[PATCH v3 0/3] usb: Add support for eUSB2v2 1024-byte Bulk MaxPacketSize

[PATCH v3 0/3] usb: Add support for eUSB2v2 1024-byte Bulk MaxPacketSize

[Pawel Laszczak via B4 Relay]

The Embedded USB2 (eUSB2) v2 specification (bcdUSB 0x0230) introduces a
mechanism that allows High-Speed Bulk endpoints to support a maximum
packet size (MPS) of 1024 bytes, expanding it from the traditional
512-byte limit defined in the standard USB 2.0 specification.
This capability is highly beneficial for Mass Storage Class (MSC)
devices, significantly improving overall throughput.

Per the eUSB2v2 specification, a peripheral device will revert its
internal Bulk MPS back to 512 bytes after major bus events, including a
bus reset, disconnect, or deconfiguration. To establish and maintain
the 1024-byte mode, the host controller must explicitly issue
a device-recipient SET_FEATURE request called
USB_DEVICE_BULK_MAX_PACKET_UPDATE. This must occur after the device enters
the ADDRESSED state, but strictly before it transitions to the CONFIGURED
state.

This patch series implements the required infrastructure on both the host
and gadget sides, and enables it for the Cadence (cdnsp) controller.

Structure of the series:
- Patch 1: Introduces host-side core support in usb_set_configuration() to
  detect eUSB2v2 devices and issue the SET_FEATURE update request prior to
  activating the device configuration. Adds eusb2v2_mps_active flag to
  struct usb_device to track negotiation state. Extends hub.c to re-issue
  SET_FEATURE after bus reset in usb_reset_and_verify_device(); on failure
  triggers re-enumeration to prevent driver from operating with
  inconsistent MPS state. Also updates xHCI endpoint initialization
  to allow 1024-byte packets for High-Speed Bulk endpoints when HCC2_E2V2C
  is present.

- Patch 2: Extends the USB Gadget Composite framework to advertise bcdUSB
  0x0230 when eUSB2v2 is supported, intercept the incoming feature
  request, and dynamically update the wMaxPacketSize fields across all
  High-Speed Bulk descriptors before configuration is completed.
  
- Patch 3: Implements peripheral-specific handling in the Cadence (cdnsp)
  driver, reading HCCPARAMS2 to discover the hardware capability and
  adapting cdnsp_endpoint_init() boundaries to allow for the larger
  packet size.

Testing:
The flow has been validated using the Cadence eUSB2v2 Host and Device
controllers, ensuring seamless transition to 1024-byte mode during early
enumeration stages and successful verification under USB-IF Compliance
Verification (CV) tooling.

Signed-off-by: Pawel Laszczak <pawell@cadence.com>
---
Changes in v3:
- message.c: Add check for bulk endpoint existence before sending
  BULK_MAX_PACKET_UPDATE. This avoids sending the request to devices that
  only use isochronous endpoints, as they do not support it.
- host & gadget: Do not overwrite ep->desc.wMaxPacketSize to 1024.
  According to eUSB2v2 spec section 5.2, the endpoint descriptor must
  always report 512 bytes regardless of the current operating mode.
- host: Rely on the eusb2v2_mps_active flag in xhci_usb_endpoint_maxp()
  to dynamically return 1024 for HS bulk endpoints when the 1KB mode
  is active.
- gadget: Introduce gadget->eusb2v2_mps_active flag to let the composite
  framework report 512 bytes while the UDC dynamically sets 1KB max packet
  size in hardware.
- Link to v2: https://patch.msgid.link/20260826-eusb2v2-packet-size-v2-0-950f19f38ef0@cadence.com

Changes in v2:
- Removed config.c change: per eUSB2v2 spec section 5.2, conformant devices
  always report wMaxPacketSize=512 in their descriptor regardless of operating
  mode, so the warning suppression was unnecessary.
- eusb_update_max_packet(): changed from void to int; added error propagation
  and eusb2v2_mps_active state tracking.
- Added hub.c handling to restore 1KB mode after bus reset, with re-enumeration
  on failure (addresses Oliver Neukum's review comments).
- xhci-mem.c: simplified HS bulk clamp logic.
- xhci.c: moved is_eusb2v2 assignment into xhci_hcd_init_usb2_data().
- composite.c: stall CLEAR_FEATURE(BULK_MAX_PACKET_UPDATE) as it is not
  defined by the eUSB2v2 spec.
- cdnsp-mem.c: simplified HS bulk clamp logic, consistent with xhci-mem.c.
- Link to v1: https://patch.msgid.link/20260717-eusb2v2-packet-size-v1-0-67c611bd0c5b@cadence.com

---
Pawel Laszczak (3):
      usb: xhci: Add support for eUSB2v2 1024-byte bulk packet size
      usb: gadget: composite: Support eUSB2v2 bulk MPS update
      usb: cdns3: cdnsp: Enable eUSB2v2 1KB bulk packet capability

 drivers/usb/cdns3/cdnsp-ep0.c    |  2 ++
 drivers/usb/cdns3/cdnsp-gadget.c |  4 +++
 drivers/usb/cdns3/cdnsp-gadget.h |  8 +++++
 drivers/usb/cdns3/cdnsp-mem.c    |  9 +++--
 drivers/usb/common/debug.c       |  2 ++
 drivers/usb/core/hub.c           | 16 +++++++++
 drivers/usb/core/message.c       | 74 ++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/usb.h           |  2 ++
 drivers/usb/gadget/composite.c   | 28 +++++++++++----
 drivers/usb/host/xhci-mem.c      | 15 ++++++--
 drivers/usb/host/xhci.c          | 10 ++++++
 include/linux/usb.h              |  7 ++++
 include/linux/usb/gadget.h       |  4 +++
 13 files changed, 169 insertions(+), 12 deletions(-)
---
base-commit: 1d52b92ce8748624f84eca7eaea3ac31ed7b09a5
change-id: 20260609-eusb2v2-packet-size-4cb668455465

Best regards,
--  
Pawel Laszczak <pawell@cadence.com>

[PATCH v3 1/3] usb: xhci: Add support for eUSB2v2 1024-byte bulk packet size

[Pawel Laszczak via B4 Relay]

From: Pawel Laszczak <pawell@cadence.com>

The eUSB2 v2 specification (bcdUSB 0x0230) introduces support for
1024-byte maximum packet sizes for Bulk endpoints in High-Speed mode.
However, an eUSB2v2 peripheral will revert its internal maximum packet
size back to 512 bytes after events like a bus reset, disconnect, or
deconfiguration.

To support 1024-byte bulk transfers on capable hosts, add a new
is_eusb2v2 flag to the usb_bus structure, populated via the HCCPARAMS2
E2V2C capability bit in the xHCI driver.

When an eUSB2v2 host configures an eUSB2v2 device, issue a specific
SET_FEATURE (USB_DEVICE_BULK_MAX_PACKET_UPDATE) request during device
configuration to switch the peripheral to 1024-byte packet mode, and
allow the xHCI endpoint initialization to accept up to 1024 bytes for
HS bulk endpoints.

Signed-off-by: Pawel Laszczak <pawell@cadence.com>
---
Changes in v3:
- Add check for bulk endpoint existence before sending
  BULK_MAX_PACKET_UPDATE. This avoids sending the request to devices that
  only use isochronous endpoints, as they do not support it.
- Do not overwrite ep->desc.wMaxPacketSize to 1024. According to eUSB2v2
  spec section 5.2, the endpoint descriptor must always report 512 bytes
  regardless of the current operating mode.
- Rely on the eusb2v2_mps_active flag in xhci_usb_endpoint_maxp() and
  xhci-mem.c to dynamically return 1024 for HS bulk endpoints when the 1KB
  mode is active.

