kernel: Import latest kkit-linux-6.5.y

- Fixes the locking warnings when configuring LEDs on 6393X - Correctly power up 3310X PHYs strapped to start powered down - Fix kernel oops when reading from blank NVMEM configured with an ONIE layout
2026-07-31 21:13:00 +02:00 · 2023-11-25 20:37:15 +01:00
parent 6c6ae844f3
commit cb0f8739ad
17 changed files with 132 additions and 22 deletions
@@ -1,7 +1,7 @@
 From e1011848d2c9476e20e6a4053b8ab46ce2a1506f Mon Sep 17 00:00:00 2001
 From: Tobias Waldekranz <tobias@waldekranz.com>
 Date: Thu, 5 Oct 2023 16:09:25 +0200
-Subject: [PATCH 01/15] eth: fwnode: Fall back to NVMEM provided MAC address
+Subject: [PATCH 01/17] eth: fwnode: Fall back to NVMEM provided MAC address
 from OF

 In the cases where the fwnode is an OF node, make use of any NVMEM
@@ -1,7 +1,7 @@
 From 1d6f30fce9f4516730439f72ee2134301d37654d Mon Sep 17 00:00:00 2001
 From: Tobias Waldekranz <tobias@waldekranz.com>
 Date: Tue, 5 Sep 2023 13:08:29 +0200
-Subject: [PATCH 02/15] net: dsa: mv88e6xxx: Create API to read a single stat
+Subject: [PATCH 02/17] net: dsa: mv88e6xxx: Create API to read a single stat
 counter

 This change contains no functional change. We simply push the hardware
@@ -1,7 +1,7 @@
 From f960475a96d2852d06ee062c970e35074f87659c Mon Sep 17 00:00:00 2001
 From: Tobias Waldekranz <tobias@waldekranz.com>
 Date: Tue, 5 Sep 2023 21:10:15 +0200
-Subject: [PATCH 03/15] net: dsa: mv88e6xxx: Give each hw stat an ID
+Subject: [PATCH 03/17] net: dsa: mv88e6xxx: Give each hw stat an ID

 With the upcoming standard counter group support, we are no longer
 reading out the whole set of counters, but rather mapping a subset to
@@ -1,7 +1,7 @@
 From 6a612f46152d8567a16b7b15fdcc3081cbdbb3fb Mon Sep 17 00:00:00 2001
 From: Tobias Waldekranz <tobias@waldekranz.com>
 Date: Tue, 5 Sep 2023 21:15:52 +0200
-Subject: [PATCH 04/15] net: dsa: mv88e6xxx: Add "eth-mac" counter group
+Subject: [PATCH 04/17] net: dsa: mv88e6xxx: Add "eth-mac" counter group
 support

 After this change, mv88e6xxx ports can report the applicable subset of
@@ -1,7 +1,7 @@
 From a85c55f3232d402780a22d42859bdc81150469b4 Mon Sep 17 00:00:00 2001
 From: Tobias Waldekranz <tobias@waldekranz.com>
 Date: Wed, 6 Sep 2023 10:55:53 +0200
-Subject: [PATCH 05/15] net: dsa: mv88e6xxx: Add "rmon" counter group support
+Subject: [PATCH 05/17] net: dsa: mv88e6xxx: Add "rmon" counter group support

 After this change, mv88e6xxx ports can report the applicable subset of
 their counters using ethtool's standardized "rmon" counter group.
@@ -1,7 +1,7 @@
 From 6f684b1b9eb8a307b77aa4baa8531e2c423112af Mon Sep 17 00:00:00 2001
 From: Tobias Waldekranz <tobias@waldekranz.com>
 Date: Thu, 21 Sep 2023 11:10:26 +0200
-Subject: [PATCH 06/15] arm64: dts: marvell: cp11x: Provide clock names for
+Subject: [PATCH 06/17] arm64: dts: marvell: cp11x: Provide clock names for
 MDIO controllers

 This will let the driver figure out the rate of the core clk, such
@@ -1,7 +1,7 @@
 From d5e03e7b44c8ac414927751097d8d65caffc33b4 Mon Sep 17 00:00:00 2001
 From: Tobias Waldekranz <tobias@waldekranz.com>
 Date: Tue, 19 Sep 2023 18:19:40 +0200
-Subject: [PATCH 07/15] net: mvmdio: Avoid excessive sleeps in polled mode
+Subject: [PATCH 07/17] net: mvmdio: Avoid excessive sleeps in polled mode

