Message-ID: <YNCxvElIL0RxreKe@google.com>
Date:   Mon, 21 Jun 2021 08:35:24 -0700
From:   Matthias Kaehlcke <mka@...omium.org>
To:     "Enrico Weigelt, metux IT consult" <lkml@...ux.net>
Cc:     Masahiro Yamada <masahiroy@...nel.org>,
        linux-kbuild@...r.kernel.org, linux-kernel@...r.kernel.org,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Douglas Anderson <dianders@...omium.org>,
        linux-usb@...r.kernel.org
Subject: Re: Looking for help with Kconfig dependencies

On Mon, Jun 21, 2021 at 01:26:01PM +0200, Enrico Weigelt, metux IT consult wrote:
> On 18.06.21 19:05, Matthias Kaehlcke wrote:
> 
> Hi,
> 
> 
> Cc'ing to linux-usb ...
> 
> > Patch https://lore.kernel.org/patchwork/patch/1444212/ adds the new
> > onboard_usb_hub driver which exports two functions,
> > onboard_hub_create_pdevs() and onboard_hub_destroy_pdevs(). It also
> > provides stubs for these functions which are used when the driver
> > is not selected (CONFIG_USB_ONBOARD_HUB=n).
> > 
> > The new exported functions are called by the xhci-plat driver
> > (https://lore.kernel.org/patchwork/patch/1444215/). Since xhci-plat
> > now depends on symbols from the onboard_hub_driver the following
> > dependency was added to its Kconfig entry:
> > 
> >   config USB_XHCI_PLATFORM
> >     tristate "Generic xHCI driver for a platform device"
> >     select USB_XHCI_RCAR if ARCH_RENESAS
> >  +  depends on USB_ONBOARD_HUB || !USB_ONBOARD_HUB
> 
> What exactly do you intent to archieve with this ?
> 
> X or !X = 1, isn't it ?
> 
> Why should something depend on something present or absent ?
> 
> Is that depends on ... statement necessary at all ?

I know, it's confusing, I had the same reaction when I first saw that
construct.

Effectively USB_XHCI_PLATFORM can be built without USB_ONBOARD_HUB.
However if USB_ONBOARD_HUB is built as a module then USB_XHCI_PLATFORM
should also be built as a module, which is what the above statement
achieves, unless there are conflicting dependencies.

The same construct is used for CONFIG_USB_XHCI_PCI.

> > This generally seems to work, however when USB_XHCI_PLATFORM is
> > forced to be builtin by another driver that depends on it (e.g.
> > USB_DWC3) it is still possible to build the onboard_hub driver
> > as a module, which results in unresolved symbols:
> > 
> > aarch64-linux-gnu-ld: drivers/usb/host/xhci-plat.o: in function
> > `xhci_plat_remove':
> > drivers/usb/host/xhci-plat.c:427: undefined reference to
> > `onboard_hub_destroy_pdevs'
> > drivers/usb/host/xhci-plat.c:427:(.text+0x82c): relocation truncated
> > to fit: R_AARCH64_CALL26 against undefined symbol
> > `onboard_hub_destroy_pdevs'
> > aarch64-linux-gnu-ld: drivers/usb/host/xhci-plat.o: in function
> > `xhci_plat_probe':
> > drivers/usb/host/xhci-plat.c:379: undefined reference to
> > `onboard_hub_create_pdevs'
> > drivers/usb/host/xhci-plat.c:379:(.text+0x131c): relocation truncated
> > to fit: R_AARCH64_CALL26 against undefined symbol
> > `onboard_hub_create_pdevs'
> > 
> > Kconfig generates the following warning with this configuration:
> > 
> > WARNING: unmet direct dependencies detected for USB_XHCI_PLATFORM
> >   Depends on [m]: USB_SUPPORT [=y] && USB [=y] && USB_XHCI_HCD [=y] && (USB_ONBOARD_HUB [=m] || !USB_ONBOARD_HUB [=m])
> >   Selected by [y]:
> >   - USB_DWC3 [=y] && USB_SUPPORT [=y] && (USB [=y] || USB_GADGET [=y]) && HAS_DMA [=y] && USB_XHCI_HCD [=y]
> >   Selected by [m]:
> >   - USB_CDNS_SUPPORT [=m] && USB_SUPPORT [=y] && (USB [=y] || USB_GADGET [=y]) && HAS_DMA [=y] && USB_XHCI_HCD [=y]
> >   - USB_BRCMSTB [=m] && USB_SUPPORT [=y] && USB [=y] && (ARCH_BRCMSTB [=y] && PHY_BRCM_USB [=m] || COMPILE_TEST [=y]) && USB_XHCI_HCD [=y]
> >   - USB_XHCI_MVEBU [=m] && USB_SUPPORT [=y] && USB [=y] && USB_XHCI_HCD [=y] && HAS_IOMEM [=y] && (ARCH_MVEBU [=y] || COMPILE_TEST [=y])
> 
> It seems that Kconfig is confused by trying to enforce contradicting
> dependencies.

yep, the purpose of my post was to sort that out :)

> Now for your driver:

TBH I don't think this is the right thread to discuss the driver, this
should be done on the corresponding patches.

> If I understand it correctly, you've got a topology like this:
> 
> 
> root hub -+--> 2ndary hub #0 -+--> usb-dev #0
>           |                   \--> usb-dev #1
>           |                     ..
>           \--> 2ndary hub #1 -+--> usb-dev #3
>                               \--> usb-dev #4
> 
> 
> And in order to get usb-dev #foo running, you need the corresponding
> hub on its path powered (which in turn is platform specific).
> 
> Correct ?

yep

> So, why not reflecting exactly this topology in the device tree ?
> In that case, the power management *IMHO* could pretty automatically
> (assuming you've implemented the corresponding pm functions on the
> 2ndary hub driver).
> 
> Okay, that could become a bit tricky when the usb-dev's are
> automatically enumerated on the root hub and would need to be
> reparented somehow ... @usb folks: it that possible ?

AFAIK the USB devices (including the secondary hubs) are all automatically
enumerated, the representation in the device tree is optional in the vast
majority of cases, so it's a bit of a chicken-egg problem.

> Another option could be implementing this as a regulator that the
> individual usb devices will be attached to. Not completely semantically
> correct (since a hub isn't exactly a regulator :o), but should at least
> do the job: the regulator will be switched on when the device is used
> and can be switched off when it isn't used anymore.

IMO the representation as a hub is preferable, also initialization might
be more complex than switching on a single regulator (e.g. multiple
regulators, GPIOs, clocks, ...)

> The cleanest approach, IMHO, might be adding an hub subsys, somewhat
> similar to the existing phy subsys. I can imagine similar cases with
> other interfaces, not just USB only, at least certainly not specific
> to xhci.
> 
> Or could existing phy subsys already be sufficient for that ?

I'll leave that to the USB maintainers, who seem to be happy/ok with
the current approach. There was discussion about other solutions,
including a revival of the pwrseq series
(https://lore.kernel.org/patchwork/project/lkml/list/?series=314989&state=%2A&archive=both),
which was discarded.

In any case the current solution isn't specific to xHCI. At this point
only xhci-plat is supported, however it could be extended to other
USB controllers if needed.

Powered by blists - more mailing lists