From: Alvin Šipraga <[email protected]>
This series adds support for Realtek's RTL8365MB-VC, a 4+1 port
10/100/1000M Ethernet switch. The driver - rtl8365mb - was developed by
Michael Ramussen and myself.
This version of the driver is relatively slim, implementing only the
standalone port functionality and no offload capabilities. It is based
on a previous RFC series [1] from August, and the main difference is the
removal of some spurious VLAN operations. Otherwise I have simply
addressed most of the feedback. Please see the respective patches for
more detail.
In parallel I am working on offloading the bridge layer capabilities,
but I would like to get the basic stuff upstreamed as soon as possible.
v2 -> v2:
- move IRQ setup earlier in probe per Florian's suggestion
- fix compilation error on some archs due to FIELD_PREP use in v1
- follow Jakub's suggestion and use the standard ethtool stats API;
NOTE: new patch in the series for relevant DSA plumbing
- following the stats change, it became apparent that the rtl8366
helper library is no longer that helpful; scrap it and implement
the ethtool ops specifically for this chip
v1 -> v2:
- drop DSA port type checks during MAC configuration
- use OF properties to configure RGMII TX/RX delay
- don't set default fwd_offload_mark if packet is trapped to CPU
- remove port mapping macros
- update device tree bindings documentation with an example
- cosmetic changes to the tagging driver using FIELD_* macros
[1] https://lore.kernel.org/netdev/[email protected]/
Alvin Šipraga (7):
ether: add EtherType for proprietary Realtek protocols
net: dsa: allow reporting of standard ethtool stats for slave devices
net: dsa: move NET_DSA_TAG_RTL4_A to right place in Kconfig/Makefile
dt-bindings: net: dsa: realtek-smi: document new compatible rtl8365mb
net: dsa: tag_rtl8_4: add realtek 8 byte protocol 4 tag
net: dsa: realtek-smi: add rtl8365mb subdriver for RTL8365MB-VC
net: phy: realtek: add support for RTL8365MB-VC internal PHYs
.../bindings/net/dsa/realtek-smi.txt | 87 +
drivers/net/dsa/Kconfig | 1 +
drivers/net/dsa/Makefile | 2 +-
drivers/net/dsa/realtek-smi-core.c | 4 +
drivers/net/dsa/realtek-smi-core.h | 1 +
drivers/net/dsa/rtl8365mb.c | 1981 +++++++++++++++++
drivers/net/phy/realtek.c | 8 +
include/net/dsa.h | 8 +
include/uapi/linux/if_ether.h | 1 +
net/dsa/Kconfig | 20 +-
net/dsa/Makefile | 3 +-
net/dsa/slave.c | 34 +
net/dsa/tag_rtl8_4.c | 185 ++
13 files changed, 2326 insertions(+), 9 deletions(-)
create mode 100644 drivers/net/dsa/rtl8365mb.c
create mode 100644 net/dsa/tag_rtl8_4.c
--
2.32.0
From: Alvin Šipraga <[email protected]>
Jakub pointed out that we have a new ethtool API for reporting device
statistics in a standardized way, via .get_eth_{phy,mac,ctrl}_stats.
Add a small amount of plumbing to allow DSA drivers to take advantage of
this when exposing statistics.
Suggested-by: Jakub Kicinski <[email protected]>
Signed-off-by: Alvin Šipraga <[email protected]>
---
v2 -> v3: this patch is new
include/net/dsa.h | 6 ++++++
net/dsa/slave.c | 34 ++++++++++++++++++++++++++++++++++
2 files changed, 40 insertions(+)
diff --git a/include/net/dsa.h b/include/net/dsa.h
index d784e76113b8..8fefd58ced8f 100644
--- a/include/net/dsa.h
+++ b/include/net/dsa.h
@@ -645,6 +645,12 @@ struct dsa_switch_ops {
int (*get_sset_count)(struct dsa_switch *ds, int port, int sset);
void (*get_ethtool_phy_stats)(struct dsa_switch *ds,
int port, uint64_t *data);
+ void (*get_eth_phy_stats)(struct dsa_switch *ds, int port,
+ struct ethtool_eth_phy_stats *phy_stats);
+ void (*get_eth_mac_stats)(struct dsa_switch *ds, int port,
+ struct ethtool_eth_mac_stats *mac_stats);
+ void (*get_eth_ctrl_stats)(struct dsa_switch *ds, int port,
+ struct ethtool_eth_ctrl_stats *ctrl_stats);
void (*get_stats64)(struct dsa_switch *ds, int port,
struct rtnl_link_stats64 *s);
void (*self_test)(struct dsa_switch *ds, int port,
diff --git a/net/dsa/slave.c b/net/dsa/slave.c
index 11ec9e689589..499f8d18c76d 100644
--- a/net/dsa/slave.c
+++ b/net/dsa/slave.c
@@ -789,6 +789,37 @@ static int dsa_slave_get_sset_count(struct net_device *dev, int sset)
return -EOPNOTSUPP;
}
+static void dsa_slave_get_eth_phy_stats(struct net_device *dev,
+ struct ethtool_eth_phy_stats *phy_stats)
+{
+ struct dsa_port *dp = dsa_slave_to_port(dev);
+ struct dsa_switch *ds = dp->ds;
+
+ if (ds->ops->get_eth_phy_stats)
+ ds->ops->get_eth_phy_stats(ds, dp->index, phy_stats);
+}
+
+static void dsa_slave_get_eth_mac_stats(struct net_device *dev,
+ struct ethtool_eth_mac_stats *mac_stats)
+{
+ struct dsa_port *dp = dsa_slave_to_port(dev);
+ struct dsa_switch *ds = dp->ds;
+
+ if (ds->ops->get_eth_mac_stats)
+ ds->ops->get_eth_mac_stats(ds, dp->index, mac_stats);
+}
+
+static void
+dsa_slave_get_eth_ctrl_stats(struct net_device *dev,
+ struct ethtool_eth_ctrl_stats *ctrl_stats)
+{
+ struct dsa_port *dp = dsa_slave_to_port(dev);
+ struct dsa_switch *ds = dp->ds;
+
+ if (ds->ops->get_eth_ctrl_stats)
+ ds->ops->get_eth_ctrl_stats(ds, dp->index, ctrl_stats);
+}
+
static void dsa_slave_net_selftest(struct net_device *ndev,
struct ethtool_test *etest, u64 *buf)
{
@@ -1695,6 +1726,9 @@ static const struct ethtool_ops dsa_slave_ethtool_ops = {
.get_strings = dsa_slave_get_strings,
.get_ethtool_stats = dsa_slave_get_ethtool_stats,
.get_sset_count = dsa_slave_get_sset_count,
+ .get_eth_phy_stats = dsa_slave_get_eth_phy_stats,
+ .get_eth_mac_stats = dsa_slave_get_eth_mac_stats,
+ .get_eth_ctrl_stats = dsa_slave_get_eth_ctrl_stats,
.set_wol = dsa_slave_set_wol,
.get_wol = dsa_slave_get_wol,
.set_eee = dsa_slave_set_eee,
--
2.32.0
From: Alvin Šipraga <[email protected]>
Add a new EtherType ETH_P_REALTEK to the if_ether.h uapi header. The
EtherType 0x8899 is used in a number of different protocols from Realtek
Semiconductor Corp [1], so no general assumptions should be made when
trying to decode such packets. Observed protocols include:
0x1 - Realtek Remote Control protocol [2]
0x2 - Echo protocol [2]
0x3 - Loop detection protocol [2]
0x4 - RTL8365MB 4- and 8-byte switch CPU tag protocols [3]
0x9 - RTL8306 switch CPU tag protocol [4]
0xA - RTL8366RB switch CPU tag protocol [4]
[1] https://lore.kernel.org/netdev/CACRpkdYQthFgjwVzHyK3DeYUOdcYyWmdjDPG=Rf9B3VrJ12Rzg@mail.gmail.com/
[2] https://www.wireshark.org/lists/ethereal-dev/200409/msg00090.html
[3] https://lore.kernel.org/netdev/[email protected]/
[4] https://lore.kernel.org/netdev/[email protected]/
Suggested-by: Andrew Lunn <[email protected]>
Signed-off-by: Alvin Šipraga <[email protected]>
Reviewed-by: Vladimir Oltean <[email protected]>
Reviewed-by: Florian Fainelli <[email protected]>
---
v2 -> v3: no change; collect Reviewed-by from Florian
v1 -> v2: no change; collect Reviewed-by from Vladimir
RFC -> v1: this patch is new
include/uapi/linux/if_ether.h | 1 +
1 file changed, 1 insertion(+)
diff --git a/include/uapi/linux/if_ether.h b/include/uapi/linux/if_ether.h
index 5f589c7a8382..5da4ee234e0b 100644
--- a/include/uapi/linux/if_ether.h
+++ b/include/uapi/linux/if_ether.h
@@ -86,6 +86,7 @@
* over Ethernet
*/
#define ETH_P_PAE 0x888E /* Port Access Entity (IEEE 802.1X) */
+#define ETH_P_REALTEK 0x8899 /* Multiple proprietary protocols */
#define ETH_P_AOE 0x88A2 /* ATA over Ethernet */
#define ETH_P_8021AD 0x88A8 /* 802.1ad Service VLAN */
#define ETH_P_802_EX1 0x88B5 /* 802.1 Local Experimental 1. */
--
2.32.0
From: Alvin Šipraga <[email protected]>
The RTL8365MB-VC ethernet switch controller has 4 internal PHYs for its
user-facing ports. All that is needed is to let the PHY driver core
pick up the IRQ made available by the switch driver.
Signed-off-by: Alvin Šipraga <[email protected]>
Reviewed-by: Florian Fainelli <[email protected]>
Reviewed-by: Linus Walleij <[email protected]>
---
v2 -> v3: no change
v1 -> v2: no change
RFC -> v1: no change; collect Reviewed-by
drivers/net/phy/realtek.c | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/drivers/net/phy/realtek.c b/drivers/net/phy/realtek.c
index 11be60333fa8..a5671ab896b3 100644
--- a/drivers/net/phy/realtek.c
+++ b/drivers/net/phy/realtek.c
@@ -1023,6 +1023,14 @@ static struct phy_driver realtek_drvs[] = {
.resume = genphy_resume,
.read_page = rtl821x_read_page,
.write_page = rtl821x_write_page,
+ }, {
+ PHY_ID_MATCH_EXACT(0x001cc942),
+ .name = "RTL8365MB-VC Gigabit Ethernet",
+ /* Interrupt handling analogous to RTL8366RB */
+ .config_intr = genphy_no_config_intr,
+ .handle_interrupt = genphy_handle_interrupt_no_ack,
+ .suspend = genphy_suspend,
+ .resume = genphy_resume,
},
};
--
2.32.0
From: Alvin Šipraga <[email protected]>
rtl8365mb is a new realtek-smi subdriver for the RTL8365MB-VC 4+1 port
10/100/1000M Ethernet switch controller. Its compatible string is
"realtek,rtl8365mb".
Signed-off-by: Alvin Šipraga <[email protected]>
Reviewed-by: Andrew Lunn <[email protected]>
Reviewed-by: Florian Fainelli <[email protected]>
Acked-by: Rob Herring <[email protected]>
Reviewed-by: Linus Walleij <[email protected]>
---
v2 -> v3: no change
v1 -> v2:
- add an example to highlight RGMII delay configuration and CPU port
reg number
RFC -> v1: no change; collect Reviewed-by and Acked-by
.../bindings/net/dsa/realtek-smi.txt | 87 +++++++++++++++++++
1 file changed, 87 insertions(+)
diff --git a/Documentation/devicetree/bindings/net/dsa/realtek-smi.txt b/Documentation/devicetree/bindings/net/dsa/realtek-smi.txt
index b6ae8541bd55..7959ec237983 100644
--- a/Documentation/devicetree/bindings/net/dsa/realtek-smi.txt
+++ b/Documentation/devicetree/bindings/net/dsa/realtek-smi.txt
@@ -9,6 +9,7 @@ SMI-based Realtek devices.
Required properties:
- compatible: must be exactly one of:
+ "realtek,rtl8365mb" (4+1 ports)
"realtek,rtl8366"
"realtek,rtl8366rb" (4+1 ports)
"realtek,rtl8366s" (4+1 ports)
@@ -62,6 +63,8 @@ and subnodes of DSA switches.
Examples:
+An example for the RTL8366RB:
+
switch {
compatible = "realtek,rtl8366rb";
/* 22 = MDIO (has input reads), 21 = MDC (clock, output only) */
@@ -151,3 +154,87 @@ switch {
};
};
};
+
+An example for the RTL8365MB-VC:
+
+switch {
+ compatible = "realtek,rtl8365mb";
+ mdc-gpios = <&gpio1 16 GPIO_ACTIVE_HIGH>;
+ mdio-gpios = <&gpio1 17 GPIO_ACTIVE_HIGH>;
+ reset-gpios = <&gpio5 0 GPIO_ACTIVE_LOW>;
+
+ switch_intc: interrupt-controller {
+ interrupt-parent = <&gpio5>;
+ interrupts = <1 IRQ_TYPE_LEVEL_LOW>;
+ interrupt-controller;
+ #address-cells = <0>;
+ #interrupt-cells = <1>;
+ };
+
+ ports {
+ #address-cells = <1>;
+ #size-cells = <0>;
+ reg = <0>;
+ port@0 {
+ reg = <0>;
+ label = "swp0";
+ phy-handle = <ðphy0>;
+ };
+ port@1 {
+ reg = <1>;
+ label = "swp1";
+ phy-handle = <ðphy1>;
+ };
+ port@2 {
+ reg = <2>;
+ label = "swp2";
+ phy-handle = <ðphy2>;
+ };
+ port@3 {
+ reg = <3>;
+ label = "swp3";
+ phy-handle = <ðphy3>;
+ };
+ port@6 {
+ reg = <6>;
+ label = "cpu";
+ ethernet = <&fec1>;
+ phy-mode = "rgmii";
+ tx-internal-delay-ps = <2000>;
+ rx-internal-delay-ps = <2000>;
+
+ fixed-link {
+ speed = <1000>;
+ full-duplex;
+ pause;
+ };
+ };
+ };
+
+ mdio {
+ compatible = "realtek,smi-mdio";
+ #address-cells = <1>;
+ #size-cells = <0>;
+
+ ethphy0: phy@0 {
+ reg = <0>;
+ interrupt-parent = <&switch_intc>;
+ interrupts = <0>;
+ };
+ ethphy1: phy@1 {
+ reg = <1>;
+ interrupt-parent = <&switch_intc>;
+ interrupts = <1>;
+ };
+ ethphy2: phy@2 {
+ reg = <2>;
+ interrupt-parent = <&switch_intc>;
+ interrupts = <2>;
+ };
+ ethphy3: phy@3 {
+ reg = <3>;
+ interrupt-parent = <&switch_intc>;
+ interrupts = <3>;
+ };
+ };
+};
--
2.32.0
From: Alvin Šipraga <[email protected]>
This commit implements a basic version of the 8 byte tag protocol used
in the Realtek RTL8365MB-VC unmanaged switch, which carries with it a
protocol version of 0x04.
The implementation itself only handles the parsing of the EtherType
value and Realtek protocol version, together with the source or
destination port fields. The rest is left unimplemented for now.
The tag format is described in a confidential document provided to my
company by Realtek Semiconductor Corp. Permission has been granted by
the vendor to publish this driver based on that material, together with
an extract from the document describing the tag format and its fields.
It is hoped that this will help future implementors who do not have
access to the material but who wish to extend the functionality of
drivers for chips which use this protocol.
In addition, two possible values of the REASON field are specified,
based on experiments on my end. Realtek does not specify what value this
field can take.