Changes in v2:
- Removed change in config.c: per eUSB2v2 spec section 5.2, conformant
  devices always report wMaxPacketSize=512 in their descriptor regardless
  of operating mode, so the warning suppression was unnecessary.
- xhci-mem.c: simplified HS bulk clamp
- xhci.c: moved is_eusb2v2 assignment into xhci_hcd_init_usb2_data()
- eusb_update_max_packet(): changed from void to int; returns error on
  SET_FEATURE failure.
- Added eusb2v2_mps_active flag to struct usb_device to track whether
  SET_FEATURE(BULK_MAX_PACKET_UPDATE) succeeded.
- Added hub.c: usb_reset_and_verify_device() now re-issues SET_FEATURE
  after bus reset to restore 1KB mode. Failure triggers re-enumeration
  to prevent a driver from operating with inconsistent MPS state.
---
 drivers/usb/core/hub.c      | 16 ++++++++++
 drivers/usb/core/message.c  | 74 +++++++++++++++++++++++++++++++++++++++++++++
 drivers/usb/core/usb.h      |  2 ++
 drivers/usb/host/xhci-mem.c | 15 +++++++--
 drivers/usb/host/xhci.c     | 10 ++++++
 include/linux/usb.h         |  7 +++++
 6 files changed, 121 insertions(+), 3 deletions(-)

diff --git a/drivers/usb/core/hub.c b/drivers/usb/core/hub.c
index 24960ba9caa9..34cfc44c5df8 100644
--- a/drivers/usb/core/hub.c
+++ b/drivers/usb/core/hub.c
@@ -6252,6 +6252,22 @@ static int usb_reset_and_verify_device(struct usb_device *udev)
 		mutex_unlock(hcd->bandwidth_mutex);
 		goto re_enumerate;
 	}
+
+	/*
+	 * Restore eUSB2v2 1KB bulk mode after reset (device reverts to 512
+	 * after any bus reset per eUSB2v2 spec section 5.2).
+	 * Only retry if the initial SET_FEATURE had succeeded.
+	 */
+	if (udev->eusb2v2_mps_active) {
+		ret = eusb_update_max_packet(udev, udev->actconfig);
+		if (ret < 0) {
+			dev_err(&udev->dev,
+				"eUSB2v2: failed to restore 1KB mode after reset (%d)\n", ret);
+			mutex_unlock(hcd->bandwidth_mutex);
+			goto re_enumerate;
+		}
+	}
+
 	ret = usb_control_msg(udev, usb_sndctrlpipe(udev, 0),
 			USB_REQ_SET_CONFIGURATION, 0,
 			udev->actconfig->desc.bConfigurationValue, 0,
diff --git a/drivers/usb/core/message.c b/drivers/usb/core/message.c
index 75e2bfd744a9..c8540d738a3e 100644
--- a/drivers/usb/core/message.c
+++ b/drivers/usb/core/message.c
@@ -2007,6 +2007,71 @@ int usb_set_wireless_status(struct usb_interface *iface,
 }
 EXPORT_SYMBOL_GPL(usb_set_wireless_status);
 
+/*
+ * eusb_update_max_packet - enable 1024-byte bulk mode for eUSB2v2 device
+ * @udev: target device
+ * @cp: configuration to be checked and enabled
+ *
+ * Per eUSB2v2 spec section 5.2, an eUSB2v2 peripheral will revert the
+ * maximum packet size to 512 for bulk endpoints after bus reset, disconnect,
+ * or deconfiguration.
+ * This function sends the BULK_MAX_PACKET_UPDATE request to restore the
+ * 1024-byte mode. It is valid only if the configuration has bulk endpoints.
+ */
+int eusb_update_max_packet(struct usb_device *udev, struct usb_host_config *cp)
+{
+	struct usb_host_config *config = cp ? cp : udev->actconfig;
+	struct usb_hcd *hcd = bus_to_hcd(udev->bus);
+	struct usb_interface_cache *intfc;
+	struct usb_host_interface *alt;
+	struct usb_host_endpoint *ep;
+	bool has_bulk = false;
+	int i, j, a;
+	int ret;
+
+	if (le16_to_cpu(udev->descriptor.bcdUSB) != 0x0230 ||
+	    !hcd->self.is_eusb2v2)
+		return 0;
+
+	if (!config)
+		return 0;
+
+	for (i = 0; i < config->desc.bNumInterfaces; i++) {
+		intfc = config->intf_cache[i];
+
+		if (!intfc)
+			continue;
+
+		for (a = 0; a < intfc->num_altsetting; a++) {
+			alt = &intfc->altsetting[a];
+
+			for (j = 0; j < alt->desc.bNumEndpoints; j++) {
+				ep = &alt->endpoint[j];
+
+				if (usb_endpoint_xfer_bulk(&ep->desc)) {
+					has_bulk = true;
+					goto found_bulk;
+				}
+			}
+		}
+	}
+
+	if (!has_bulk)
+		return 0;
+
+found_bulk:
+	ret = usb_control_msg(udev, usb_sndctrlpipe(udev, 0),
+			      USB_REQ_SET_FEATURE, USB_RECIP_DEVICE,
+			      USB_DEVICE_BULK_MAX_PACKET_UPDATE, 0, NULL, 0,
+			      USB_CTRL_SET_TIMEOUT);
+	if (ret < 0) {
+		dev_warn(&udev->dev, "eUSB2v2 1KB update failed: %d\n", ret);
+		return ret;
+	}
+
+	return 0;
+}
+
 /*
  * usb_set_configuration - Makes a particular device setting be current
  * @dev: the device whose configuration is being updated
@@ -2123,6 +2188,15 @@ int usb_set_configuration(struct usb_device *dev, int configuration)
 	if (dev->state != USB_STATE_ADDRESS)
 		usb_disable_device(dev, 1);	/* Skip ep0 */
 
+	ret = eusb_update_max_packet(dev, cp);
+	if (ret < 0)
+		dev->eusb2v2_mps_active = 0;
+	else if (cp && le16_to_cpu(dev->descriptor.bcdUSB) == 0x0230 &&
+		 hcd->self.is_eusb2v2)
+		dev->eusb2v2_mps_active = 1;
+	else
+		dev->eusb2v2_mps_active = 0;
+
 	/* Get rid of pending async Set-Config requests for this device */
 	cancel_async_set_config(dev);
 
diff --git a/drivers/usb/core/usb.h b/drivers/usb/core/usb.h
index a9b37aeb515b..c51e4261085c 100644
--- a/drivers/usb/core/usb.h
+++ b/drivers/usb/core/usb.h
@@ -89,6 +89,8 @@ extern int usb_major_init(void);
 extern void usb_major_cleanup(void);
 extern int usb_device_supports_lpm(struct usb_device *udev);
 extern int usb_port_disable(struct usb_device *udev);
+int eusb_update_max_packet(struct usb_device *udev,
+			   struct usb_host_config *cp);
 
 #ifdef	CONFIG_PM
 
diff --git a/drivers/usb/host/xhci-mem.c b/drivers/usb/host/xhci-mem.c
index 997fe90f54e5..37a5f9e9bd6d 100644
--- a/drivers/usb/host/xhci-mem.c
+++ b/drivers/usb/host/xhci-mem.c
@@ -1479,10 +1479,19 @@ int xhci_endpoint_init(struct xhci_hcd *xhci,
 	/* Allow 3 retries for everything but isoc, set CErr = 3 */
 	if (!usb_endpoint_xfer_isoc(&ep->desc))
 		err_count = 3;
-	/* HS bulk max packet should be 512, FS bulk supports 8, 16, 32 or 64 */
+
+	/*
+	 * HS bulk max packet should be 512 (or 1024 for eUSB2v2),
+	 * FS bulk supports 8, 16, 32 or 64.
+	 */
 	if (usb_endpoint_xfer_bulk(&ep->desc)) {
-		if (udev->speed == USB_SPEED_HIGH)
-			max_packet = 512;
+		if (udev->speed == USB_SPEED_HIGH) {
+			if (udev->eusb2v2_mps_active)
+				max_packet = 1024;
+			else
+				max_packet = 512;
+		}
+
 		if (udev->speed == USB_SPEED_FULL) {
 			max_packet = rounddown_pow_of_two(max_packet);
 			max_packet = clamp_val(max_packet, 8, 64);
diff --git a/drivers/usb/host/xhci.c b/drivers/usb/host/xhci.c
index a54f5b57f205..b4a3bdf88fb4 100644
--- a/drivers/usb/host/xhci.c
+++ b/drivers/usb/host/xhci.c
@@ -2951,6 +2951,12 @@ int xhci_usb_endpoint_maxp(struct usb_device *udev,
 {
 	if (usb_endpoint_is_hs_isoc_double(udev, host_ep))
 		return le16_to_cpu(host_ep->eusb2_isoc_ep_comp.wMaxPacketSize);
+
+	if (udev->eusb2v2_mps_active &&
+	    udev->speed == USB_SPEED_HIGH &&
+	    usb_endpoint_xfer_bulk(&host_ep->desc))
+		return 1024;
+
 	return usb_endpoint_maxp(&host_ep->desc);
 }
 
