc_rxbd(rx_ring, i);
 		bd_status = le32_to_cpu(rxbd->r.lstatus);
-		if (!bd_status) {
-			enetc_unlock_mdio();
+		if (!bd_status)
 			break;
-		}
 
 		enetc_wr_reg_hot(rx_ring->idr, BIT(rx_ring->index));
 		dma_rmb(); /* for reading other rxbd fields */
 		size = le16_to_cpu(rxbd->r.buf_len);
 		skb = enetc_map_rx_buff_to_skb(rx_ring, i, size);
-		if (!skb) {
-			enetc_unlock_mdio();
+		if (!skb)
 			break;
-		}
 
 		enetc_get_offloads(rx_ring, rxbd, skb);
 
@@ -696,7 +686,6 @@ static int enetc_clean_rx_ring(struct enetc_bdr *rx_ring,
 
 		if (unlikely(bd_status &
 			     ENETC_RXBD_LSTATUS(ENETC_RXBD_ERR_MASK))) {
-			enetc_unlock_mdio();
 			dev_kfree_skb(skb);
 			while (!(bd_status & ENETC_RXBD_LSTATUS_F)) {
 				dma_rmb();
@@ -736,8 +725,6 @@ static int enetc_clean_rx_ring(struct enetc_bdr *rx_ring,
 
 		enetc_process_skb(rx_ring, skb);
 
-		enetc_unlock_mdio();
-
 		napi_gro_receive(napi, skb);
 
 		rx_frm_cnt++;
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_hw.h b/drivers/net/ethernet/freescale/enetc/enetc_hw.h
index c71fe8d751d5..8b54562f5da6 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_hw.h
+++ b/drivers/net/ethernet/freescale/enetc/enetc_hw.h
@@ -453,6 +453,8 @@ static inline u64 _enetc_rd_reg64_wa(void __iomem *reg)
 #define enetc_wr_reg(reg, val)		_enetc_wr_reg_wa((reg), (val))
 #define enetc_rd(hw, off)		enetc_rd_reg((hw)->reg + (off))
 #define enetc_wr(hw, off, val)		enetc_wr_reg((hw)->reg + (off), val)
+#define enetc_rd_hot(hw, off)		enetc_rd_reg_hot((hw)->reg + (off))
+#define enetc_wr_hot(hw, off, val)	enetc_wr_reg_hot((hw)->reg + (off), val)
 #define enetc_rd64(hw, off)		_enetc_rd_reg64_wa((hw)->reg + (off))
 /* port register accessors - PF only */
 #define enetc_port_rd(hw, off)		enetc_rd_reg((hw)->port + (off))
-- 
2.25.1

[PATCH net 4/6] net: enetc: fix incorrect TPID when receiving 802.1ad tagged packets

[Vladimir Oltean]

From: Vladimir Oltean <vladimir.oltean@nxp.com>

When the enetc ports have rx-vlan-offload enabled, they report a TPID of
ETH_P_8021Q regardless of what was actually in the packet. When
rx-vlan-offload is disabled, packets have the proper TPID. Fix this
inconsistency by finishing the TODO left in the code.

Fixes: d4fd0404c1c9 ("enetc: Introduce basic PF and VF ENETC ethernet drivers")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
 drivers/net/ethernet/freescale/enetc/enetc.c  | 31 +++++++++++++++----
 .../net/ethernet/freescale/enetc/enetc_hw.h   |  3 ++
 2 files changed, 28 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/freescale/enetc/enetc.c b/drivers/net/ethernet/freescale/enetc/enetc.c
index eebe08a99270..25284bf01c16 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc.c
@@ -534,12 +534,31 @@ static void enetc_get_offloads(struct enetc_bdr *rx_ring,
 		skb->ip_summed = CHECKSUM_COMPLETE;
 	}
 
