[PATCH linux dev-6.6 0/3] Backport MCTP Over USB Binding

[PATCH linux dev-6.6 0/3] Backport MCTP Over USB Binding

[Santosh Puranik]

This series backports the MCTP over USB binding driver and associated
definitions.

Patches 1 and 2 were cherry-picked from upstream linux master and
patch 3 includes fixes needed to backport the binding to 6.6.

This was tested against NVIDIA BMC (machine gb200nvl-obmc) running
on ASPEED AST2600.

Jeremy Kerr (2):
  usb: Add base USB MCTP definitions
  net: mctp: Add MCTP USB transport driver

Santosh Puranik (1):
  drivers: net: mctp: usb: Port for kernel 6.6

 MAINTAINERS                  |   1 +
 drivers/net/mctp/Kconfig     |  10 +
 drivers/net/mctp/Makefile    |   1 +
 drivers/net/mctp/mctp-usb.c  | 387 +++++++++++++++++++++++++++++++++++
 include/linux/usb/mctp-usb.h |  30 +++
 include/uapi/linux/usb/ch9.h |   1 +
 6 files changed, 430 insertions(+)
 create mode 100644 drivers/net/mctp/mctp-usb.c
 create mode 100644 include/linux/usb/mctp-usb.h

-- 
2.39.5

[PATCH linux dev-6.6 1/3] usb: Add base USB MCTP definitions

[Santosh Puranik]

From: Jeremy Kerr <jk@codeconstruct.com.au>

Upcoming changes will add a USB host (and later gadget) driver for the
MCTP-over-USB protocol. Add a header that provides common definitions
for protocol support: the packet header format and a few framing
definitions. Add a define for the MCTP class code, as per
https://usb.org/defined-class-codes.

Signed-off-by: Jeremy Kerr <jk@codeconstruct.com.au>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Link: https://patch.msgid.link/20250221-dev-mctp-usb-v3-1-3353030fe9cc@codeconstruct.com.au
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
(cherry picked from commit dcc35baae732b9079b2c6595cfd86da02b34a4e6)
Signed-off-by: Santosh Puranik <santosh.puranik.ibm@gmail.com>
---
 MAINTAINERS                  |  1 +
 include/linux/usb/mctp-usb.h | 30 ++++++++++++++++++++++++++++++
 include/uapi/linux/usb/ch9.h |  1 +
 3 files changed, 32 insertions(+)
 create mode 100644 include/linux/usb/mctp-usb.h

diff --git a/MAINTAINERS b/MAINTAINERS
index 6afbb4f7b33c..5da10b86be0b 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -12647,6 +12647,7 @@ L:	netdev@vger.kernel.org
 S:	Maintained
 F:	Documentation/networking/mctp.rst
 F:	drivers/net/mctp/
+F:	include/linux/usb/mctp-usb.h
 F:	include/net/mctp.h
 F:	include/net/mctpdevice.h
 F:	include/net/netns/mctp.h
diff --git a/include/linux/usb/mctp-usb.h b/include/linux/usb/mctp-usb.h
new file mode 100644
index 000000000000..a2f6f1e04efb
--- /dev/null
+++ b/include/linux/usb/mctp-usb.h
@@ -0,0 +1,30 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * mctp-usb.h - MCTP USB transport binding: common definitions,
+ * based on DMTF0283 specification:
+ * https://www.dmtf.org/sites/default/files/standards/documents/DSP0283_1.0.1.pdf
+ *
+ * These are protocol-level definitions, that may be shared between host
+ * and gadget drivers.
+ *
+ * Copyright (C) 2024-2025 Code Construct Pty Ltd
+ */
+
+#ifndef __LINUX_USB_MCTP_USB_H
+#define __LINUX_USB_MCTP_USB_H
+
+#include <linux/types.h>
+
+struct mctp_usb_hdr {
+	__be16	id;
+	u8	rsvd;
+	u8	len;
+} __packed;
+
+#define MCTP_USB_XFER_SIZE	512
+#define MCTP_USB_BTU		68
+#define MCTP_USB_MTU_MIN	MCTP_USB_BTU
+#define MCTP_USB_MTU_MAX	(U8_MAX - sizeof(struct mctp_usb_hdr))
+#define MCTP_USB_DMTF_ID	0x1ab4
+
+#endif /*  __LINUX_USB_MCTP_USB_H */
diff --git a/include/uapi/linux/usb/ch9.h b/include/uapi/linux/usb/ch9.h
index 8a147abfc680..dce954da45ac 100644
--- a/include/uapi/linux/usb/ch9.h
+++ b/include/uapi/linux/usb/ch9.h
@@ -327,6 +327,7 @@ struct usb_device_descriptor {
 #define USB_CLASS_AUDIO_VIDEO		0x10
 #define USB_CLASS_BILLBOARD		0x11
 #define USB_CLASS_USB_TYPE_C_BRIDGE	0x12
+#define USB_CLASS_MCTP			0x14
 #define USB_CLASS_MISC			0xef
 #define USB_CLASS_APP_SPEC		0xfe
 #define USB_CLASS_VENDOR_SPEC		0xff
-- 
2.39.5

