Re: [PATCH v7 00/24] usb: dwc3: part 2, make the resynced code work

[Love Kumar <love.kumar@amd.com>]

Hi,

On 8/24/26 9:11 PM, Anders Roxell wrote:
> Hi,
> 
> This is part two of two. It applies on top of part one [5], which resets
> the DWC3 driver and imports it up to Linux v6.16. Part one alone breaks
> the build for every board that enables DWC3. This part adds back the
> U-Boot glue, the XHCI and UDC updates and the build fixes, so the tree
> builds and the gadget works again. It builds on Jerome's work [2].
> 
> The two parts have to go in together.
> 
> The resynced epautoconf matches endpoints on their caps field instead of
> on their name. Five UDC drivers never filled that in, so each one gets a
> patch next to its other fixup: at91_udc, atmel_usba_udc, ci_udc, musb-new
> and mtu3. With those and dwc2, every UDC driver in the tree sets caps.
> 
> Build tested one board per driver: every DWC3 glue layer in the tree,
> dwc2 with and without DM_USB_GADGET, at91_udc, atmel_usba_udc, ci_udc,
> musb-new and mtu3. 17 boards in total.
> 
> DFU is tested on hardware on BeagleBoard-X15, and on BeaglePlay by
> Anshul. ci_udc is tested on an i.MX93 board with ums. The rest is build
> tested only.
> 
> The branch has both parts [4].
> 
> [1] v5: https://lore.kernel.org/u-boot/20260716134305.614278-1-anders.roxell@linaro.org/
> [2] Jerome's adaptation work: https://lore.kernel.org/all/20251121153812.2076440-1-jerome.forissier@linaro.org/
> [3] dwc2 prep: https://lore.kernel.org/all/20260703-usb-prep-dwc3-sync-v1-0-1352bc238c9b@kernel.org/
> [4] https://github.com/roxell/u-boot/tree/usb_dwc3_sync_v7
> [5] part one, subject: [PATCH v7 00/63] usb: dwc3: part 1, reset and
>      import from Linux v6.16
> 
> v6 -> v7:
> - Split the series in two. Tom asked for this, so review and iteration
>    gets easier. Part one is the reset and the imports, part two makes it
>    work.
> - No code changes in these 24 patches.
> - v6 never reached the list in full. My send died in the middle, so the
>    cover letter and the last patches went missing.
> 
> v5 -> v6:
> - Split the "add helpers" patch into a dm patch and a usb phy patch.
>    Mattijs asked for this.
> - Kept the U-Boot dwc3 Makefile instead of the kernel one. This is what
>    Marek meant with glue drivers that are never compiled or used.
> - Brought back samsung_usb_phy.c in an own commit. The restore commit
>    deleted it and nothing added it back, so odroid-xu3 lost its USB 3.0
>    phy setup.
> - Stopped moving functions around in dwc3 gadget.c. Where a function
>    still had to move, the commit message now says why. Mattijs said the
>    last patches were hard to review because of this.
> - Split "fix up musb-new and mtu3" and "fix up xhci and cdns3" into one
>    patch per controller. David asked for this.
> - Set endpoint capabilities in at91_udc, atmel_usba_udc, ci_udc, musb-new
>    and mtu3. Mattijs asked for this on v4, after the VIM3 fastboot
>    regression. Each patch sits next to the other fixup for that driver.
> - at91_udc now calls usb_add_gadget_udc(). Without it udc-core never sees
>    the controller.
> - Enabled async callbacks on bind, and added a cache flush and invalidate
>    around ep0 transfers. Without these no dwc3 gadget enumerates. Anshul
>    found them on BeaglePlay, I hit them on BeagleBoard-X15. DFU works on
>    both now.
> - ci_udc got udc_start() and udc_stop(). udc-core calls udc_start() on
>    bind, ci_udc did not have it, so the board crashed with a synchronous
>    abort when ums started. Found on i.MX93, ums works there now.
> - udc-core got two more fixes while I looked at that. It checks
>    udc_start(), udc_stop() and get_frame() before calling them, they are
>    mandatory in the kernel but not in u-boot. And it reports what
>    usb_gadget_udc_bringup() returned, before it only said "No UDC
>    available".
> - Rebased on current mainline. The ci_udc patch shrank a lot: mainline
