DWC OTG TX path optimisation for 11-current

[Hans Petter Selasky]

On 08/18/15 15:34, Svatopluk Kraus wrote:
> On Tue, Aug 18, 2015 at 3:23 PM, Hans Petter Selasky <hps at selasky.org> wrote:
>> On 08/18/15 15:12, Svatopluk Kraus wrote:
>>>
>>> On Tue, Aug 18, 2015 at 2:46 PM, Hans Petter Selasky <hps at selasky.org>
>>> wrote:
>>>>
>>>> On 08/18/15 14:31, Svatopluk Kraus wrote:
>>>>>
>>>>>
>>>>> On Tue, Aug 18, 2015 at 1:54 PM, Hans Petter Selasky <hps at selasky.org>
>>>>> wrote:
>>>>>>
>>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> On 08/18/15 13:22, Svatopluk Kraus wrote:
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> Do I understand it well that you are not able to trigger the problem?
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> Maybe I haven't been trying hard enough. I've been using a kernel built
>>>>>> from
>>>>>> "projects/hps_head" which I just updated to include the latest DWC OTG
>>>>>> patches, instead of "head". There was today a timer subsystem fix, not
>>>>>> sure
>>>>>> if it makes any difference:
>>>>>>
>>>>>> https://svnweb.freebsd.org/changeset/base/286880
>>>>>>
>>>>>> I will try to buildworld from an external USB disk later today and see
>>>>>> what
>>>>>> happens.
>>>>>>
>>>>>>>
>>>>>>> and still the system response time is terrible.
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> If you can do:
>>>>>>
>>>>>> usbconfig -d 0.1 set_config 255
>>>>>>
>>>>>> And verify that the IRQ rate goes to zero for DWC OTG, then we maybe
>>>>>> could
>>>>>> rule out USB.
>>>>>>
>>>>>
>>>>> Note that my root filesystem was on usb disk.
>>>>>
>>>>
>>>> Is the console still not responding properly after this command?
>>>>
>>>
>>> System response time is still bad.
>>>
>>
>> Is "kgdb" working on your setup? Can you dump all kernel threads?
>
> db> ps
>    pid  ppid  pgrp   uid   state   wmesg     wchan    cmd

>

Not "ps", but "alltrace" I think.

--HPS