@@ -5371,6 +5377,10 @@ static void xhci_hcd_init_usb2_data(struct xhci_hcd *xhci, struct usb_hcd *hcd)
 	xhci->usb2_rhub.hcd = hcd;
 	hcd->speed = HCD_USB2;
 	hcd->self.root_hub->speed = USB_SPEED_HIGH;
+
+	if (xhci->hcc_params2 & HCC2_E2V2C)
+		hcd->self.is_eusb2v2 = 1;
+
 	/*
 	 * USB 2.0 roothub under xHCI has an integrated TT,
 	 * (rate matching hub) as opposed to having an OHCI/UHCI
diff --git a/include/linux/usb.h b/include/linux/usb.h
index 25a203ac7a7e..57fb4c552740 100644
--- a/include/linux/usb.h
+++ b/include/linux/usb.h
@@ -464,6 +464,10 @@ struct usb_bus {
 					 * the ep queue on a short transfer
 					 * with the URB_SHORT_NOT_OK flag set.
 					 */
+	unsigned is_eusb2v2:1;		/*
+					 * true when HC controller supports
+					 * eusb2v2
+					 */
 	unsigned no_sg_constraint:1;	/* no sg constraint */
 	unsigned sg_tablesize;		/* 0 or largest number of sg list entries */
 
@@ -625,6 +629,8 @@ struct usb3_lpm_parameters {
  * @usb2_hw_lpm_allowed: Userspace allows USB 2.0 LPM to be enabled
  * @usb3_lpm_u1_enabled: USB3 hardware U1 LPM enabled
  * @usb3_lpm_u2_enabled: USB3 hardware U2 LPM enabled
+ * @eusb2v2_mps_active: 1024-byte bulk mode is active and must be restored
+ *	 after bus reset.
  * @string_langid: language ID for strings
  * @product: iProduct string, if present (static)
  * @manufacturer: iManufacturer string, if present (static)
@@ -708,6 +714,7 @@ struct usb_device {
 	unsigned usb2_hw_lpm_allowed:1;
 	unsigned usb3_lpm_u1_enabled:1;
 	unsigned usb3_lpm_u2_enabled:1;
+	unsigned eusb2v2_mps_active:1;
 	int string_langid;
 
 	/* static strings from the device */

-- 
2.43.0

Re: [PATCH v3 1/3] usb: xhci: Add support for eUSB2v2 1024-byte bulk packet size

[Michal Pecio]

On Mon, 05 Oct 2026 10:18:24 +0200, Pawel Laszczak via B4 Relay wrote:
> From: Pawel Laszczak <pawell@cadence.com>
> 
> The eUSB2 v2 specification (bcdUSB 0x0230) introduces support for
> 1024-byte maximum packet sizes for Bulk endpoints in High-Speed mode.
> However, an eUSB2v2 peripheral will revert its internal maximum packet
> size back to 512 bytes after events like a bus reset, disconnect, or
> deconfiguration.
> 
> To support 1024-byte bulk transfers on capable hosts, add a new
> is_eusb2v2 flag to the usb_bus structure, populated via the HCCPARAMS2
> E2V2C capability bit in the xHCI driver.
> 
> When an eUSB2v2 host configures an eUSB2v2 device, issue a specific
> SET_FEATURE (USB_DEVICE_BULK_MAX_PACKET_UPDATE) request during device
> configuration to switch the peripheral to 1024-byte packet mode, and
> allow the xHCI endpoint initialization to accept up to 1024 bytes for
> HS bulk endpoints.
> 
> Signed-off-by: Pawel Laszczak <pawell@cadence.com>
> ---
> Changes in v3:
> - Add check for bulk endpoint existence before sending
>   BULK_MAX_PACKET_UPDATE. This avoids sending the request to devices that
>   only use isochronous endpoints, as they do not support it.
> - Do not overwrite ep->desc.wMaxPacketSize to 1024. According to eUSB2v2
>   spec section 5.2, the endpoint descriptor must always report 512 bytes
>   regardless of the current operating mode.
> - Rely on the eusb2v2_mps_active flag in xhci_usb_endpoint_maxp() and
>   xhci-mem.c to dynamically return 1024 for HS bulk endpoints when the 1KB
>   mode is active.
> 
> Changes in v2:
> - Removed change in config.c: per eUSB2v2 spec section 5.2, conformant
>   devices always report wMaxPacketSize=512 in their descriptor regardless
>   of operating mode, so the warning suppression was unnecessary.
> - xhci-mem.c: simplified HS bulk clamp
> - xhci.c: moved is_eusb2v2 assignment into xhci_hcd_init_usb2_data()
> - eusb_update_max_packet(): changed from void to int; returns error on
>   SET_FEATURE failure.
> - Added eusb2v2_mps_active flag to struct usb_device to track whether
>   SET_FEATURE(BULK_MAX_PACKET_UPDATE) succeeded.
> - Added hub.c: usb_reset_and_verify_device() now re-issues SET_FEATURE
>   after bus reset to restore 1KB mode. Failure triggers re-enumeration
>   to prevent a driver from operating with inconsistent MPS state.
> ---
>  drivers/usb/core/hub.c      | 16 ++++++++++
>  drivers/usb/core/message.c  | 74 +++++++++++++++++++++++++++++++++++++++++++++
>  drivers/usb/core/usb.h      |  2 ++
>  drivers/usb/host/xhci-mem.c | 15 +++++++--
>  drivers/usb/host/xhci.c     | 10 ++++++
>  include/linux/usb.h         |  7 +++++
>  6 files changed, 121 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/usb/core/hub.c b/drivers/usb/core/hub.c
> index 24960ba9caa9..34cfc44c5df8 100644
> --- a/drivers/usb/core/hub.c
> +++ b/drivers/usb/core/hub.c
> @@ -6252,6 +6252,22 @@ static int usb_reset_and_verify_device(struct usb_device *udev)
>  		mutex_unlock(hcd->bandwidth_mutex);
>  		goto re_enumerate;
>  	}
> +
> +	/*
> +	 * Restore eUSB2v2 1KB bulk mode after reset (device reverts to 512
> +	 * after any bus reset per eUSB2v2 spec section 5.2).
> +	 * Only retry if the initial SET_FEATURE had succeeded.
> +	 */
> +	if (udev->eusb2v2_mps_active) {
> +		ret = eusb_update_max_packet(udev, udev->actconfig);
> +		if (ret < 0) {
> +			dev_err(&udev->dev,
> +				"eUSB2v2: failed to restore 1KB mode after reset (%d)\n", ret);

This seems redundant because the function already logs errors.
Or maybe the function shouldn't log them, if we want different
messages for different call sites?

> +			mutex_unlock(hcd->bandwidth_mutex);
> +			goto re_enumerate;
> +		}
> +	}
> +
>  	ret = usb_control_msg(udev, usb_sndctrlpipe(udev, 0),
>  			USB_REQ_SET_CONFIGURATION, 0,
>  			udev->actconfig->desc.bConfigurationValue, 0,
> diff --git a/drivers/usb/core/message.c b/drivers/usb/core/message.c
> index 75e2bfd744a9..c8540d738a3e 100644
> --- a/drivers/usb/core/message.c
> +++ b/drivers/usb/core/message.c
> @@ -2007,6 +2007,71 @@ int usb_set_wireless_status(struct usb_interface *iface,
>  }
>  EXPORT_SYMBOL_GPL(usb_set_wireless_status);
>  
> +/*
> + * eusb_update_max_packet - enable 1024-byte bulk mode for eUSB2v2 device
> + * @udev: target device
> + * @cp: configuration to be checked and enabled
> + *
> + * Per eUSB2v2 spec section 5.2, an eUSB2v2 peripheral will revert the
> + * maximum packet size to 512 for bulk endpoints after bus reset, disconnect,
> + * or deconfiguration.
> + * This function sends the BULK_MAX_PACKET_UPDATE request to restore the
> + * 1024-byte mode. It is valid only if the configuration has bulk endpoints.
> + */
> +int eusb_update_max_packet(struct usb_device *udev, struct usb_host_config *cp)
> +{
> +	struct usb_host_config *config = cp ? cp : udev->actconfig;
> +	struct usb_hcd *hcd = bus_to_hcd(udev->bus);
> +	struct usb_interface_cache *intfc;
> +	struct usb_host_interface *alt;
> +	struct usb_host_endpoint *ep;
> +	bool has_bulk = false;
> +	int i, j, a;
> +	int ret;
> +
> +	if (le16_to_cpu(udev->descriptor.bcdUSB) != 0x0230 ||
> +	    !hcd->self.is_eusb2v2)
> +		return 0;
> +
> +	if (!config)
> +		return 0;
> +
> +	for (i = 0; i < config->desc.bNumInterfaces; i++) {
> +		intfc = config->intf_cache[i];
> +
> +		if (!intfc)
> +			continue;
> +
> +		for (a = 0; a < intfc->num_altsetting; a++) {
> +			alt = &intfc->altsetting[a];
> +
> +			for (j = 0; j < alt->desc.bNumEndpoints; j++) {
> +				ep = &alt->endpoint[j];
> +
> +				if (usb_endpoint_xfer_bulk(&ep->desc)) {
> +					has_bulk = true;
> +					goto found_bulk;
> +				}
> +			}
> +		}
> +	}
> +
> +	if (!has_bulk)
> +		return 0;

