drivers: usb: udc: dwc3: initial implementation by josuah · Pull Request #84544 · zephyrproject-rtos/zephyr

josuah · 2025-01-25T04:46:23Z

Add support for Synopsys DWC3 core.

There is no hardware implementing this core in this PR, and users are encouraged to include this commit it on downstream PRs, then I will periodically re-integrate the modifications they needed in this PR.

This can help synchronizing the work of everyone on a common basis.

Active branches using this:

Related issues/PRs:

Add USB support for TI AM62 SK EVM #104675

github-actions · 2025-05-05T00:40:17Z

This pull request has been marked as stale because it has been open (more than) 60 days with no activity. Remove the stale label or add a comment saying that you would like to have the label removed otherwise this pull request will automatically be closed in 14 days. Note, that you can always re-open a closed pull request at any time.

josuah · 2025-05-05T08:38:26Z

Once USB3 support lands (I anticipate after device -> device_next migration?) and once we could test it with IRQ support, I will reopen this PR.

If anybody need this before that happen, do feel free to ping me about it.

github-actions · 2025-07-06T00:44:11Z

This pull request has been marked as stale because it has been open (more than) 60 days with no activity. Remove the stale label or add a comment saying that you would like to have the label removed otherwise this pull request will automatically be closed in 14 days. Note, that you can always re-open a closed pull request at any time.

natto1784 · 2026-01-28T08:46:07Z

can this (and the dependency) be reopened? i feel like this should be a welcome addition

josuah · 2026-01-28T12:26:56Z

@natto1784 yes absolutely. However it might need a platform with which this can be tested with, otherwise it is dead code.

Do you know of any available hardware that can be picked for testing it?

josuah · 2026-01-28T12:45:50Z

It seems like you are working on AM243x?

Here is the current state of this driver:

It is still being developed in rare occasions internally at tinyVision.ai for the needs of the https://tinyclunx33.tinyvision.ai/ (which remains a downstream platform with no current plan of upstreaming, not my choice and could change some day maybe)
Isochronous endpoint support might be on the roadmap
It lacks essential features but works well for as long as we used it with control and bulk endpoints
It requires the USB 3 support to be merged, but it is also possible to upstream an USB2-only version in the meantime.

I am @josuah.demangeon on Discord if that helps.

natto1784 · 2026-02-12T13:38:02Z

Hi, @josuah, after some efforts with minimal changes/fixes to your driver (+ a small glue driver of our own), I have managed to get the CDC ACM sample working on AM62x SK platform which is already in the Zephyr tree, so perhaps we can get this PR re-opened, or is there some other approach to compiling patches in one patchset that you would prefer? I am open to talking on discord as well if that is what you prefer, my nickname is @ amneesh on the Zephyr discord.

In the meantime can you tell me the rationale behind adding the ev_* DT regions? They do not seem to be a standard thing and I have also removed them in my code.

Note that I got it working with the next stack and have not tested with the legacy one, but it should work regardless.

josuah · 2026-02-12T15:54:27Z

Thank you for trying and making it so far!

It looks like there are two parallel efforts to get DWC3 working:

https://github.com/piterzhang/zephyr/tree/dwc3-rk3588

Luckily, both platforms use the same basis for the driver, so the implementation should be relatively similar.

can get this PR re-opened

If it was, I would need to be the one pushing the commits, so best if you open a new one.

have not tested with the legacy one, but it should work regardless.

The legacy stack uses different driver APIs, so it will not work with the legacy stack, but this is not a bad thing as it is expected to be deprecated relatively soon.

josuah · 2026-02-12T16:01:39Z

DT_INST_REG_ADDR_BY_NAME() "REG" here stands for "Register", as in reg = <0x.....>;.

ev_* is what LiteX uses, an FPGA-based SoC architecture, so not relevant to DWC3 driver.

It is best to ignore it.

josuah · 2026-02-12T16:02:11Z

I will ping you in the "DWC3 Upstreaming" thread...

jfischer-no · 2026-02-12T16:16:53Z

can get this PR re-opened

If it was, I would need to be the one pushing the commits, so best if you open a new one.

Let's keep it open until there is a new one. Alternatively, you can allow to push to this branch.

jfischer-no · 2026-02-12T16:17:50Z

@@ -0,0 +1,1593 @@
+/*
+ * SPDX-License-Identifier: MIT


@josuah Why is license MIT?

It used to be the fallback license that tinyVision.ai used for everything, I can request it to be changed.

natto1784 · 2026-02-12T17:15:41Z

I understand that. I meant to ask whether it would be alright to get rid of it as it's vendor specific code. It would be alright to add it as some sort of quirk I think. Let's discuss more on discord.