 Before this change, when operating in polled mode, i.e. no IRQ is
 available, every individual C45 access would be hit with a 150us sleep
@@ -1,7 +1,7 @@
 From d389cc01295ebd7be228c8240f0061134b27f215 Mon Sep 17 00:00:00 2001
 From: Tobias Waldekranz <tobias@waldekranz.com>
 Date: Thu, 21 Sep 2023 11:12:55 +0200
-Subject: [PATCH 08/15] net: mvmdio: Support setting the MDC frequency on XSMI
+Subject: [PATCH 08/17] net: mvmdio: Support setting the MDC frequency on XSMI
 controllers

 Support the standard "clock-frequency" attribute to set the generated
@@ -1,7 +1,7 @@
 From 46bbea6b9d692ac7460e4e278cedee9f88b6da0b Mon Sep 17 00:00:00 2001
 From: Tobias Waldekranz <tobias@waldekranz.com>
 Date: Tue, 19 Sep 2023 18:38:10 +0200
-Subject: [PATCH 09/15] net: phy: marvell10g: Support firmware loading on
+Subject: [PATCH 09/17] net: phy: marvell10g: Support firmware loading on
 88X3310

 If a device is found waiting for firmware to be loaded into its RAM,
@@ -1,7 +1,7 @@
 From 312d3efde0fbe7cec3bf7ba1908c213c9cf69337 Mon Sep 17 00:00:00 2001
 From: Andrew Lunn <andrew@lunn.ch>
 Date: Tue, 8 Aug 2023 23:04:33 +0200
-Subject: [PATCH 10/15] led: trig: netdev: Fix requesting offload device
+Subject: [PATCH 10/17] led: trig: netdev: Fix requesting offload device

 When the netdev trigger is activates, it tries to determine what
 device the LED blinks for, and what the current blink mode is.
@@ -1,7 +1,7 @@
 From 83322015e7960ba0f1f3698e36333cfb6aecb931 Mon Sep 17 00:00:00 2001
 From: Andrew Lunn <andrew@lunn.ch>
 Date: Tue, 8 Aug 2023 23:04:36 +0200
-Subject: [PATCH 11/15] leds: trig-netdev: Disable offload on deactivation of
+Subject: [PATCH 11/17] leds: trig-netdev: Disable offload on deactivation of
 trigger

 Ensure that the offloading of blinking is stopped when the trigger is
@@ -1,7 +1,7 @@
 From a4f09c743da21f09b0fcc6e5842d665490f9f70e Mon Sep 17 00:00:00 2001
 From: Andrew Lunn <andrew@lunn.ch>
 Date: Tue, 8 Aug 2023 23:04:34 +0200
-Subject: [PATCH 12/15] net: phy: phy_device: Call into the PHY driver to set
+Subject: [PATCH 12/17] net: phy: phy_device: Call into the PHY driver to set
 LED offload

 Linux LEDs can be requested to perform hardware accelerated blinking
@@ -1,7 +1,7 @@
 From 5adf394404d4c8ac25fb1e36d7d3f23b699b3104 Mon Sep 17 00:00:00 2001
 From: Tobias Waldekranz <tobias@waldekranz.com>
 Date: Wed, 15 Nov 2023 20:58:42 +0100
-Subject: [PATCH 13/15] net: phy: marvell10g: Add LED support for 88X3310
+Subject: [PATCH 13/17] net: phy: marvell10g: Add LED support for 88X3310

 Pickup the LEDs from the state in which the hardware reset or
 bootloader left them, but also support further configuration via
@@ -1,7 +1,7 @@
 From 913a57a304e8024f4cea872fa904cee867c469a2 Mon Sep 17 00:00:00 2001
 From: Tobias Waldekranz <tobias@waldekranz.com>
 Date: Thu, 16 Nov 2023 19:44:32 +0100
-Subject: [PATCH 14/15] net: dsa: mv88e6xxx: Add LED infrastructure
+Subject: [PATCH 14/17] net: dsa: mv88e6xxx: Add LED infrastructure

 Parse DT for LEDs and register them for devices that support it,
 though no actual implementations exist yet.
@@ -1,7 +1,7 @@
-From 09fca5d7f66b1bdc6a5f72eb3d36bfd8c88bebcc Mon Sep 17 00:00:00 2001
+From 8510028fa738493a46b235057ed166f9bbbf0ed4 Mon Sep 17 00:00:00 2001
 From: Tobias Waldekranz <tobias@waldekranz.com>
 Date: Thu, 16 Nov 2023 21:59:35 +0100
-Subject: [PATCH 15/15] net: dsa: mv88e6xxx: Add LED support for 6393X
+Subject: [PATCH 15/17] net: dsa: mv88e6xxx: Add LED support for 6393X

