Skip to content

drivers: usb: uhc_dwc2: add initial support #102967

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
roma-jam wants to merge 2 commits into
base: main
Choose a base branch
from
Open

drivers: usb: uhc_dwc2: add initial support #102967

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
e196de8
drivers: usb: common: Add dwc2 core get revision
roma-jam May 7, 2026
b2e731d
drivers: usb: uhc_dwc2: add initial support
roma-jam Jan 8, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 6 additions & 0 deletions drivers/usb/common/usb_dwc2_hw.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -374,7 +374,13 @@ USB_DWC2_SET_FIELD_DEFINE(gnptxfsiz_nptxfstaddr, GNPTXFSIZ_NPTXFSTADDR)
#define USB_DWC2_GGPIO_STM32_PWRDWN BIT(USB_DWC2_GGPIO_STM32_PWRDWN_POS)

/* GSNPSID register */
#define USB_DWC2_GSNPSID_REV_2_92A 0x4F54292AUL
#define USB_DWC2_GSNPSID_REV_4_20A 0x4F54420AUL
#define USB_DWC2_GSNPSID_REV_5_00A 0x4F54500AUL
#define USB_DWC2_GSNPSID_REV_POS 0UL
#define USB_DWC2_GSNPSID_REV_MASK (0xFFFFUL << USB_DWC2_GSNPSID_REV_POS)

USB_DWC2_GET_FIELD_DEFINE(gsnpsid_rev, GSNPSID_REV)
Comment on lines +377 to +383
Copy link
Copy Markdown
Contributor

@tmon-nordic tmon-nordic May 7, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This should go into separate commit (same PR is fine, just do it in a first commit).

Copy link
Copy Markdown
Contributor Author

@roma-jam roma-jam May 7, 2026 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Done, now this PR contains two commits.

Just for me to the future: what is the reason for that and which question should I ask myself to understand, do I need to split changes in several commits or can keep them in one commit?

Copy link
Copy Markdown
Contributor

@tmon-nordic tmon-nordic May 8, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The reason is to make things cherry-pickable and also to make things easily revertable. The register header is used by both UDC and UHC. If there were some changes to be backported to UDC (especially if there were some that also touched the register definitions), then keeping the commits separate makes it possible to easily omit the UHC part.


/* GHWCFG1 register */
#define USB_DWC2_GHWCFG1 0x0044UL
Expand Down
2 changes: 2 additions & 0 deletions drivers/usb/uhc/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,8 +3,10 @@
# SPDX-License-Identifier: Apache-2.0

zephyr_library()
zephyr_library_include_directories(${ZEPHYR_BASE}/drivers/usb/common/)
Copy link
Copy Markdown
Contributor

@sylvioalves sylvioalves Mar 2, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Perhaps this should be inside if(CONFIG_UHC_DWC2)?

Copy link
Copy Markdown
Contributor

@josuah josuah Apr 24, 2026 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The current implementation seems to match UDC. This probably matches the maintainer/collaborators preferred organization.

zephyr/drivers/usb/udc/CMakeLists.txt

Lines 6 to 7 in 7edd883

zephyr_library_sources(udc_common.c)
zephyr_library_include_directories(${ZEPHYR_BASE}/drivers/usb/common/)

On my side no preference.

Copy link
Copy Markdown
Contributor Author

@roma-jam roma-jam May 4, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

There are some functions, that might be merged with their implementation in udc.

For example:

  • dwc2_core_soft_reset()
  • dwc2_flush_rx_fifo()
  • dwc2_flush_tx_fifo()

So, after the uhc will be merged, I can check the possibility to merge them and propose the dwc2_common.c for such functions and share them between udc and uhc.

During this, it will be possible to consider adding the common/dwc2 folder and include it as if(CONFIG_UDC_DRIVER OR CONFIG_UHC_DRIVER).

Otherwise, I do not mind to keep it similar in udc and uhc for now.


zephyr_library_sources(uhc_common.c)
zephyr_library_sources_ifdef(CONFIG_UHC_DWC2 uhc_dwc2.c)
zephyr_library_sources_ifdef(CONFIG_UHC_MAX3421E uhc_max3421e.c)
zephyr_library_sources_ifdef(CONFIG_UHC_VIRTUAL uhc_virtual.c)
zephyr_library_sources_ifdef(CONFIG_UHC_NXP_EHCI uhc_mcux_common.c uhc_mcux_ehci.c)
Expand Down
1 change: 1 addition & 0 deletions drivers/usb/uhc/Kconfig
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -43,6 +43,7 @@ module = UHC_DRIVER
module-str = uhc drv
source "subsys/logging/Kconfig.template.log_config"

source "drivers/usb/uhc/Kconfig.dwc2"
source "drivers/usb/uhc/Kconfig.max3421e"
source "drivers/usb/uhc/Kconfig.virtual"
source "drivers/usb/uhc/Kconfig.mcux"
Expand Down
90 changes: 90 additions & 0 deletions drivers/usb/uhc/Kconfig.dwc2
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,90 @@
# Copyright (c) 2026 Roman Leonov <jam_roma@yahoo.com>
# SPDX-License-Identifier: Apache-2.0

config UHC_DWC2
bool "DWC2 USB host controller driver"
depends on DT_HAS_SNPS_DWC2_ENABLED
default y
Copy link
Copy Markdown
Contributor

@sylvioalves sylvioalves May 10, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

In PR description, you wrote that building needs -DCONFIG_UDC_DWC2=n. If that is the case, why not removing default y from here?

