u-boot.git/drivers/usb/gadget, branch next

usb: ci_udc: verify all dtds are inactive before completing request

2026-07-22T07:16:16Z

According to device mode spec, the ACTIVE status field of dtds should be check to determine whether the transfers completed successfully. However, this is not implemented in handle_ep_complete. When two EPs are enabled and transferring, EPa requests with multiple dtds and EPb request with one dtd. Irq is triggred on EPb. The udc_irq handler finds both EPb's and EPa's ENDPTCOMPLETE=1 while not all of EPa's dtds have been completed. Because ACTIVE status is not checked, this case causes crash in ci_udc driver. Signed-off-by: Jason He Signed-off-by: Ye Li Link: https://patch.msgid.link/20260522075512.1291485-3-ye.li@nxp.com Signed-off-by: Mattijs Korpershoek

usb: ci_udc: Update usb request status

2026-07-22T07:16:16Z

usb request status is not set in ci_udc driver. However in ci_ep_dequeue, it checks the request status for executing complete callback, so upper layer is not able to free these enqueued requests when calling usb_ep_dequeue. Signed-off-by: Ye Li Fixes: 26cc5129ee64 ("USB: gadaget: add Marvell controller support") Reviewed-by: Mattijs Korpershoek Link: https://patch.msgid.link/20260522075512.1291485-2-ye.li@nxp.com Signed-off-by: Mattijs Korpershoek

usb: ci_udc: set correct ep type in ep_enable()

2026-07-22T07:16:16Z

The Endpoint Type in ENDPTCTRL register should be correctly set according to endpoint descriptor, not fixed to bulk. Otherwise, unexpected behaviors may happen. Fixes: 26cc5129ee64 ("USB: gadaget: add Marvell controller support") Signed-off-by: Xu Yang Signed-off-by: Ye Li Reviewed-by: Mattijs Korpershoek Link: https://patch.msgid.link/20260522075512.1291485-1-ye.li@nxp.com Signed-off-by: Mattijs Korpershoek

usb: gadget: dwc2: Set endpoint capabilities via .caps field

2026-07-22T06:54:16Z

Endpoint capabilities can be described via the .caps field of struct usb_ep since commit 77dcbdf3c1ce ("usb: gadget: Add match_ep() op to usb_gadget_ops") In linux, the following commit introduces it: b86f33a3a371 ("usb: gadget: epautoconf: add endpoint capabilities flags verification") Linux's change also reworks the ep matching function by relying on these new capabilities instead of guessing based on the endpoint name. Prepare the dwc2 UDC driver for capability-based matching by adding the required capabilities for each endpoint. Acked-by: Ilias Apalodimas Link: https://patch.msgid.link/20260703-usb-prep-dwc3-sync-v1-2-1352bc238c9b@kernel.org Signed-off-by: Mattijs Korpershoek

usb: gadget: dwc2: Use usb_ep_set_maxpacket_limit() for fifo sizes

2026-07-22T06:54:16Z

More recent versions of usb_ep_autoconfig() use the maxpacket_limit field from struct usb_ep. This field has been introduced in the linux kernel via commit e117e742d310 ("usb: gadget: add "maxpacket_limit" field to struct usb_ep") And has been ported to U-Boot via commit 747a0a5b387f ("usb: dwc3: gadget: make dwc3 gadget build in uboot") However, the dwc2 gadget driver was not converted to use usb_ep_set_maxpacket_limit(). This is harmless for now, but the driver will break once usb_ep_autoconfig() is updated to a more recent version. Set maxpacket_limit via the recommended helper function to prepare for a core gadget update. Acked-by: Ilias Apalodimas Link: https://patch.msgid.link/20260703-usb-prep-dwc3-sync-v1-1-1352bc238c9b@kernel.org Signed-off-by: Mattijs Korpershoek