Signed-off-by: Alvin Šipraga <[email protected]>
Reviewed-by: Vladimir Oltean <[email protected]>
Reviewed-by: Linus Walleij <[email protected]>
Reviewed-by: Florian Fainelli <[email protected]>
---
v2 -> v3:
- add bitfield.h include
- collect Reviewed-by from Florian
v1 -> v2:
- collect Reviewed-by from Vladimir and Linus
- don't set default fwd_offload_mark if packet is trapped to CPU
- use #defines and FIELD_* macros to make Linus happy
RFC -> v1:
- minor changes to the big comment at the top, including some
empirical information about the REASON code
- use dev_*_ratelimited() instead of netdev_*() for logging
- use warning instead of debug messages
- use ETH_P_REALTEK from if_ether.h
- set LEARN_DIS on xmit
- remove superfluous variables/expressions and use __b16 for tag
variable
- use new helper functions to insert/remove CPU tag
- set offload_fwd_mark properly using helper function
include/net/dsa.h | 2 +
net/dsa/Kconfig | 6 ++
net/dsa/Makefile | 1 +
net/dsa/tag_rtl8_4.c | 185 +++++++++++++++++++++++++++++++++++++++++++
4 files changed, 194 insertions(+)
create mode 100644 net/dsa/tag_rtl8_4.c
diff --git a/include/net/dsa.h b/include/net/dsa.h
index 8fefd58ced8f..05ebdd8d5321 100644
--- a/include/net/dsa.h
+++ b/include/net/dsa.h
@@ -51,6 +51,7 @@ struct phylink_link_state;
#define DSA_TAG_PROTO_SEVILLE_VALUE 21
#define DSA_TAG_PROTO_BRCM_LEGACY_VALUE 22
#define DSA_TAG_PROTO_SJA1110_VALUE 23
+#define DSA_TAG_PROTO_RTL8_4_VALUE 24
enum dsa_tag_protocol {
DSA_TAG_PROTO_NONE = DSA_TAG_PROTO_NONE_VALUE,
@@ -77,6 +78,7 @@ enum dsa_tag_protocol {
DSA_TAG_PROTO_OCELOT_8021Q = DSA_TAG_PROTO_OCELOT_8021Q_VALUE,
DSA_TAG_PROTO_SEVILLE = DSA_TAG_PROTO_SEVILLE_VALUE,
DSA_TAG_PROTO_SJA1110 = DSA_TAG_PROTO_SJA1110_VALUE,
+ DSA_TAG_PROTO_RTL8_4 = DSA_TAG_PROTO_RTL8_4_VALUE,
};
struct dsa_switch;
diff --git a/net/dsa/Kconfig b/net/dsa/Kconfig
index 3a09784dae63..8cb87b5067ee 100644
--- a/net/dsa/Kconfig
+++ b/net/dsa/Kconfig
@@ -126,6 +126,12 @@ config NET_DSA_TAG_RTL4_A
Realtek switches with 4 byte protocol A tags, sich as found in
the Realtek RTL8366RB.
+config NET_DSA_TAG_RTL8_4
+ tristate "Tag driver for Realtek 8 byte protocol 4 tags"
+ help
+ Say Y or M if you want to enable support for tagging frames for Realtek
+ switches with 8 byte protocol 4 tags, such as the Realtek RTL8365MB-VC.
+
config NET_DSA_TAG_LAN9303
tristate "Tag driver for SMSC/Microchip LAN9303 family of switches"
help
diff --git a/net/dsa/Makefile b/net/dsa/Makefile
index f78d537044db..9f75820e7c98 100644
--- a/net/dsa/Makefile
+++ b/net/dsa/Makefile
@@ -16,6 +16,7 @@ obj-$(CONFIG_NET_DSA_TAG_OCELOT) += tag_ocelot.o
obj-$(CONFIG_NET_DSA_TAG_OCELOT_8021Q) += tag_ocelot_8021q.o
obj-$(CONFIG_NET_DSA_TAG_QCA) += tag_qca.o
obj-$(CONFIG_NET_DSA_TAG_RTL4_A) += tag_rtl4_a.o
+obj-$(CONFIG_NET_DSA_TAG_RTL8_4) += tag_rtl8_4.o
obj-$(CONFIG_NET_DSA_TAG_SJA1105) += tag_sja1105.o
obj-$(CONFIG_NET_DSA_TAG_TRAILER) += tag_trailer.o
obj-$(CONFIG_NET_DSA_TAG_XRS700X) += tag_xrs700x.o
diff --git a/net/dsa/tag_rtl8_4.c b/net/dsa/tag_rtl8_4.c
new file mode 100644
index 000000000000..831adda84610
--- /dev/null
+++ b/net/dsa/tag_rtl8_4.c
@@ -0,0 +1,185 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Handler for Realtek 8 byte switch tags
+ *
+ * Copyright (C) 2021 Alvin Šipraga <[email protected]>
+ *
+ * NOTE: Currently only supports protocol "4" found in the RTL8365MB, hence
+ * named tag_rtl8_4.
+ *
+ * This tag header has the following format:
+ *
+ * -------------------------------------------
+ * | MAC DA | MAC SA | 8 byte tag | Type | ...
+ * -------------------------------------------
+ * _______________/ \______________________________________
+ * / \
+ * 0 7|8 15
+ * |-----------------------------------+-----------------------------------|---
+ * | (16-bit) | ^
+ * | Realtek EtherType [0x8899] | |
+ * |-----------------------------------+-----------------------------------| 8
+ * | (8-bit) | (8-bit) |
+ * | Protocol [0x04] | REASON | b
+ * |-----------------------------------+-----------------------------------| y
+ * | (1) | (1) | (2) | (1) | (3) | (1) | (1) | (1) | (5) | t
+ * | FID_EN | X | FID | PRI_EN | PRI | KEEP | X | LEARN_DIS | X | e
+ * |-----------------------------------+-----------------------------------| s
+ * | (1) | (15-bit) | |
+ * | ALLOW | TX/RX | v
+ * |-----------------------------------+-----------------------------------|---
+ *
+ * With the following field descriptions:
+ *
+ * field | description
+ * ------------+-------------
+ * Realtek | 0x8899: indicates that this is a proprietary Realtek tag;
+ * EtherType | note that Realtek uses the same EtherType for
+ * | other incompatible tag formats (e.g. tag_rtl4_a.c)
+ * Protocol | 0x04: indicates that this tag conforms to this format
+ * X | reserved
+ * ------------+-------------
+ * REASON | reason for forwarding packet to CPU
+ * | 0: packet was forwarded or flooded to CPU
+ * | 80: packet was trapped to CPU
+ * FID_EN | 1: packet has an FID
+ * | 0: no FID
+ * FID | FID of packet (if FID_EN=1)
+ * PRI_EN | 1: force priority of packet
+ * | 0: don't force priority
+ * PRI | priority of packet (if PRI_EN=1)
+ * KEEP | preserve packet VLAN tag format
+ * LEARN_DIS | don't learn the source MAC address of the packet
+ * ALLOW | 1: treat TX/RX field as an allowance port mask, meaning the
+ * | packet may only be forwarded to ports specified in the
+ * | mask
+ * | 0: no allowance port mask, TX/RX field is the forwarding
+ * | port mask
+ * TX/RX | TX (switch->CPU): port number the packet was received on
+ * | RX (CPU->switch): forwarding port mask (if ALLOW=0)
+ * | allowance port mask (if ALLOW=1)
+ */
+
+#include <linux/bitfield.h>
+#include <linux/bits.h>
+#include <linux/etherdevice.h>
+
+#include "dsa_priv.h"
+
+/* Protocols supported:
+ *
+ * 0x04 = RTL8365MB DSA protocol
+ */
+
+#define RTL8_4_TAG_LEN 8
+
+#define RTL8_4_PROTOCOL GENMASK(15, 8)
+#define RTL8_4_PROTOCOL_RTL8365MB 0x04
+#define RTL8_4_REASON GENMASK(7, 0)
+#define RTL8_4_REASON_FORWARD 0
+#define RTL8_4_REASON_TRAP 80
+
+#define RTL8_4_LEARN_DIS BIT(5)
+
+#define RTL8_4_TX GENMASK(3, 0)
+#define RTL8_4_RX GENMASK(10, 0)
+
+static struct sk_buff *rtl8_4_tag_xmit(struct sk_buff *skb,
+ struct net_device *dev)
+{
+ struct dsa_port *dp = dsa_slave_to_port(dev);
+ __be16 *tag;
+
+ /* Pad out so the (stripped) packet is at least 64 bytes long
+ * (including FCS), otherwise the switch will drop the packet.
+ * Then we need an additional 8 bytes for the Realtek tag.
+ */
+ if (unlikely(__skb_put_padto(skb, ETH_ZLEN + RTL8_4_TAG_LEN, false)))
+ return NULL;
+
+ skb_push(skb, RTL8_4_TAG_LEN);
+
+ dsa_alloc_etype_header(skb, RTL8_4_TAG_LEN);
+ tag = dsa_etype_header_pos_tx(skb);
+
+ /* Set Realtek EtherType */
+ tag[0] = htons(ETH_P_REALTEK);
+
+ /* Set Protocol; zero REASON */
+ tag[1] = htons(FIELD_PREP(RTL8_4_PROTOCOL, RTL8_4_PROTOCOL_RTL8365MB));
+
+ /* Zero FID_EN, FID, PRI_EN, PRI, KEEP; set LEARN_DIS */
+ tag[2] = htons(FIELD_PREP(RTL8_4_LEARN_DIS, 1));
+
+ /* Zero ALLOW; set RX (CPU->switch) forwarding port mask */
+ tag[3] = htons(FIELD_PREP(RTL8_4_RX, BIT(dp->index)));
+
+ return skb;
+}
+
+static struct sk_buff *rtl8_4_tag_rcv(struct sk_buff *skb,
+ struct net_device *dev)
+{
+ __be16 *tag;
+ u16 etype;
+ u8 reason;
+ u8 proto;
+ u8 port;
+
+ if (unlikely(!pskb_may_pull(skb, RTL8_4_TAG_LEN)))
+ return NULL;
+
+ tag = dsa_etype_header_pos_rx(skb);
+
+ /* Parse Realtek EtherType */
+ etype = ntohs(tag[0]);
+ if (unlikely(etype != ETH_P_REALTEK)) {
+ dev_warn_ratelimited(&dev->dev,
+ "non-realtek ethertype 0x%04x\n", etype);
+ return NULL;
+ }
+
+ /* Parse Protocol */
+ proto = FIELD_GET(RTL8_4_PROTOCOL, ntohs(tag[1]));
+ if (unlikely(proto != RTL8_4_PROTOCOL_RTL8365MB)) {
+ dev_warn_ratelimited(&dev->dev,
+ "unknown realtek protocol 0x%02x\n",
+ proto);
+ return NULL;
+ }
+
+ /* Parse REASON */
+ reason = FIELD_GET(RTL8_4_REASON, ntohs(tag[1]));
+
+ /* Parse TX (switch->CPU) */
+ port = FIELD_GET(RTL8_4_TX, ntohs(tag[3]));
+ skb->dev = dsa_master_find_slave(dev, 0, port);
+ if (!skb->dev) {
+ dev_warn_ratelimited(&dev->dev,
+ "could not find slave for port %d\n",
+ port);
+ return NULL;
+ }
+
+ /* Remove tag and recalculate checksum */
+ skb_pull_rcsum(skb, RTL8_4_TAG_LEN);
+
+ dsa_strip_etype_header(skb, RTL8_4_TAG_LEN);
+
+ if (reason != RTL8_4_REASON_TRAP)
+ dsa_default_offload_fwd_mark(skb);
+
+ return skb;
+}
+
+static const struct dsa_device_ops rtl8_4_netdev_ops = {
+ .name = "rtl8_4",
+ .proto = DSA_TAG_PROTO_RTL8_4,
+ .xmit = rtl8_4_tag_xmit,
+ .rcv = rtl8_4_tag_rcv,
+ .needed_headroom = RTL8_4_TAG_LEN,
+};
+module_dsa_tag_driver(rtl8_4_netdev_ops);
+
+MODULE_LICENSE("GPL");
+MODULE_ALIAS_DSA_TAG_DRIVER(DSA_TAG_PROTO_RTL8_4);
--
2.32.0
From: Alvin Šipraga <[email protected]>
This patch adds a realtek-smi subdriver for the RTL8365MB-VC 4+1 port
10/100/1000M switch controller. The driver has been developed based on a
GPL-licensed OS-agnostic Realtek vendor driver known as rtl8367c found
in the OpenWrt source tree.
Despite the name, the RTL8365MB-VC has an entirely different register
layout to the already-supported RTL8366RB ASIC. Notwithstanding this,
the structure of the rtl8365mb subdriver is loosely based on the rtl8366rb
subdriver. Like the 'rb, it establishes its own irqchip to handle
cascaded PHY link status interrupts.
The RTL8365MB-VC switch is capable of offloading a large number of
features from the software, but this patch introduces only the most
basic DSA driver functionality. The ports always function as standalone
ports, with bridging handled in software.
One more thing. Realtek's nomenclature for switches makes it hard to
know exactly what other ASICs might be supported by this driver. The
vendor driver goes by the name rtl8367c, but as far as I can tell, no
chip actually exists under this name. As such, the subdriver is named
rtl8365mb to emphasize the potentially limited support. But it is clear
from the vendor sources that a number of other more advanced switches
share a similar register layout, and further support should not be too
hard to add given access to the relevant hardware. With this in mind,
the subdriver has been written with as few assumptions about the
particular chip as is reasonable. But the RTL8365MB-VC is the only
hardware I have available, so some further work is surely needed.
Co-developed-by: Michael Rasmussen <[email protected]>
Signed-off-by: Michael Rasmussen <[email protected]>
Signed-off-by: Alvin Šipraga <[email protected]>
Reviewed-by: Vladimir Oltean <[email protected]>
Reviewed-by: Linus Walleij <[email protected]>
Reviewed-by: Florian Fainelli <[email protected]>
---
v2 -> v3:
- collect Florian's Reviewed-by
- move IRQ setup earlier in probe per Florian's suggestion
- follow Jakub's suggestion and use the standard ethtool stats API
- following the stats change, it became apparent that the rtl8366
helper library is no longer that helpful; scrap it and implement
the ethtool ops specifically for this chip
v1 -> v2:
- collect Vladimir's Reviewed-by
- drop DSA port type checks during MAC configuration
- use OF properties to configure RGMII TX/RX delay
- remove port mapping macros and clean up some private data that is no
longer needed as a result
RFC -> v1:
- clean up duplicate MASK #defines for interrupt registers
- removed 1000baseT_Half
- make port domain mapping macros call real functions
- add STP state setter
- set STP state DISABLED for all ports on setup
- disable learning on all ports on setup
- remove VLAN ops
- remove .port_{enable,disable} ops; no longer needed with STP state
- small cosmetic changes
drivers/net/dsa/Kconfig | 1 +
drivers/net/dsa/Makefile | 2 +-
drivers/net/dsa/realtek-smi-core.c | 4 +
drivers/net/dsa/realtek-smi-core.h | 1 +
drivers/net/dsa/rtl8365mb.c | 1981 ++++++++++++++++++++++++++++
5 files changed, 1988 insertions(+), 1 deletion(-)
create mode 100644 drivers/net/dsa/rtl8365mb.c
diff --git a/drivers/net/dsa/Kconfig b/drivers/net/dsa/Kconfig
index a5f1aa911fe2..7b1457a6e327 100644
--- a/drivers/net/dsa/Kconfig
+++ b/drivers/net/dsa/Kconfig
@@ -70,6 +70,7 @@ config NET_DSA_QCA8K
config NET_DSA_REALTEK_SMI
tristate "Realtek SMI Ethernet switch family support"
select NET_DSA_TAG_RTL4_A
+ select NET_DSA_TAG_RTL8_4
select FIXED_PHY
select IRQ_DOMAIN
select REALTEK_PHY
diff --git a/drivers/net/dsa/Makefile b/drivers/net/dsa/Makefile
index f3598c040994..8da1569a34e6 100644
--- a/drivers/net/dsa/Makefile
+++ b/drivers/net/dsa/Makefile
@@ -10,7 +10,7 @@ obj-$(CONFIG_NET_DSA_MT7530) += mt7530.o
obj-$(CONFIG_NET_DSA_MV88E6060) += mv88e6060.o
obj-$(CONFIG_NET_DSA_QCA8K) += qca8k.o
obj-$(CONFIG_NET_DSA_REALTEK_SMI) += realtek-smi.o
-realtek-smi-objs := realtek-smi-core.o rtl8366.o rtl8366rb.o
+realtek-smi-objs := realtek-smi-core.o rtl8366.o rtl8366rb.o rtl8365mb.o
obj-$(CONFIG_NET_DSA_SMSC_LAN9303) += lan9303-core.o
obj-$(CONFIG_NET_DSA_SMSC_LAN9303_I2C) += lan9303_i2c.o
obj-$(CONFIG_NET_DSA_SMSC_LAN9303_MDIO) += lan9303_mdio.o
diff --git a/drivers/net/dsa/realtek-smi-core.c b/drivers/net/dsa/realtek-smi-core.c
index 2fcfd917b876..c66ebd0ee217 100644
--- a/drivers/net/dsa/realtek-smi-core.c
+++ b/drivers/net/dsa/realtek-smi-core.c
@@ -501,6 +501,10 @@ static const struct of_device_id realtek_smi_of_match[] = {
.compatible = "realtek,rtl8366s",
.data = NULL,
},
+ {
+ .compatible = "realtek,rtl8365mb",
+ .data = &rtl8365mb_variant,
+ },
{ /* sentinel */ },
};
MODULE_DEVICE_TABLE(of, realtek_smi_of_match);
diff --git a/drivers/net/dsa/realtek-smi-core.h b/drivers/net/dsa/realtek-smi-core.h
index 214f710d7dd5..5bfa53e2480a 100644
--- a/drivers/net/dsa/realtek-smi-core.h
+++ b/drivers/net/dsa/realtek-smi-core.h
@@ -140,5 +140,6 @@ int rtl8366_get_sset_count(struct dsa_switch *ds, int port, int sset);
void rtl8366_get_ethtool_stats(struct dsa_switch *ds, int port, uint64_t *data);
extern const struct realtek_smi_variant rtl8366rb_variant;
+extern const struct realtek_smi_variant rtl8365mb_variant;
#endif /* _REALTEK_SMI_H */
diff --git a/drivers/net/dsa/rtl8365mb.c b/drivers/net/dsa/rtl8365mb.c
new file mode 100644
index 000000000000..45e996867c42
--- /dev/null
+++ b/drivers/net/dsa/rtl8365mb.c
@@ -0,0 +1,1981 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Realtek SMI subdriver for the Realtek RTL8365MB-VC ethernet switch.
