[RFC PATCH v2 30/64] usb: dwc3: import from kernel v5.6

[Jens Wiklander]

Hi Simon,

On Fri, May 8, 2026 at 5:36 PM Simon Glass <sjg at chromium.org> wrote:
>
> Hi Jens,
>
> On 2026-05-07T09:27:07, Jens Wiklander <jens.wiklander at linaro.org> wrote:
> > usb: dwc3: import from kernel v5.6
> >
> > Sync Linux kernel dwc3 changes from v5.5 to v5.6.
> >
> > The following files are preserved accross the import:
> > Makefile Kconfig dwc3-meson-g12a.c dwc3-meson-gxl.c dwc3-omap.c
> > dwc3-uniphier.c dwc3-generic.h dwc3-generic.c dwc3-generic-sti.c
> > dwc3-layerscape.c ti_usb_phy.c
> >
> > Skipping unused files:
> > debugfs.c drd.c dwc3-exynos.c dwc3-haps.c dwc3-imx8mp.c dwc3-keystone.c
> > dwc3-octeon.c dwc3-of-simple.c dwc3-pci.c dwc3-qcom.c dwc3-qcom-legacy.c
> > dwc3-rtk.c dwc3-st.c dwc3-xilinx.c host.c trace.c trace.h ulpi.c
> >
> > Note that this is a raw import and doesn't build.
> > A fixup commit at the end of the series fixes that.
> >
> > List of commits: git log --oneline v5.5..v5.6
> > Commits imported:
> > dad2aff3e827 usb: dwc3: gadget: Update chain bit correctly when using sg list
> > [...]
> >
> > drivers/usb/dwc3/core.c   |  3 +++
> >  drivers/usb/dwc3/core.h   |  2 ++
> >  drivers/usb/dwc3/debug.h  | 39 +++++++++++++---------------------
> >  drivers/usb/dwc3/ep0.c    |  4 +++-
> >  drivers/usb/dwc3/gadget.c | 54 ++++++++++++++++++++++++++++++++---------------
> >  drivers/usb/dwc3/gadget.h | 14 ++++++++++++
> >  6 files changed, 74 insertions(+), 42 deletions(-)
>
> > diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
> > @@ -167,6 +167,9 @@ static void dwc3_ep_inc_enq(struct dwc3_ep *dep)
> >  static void dwc3_ep_inc_deq(struct dwc3_ep *dep)
> >  {
> >       dwc3_ep_inc_trb(&dep->trb_dequeue);
> > +     /* set no action before sending new link state change */
> > +     dwc3_writel(dwc->regs, DWC3_DCTL, reg);
> > +
> >  }
>
> This hunk is in the wrong function - it should be dwc3_gadget_set_link_state()
>
> I'm wondering whether this is intentional, i.e. that you are not too
> worried about the accuracy of these individual commits and are mostly
> just concerned with the final result.

That's right. I'm copying the end result of each kernel version. The
commits listed in the commit message have not been treated
individually. I'm listing them in the commit message because I think
it might help in some cases when searching for something later.

> That seems OK to me, since this
> is just such a large effort, so I will skip past these.

Thanks.

Cheers,
Jens

>
> Regards,
> Simon