-	/* copy VLAN to skb, if one is extracted, for now we assume it's a
-	 * standard TPID, but HW also supports custom values
-	 */
-	if (le16_to_cpu(rxbd->r.flags) & ENETC_RXBD_FLAG_VLAN)
-		__vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q),
-				       le16_to_cpu(rxbd->r.vlan_opt));
+	if (le16_to_cpu(rxbd->r.flags) & ENETC_RXBD_FLAG_VLAN) {
+		__be16 tpid = 0;
+
+		switch (le16_to_cpu(rxbd->r.flags) & ENETC_RXBD_FLAG_TPID) {
+		case 0:
+			tpid = htons(ETH_P_8021Q);
+			break;
+		case 1:
+			tpid = htons(ETH_P_8021AD);
+			break;
+		case 2:
+			tpid = htons(enetc_port_rd(&priv->si->hw,
+						   ENETC_PCVLANR1));
+			break;
+		case 3:
+			tpid = htons(enetc_port_rd(&priv->si->hw,
+						   ENETC_PCVLANR2));
+			break;
+		default:
+			break;
+		}
+
+		__vlan_hwaccel_put_tag(skb, tpid, le16_to_cpu(rxbd->r.vlan_opt));
+	}
+
 #ifdef CONFIG_FSL_ENETC_PTP_CLOCK
 	if (priv->active_offloads & ENETC_F_RX_TSTAMP)
 		enetc_get_rx_tstamp(rx_ring->ndev, rxbd, skb);
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_hw.h b/drivers/net/ethernet/freescale/enetc/enetc_hw.h
index 8b54562f5da6..a62604a1e54e 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_hw.h
+++ b/drivers/net/ethernet/freescale/enetc/enetc_hw.h
@@ -172,6 +172,8 @@ enum enetc_bdr_type {TX, RX};
 #define ENETC_PSIPMAR0(n)	(0x0100 + (n) * 0x8) /* n = SI index */
 #define ENETC_PSIPMAR1(n)	(0x0104 + (n) * 0x8)
 #define ENETC_PVCLCTR		0x0208
+#define ENETC_PCVLANR1		0x0210
+#define ENETC_PCVLANR2		0x0214
 #define ENETC_VLAN_TYPE_C	BIT(0)
 #define ENETC_VLAN_TYPE_S	BIT(1)
 #define ENETC_PVCLCTR_OVTPIDL(bmp)	((bmp) & 0xff) /* VLAN_TYPE */
@@ -570,6 +572,7 @@ union enetc_rx_bd {
 #define ENETC_RXBD_LSTATUS(flags)	((flags) << 16)
 #define ENETC_RXBD_FLAG_VLAN	BIT(9)
 #define ENETC_RXBD_FLAG_TSTMP	BIT(10)
+#define ENETC_RXBD_FLAG_TPID	GENMASK(1, 0)
 
 #define ENETC_MAC_ADDR_FILT_CNT	8 /* # of supported entries per port */
 #define EMETC_MAC_ADDR_FILT_RES	3 /* # of reserved entries at the beginning */
-- 
2.25.1

[PATCH net 5/6] net: enetc: don't disable VLAN filtering in IFF_PROMISC mode

[Vladimir Oltean]

From: Vladimir Oltean <vladimir.oltean@nxp.com>

Quoting from the blamed commit:

    In promiscuous mode, it is more intuitive that all traffic is received,
    including VLAN tagged traffic. It appears that it is necessary to set
    the flag in PSIPVMR for that to be the case, so VLAN promiscuous mode is
    also temporarily enabled. On exit from promiscuous mode, the setting
    made by ethtool is restored.

Intuitive or not, there isn't any definition issued by a standards body
which says that promiscuity has anything to do with VLAN filtering - it
only has to do with accepting packets regardless of destination MAC address.

In fact people are already trying to use this misunderstanding/bug of
the enetc driver as a justification to transform promiscuity into
something it never was about: accepting every packet (maybe that would
be the "rx-all" netdev feature?):
https://lore.kernel.org/netdev/20201110153958.ci5ekor3o2ekg3ky@ipetronik.com/

So we should avoid that and delete the bogus logic in enetc.