Copy link
Copy Markdown
Contributor

@tmon-nordic tmon-nordic May 11, 2026 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What about simply add depends on !CONFIG_UDC_DWC2 line for the time being? This could be removed once (if at all) there is dynamic role switching implemented.

Copy link
Copy Markdown
Contributor Author

@roma-jam roma-jam May 11, 2026 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

According to the uhc driver implementation, there is not strict dependency from UDC and it is possible to have both at the same time.
For example, there should be no problem in having both drivers complied, but not running at the same time. That means, we can switch between them, via uhc_init, uhc_enable, then uhc_disable, uhc_shutdown and then udc_init and udc_enable (but uhc_shutdown is not implemented yet).

I added the -DCONFIG_UDC_DWC2=n to the build command at a very early stage, when there was a conflict during the build (multiple definition of `__device_dts_ord_75').

Right now I know a bit more about the environment so I'll double check the conflict, maybe it is possible to make it right. But if it is a DT- conflict in utilizing the same peripheral (usb_otg), then let me know.

UPD: OK, seems that it is technically impossible. So seems that right now it gives us only two options: make them mutually exclusive or assign the node to UDC/UHC via menuconfig.

I'll think about it a bit more, but let me know if there are any preferences.

Copy link
Copy Markdown
Contributor

@josuah josuah May 12, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It might only have been a problem before this was merged (which required manually turning off the device mode): #105005

As enabling host/device is still happening manually so it might work to enable both by default, and the user would then be responsible to enable host or device stack.

Copy link
Copy Markdown
Contributor Author

@roma-jam roma-jam May 12, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If we enable both at the same time - we get the multiple definition of the node during the build.

I don't think (please, correct me if I am missing something) that changing the device tree description to having the assign of the role there something like this does make sense:

usb_otg: usb@40080000 {
    compatible = "snps,dwc2";
    reg = <0x40080000 0x40000>;
    interrupts = <...>;
    zephyr,usb-role = "host";
    status = "okay";
};

What about simply add depends on !CONFIG_UDC_DWC2 line for the time being?

So, I think that simple mutual exclusive like above is the best option at the moment.

Copy link
Copy Markdown
Contributor

@josuah josuah May 12, 2026 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The devicetree is there to describe what exists in the hardware, and things like "role" is how this hardware is used by drivers, which I think is a different concern.

If a device needs multiple drivers (i.e. host and device mode), indeed it needs multiple devicetree nodes, such as having child nodes, one for device and one for host, both always enabled and Kconfig used to select what to leverage, as usual?

Though this seems a long-term issue, not needing a solution on this PR? Still glad to discuss it.

select EVENTS
help
DWC2 USB host controller driver.

if UHC_DWC2

config UHC_DWC2_MAX_CHANNELS
Comment thread
roma-jam marked this conversation as resolved.
int "UHC DWC2 driver maximum supported number of channels"
default 16
range 1 16
help
Number of channels driver can support.
Reduce this value to save memory.
Increase to support more concurrent transfers.

config UHC_DRIVER_HIGH_SPEED_SUPPORT_ENABLED
bool "UHC DWC2 High speed"
default y if SOC_SERIES_NRF54LX
help
Enables HS support code if the SoC/board can support it.

config UHC_DWC2_STACK_SIZE
int "UHC DWC2 driver thread stack size"
default 512
help
DWC2 driver thread stack size.

config UHC_DWC2_THREAD_PRIORITY
int "UHC DWC2 driver thread priority"
default 8
help
DWC2 driver thread priority.

config UHC_DWC2_DEBOUNCE_DELAY_MS
int "Debounce delay in ms"
default 250
help
On connection of a USB device, the USB 2.0 specification requires
a "debounce interval with a minimum duration of 100ms" to allow the connection to stabilize
(see USB 2.0 chapter 7.1.7.3 for more details).
During the debounce interval, no new connection/disconnection events are registered.

The default value is set to 250 ms to be safe.

config UHC_DWC2_RESET_HOLD_MS
int "Reset hold in ms"
default 50
help
Root-port reset hold time in milliseconds.

USB 2.0 requires ordinary reset signaling to be asserted for at least
10 ms, but reset issued from a root port must last at least 50 ms..

The default value is set to 50 ms.

config UHC_DWC2_RESET_RECOVERY_MS
int "Reset recovery delay in ms"
default 30
help
Post-reset recovery delay in milliseconds.

USB 2.0 requires system software to allow at least 10 ms for reset
recovery after the port stops driving reset, before the attached
device is expected to respond to data transfers.

The attached device may ignore data transfers during this recovery
interval.

The default value is set to 30 ms.

config UHC_DWC2_SET_ADDR_DELAY_MS
int "Delay after set device address"
default 10
help
"After successful completion of the Status stage, the device is allowed a SetAddress()
recovery interval of 2 ms. At the end of this interval, the device must be able to accept
Setup packets addressed to the new address. Also, at the end of the recovery interval, the
device must not respond to tokens sent to the old address (unless, of course, the old and new
address is the same)." See USB 2.0 chapter 9.2.6.3 for more details.

The default value is set to 10 ms to be safe.

endif # UHC_DWC2
Loading
Loading