+ *
+ * Copyright (C) 2021 Alvin Šipraga <[email protected]>
+ * Copyright (C) 2021 Michael Rasmussen <[email protected]>
+ *
+ * The RTL8365MB-VC is a 4+1 port 10/100/1000M switch controller. It includes 4
+ * integrated PHYs for the user facing ports, and an extension interface which
+ * can be connected to the CPU - or another PHY - via either MII, RMII, or
+ * RGMII. The switch is configured via the Realtek Simple Management Interface
+ * (SMI), which uses the MDIO/MDC lines.
+ *
+ * Below is a simplified block diagram of the chip and its relevant interfaces.
+ *
+ * .-----------------------------------.
+ * | |
+ * UTP <---------------> Giga PHY <-> PCS <-> P0 GMAC |
+ * UTP <---------------> Giga PHY <-> PCS <-> P1 GMAC |
+ * UTP <---------------> Giga PHY <-> PCS <-> P2 GMAC |
+ * UTP <---------------> Giga PHY <-> PCS <-> P3 GMAC |
+ * | |
+ * CPU/PHY <-MII/RMII/RGMII---> Extension <---> Extension |
+ * | interface 1 GMAC 1 |
+ * | |
+ * SMI driver/ <-MDC/SCL---> Management ~~~~~~~~~~~~~~ |
+ * EEPROM <-MDIO/SDA--> interface ~REALTEK ~~~~~ |
+ * | ~RTL8365MB ~~~ |
+ * | ~GXXXC TAIWAN~ |
+ * GPIO <--------------> Reset ~~~~~~~~~~~~~~ |
+ * | |
+ * Interrupt <----------> Link UP/DOWN events |
+ * controller | |
+ * '-----------------------------------'
+ *
+ * The driver uses DSA to integrate the 4 user and 1 extension ports into the
+ * kernel. Netdevices are created for the user ports, as are PHY devices for
+ * their integrated PHYs. The device tree firmware should also specify the link
+ * partner of the extension port - either via a fixed-link or other phy-handle.
+ * See the device tree bindings for more detailed information. Note that the
+ * driver has only been tested with a fixed-link, but in principle it should not
+ * matter.
+ *
+ * NOTE: Currently, only the RGMII interface is implemented in this driver.
+ *
+ * The interrupt line is asserted on link UP/DOWN events. The driver creates a
+ * custom irqchip to handle this interrupt and demultiplex the events by reading
+ * the status registers via SMI. Interrupts are then propagated to the relevant
+ * PHY device.
+ *
+ * The EEPROM contains initial register values which the chip will read over I2C
+ * upon hardware reset. It is also possible to omit the EEPROM. In both cases,
+ * the driver will manually reprogram some registers using jam tables to reach
+ * an initial state defined by the vendor driver.
+ *
+ * This Linux driver is written based on an OS-agnostic vendor driver from
+ * Realtek. The reference GPL-licensed sources can be found in the OpenWrt
+ * source tree under the name rtl8367c. The vendor driver claims to support a
+ * number of similar switch controllers from Realtek, but the only hardware we
+ * have is the RTL8365MB-VC. Moreover, there does not seem to be any chip under
+ * the name RTL8367C. Although one wishes that the 'C' stood for some kind of
+ * common hardware revision, there exist examples of chips with the suffix -VC
+ * which are explicitly not supported by the rtl8367c driver and which instead
+ * require the rtl8367d vendor driver. With all this uncertainty, the driver has
+ * been modestly named rtl8365mb. Future implementors may wish to rename things
+ * accordingly.
+ *
+ * In the same family of chips, some carry up to 8 user ports and up to 2
+ * extension ports. Where possible this driver tries to make things generic, but
+ * more work must be done to support these configurations. According to
+ * documentation from Realtek, the family should include the following chips:
+ *
+ * - RTL8363NB
+ * - RTL8363NB-VB
+ * - RTL8363SC
+ * - RTL8363SC-VB
+ * - RTL8364NB
+ * - RTL8364NB-VB
+ * - RTL8365MB-VC
+ * - RTL8366SC
+ * - RTL8367RB-VB
+ * - RTL8367SB
+ * - RTL8367S
+ * - RTL8370MB
+ * - RTL8310SR
+ *
+ * Some of the register logic for these additional chips has been skipped over
+ * while implementing this driver. It is therefore not possible to assume that
+ * things will work out-of-the-box for other chips, and a careful review of the
+ * vendor driver may be needed to expand support. The RTL8365MB-VC seems to be
+ * one of the simpler chips.
+ */
+
+#include <linux/bitfield.h>
+#include <linux/bitops.h>
+#include <linux/interrupt.h>
+#include <linux/irqdomain.h>
+#include <linux/of_irq.h>
+#include <linux/regmap.h>
+#include <linux/if_bridge.h>
+
+#include "realtek-smi-core.h"
+
+/* Chip-specific data and limits */
+#define RTL8365MB_CHIP_ID_8365MB_VC 0x6367
+#define RTL8365MB_CPU_PORT_NUM_8365MB_VC 6
+#define RTL8365MB_LEARN_LIMIT_MAX_8365MB_VC 2112
+
+/* Family-specific data and limits */
+#define RTL8365MB_NUM_PHYREGS 32
+#define RTL8365MB_PHYREGMAX (RTL8365MB_NUM_PHYREGS - 1)
+#define RTL8365MB_MAX_NUM_PORTS (RTL8365MB_CPU_PORT_NUM_8365MB_VC + 1)
+
+/* Chip identification registers */
+#define RTL8365MB_CHIP_ID_REG 0x1300
+
+#define RTL8365MB_CHIP_VER_REG 0x1301
+
+#define RTL8365MB_MAGIC_REG 0x13C2
+#define RTL8365MB_MAGIC_VALUE 0x0249
+
+/* Chip reset register */
+#define RTL8365MB_CHIP_RESET_REG 0x1322
+#define RTL8365MB_CHIP_RESET_SW_MASK 0x0002
+#define RTL8365MB_CHIP_RESET_HW_MASK 0x0001
+
+/* Interrupt polarity register */
+#define RTL8365MB_INTR_POLARITY_REG 0x1100
+#define RTL8365MB_INTR_POLARITY_MASK 0x0001
+#define RTL8365MB_INTR_POLARITY_HIGH 0
+#define RTL8365MB_INTR_POLARITY_LOW 1
+
+/* Interrupt control/status register - enable/check specific interrupt types */
+#define RTL8365MB_INTR_CTRL_REG 0x1101
+#define RTL8365MB_INTR_STATUS_REG 0x1102
+#define RTL8365MB_INTR_SLIENT_START_2_MASK 0x1000
+#define RTL8365MB_INTR_SLIENT_START_MASK 0x0800
+#define RTL8365MB_INTR_ACL_ACTION_MASK 0x0200
+#define RTL8365MB_INTR_CABLE_DIAG_FIN_MASK 0x0100
+#define RTL8365MB_INTR_INTERRUPT_8051_MASK 0x0080
+#define RTL8365MB_INTR_LOOP_DETECTION_MASK 0x0040
+#define RTL8365MB_INTR_GREEN_TIMER_MASK 0x0020
+#define RTL8365MB_INTR_SPECIAL_CONGEST_MASK 0x0010
+#define RTL8365MB_INTR_SPEED_CHANGE_MASK 0x0008
+#define RTL8365MB_INTR_LEARN_OVER_MASK 0x0004
+#define RTL8365MB_INTR_METER_EXCEEDED_MASK 0x0002
+#define RTL8365MB_INTR_LINK_CHANGE_MASK 0x0001
+#define RTL8365MB_INTR_ALL_MASK \
+ (RTL8365MB_INTR_SLIENT_START_2_MASK | \
+ RTL8365MB_INTR_SLIENT_START_MASK | \
+ RTL8365MB_INTR_ACL_ACTION_MASK | \
+ RTL8365MB_INTR_CABLE_DIAG_FIN_MASK | \
+ RTL8365MB_INTR_INTERRUPT_8051_MASK | \
+ RTL8365MB_INTR_LOOP_DETECTION_MASK | \
+ RTL8365MB_INTR_GREEN_TIMER_MASK | \
+ RTL8365MB_INTR_SPECIAL_CONGEST_MASK | \
+ RTL8365MB_INTR_SPEED_CHANGE_MASK | \
+ RTL8365MB_INTR_LEARN_OVER_MASK | \
+ RTL8365MB_INTR_METER_EXCEEDED_MASK | \
+ RTL8365MB_INTR_LINK_CHANGE_MASK)
+
+/* Per-port interrupt type status registers */
+#define RTL8365MB_PORT_LINKDOWN_IND_REG 0x1106
+#define RTL8365MB_PORT_LINKDOWN_IND_MASK 0x07FF
+
+#define RTL8365MB_PORT_LINKUP_IND_REG 0x1107
+#define RTL8365MB_PORT_LINKUP_IND_MASK 0x07FF
+
+/* PHY indirect access registers */
+#define RTL8365MB_INDIRECT_ACCESS_CTRL_REG 0x1F00
+#define RTL8365MB_INDIRECT_ACCESS_CTRL_RW_MASK 0x0002
+#define RTL8365MB_INDIRECT_ACCESS_CTRL_RW_READ 0
+#define RTL8365MB_INDIRECT_ACCESS_CTRL_RW_WRITE 1
+#define RTL8365MB_INDIRECT_ACCESS_CTRL_CMD_MASK 0x0001
+#define RTL8365MB_INDIRECT_ACCESS_CTRL_CMD_VALUE 1
+#define RTL8365MB_INDIRECT_ACCESS_STATUS_REG 0x1F01
+#define RTL8365MB_INDIRECT_ACCESS_ADDRESS_REG 0x1F02
+#define RTL8365MB_INDIRECT_ACCESS_ADDRESS_OCPADR_5_1_MASK GENMASK(4, 0)
+#define RTL8365MB_INDIRECT_ACCESS_ADDRESS_PHYNUM_MASK GENMASK(6, 5)
+#define RTL8365MB_INDIRECT_ACCESS_ADDRESS_OCPADR_9_6_MASK GENMASK(11, 8)
+#define RTL8365MB_PHY_BASE 0x2000
+#define RTL8365MB_INDIRECT_ACCESS_WRITE_DATA_REG 0x1F03
+#define RTL8365MB_INDIRECT_ACCESS_READ_DATA_REG 0x1F04
+
+/* PHY OCP address prefix register */
+#define RTL8365MB_GPHY_OCP_MSB_0_REG 0x1D15
+#define RTL8365MB_GPHY_OCP_MSB_0_CFG_CPU_OCPADR_MASK 0x0FC0
+#define RTL8365MB_PHY_OCP_ADDR_PREFIX_MASK 0xFC00
+
+/* The PHY OCP addresses of PHY registers 0~31 start here */
+#define RTL8365MB_PHY_OCP_ADDR_PHYREG_BASE 0xA400
+
+/* EXT port interface mode values - used in DIGITAL_INTERFACE_SELECT */
+#define RTL8365MB_EXT_PORT_MODE_DISABLE 0
+#define RTL8365MB_EXT_PORT_MODE_RGMII 1
+#define RTL8365MB_EXT_PORT_MODE_MII_MAC 2
+#define RTL8365MB_EXT_PORT_MODE_MII_PHY 3
+#define RTL8365MB_EXT_PORT_MODE_TMII_MAC 4
+#define RTL8365MB_EXT_PORT_MODE_TMII_PHY 5
+#define RTL8365MB_EXT_PORT_MODE_GMII 6
+#define RTL8365MB_EXT_PORT_MODE_RMII_MAC 7
+#define RTL8365MB_EXT_PORT_MODE_RMII_PHY 8
+#define RTL8365MB_EXT_PORT_MODE_SGMII 9
+#define RTL8365MB_EXT_PORT_MODE_HSGMII 10
+#define RTL8365MB_EXT_PORT_MODE_1000X_100FX 11
+#define RTL8365MB_EXT_PORT_MODE_1000X 12
+#define RTL8365MB_EXT_PORT_MODE_100FX 13
+
+/* EXT port interface mode configuration registers 0~1 */
+#define RTL8365MB_DIGITAL_INTERFACE_SELECT_REG0 0x1305
+#define RTL8365MB_DIGITAL_INTERFACE_SELECT_REG1 0x13C3
+#define RTL8365MB_DIGITAL_INTERFACE_SELECT_REG(_extport) \
+ (RTL8365MB_DIGITAL_INTERFACE_SELECT_REG0 + \
+ ((_extport) >> 1) * (0x13C3 - 0x1305))
+#define RTL8365MB_DIGITAL_INTERFACE_SELECT_MODE_MASK(_extport) \
+ (0xF << (((_extport) % 2)))
+#define RTL8365MB_DIGITAL_INTERFACE_SELECT_MODE_OFFSET(_extport) \
+ (((_extport) % 2) * 4)
+
+/* EXT port RGMII TX/RX delay configuration registers 1~2 */
+#define RTL8365MB_EXT_RGMXF_REG1 0x1307
+#define RTL8365MB_EXT_RGMXF_REG2 0x13C5
+#define RTL8365MB_EXT_RGMXF_REG(_extport) \
+ (RTL8365MB_EXT_RGMXF_REG1 + \
+ (((_extport) >> 1) * (0x13C5 - 0x1307)))
+#define RTL8365MB_EXT_RGMXF_RXDELAY_MASK 0x0007
+#define RTL8365MB_EXT_RGMXF_TXDELAY_MASK 0x0008
+
+/* External port speed values - used in DIGITAL_INTERFACE_FORCE */
+#define RTL8365MB_PORT_SPEED_10M 0
+#define RTL8365MB_PORT_SPEED_100M 1
+#define RTL8365MB_PORT_SPEED_1000M 2
+
+/* EXT port force configuration registers 0~2 */
+#define RTL8365MB_DIGITAL_INTERFACE_FORCE_REG0 0x1310
+#define RTL8365MB_DIGITAL_INTERFACE_FORCE_REG1 0x1311
+#define RTL8365MB_DIGITAL_INTERFACE_FORCE_REG2 0x13C4
+#define RTL8365MB_DIGITAL_INTERFACE_FORCE_REG(_extport) \
+ (RTL8365MB_DIGITAL_INTERFACE_FORCE_REG0 + \
+ ((_extport) & 0x1) + \
+ ((((_extport) >> 1) & 0x1) * (0x13C4 - 0x1310)))
+#define RTL8365MB_DIGITAL_INTERFACE_FORCE_EN_MASK 0x1000
+#define RTL8365MB_DIGITAL_INTERFACE_FORCE_NWAY_MASK 0x0080
+#define RTL8365MB_DIGITAL_INTERFACE_FORCE_TXPAUSE_MASK 0x0040
+#define RTL8365MB_DIGITAL_INTERFACE_FORCE_RXPAUSE_MASK 0x0020
+#define RTL8365MB_DIGITAL_INTERFACE_FORCE_LINK_MASK 0x0010
+#define RTL8365MB_DIGITAL_INTERFACE_FORCE_DUPLEX_MASK 0x0004
+#define RTL8365MB_DIGITAL_INTERFACE_FORCE_SPEED_MASK 0x0003
+
+/* CPU port mask register - controls which ports are treated as CPU ports */
+#define RTL8365MB_CPU_PORT_MASK_REG 0x1219
+#define RTL8365MB_CPU_PORT_MASK_MASK 0x07FF
+
+/* CPU control register */
+#define RTL8365MB_CPU_CTRL_REG 0x121A
+#define RTL8365MB_CPU_CTRL_TRAP_PORT_EXT_MASK 0x0400
+#define RTL8365MB_CPU_CTRL_TAG_FORMAT_MASK 0x0200
+#define RTL8365MB_CPU_CTRL_RXBYTECOUNT_MASK 0x0080
+#define RTL8365MB_CPU_CTRL_TAG_POSITION_MASK 0x0040
+#define RTL8365MB_CPU_CTRL_TRAP_PORT_MASK 0x0038
+#define RTL8365MB_CPU_CTRL_INSERTMODE_MASK 0x0006
+#define RTL8365MB_CPU_CTRL_EN_MASK 0x0001
+
+/* Maximum packet length register */
+#define RTL8365MB_CFG0_MAX_LEN_REG 0x088C
+#define RTL8365MB_CFG0_MAX_LEN_MASK 0x3FFF
+
+/* Port learning limit registers */
+#define RTL8365MB_LUT_PORT_LEARN_LIMIT_BASE 0x0A20
+#define RTL8365MB_LUT_PORT_LEARN_LIMIT_REG(_physport) \
+ (RTL8365MB_LUT_PORT_LEARN_LIMIT_BASE + (_physport))
+
+/* Port isolation (forwarding mask) registers */
+#define RTL8365MB_PORT_ISOLATION_REG_BASE 0x08A2
+#define RTL8365MB_PORT_ISOLATION_REG(_physport) \
+ (RTL8365MB_PORT_ISOLATION_REG_BASE + (_physport))
+#define RTL8365MB_PORT_ISOLATION_MASK 0x07FF
+
+/* MSTP port state registers - indexed by tree instancrSTI (tree ine */
+#define RTL8365MB_MSTI_CTRL_BASE 0x0A00
+#define RTL8365MB_MSTI_CTRL_REG(_msti, _physport) \
+ (RTL8365MB_MSTI_CTRL_BASE + ((_msti) << 1) + ((_physport) >> 3))
+#define RTL8365MB_MSTI_CTRL_PORT_STATE_OFFSET(_physport) ((_physport) << 1)
+#define RTL8365MB_MSTI_CTRL_PORT_STATE_MASK(_physport) \
+ (0x3 << RTL8365MB_MSTI_CTRL_PORT_STATE_OFFSET((_physport)))
+
+/* MIB counter value registers */
+#define RTL8365MB_MIB_COUNTER_BASE 0x1000
+#define RTL8365MB_MIB_COUNTER_REG(_x) (RTL8365MB_MIB_COUNTER_BASE + (_x))
+
+/* MIB counter address register */
+#define RTL8365MB_MIB_ADDRESS_REG 0x1004
+#define RTL8365MB_MIB_ADDRESS_PORT_OFFSET 0x007C
+#define RTL8365MB_MIB_ADDRESS(_p, _x) \
+ (((RTL8365MB_MIB_ADDRESS_PORT_OFFSET) * (_p) + (_x)) >> 2)
+
+#define RTL8365MB_MIB_CTRL0_REG 0x1005
+#define RTL8365MB_MIB_CTRL0_RESET_MASK 0x0002
+#define RTL8365MB_MIB_CTRL0_BUSY_MASK 0x0001
+
+/* The DSA callback .get_stats64 runs in atomic context, so we are not allowed
+ * to sleep. On the other hand, accessing MIB counters absolutely requires us to
+ * sleep. The solution is thus to schedule work which polls the MIB counters
+ * asynchronously and updates some private data, which the callback can then
+ * fetch atomically. Three seconds should be a good enough polling interval.