cmd: ums: Set Device Descriptor iSerial based on serial# env variable

2026-07-22T06:53:37Z

The DFU code currently sets the content of "serial#" environment variable into Device Descriptor iSerial field. This is useful when there are multiple devices connected to the same host PC and they have to be told apart. Replicate the same behavior in the UMS code. Example without serial# set, which is the original behavior: " u-boot=> ums 0 mmc 1 UMS: LUN 0, dev mmc 1, hwpart 0, sector 0x0, count 0x1d5a000 $ dmesg -w ... usb 5-2.3.7: New USB device found, idVendor=0525, idProduct=a4a5, bcdDevice=7e.a7 usb 5-2.3.7: New USB device strings: Mfr=1, Product=2, SerialNumber=0 usb 5-2.3.7: Product: USB download gadget usb 5-2.3.7: Manufacturer: Vendor <------------------------------- NOT HERE usb-storage 5-2.3.7:1.0: USB Mass Storage device detected ... $ lsusb -vd 0525:a4a5 Bus 005 Device 051: ID 0525:a4a5 Netchip Technology, Inc. Linux-USB File-backed Storage Gadget Negotiated speed: High Speed (480Mbps) Device Descriptor: bLength 18 bDescriptorType 1 bcdUSB 2.00 bDeviceClass 0 [unknown] bDeviceSubClass 0 [unknown] bDeviceProtocol 0 bMaxPacketSize0 64 idVendor 0x0525 Netchip Technology, Inc. idProduct 0xa4a5 Linux-USB File-backed Storage Gadget bcdDevice 7e.a7 iManufacturer 1 Vendor iProduct 2 USB download gadget iSerial 0 <------------------------------- NOT HERE ... " Example with serial# set to 1234abcd, which is the optional new behavior: " u-boot=> setenv serial# 1234abcd <------------------------------ HERE u-boot=> ums 0 mmc 1 UMS: LUN 0, dev mmc 1, hwpart 0, sector 0x0, count 0x1d5a000 $ dmesg -w ... usb 5-2.3.7: New USB device found, idVendor=0525, idProduct=a4a5, bcdDevice=7e.a7 usb 5-2.3.7: New USB device strings: Mfr=1, Product=2, SerialNumber=3 usb 5-2.3.7: Product: USB download gadget usb 5-2.3.7: Manufacturer: Vendor usb 5-2.3.7: SerialNumber: 1234abcd <--------------------------- HERE usb-storage 5-2.3.7:1.0: USB Mass Storage device detected ... $ lsusb -vd 0525:a4a5 Bus 005 Device 052: ID 0525:a4a5 Netchip Technology, Inc. Linux-USB File-backed Storage Gadget Negotiated speed: High Speed (480Mbps) Device Descriptor: bLength 18 bDescriptorType 1 bcdUSB 2.00 bDeviceClass 0 [unknown] bDeviceSubClass 0 [unknown] bDeviceProtocol 0 bMaxPacketSize0 64 idVendor 0x0525 Netchip Technology, Inc. idProduct 0xa4a5 Linux-USB File-backed Storage Gadget bcdDevice 7e.a7 iManufacturer 1 Vendor iProduct 2 USB download gadget iSerial 3 1234abcd <-------------------------- HERE ... " Signed-off-by: Marek Vasut Reviewed-by: Mattijs Korpershoek Link: https://patch.msgid.link/20260626142840.305478-1-marex@nabladev.com Signed-off-by: Mattijs Korpershoek

usb: gadget: f_mass_storage: Disable eps during disconnect

2026-06-26T09:05:21Z

When trying two ums commands in a row, the second one no longer enumerates properly from the host. This happens since commit 59310d1ecb9f ("usb: gadget: introduce 'enabled' flag in struct usb_ep") causing usb_ep_enable() to return early when ep->enabled is already set. Gadget function drivers (such as f_fastboot or f_mass_storage) implement a disable() function which is called whenever we are done using the gadget. Because f_mass_storage driver does not disable the endpoints, ep->enabled will never be set to false again. This can be reproduced on the STM32MP157C-DK2 or the Khadas VIM3 boards. Add calls to usb_ep_disable() as done in linux [1] to fix this. [1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=9fff139aeb11186fd8e75860c959c86cb43ab2f6 Fixes: 59310d1ecb9f ("usb: gadget: introduce 'enabled' flag in struct usb_ep") Reported-by: Patrice Chotard Reviewed-by: Patrice Chotard Link: https://patch.msgid.link/20260617-ums-disconnect-v2-1-6e52d9de1d36@kernel.org Signed-off-by: Mattijs Korpershoek

usb: f_sdp: handle the spl load function failure

2026-06-26T09:01:42Z

Current implementation does not check the return value of spl load function. If the spl load is failed, SPL may meet crash due to spl_image variable is not initialized. Add the failure check, so SPL can print and stop with error. Signed-off-by: Ye Li Fixes: 2c72ead73874 ("usb: gadget: f_sdp: Allow SPL to load and boot FIT via SDP") Reviewed-by: Mattijs Korpershoek Link: https://patch.msgid.link/20260522072705.1156220-1-ye.li@nxp.com Signed-off-by: Mattijs Korpershoek

Kconfig: usb: restyle

2026-06-25T20:58:13Z

Restyle all Kconfigs for "usb": Menu entries : no space left Menu attributes: 1 TAB Help text : 1 TAB + 2 spaces Replace '---help---' by 'help' Signed-off-by: Johan Jonker

usb: gadget: atmel: use calloc() to allocate endpoint list

2026-06-15T07:15:25Z

malloc() doesn't zero out memory, leaving ep->ep.enabled uninitiated, which could make this flag falsely true. In next usb_ep_enable() call since this flag is true, ep->ops->enable() will be skipped. Then usb_ep_queue() will fail on uninitialized endpoint. Fixes: 59310d1ecb9f ("usb: gadget: introduce 'enabled' flag in struct usb_ep") Signed-off-by: Zixun LI Reviewed-by: Marek Vasut Reviewed-by: Mattijs Korpershoek Link: https://patch.msgid.link/20260606-udc_malloc-v2-1-1155326b182b@hifiphile.com Signed-off-by: Mattijs Korpershoek