[RFC 0/9] Regmap over USB for Multifunction USB Device (gpio, display, ...)

[Noralf Trønnes]

Den 17.02.2020 11.32, skrev Neil Armstrong:
> Hi,
> 
> On 16/02/2020 18:21, Noralf Trønnes wrote:
>> Hi,
>>
>> A while back I had the idea to turn a Raspberry Pi Zero into a $5
>> USB to HDMI/SDTV/DSI/DPI display adapter.
>>
>> Thinking about how to represent the display to the driver I realised
>> that hardware use registers as API. And Linux does have a generic
>> register abstraction: regmap. Furthermore this means that if I can do a
>> regmap over USB implementation, it will be easy to do other functions
>> like gpio, adc and others. After a few iterations trying to understand
>> the USB subsystem and satisfying driver requirements, I now have
>> something that looks promising.
>>
>> I'm sending out an early version hoping to get feedback especially on
>> the core parts that handles regmap and interrupts.
>>
>> Overview:
>>
>>           USB Host          :         USB Device
>>                             :
>>             --------------  :  ------------------
>> ----------  | mfd: mud   |  :  | f_mud          |  ----------
>> | Driver |  --------------  :  |                |  | Driver |
>> ----------  | regmap-usb |  :  | (mud_regmap)   |  ----------
>>             --------------  :  ------------------
>>
> 
> The idea is really like ARA's greybus, but much simpler !
> Anyway nice idea, do you have good performance over USB2 and
> RPi's awful DWC2 gagdet controller ?
> 

Not as good as I was hoping for. If I disable compression I'm getting 5
fps for a 1.5MB framebuffer (7800 kB/s):

$ modetest -M mud_drm -s 35:1024x768 at RG16 -v
setting mode 1024x768-60.00Hz at RG16 on connectors 35, crtc 33
freq: 5.07Hz

When I tried reading I discovered that it was almost 3 times faster than
writing.

The zero gadget (loop testing) confirmed my findings:

Device:
$ sudo modprobe g_zero
[   44.221890] zero gadget: Gadget Zero, version: Cinco de Mayo 2008
[   44.221906] zero gadget: zero ready
[   60.751451] zero gadget: high-speed config #3: source/sink

Host:

$ sudo ~/testusb -a -t <n> -g 64 -s 16384
/dev/bus/usb/001/010 test 27,  107.230669 secs	-> 1000 / 107 =  9MB/s
/dev/bus/usb/001/010 test 28,   37.791292 secs	-> 1000 / 37  = 27MB/s
[73983.796552] usbtest 1-1.3:3.0: TEST 27: bulk write 1000Mbytes
[74205.060204] usbtest 1-1.3:3.0: TEST 28: bulk read 1000Mbytes

$ sudo ~/testusb -a -t <n> -g 64 -s 16384
/dev/bus/usb/001/010 test 5,  107.421535 secs
/dev/bus/usb/001/010 test 6,   38.189712 secs
[74893.204170] usbtest 1-1.3:3.0: TEST 5:  write 1000 sglists 64 entries
of 16384 bytes
[75012.592222] usbtest 1-1.3:3.0: TEST 6:  read 1000 sglists 64 entries
of 16384 bytes


I have tried Raspberry Pi1 and Pi4 as host (2 different controllers)
and Pi Zero and Beaglebone Black as device, getting similar result.

I found this post having the same issue:

Re: Asymmetric speed results with testusb/usbtest/g_zero
https://www.spinics.net/lists/linux-usb/msg100588.html

I haven't got a usb analyzer, but adding printk to
dwc2_assign_and_init_hc() showed that IN interrupts were 2-3 ms apart
but OUT interrupts were ~8 ms apart.

Noralf.