+ */
+#define RTL8365MB_STATS_INTERVAL_JIFFIES (3 * HZ)
+
+enum rtl8365mb_mib_counter_index {
+ RTL8365MB_MIB_ifInOctets,
+ RTL8365MB_MIB_dot3StatsFCSErrors,
+ RTL8365MB_MIB_dot3StatsSymbolErrors,
+ RTL8365MB_MIB_dot3InPauseFrames,
+ RTL8365MB_MIB_dot3ControlInUnknownOpcodes,
+ RTL8365MB_MIB_etherStatsFragments,
+ RTL8365MB_MIB_etherStatsJabbers,
+ RTL8365MB_MIB_ifInUcastPkts,
+ RTL8365MB_MIB_etherStatsDropEvents,
+ RTL8365MB_MIB_ifInMulticastPkts,
+ RTL8365MB_MIB_ifInBroadcastPkts,
+ RTL8365MB_MIB_inMldChecksumError,
+ RTL8365MB_MIB_inIgmpChecksumError,
+ RTL8365MB_MIB_inMldSpecificQuery,
+ RTL8365MB_MIB_inMldGeneralQuery,
+ RTL8365MB_MIB_inIgmpSpecificQuery,
+ RTL8365MB_MIB_inIgmpGeneralQuery,
+ RTL8365MB_MIB_inMldLeaves,
+ RTL8365MB_MIB_inIgmpLeaves,
+ RTL8365MB_MIB_etherStatsOctets,
+ RTL8365MB_MIB_etherStatsUnderSizePkts,
+ RTL8365MB_MIB_etherOversizeStats,
+ RTL8365MB_MIB_etherStatsPkts64Octets,
+ RTL8365MB_MIB_etherStatsPkts65to127Octets,
+ RTL8365MB_MIB_etherStatsPkts128to255Octets,
+ RTL8365MB_MIB_etherStatsPkts256to511Octets,
+ RTL8365MB_MIB_etherStatsPkts512to1023Octets,
+ RTL8365MB_MIB_etherStatsPkts1024to1518Octets,
+ RTL8365MB_MIB_ifOutOctets,
+ RTL8365MB_MIB_dot3StatsSingleCollisionFrames,
+ RTL8365MB_MIB_dot3StatsMultipleCollisionFrames,
+ RTL8365MB_MIB_dot3StatsDeferredTransmissions,
+ RTL8365MB_MIB_dot3StatsLateCollisions,
+ RTL8365MB_MIB_etherStatsCollisions,
+ RTL8365MB_MIB_dot3StatsExcessiveCollisions,
+ RTL8365MB_MIB_dot3OutPauseFrames,
+ RTL8365MB_MIB_ifOutDiscards,
+ RTL8365MB_MIB_dot1dTpPortInDiscards,
+ RTL8365MB_MIB_ifOutUcastPkts,
+ RTL8365MB_MIB_ifOutMulticastPkts,
+ RTL8365MB_MIB_ifOutBroadcastPkts,
+ RTL8365MB_MIB_outOampduPkts,
+ RTL8365MB_MIB_inOampduPkts,
+ RTL8365MB_MIB_inIgmpJoinsSuccess,
+ RTL8365MB_MIB_inIgmpJoinsFail,
+ RTL8365MB_MIB_inMldJoinsSuccess,
+ RTL8365MB_MIB_inMldJoinsFail,
+ RTL8365MB_MIB_inReportSuppressionDrop,
+ RTL8365MB_MIB_inLeaveSuppressionDrop,
+ RTL8365MB_MIB_outIgmpReports,
+ RTL8365MB_MIB_outIgmpLeaves,
+ RTL8365MB_MIB_outIgmpGeneralQuery,
+ RTL8365MB_MIB_outIgmpSpecificQuery,
+ RTL8365MB_MIB_outMldReports,
+ RTL8365MB_MIB_outMldLeaves,
+ RTL8365MB_MIB_outMldGeneralQuery,
+ RTL8365MB_MIB_outMldSpecificQuery,
+ RTL8365MB_MIB_inKnownMulticastPkts,
+ RTL8365MB_MIB_END,
+};
+
+struct rtl8365mb_mib_counter {
+ u32 offset;
+ u32 length;
+ const char *name;
+};
+
+#define RTL8365MB_MAKE_MIB_COUNTER(_offset, _length, _name) \
+ [RTL8365MB_MIB_ ## _name] = { _offset, _length, #_name }
+
+static struct rtl8365mb_mib_counter rtl8365mb_mib_counters[] = {
+ RTL8365MB_MAKE_MIB_COUNTER(0, 4, ifInOctets),
+ RTL8365MB_MAKE_MIB_COUNTER(4, 2, dot3StatsFCSErrors),
+ RTL8365MB_MAKE_MIB_COUNTER(6, 2, dot3StatsSymbolErrors),
+ RTL8365MB_MAKE_MIB_COUNTER(8, 2, dot3InPauseFrames),
+ RTL8365MB_MAKE_MIB_COUNTER(10, 2, dot3ControlInUnknownOpcodes),
+ RTL8365MB_MAKE_MIB_COUNTER(12, 2, etherStatsFragments),
+ RTL8365MB_MAKE_MIB_COUNTER(14, 2, etherStatsJabbers),
+ RTL8365MB_MAKE_MIB_COUNTER(16, 2, ifInUcastPkts),
+ RTL8365MB_MAKE_MIB_COUNTER(18, 2, etherStatsDropEvents),
+ RTL8365MB_MAKE_MIB_COUNTER(20, 2, ifInMulticastPkts),
+ RTL8365MB_MAKE_MIB_COUNTER(22, 2, ifInBroadcastPkts),
+ RTL8365MB_MAKE_MIB_COUNTER(24, 2, inMldChecksumError),
+ RTL8365MB_MAKE_MIB_COUNTER(26, 2, inIgmpChecksumError),
+ RTL8365MB_MAKE_MIB_COUNTER(28, 2, inMldSpecificQuery),
+ RTL8365MB_MAKE_MIB_COUNTER(30, 2, inMldGeneralQuery),
+ RTL8365MB_MAKE_MIB_COUNTER(32, 2, inIgmpSpecificQuery),
+ RTL8365MB_MAKE_MIB_COUNTER(34, 2, inIgmpGeneralQuery),
+ RTL8365MB_MAKE_MIB_COUNTER(36, 2, inMldLeaves),
+ RTL8365MB_MAKE_MIB_COUNTER(38, 2, inIgmpLeaves),
+ RTL8365MB_MAKE_MIB_COUNTER(40, 4, etherStatsOctets),
+ RTL8365MB_MAKE_MIB_COUNTER(44, 2, etherStatsUnderSizePkts),
+ RTL8365MB_MAKE_MIB_COUNTER(46, 2, etherOversizeStats),
+ RTL8365MB_MAKE_MIB_COUNTER(48, 2, etherStatsPkts64Octets),
+ RTL8365MB_MAKE_MIB_COUNTER(50, 2, etherStatsPkts65to127Octets),
+ RTL8365MB_MAKE_MIB_COUNTER(52, 2, etherStatsPkts128to255Octets),
+ RTL8365MB_MAKE_MIB_COUNTER(54, 2, etherStatsPkts256to511Octets),
+ RTL8365MB_MAKE_MIB_COUNTER(56, 2, etherStatsPkts512to1023Octets),
+ RTL8365MB_MAKE_MIB_COUNTER(58, 2, etherStatsPkts1024to1518Octets),
+ RTL8365MB_MAKE_MIB_COUNTER(60, 4, ifOutOctets),
+ RTL8365MB_MAKE_MIB_COUNTER(64, 2, dot3StatsSingleCollisionFrames),
+ RTL8365MB_MAKE_MIB_COUNTER(66, 2, dot3StatsMultipleCollisionFrames),
+ RTL8365MB_MAKE_MIB_COUNTER(68, 2, dot3StatsDeferredTransmissions),
+ RTL8365MB_MAKE_MIB_COUNTER(70, 2, dot3StatsLateCollisions),
+ RTL8365MB_MAKE_MIB_COUNTER(72, 2, etherStatsCollisions),
+ RTL8365MB_MAKE_MIB_COUNTER(74, 2, dot3StatsExcessiveCollisions),
+ RTL8365MB_MAKE_MIB_COUNTER(76, 2, dot3OutPauseFrames),
+ RTL8365MB_MAKE_MIB_COUNTER(78, 2, ifOutDiscards),
+ RTL8365MB_MAKE_MIB_COUNTER(80, 2, dot1dTpPortInDiscards),
+ RTL8365MB_MAKE_MIB_COUNTER(82, 2, ifOutUcastPkts),
+ RTL8365MB_MAKE_MIB_COUNTER(84, 2, ifOutMulticastPkts),
+ RTL8365MB_MAKE_MIB_COUNTER(86, 2, ifOutBroadcastPkts),
+ RTL8365MB_MAKE_MIB_COUNTER(88, 2, outOampduPkts),
+ RTL8365MB_MAKE_MIB_COUNTER(90, 2, inOampduPkts),
+ RTL8365MB_MAKE_MIB_COUNTER(92, 4, inIgmpJoinsSuccess),
+ RTL8365MB_MAKE_MIB_COUNTER(96, 2, inIgmpJoinsFail),
+ RTL8365MB_MAKE_MIB_COUNTER(98, 2, inMldJoinsSuccess),
+ RTL8365MB_MAKE_MIB_COUNTER(100, 2, inMldJoinsFail),
+ RTL8365MB_MAKE_MIB_COUNTER(102, 2, inReportSuppressionDrop),
+ RTL8365MB_MAKE_MIB_COUNTER(104, 2, inLeaveSuppressionDrop),
+ RTL8365MB_MAKE_MIB_COUNTER(106, 2, outIgmpReports),
+ RTL8365MB_MAKE_MIB_COUNTER(108, 2, outIgmpLeaves),
+ RTL8365MB_MAKE_MIB_COUNTER(110, 2, outIgmpGeneralQuery),
+ RTL8365MB_MAKE_MIB_COUNTER(112, 2, outIgmpSpecificQuery),
+ RTL8365MB_MAKE_MIB_COUNTER(114, 2, outMldReports),
+ RTL8365MB_MAKE_MIB_COUNTER(116, 2, outMldLeaves),
+ RTL8365MB_MAKE_MIB_COUNTER(118, 2, outMldGeneralQuery),
+ RTL8365MB_MAKE_MIB_COUNTER(120, 2, outMldSpecificQuery),
+ RTL8365MB_MAKE_MIB_COUNTER(122, 2, inKnownMulticastPkts),
+};
+
+static_assert(ARRAY_SIZE(rtl8365mb_mib_counters) == RTL8365MB_MIB_END);
+
+struct rtl8365mb_jam_tbl_entry {
+ u16 reg;
+ u16 val;
+};
+
+/* Lifted from the vendor driver sources */
+static const struct rtl8365mb_jam_tbl_entry rtl8365mb_init_jam_8365mb_vc[] = {
+ { 0x13EB, 0x15BB }, { 0x1303, 0x06D6 }, { 0x1304, 0x0700 },
+ { 0x13E2, 0x003F }, { 0x13F9, 0x0090 }, { 0x121E, 0x03CA },
+ { 0x1233, 0x0352 }, { 0x1237, 0x00A0 }, { 0x123A, 0x0030 },
+ { 0x1239, 0x0084 }, { 0x0301, 0x1000 }, { 0x1349, 0x001F },
+ { 0x18E0, 0x4004 }, { 0x122B, 0x241C }, { 0x1305, 0xC000 },
+ { 0x13F0, 0x0000 },
+};
+
+static const struct rtl8365mb_jam_tbl_entry rtl8365mb_init_jam_common[] = {
+ { 0x1200, 0x7FCB }, { 0x0884, 0x0003 }, { 0x06EB, 0x0001 },
+ { 0x03Fa, 0x0007 }, { 0x08C8, 0x00C0 }, { 0x0A30, 0x020E },
+ { 0x0800, 0x0000 }, { 0x0802, 0x0000 }, { 0x09DA, 0x0013 },
+ { 0x1D32, 0x0002 },
+};
+
+enum rtl8365mb_stp_state {
+ RTL8365MB_STP_STATE_DISABLED = 0,
+ RTL8365MB_STP_STATE_BLOCKING = 1,
+ RTL8365MB_STP_STATE_LEARNING = 2,
+ RTL8365MB_STP_STATE_FORWARDING = 3,
+};
+
+enum rtl8365mb_cpu_insert {
+ RTL8365MB_CPU_INSERT_TO_ALL = 0,
+ RTL8365MB_CPU_INSERT_TO_TRAPPING = 1,
+ RTL8365MB_CPU_INSERT_TO_NONE = 2,
+};
+
+enum rtl8365mb_cpu_position {
+ RTL8365MB_CPU_POS_AFTER_SA = 0,
+ RTL8365MB_CPU_POS_BEFORE_CRC = 1,
+};
+
+enum rtl8365mb_cpu_format {
+ RTL8365MB_CPU_FORMAT_8BYTES = 0,
+ RTL8365MB_CPU_FORMAT_4BYTES = 1,
+};
+
+enum rtl8365mb_cpu_rxlen {
+ RTL8365MB_CPU_RXLEN_72BYTES = 0,
+ RTL8365MB_CPU_RXLEN_64BYTES = 1,
+};
+
+/**
+ * struct rtl8365mb_cpu - CPU port configuration
+ * @enable: enable/disable hardware insertion of CPU tag in switch->CPU frames
+ * @mask: port mask of ports that parse should parse CPU tags
+ * @trap_port: forward trapped frames to this port
+ * @insert: CPU tag insertion mode in switch->CPU frames
+ * @position: position of CPU tag in frame
+ * @rx_length: minimum CPU RX length
+ * @format: CPU tag format
+ *
+ * Represents the CPU tagging and CPU port configuration of the switch. These
+ * settings are configurable at runtime.
