[U-Boot] [PATCH v3 3/3] arm: odroid: usb: add support for usb host including ethernet

[Przemyslaw Marczak]

Hello Suriyan,

On 11/18/2014 06:34 PM, Suriyan Ramasami wrote:
> Hello Przemyslaw,
>
> On Tue, Nov 18, 2014 at 3:09 AM, Przemyslaw Marczak
> <p.marczak at samsung.com> wrote:
>> Hello Suriyan,
>>
>> On 11/18/2014 04:30 AM, Suriyan Ramasami wrote:
>>>
>>> Hello Minkyu Kang/Przemyslaw and of course anyone else who can direct me,
>>>
>>> Thanks for the commit.
>>>
>>>       I just realized that the code in this particular patch is missing
>>> the gpio_request() call before calling the gpio_direction_output()
>>> call, and hence fails to initialize the USB3503A hub. This was pointed
>>> out by Przemyslaw in another patch of mine (odroid: blue LED)
>>>
>>>      So, my question is, this patch is in u-boot-samsung. To correct the
>>> above mentioned error, do I submit a patch directed to u-boot-samsung?
>>> If so how do I do it? I mean, do I elaborate this in the subject
>>> /.body of the mail?
>>>
>>> Thanks and regards,
>>> - Suriyan
>>>
>>>
>>
>> Yes, the gpio_request() was missed in your code - it prints errors about it
>> but the HUB works fine. So when you send the fixed version this should be
>> okay. Please check ums command before send.
>>
>
> Sure, will do!
>
>> Please also check setting buck value at function board_usb_init(),
>> the buck is disabled and before enable - the value changes two times -
>> probably by mistake.
>>
>
> I think there is an error in the implementation of
> max77686_set_buck_mode(..., buck #, OPMODE_OFF), or I have some gap in
> my understanding of it. I believe that when you set OPMODE_OFF it
> should turn it OFF right? as in no voltage? This does not seem to be
> the case.
> I have to turn the power off and then on, on BUCK8 (LAN9730 feeds off
> of it), to reset the LAN when a usb reset is issued, else it will not
> be detected again when usb start is re-issued.
> Initially, I had max77686_set_buck_mode(p_pmic, 9, OPMODE_OFF); delay;
> max77686(set_buck_mode(p_pmic, 9, OPMODE_ON); but it did not power
> recycle the LAN9730 chip.
> I do not have access to the MAX77686 data sheets to debug this. Hence,
> in my current code I set the voltage to the lowest possible value, and
> then bring it back up to 3.3V to simulate a reset. In fact, I can
> remove the OPMODE_OFF and OPMODE_ON code, and it works.
> Please let me know if there is any change needed in the OPMODE_OFF
> code for it to work the way it should, or correct my understanding of
> what it is supposed to do.
>
> Currently, I am thinking of just having:
>                  max77686_set_buck_voltage(p_pmic, 8, 750000);
>                  max77686_set_buck_voltage(p_pmic, 8, 3300000);
> and leave the OPMODE_OFF and OPMODE_ON out of it.
>
Ok, now I see the issue. So I looked into the documentation and there is 
a "ENB8" pin in PMIC package - this pin allows steering BUCK8 ON/OFF by 
the hardware. If ENB8 is set to "low", then you can do on/off - when 
"high" then you can't change it's state by I2C write. This PMIC has a 
lot of such gpio for setting the default power-on state of ldo/buck 
regulators.

The signal of ENB8 in OdroidU3 schematic is missed, I can suppose that 
it is connected to one of SOC RST signal as it is on Odroid X2.

So doing the reset by setting low/high voltage value - seems to be good 
in this case.

>> And the gpio_request(...) you can put into board_gpio_init - request gpio
>> only one.
>>
>
> Yes.
>
>> The odroid is available on the market, so I would prefer keep some
>> documentation in the doc/README.odroid.
>> Could you please add short section about the USB usage in U-Boot? This would
>> be useful.
>
> OK!
>
>>
>>
>>> On Mon, Nov 17, 2014 at 6:02 AM, Minkyu Kang <mk7.kang at samsung.com> wrote:
>>>>
>>>> On 30/10/14 01:22, Suriyan Ramasami wrote:
>>>>>
>>>>> This change adds support for enabling the USB host features of the
>>>>> board.
>>>>> This includes the USB3503A hub and the SMC LAN9730 ethernet controller
>>>>> as well.
>>>>>
>>>>> Signed-off-by: Suriyan Ramasami <suriyan.r at gmail.com>
>>>>>
>>>>> ---
>>>>>
>>>>> Changes in v3:
>>>>> * removed set_usb_ethaddr() and related code as the GUID registers do
>>>>> not
>>>>>     seem to be documented anywhere. This is sad, as this mechanism allows
>>>>>     for each Odroid to boot up with the same MAC address every time, but
>>>>> no
>>>>>     two odroids shall have the same MAC address on boot. This ensures
>>>>> multiple
>>>>>     odroids in the same LAN to come up without conflicting MAC addresses.
>>>>> * Minkyu - Do not mix cpu_is... and proid_is...
>>>>>
>>>>> Changes in v2:
>>>>> * Jaehoon - Split power.[ch] as a separate patch
>>>>> * Removed an unneeded header file from ehci-exynos.c
>>>>> * Jaehoon - Fix indentation in the dts file
>>>>>
>>>>> Changes in v1:
>>>>> * First try
>>>>>
>>>>>    arch/arm/dts/exynos4412-odroid.dts      | 11 +++++++
>>>>>    arch/arm/include/asm/arch-exynos/ehci.h | 13 ++++++++
>>>>>    board/samsung/odroid/odroid.c           | 32 +++++++++++++++++++
>>>>>    drivers/usb/host/ehci-exynos.c          | 55
>>>>> ++++++++++++++++++++++++++++-----
>>>>>    include/configs/odroid.h                | 13 ++++++++
>>>>>    5 files changed, 116 insertions(+), 8 deletions(-)
>>>>>
>>>>
>>>> applied to u-boot-samsung.
>>>>
>>>> Thanks,
>>>> Minkyu Kang.
>>>>
>>>
>>
>> Best regards,
>> --
>> Przemyslaw Marczak
>> Samsung R&D Institute Poland
>> Samsung Electronics
>> p.marczak at samsung.com

Best regards,
-- 
Przemyslaw Marczak
Samsung R&D Institute Poland
Samsung Electronics
p.marczak at samsung.com