Control flow ensures that has_bulk is false here and true below.
The variable is redundant.

> +found_bulk:
> +	ret = usb_control_msg(udev, usb_sndctrlpipe(udev, 0),
> +			      USB_REQ_SET_FEATURE, USB_RECIP_DEVICE,
> +			      USB_DEVICE_BULK_MAX_PACKET_UPDATE, 0, NULL, 0,
> +			      USB_CTRL_SET_TIMEOUT);
> +	if (ret < 0) {
> +		dev_warn(&udev->dev, "eUSB2v2 1KB update failed: %d\n", ret);
> +		return ret;
> +	}
> +
> +	return 0;
> +}
> +
>  /*
>   * usb_set_configuration - Makes a particular device setting be current
>   * @dev: the device whose configuration is being updated
> @@ -2123,6 +2188,15 @@ int usb_set_configuration(struct usb_device *dev, int configuration)
>  	if (dev->state != USB_STATE_ADDRESS)
>  		usb_disable_device(dev, 1);	/* Skip ep0 */
>  
> +	ret = eusb_update_max_packet(dev, cp);
> +	if (ret < 0)
> +		dev->eusb2v2_mps_active = 0;
> +	else if (cp && le16_to_cpu(dev->descriptor.bcdUSB) == 0x0230 &&
> +		 hcd->self.is_eusb2v2)

Hmm, this code could be simpler if the function returned errors on
unsupported devices or host controllers. Would that cause problems
with anything?

> +		dev->eusb2v2_mps_active = 1;
> +	else
> +		dev->eusb2v2_mps_active = 0;
> +
>  	/* Get rid of pending async Set-Config requests for this device */
>  	cancel_async_set_config(dev);
>  
> diff --git a/drivers/usb/core/usb.h b/drivers/usb/core/usb.h
> index a9b37aeb515b..c51e4261085c 100644
> --- a/drivers/usb/core/usb.h
> +++ b/drivers/usb/core/usb.h
> @@ -89,6 +89,8 @@ extern int usb_major_init(void);
>  extern void usb_major_cleanup(void);
>  extern int usb_device_supports_lpm(struct usb_device *udev);
>  extern int usb_port_disable(struct usb_device *udev);
> +int eusb_update_max_packet(struct usb_device *udev,
> +			   struct usb_host_config *cp);
>  
>  #ifdef	CONFIG_PM
>  
> diff --git a/drivers/usb/host/xhci-mem.c b/drivers/usb/host/xhci-mem.c
> index 997fe90f54e5..37a5f9e9bd6d 100644
> --- a/drivers/usb/host/xhci-mem.c
> +++ b/drivers/usb/host/xhci-mem.c
> @@ -1479,10 +1479,19 @@ int xhci_endpoint_init(struct xhci_hcd *xhci,
>  	/* Allow 3 retries for everything but isoc, set CErr = 3 */
>  	if (!usb_endpoint_xfer_isoc(&ep->desc))
>  		err_count = 3;
> -	/* HS bulk max packet should be 512, FS bulk supports 8, 16, 32 or 64 */
> +
> +	/*
> +	 * HS bulk max packet should be 512 (or 1024 for eUSB2v2),
> +	 * FS bulk supports 8, 16, 32 or 64.
> +	 */
>  	if (usb_endpoint_xfer_bulk(&ep->desc)) {
> -		if (udev->speed == USB_SPEED_HIGH)
> -			max_packet = 512;
> +		if (udev->speed == USB_SPEED_HIGH) {
> +			if (udev->eusb2v2_mps_active)
> +				max_packet = 1024;
> +			else
> +				max_packet = 512;
> +		}
> +
>  		if (udev->speed == USB_SPEED_FULL) {
>  			max_packet = rounddown_pow_of_two(max_packet);
>  			max_packet = clamp_val(max_packet, 8, 64);
> diff --git a/drivers/usb/host/xhci.c b/drivers/usb/host/xhci.c
> index a54f5b57f205..b4a3bdf88fb4 100644
> --- a/drivers/usb/host/xhci.c
> +++ b/drivers/usb/host/xhci.c
> @@ -2951,6 +2951,12 @@ int xhci_usb_endpoint_maxp(struct usb_device *udev,
>  {
>  	if (usb_endpoint_is_hs_isoc_double(udev, host_ep))
>  		return le16_to_cpu(host_ep->eusb2_isoc_ep_comp.wMaxPacketSize);
> +
> +	if (udev->eusb2v2_mps_active &&
> +	    udev->speed == USB_SPEED_HIGH &&
> +	    usb_endpoint_xfer_bulk(&host_ep->desc))
> +		return 1024;

Is the check for USB_SPEED_HIGH necessary?
Is eusb2v2_mps_active supposed to be set at other speeds?

One tricky edge case is a high-speed device failing to restore
high-speed signaling after reset, but at a quick glance this seems
to result in hub_port_init() error and abort of the reset flow.

> +
>  	return usb_endpoint_maxp(&host_ep->desc);
>  }
>  
> @@ -5371,6 +5377,10 @@ static void xhci_hcd_init_usb2_data(struct xhci_hcd *xhci, struct usb_hcd *hcd)
>  	xhci->usb2_rhub.hcd = hcd;
>  	hcd->speed = HCD_USB2;
>  	hcd->self.root_hub->speed = USB_SPEED_HIGH;
> +
> +	if (xhci->hcc_params2 & HCC2_E2V2C)
> +		hcd->self.is_eusb2v2 = 1;
> +
>  	/*
>  	 * USB 2.0 roothub under xHCI has an integrated TT,
>  	 * (rate matching hub) as opposed to having an OHCI/UHCI
> diff --git a/include/linux/usb.h b/include/linux/usb.h
> index 25a203ac7a7e..57fb4c552740 100644
> --- a/include/linux/usb.h
> +++ b/include/linux/usb.h
> @@ -464,6 +464,10 @@ struct usb_bus {
>  					 * the ep queue on a short transfer
>  					 * with the URB_SHORT_NOT_OK flag set.
>  					 */
> +	unsigned is_eusb2v2:1;		/*
> +					 * true when HC controller supports
> +					 * eusb2v2
> +					 */
>  	unsigned no_sg_constraint:1;	/* no sg constraint */
>  	unsigned sg_tablesize;		/* 0 or largest number of sg list entries */
>  
> @@ -625,6 +629,8 @@ struct usb3_lpm_parameters {
>   * @usb2_hw_lpm_allowed: Userspace allows USB 2.0 LPM to be enabled
>   * @usb3_lpm_u1_enabled: USB3 hardware U1 LPM enabled
>   * @usb3_lpm_u2_enabled: USB3 hardware U2 LPM enabled
> + * @eusb2v2_mps_active: 1024-byte bulk mode is active and must be restored
> + *	 after bus reset.
>   * @string_langid: language ID for strings
>   * @product: iProduct string, if present (static)
>   * @manufacturer: iManufacturer string, if present (static)
> @@ -708,6 +714,7 @@ struct usb_device {
>  	unsigned usb2_hw_lpm_allowed:1;
>  	unsigned usb3_lpm_u1_enabled:1;
>  	unsigned usb3_lpm_u2_enabled:1;
> +	unsigned eusb2v2_mps_active:1;
>  	int string_langid;
>  
>  	/* static strings from the device */
> 
> -- 
> 2.43.0
> 
> 