[PATCH linux dev-6.6 2/3] net: mctp: Add MCTP USB transport driver

[Santosh Puranik]

From: Jeremy Kerr <jk@codeconstruct.com.au>

Add an implementation for DMTF DSP0283, which defines a MCTP-over-USB
transport. As per that spec, we're restricted to full speed mode,
requiring 512-byte transfers.

Each MCTP-over-USB interface is a peer-to-peer link to a single MCTP
endpoint, so no physical addressing is required (of course, that MCTP
endpoint may then bridge to further MCTP endpoints). Consequently,
interfaces will report with no lladdr data:

    # mctp link
    dev lo index 1 address 00:00:00:00:00:00 net 1 mtu 65536 up
    dev mctpusb0 index 6 address none net 1 mtu 68 up

This is a simple initial implementation, with single rx & tx urbs, and
no multi-packet tx transfers - although we do accept multi-packet rx
from the device.

Includes suggested fixes from Santosh Puranik <spuranik@nvidia.com>.

Signed-off-by: Jeremy Kerr <jk@codeconstruct.com.au>
Cc: Santosh Puranik <spuranik@nvidia.com>
Link: https://patch.msgid.link/20250221-dev-mctp-usb-v3-2-3353030fe9cc@codeconstruct.com.au
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
(cherry picked from commit 0791c0327a6e4e7691d6fc5ad334c215de04dcc9)
Signed-off-by: Santosh Puranik <santosh.puranik.ibm@gmail.com>
---
 drivers/net/mctp/Kconfig    |  10 +
 drivers/net/mctp/Makefile   |   1 +
 drivers/net/mctp/mctp-usb.c | 385 ++++++++++++++++++++++++++++++++++++
 3 files changed, 396 insertions(+)
 create mode 100644 drivers/net/mctp/mctp-usb.c

diff --git a/drivers/net/mctp/Kconfig b/drivers/net/mctp/Kconfig
index ce9d2d2ccf3b..569cb5cfba62 100644
--- a/drivers/net/mctp/Kconfig
+++ b/drivers/net/mctp/Kconfig
@@ -42,6 +42,16 @@ config MCTP_TRANSPORT_I3C
 	  A MCTP protocol network device is created for each I3C bus
 	  having a "mctp-controller" devicetree property.
 
+config MCTP_TRANSPORT_USB
+	tristate "MCTP USB transport"
+	depends on USB
+	help
+	  Provides a driver to access MCTP devices over USB transport,
+	  defined by DMTF specification DSP0283.
+
+	  MCTP-over-USB interfaces are peer-to-peer, so each interface
+	  represents a physical connection to one remote MCTP endpoint.
+
 endmenu
 
