sunxi: Endless musb_bus_suspend warnings when booting A23 tablet with OTG

sunxi: Endless musb_bus_suspend warnings when booting A23 tablet with OTG

[wens Tsai]

Hi Hans,

I've built the latest sunxi-next branches of Linux and U-boot and
tried it on my Q8 A23 tablet. My tablet has an OTG adapter with a
USB ethernet adapter connected all the time. Upon booting into
Linux I get an endless stream of:

    musb_bus_suspend 2586: trying to suspend as a_wait_bcon while active

messages. To get out of this I have to physically remove the OTG
adapter. Booting without the OTG adapter to begin with also gets
rid of the problem.

Have you run into this problem? I wonder if we should disable VBUS
just before leaving U-boot? Now that we support DRIVEVBUS this
shouldn't be a problem.

Regards
ChenYu

[linux-sunxi] sunxi: Endless musb_bus_suspend warnings when booting A23 tablet with OTG

[Hans de Goede]

Hi,

On 06-09-16 13:02, wens Tsai wrote:
> Hi Hans,
>
> I've built the latest sunxi-next branches of Linux and U-boot and
> tried it on my Q8 A23 tablet. My tablet has an OTG adapter with a
> USB ethernet adapter connected all the time. Upon booting into
> Linux I get an endless stream of:
>
>     musb_bus_suspend 2586: trying to suspend as a_wait_bcon while active
>
> messages. To get out of this I have to physically remove the OTG
> adapter. Booting without the OTG adapter to begin with also gets
> rid of the problem.
>
> Have you run into this problem? I wonder if we should disable VBUS
> just before leaving U-boot? Now that we support DRIVEVBUS this
> shouldn't be a problem.

I've not seen such a problem. I'm fine with disabling Vbus when
leaving u-boot, that certainly is the safe thing to do anyways.

I guess I'll get a u-boot patch from you if that indeed fixes
things ?

Regards,

Hans

[linux-sunxi] sunxi: Endless musb_bus_suspend warnings when booting A23 tablet with OTG

[wens Tsai]

On Tue, Sep 6, 2016 at 7:28 PM, Hans de Goede <hdegoede@redhat.com> wrote:
> Hi,
>
> On 06-09-16 13:02, wens Tsai wrote:
>>
>> Hi Hans,
>>
>> I've built the latest sunxi-next branches of Linux and U-boot and
>> tried it on my Q8 A23 tablet. My tablet has an OTG adapter with a
>> USB ethernet adapter connected all the time. Upon booting into
>> Linux I get an endless stream of:
>>
>>     musb_bus_suspend 2586: trying to suspend as a_wait_bcon while active
>>
>> messages. To get out of this I have to physically remove the OTG
>> adapter. Booting without the OTG adapter to begin with also gets
>> rid of the problem.
>>
>> Have you run into this problem? I wonder if we should disable VBUS
>> just before leaving U-boot? Now that we support DRIVEVBUS this
>> shouldn't be a problem.
>
>
> I've not seen such a problem. I'm fine with disabling Vbus when
> leaving u-boot, that certainly is the safe thing to do anyways.

Seems there's no distinction between "USB reset" and leaving
U-boot, as you mentioned in your patch removing the "power off"
code from the USB drivers. Are you OK with power cycling during
USB reset as well?

> I guess I'll get a u-boot patch from you if that indeed fixes
> things ?

Yup. In progress. Though my ethernet adapter is acting up, so I'll
need to swap in something else to test it properly.

ChenYu

[linux-sunxi] sunxi: Endless musb_bus_suspend warnings when booting A23 tablet with OTG

[Hans de Goede]

Hi,

On 06-09-16 17:31, wens Tsai wrote:
> On Tue, Sep 6, 2016 at 7:28 PM, Hans de Goede <hdegoede@redhat.com> wrote:
>> Hi,
>>
>> On 06-09-16 13:02, wens Tsai wrote:
>>>
>>> Hi Hans,
>>>
>>> I've built the latest sunxi-next branches of Linux and U-boot and
>>> tried it on my Q8 A23 tablet. My tablet has an OTG adapter with a
>>> USB ethernet adapter connected all the time. Upon booting into
>>> Linux I get an endless stream of:
>>>
>>>     musb_bus_suspend 2586: trying to suspend as a_wait_bcon while active
>>>
>>> messages. To get out of this I have to physically remove the OTG
>>> adapter. Booting without the OTG adapter to begin with also gets
>>> rid of the problem.
>>>
>>> Have you run into this problem? I wonder if we should disable VBUS
>>> just before leaving U-boot? Now that we support DRIVEVBUS this
>>> shouldn't be a problem.
>>
>>
>> I've not seen such a problem. I'm fine with disabling Vbus when
>> leaving u-boot, that certainly is the safe thing to do anyways.
>
> Seems there's no distinction between "USB reset" and leaving
> U-boot, as you mentioned in your patch removing the "power off"
> code from the USB drivers. Are you OK with power cycling during
> USB reset as well?

I would prefer not too, but if it is hard to get around that
I can live with it.

>
>> I guess I'll get a u-boot patch from you if that indeed fixes
>> things ?
>
> Yup. In progress. Though my ethernet adapter is acting up, so I'll
> need to swap in something else to test it properly.

Cool.

Thanks & Regards,

Hans

[linux-sunxi] sunxi: Endless musb_bus_suspend warnings when booting A23 tablet with OTG

[wens Tsai]

On Wed, Sep 7, 2016 at 6:44 PM, Hans de Goede <hdegoede@redhat.com> wrote:
> Hi,
>
> On 06-09-16 17:31, wens Tsai wrote:
>>
>> On Tue, Sep 6, 2016 at 7:28 PM, Hans de Goede <hdegoede@redhat.com> wrote:
>>>
>>> Hi,
>>>
>>> On 06-09-16 13:02, wens Tsai wrote:
>>>>
>>>>
>>>> Hi Hans,
>>>>
>>>> I've built the latest sunxi-next branches of Linux and U-boot and
>>>> tried it on my Q8 A23 tablet. My tablet has an OTG adapter with a
>>>> USB ethernet adapter connected all the time. Upon booting into
>>>> Linux I get an endless stream of:
>>>>
>>>>     musb_bus_suspend 2586: trying to suspend as a_wait_bcon while active
>>>>
>>>> messages. To get out of this I have to physically remove the OTG
>>>> adapter. Booting without the OTG adapter to begin with also gets
>>>> rid of the problem.
>>>>
>>>> Have you run into this problem? I wonder if we should disable VBUS
>>>> just before leaving U-boot? Now that we support DRIVEVBUS this
>>>> shouldn't be a problem.
>>>
>>>
>>>
>>> I've not seen such a problem. I'm fine with disabling Vbus when
>>> leaving u-boot, that certainly is the safe thing to do anyways.
>>
>>
>> Seems there's no distinction between "USB reset" and leaving
>> U-boot, as you mentioned in your patch removing the "power off"
>> code from the USB drivers. Are you OK with power cycling during
>> USB reset as well?
>
>
> I would prefer not too, but if it is hard to get around that
> I can live with it.

Unfortunately musb USB reset does not work. I tested without my
patch (the one I sent out today). So I think this point is
irrelevant, at least until someone fixes musb.

ChenYu

>>
>>> I guess I'll get a u-boot patch from you if that indeed fixes
>>> things ?
>>
>>
>> Yup. In progress. Though my ethernet adapter is acting up, so I'll
>> need to swap in something else to test it properly.
>
>
> Cool.
>
> Thanks & Regards,
>
> Hans