 Trigger support:
 - "none"
@@ -11,11 +11,11 @@ Trigger support:
 Signed-off-by: Tobias Waldekranz <tobias@waldekranz.com>
 ---
 drivers/net/dsa/mv88e6xxx/chip.c |   1 +
- drivers/net/dsa/mv88e6xxx/leds.c | 219 +++++++++++++++++++++++++++++++
+ drivers/net/dsa/mv88e6xxx/leds.c | 226 +++++++++++++++++++++++++++++++
 drivers/net/dsa/mv88e6xxx/leds.h |   2 +
 drivers/net/dsa/mv88e6xxx/port.c |  33 +++++
 drivers/net/dsa/mv88e6xxx/port.h |   7 +
- 5 files changed, 262 insertions(+)
+ 5 files changed, 269 insertions(+)

 diff --git a/drivers/net/dsa/mv88e6xxx/chip.c b/drivers/net/dsa/mv88e6xxx/chip.c
 index d343ea5cc4a1..7d6b2b85fc31 100644
@@ -30,7 +30,7 @@ index d343ea5cc4a1..7d6b2b85fc31 100644
 };
 
 diff --git a/drivers/net/dsa/mv88e6xxx/leds.c b/drivers/net/dsa/mv88e6xxx/leds.c
-index 42e1661e1f47..02daf976100d 100644
+index 42e1661e1f47..b2a55d96b506 100644
 --- a/drivers/net/dsa/mv88e6xxx/leds.c
 +++ b/drivers/net/dsa/mv88e6xxx/leds.c
@@ -4,6 +4,13 @@
@@ -47,7 +47,7 @@ index 42e1661e1f47..02daf976100d 100644
 struct mv88e6xxx_led {
 	struct mv88e6xxx_chip *chip;
 	int port;
-@@ -12,6 +19,218 @@ struct mv88e6xxx_led {
+@@ -12,6 +19,225 @@ struct mv88e6xxx_led {
 	struct led_classdev ldev;
 };
 