 endif
diff --git a/drivers/net/mctp/Makefile b/drivers/net/mctp/Makefile
index e1cb99ced54a..c36006849a1e 100644
--- a/drivers/net/mctp/Makefile
+++ b/drivers/net/mctp/Makefile
@@ -1,3 +1,4 @@
 obj-$(CONFIG_MCTP_SERIAL) += mctp-serial.o
 obj-$(CONFIG_MCTP_TRANSPORT_I2C) += mctp-i2c.o
 obj-$(CONFIG_MCTP_TRANSPORT_I3C) += mctp-i3c.o
+obj-$(CONFIG_MCTP_TRANSPORT_USB) += mctp-usb.o
diff --git a/drivers/net/mctp/mctp-usb.c b/drivers/net/mctp/mctp-usb.c
new file mode 100644
index 000000000000..e8d4b01c3f34
--- /dev/null
+++ b/drivers/net/mctp/mctp-usb.c
@@ -0,0 +1,385 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * mctp-usb.c - MCTP-over-USB (DMTF DSP0283) transport binding driver.
+ *
+ * DSP0283 is available at:
+ * https://www.dmtf.org/sites/default/files/standards/documents/DSP0283_1.0.1.pdf
+ *
+ * Copyright (C) 2024-2025 Code Construct Pty Ltd
+ */
+
+#include <linux/module.h>
+#include <linux/netdevice.h>
+#include <linux/usb.h>
+#include <linux/usb/mctp-usb.h>
+
+#include <net/mctp.h>
+#include <net/mctpdevice.h>
+#include <net/pkt_sched.h>
+
+#include <uapi/linux/if_arp.h>
+
+struct mctp_usb {
+	struct usb_device *usbdev;
+	struct usb_interface *intf;
+	bool stopped;
+
+	struct net_device *netdev;
+
+	u8 ep_in;
+	u8 ep_out;
+
+	struct urb *tx_urb;
+	struct urb *rx_urb;
+
+	struct delayed_work rx_retry_work;
+};
+
+static void mctp_usb_out_complete(struct urb *urb)
+{
+	struct sk_buff *skb = urb->context;
+	struct net_device *netdev = skb->dev;
+	int status;
+
+	status = urb->status;
+
+	switch (status) {
+	case -ENOENT:
+	case -ECONNRESET:
+	case -ESHUTDOWN:
+	case -EPROTO:
+		dev_dstats_tx_dropped(netdev);
+		break;
+	case 0:
+		dev_dstats_tx_add(netdev, skb->len);
+		netif_wake_queue(netdev);
+		consume_skb(skb);
+		return;
+	default:
+		netdev_dbg(netdev, "unexpected tx urb status: %d\n", status);
+		dev_dstats_tx_dropped(netdev);
+	}
+
+	kfree_skb(skb);
+}
+
+static netdev_tx_t mctp_usb_start_xmit(struct sk_buff *skb,
+				       struct net_device *dev)
+{
+	struct mctp_usb *mctp_usb = netdev_priv(dev);
+	struct mctp_usb_hdr *hdr;
+	unsigned int plen;
+	struct urb *urb;
+	int rc;
+
+	plen = skb->len;
+
+	if (plen + sizeof(*hdr) > MCTP_USB_XFER_SIZE)
+		goto err_drop;
+
+	rc = skb_cow_head(skb, sizeof(*hdr));
+	if (rc)
+		goto err_drop;
+
+	hdr = skb_push(skb, sizeof(*hdr));
+	if (!hdr)
+		goto err_drop;
+
+	hdr->id = cpu_to_be16(MCTP_USB_DMTF_ID);
+	hdr->rsvd = 0;
+	hdr->len = plen + sizeof(*hdr);
+
+	urb = mctp_usb->tx_urb;
+
+	usb_fill_bulk_urb(urb, mctp_usb->usbdev,
+			  usb_sndbulkpipe(mctp_usb->usbdev, mctp_usb->ep_out),
+			  skb->data, skb->len,
+			  mctp_usb_out_complete, skb);
+
+	rc = usb_submit_urb(urb, GFP_ATOMIC);
+	if (rc)
+		goto err_drop;
+	else
+		netif_stop_queue(dev);
+
+	return NETDEV_TX_OK;
+
+err_drop:
+	dev_dstats_tx_dropped(dev);
+	kfree_skb(skb);
+	return NETDEV_TX_OK;
+}
+
+static void mctp_usb_in_complete(struct urb *urb);
+
+/* If we fail to queue an in urb atomically (either due to skb allocation or
+ * urb submission), we will schedule a rx queue in nonatomic context
+ * after a delay, specified in jiffies
+ */