RE: [PATCH v3 1/3] usb: xhci: Add support for eUSB2v2 1024-byte bulk packet size

[Pawel Laszczak]

>
>On Mon, 05 Oct 2026 10:18:24 +0200, Pawel Laszczak via B4 Relay wrote:
>> From: Pawel Laszczak <pawell@cadence.com>
>>
>> The eUSB2 v2 specification (bcdUSB 0x0230) introduces support for
>> 1024-byte maximum packet sizes for Bulk endpoints in High-Speed mode.
>> However, an eUSB2v2 peripheral will revert its internal maximum packet
>> size back to 512 bytes after events like a bus reset, disconnect, or
>> deconfiguration.
>>
>> To support 1024-byte bulk transfers on capable hosts, add a new
>> is_eusb2v2 flag to the usb_bus structure, populated via the HCCPARAMS2
>> E2V2C capability bit in the xHCI driver.
>>
>> When an eUSB2v2 host configures an eUSB2v2 device, issue a specific
>> SET_FEATURE (USB_DEVICE_BULK_MAX_PACKET_UPDATE) request during
>device
>> configuration to switch the peripheral to 1024-byte packet mode, and
>> allow the xHCI endpoint initialization to accept up to 1024 bytes for
>> HS bulk endpoints.
>>
>> Signed-off-by: Pawel Laszczak <pawell@cadence.com>
>> ---
>> Changes in v3:
>> - Add check for bulk endpoint existence before sending
>>   BULK_MAX_PACKET_UPDATE. This avoids sending the request to devices that
>>   only use isochronous endpoints, as they do not support it.
>> - Do not overwrite ep->desc.wMaxPacketSize to 1024. According to eUSB2v2
>>   spec section 5.2, the endpoint descriptor must always report 512 bytes
>>   regardless of the current operating mode.
>> - Rely on the eusb2v2_mps_active flag in xhci_usb_endpoint_maxp() and
>>   xhci-mem.c to dynamically return 1024 for HS bulk endpoints when the 1KB
>>   mode is active.
>>
>> Changes in v2:
>> - Removed change in config.c: per eUSB2v2 spec section 5.2, conformant
>>   devices always report wMaxPacketSize=512 in their descriptor regardless
>>   of operating mode, so the warning suppression was unnecessary.
>> - xhci-mem.c: simplified HS bulk clamp
>> - xhci.c: moved is_eusb2v2 assignment into xhci_hcd_init_usb2_data()
>> - eusb_update_max_packet(): changed from void to int; returns error on
>>   SET_FEATURE failure.
>> - Added eusb2v2_mps_active flag to struct usb_device to track whether
>>   SET_FEATURE(BULK_MAX_PACKET_UPDATE) succeeded.
>> - Added hub.c: usb_reset_and_verify_device() now re-issues SET_FEATURE
>>   after bus reset to restore 1KB mode. Failure triggers re-enumeration
>>   to prevent a driver from operating with inconsistent MPS state.
>> ---
>>  drivers/usb/core/hub.c      | 16 ++++++++++
>>  drivers/usb/core/message.c  | 74
>+++++++++++++++++++++++++++++++++++++++++++++
>>  drivers/usb/core/usb.h      |  2 ++
>>  drivers/usb/host/xhci-mem.c | 15 +++++++--
>>  drivers/usb/host/xhci.c     | 10 ++++++
>>  include/linux/usb.h         |  7 +++++
>>  6 files changed, 121 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/usb/core/hub.c b/drivers/usb/core/hub.c index
>> 24960ba9caa9..34cfc44c5df8 100644
>> --- a/drivers/usb/core/hub.c
>> +++ b/drivers/usb/core/hub.c
>> @@ -6252,6 +6252,22 @@ static int usb_reset_and_verify_device(struct
>usb_device *udev)
>>  		mutex_unlock(hcd->bandwidth_mutex);
>>  		goto re_enumerate;
>>  	}
>> +
>> +	/*
>> +	 * Restore eUSB2v2 1KB bulk mode after reset (device reverts to 512
>> +	 * after any bus reset per eUSB2v2 spec section 5.2).
>> +	 * Only retry if the initial SET_FEATURE had succeeded.
>> +	 */
>> +	if (udev->eusb2v2_mps_active) {
>> +		ret = eusb_update_max_packet(udev, udev->actconfig);
>> +		if (ret < 0) {
>> +			dev_err(&udev->dev,
>> +				"eUSB2v2: failed to restore 1KB mode after
>reset (%d)\n", ret);
>
>This seems redundant because the function already logs errors.
>Or maybe the function shouldn't log them, if we want different messages for
>different call sites?

I will move dev_warn from eusb_update_max_packet to usb_set_configuration
to distinguish these two cases in the log.

>
>> +			mutex_unlock(hcd->bandwidth_mutex);
>> +			goto re_enumerate;
>> +		}
>> +	}
>> +
>>  	ret = usb_control_msg(udev, usb_sndctrlpipe(udev, 0),
>>  			USB_REQ_SET_CONFIGURATION, 0,
>>  			udev->actconfig->desc.bConfigurationValue, 0, diff --git
>> a/drivers/usb/core/message.c b/drivers/usb/core/message.c index
>> 75e2bfd744a9..c8540d738a3e 100644
>> --- a/drivers/usb/core/message.c
>> +++ b/drivers/usb/core/message.c
>> @@ -2007,6 +2007,71 @@ int usb_set_wireless_status(struct
>> usb_interface *iface,  }  EXPORT_SYMBOL_GPL(usb_set_wireless_status);
>>
>> +/*
>> + * eusb_update_max_packet - enable 1024-byte bulk mode for eUSB2v2
>> +device
>> + * @udev: target device
>> + * @cp: configuration to be checked and enabled
>> + *
>> + * Per eUSB2v2 spec section 5.2, an eUSB2v2 peripheral will revert
>> +the
>> + * maximum packet size to 512 for bulk endpoints after bus reset,
>> +disconnect,
>> + * or deconfiguration.
>> + * This function sends the BULK_MAX_PACKET_UPDATE request to restore
>> +the
>> + * 1024-byte mode. It is valid only if the configuration has bulk endpoints.
>> + */
>> +int eusb_update_max_packet(struct usb_device *udev, struct
>> +usb_host_config *cp) {
>> +	struct usb_host_config *config = cp ? cp : udev->actconfig;
>> +	struct usb_hcd *hcd = bus_to_hcd(udev->bus);
>> +	struct usb_interface_cache *intfc;
>> +	struct usb_host_interface *alt;
>> +	struct usb_host_endpoint *ep;
>> +	bool has_bulk = false;
>> +	int i, j, a;
>> +	int ret;
>> +
>> +	if (le16_to_cpu(udev->descriptor.bcdUSB) != 0x0230 ||
>> +	    !hcd->self.is_eusb2v2)
>> +		return 0;
>> +
>> +	if (!config)
>> +		return 0;
>> +
>> +	for (i = 0; i < config->desc.bNumInterfaces; i++) {
>> +		intfc = config->intf_cache[i];
>> +
>> +		if (!intfc)
>> +			continue;
>> +
>> +		for (a = 0; a < intfc->num_altsetting; a++) {
>> +			alt = &intfc->altsetting[a];
>> +
>> +			for (j = 0; j < alt->desc.bNumEndpoints; j++) {
>> +				ep = &alt->endpoint[j];
>> +
>> +				if (usb_endpoint_xfer_bulk(&ep->desc)) {
>> +					has_bulk = true;
>> +					goto found_bulk;
>> +				}
>> +			}
>> +		}
>> +	}
>> +
>> +	if (!has_bulk)
>> +		return 0;
>
>Control flow ensures that has_bulk is false here and true below.
>The variable is redundant.