…

On 12 February 2026 9:33:27 pm IST, josuah ***@***.***> wrote: josuah left a comment (zephyrproject-rtos/zephyr#84544) `DT_INST_REG_ADDR_BY_NAME()` "REG" here stands for "Register", as in `reg = <0x.....>;`. `ev_*` is what LiteX uses, an FPGA-based SoC architecture, so not relevant to DWC3 driver. It is best to ignore it. -- Reply to this email directly or view it on GitHub: #84544 (comment) You are receiving this because you were mentioned. Message ID: ***@***.***>

josuah · 2026-02-12T17:25:37Z

get rid of it as it's vendor specific code

Yes, totally, it better not be there and was planned to be removed for something more generic.

jfischer-no · 2026-04-01T15:35:43Z

@josuah Do you want to have USB3 support in the stack, or will you limit this driver to support high-speed only and add USB3 support in the stack later?

josuah · 2026-04-01T18:11:14Z

@jfischer-no I was thinking of keeping it USB2, and have a separate USB3 PR based on top of this DWC3 PR.

This way all the noise about DWC3 hardware can be kept away from USB3 protocol discussions.

If anything else is preferred let me know!

natto1784 · 2026-04-06T10:30:53Z

I for one, do not mind HS-only DWC3 driver first before adding SS support. Also I think we can open this for review soon, would be great if we could get this merged for 4.5.

natto1784 · 2026-05-05T10:45:58Z

Hi, has there been some progress on using udc_setup_received yet? I can volunteer to make the change if you're bit tied up with stuff.

josuah · 2026-05-09T00:50:41Z

Force-push: refactor according to #103493

Tested on a CrossLinkU-NX33 FPGA, but no result so far (needs vendor quirks for it).

natto1784 · 2026-05-09T04:15:50Z

+		 * ACK/NAK the first packets from the host with the new address, otherwise the host
+		 * issue a reset
+		 */
+		if (cfg->setup_pkt->bmRequestType == 0x00 && cfg->setup_pkt->bRequest == 0x05) {


Suggested change

if (cfg->setup_pkt->bmRequestType == 0x00 && cfg->setup_pkt->bRequest == 0x05) {

if (cfg->setup_pkt->bmRequestType == USB_REQTYPE_TYPE_STANDARD && cfg->setup_pkt->bRequest == USB_SREQ_SET_ADDRESS) {

Thank you. This is still WIP but that is one thing less to lookup :)

natto1784 · 2026-05-09T04:18:22Z

+			udc_dwc3_do_set_address(dev, cfg->setup_pkt->wValue);
+		}
+
+		udc_setup_received(dev, cfg->setup_pkt);


Note: You can also directly use the programmed setup buffer in your EP0 queue by passing NULL here, I wonder which one is a better approach

I realized that the stack would submit the setup buffer, so yes this can be simplified by one less special case.

Maybe drivers that use a local buffer for setup are those which constantly need a buffer ready for setup.

Yes, a lot better thank you.

jfischer-no · 2026-05-09T23:49:21Z

+		/* Latency optimization: set the address immediately to be able to be able to
+		 * ACK/NAK the first packets from the host with the new address, otherwise the host
+		 * issue a reset


Set addr_before_status controller controller capability.

This was missing thank you.

josuah · 2026-05-11T22:18:57Z

Force-push: fix all runtime bugs I could find when running CDC-ACM. Still some TODO left through the code.

migrate to the new UDC API
get CDC ACM working and fix the bugs
Remove all the Lattice-specific registers to make our own vendor quirks.
I import the changes suggested by @piterzhang, @natto1784 and test again on tinyCLUXN33
opening the PR for early review (with In Progress label still)
@piterzhang, @natto1784 can suggest a vendor quirk ready to merge
run usbtest

Add support for Synopsys DWC3 core, with only BULK endpoints supported at the moment, as well as an initial vendor quirks system. Signed-off-by: Josuah Demangeon <me@josuah.net>

josuah · 2026-05-12T01:19:42Z

I finished migrating this to be using IRQs only without a workqueue thread.

The tinyCLUNX33 platform still does not have working IRQs, and uses a 1ms timer to simulate this (called from the tinyCLUNX33 vendor quirk, not part of this PR).

This means all the event handlers are only tested in non-IRQ context, and some modification might be needed to let all of this run from IRQ, such as reintroducing a work queue if needed.