+ */
+struct rtl8365mb_cpu {
+ bool enable;
+ u32 mask;
+ u32 trap_port;
+ enum rtl8365mb_cpu_insert insert;
+ enum rtl8365mb_cpu_position position;
+ enum rtl8365mb_cpu_rxlen rx_length;
+ enum rtl8365mb_cpu_format format;
+};
+
+/**
+ * struct rtl8365mb_port - private per-port data
+ * @smi: pointer to parent realtek_smi data
+ * @index: DSA port index, same as dsa_port::index
+ * @stats: link statistics populated by rtl8365mb_stats_poll, ready for atomic
+ * access via rtl8365mb_get_stats64
+ * @stats_lock: spinlock to protect the stats structure during read/update
+ * @mib_work: delayed work for polling MIB counters
+ */
+struct rtl8365mb_port {
+ struct realtek_smi *smi;
+ unsigned int index;
+ struct rtnl_link_stats64 stats;
+ spinlock_t stats_lock;
+ struct delayed_work mib_work;
+};
+
+/**
+ * struct rtl8365mb - private chip-specific driver data
+ * @smi: pointer to parent realtek_smi data
+ * @irq: registered IRQ or zero
+ * @chip_id: chip identifier
+ * @chip_ver: chip silicon revision
+ * @port_mask: mask of all ports
+ * @learn_limit_max: maximum number of L2 addresses the chip can learn
+ * @cpu: CPU tagging and CPU port configuration for this chip
+ * @mib_lock: spinlock to prevent concurrent reads of MIB counters
+ * @ports: per-port data
+ * @jam_table: chip-specific initialization jam table
+ * @jam_size: size of the chip's jam table
+ *
+ * Private data for this driver.
+ */
+struct rtl8365mb {
+ struct realtek_smi *smi;
+ int irq;
+ u32 chip_id;
+ u32 chip_ver;
+ u32 port_mask;
+ u32 learn_limit_max;
+ struct rtl8365mb_cpu cpu;
+ spinlock_t mib_lock;
+ struct rtl8365mb_port ports[RTL8365MB_MAX_NUM_PORTS];
+ const struct rtl8365mb_jam_tbl_entry *jam_table;
+ size_t jam_size;
+};
+
+static int rtl8365mb_phy_poll_busy(struct realtek_smi *smi)
+{
+ u32 val;
+
+ return regmap_read_poll_timeout(smi->map,
+ RTL8365MB_INDIRECT_ACCESS_STATUS_REG,
+ val, !val, 10, 100);
+}
+
+static int rtl8365mb_phy_ocp_prepare(struct realtek_smi *smi, int phy,
+ u32 ocp_addr)
+{
+ u32 val;
+ int ret;
+
+ /* Set OCP prefix */
+ val = FIELD_GET(RTL8365MB_PHY_OCP_ADDR_PREFIX_MASK, ocp_addr);
+ ret = regmap_update_bits(
+ smi->map, RTL8365MB_GPHY_OCP_MSB_0_REG,
+ RTL8365MB_GPHY_OCP_MSB_0_CFG_CPU_OCPADR_MASK,
+ FIELD_PREP(RTL8365MB_GPHY_OCP_MSB_0_CFG_CPU_OCPADR_MASK, val));
+ if (ret)
+ return ret;
+
+ /* Set PHY register address */
+ val = RTL8365MB_PHY_BASE;
+ val |= FIELD_PREP(RTL8365MB_INDIRECT_ACCESS_ADDRESS_PHYNUM_MASK, phy);
+ val |= FIELD_PREP(RTL8365MB_INDIRECT_ACCESS_ADDRESS_OCPADR_5_1_MASK,
+ ocp_addr >> 1);
+ val |= FIELD_PREP(RTL8365MB_INDIRECT_ACCESS_ADDRESS_OCPADR_9_6_MASK,
+ ocp_addr >> 6);
+ ret = regmap_write(smi->map, RTL8365MB_INDIRECT_ACCESS_ADDRESS_REG,
+ val);
+ if (ret)
+ return ret;
+
+ return 0;
+}
+
+static int rtl8365mb_phy_ocp_read(struct realtek_smi *smi, int phy,
+ u32 ocp_addr, u16 *data)
+{
+ u32 val;
+ int ret;
+
+ ret = rtl8365mb_phy_poll_busy(smi);
+ if (ret)
+ return ret;
+
+ ret = rtl8365mb_phy_ocp_prepare(smi, phy, ocp_addr);
+ if (ret)
+ return ret;
+
+ /* Execute read operation */
+ val = FIELD_PREP(RTL8365MB_INDIRECT_ACCESS_CTRL_CMD_MASK,
+ RTL8365MB_INDIRECT_ACCESS_CTRL_CMD_VALUE) |
+ FIELD_PREP(RTL8365MB_INDIRECT_ACCESS_CTRL_RW_MASK,
+ RTL8365MB_INDIRECT_ACCESS_CTRL_RW_READ);
+ ret = regmap_write(smi->map, RTL8365MB_INDIRECT_ACCESS_CTRL_REG, val);
+ if (ret)
+ return ret;
+
+ ret = rtl8365mb_phy_poll_busy(smi);
+ if (ret)
+ return ret;
+
+ /* Get PHY register data */
+ ret = regmap_read(smi->map, RTL8365MB_INDIRECT_ACCESS_READ_DATA_REG,
+ &val);
+ if (ret)
+ return ret;
+
+ *data = val & 0xFFFF;
+
+ return 0;
+}
+
+static int rtl8365mb_phy_ocp_write(struct realtek_smi *smi, int phy,
+ u32 ocp_addr, u16 data)
+{
+ u32 val;
+ int ret;
+
+ ret = rtl8365mb_phy_poll_busy(smi);
+ if (ret)
+ return ret;
+
+ ret = rtl8365mb_phy_ocp_prepare(smi, phy, ocp_addr);
+ if (ret)
+ return ret;
+
+ /* Set PHY register data */
+ ret = regmap_write(smi->map, RTL8365MB_INDIRECT_ACCESS_WRITE_DATA_REG,
+ data);
+ if (ret)
+ return ret;
+
+ /* Execute write operation */
+ val = FIELD_PREP(RTL8365MB_INDIRECT_ACCESS_CTRL_CMD_MASK,
+ RTL8365MB_INDIRECT_ACCESS_CTRL_CMD_VALUE) |
+ FIELD_PREP(RTL8365MB_INDIRECT_ACCESS_CTRL_RW_MASK,
+ RTL8365MB_INDIRECT_ACCESS_CTRL_RW_WRITE);
+ ret = regmap_write(smi->map, RTL8365MB_INDIRECT_ACCESS_CTRL_REG, val);
+ if (ret)
+ return ret;
+
+ ret = rtl8365mb_phy_poll_busy(smi);
+ if (ret)
+ return ret;
+
+ return 0;
+}
+
+static int rtl8365mb_phy_read(struct realtek_smi *smi, int phy, int regnum)
+{
+ u32 ocp_addr;
+ u16 val;
+ int ret;
+
+ if (regnum > RTL8365MB_PHYREGMAX)
+ return -EINVAL;
+
+ ocp_addr = RTL8365MB_PHY_OCP_ADDR_PHYREG_BASE + regnum * 2;
+
+ ret = rtl8365mb_phy_ocp_read(smi, phy, ocp_addr, &val);
+ if (ret) {
+ dev_err(smi->dev,
+ "failed to read PHY%d reg %02x @ %04x, ret %d\n", phy,
+ regnum, ocp_addr, ret);
+ return ret;
+ }
+
+ dev_dbg(smi->dev, "read PHY%d register 0x%02x @ %04x, val <- %04x\n",
+ phy, regnum, ocp_addr, val);
+
+ return val;
+}
+
+static int rtl8365mb_phy_write(struct realtek_smi *smi, int phy, int regnum,
+ u16 val)
+{
+ u32 ocp_addr;
+ int ret;
+
+ if (regnum > RTL8365MB_PHYREGMAX)
+ return -EINVAL;
+
+ ocp_addr = RTL8365MB_PHY_OCP_ADDR_PHYREG_BASE + regnum * 2;
+
+ ret = rtl8365mb_phy_ocp_write(smi, phy, ocp_addr, val);
+ if (ret) {
+ dev_err(smi->dev,
+ "failed to write PHY%d reg %02x @ %04x, ret %d\n", phy,
+ regnum, ocp_addr, ret);
+ return ret;
+ }
+
+ dev_dbg(smi->dev, "write PHY%d register 0x%02x @ %04x, val -> %04x\n",
+ phy, regnum, ocp_addr, val);
+
+ return 0;
+}
+
+static enum dsa_tag_protocol
+rtl8365mb_get_tag_protocol(struct dsa_switch *ds, int port,
+ enum dsa_tag_protocol mp)
+{
+ return DSA_TAG_PROTO_RTL8_4;
+}
+
+static int rtl8365mb_ext_config_rgmii(struct realtek_smi *smi, int port,
+ phy_interface_t interface)
+{
+ struct device_node *dn;
+ struct dsa_port *dp;
+ int tx_delay = 0;
+ int rx_delay = 0;
+ int ext_port;
+ u32 val;
+ int ret;
+
+ if (port == smi->cpu_port) {
+ ext_port = 1;
+ } else {
+ dev_err(smi->dev, "only one EXT port is currently supported\n");
+ return -EINVAL;
+ }
+
+ dp = dsa_to_port(smi->ds, port);
+ dn = dp->dn;
+
+ /* Set the RGMII TX/RX delay
+ *
+ * The Realtek vendor driver indicates the following possible
+ * configuration settings:
+ *
+ * TX delay:
+ * 0 = no delay, 1 = 2 ns delay
+ * RX delay:
+ * 0 = no delay, 7 = maximum delay
+ * No units are specified, but there are a total of 8 steps.
+ *
+ * The vendor driver also states that this must be configured *before*
+ * forcing the external interface into a particular mode, which is done
+ * in the rtl8365mb_phylink_mac_link_{up,down} functions.
+ *
+ * Only configure an RGMII TX (resp. RX) delay if the
+ * tx-internal-delay-ps (resp. rx-internal-delay-ps) OF property is
+ * specified. We ignore the detail of the RGMII interface mode
+ * (RGMII_{RXID, TXID, etc.}), as this is considered to be a PHY-only
+ * property.
+ *
+ * For the RX delay, we assume that a register value of 4 corresponds to
+ * 2 ns. But this is just an educated guess, so ignore all other values
+ * to avoid too much confusion.
+ */
+ if (!of_property_read_u32(dn, "tx-internal-delay-ps", &val)) {
+ val = val / 1000; /* convert to ns */
+
+ if (val == 0 || val == 2)
+ tx_delay = val / 2;
+ else
+ dev_warn(smi->dev,
+ "EXT port TX delay must be 0 or 2 ns\n");
+ }
+
+ if (!of_property_read_u32(dn, "rx-internal-delay-ps", &val)) {
+ val = val / 1000; /* convert to ns */
+
+ if (val == 0 || val == 2)
+ rx_delay = val * 2;
+ else
+ dev_warn(smi->dev,
+ "EXT port RX delay must be 0 to 2 ns\n");
+ }
+
+ ret = regmap_update_bits(
+ smi->map, RTL8365MB_EXT_RGMXF_REG(ext_port),
+ RTL8365MB_EXT_RGMXF_TXDELAY_MASK |
+ RTL8365MB_EXT_RGMXF_RXDELAY_MASK,
+ FIELD_PREP(RTL8365MB_EXT_RGMXF_TXDELAY_MASK, tx_delay) |
+ FIELD_PREP(RTL8365MB_EXT_RGMXF_RXDELAY_MASK, rx_delay));
+ if (ret)
+ return ret;
+
+ ret = regmap_update_bits(
+ smi->map, RTL8365MB_DIGITAL_INTERFACE_SELECT_REG(ext_port),
+ RTL8365MB_DIGITAL_INTERFACE_SELECT_MODE_MASK(ext_port),
+ RTL8365MB_EXT_PORT_MODE_RGMII
+ << RTL8365MB_DIGITAL_INTERFACE_SELECT_MODE_OFFSET(
+ ext_port));
+ if (ret)
+ return ret;
+
+ return 0;
+}
+
+static int rtl8365mb_ext_config_forcemode(struct realtek_smi *smi, int port,
+ bool link, int speed, int duplex,
+ bool tx_pause, bool rx_pause)
+{
+ u32 r_tx_pause;
+ u32 r_rx_pause;
+ u32 r_duplex;
+ u32 r_speed;
+ u32 r_link;
+ int ext_port;
+ int val;
+ int ret;
+
+ if (port == smi->cpu_port) {
+ ext_port = 1;
+ } else {
+ dev_err(smi->dev, "only one EXT port is currently supported\n");
+ return -EINVAL;
+ }
+
+ if (link) {
+ /* Force the link up with the desired configuration */
+ r_link = 1;
+ r_rx_pause = rx_pause ? 1 : 0;
+ r_tx_pause = tx_pause ? 1 : 0;
+
+ if (speed == SPEED_1000) {
+ r_speed = RTL8365MB_PORT_SPEED_1000M;
+ } else if (speed == SPEED_100) {
+ r_speed = RTL8365MB_PORT_SPEED_100M;
+ } else if (speed == SPEED_10) {
+ r_speed = RTL8365MB_PORT_SPEED_10M;
+ } else {
+ dev_err(smi->dev, "unsupported port speed %s\n",
+ phy_speed_to_str(speed));
+ return -EINVAL;
+ }
+
+ if (duplex == DUPLEX_FULL) {
+ r_duplex = 1;
+ } else if (duplex == DUPLEX_HALF) {
+ r_duplex = 0;
+ } else {
+ dev_err(smi->dev, "unsupported duplex %s\n",
+ phy_duplex_to_str(duplex));
+ return -EINVAL;
+ }
+ } else {
+ /* Force the link down and reset any programmed configuration */
+ r_link = 0;
+ r_tx_pause = 0;
+ r_rx_pause = 0;
+ r_speed = 0;
+ r_duplex = 0;
+ }
+
+ val = FIELD_PREP(RTL8365MB_DIGITAL_INTERFACE_FORCE_EN_MASK, 1) |
+ FIELD_PREP(RTL8365MB_DIGITAL_INTERFACE_FORCE_TXPAUSE_MASK,
+ r_tx_pause) |
+ FIELD_PREP(RTL8365MB_DIGITAL_INTERFACE_FORCE_RXPAUSE_MASK,
+ r_rx_pause) |
+ FIELD_PREP(RTL8365MB_DIGITAL_INTERFACE_FORCE_LINK_MASK, r_link) |
+ FIELD_PREP(RTL8365MB_DIGITAL_INTERFACE_FORCE_DUPLEX_MASK,
+ r_duplex) |
+ FIELD_PREP(RTL8365MB_DIGITAL_INTERFACE_FORCE_SPEED_MASK, r_speed);
+ ret = regmap_write(smi->map,
+ RTL8365MB_DIGITAL_INTERFACE_FORCE_REG(ext_port),
+ val);
+ if (ret)
+ return ret;
+
+ return 0;
+}
+
+static bool rtl8365mb_phy_mode_supported(struct dsa_switch *ds, int port,
+ phy_interface_t interface)
+{
+ if (dsa_is_user_port(ds, port) &&
+ (interface == PHY_INTERFACE_MODE_NA ||
+ interface == PHY_INTERFACE_MODE_INTERNAL))
+ /* Internal PHY */
+ return true;
+ else if (dsa_is_cpu_port(ds, port) &&
+ phy_interface_mode_is_rgmii(interface))
+ /* Extension MAC */
+ return true;
+
+ return false;
+}
+
+static void rtl8365mb_phylink_validate(struct dsa_switch *ds, int port,
+ unsigned long *supported,
+ struct phylink_link_state *state)
+{
+ struct realtek_smi *smi = ds->priv;
+ __ETHTOOL_DECLARE_LINK_MODE_MASK(mask) = { 0 };
+
+ /* include/linux/phylink.h says:
+ * When @state->interface is %PHY_INTERFACE_MODE_NA, phylink
+ * expects the MAC driver to return all supported link modes.