Fixes: 7070eea5e95a ("enetc: permit configuration of rx-vlan-filter with ethtool")
Cc: Markus Blöchl <Markus.Bloechl@ipetronik.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
 drivers/net/ethernet/freescale/enetc/enetc_pf.c | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pf.c b/drivers/net/ethernet/freescale/enetc/enetc_pf.c
index 62ba4bf56f0d..49681a0566ed 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_pf.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_pf.c
@@ -190,7 +190,6 @@ static void enetc_pf_set_rx_mode(struct net_device *ndev)
 {
 	struct enetc_ndev_priv *priv = netdev_priv(ndev);
 	struct enetc_pf *pf = enetc_si_priv(priv->si);
-	char vlan_promisc_simap = pf->vlan_promisc_simap;
 	struct enetc_hw *hw = &priv->si->hw;
 	bool uprom = false, mprom = false;
 	struct enetc_mac_filter *filter;
@@ -203,16 +202,12 @@ static void enetc_pf_set_rx_mode(struct net_device *ndev)
 		psipmr = ENETC_PSIPMR_SET_UP(0) | ENETC_PSIPMR_SET_MP(0);
 		uprom = true;
 		mprom = true;
-		/* Enable VLAN promiscuous mode for SI0 (PF) */
-		vlan_promisc_simap |= BIT(0);
 	} else if (ndev->flags & IFF_ALLMULTI) {
 		/* enable multi cast promisc mode for SI0 (PF) */
 		psipmr = ENETC_PSIPMR_SET_MP(0);
 		mprom = true;
 	}
 
-	enetc_set_vlan_promisc(&pf->si->hw, vlan_promisc_simap);
-
 	/* first 2 filter entries belong to PF */
 	if (!uprom) {
 		/* Update unicast filters */
-- 
2.25.1

[PATCH net 6/6] net: enetc: force the RGMII speed and duplex instead of operating in inband mode

[Vladimir Oltean]

From: Vladimir Oltean <vladimir.oltean@nxp.com>

The ENETC port 0 MAC supports in-band status signaling coming from a PHY
when operating in RGMII mode, and this feature is enabled by default.

It has been reported that RGMII is broken in fixed-link, and that is not
surprising considering the fact that no PHY is attached to the MAC in
that case, but a switch.

This brings us to the topic of the patch: the enetc driver should have
not enabled the optional in-band status signaling for RGMII unconditionally,
but should have forced the speed and duplex to what was resolved by
phylink.

Note that phylink does not accept the RGMII modes as valid for in-band
signaling, and these operate a bit differently than 1000base-x and SGMII
(notably there is no clause 37 state machine so no ACK required from the
MAC, instead the PHY sends extra code words on RXD[3:0] whenever it is
not transmitting something else, so it should be safe to leave a PHY
with this option unconditionally enabled even if we ignore it). The spec
talks about this here:
https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/138/RGMIIv1_5F00_3.pdf

Fixes: 71b77a7a27a3 ("enetc: Migrate to PHYLINK and PCS_LYNX")
Cc: Florian Fainelli <f.fainelli@gmail.com>
Cc: Andrew Lunn <andrew@lunn.ch>
Cc: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
 .../net/ethernet/freescale/enetc/enetc_hw.h   | 13 ++++-
 .../net/ethernet/freescale/enetc/enetc_pf.c   | 58 ++++++++++++++++---
 2 files changed, 59 insertions(+), 12 deletions(-)

diff --git a/drivers/net/ethernet/freescale/enetc/enetc_hw.h b/drivers/net/ethernet/freescale/enetc/enetc_hw.h
index a62604a1e54e..de0d20b0f489 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_hw.h
+++ b/drivers/net/ethernet/freescale/enetc/enetc_hw.h
@@ -238,10 +238,17 @@ enum enetc_bdr_type {TX, RX};
 #define ENETC_PM_IMDIO_BASE	0x8030
 
