Re: [PATCH 1/2] ARM: dts: meson8b: odroidc1: Enable usb phy node

[Anand Moon]

Hi Martin,

On Tue, 22 Jan 2019 at 00:22, Martin Blumenstingl
<martin.blumenstingl@xxxxxxxxxxxxxx> wrote:
>
> Hi Anand,
>
> On Mon, Jan 21, 2019 at 7:37 AM Anand Moon <linux.amoon@xxxxxxxxx> wrote:
> >
> > Hi Martin,
> >
> > Thanks for you comments.
> >
> > On Mon, 21 Jan 2019 at 03:21, Martin Blumenstingl
> > <martin.blumenstingl@xxxxxxxxxxxxxx> wrote:
> > >
> > > Hi Anand,
> > >
> > > On Sun, Jan 20, 2019 at 7:44 PM Anand Moon <linux.amoon@xxxxxxxxx> wrote:
> > > >
> > > > Hi Martin,
> > > >
> > > > Thanks for your review comments.
> > > >
> > > > On Sat, 19 Jan 2019 at 03:59, Martin Blumenstingl
> > > > <martin.blumenstingl@xxxxxxxxxxxxxx> wrote:
> > > > >
> > > > > Hi Anand,
> > > > >
> > > > > thank you for this patch!
> > > > >
> > > > > On Sun, Jan 13, 2019 at 7:18 PM Anand Moon <linux.amoon@xxxxxxxxx> wrote:
> > > > > >
> > > > > > Add missing vbus-supply link to phy controller for usb_phy0
> > > > > > and usb_phy1 nodes, this changes fixed the power issue
> > > > > > on usb ports usb, changes help fix usb reset warning.
> > > > > I trust you on the fact that GPIOAO_5 controls VBUS of USB0 and USB1
> > > > >
> > > > > to me the schematics are not 100% clear or I'm bad at interpreting them:
> > > > > - I'll refer to odroid-c1+_rev0.4_20150615.pdf
> > > > > - page 1 lists GPIOAO.BIT5 as input for the PWREN signal of the
> > > > > USB_OTG controller (usb0)
> > > > > - page 1 also shows a PWREN signal in the USB_HOST controller (usb1)
> > > > > but not it's input
> > > > > - based on your patch I assume that PWREN is the same signal for both
> > > > > USB controllers
> > > > >
> > > > As per the schematics odroid-c1+_rev0.4_20150615.pdf
> > > >
> > > > It seem that their  vbus setting for *OTG* port. *Bus 01.Port 1* using
> > > > *GPIOAO.BIT5* gpio
> > > that matches my interpretation of the schematics
> >
> > Yep, I feel that GPIOAO.BIT5 enable the PWREN power on  Bus 02.Port (usb1)
> > But I could not get this to enable on Bus 01.Port (usb0).
> >
> > usb hot-plug of devices is working fine on *usb1* but not on *usb0*.
> > I have tried different combination but it's not working for me.
> >
> > I am attaching small changes to phy which enables the Bus 01.Port
> > (usb0) on Odroid C1+/Odroidc2.
> > we need to set dr_mode="otg" to enable this bus in dts.
> I need to get my multimeter out and measure on my own Odroid-C1
> I *believe* to have observed that toggling GPIOAO.BIT5 changes the
> voltage at at the OTG port, but it's been a few weeks since then...
> let me know if you want me to do the measurements on my Odroid-C1+ again
>
> > > > But we also need set power supply for *USB_B* port. *Bus 02.Port 1* ?
> > > > I don't know but I will try to find solution.
> > > I assume we also have to enable VBUS for the host only port, but I
> > > don't know how
> > >
> > > > We need to enable some bit in phy driver ?
> > > I'm not aware of any bit in the PHY driver which can toggle VBUS
> > >
> > > (trying to be creative here) maybe the GPIO is incorrectly configured
> > > to pull up or down?
> >
> > I have tried this but no luck.
> OK, too bad
>
> [...]
> > > > After fix the phy driver code and dts changes I am getting this changes.
> > > > # cat /sys/kernel/debug/regulator/regulator_summary | grep USB_VBUS
> > > >     USB_VBUS                      1    2      0 unknown  5000mV
> > > > 0mA  5000mV  5000mV
> > > >
> > > > Now I am able to enable both *Bus 01.Port 1* and *Bus 02.Port 1*
> > > >
> > > > # lsusb -t
> > > > /:  Bus 02.Port 1: Dev 1, Class=root_hub, Driver=dwc2/1p, 480M
> > > >     |__ Port 1: Dev 2, If 0, Class=Hub, Driver=hub/4p, 480M
> > > >         |__ Port 1: Dev 4, If 0, Class=Mass Storage, Driver=usb-storage, 480M
> > > >         |__ Port 3: Dev 3, If 0, Class=Video, Driver=uvcvideo, 480M
> > > >         |__ Port 3: Dev 3, If 1, Class=Video, Driver=uvcvideo, 480M
> > > >         |__ Port 3: Dev 3, If 2, Class=Audio, Driver=snd-usb-audio, 480M
> > > >         |__ Port 3: Dev 3, If 3, Class=Audio, Driver=snd-usb-audio, 480M
> > > so this is the host-only controller (usb1) and it's working fine for
> > > you - very nice!
> > >
> > > > /:  Bus 01.Port 1: Dev 1, Class=root_hub, Driver=dwc2/1p, 480M
> > > >
> > > > But not usb device are getting enable on *Bus 01.Port 1*
> > > so you have something plugged into the OTG capable controller (usb0)
> > > using an OTG cable?
> >
> > Yes I have connected USB-OTG cable with storage device attached,
> > led lights on the device is blinking, but no power to the port is enable.
> > so not hotplug of device is working on this port.
> thanks for confirming this!
> can you please send an updated version of both patches? if you want
> someone else to test and/or review the patches before they are applied
> you can use the subject prefix "RFT PATCH n/m" or "RFC PATCH n/m" (RFT
> means "request for testing", "RFC" means "request for comments")
>