+ */
+ if (state->interface != PHY_INTERFACE_MODE_NA &&
+ !rtl8365mb_phy_mode_supported(ds, port, state->interface)) {
+ dev_err(smi->dev, "phy mode %s is unsupported on port %d\n",
+ phy_modes(state->interface), port);
+ linkmode_zero(supported);
+ return;
+ }
+
+ phylink_set_port_modes(mask);
+
+ phylink_set(mask, Autoneg);
+ phylink_set(mask, Pause);
+ phylink_set(mask, Asym_Pause);
+
+ phylink_set(mask, 10baseT_Half);
+ phylink_set(mask, 10baseT_Full);
+ phylink_set(mask, 100baseT_Half);
+ phylink_set(mask, 100baseT_Full);
+ phylink_set(mask, 1000baseT_Full);
+
+ linkmode_and(supported, supported, mask);
+ linkmode_and(state->advertising, state->advertising, mask);
+}
+
+static void rtl8365mb_phylink_mac_config(struct dsa_switch *ds, int port,
+ unsigned int mode,
+ const struct phylink_link_state *state)
+{
+ struct realtek_smi *smi = ds->priv;
+ int ret;
+
+ if (!rtl8365mb_phy_mode_supported(ds, port, state->interface)) {
+ dev_err(smi->dev, "phy mode %s is unsupported on port %d\n",
+ phy_modes(state->interface), port);
+ return;
+ }
+
+ if (mode != MLO_AN_PHY && mode != MLO_AN_FIXED) {
+ dev_err(smi->dev,
+ "port %d supports only conventional PHY or fixed-link\n",
+ port);
+ return;
+ }
+
+ if (phy_interface_mode_is_rgmii(state->interface)) {
+ ret = rtl8365mb_ext_config_rgmii(smi, port, state->interface);
+ if (ret)
+ dev_err(smi->dev,
+ "failed to configure RGMII mode on port %d: %d\n",
+ port, ret);
+ return;
+ }
+
+ /* TODO: Implement MII and RMII modes, which the RTL8365MB-VC also
+ * supports
+ */
+}
+
+static void rtl8365mb_phylink_mac_link_down(struct dsa_switch *ds, int port,
+ unsigned int mode,
+ phy_interface_t interface)
+{
+ struct realtek_smi *smi = ds->priv;
+ struct rtl8365mb_port *p;
+ struct rtl8365mb *mb;
+ int ret;
+
+ mb = smi->chip_data;
+ p = &mb->ports[port];
+ cancel_delayed_work_sync(&p->mib_work);
+
+ if (phy_interface_mode_is_rgmii(interface)) {
+ ret = rtl8365mb_ext_config_forcemode(smi, port, false, 0, 0,
+ false, false);
+ if (ret)
+ dev_err(smi->dev,
+ "failed to reset forced mode on port %d: %d\n",
+ port, ret);
+
+ return;
+ }
+}
+
+static void rtl8365mb_phylink_mac_link_up(struct dsa_switch *ds, int port,
+ unsigned int mode,
+ phy_interface_t interface,
+ struct phy_device *phydev, int speed,
+ int duplex, bool tx_pause,
+ bool rx_pause)
+{
+ struct realtek_smi *smi = ds->priv;
+ struct rtl8365mb_port *p;
+ struct rtl8365mb *mb;
+ int ret;
+
+ mb = smi->chip_data;
+ p = &mb->ports[port];
+ schedule_delayed_work(&p->mib_work, 0);
+
+ if (phy_interface_mode_is_rgmii(interface)) {
+ ret = rtl8365mb_ext_config_forcemode(smi, port, true, speed,
+ duplex, tx_pause,
+ rx_pause);
+ if (ret)
+ dev_err(smi->dev,
+ "failed to force mode on port %d: %d\n", port,
+ ret);
+
+ return;
+ }
+}
+
+static void rtl8365mb_port_stp_state_set(struct dsa_switch *ds, int port,
+ u8 state)
+{
+ struct realtek_smi *smi = ds->priv;
+ enum rtl8365mb_stp_state val;
+ int msti = 0;
+
+ switch (state) {
+ case BR_STATE_DISABLED:
+ val = RTL8365MB_STP_STATE_DISABLED;
+ break;
+ case BR_STATE_BLOCKING:
+ case BR_STATE_LISTENING:
+ val = RTL8365MB_STP_STATE_BLOCKING;
+ break;
+ case BR_STATE_LEARNING:
+ val = RTL8365MB_STP_STATE_LEARNING;
+ break;
+ case BR_STATE_FORWARDING:
+ val = RTL8365MB_STP_STATE_FORWARDING;
+ break;
+ default:
+ dev_err(smi->dev, "invalid STP state: %u\n", state);
+ return;
+ }
+
+ regmap_update_bits(smi->map, RTL8365MB_MSTI_CTRL_REG(msti, port),
+ RTL8365MB_MSTI_CTRL_PORT_STATE_MASK(port),
+ val << RTL8365MB_MSTI_CTRL_PORT_STATE_OFFSET(port));
+}
+
+static int rtl8365mb_port_set_learning(struct realtek_smi *smi, int port,
+ bool enable)
+{
+ struct rtl8365mb *mb = smi->chip_data;
+
+ /* Enable/disable learning by limiting the number of L2 addresses the
+ * port can learn. Realtek documentation states that a limit of zero
+ * disables learning. When enabling learning, set it to the chip's
+ * maximum.
+ */
+ return regmap_write(smi->map, RTL8365MB_LUT_PORT_LEARN_LIMIT_REG(port),
+ enable ? mb->learn_limit_max : 0);
+}
+
+static int rtl8365mb_port_set_isolation(struct realtek_smi *smi, int port,
+ u32 mask)
+{
+ return regmap_write(smi->map, RTL8365MB_PORT_ISOLATION_REG(port), mask);
+}
+
+static int rtl8365mb_mib_counter_read(struct realtek_smi *smi, int port,
+ u32 offset, u32 length, u64 *mibvalue)
+{
+ u64 tmpvalue = 0;
+ u32 val;
+ int ret;
+ int i;
+
+ /* The MIB address is an SRAM address. We request a particular address
+ * and then poll the control register before reading the value from some
+ * counter registers.
+ */
+ ret = regmap_write(smi->map, RTL8365MB_MIB_ADDRESS_REG,
+ RTL8365MB_MIB_ADDRESS(port, offset));
+ if (ret)
+ return ret;
+
+ /* Poll for completion */
+ ret = regmap_read_poll_timeout(smi->map, RTL8365MB_MIB_CTRL0_REG, val,
+ !(val & RTL8365MB_MIB_CTRL0_BUSY_MASK),
+ 10, 100);
+ if (ret)
+ return ret;
+
+ /* Presumably this indicates a MIB counter read failure */
+ if (val & RTL8365MB_MIB_CTRL0_RESET_MASK)
+ return -EIO;
+
+ /* There are four MIB counter registers each holding a 16 bit word of a
+ * MIB counter. Depending on the offset, we should read from the upper
+ * two or lower two registers. In case the MIB counter is 4 words, we
+ * read from all four registers.
+ */
+ if (length == 4)
+ offset = 3;
+ else
+ offset = (offset + 1) % 4;
+
+ /* Read the MIB counter 16 bits at a time */
+ for (i = 0; i < length; i++) {
+ ret = regmap_read(smi->map,
+ RTL8365MB_MIB_COUNTER_REG(offset - i), &val);
+ if (ret)
+ return ret;
+
+ tmpvalue = ((tmpvalue) << 16) | (val & 0xFFFF);
+ }
+
+ /* Only commit the result if no error occurred */
+ *mibvalue = tmpvalue;
+
+ return 0;
+}
+
+static void rtl8365mb_get_ethtool_stats(struct dsa_switch *ds, int port, u64 *data)
+{
+ struct realtek_smi *smi = ds->priv;
+ struct rtl8365mb *mb;
+ int ret;
+ int i;
+
+ mb = smi->chip_data;
+
+ spin_lock(&mb->mib_lock);
+ for (i = 0; i < RTL8365MB_MIB_END; i++) {
+ struct rtl8365mb_mib_counter *mib = &rtl8365mb_mib_counters[i];
+
+ ret = rtl8365mb_mib_counter_read(smi, port, mib->offset,
+ mib->length, &data[i]);
+ if (ret) {
+ dev_err(smi->dev,
+ "failed to read port %d counters: %d\n", port,
+ ret);
+ break;
+ }
+ }
+ spin_unlock(&mb->mib_lock);
+}
+
+static void rtl8365mb_get_strings(struct dsa_switch *ds, int port, u32 stringset, u8 *data)
+{
+ int i;
+
+ if (stringset != ETH_SS_STATS)
+ return;
+
+ for (i = 0; i < RTL8365MB_MIB_END; i++) {
+ struct rtl8365mb_mib_counter *mib = &rtl8365mb_mib_counters[i];
+
+ strncpy(data + i * ETH_GSTRING_LEN, mib->name, ETH_GSTRING_LEN);
+ }
+}
+
+static int rtl8365mb_get_sset_count(struct dsa_switch *ds, int port, int sset)
+{
+ if (sset != ETH_SS_STATS)
+ return -EOPNOTSUPP;
+
+ return RTL8365MB_MIB_END;
+}
+
+static void rtl8365mb_get_phy_stats(struct dsa_switch *ds, int port,
+ struct ethtool_eth_phy_stats *phy_stats)
+{
+ struct realtek_smi *smi = ds->priv;
+ struct rtl8365mb_mib_counter *mib;
+ struct rtl8365mb *mb;
+
+ mb = smi->chip_data;
+ mib = &rtl8365mb_mib_counters[RTL8365MB_MIB_dot3StatsSymbolErrors];
+
+ spin_lock(&mb->mib_lock);
+ rtl8365mb_mib_counter_read(smi, port, mib->offset, mib->length,
+ &phy_stats->SymbolErrorDuringCarrier);
+ spin_unlock(&mb->mib_lock);
+}
+
+static void rtl8365mb_get_mac_stats(struct dsa_switch *ds, int port,
+ struct ethtool_eth_mac_stats *mac_stats)
+{
+ u64 cnt[RTL8365MB_MIB_END] = {
+ [RTL8365MB_MIB_ifOutOctets] = 1,
+ [RTL8365MB_MIB_ifOutUcastPkts] = 1,
+ [RTL8365MB_MIB_ifOutMulticastPkts] = 1,
+ [RTL8365MB_MIB_ifOutBroadcastPkts] = 1,
+ [RTL8365MB_MIB_dot3OutPauseFrames] = 1,
+ [RTL8365MB_MIB_ifOutDiscards] = 1,
+ [RTL8365MB_MIB_ifInOctets] = 1,
+ [RTL8365MB_MIB_ifInUcastPkts] = 1,
+ [RTL8365MB_MIB_ifInMulticastPkts] = 1,
+ [RTL8365MB_MIB_ifInBroadcastPkts] = 1,
+ [RTL8365MB_MIB_dot3InPauseFrames] = 1,
+ [RTL8365MB_MIB_dot3StatsSingleCollisionFrames] = 1,
+ [RTL8365MB_MIB_dot3StatsMultipleCollisionFrames] = 1,
+ [RTL8365MB_MIB_dot3StatsFCSErrors] = 1,
+ [RTL8365MB_MIB_dot3StatsDeferredTransmissions] = 1,
+ [RTL8365MB_MIB_dot3StatsLateCollisions] = 1,
+ [RTL8365MB_MIB_dot3StatsExcessiveCollisions] = 1,
+
+ };
+ struct realtek_smi *smi = ds->priv;
+ struct rtl8365mb *mb;
+ int ret;
+ int i;
+
+ mb = smi->chip_data;
+
+ spin_lock(&mb->mib_lock);
+ for (i = 0; i < RTL8365MB_MIB_END; i++) {
+ struct rtl8365mb_mib_counter *mib = &rtl8365mb_mib_counters[i];
+
+ /* Only fetch required MIB counters (marked = 1 above) */
+ if (!cnt[i])
+ continue;
+
+ ret = rtl8365mb_mib_counter_read(smi, port, mib->offset,
+ mib->length, &cnt[i]);
+ if (ret)
+ break;
+ }
+ spin_unlock(&mb->mib_lock);
+
+ /* The RTL8365MB-VC exposes MIB objects, which we have to translate into
+ * IEEE 802.3 Managed Objects. This is not always completely faithful,
+ * but we try out best. See RFC 3635 for a detailed treatment of the
+ * subject.