+static const unsigned long RX_RETRY_DELAY = HZ / 4;
+
+static int mctp_usb_rx_queue(struct mctp_usb *mctp_usb, gfp_t gfp)
+{
+	struct sk_buff *skb;
+	int rc;
+
+	skb = __netdev_alloc_skb(mctp_usb->netdev, MCTP_USB_XFER_SIZE, gfp);
+	if (!skb) {
+		rc = -ENOMEM;
+		goto err_retry;
+	}
+
+	usb_fill_bulk_urb(mctp_usb->rx_urb, mctp_usb->usbdev,
+			  usb_rcvbulkpipe(mctp_usb->usbdev, mctp_usb->ep_in),
+			  skb->data, MCTP_USB_XFER_SIZE,
+			  mctp_usb_in_complete, skb);
+
+	rc = usb_submit_urb(mctp_usb->rx_urb, gfp);
+	if (rc) {
+		netdev_dbg(mctp_usb->netdev, "rx urb submit failure: %d\n", rc);
+		kfree_skb(skb);
+		if (rc == -ENOMEM)
+			goto err_retry;
+	}
+
+	return rc;
+
+err_retry:
+	schedule_delayed_work(&mctp_usb->rx_retry_work, RX_RETRY_DELAY);
+	return rc;
+}
+
+static void mctp_usb_in_complete(struct urb *urb)
+{
+	struct sk_buff *skb = urb->context;
+	struct net_device *netdev = skb->dev;
+	struct mctp_usb *mctp_usb = netdev_priv(netdev);
+	struct mctp_skb_cb *cb;
+	unsigned int len;
+	int status;
+
+	status = urb->status;
+
+	switch (status) {
+	case -ENOENT:
+	case -ECONNRESET:
+	case -ESHUTDOWN:
+	case -EPROTO:
+		kfree_skb(skb);
+		return;
+	case 0:
+		break;
+	default:
+		netdev_dbg(netdev, "unexpected rx urb status: %d\n", status);
+		kfree_skb(skb);
+		return;
+	}
+
+	len = urb->actual_length;
+	__skb_put(skb, len);
+
+	while (skb) {
+		struct sk_buff *skb2 = NULL;
+		struct mctp_usb_hdr *hdr;
+		u8 pkt_len; /* length of MCTP packet, no USB header */
+
+		hdr = skb_pull_data(skb, sizeof(*hdr));
+		if (!hdr)
+			break;
+
+		if (be16_to_cpu(hdr->id) != MCTP_USB_DMTF_ID) {
+			netdev_dbg(netdev, "rx: invalid id %04x\n",
+				   be16_to_cpu(hdr->id));
+			break;
+		}
+
+		if (hdr->len <
+		    sizeof(struct mctp_hdr) + sizeof(struct mctp_usb_hdr)) {
+			netdev_dbg(netdev, "rx: short packet (hdr) %d\n",
+				   hdr->len);
+			break;
+		}
+
+		/* we know we have at least sizeof(struct mctp_usb_hdr) here */
+		pkt_len = hdr->len - sizeof(struct mctp_usb_hdr);
+		if (pkt_len > skb->len) {
+			netdev_dbg(netdev,
+				   "rx: short packet (xfer) %d, actual %d\n",
+				   hdr->len, skb->len);
+			break;
+		}
+
+		if (pkt_len < skb->len) {
+			/* more packets may follow - clone to a new
+			 * skb to use on the next iteration
+			 */
+			skb2 = skb_clone(skb, GFP_ATOMIC);
+			if (skb2) {
+				if (!skb_pull(skb2, pkt_len)) {
+					kfree_skb(skb2);
+					skb2 = NULL;
+				}
+			}
+			skb_trim(skb, pkt_len);
+		}
+
+		dev_dstats_rx_add(netdev, skb->len);
+
+		skb->protocol = htons(ETH_P_MCTP);
+		skb_reset_network_header(skb);
+		cb = __mctp_cb(skb);
+		cb->halen = 0;
+		netif_rx(skb);
+
+		skb = skb2;
+	}
+
+	if (skb)
+		kfree_skb(skb);
+
+	mctp_usb_rx_queue(mctp_usb, GFP_ATOMIC);
+}
+
+static void mctp_usb_rx_retry_work(struct work_struct *work)
+{
+	struct mctp_usb *mctp_usb = container_of(work, struct mctp_usb,
+						 rx_retry_work.work);
+
+	if (READ_ONCE(mctp_usb->stopped))
+		return;
+
+	mctp_usb_rx_queue(mctp_usb, GFP_KERNEL);
+}
+
+static int mctp_usb_open(struct net_device *dev)
+{
+	struct mctp_usb *mctp_usb = netdev_priv(dev);
+
+	WRITE_ONCE(mctp_usb->stopped, false);
+