Ok

>
>> +found_bulk:
>> +	ret = usb_control_msg(udev, usb_sndctrlpipe(udev, 0),
>> +			      USB_REQ_SET_FEATURE, USB_RECIP_DEVICE,
>> +			      USB_DEVICE_BULK_MAX_PACKET_UPDATE, 0, NULL,
>0,
>> +			      USB_CTRL_SET_TIMEOUT);
>> +	if (ret < 0) {
>> +		dev_warn(&udev->dev, "eUSB2v2 1KB update failed: %d\n",
>ret);
>> +		return ret;
>> +	}
>> +
>> +	return 0;
>> +}
>> +
>>  /*
>>   * usb_set_configuration - Makes a particular device setting be current
>>   * @dev: the device whose configuration is being updated @@ -2123,6
>> +2188,15 @@ int usb_set_configuration(struct usb_device *dev, int
>configuration)
>>  	if (dev->state != USB_STATE_ADDRESS)
>>  		usb_disable_device(dev, 1);	/* Skip ep0 */
>>
>> +	ret = eusb_update_max_packet(dev, cp);
>> +	if (ret < 0)
>> +		dev->eusb2v2_mps_active = 0;
>> +	else if (cp && le16_to_cpu(dev->descriptor.bcdUSB) == 0x0230 &&
>> +		 hcd->self.is_eusb2v2)
>
>Hmm, this code could be simpler if the function returned errors on unsupported
>devices or host controllers. Would that cause problems with anything?

I'll simplify this

>
>> +		dev->eusb2v2_mps_active = 1;
>> +	else
>> +		dev->eusb2v2_mps_active = 0;
>> +
>>  	/* Get rid of pending async Set-Config requests for this device */
>>  	cancel_async_set_config(dev);
>>
>> diff --git a/drivers/usb/core/usb.h b/drivers/usb/core/usb.h index
>> a9b37aeb515b..c51e4261085c 100644
>> --- a/drivers/usb/core/usb.h
>> +++ b/drivers/usb/core/usb.h
>> @@ -89,6 +89,8 @@ extern int usb_major_init(void);  extern void
>> usb_major_cleanup(void);  extern int usb_device_supports_lpm(struct
>> usb_device *udev);  extern int usb_port_disable(struct usb_device
>> *udev);
>> +int eusb_update_max_packet(struct usb_device *udev,
>> +			   struct usb_host_config *cp);
>>
>>  #ifdef	CONFIG_PM
>>
>> diff --git a/drivers/usb/host/xhci-mem.c b/drivers/usb/host/xhci-mem.c
>> index 997fe90f54e5..37a5f9e9bd6d 100644
>> --- a/drivers/usb/host/xhci-mem.c
>> +++ b/drivers/usb/host/xhci-mem.c
>> @@ -1479,10 +1479,19 @@ int xhci_endpoint_init(struct xhci_hcd *xhci,
>>  	/* Allow 3 retries for everything but isoc, set CErr = 3 */
>>  	if (!usb_endpoint_xfer_isoc(&ep->desc))
>>  		err_count = 3;
>> -	/* HS bulk max packet should be 512, FS bulk supports 8, 16, 32 or 64 */
>> +
>> +	/*
>> +	 * HS bulk max packet should be 512 (or 1024 for eUSB2v2),
>> +	 * FS bulk supports 8, 16, 32 or 64.
>> +	 */
>>  	if (usb_endpoint_xfer_bulk(&ep->desc)) {
>> -		if (udev->speed == USB_SPEED_HIGH)
>> -			max_packet = 512;
>> +		if (udev->speed == USB_SPEED_HIGH) {
>> +			if (udev->eusb2v2_mps_active)
>> +				max_packet = 1024;
>> +			else
>> +				max_packet = 512;
>> +		}
>> +
>>  		if (udev->speed == USB_SPEED_FULL) {
>>  			max_packet = rounddown_pow_of_two(max_packet);
>>  			max_packet = clamp_val(max_packet, 8, 64); diff --git
>> a/drivers/usb/host/xhci.c b/drivers/usb/host/xhci.c index
>> a54f5b57f205..b4a3bdf88fb4 100644
>> --- a/drivers/usb/host/xhci.c
>> +++ b/drivers/usb/host/xhci.c
>> @@ -2951,6 +2951,12 @@ int xhci_usb_endpoint_maxp(struct usb_device
>> *udev,  {
>>  	if (usb_endpoint_is_hs_isoc_double(udev, host_ep))
>>  		return le16_to_cpu(host_ep-
>>eusb2_isoc_ep_comp.wMaxPacketSize);
>> +
>> +	if (udev->eusb2v2_mps_active &&
>> +	    udev->speed == USB_SPEED_HIGH &&
>> +	    usb_endpoint_xfer_bulk(&host_ep->desc))
>> +		return 1024;
>
>Is the check for USB_SPEED_HIGH necessary?
>Is eusb2v2_mps_active supposed to be set at other speeds?

This check is purely defensive to guarantee compliance (since according
to the spec, 1KB is strictly an addition for HS).
Generally, it is not needed. I will remove it.

Thanks,
Pawel

>
>One tricky edge case is a high-speed device failing to restore high-speed signaling
>after reset, but at a quick glance this seems to result in hub_port_init() error and
>abort of the reset flow.
>
>> +
>>  	return usb_endpoint_maxp(&host_ep->desc);
>>  }
>>
>> @@ -5371,6 +5377,10 @@ static void xhci_hcd_init_usb2_data(struct
>xhci_hcd *xhci, struct usb_hcd *hcd)
>>  	xhci->usb2_rhub.hcd = hcd;
>>  	hcd->speed = HCD_USB2;
>>  	hcd->self.root_hub->speed = USB_SPEED_HIGH;
>> +
>> +	if (xhci->hcc_params2 & HCC2_E2V2C)
>> +		hcd->self.is_eusb2v2 = 1;
>> +
>>  	/*
>>  	 * USB 2.0 roothub under xHCI has an integrated TT,
>>  	 * (rate matching hub) as opposed to having an OHCI/UHCI diff --git
>> a/include/linux/usb.h b/include/linux/usb.h index
>> 25a203ac7a7e..57fb4c552740 100644
>> --- a/include/linux/usb.h
>> +++ b/include/linux/usb.h
>> @@ -464,6 +464,10 @@ struct usb_bus {
>>  					 * the ep queue on a short transfer
>>  					 * with the URB_SHORT_NOT_OK flag
>set.
>>  					 */
>> +	unsigned is_eusb2v2:1;		/*
>> +					 * true when HC controller supports
>> +					 * eusb2v2
>> +					 */
>>  	unsigned no_sg_constraint:1;	/* no sg constraint */
>>  	unsigned sg_tablesize;		/* 0 or largest number of sg list entries
>*/
>>
>> @@ -625,6 +629,8 @@ struct usb3_lpm_parameters {
>>   * @usb2_hw_lpm_allowed: Userspace allows USB 2.0 LPM to be enabled
>>   * @usb3_lpm_u1_enabled: USB3 hardware U1 LPM enabled
>>   * @usb3_lpm_u2_enabled: USB3 hardware U2 LPM enabled
>> + * @eusb2v2_mps_active: 1024-byte bulk mode is active and must be
>restored
>> + *	 after bus reset.
>>   * @string_langid: language ID for strings
>>   * @product: iProduct string, if present (static)
>>   * @manufacturer: iManufacturer string, if present (static) @@ -708,6
>> +714,7 @@ struct usb_device {
>>  	unsigned usb2_hw_lpm_allowed:1;
>>  	unsigned usb3_lpm_u1_enabled:1;
>>  	unsigned usb3_lpm_u2_enabled:1;
>> +	unsigned eusb2v2_mps_active:1;
>>  	int string_langid;
>>
>>  	/* static strings from the device */
>>
>> --
>> 2.43.0
>>
>>

Re: [PATCH v3 1/3] usb: xhci: Add support for eUSB2v2 1024-byte bulk packet size

[Oliver Neukum]

On 05.10.26 10:18, Pawel Laszczak via B4 Relay wrote:
> From: Pawel Laszczak <pawell@cadence.com>
> 
> The eUSB2 v2 specification (bcdUSB 0x0230) introduces support for
> 1024-byte maximum packet sizes for Bulk endpoints in High-Speed mode.
> However, an eUSB2v2 peripheral will revert its internal maximum packet
> size back to 512 bytes after events like a bus reset, disconnect, or
> deconfiguration.
> 
> To support 1024-byte bulk transfers on capable hosts, add a new
> is_eusb2v2 flag to the usb_bus structure, populated via the HCCPARAMS2
> E2V2C capability bit in the xHCI driver.
> 
> When an eUSB2v2 host configures an eUSB2v2 device, issue a specific
> SET_FEATURE (USB_DEVICE_BULK_MAX_PACKET_UPDATE) request during device
> configuration to switch the peripheral to 1024-byte packet mode, and
> allow the xHCI endpoint initialization to accept up to 1024 bytes for
> HS bulk endpoints.