I have tried to address the issue in the next series of patches.

> > > I've not tried this before on my Odroid-C1 because I'm using the micro
> > > USB connector to power the board
> >
> > Correct me if I am wrong, we need to tune USB phy driver to enable as
> > power the board ?
> so far I have seen two designs on Amlogic boards:
> - VBUS is hardwired to +5V
> - VBUS is controlled by a regulator/current limiting switch/etc. which
> can be toggled using a GPIO
>
> the PHY is involved in this because for the OTG port it senses the ID
> line to decide whether the controller has to go into host or device
> (peripheral) mode.
> however, this part is not implemented in the mainline phy-meson8b-usb2
> driver yet
>
> in case someone is interested in *why* dwc2 doesn't do OTG on the
> Amlogic SoCs yet - here are the notes from the top of my head:
> - the dwc2 controller doesn't get an interrupt when switching between
> host and peripheral mode
> - instead the Amlogic vendor kernel has a timer which polls the PHY registers
> - if the timer handler detects a "mode change" in the PHY registers it
> changes the dwc2 mode
> - however, in mainline the PHY can't notify the USB controller driver
> (dwc2) to change the mode (yet)
>
Thanks for your input.

Yep I tired to study the code changes in the u-boot from amlogic.
I tried to enable few bit in the phy to test enable of usb otg port
but no success.
I have tested that otg port is used to power on the board.

> > >
> > > do you get VBUS (maybe you have a USB thumb drive with an LED which
> > > lights up when you connect it)?
> > > do you get any errors or warnings from the dwc2 driver during boot?
> > >
> >
> > Please find the boot log of my Odroid C1+ here
> > [0]: https://pastebin.com/PiDmVmEC
> >
> > I dont see any major error on usb side.
> that looks very good indeed. as I already mentioned above: please send
> an updated version of the patches so I can try to reproduce your test
> results!
>
> > > > > on which board revision have you tested this patch?
> > > > >
> > > >
> > > > I have with me for testing
> > > >
> > > > ODROID-C0 REV 0.4 20151217
> > > > ODROID-C1+ REV 0.4 20150615 ----> current testing on
> > > I have the exact same Odroid-C1+ board revision
> > > this is my u-boot version:
> > > U-boot-00000-gb7b8dc2-dirty(odroidc@b7b8dc21) (Sep 01 2017 - 01:18:22)
> > > -> b7b8dc21 is the lastest commit of the hardkernel odroidc-v2011.03
> > > u-boot branch at the time of writing
> > >
> >
> > I am using the the u-boot from archlinux.
> > If their is another u-boot for Odroid c1 plz let me know, I will try
> > to install this on my system.
> I don't remember where I got my u-boot version from.
> maybe I should switch to the Arch Linux ARM one instead (so I'm back
> to a well-known state). but let's wait for the test results from your
> updated patches first.
>
>
> Regards
> Martin

Best Regards
-Anand