+	return mctp_usb_rx_queue(mctp_usb, GFP_KERNEL);
+}
+
+static int mctp_usb_stop(struct net_device *dev)
+{
+	struct mctp_usb *mctp_usb = netdev_priv(dev);
+
+	netif_stop_queue(dev);
+
+	/* prevent RX submission retry */
+	WRITE_ONCE(mctp_usb->stopped, true);
+
+	usb_kill_urb(mctp_usb->rx_urb);
+	usb_kill_urb(mctp_usb->tx_urb);
+
+	cancel_delayed_work_sync(&mctp_usb->rx_retry_work);
+
+	return 0;
+}
+
+static const struct net_device_ops mctp_usb_netdev_ops = {
+	.ndo_start_xmit = mctp_usb_start_xmit,
+	.ndo_open = mctp_usb_open,
+	.ndo_stop = mctp_usb_stop,
+};
+
+static void mctp_usb_netdev_setup(struct net_device *dev)
+{
+	dev->type = ARPHRD_MCTP;
+
+	dev->mtu = MCTP_USB_MTU_MIN;
+	dev->min_mtu = MCTP_USB_MTU_MIN;
+	dev->max_mtu = MCTP_USB_MTU_MAX;
+
+	dev->hard_header_len = sizeof(struct mctp_usb_hdr);
+	dev->tx_queue_len = DEFAULT_TX_QUEUE_LEN;
+	dev->flags = IFF_NOARP;
+	dev->netdev_ops = &mctp_usb_netdev_ops;
+	dev->pcpu_stat_type = NETDEV_PCPU_STAT_DSTATS;
+}
+
+static int mctp_usb_probe(struct usb_interface *intf,
+			  const struct usb_device_id *id)
+{
+	struct usb_endpoint_descriptor *ep_in, *ep_out;
+	struct usb_host_interface *iface_desc;
+	struct net_device *netdev;
+	struct mctp_usb *dev;
+	int rc;
+
+	/* only one alternate */
+	iface_desc = intf->cur_altsetting;
+
+	rc = usb_find_common_endpoints(iface_desc, &ep_in, &ep_out, NULL, NULL);
+	if (rc) {
+		dev_err(&intf->dev, "invalid endpoints on device?\n");
+		return rc;
+	}
+
+	netdev = alloc_netdev(sizeof(*dev), "mctpusb%d", NET_NAME_ENUM,
+			      mctp_usb_netdev_setup);
+	if (!netdev)
+		return -ENOMEM;
+
+	SET_NETDEV_DEV(netdev, &intf->dev);
+	dev = netdev_priv(netdev);
+	dev->netdev = netdev;
+	dev->usbdev = usb_get_dev(interface_to_usbdev(intf));
+	dev->intf = intf;
+	usb_set_intfdata(intf, dev);
+
+	dev->ep_in = ep_in->bEndpointAddress;
+	dev->ep_out = ep_out->bEndpointAddress;
+
+	dev->tx_urb = usb_alloc_urb(0, GFP_KERNEL);
+	dev->rx_urb = usb_alloc_urb(0, GFP_KERNEL);
+	if (!dev->tx_urb || !dev->rx_urb) {
+		rc = -ENOMEM;
+		goto err_free_urbs;
+	}
+
+	INIT_DELAYED_WORK(&dev->rx_retry_work, mctp_usb_rx_retry_work);
+
+	rc = mctp_register_netdev(netdev, NULL, MCTP_PHYS_BINDING_USB);
+	if (rc)
+		goto err_free_urbs;
+
+	return 0;
+
+err_free_urbs:
+	usb_free_urb(dev->tx_urb);
+	usb_free_urb(dev->rx_urb);
+	free_netdev(netdev);
+	return rc;
+}
+
+static void mctp_usb_disconnect(struct usb_interface *intf)
+{
+	struct mctp_usb *dev = usb_get_intfdata(intf);
+
+	mctp_unregister_netdev(dev->netdev);
+	usb_free_urb(dev->tx_urb);
+	usb_free_urb(dev->rx_urb);
+	usb_put_dev(dev->usbdev);
+	free_netdev(dev->netdev);
+}
+
+static const struct usb_device_id mctp_usb_devices[] = {
+	{ USB_INTERFACE_INFO(USB_CLASS_MCTP, 0x0, 0x1) },
+	{ 0 },
+};
+
+MODULE_DEVICE_TABLE(usb, mctp_usb_devices);
+
+static struct usb_driver mctp_usb_driver = {
+	.name		= "mctp-usb",
+	.id_table	= mctp_usb_devices,
+	.probe		= mctp_usb_probe,
+	.disconnect	= mctp_usb_disconnect,
+};
+
+module_usb_driver(mctp_usb_driver)
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Jeremy Kerr <jk@codeconstruct.com.au>");
+MODULE_DESCRIPTION("MCTP USB transport");
-- 
2.39.5