+ */
+
+ mac_stats->FramesTransmittedOK = cnt[RTL8365MB_MIB_ifOutUcastPkts] +
+ cnt[RTL8365MB_MIB_ifOutMulticastPkts] +
+ cnt[RTL8365MB_MIB_ifOutBroadcastPkts] +
+ cnt[RTL8365MB_MIB_dot3OutPauseFrames] -
+ cnt[RTL8365MB_MIB_ifOutDiscards];
+ mac_stats->SingleCollisionFrames =
+ cnt[RTL8365MB_MIB_dot3StatsSingleCollisionFrames];
+ mac_stats->MultipleCollisionFrames =
+ cnt[RTL8365MB_MIB_dot3StatsMultipleCollisionFrames];
+ mac_stats->FramesReceivedOK = cnt[RTL8365MB_MIB_ifInUcastPkts] +
+ cnt[RTL8365MB_MIB_ifInMulticastPkts] +
+ cnt[RTL8365MB_MIB_ifInBroadcastPkts] +
+ cnt[RTL8365MB_MIB_dot3InPauseFrames];
+ mac_stats->FrameCheckSequenceErrors =
+ cnt[RTL8365MB_MIB_dot3StatsFCSErrors];
+ mac_stats->OctetsTransmittedOK = cnt[RTL8365MB_MIB_ifOutOctets] -
+ 18 * mac_stats->FramesTransmittedOK;
+ mac_stats->FramesWithDeferredXmissions =
+ cnt[RTL8365MB_MIB_dot3StatsDeferredTransmissions];
+ mac_stats->LateCollisions = cnt[RTL8365MB_MIB_dot3StatsLateCollisions];
+ mac_stats->FramesAbortedDueToXSColls =
+ cnt[RTL8365MB_MIB_dot3StatsExcessiveCollisions];
+ mac_stats->OctetsReceivedOK = cnt[RTL8365MB_MIB_ifInOctets] -
+ 18 * mac_stats->FramesReceivedOK;
+ mac_stats->MulticastFramesXmittedOK =
+ cnt[RTL8365MB_MIB_ifOutMulticastPkts];
+ mac_stats->BroadcastFramesXmittedOK =
+ cnt[RTL8365MB_MIB_ifOutBroadcastPkts];
+ mac_stats->MulticastFramesReceivedOK =
+ cnt[RTL8365MB_MIB_ifInMulticastPkts];
+ mac_stats->BroadcastFramesReceivedOK =
+ cnt[RTL8365MB_MIB_ifInBroadcastPkts];
+}
+
+static void rtl8365mb_get_ctrl_stats(struct dsa_switch *ds, int port,
+ struct ethtool_eth_ctrl_stats *ctrl_stats)
+{
+ struct realtek_smi *smi = ds->priv;
+ struct rtl8365mb_mib_counter *mib;
+ struct rtl8365mb *mb;
+
+ mb = smi->chip_data;
+ mib = &rtl8365mb_mib_counters[RTL8365MB_MIB_dot3ControlInUnknownOpcodes];
+
+ spin_lock(&mb->mib_lock);
+ rtl8365mb_mib_counter_read(smi, port, mib->offset, mib->length,
+ &ctrl_stats->UnsupportedOpcodesReceived);
+ spin_unlock(&mb->mib_lock);
+}
+
+static void rtl8365mb_stats_update(struct realtek_smi *smi, int port)
+{
+ u64 cnt[RTL8365MB_MIB_END] = {
+ [RTL8365MB_MIB_ifOutOctets] = 1,
+ [RTL8365MB_MIB_ifOutUcastPkts] = 1,
+ [RTL8365MB_MIB_ifOutMulticastPkts] = 1,
+ [RTL8365MB_MIB_ifOutBroadcastPkts] = 1,
+ [RTL8365MB_MIB_ifOutDiscards] = 1,
+ [RTL8365MB_MIB_ifInOctets] = 1,
+ [RTL8365MB_MIB_ifInUcastPkts] = 1,
+ [RTL8365MB_MIB_ifInMulticastPkts] = 1,
+ [RTL8365MB_MIB_ifInBroadcastPkts] = 1,
+ [RTL8365MB_MIB_etherStatsDropEvents] = 1,
+ [RTL8365MB_MIB_etherStatsCollisions] = 1,
+ [RTL8365MB_MIB_etherStatsFragments] = 1,
+ [RTL8365MB_MIB_etherStatsJabbers] = 1,
+ [RTL8365MB_MIB_dot3StatsFCSErrors] = 1,
+ [RTL8365MB_MIB_dot3StatsLateCollisions] = 1,
+ };
+ struct rtl8365mb *mb = smi->chip_data;
+ struct rtnl_link_stats64 *stats;
+ int ret;
+ int i;
+
+ stats = &mb->ports[port].stats;
+
+ spin_lock(&mb->mib_lock);
+ for (i = 0; i < RTL8365MB_MIB_END; i++) {
+ struct rtl8365mb_mib_counter *c = &rtl8365mb_mib_counters[i];
+
+ /* Only fetch required MIB counters (marked = 1 above) */
+ if (!cnt[i])
+ continue;
+
+ ret = rtl8365mb_mib_counter_read(smi, port, c->offset,
+ c->length, &cnt[i]);
+ if (ret)
+ break;
+ }
+ spin_unlock(&mb->mib_lock);
+
+ /* Don't update statistics if there was an error reading the counters */
+ if (ret)
+ return;
+
+ spin_lock(&mb->ports[port].stats_lock);
+
+ stats->rx_packets = cnt[RTL8365MB_MIB_ifInUcastPkts] +
+ cnt[RTL8365MB_MIB_ifInMulticastPkts] +
+ cnt[RTL8365MB_MIB_ifInBroadcastPkts] -
+ cnt[RTL8365MB_MIB_ifOutDiscards];
+
+ stats->tx_packets = cnt[RTL8365MB_MIB_ifOutUcastPkts] +
+ cnt[RTL8365MB_MIB_ifOutMulticastPkts] +
+ cnt[RTL8365MB_MIB_ifOutBroadcastPkts];
+
+ /* if{In,Out}Octets includes FCS - remove it */
+ stats->rx_bytes = cnt[RTL8365MB_MIB_ifInOctets] - 4 * stats->rx_packets;
+ stats->tx_bytes =
+ cnt[RTL8365MB_MIB_ifOutOctets] - 4 * stats->tx_packets;
+
+ stats->rx_dropped = cnt[RTL8365MB_MIB_etherStatsDropEvents];
+ stats->tx_dropped = cnt[RTL8365MB_MIB_ifOutDiscards];
+
+ stats->multicast = cnt[RTL8365MB_MIB_ifInMulticastPkts];
+ stats->collisions = cnt[RTL8365MB_MIB_etherStatsCollisions];
+
+ stats->rx_length_errors = cnt[RTL8365MB_MIB_etherStatsFragments] +
+ cnt[RTL8365MB_MIB_etherStatsJabbers];
+ stats->rx_crc_errors = cnt[RTL8365MB_MIB_dot3StatsFCSErrors];
+ stats->rx_errors = stats->rx_length_errors + stats->rx_crc_errors;
+
+ stats->tx_aborted_errors = cnt[RTL8365MB_MIB_ifOutDiscards];
+ stats->tx_window_errors = cnt[RTL8365MB_MIB_dot3StatsLateCollisions];
+ stats->tx_errors = stats->tx_aborted_errors + stats->tx_window_errors;
+
+ spin_unlock(&mb->ports[port].stats_lock);
+}
+
+static void rtl8365mb_stats_poll(struct work_struct *work)
+{
+ struct rtl8365mb_port *p = container_of(to_delayed_work(work),
+ struct rtl8365mb_port,
+ mib_work);
+ struct realtek_smi *smi = p->smi;
+
+ rtl8365mb_stats_update(smi, p->index);
+
+ schedule_delayed_work(&p->mib_work, RTL8365MB_STATS_INTERVAL_JIFFIES);
+}
+
+static void rtl8365mb_get_stats64(struct dsa_switch *ds, int port,
+ struct rtnl_link_stats64 *s)
+{
+ struct realtek_smi *smi = ds->priv;
+ struct rtl8365mb_port *p;
+ struct rtl8365mb *mb;
+
+ mb = smi->chip_data;
+ p = &mb->ports[port];
+
+ spin_lock(&p->stats_lock);
+ memcpy(s, &p->stats, sizeof(*s));
+ spin_unlock(&p->stats_lock);
+}
+
+static void rtl8365mb_stats_setup(struct realtek_smi *smi)
+{
+ struct rtl8365mb *mb = smi->chip_data;
+ int i;
+
+ /* Per-chip global spinlock to protect MIB counter access, since doing
+ * so requires accessing a series of registers in a particular order.
+ */
+ spin_lock_init(&mb->mib_lock);
+
+ for (i = 0; i < smi->num_ports; i++) {
+ struct rtl8365mb_port *p = &mb->ports[i];
+
+ if (dsa_is_unused_port(smi->ds, i))
+ continue;
+
+ /* Per-port spinlock to protect the stats64 data */
+ spin_lock_init(&p->stats_lock);
+
+ /* This work polls the MIB counters and keeps the stats64 data
+ * up-to-date.
+ */
+ INIT_DELAYED_WORK(&p->mib_work, rtl8365mb_stats_poll);
+ }
+}
+
+static void rtl8365mb_stats_teardown(struct realtek_smi *smi)
+{
+ struct rtl8365mb *mb = smi->chip_data;
+ int i;
+
+ for (i = 0; i < smi->num_ports; i++) {
+ struct rtl8365mb_port *p = &mb->ports[i];
+
+ if (dsa_is_unused_port(smi->ds, i))
+ continue;
+
+ cancel_delayed_work_sync(&p->mib_work);
+ }
+}
+
+static int rtl8365mb_get_and_clear_status_reg(struct realtek_smi *smi, u32 reg,
+ u32 *val)
+{
+ int ret;
+
+ ret = regmap_read(smi->map, reg, val);
+ if (ret)
+ return ret;
+
+ return regmap_write(smi->map, reg, *val);
+}
+
+static irqreturn_t rtl8365mb_irq(int irq, void *data)
+{
+ struct realtek_smi *smi = data;
+ unsigned long line_changes = 0;
+ struct rtl8365mb *mb;
+ u32 stat;
+ int line;
+ int ret;
+
+ mb = smi->chip_data;
+
+ ret = rtl8365mb_get_and_clear_status_reg(smi, RTL8365MB_INTR_STATUS_REG,
+ &stat);
+ if (ret)
+ goto out_error;
+
+ if (stat & RTL8365MB_INTR_LINK_CHANGE_MASK) {
+ u32 linkdown_ind;
+ u32 linkup_ind;
+ u32 val;
+
+ ret = rtl8365mb_get_and_clear_status_reg(
+ smi, RTL8365MB_PORT_LINKUP_IND_REG, &val);
+ if (ret)
+ goto out_error;
+
+ linkup_ind = FIELD_GET(RTL8365MB_PORT_LINKUP_IND_MASK, val);
+
+ ret = rtl8365mb_get_and_clear_status_reg(
+ smi, RTL8365MB_PORT_LINKDOWN_IND_REG, &val);
+ if (ret)
+ goto out_error;
+
+ linkdown_ind = FIELD_GET(RTL8365MB_PORT_LINKDOWN_IND_MASK, val);
+
+ line_changes = (linkup_ind | linkdown_ind) & mb->port_mask;
+ }
+
+ if (!line_changes)
+ goto out_none;
+
+ for_each_set_bit(line, &line_changes, smi->num_ports) {
+ int child_irq = irq_find_mapping(smi->irqdomain, line);
+
+ handle_nested_irq(child_irq);
+ }
+
+ return IRQ_HANDLED;
+
+out_error:
+ dev_err(smi->dev, "failed to read interrupt status: %d\n", ret);
+
+out_none:
+ return IRQ_NONE;
+}
+
+static struct irq_chip rtl8365mb_irq_chip = {
+ .name = "rtl8365mb",
+ /* The hardware doesn't support masking IRQs on a per-port basis */
+};
+
+static int rtl8365mb_irq_map(struct irq_domain *domain, unsigned int irq,
+ irq_hw_number_t hwirq)
+{
+ irq_set_chip_data(irq, domain->host_data);
+ irq_set_chip_and_handler(irq, &rtl8365mb_irq_chip, handle_simple_irq);
+ irq_set_nested_thread(irq, 1);
+ irq_set_noprobe(irq);
+
+ return 0;
+}
+
+static void rtl8365mb_irq_unmap(struct irq_domain *d, unsigned int irq)
+{
+ irq_set_nested_thread(irq, 0);
+ irq_set_chip_and_handler(irq, NULL, NULL);
+ irq_set_chip_data(irq, NULL);
+}
+
+static const struct irq_domain_ops rtl8365mb_irqdomain_ops = {
+ .map = rtl8365mb_irq_map,
+ .unmap = rtl8365mb_irq_unmap,
+ .xlate = irq_domain_xlate_onecell,
+};
+
+static int rtl8365mb_set_irq_enable(struct realtek_smi *smi, bool enable)
+{
+ return regmap_update_bits(smi->map, RTL8365MB_INTR_CTRL_REG,
+ RTL8365MB_INTR_LINK_CHANGE_MASK,
+ FIELD_PREP(RTL8365MB_INTR_LINK_CHANGE_MASK,
+ enable ? 1 : 0));
+}
+
+static int rtl8365mb_irq_enable(struct realtek_smi *smi)
+{
+ return rtl8365mb_set_irq_enable(smi, true);
+}
+
+static int rtl8365mb_irq_disable(struct realtek_smi *smi)
+{
+ return rtl8365mb_set_irq_enable(smi, false);
+}
+
+static int rtl8365mb_irq_setup(struct realtek_smi *smi)
+{
+ struct rtl8365mb *mb = smi->chip_data;
+ struct device_node *intc;
+ u32 irq_trig;
+ int virq;
+ int irq;
+ u32 val;
+ int ret;
+ int i;
+
+ intc = of_get_child_by_name(smi->dev->of_node, "interrupt-controller");
+ if (!intc) {
+ dev_err(smi->dev, "missing child interrupt-controller node\n");
+ return -EINVAL;
+ }
+
+ smi->irqdomain = irq_domain_add_linear(intc, smi->num_ports,
+ &rtl8365mb_irqdomain_ops, smi);
+ if (!smi->irqdomain) {
+ dev_err(smi->dev, "failed to add irq domain\n");
+ ret = -ENOMEM;
+ goto out_put_node;
+ }
+
+ for (i = 0; i < smi->num_ports; i++) {
+ virq = irq_create_mapping(smi->irqdomain, i);
+ if (!virq) {
+ dev_err(smi->dev,
+ "failed to create irq domain mapping\n");
+ ret = -EINVAL;
+ goto out_remove_irqdomain;
+ }
+
+ irq_set_parent(virq, irq);
+ }
+
+ /* rtl8365mb IRQs cascade off this one */
+ irq = of_irq_get(intc, 0);
+ if (irq <= 0) {
+ if (irq != -EPROBE_DEFER)
+ dev_err(smi->dev, "failed to get parent irq: %d\n",
+ irq);
+ ret = irq ? irq : -EINVAL;
+ goto out_remove_irqdomain;
+ }
+
+ /* Configure chip interrupt signal polarity */
+ irq_trig = irqd_get_trigger_type(irq_get_irq_data(irq));
+ switch (irq_trig) {
+ case IRQF_TRIGGER_RISING:
+ case IRQF_TRIGGER_HIGH:
+ val = RTL8365MB_INTR_POLARITY_HIGH;
+ break;
+ case IRQF_TRIGGER_FALLING:
+ case IRQF_TRIGGER_LOW:
+ val = RTL8365MB_INTR_POLARITY_LOW;
+ break;
+ default:
+ dev_err(smi->dev, "unsupported irq trigger type %u\n",
+ irq_trig);
+ ret = -EINVAL;
+ goto out_remove_irqdomain;
+ }
+
+ ret = regmap_update_bits(smi->map, RTL8365MB_INTR_POLARITY_REG,
+ RTL8365MB_INTR_POLARITY_MASK,
+ FIELD_PREP(RTL8365MB_INTR_POLARITY_MASK, val));
+ if (ret)
+ goto out_remove_irqdomain;
+
+ /* Disable the interrupt in case the chip has it enabled on reset */
+ ret = rtl8365mb_irq_disable(smi);
+ if (ret)
+ goto out_remove_irqdomain;
+
+ /* Clear the interrupt status register */
+ ret = regmap_write(smi->map, RTL8365MB_INTR_STATUS_REG,
+ RTL8365MB_INTR_ALL_MASK);
+ if (ret)
+ goto out_remove_irqdomain;
+
+ ret = request_threaded_irq(irq, NULL, rtl8365mb_irq, IRQF_ONESHOT,
+ "rtl8365mb", smi);
+ if (ret) {
+ dev_err(smi->dev, "failed to request irq: %d\n", ret);
+ goto out_remove_irqdomain;
+ }
+
+ /* Store the irq so that we know to free it during teardown */
+ mb->irq = irq;
+
+ ret = rtl8365mb_irq_enable(smi);
+ if (ret)
+ goto out_free_irq;
+
+ of_node_put(intc);
+
+ return 0;
+
+out_free_irq:
+ free_irq(mb->irq, smi);
+ mb->irq = 0;
+
+out_remove_irqdomain:
+ for (i = 0; i < smi->num_ports; i++) {
+ virq = irq_find_mapping(smi->irqdomain, i);
+ irq_dispose_mapping(virq);
+ }
+
+ irq_domain_remove(smi->irqdomain);
+ smi->irqdomain = NULL;
+
+out_put_node:
+ of_node_put(intc);
+
+ return ret;
+}
+
+static void rtl8365mb_irq_teardown(struct realtek_smi *smi)
+{
+ struct rtl8365mb *mb = smi->chip_data;
+ int virq;
+ int i;
+
+ if (mb->irq) {
+ free_irq(mb->irq, smi);
+ mb->irq = 0;
+ }
+
+ if (smi->irqdomain) {
+ for (i = 0; i < smi->num_ports; i++) {
+ virq = irq_find_mapping(smi->irqdomain, i);
+ irq_dispose_mapping(virq);
+ }
+
+ irq_domain_remove(smi->irqdomain);
+ smi->irqdomain = NULL;
+ }
+}
+
+static int rtl8365mb_cpu_config(struct realtek_smi *smi)
+{
+ struct rtl8365mb *mb = smi->chip_data;
+ struct rtl8365mb_cpu *cpu = &mb->cpu;
+ u32 val;
+ int ret;
+
+ ret = regmap_update_bits(smi->map, RTL8365MB_CPU_PORT_MASK_REG,
+ RTL8365MB_CPU_PORT_MASK_MASK,
+ FIELD_PREP(RTL8365MB_CPU_PORT_MASK_MASK,
+ cpu->mask));
+ if (ret)
+ return ret;
+
+ val = FIELD_PREP(RTL8365MB_CPU_CTRL_EN_MASK, cpu->enable ? 1 : 0) |
+ FIELD_PREP(RTL8365MB_CPU_CTRL_INSERTMODE_MASK, cpu->insert) |
+ FIELD_PREP(RTL8365MB_CPU_CTRL_TAG_POSITION_MASK, cpu->position) |
+ FIELD_PREP(RTL8365MB_CPU_CTRL_RXBYTECOUNT_MASK, cpu->rx_length) |
+ FIELD_PREP(RTL8365MB_CPU_CTRL_TAG_FORMAT_MASK, cpu->format) |
+ FIELD_PREP(RTL8365MB_CPU_CTRL_TRAP_PORT_MASK, cpu->trap_port) |
+ FIELD_PREP(RTL8365MB_CPU_CTRL_TRAP_PORT_EXT_MASK,
+ cpu->trap_port >> 3);
+ ret = regmap_write(smi->map, RTL8365MB_CPU_CTRL_REG, val);
+ if (ret)
+ return ret;
+
+ return 0;
+}
+
+static int rtl8365mb_switch_init(struct realtek_smi *smi)
+{
+ struct rtl8365mb *mb = smi->chip_data;
+ int ret;
+ int i;
+
+ /* Do any chip-specific init jam before getting to the common stuff */
+ if (mb->jam_table) {
+ for (i = 0; i < mb->jam_size; i++) {
+ ret = regmap_write(smi->map, mb->jam_table[i].reg,
+ mb->jam_table[i].val);
+ if (ret)
+ return ret;
+ }
+ }
+
+ /* Common init jam */
+ for (i = 0; i < ARRAY_SIZE(rtl8365mb_init_jam_common); i++) {
+ ret = regmap_write(smi->map, rtl8365mb_init_jam_common[i].reg,
+ rtl8365mb_init_jam_common[i].val);
+ if (ret)
+ return ret;
+ }
+
+ return 0;
+}
+
+static int rtl8365mb_reset_chip(struct realtek_smi *smi)
+{
+ u32 val;
+
+ realtek_smi_write_reg_noack(smi, RTL8365MB_CHIP_RESET_REG,
+ FIELD_PREP(RTL8365MB_CHIP_RESET_HW_MASK,
+ 1));
+
+ /* Realtek documentation says the chip needs 1 second to reset. Sleep
+ * for 100 ms before accessing any registers to prevent ACK timeouts.