@@ -163,9 +163,11 @@ index 42e1661e1f47..02daf976100d 100644
 +	if (mode < 0)
 +		return mode;
 +
+	mv88e6xxx_reg_lock(led->chip);
+
 +	err = mv88e6393x_port_led_read(led->chip, led->port, 0, &ctrl);
 +	if (err)
-+		return err;
+		goto out;
 +
 +	switch (led->index) {
 +	case 0:
@@ -177,7 +179,10 @@ index 42e1661e1f47..02daf976100d 100644
 +		ctrl |= mode << 4;
 +	}
 +
-+	return mv88e6393x_port_led_write(led->chip, led->port, 0, ctrl);
+	err = mv88e6393x_port_led_write(led->chip, led->port, 0, ctrl);
+out:
+	mv88e6xxx_reg_unlock(led->chip);
+	return err;
 +}
 +
 +static int mv88e6393x_led_get(struct mv88e6xxx_led *led)
@@ -185,7 +190,9 @@ index 42e1661e1f47..02daf976100d 100644
 +	u16 ctrl;
 +	int err;
 +
+	mv88e6xxx_reg_lock(led->chip);
 +	err = mv88e6393x_port_led_read(led->chip, led->port, 0, &ctrl);
+	mv88e6xxx_reg_unlock(led->chip);
 +	if (err)
 +		return err;
 +
@@ -0,0 +1,49 @@
+From b70f0aa3eb77567a057c2f42f5012a3e009a95c8 Mon Sep 17 00:00:00 2001
+From: Tobias Waldekranz <tobias@waldekranz.com>
+Date: Tue, 21 Nov 2023 20:15:24 +0100
+Subject: [PATCH 16/17] net: phy: marvell10g: Fix power-up when strapped to
+ start powered down
+
+On devices which are hardware strapped to start powered down (PDSTATE
+== 1), make sure that we clear the power-down bit on all units
+affected by this setting.
+
+Signed-off-by: Tobias Waldekranz <tobias@waldekranz.com>
+---
+ drivers/net/phy/marvell10g.c | 18 +++++++++++++++---
+ 1 file changed, 15 insertions(+), 3 deletions(-)
+
+diff --git a/drivers/net/phy/marvell10g.c b/drivers/net/phy/marvell10g.c
+index 05ad7c769e24..1d86ee10650e 100644
+--- a/drivers/net/phy/marvell10g.c
+++ b/drivers/net/phy/marvell10g.c
+@@ -782,11 +782,23 @@ static int mv3310_power_down(struct phy_device *phydev)
+ 
+ static int mv3310_power_up(struct phy_device *phydev)
+ {
+	static const u16 resets[][2] = {
+		{ MDIO_MMD_PCS,    MV_PCS_BASE_R    + MDIO_CTRL1 },
+		{ MDIO_MMD_PCS,    MV_PCS_1000BASEX + MDIO_CTRL1 },
+		{ MDIO_MMD_PCS,    MV_PCS_BASE_T    + MDIO_CTRL1 },
+		{ MDIO_MMD_PMAPMD, MDIO_CTRL1 },
+		{ MDIO_MMD_VEND2,  MV_V2_PORT_CTRL },
+	};
+ 	struct mv3310_priv *priv = dev_get_drvdata(&phydev->mdio.dev);
+-	int ret;
+	int i, ret;
+ 
+-	ret = phy_clear_bits_mmd(phydev, MDIO_MMD_VEND2, MV_V2_PORT_CTRL,
+-				 MV_V2_PORT_CTRL_PWRDOWN);
+	for (i = 0; i < ARRAY_SIZE(resets); i++) {
+		ret = phy_clear_bits_mmd(phydev, resets[i][0], resets[i][1],
+					 MV_V2_PORT_CTRL_PWRDOWN);
+		if (ret)
+			break;
+
+	}
+ 
+ 	/* Sometimes, the power down bit doesn't clear immediately, and
+ 	 * a read of this register causes the bit not to clear. Delay
+-- 
+2.34.1
+
@@ -0,0 +1,54 @@
+From dd693e037f0dd799e77af397b4d23190b66e9cf1 Mon Sep 17 00:00:00 2001
+From: Tobias Waldekranz <tobias@waldekranz.com>
+Date: Fri, 24 Nov 2023 23:29:55 +0100
+Subject: [PATCH 17/17] nvmem: layouts: onie-tlv: Let device probe even when
+ TLV is invalid
+
+Before this change, probing an NVMEM device, expected to contain a
+valid TLV, would fail if it had not been provisioned yet. But an
+obvious way of provisioning is to write the TLV to the nvmem
+attribute, which would not be possible because the device would not
+be successfully probed.
+
+Therefore, settle for reporting data corruption issues in the log, and
+simply refrain from registering any cells in those cases.
+
+Signed-off-by: Tobias Waldekranz <tobias@waldekranz.com>
+---
+ drivers/nvmem/layouts/onie-tlv.c | 6 +++---
+ 1 file changed, 3 insertions(+), 3 deletions(-)
+
+diff --git a/drivers/nvmem/layouts/onie-tlv.c b/drivers/nvmem/layouts/onie-tlv.c
+index 59fc87ccfcff..834c46e46d0a 100644
+--- a/drivers/nvmem/layouts/onie-tlv.c
+++ b/drivers/nvmem/layouts/onie-tlv.c
+@@ -196,7 +196,7 @@ static int onie_tlv_parse_table(struct device *dev, struct nvmem_device *nvmem,
+ 
+ 	if (!onie_tlv_hdr_is_valid(dev, &hdr)) {
+ 		dev_err(dev, "Invalid ONIE TLV header\n");
+-		return -EINVAL;
+		return 0;
+ 	}
+ 
+ 	hdr_len = sizeof(hdr.id) + sizeof(hdr.version) + sizeof(hdr.data_len);
+@@ -204,7 +204,7 @@ static int onie_tlv_parse_table(struct device *dev, struct nvmem_device *nvmem,
+ 	table_len = hdr_len + data_len;
+ 	if (table_len > ONIE_TLV_MAX_LEN) {
+ 		dev_err(dev, "Invalid ONIE TLV data length\n");
+-		return -EINVAL;
+		return 0;
+ 	}
+ 
+ 	table = devm_kmalloc(dev, table_len, GFP_KERNEL);
+@@ -216,7 +216,7 @@ static int onie_tlv_parse_table(struct device *dev, struct nvmem_device *nvmem,
+ 		return ret;
+ 
+ 	if (!onie_tlv_crc_is_valid(dev, table_len, table))
+-		return -EINVAL;
+		return 0;
+ 
+ 	data = table + hdr_len;
+ 	ret = onie_tlv_add_cells(dev, nvmem, data_len, data);
+-- 
+2.34.1
+