[PATCH linux dev-6.6 3/3] net: mctp: usb: Port for kernel 6.6

[Santosh Puranik]

From: Santosh Puranik <spuranik@nvidia.com>

Fix the way the mctp usb binding driver records netdev stats.
Fix the netdev register call.

Signed-off-by: Santosh Puranik <santosh.puranik.ibm@gmail.com>
---
 drivers/net/mctp/mctp-usb.c | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)

diff --git a/drivers/net/mctp/mctp-usb.c b/drivers/net/mctp/mctp-usb.c
index e8d4b01c3f34..0c293d6d3607 100644
--- a/drivers/net/mctp/mctp-usb.c
+++ b/drivers/net/mctp/mctp-usb.c
@@ -48,16 +48,17 @@ static void mctp_usb_out_complete(struct urb *urb)
 	case -ECONNRESET:
 	case -ESHUTDOWN:
 	case -EPROTO:
-		dev_dstats_tx_dropped(netdev);
+        netdev->stats.tx_dropped++;
 		break;
 	case 0:
-		dev_dstats_tx_add(netdev, skb->len);
+        netdev->stats.tx_packets++;
+        netdev->stats.tx_bytes += skb->len;
 		netif_wake_queue(netdev);
 		consume_skb(skb);
 		return;
 	default:
 		netdev_dbg(netdev, "unexpected tx urb status: %d\n", status);
-		dev_dstats_tx_dropped(netdev);
+        netdev->stats.tx_dropped++;
 	}
 
 	kfree_skb(skb);
@@ -105,7 +106,7 @@ static netdev_tx_t mctp_usb_start_xmit(struct sk_buff *skb,
 	return NETDEV_TX_OK;
 
 err_drop:
-	dev_dstats_tx_dropped(dev);
+    dev->stats.tx_dropped++;
 	kfree_skb(skb);
 	return NETDEV_TX_OK;
 }
@@ -223,7 +224,8 @@ static void mctp_usb_in_complete(struct urb *urb)
 			skb_trim(skb, pkt_len);
 		}
 
-		dev_dstats_rx_add(netdev, skb->len);
+		netdev->stats.rx_packets++;
+        netdev->stats.rx_bytes += skb->len;
 
 		skb->protocol = htons(ETH_P_MCTP);
 		skb_reset_network_header(skb);
@@ -340,7 +342,7 @@ static int mctp_usb_probe(struct usb_interface *intf,
 
 	INIT_DELAYED_WORK(&dev->rx_retry_work, mctp_usb_rx_retry_work);
 
-	rc = mctp_register_netdev(netdev, NULL, MCTP_PHYS_BINDING_USB);
+	rc = mctp_register_netdev(netdev, NULL);
 	if (rc)
 		goto err_free_urbs;
 
-- 
2.39.5

Re: [PATCH linux dev-6.6 0/3] Backport MCTP Over USB Binding

[Jeremy Kerr]

Hi Santosh,

> This series backports the MCTP over USB binding driver and associated
> definitions.
> 
> Patches 1 and 2 were cherry-picked from upstream linux master and
> patch 3 includes fixes needed to backport the binding to 6.6.

We'll want to have f5d83cf0eeb9 ("net: mctp: unshare packets when
reassembling") included too:

  https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/net/mctp?id=f5d83cf0eeb90fade4d5c4d17d24b8bee9ceeecc

- that should come from the 6.6 stable updates, but doesn't look like it
has hit the openbmc tree yet. Could just be a matter of sequencing vs.
that update, or you could include explicitly.

(Andrew or Joel may have opinions one way or the other)

A couple of comments on 3/3 too.

Cheers,


Jeremy

Re: [PATCH linux dev-6.6 3/3] net: mctp: usb: Port for kernel 6.6

[Jeremy Kerr]

Hi Santosh,

> --- a/drivers/net/mctp/mctp-usb.c
> +++ b/drivers/net/mctp/mctp-usb.c
> @@ -48,16 +48,17 @@ static void mctp_usb_out_complete(struct urb *urb)
>         case -ECONNRESET:
>         case -ESHUTDOWN:
>         case -EPROTO:
> -               dev_dstats_tx_dropped(netdev);
> +        netdev->stats.tx_dropped++;

Some weird indenting happening here (and with most of the other
additions too).

If we don't want to use dstats, we should also remove

    dev->pcpu_stat_type = NETDEV_PCPU_STAT_DSTATS;

Have you checked that the stats look correct in your backport?