[..]
> +int eusb_update_max_packet(struct usb_device *udev, struct usb_host_config *cp)
> +{
> +	struct usb_host_config *config = cp ? cp : udev->actconfig;
> +	struct usb_hcd *hcd = bus_to_hcd(udev->bus);
> +	struct usb_interface_cache *intfc;
> +	struct usb_host_interface *alt;
> +	struct usb_host_endpoint *ep;
> +	bool has_bulk = false;
> +	int i, j, a;
> +	int ret;
> +
> +	if (le16_to_cpu(udev->descriptor.bcdUSB) != 0x0230 ||
> +	    !hcd->self.is_eusb2v2)
> +		return 0;
> +
> +	if (!config)
> +		return 0;
> +
> +	for (i = 0; i < config->desc.bNumInterfaces; i++) {
> +		intfc = config->intf_cache[i];
> +
> +		if (!intfc)
> +			continue;
> +
> +		for (a = 0; a < intfc->num_altsetting; a++) {
> +			alt = &intfc->altsetting[a];
> +
> +			for (j = 0; j < alt->desc.bNumEndpoints; j++) {
> +				ep = &alt->endpoint[j];
> +
> +				if (usb_endpoint_xfer_bulk(&ep->desc)) {
> +					has_bulk = true;
> +					goto found_bulk;
> +				}
> +			}
> +		}
> +	}
> +
> +	if (!has_bulk)
> +		return 0;
> +
> +found_bulk:
> +	ret = usb_control_msg(udev, usb_sndctrlpipe(udev, 0),
> +			      USB_REQ_SET_FEATURE, USB_RECIP_DEVICE,
> +			      USB_DEVICE_BULK_MAX_PACKET_UPDATE, 0, NULL, 0,
> +			      USB_CTRL_SET_TIMEOUT);
> +	if (ret < 0) {
> +		dev_warn(&udev->dev, "eUSB2v2 1KB update failed: %d\n", ret);
> +		return ret;
> +	}

Why not handle eusb2v2_mps_active here?

> +
> +	return 0;
> +}
> +
>   /*
>    * usb_set_configuration - Makes a particular device setting be current
>    * @dev: the device whose configuration is being updated
> @@ -2123,6 +2188,15 @@ int usb_set_configuration(struct usb_device *dev, int configuration)
>   	if (dev->state != USB_STATE_ADDRESS)
>   		usb_disable_device(dev, 1);	/* Skip ep0 */
>   
> +	ret = eusb_update_max_packet(dev, cp);
> +	if (ret < 0)
> +		dev->eusb2v2_mps_active = 0;
> +	else if (cp && le16_to_cpu(dev->descriptor.bcdUSB) == 0x0230 &&
> +		 hcd->self.is_eusb2v2)
> +		dev->eusb2v2_mps_active = 1;
> +	else
> +		dev->eusb2v2_mps_active = 0;
> +

You are effectively duplicating the logic of eusb_update_max_packet()

	Regards
		Oliver

[PATCH v3 2/3] usb: gadget: composite: Support eUSB2v2 bulk MPS update

[Pawel Laszczak via B4 Relay]

From: Pawel Laszczak <pawell@cadence.com>

Add support for eUSB2v2 1024-byte Bulk MPS negotiation to the Gadget
Composite framework.

If 'gadget->is_eusb2v2' is set, force bcdUSB to 0x0230 and bMaxPacketSize0
to 64 bytes. Handle the USB_DEVICE_BULK_MAX_PACKET_UPDATE Feature Selector
request by activating the new 'eusb2v2_mps_active' gadget flag.

This flag allows the UDC (USB Device Controller) driver to dynamically
configure its hardware endpoints for 1024-byte packet mode, while the
Composite framework continues to report 512 bytes in the Configuration
Descriptor, ensuring strict compliance with eUSB2v2 specification
section 5.2. The flag is cleared upon device reset or disconnect.

Signed-off-by: Pawel Laszczak <pawell@cadence.com>
---
Changes in v3:
- composite.c: Do not overwrite ep->desc.wMaxPacketSize to 1024, as the eUSB2v2
  spec section 5.2 mandates that the endpoint descriptor always reports 512 bytes
  to the host. Instead, introduce gadget->eusb2v2_mps_active flag.
- gadget.h: Add eusb2v2_mps_active to struct usb_gadget.

Changes in v2:
- composite.c: CLEAR_FEATURE(BULK_MAX_PACKET_UPDATE) is not defined in the
  eUSB2v2 spec; stall it instead of incorrectly setting 1024-byte mode.
- restore MPS in __composite_disconnect
---
 drivers/usb/gadget/composite.c | 28 +++++++++++++++++++++-------
 include/linux/usb/gadget.h     |  4 ++++
 2 files changed, 25 insertions(+), 7 deletions(-)

diff --git a/drivers/usb/gadget/composite.c b/drivers/usb/gadget/composite.c
index dc3664374596..9678ed4e4aee 100644
--- a/drivers/usb/gadget/composite.c
+++ b/drivers/usb/gadget/composite.c
@@ -924,7 +924,6 @@ static void device_qual(struct usb_composite_dev *cdev)
 }
 
 /*-------------------------------------------------------------------------*/
-
 static void reset_config(struct usb_composite_dev *cdev)
 {
 	struct usb_function		*f;
@@ -971,8 +970,10 @@ static int set_config(struct usb_composite_dev *cdev,
 		if (result < 0)
 			goto done;
 	} else { /* Zero configuration value - need to reset the config */
-		if (cdev->config)
+		if (cdev->config) {
 			reset_config(cdev);
+			cdev->gadget->eusb2v2_mps_active = 0;
+		}
 		result = 0;
 	}
 
@@ -1807,7 +1808,11 @@ composite_setup(struct usb_gadget *gadget, const struct usb_ctrlrequest *ctrl)
 				count_configs(cdev, USB_DT_DEVICE);
 			cdev->desc.bMaxPacketSize0 =
 				cdev->gadget->ep0->maxpacket;
-			if (gadget_is_superspeed(gadget)) {
+
+			if (gadget->is_eusb2v2) {
+				cdev->desc.bcdUSB = cpu_to_le16(0x0230);
+				cdev->desc.bMaxPacketSize0 = 64;
+			} else if (gadget_is_superspeed(gadget)) {
 				if (gadget->speed >= USB_SPEED_SUPER) {
 					cdev->desc.bcdUSB = cpu_to_le16(0x0320);
 					cdev->desc.bMaxPacketSize0 = 9;
@@ -2005,12 +2010,20 @@ composite_setup(struct usb_gadget *gadget, const struct usb_ctrlrequest *ctrl)
 	 */
 	case USB_REQ_CLEAR_FEATURE:
 	case USB_REQ_SET_FEATURE:
-		if (!gadget_is_superspeed(gadget))
-			goto unknown;
-		if (ctrl->bRequestType != (USB_DIR_OUT | USB_RECIP_INTERFACE))
-			goto unknown;
 		switch (w_value) {
+		case USB_DEVICE_BULK_MAX_PACKET_UPDATE:
+			if (ctrl->bRequestType != (USB_DIR_OUT | USB_RECIP_DEVICE))
+				goto unknown;
+			if (ctrl->bRequest != USB_REQ_SET_FEATURE)
+				goto unknown;
+			gadget->eusb2v2_mps_active = 1;
+			value = 0;
+			break;
 		case USB_INTRF_FUNC_SUSPEND:
+			if (!gadget_is_superspeed(gadget))
+				goto unknown;
+			if (ctrl->bRequestType != (USB_DIR_OUT | USB_RECIP_INTERFACE))
+				goto unknown;
 			if (!cdev->config || intf >= MAX_CONFIG_INTERFACES)
 				break;
 			f = cdev->config->interface[intf];
@@ -2293,6 +2306,7 @@ static void __composite_disconnect(struct usb_gadget *gadget)
 	 * disconnect callbacks?
 	 */
 	spin_lock_irqsave(&cdev->lock, flags);
+	cdev->gadget->eusb2v2_mps_active = 0;
 	cdev->suspended = 0;
 	if (cdev->config)
 		reset_config(cdev);
diff --git a/include/linux/usb/gadget.h b/include/linux/usb/gadget.h
index 8285b19a25e0..5f691f80fd68 100644
--- a/include/linux/usb/gadget.h
+++ b/include/linux/usb/gadget.h
@@ -420,6 +420,8 @@ struct usb_gadget_ops {
  * @wakeup_armed: True if gadget is armed by the host for remote wakeup.
  * @irq: the interrupt number for device controller.
  * @id_number: a unique ID number for ensuring that gadget names are distinct
+ * @is_eusb2v2: True if controller is Embedded usb2.
+ * @eusb2v2_mps_active: 1024-byte bulk mode is active.
  *
  * Gadgets have a mostly-portable "gadget driver" implementing device
  * functions, handling all usb configurations and interfaces.  Gadget
@@ -483,6 +485,8 @@ struct usb_gadget {
 	unsigned			lpm_capable:1;
 	unsigned			wakeup_capable:1;
 	unsigned			wakeup_armed:1;
+	unsigned			is_eusb2v2:1;
+	unsigned			eusb2v2_mps_active:1;
 	int				irq;
 	int				id_number;
 };