 #define ENETC_PM0_IF_MODE	0x8300
-#define ENETC_PMO_IFM_RG	BIT(2)
+#define ENETC_PM0_IFM_RG	BIT(2)
 #define ENETC_PM0_IFM_RLP	(BIT(5) | BIT(11))
-#define ENETC_PM0_IFM_RGAUTO	(BIT(15) | ENETC_PMO_IFM_RG | BIT(1))
-#define ENETC_PM0_IFM_XGMII	BIT(12)
+#define ENETC_PM0_IFM_EN_AUTO	BIT(15)
+#define ENETC_PM0_IFM_SSP_MASK	GENMASK(14, 13)
+#define ENETC_PM0_IFM_SSP_1000	(2 << 13)
+#define ENETC_PM0_IFM_SSP_100	(0 << 13)
+#define ENETC_PM0_IFM_SSP_10	(1 << 13)
+#define ENETC_PM0_IFM_FULL_DPX	BIT(12)
+#define ENETC_PM0_IFM_IFMODE_MASK GENMASK(1, 0)
+#define ENETC_PM0_IFM_IFMODE_XGMII 0
+#define ENETC_PM0_IFM_IFMODE_GMII 2
 #define ENETC_PSIDCAPR		0x1b08
 #define ENETC_PSIDCAPR_MSK	GENMASK(15, 0)
 #define ENETC_PSFCAPR		0x1b18
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pf.c b/drivers/net/ethernet/freescale/enetc/enetc_pf.c
index 49681a0566ed..796f19b00e1a 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_pf.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_pf.c
@@ -315,7 +315,7 @@ static void enetc_set_loopback(struct net_device *ndev, bool en)
 	u32 reg;
 
 	reg = enetc_port_rd(hw, ENETC_PM0_IF_MODE);
-	if (reg & ENETC_PMO_IFM_RG) {
+	if (reg & ENETC_PM0_IFM_RG) {
 		/* RGMII mode */
 		reg = (reg & ~ENETC_PM0_IFM_RLP) |
 		      (en ? ENETC_PM0_IFM_RLP : 0);
@@ -492,15 +492,26 @@ static void enetc_configure_port_mac(struct enetc_hw *hw)
 		      ENETC_PM0_CMD_TXP	| ENETC_PM0_PROMISC);
 }
 
-static void enetc_mac_config(struct enetc_hw *hw, phy_interface_t phy_mode)
+static void enetc_mac_config(struct enetc_hw *hw, phy_interface_t phy_mode,
+			     unsigned int mode)
 {
-	/* set auto-speed for RGMII */
-	if (enetc_port_rd(hw, ENETC_PM0_IF_MODE) & ENETC_PMO_IFM_RG ||
-	    phy_interface_mode_is_rgmii(phy_mode))
-		enetc_port_wr(hw, ENETC_PM0_IF_MODE, ENETC_PM0_IFM_RGAUTO);
+	u32 val;
+
+	if (phy_interface_mode_is_rgmii(phy_mode)) {
+		val = enetc_port_rd(hw, ENETC_PM0_IF_MODE);
+		val &= ENETC_PM0_IFM_IFMODE_MASK;
+		val |= ENETC_PM0_IFM_IFMODE_GMII | ENETC_PM0_IFM_RG;
+		if (phylink_autoneg_inband(mode))
+			val |= ENETC_PM0_IFM_EN_AUTO;
+		else
+			val &= ~ENETC_PM0_IFM_EN_AUTO;
+		enetc_port_wr(hw, ENETC_PM0_IF_MODE, val);
+	}
 
-	if (phy_mode == PHY_INTERFACE_MODE_USXGMII)
-		enetc_port_wr(hw, ENETC_PM0_IF_MODE, ENETC_PM0_IFM_XGMII);
+	if (phy_mode == PHY_INTERFACE_MODE_USXGMII) {
+		val = ENETC_PM0_IFM_FULL_DPX | ENETC_PM0_IFM_IFMODE_XGMII;
+		enetc_port_wr(hw, ENETC_PM0_IF_MODE, val);
+	}
 }
 
 static void enetc_mac_enable(struct enetc_hw *hw, bool en)
@@ -925,13 +936,38 @@ static void enetc_pl_mac_config(struct phylink_config *config,
 	struct enetc_pf *pf = phylink_to_enetc_pf(config);
 	struct enetc_ndev_priv *priv;
 
-	enetc_mac_config(&pf->si->hw, state->interface);
+	enetc_mac_config(&pf->si->hw, state->interface, mode);
 
 	priv = netdev_priv(pf->si->ndev);
 	if (pf->pcs)
 		phylink_set_pcs(priv->phylink, &pf->pcs->pcs);
 }
 