> @@ -340,7 +342,7 @@ static int mctp_usb_probe(struct usb_interface *intf,
>  
>         INIT_DELAYED_WORK(&dev->rx_retry_work, mctp_usb_rx_retry_work);
>  
> -       rc = mctp_register_netdev(netdev, NULL, MCTP_PHYS_BINDING_USB);
> +       rc = mctp_register_netdev(netdev, NULL);
>         if (rc)
>                 goto err_free_urbs;

Alternatively, we could consider backporting 580db513b4a9 ("net: mctp:
Expose transport binding identifier via IFLA attribute") too. This would
be user-visible, as we would now have an new IFLA_MCTP_PHYS_BINDING
attribute on netlink update messages, but that should be entirely
backward-compatible for applications.

Cheers,


Jeremy

Re: [PATCH linux dev-6.6 3/3] net: mctp: usb: Port for kernel 6.6

[Santosh Puranik]

Hi Jeremy,

Thank you for the review.

On 28/04/25 7:13 AM, Jeremy Kerr wrote:
> Hi Santosh,
> 
>> --- a/drivers/net/mctp/mctp-usb.c
>> +++ b/drivers/net/mctp/mctp-usb.c
>> @@ -48,16 +48,17 @@ static void mctp_usb_out_complete(struct urb *urb)
>>          case -ECONNRESET:
>>          case -ESHUTDOWN:
>>          case -EPROTO:
>> -               dev_dstats_tx_dropped(netdev);
>> +        netdev->stats.tx_dropped++;
> 
> Some weird indenting happening here (and with most of the other
> additions too).

Will fix in v2. Thanks.

> 
> If we don't want to use dstats, we should also remove
> 
>      dev->pcpu_stat_type = NETDEV_PCPU_STAT_DSTATS;
> 
> Have you checked that the stats look correct in your backport?

Yes, they did, for ex:

```
mctpusb0  Link encap:UNSPEC  HWaddr 
00-00-00-00-00-00-00-00-00-00-00-00-00-00-00-00
           UP RUNNING NOARP  MTU:68  Metric:1
           RX packets:1 errors:0 dropped:0 overruns:0 frame:0
           TX packets:1 errors:0 dropped:0 overruns:0 carrier:0
           collisions:0 txqueuelen:1000
           RX bytes:11 (11.0 B)  TX bytes:13 (13.0 B)
```
I will remove the stat_type assignment and re-run the test before 
submitting v2.

> 
>> @@ -340,7 +342,7 @@ static int mctp_usb_probe(struct usb_interface *intf,
>>   
>>          INIT_DELAYED_WORK(&dev->rx_retry_work, mctp_usb_rx_retry_work);
>>   
>> -       rc = mctp_register_netdev(netdev, NULL, MCTP_PHYS_BINDING_USB);
>> +       rc = mctp_register_netdev(netdev, NULL);
>>          if (rc)
>>                  goto err_free_urbs;
> 
> Alternatively, we could consider backporting 580db513b4a9 ("net: mctp:
> Expose transport binding identifier via IFLA attribute") too. This would
> be user-visible, as we would now have an new IFLA_MCTP_PHYS_BINDING
> attribute on netlink update messages, but that should be entirely
> backward-compatible for applications.

Ack. Please Let me know if you have a preference.

> 
> Cheers,
> 
> 
> Jeremy

Re: [PATCH linux dev-6.6 0/3] Backport MCTP Over USB Binding

[Santosh Puranik]

Hi Jeremy,

On 28/04/25 7:13 AM, Jeremy Kerr wrote:
> Hi Santosh,
> 
>> This series backports the MCTP over USB binding driver and associated
>> definitions.
>>
>> Patches 1 and 2 were cherry-picked from upstream linux master and
>> patch 3 includes fixes needed to backport the binding to 6.6.
> 
> We'll want to have f5d83cf0eeb9 ("net: mctp: unshare packets when
> reassembling") included too:
> 
>    https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/net/mctp?id=f5d83cf0eeb90fade4d5c4d17d24b8bee9ceeecc
> 
> - that should come from the 6.6 stable updates, but doesn't look like it
> has hit the openbmc tree yet. Could just be a matter of sequencing vs.
> that update, or you could include explicitly.
> 
> (Andrew or Joel may have opinions one way or the other)

OK, will wait to hear before cherry-picking that commit.

> 
> A couple of comments on 3/3 too.

Ack, thanks.