Let me know what error you encounter.

sonarqubecloud · 2026-05-12T01:42:35Z

Quality Gate passed

Issues
15 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

josuah marked this pull request as draft January 25, 2025 04:46

zephyrbot added the area: USB Universal Serial Bus label Jan 25, 2025

zephyrbot requested review from jfischer-no and tmon-nordic January 25, 2025 04:46

zephyrbot assigned jfischer-no Jan 25, 2025

josuah added area: Drivers Experimental Experimental features not enabled by default labels Jan 25, 2025

carlescufi requested a review from XenuIsWatching March 5, 2025 19:15

github-actions Bot added the Stale label May 5, 2025

github-actions Bot removed the Stale label May 6, 2025

github-actions Bot added the Stale label Jul 6, 2025

github-actions Bot closed this Jul 20, 2025

jfischer-no reopened this Feb 12, 2026

jfischer-no removed the Stale label Feb 12, 2026

jfischer-no reviewed Feb 12, 2026

View reviewed changes

josuah force-pushed the pr-dwc3 branch 2 times, most recently from eb1629d to 3ee1194 Compare February 12, 2026 19:08

kartben assigned tmon-nordic May 4, 2026

josuah force-pushed the pr-dwc3 branch 2 times, most recently from 81a1786 to 1dadb54 Compare May 9, 2026 00:48

natto1784 reviewed May 9, 2026

View reviewed changes

josuah force-pushed the pr-dwc3 branch from 1dadb54 to 3c8c2ac Compare May 9, 2026 11:06

jfischer-no requested changes May 9, 2026

View reviewed changes

josuah force-pushed the pr-dwc3 branch from 3c8c2ac to 64e0c92 Compare May 10, 2026 21:57

josuah added the In progress For PRs: is work in progress and should not be merged yet. For issues: Is being worked on label May 10, 2026

josuah force-pushed the pr-dwc3 branch 4 times, most recently from ec56b9f to 18f7fba Compare May 11, 2026 21:59

josuah force-pushed the pr-dwc3 branch 2 times, most recently from c178cfa to 4e6960f Compare May 12, 2026 01:12

drivers: usb: udc: dwc3: initial implementation

4e6960f

Add support for Synopsys DWC3 core, with only BULK endpoints supported at the moment, as well as an initial vendor quirks system. Signed-off-by: Josuah Demangeon <me@josuah.net>

josuah marked this pull request as ready for review May 12, 2026 01:23

zephyrbot added area: Tests Issues related to a particular existing or missing test area: Devicetree Binding PR modifies or adds a Device Tree binding labels May 12, 2026

zephyrbot requested review from JarmouniA, MarkWangChinese and nashif May 12, 2026 01:25

	if (cfg->setup_pkt->bmRequestType == 0x00 && cfg->setup_pkt->bRequest == 0x05) {
	if (cfg->setup_pkt->bmRequestType == USB_REQTYPE_TYPE_STANDARD && cfg->setup_pkt->bRequest == USB_SREQ_SET_ADDRESS) {

Conversation

josuah commented Jan 25, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions Bot commented May 5, 2025

Uh oh!

josuah commented May 5, 2025

Uh oh!

github-actions Bot commented Jul 6, 2025

Uh oh!

natto1784 commented Jan 28, 2026

Uh oh!

josuah commented Jan 28, 2026

Uh oh!

josuah commented Jan 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

natto1784 commented Feb 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

josuah commented Feb 12, 2026

Uh oh!

josuah commented Feb 12, 2026

Uh oh!

josuah commented Feb 12, 2026

Uh oh!

jfischer-no commented Feb 12, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

natto1784 commented Feb 12, 2026 via email

Uh oh!

josuah commented Feb 12, 2026

Uh oh!

jfischer-no commented Apr 1, 2026

Uh oh!

josuah commented Apr 1, 2026

Uh oh!

natto1784 commented Apr 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

natto1784 commented May 5, 2026

Uh oh!

josuah commented May 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

natto1784 May 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

josuah commented May 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

josuah commented May 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

sonarqubecloud Bot commented May 12, 2026

Quality Gate passed

Uh oh!

Reviewers

Assignees

Labels

josuah commented Jan 25, 2025 •

edited

Loading

josuah commented Jan 28, 2026 •

edited

Loading

natto1784 commented Feb 12, 2026 •

edited

Loading

natto1784 commented Apr 6, 2026 •

edited

Loading

josuah commented May 9, 2026 •

edited

Loading

natto1784 May 9, 2026 •

edited

Loading

josuah commented May 11, 2026 •

edited

Loading

josuah commented May 12, 2026 •

edited

Loading