On 12/14/2018 2:26 PM, Andrzej Pietrasiewicz wrote:
> W dniu 14.12.2018 o 09:55, Minas Harutyunyan pisze:
>> Hi Andrzej,
>>
>> On 12/4/2018 9:54 PM, Andrzej Pietrasiewicz wrote:
>>> Hi All,
>>>
>>> I have recently noticed a problem with dwc2 and audio gadget.
>>>
>>> TICK_TIME: 0
>>>
>>> Any ideas on how to test it further to identify the problem?
>>>
>>> Definitely the commit identified with bisect was a problem at least
>>> until 4.13. Perhaps it still is the root cause, but I can't tell.
>>> It is difficult to revert it, because there were later changes which
>>> depend on it.
>>>
>>> Andrzej
>>>
>>
>> Which DMA (BDMA or DDMA) or Slave mode in dwc2 you use for testing?
>>
>
> Currently the default available in upstream dwc2, which is, I guess, BDMA.
Currently gadget default mode is DDMA. DDMA support by default set
starting from 4.10, of course, if core supported DDMA.
>
>
>> By changing UAC2_DEF_REQ_NUM to 4 all audio gadget tests passed smoothly.
>> Could you please apply this patch and run tests in DDMA mode:
>>
>> diff --git a/drivers/usb/gadget/function/u_uac2.h
>> b/drivers/usb/gadget/function/u_uac2.h
>> index 8362ee572e1e..5e649259ab76 100644
>> --- a/drivers/usb/gadget/function/u_uac2.h
>> +++ b/drivers/usb/gadget/function/u_uac2.h
>> @@ -21,7 +21,7 @@
>> #define UAC2_DEF_CCHMASK 0x3
>> #define UAC2_DEF_CSRATE 64000
>> #define UAC2_DEF_CSSIZE 2
>> -#define UAC2_DEF_REQ_NUM 2
>> +#define UAC2_DEF_REQ_NUM 4
>>
>> struct f_uac2_opts {
>> struct usb_function_instance func_inst;
>>
> This patch applied to current mainline seems to help in case the board
> boots correctly - no problems playing from the host using the USB gadget -
> but it still sometimes hangs at boot. I tried tracing the problems
> back and identified 381fc8f8228923026b3d75c8230fa2ee4d688f32 at 4.8 as the first
It's for BDMA and that commit have a bug which fixed by
d5d5f079138254e4d641de2f36760d0d0fc9f510. Please apply and test.
DDMA mode for ISOCs support in gadget added starting from 4.10
> commit at which they happen. Applying the above patch is not possible at 4.8,
4.8 should be BDMA
> because at that point there was no such #define. It does apply though at 4.11,
> but does not help, so I it seems there are some more co-existing problems.
DDMA ISOC support for gadget cardinally changed starting from 4.18 by
follow commits: 729cac693eecfebdb9e152eaddddd358ae2decb7,
37981e00503f6bdeb99108ff90199002c3eec2f5,1ffba9058737af2ddeebc813faa8ea9b16bc892a,
21cbbc6bb7b4ba6aee303a3b4aef1578253724d8,1d8e5c00275825fc42aaa5597dab1d0b5b26bb64,
4d4f1e79b800c80313280eec1af25a242a50dce5
Finally fully fixed from 4.19
Thanks,
Minas
>
> The board is an Odroid U3 (Exynos 4412).
>
> Andrzej
>