+static void enetc_force_rgmii_mac(struct enetc_hw *hw, int speed, int duplex)
+{
+	u32 val;
+
+	val = enetc_port_rd(hw, ENETC_PM0_IF_MODE);
+
+	if (speed == SPEED_1000) {
+		val &= ~ENETC_PM0_IFM_SSP_MASK;
+		val |= ENETC_PM0_IFM_SSP_1000;
+	} else if (speed == SPEED_100) {
+		val &= ~ENETC_PM0_IFM_SSP_MASK;
+		val |= ENETC_PM0_IFM_SSP_100;
+	} else if (speed == SPEED_10) {
+		val &= ~ENETC_PM0_IFM_SSP_MASK;
+		val |= ENETC_PM0_IFM_SSP_10;
+	}
+
+	if (duplex == DUPLEX_FULL)
+		val |= ENETC_PM0_IFM_FULL_DPX;
+	else
+		val &= ~ENETC_PM0_IFM_FULL_DPX;
+
+	enetc_port_wr(hw, ENETC_PM0_IF_MODE, val);
+}
+
 static void enetc_pl_mac_link_up(struct phylink_config *config,
 				 struct phy_device *phy, unsigned int mode,
 				 phy_interface_t interface, int speed,
@@ -945,6 +981,10 @@ static void enetc_pl_mac_link_up(struct phylink_config *config,
 		enetc_sched_speed_set(priv, speed);
 
 	enetc_mac_enable(&pf->si->hw, true);
+
+	if (!phylink_autoneg_inband(mode) &&
+	    phy_interface_mode_is_rgmii(interface))
+		enetc_force_rgmii_mac(&pf->si->hw, speed, duplex);
 }
 
 static void enetc_pl_mac_link_down(struct phylink_config *config,
-- 
2.25.1

Re: [PATCH net 6/6] net: enetc: force the RGMII speed and duplex instead of operating in inband mode

[Russell King - ARM Linux admin]

On Thu, Feb 25, 2021 at 01:23:57PM +0200, Vladimir Oltean wrote:
>  static void enetc_pl_mac_link_up(struct phylink_config *config,
>  				 struct phy_device *phy, unsigned int mode,
>  				 phy_interface_t interface, int speed,
> @@ -945,6 +981,10 @@ static void enetc_pl_mac_link_up(struct phylink_config *config,
>  		enetc_sched_speed_set(priv, speed);
>  
>  	enetc_mac_enable(&pf->si->hw, true);
> +
> +	if (!phylink_autoneg_inband(mode) &&
> +	    phy_interface_mode_is_rgmii(interface))
> +		enetc_force_rgmii_mac(&pf->si->hw, speed, duplex);

Does it matter that you're forcing the RGMII setup after having enabled
the MAC?

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!

Re: [PATCH net 6/6] net: enetc: force the RGMII speed and duplex instead of operating in inband mode

[Vladimir Oltean]

On Thu, Feb 25, 2021 at 11:29:42AM +0000, Russell King - ARM Linux admin wrote:
> On Thu, Feb 25, 2021 at 01:23:57PM +0200, Vladimir Oltean wrote:
> >  static void enetc_pl_mac_link_up(struct phylink_config *config,
> >  				 struct phy_device *phy, unsigned int mode,
> >  				 phy_interface_t interface, int speed,
> > @@ -945,6 +981,10 @@ static void enetc_pl_mac_link_up(struct phylink_config *config,
> >  		enetc_sched_speed_set(priv, speed);
> >  
> >  	enetc_mac_enable(&pf->si->hw, true);
> > +
> > +	if (!phylink_autoneg_inband(mode) &&
> > +	    phy_interface_mode_is_rgmii(interface))
> > +		enetc_force_rgmii_mac(&pf->si->hw, speed, duplex);
> 
> Does it matter that you're forcing the RGMII setup after having enabled
> the MAC?

Thanks, this is a good point. I had tested only at gigabit on the
LS1028A-QDS, but at 10/100 things seem to be broken if I update the
speed with the MAC being enabled. I'll send a v2 with this change and
also with a build failure fix for patch 4/6.