-- 
2.43.0

[PATCH v3 3/3] usb: cdns3: cdnsp: Enable eUSB2v2 1KB bulk packet capability

[Pawel Laszczak via B4 Relay]

From: Pawel Laszczak <pawell@cadence.com>

Implement peripheral-side changes in the Cadence (cdnsp) driver
to support eUSB2v2 1024-byte Bulk MaxPacketSize.

Check the HCC2_E2V2C bit in HCCPARAMS2 during setup and set
'pdev->gadget.is_eusb2v2 = 1' if supported. Delegate the incoming
USB_DEVICE_BULK_MAX_PACKET_UPDATE setup request from ep0 to the
composite layer.

Update cdnsp_endpoint_init() to check the 'eusb2v2_mps_active' flag
set by the gadget framework. If active, dynamically configure the
hardware endpoint for 1024-byte maximum packet size, overriding
the 512-byte value reported in the endpoint descriptor.

Signed-off-by: Pawel Laszczak <pawell@cadence.com>
---
Changes in v3:
- cdnsp-mem.c: Use the new gadget->eusb2v2_mps_active flag instead of modifying
  ep->desc.wMaxPacketSize. This allows the UDC to determine the correct HS bulk
  maximum packet size while letting the composite framework report 512 bytes in
  the endpoint descriptor per eUSB2v2 spec section 5.2.

Changes in v2:
- cdnsp-mem.c: simplified HS bulk clamp logic to: if (max_packet != 1024)
  max_packet = 512, consistent with xhci-mem.c.
---
 drivers/usb/cdns3/cdnsp-ep0.c    | 2 ++
 drivers/usb/cdns3/cdnsp-gadget.c | 4 ++++
 drivers/usb/cdns3/cdnsp-gadget.h | 8 ++++++++
 drivers/usb/cdns3/cdnsp-mem.c    | 9 +++++++--
 drivers/usb/common/debug.c       | 2 ++
 5 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/drivers/usb/cdns3/cdnsp-ep0.c b/drivers/usb/cdns3/cdnsp-ep0.c
index 5cd9b898ce97..090d20b307ee 100644
--- a/drivers/usb/cdns3/cdnsp-ep0.c
+++ b/drivers/usb/cdns3/cdnsp-ep0.c
@@ -253,6 +253,8 @@ static int cdnsp_ep0_handle_feature_device(struct cdnsp_device *pdev,
 		 */
 		cdnsp_enter_test_mode(pdev);
 		break;
+	case USB_DEVICE_BULK_MAX_PACKET_UPDATE:
+		return cdnsp_ep0_delegate_req(pdev, ctrl);
 	default:
 		return -EINVAL;
 	}
diff --git a/drivers/usb/cdns3/cdnsp-gadget.c b/drivers/usb/cdns3/cdnsp-gadget.c
index a5275c2fb43b..ac1dde43ce1d 100644
--- a/drivers/usb/cdns3/cdnsp-gadget.c
+++ b/drivers/usb/cdns3/cdnsp-gadget.c
@@ -1853,6 +1853,10 @@ static int cdnsp_gen_setup(struct cdnsp_device *pdev)
 	pdev->hcc_params = readl(&pdev->cap_regs->hc_capbase);
 	pdev->hci_version = HC_VERSION(pdev->hcc_params);
 	pdev->hcc_params = readl(&pdev->cap_regs->hcc_params);
+	pdev->hcc_params2 = readl(&pdev->cap_regs->hcc_params2);
+
+	if (pdev->hcc_params2 & HCC2_E2V2C)
+		pdev->gadget.is_eusb2v2 = 1;
 
 	/*
 	 * Override the APB timeout value to give the controller more time for
diff --git a/drivers/usb/cdns3/cdnsp-gadget.h b/drivers/usb/cdns3/cdnsp-gadget.h
index c44bca348a41..75bc3f15bde5 100644
--- a/drivers/usb/cdns3/cdnsp-gadget.h
+++ b/drivers/usb/cdns3/cdnsp-gadget.h
@@ -89,6 +89,12 @@ struct cdnsp_cap_regs {
 
 #define CTX_SIZE(_hcc)		(HCC_64BYTE_CONTEXT(_hcc) ? 64 : 32)
 
+/* HCCPARAMS2 - hcc_params2 - bitmasks */
+/* bit 11 - DC support Double BW on a eUSB2 HS ISOC EP */
+#define HCC2_EUSB2_DIC		BIT(11)
+/* bit 12 - DC support eUSB2V2 capability */
+#define HCC2_E2V2C		BIT(12)
+
 /* db_off bitmask - bits 0:1 reserved. */
 #define DBOFF_MASK	GENMASK(31, 2)
 
@@ -1370,6 +1376,7 @@ struct cdnsp_port {
  * @rev_cap: Controller Capabilities Registers.
  * @hcs_params1: Cached register copies of read-only HCSPARAMS1
  * @hcc_params: Cached register copies of read-only HCCPARAMS1
+ * @hcc_params2: Cached register copies of read-only HCCPARAMS2
  * @rtl_revision: Cached controller rtl revision.
  * @setup: Temporary buffer for setup packet.
  * @ep0_preq: Internal allocated request used during enumeration.
@@ -1425,6 +1432,7 @@ struct cdnsp_device {
 	__u32 hcs_params1;
 	__u32 hcs_params3;
 	__u32 hcc_params;
+	__u32 hcc_params2;
 	#define RTL_REVISION_NEW_LPM 0x2700
 	__u32 rtl_revision;
 	/* Lock used in interrupt thread context. */
diff --git a/drivers/usb/cdns3/cdnsp-mem.c b/drivers/usb/cdns3/cdnsp-mem.c
index 5d8cdc91927d..8892767942a7 100644
--- a/drivers/usb/cdns3/cdnsp-mem.c
+++ b/drivers/usb/cdns3/cdnsp-mem.c
@@ -978,8 +978,13 @@ int cdnsp_endpoint_init(struct cdnsp_device *pdev,
 	if (!usb_endpoint_xfer_isoc(pep->endpoint.desc))
 		err_count = 3;
 	if (usb_endpoint_xfer_bulk(pep->endpoint.desc) &&
-	    pdev->gadget.speed == USB_SPEED_HIGH)
-		max_packet = 512;
+	    pdev->gadget.speed == USB_SPEED_HIGH) {
+		if (pdev->gadget.eusb2v2_mps_active)
+			max_packet = 1024;
+		else
+			max_packet = 512;
+	}
+
 	/* Controller spec indicates that ctrl ep avg TRB Length should be 8. */
 	if (usb_endpoint_xfer_control(pep->endpoint.desc))
 		avg_trb_len = 8;
diff --git a/drivers/usb/common/debug.c b/drivers/usb/common/debug.c
index f204cec8d380..b9696fae7ef5 100644
--- a/drivers/usb/common/debug.c
+++ b/drivers/usb/common/debug.c
@@ -46,6 +46,8 @@ static const char *usb_decode_device_feature(u16 wValue)
 		return "U2 Enable";
 	case USB_DEVICE_LTM_ENABLE:
 		return "LTM Enable";
+	case USB_DEVICE_BULK_MAX_PACKET_UPDATE:
+		return "Bulk mps update";
 	default:
 		return "UNKNOWN";
 	}

-- 
2.43.0