> 
> Cheers,
> 
> 
> Jeremy

Re: [PATCH linux dev-6.6 0/3] Backport MCTP Over USB Binding

[Andrew Jeffery]

On Mon, 2025-04-28 at 09:43 +0800, Jeremy Kerr wrote:
> Hi Santosh,
> 
> > This series backports the MCTP over USB binding driver and associated
> > definitions.
> > 
> > Patches 1 and 2 were cherry-picked from upstream linux master and
> > patch 3 includes fixes needed to backport the binding to 6.6.
> 
> We'll want to have f5d83cf0eeb9 ("net: mctp: unshare packets when
> reassembling") included too:
> 
>   https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/net/mctp?id=f5d83cf0eeb90fade4d5c4d17d24b8bee9ceeecc
> 
> - that should come from the 6.6 stable updates, but doesn't look like it
> has hit the openbmc tree yet. Could just be a matter of sequencing vs.
> that update, or you could include explicitly.
> 
> (Andrew or Joel may have opinions one way or the other)

I can take care of it: There are some other related changes that need
backporting as well as they haven't come in via the stable trees (I've
just pushed recipe bumps integrating v6.6.88 for openbmc/linux).

Andrew

Re: [PATCH linux dev-6.6 0/3] Backport MCTP Over USB Binding

[Andrew Jeffery]

On Wed, 2025-04-30 at 11:58 +0800, Andrew Jeffery wrote:
> On Mon, 2025-04-28 at 09:43 +0800, Jeremy Kerr wrote:
> > Hi Santosh,
> > 
> > > This series backports the MCTP over USB binding driver and associated
> > > definitions.
> > > 
> > > Patches 1 and 2 were cherry-picked from upstream linux master and
> > > patch 3 includes fixes needed to backport the binding to 6.6.
> > 
> > We'll want to have f5d83cf0eeb9 ("net: mctp: unshare packets when
> > reassembling") included too:
> > 
> >   https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/net/mctp?id=f5d83cf0eeb90fade4d5c4d17d24b8bee9ceeecc
> > 
> > - that should come from the 6.6 stable updates, but doesn't look like it
> > has hit the openbmc tree yet. Could just be a matter of sequencing vs.
> > that update, or you could include explicitly.
> > 
> > (Andrew or Joel may have opinions one way or the other)
> 
> I can take care of it: There are some other related changes that need
> backporting as well as they haven't come in via the stable trees (I've
> just pushed recipe bumps integrating v6.6.88 for openbmc/linux).

I've backported several fixes now, including the one above:

- https://gerrit.openbmc.org/c/openbmc/openbmc/+/80040
- https://gerrit.openbmc.org/c/openbmc/openbmc/+/80041

Santosh: Feel free to rebase your series and re-send it after
addressing Jeremy's concerns.

Andrew

Re: [PATCH linux dev-6.6 0/3] Backport MCTP Over USB Binding

[Santosh Puranik]

On 06/05/25 12:41 PM, Andrew Jeffery wrote:
> On Wed, 2025-04-30 at 11:58 +0800, Andrew Jeffery wrote:
>> On Mon, 2025-04-28 at 09:43 +0800, Jeremy Kerr wrote:
>>> Hi Santosh,
>>>
>>>> This series backports the MCTP over USB binding driver and associated
>>>> definitions.
>>>>
>>>> Patches 1 and 2 were cherry-picked from upstream linux master and
>>>> patch 3 includes fixes needed to backport the binding to 6.6.
>>>
>>> We'll want to have f5d83cf0eeb9 ("net: mctp: unshare packets when
>>> reassembling") included too:
>>>
>>>    https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/net/mctp?id=f5d83cf0eeb90fade4d5c4d17d24b8bee9ceeecc
>>>
>>> - that should come from the 6.6 stable updates, but doesn't look like it
>>> has hit the openbmc tree yet. Could just be a matter of sequencing vs.
>>> that update, or you could include explicitly.
>>>
>>> (Andrew or Joel may have opinions one way or the other)
>>
>> I can take care of it: There are some other related changes that need
>> backporting as well as they haven't come in via the stable trees (I've
>> just pushed recipe bumps integrating v6.6.88 for openbmc/linux).
> 
> I've backported several fixes now, including the one above:
> 
> - https://gerrit.openbmc.org/c/openbmc/openbmc/+/80040
> - https://gerrit.openbmc.org/c/openbmc/openbmc/+/80041
> 
> Santosh: Feel free to rebase your series and re-send it after
> addressing Jeremy's concerns.

Thanks, Andrew. I have rebased, addressed the review comments and sent a v2.

> 
> Andrew