+ */
+ msleep(100);
+ return regmap_read_poll_timeout(smi->map, RTL8365MB_CHIP_RESET_REG, val,
+ !(val & RTL8365MB_CHIP_RESET_HW_MASK),
+ 20000, 1e6);
+}
+
+static int rtl8365mb_setup(struct dsa_switch *ds)
+{
+ struct realtek_smi *smi = ds->priv;
+ struct rtl8365mb *mb;
+ int ret;
+ int i;
+
+ mb = smi->chip_data;
+
+ ret = rtl8365mb_reset_chip(smi);
+ if (ret) {
+ dev_err(smi->dev, "failed to reset chip: %d\n", ret);
+ goto out_error;
+ }
+
+ /* Configure switch to vendor-defined initial state */
+ ret = rtl8365mb_switch_init(smi);
+ if (ret) {
+ dev_err(smi->dev, "failed to initialize switch: %d\n", ret);
+ goto out_error;
+ }
+
+ /* Set up cascading IRQs */
+ ret = rtl8365mb_irq_setup(smi);
+ if (ret == -EPROBE_DEFER)
+ return ret;
+ else if (ret)
+ dev_info(smi->dev, "no interrupt support\n");
+
+ /* Configure CPU tagging */
+ ret = rtl8365mb_cpu_config(smi);
+ if (ret)
+ goto out_teardown_irq;
+
+ /* Configure ports */
+ for (i = 0; i < smi->num_ports; i++) {
+ struct rtl8365mb_port *p = &mb->ports[i];
+
+ if (dsa_is_unused_port(smi->ds, i))
+ continue;
+
+ /* Set up per-port private data */
+ p->smi = smi;
+ p->index = i;
+
+ /* Forward only to the CPU */
+ ret = rtl8365mb_port_set_isolation(smi, i, BIT(smi->cpu_port));
+ if (ret)
+ goto out_teardown_irq;
+
+ /* Disable learning */
+ ret = rtl8365mb_port_set_learning(smi, i, false);
+ if (ret)
+ goto out_teardown_irq;
+
+ /* Set the initial STP state of all ports to DISABLED, otherwise
+ * ports will still forward frames to the CPU despite being
+ * administratively down by default.
+ */
+ rtl8365mb_port_stp_state_set(smi->ds, i, BR_STATE_DISABLED);
+ }
+
+ /* Set maximum packet length to 1536 bytes */
+ ret = regmap_update_bits(smi->map, RTL8365MB_CFG0_MAX_LEN_REG,
+ RTL8365MB_CFG0_MAX_LEN_MASK,
+ FIELD_PREP(RTL8365MB_CFG0_MAX_LEN_MASK, 1536));
+ if (ret)
+ goto out_teardown_irq;
+
+ ret = realtek_smi_setup_mdio(smi);
+ if (ret) {
+ dev_err(smi->dev, "could not set up MDIO bus\n");
+ goto out_teardown_irq;
+ }
+
+ /* Start statistics counter polling */
+ rtl8365mb_stats_setup(smi);
+
+ return 0;
+
+out_teardown_irq:
+ rtl8365mb_irq_teardown(smi);
+
+out_error:
+ return ret;
+}
+
+static void rtl8365mb_teardown(struct dsa_switch *ds)
+{
+ struct realtek_smi *smi = ds->priv;
+
+ rtl8365mb_stats_teardown(smi);
+ rtl8365mb_irq_teardown(smi);
+}
+
+static int rtl8365mb_get_chip_id_and_ver(struct regmap *map, u32 *id, u32 *ver)
+{
+ int ret;
+
+ /* For some reason we have to write a magic value to an arbitrary
+ * register whenever accessing the chip ID/version registers.
+ */
+ ret = regmap_write(map, RTL8365MB_MAGIC_REG, RTL8365MB_MAGIC_VALUE);
+ if (ret)
+ return ret;
+
+ ret = regmap_read(map, RTL8365MB_CHIP_ID_REG, id);
+ if (ret)
+ return ret;
+
+ ret = regmap_read(map, RTL8365MB_CHIP_VER_REG, ver);
+ if (ret)
+ return ret;
+
+ /* Reset magic register */
+ ret = regmap_write(map, RTL8365MB_MAGIC_REG, 0);
+ if (ret)
+ return ret;
+
+ return 0;
+}
+
+static int rtl8365mb_detect(struct realtek_smi *smi)
+{
+ struct rtl8365mb *mb = smi->chip_data;
+ u32 chip_id;
+ u32 chip_ver;
+ int ret;
+
+ ret = rtl8365mb_get_chip_id_and_ver(smi->map, &chip_id, &chip_ver);
+ if (ret) {
+ dev_err(smi->dev, "failed to read chip id and version: %d\n",
+ ret);
+ return ret;
+ }
+
+ switch (chip_id) {
+ case RTL8365MB_CHIP_ID_8365MB_VC:
+ dev_info(smi->dev,
+ "found an RTL8365MB-VC switch (ver=0x%04x)\n",
+ chip_ver);
+
+ smi->cpu_port = RTL8365MB_CPU_PORT_NUM_8365MB_VC;
+ smi->num_ports = smi->cpu_port + 1;
+
+ mb->smi = smi;
+ mb->chip_id = chip_id;
+ mb->chip_ver = chip_ver;
+ mb->port_mask = BIT(smi->num_ports) - 1;
+ mb->learn_limit_max = RTL8365MB_LEARN_LIMIT_MAX_8365MB_VC;
+ mb->jam_table = rtl8365mb_init_jam_8365mb_vc;
+ mb->jam_size = ARRAY_SIZE(rtl8365mb_init_jam_8365mb_vc);
+
+ mb->cpu.enable = 1;
+ mb->cpu.mask = BIT(smi->cpu_port);
+ mb->cpu.trap_port = smi->cpu_port;
+ mb->cpu.insert = RTL8365MB_CPU_INSERT_TO_ALL;
+ mb->cpu.position = RTL8365MB_CPU_POS_AFTER_SA;
+ mb->cpu.rx_length = RTL8365MB_CPU_RXLEN_64BYTES;
+ mb->cpu.format = RTL8365MB_CPU_FORMAT_8BYTES;
+
+ break;
+ default:
+ dev_err(smi->dev,
+ "found an unknown Realtek switch (id=0x%04x, ver=0x%04x)\n",
+ chip_id, chip_ver);
+ return -ENODEV;
+ }
+
+ return 0;
+}
+
+static const struct dsa_switch_ops rtl8365mb_switch_ops = {
+ .get_tag_protocol = rtl8365mb_get_tag_protocol,
+ .setup = rtl8365mb_setup,
+ .teardown = rtl8365mb_teardown,
+ .phylink_validate = rtl8365mb_phylink_validate,
+ .phylink_mac_config = rtl8365mb_phylink_mac_config,
+ .phylink_mac_link_down = rtl8365mb_phylink_mac_link_down,
+ .phylink_mac_link_up = rtl8365mb_phylink_mac_link_up,
+ .port_stp_state_set = rtl8365mb_port_stp_state_set,
+ .get_strings = rtl8365mb_get_strings,
+ .get_ethtool_stats = rtl8365mb_get_ethtool_stats,
+ .get_sset_count = rtl8365mb_get_sset_count,
+ .get_eth_phy_stats = rtl8365mb_get_phy_stats,
+ .get_eth_mac_stats = rtl8365mb_get_mac_stats,
+ .get_eth_ctrl_stats = rtl8365mb_get_ctrl_stats,
+ .get_stats64 = rtl8365mb_get_stats64,
+};
+
+static const struct realtek_smi_ops rtl8365mb_smi_ops = {
+ .detect = rtl8365mb_detect,
+ .phy_read = rtl8365mb_phy_read,
+ .phy_write = rtl8365mb_phy_write,
+};
+
+const struct realtek_smi_variant rtl8365mb_variant = {
+ .ds_ops = &rtl8365mb_switch_ops,
+ .ops = &rtl8365mb_smi_ops,
+ .clk_delay = 10,
+ .cmd_read = 0xb9,
+ .cmd_write = 0xb8,
+ .chip_data_sz = sizeof(struct rtl8365mb),
+};
+EXPORT_SYMBOL_GPL(rtl8365mb_variant);
--
2.32.0
On Fri, 15 Oct 2021 19:10:27 +0200 Alvin Šipraga wrote:
> v2 -> v3:
> - collect Florian's Reviewed-by
> - move IRQ setup earlier in probe per Florian's suggestion
> - follow Jakub's suggestion and use the standard ethtool stats API
Thanks a lot for doing this. The code LGTM, the only thing that stands
out is the use of spin_locks(). I couldn't quickly parse out what bus
this device hangs off, if it's MMIO and registers can be read without
sleeping you could potentially get rid of the delayed work to read
stats, but I think you need to switch to
regmap_read_poll_timeout_atomic() because regmap_read_poll_timeout()
itself can sleep.
If the register access sleeps (I2C, SPI, MDIO etc) you need to switch
from a spin lock to a mutex.
Either way CONFIG_DEBUG_ATOMIC_SLEEP is your friend.
On Fri, Oct 15, 2021 at 07:10:26PM +0200, Alvin Šipraga wrote:
> +static struct sk_buff *rtl8_4_tag_xmit(struct sk_buff *skb,
> + struct net_device *dev)
> +{
> + struct dsa_port *dp = dsa_slave_to_port(dev);
> + __be16 *tag;
> +
> + /* Pad out so the (stripped) packet is at least 64 bytes long
> + * (including FCS), otherwise the switch will drop the packet.
> + * Then we need an additional 8 bytes for the Realtek tag.
> + */
> + if (unlikely(__skb_put_padto(skb, ETH_ZLEN + RTL8_4_TAG_LEN, false)))
> + return NULL;
Is this still required if you set rtl8365mb_cpu_rxlen to 64 bytes already?
> +
> + skb_push(skb, RTL8_4_TAG_LEN);
> +
Tested on net-next master branch kernel on Asus RT-AC88U router using
OpenWrt.
DT specification:
ethernet-switch {
compatible = "realtek,rtl8365mb";
mdc-gpios = <&chipcommon 6 GPIO_ACTIVE_HIGH>;
mdio-gpios = <&chipcommon 7 GPIO_ACTIVE_HIGH>;
reset-gpios = <&chipcommon 10 GPIO_ACTIVE_LOW>;
realtek,disable-leds;
dsa,member = <1 0>;
ports {
#address-cells = <1>;
#size-cells = <0>;
reg = <0>;
port@0 {
reg = <0>;
label = "lan5";
phy-handle = <ðphy0>;
};
port@1 {
reg = <1>;
label = "lan6";
phy-handle = <ðphy1>;
};
port@2 {
reg = <2>;
label = "lan7";
phy-handle = <ðphy2>;
};
port@3 {
reg = <3>;
label = "lan8";
phy-handle = <ðphy3>;
};
port@6 {
reg = <6>;
label = "cpu";
ethernet = <&sw0_p5>;
phy-mode = "rgmii";
tx-internal-delay-ps = <2000>;
rx-internal-delay-ps = <2000>;
fixed-link {
speed = <1000>;
full-duplex;
pause;
};
};
};
mdio {
compatible = "realtek,smi-mdio";
#address-cells = <1>;
#size-cells = <0>;
ethphy0: phy@0 {
reg = <0>;
};
ethphy1: phy@1 {
reg = <1>;
};
ethphy2: phy@2 {
reg = <2>;
};
ethphy3: phy@3 {
reg = <3>;
};
};
};
Tested-by: Arınç ÜNAL <[email protected]>
On 10/16/21 2:40 AM, Jakub Kicinski wrote:
> On Fri, 15 Oct 2021 19:10:27 +0200 Alvin Šipraga wrote:
>> v2 -> v3:
>> - collect Florian's Reviewed-by
>> - move IRQ setup earlier in probe per Florian's suggestion
>> - follow Jakub's suggestion and use the standard ethtool stats API
>
> Thanks a lot for doing this. The code LGTM, the only thing that stands
> out is the use of spin_locks(). I couldn't quickly parse out what bus
> this device hangs off, if it's MMIO and registers can be read without
> sleeping you could potentially get rid of the delayed work to read
> stats, but I think you need to switch to
> regmap_read_poll_timeout_atomic() because regmap_read_poll_timeout()
> itself can sleep.
It's an "SMI" bus bit-banged with GPIO - details are in
realtek-smi-core.c - a Realtek peculiarity. Initially I thought this was
implementation was sleeping, but actually it's using ndelay() which is
OK in atomic context if I'm not mistaken? So I guess you're right, I can
get rid of the delayed work. :)
>
> If the register access sleeps (I2C, SPI, MDIO etc) you need to switch
> from a spin lock to a mutex.
>
> Either way CONFIG_DEBUG_ATOMIC_SLEEP is your friend.
Thanks, I'll test with this and follow up. I see a kernel test robot
warning so I will be sending v4 anyway.
On Fri, Oct 15, 2021 at 7:11 PM Alvin Šipraga <[email protected]> wrote:
> This patch adds a realtek-smi subdriver for the RTL8365MB-VC 4+1 port
> 10/100/1000M switch controller. The driver has been developed based on a
> GPL-licensed OS-agnostic Realtek vendor driver known as rtl8367c found
> in the OpenWrt source tree.
This version is really nice!
> v2 -> v3:
> - following the stats change, it became apparent that the rtl8366
> helper library is no longer that helpful; scrap it and implement
> the ethtool ops specifically for this chip
I wonder if I should just scrap it in general when I convert RTL8366RB over
to use your new nice callbacks? Just push that part down into the
driver for each chip.
I vaguely remember the RTL8366S having the same counters though...
(I think this is why it is in the shared library at all.)
Yours,
Linus Walleij
On 10/16/21 8:04 AM, DENG Qingfang wrote:
> On Fri, Oct 15, 2021 at 07:10:26PM +0200, Alvin Šipraga wrote:
>> +static struct sk_buff *rtl8_4_tag_xmit(struct sk_buff *skb,
>> + struct net_device *dev)
>> +{
>> + struct dsa_port *dp = dsa_slave_to_port(dev);
>> + __be16 *tag;
>> +
>> + /* Pad out so the (stripped) packet is at least 64 bytes long
>> + * (including FCS), otherwise the switch will drop the packet.
>> + * Then we need an additional 8 bytes for the Realtek tag.
>> + */
>> + if (unlikely(__skb_put_padto(skb, ETH_ZLEN + RTL8_4_TAG_LEN, false)))
>> + return NULL;
>
> Is this still required if you set rtl8365mb_cpu_rxlen to 64 bytes already?
Apparently not - nice catch. I've removed it for what will be v4. :)
>
>> +
>> + skb_push(skb, RTL8_4_TAG_LEN);
>> +