>    gained three ci_udc fixes and the old patch deleted the functions they
>    touch.
> - Mattijs' dwc2 endpoint fixes [3] are in mainline now, so this series
>    has no dependencies left.
> 
> Older changes are listed in v5 [1].
> 
> Cheers,
> Anders
> Anders Roxell (24):
>    dm: add helpers needed by the resynced DWC3 code
>    usb: phy: add the usb_phy helpers needed by the resynced DWC3 code
>    usb: gadget: adapt the gadget API after the resync
>    usb: gadget: udc: make udc-core build in u-boot
>    usb: dwc3: make the core driver build in u-boot
>    usb: dwc3: bring back the samsung usb phy driver
>    usb: dwc3: make gadget and ep0 build in u-boot
>    usb: gadget: udc: enable async callbacks on bind
>    usb: dwc3: ep0: flush and invalidate cache around ep0 transfers
>    usb: dwc3: make the am62 glue build in u-boot
>    usb: gadget: ci_udc: move to the udc-core registration
>    usb: gadget: ci_udc: set endpoint capabilities
>    usb: gadget: dwc2_udc_otg: fix up after the resync
>    usb: gadget: fix up the function drivers after the resync
>    usb: gadget: fix up the last UDC drivers after the resync
>    usb: gadget: atmel_usba_udc: set endpoint capabilities
>    usb: gadget: at91_udc: register with udc-core
>    usb: gadget: at91_udc: set endpoint capabilities
>    usb: musb-new: fix up after the resync
>    usb: musb-new: set endpoint capabilities
>    usb: mtu3: fix up after the resync
>    usb: mtu3: set endpoint capabilities
>    usb: host: xhci: fix up after the resync
>    usb: cdns3: fix up after the resync
> 
>   drivers/usb/cdns3/ep0.c                |    8 +-
>   drivers/usb/dwc3/Makefile              |   16 +-
>   drivers/usb/dwc3/core.c                | 1761 ++++++------------------
>   drivers/usb/dwc3/core.h                |   39 +-
>   drivers/usb/dwc3/dwc3-am62.c           |  424 +-----
>   drivers/usb/dwc3/dwc3-omap.c           |    2 -
>   drivers/usb/dwc3/ep0.c                 |  117 +-
>   drivers/usb/dwc3/gadget.c              |  719 ++--------
>   drivers/usb/dwc3/gadget.h              |    4 +-
>   drivers/usb/dwc3/io.h                  |   36 +-
>   drivers/usb/dwc3/samsung_usb_phy.c     |   77 ++
>   drivers/usb/dwc3/ti_usb_phy.c          |    2 -
>   drivers/usb/gadget/at91_udc.c          |   83 +-
>   drivers/usb/gadget/atmel_usba_udc.c    |  114 +-
>   drivers/usb/gadget/ci_udc.c            |   86 +-
>   drivers/usb/gadget/composite.c         |    2 +-
>   drivers/usb/gadget/dwc2_udc_otg.c      |   86 +-
>   drivers/usb/gadget/epautoconf.c        |    2 -
>   drivers/usb/gadget/ether.c             |   21 +-
>   drivers/usb/gadget/f_acm.c             |   20 +-
>   drivers/usb/gadget/f_fastboot.c        |   10 +-
>   drivers/usb/gadget/f_mass_storage.c    |    5 +-
>   drivers/usb/gadget/f_rockusb.c         |   11 +-
>   drivers/usb/gadget/f_sdp.c             |   12 +-
>   drivers/usb/gadget/f_thor.c            |   19 +-
>   drivers/usb/gadget/udc/Makefile        |    1 +
>   drivers/usb/gadget/udc/udc-core.c      |  933 ++-----------
>   drivers/usb/host/xhci-dwc3.c           |    4 +-
>   drivers/usb/mtu3/mtu3_gadget.c         |   10 +
>   drivers/usb/mtu3/mtu3_gadget_ep0.c     |   16 +-
>   drivers/usb/musb-new/musb_gadget.c     |   21 +-
>   drivers/usb/musb-new/musb_gadget_ep0.c |   24 +-
>   drivers/usb/musb-new/musb_uboot.c      |   39 -
>   include/dm/device_compat.h             |   13 +
>   include/dm/read.h                      |   51 +
>   include/linux/compat.h                 |   15 +
>   include/linux/usb/ch9.h                |   25 +-
>   include/linux/usb/gadget.h             |  529 ++-----
>   include/linux/usb/phy.h                |   56 +
>   39 files changed, 1406 insertions(+), 4007 deletions(-)
>   create mode 100644 drivers/usb/dwc3/samsung_usb_phy.c
> 

I tested this on AMD/Xilinx boards and didn't observe any issues. So:

Tested-by: Love Kumar <love.kumar@amd.com>

Regards,
Love Kumar