Black Pill boards - F411 USB hardware -- the Arduino_STM32 sources

It may seem like there is a lot of code in the "duino" source tree, but there really isn't. Take a look at this section extracted from the Hydra build:

cd usbF4; make
make[1]: Entering directory '/u1/Projects/STM32/Hydra/usbF4'
arm-none-eabi-gcc -mcpu=cortex-m3 -mthumb -Wno-implicit-function-declaration -fno-builtin -DHYDRA -O -I. -o usb.o -c usb.c
arm-none-eabi-gcc -mcpu=cortex-m3 -mthumb -Wno-implicit-function-declaration -fno-builtin -DHYDRA -O -I. -o misc.o -c vcp/misc.c
arm-none-eabi-gcc -mcpu=cortex-m3 -mthumb -Wno-implicit-function-declaration -fno-builtin -DHYDRA -O -I. -o usb_bsp.o -c vcp/usb_bsp.c
arm-none-eabi-gcc -mcpu=cortex-m3 -mthumb -Wno-implicit-function-declaration -fno-builtin -DHYDRA -O -I. -o usbd_cdc_vcp.o -c vcp/usbd_cdc_vcp.c
arm-none-eabi-gcc -mcpu=cortex-m3 -mthumb -Wno-implicit-function-declaration -fno-builtin -DHYDRA -O -I. -o usbd_desc.o -c vcp/usbd_desc.c
arm-none-eabi-gcc -mcpu=cortex-m3 -mthumb -Wno-implicit-function-declaration -fno-builtin -DHYDRA -O -I. -o usbd_usr.o -c vcp/usbd_usr.c
arm-none-eabi-gcc -mcpu=cortex-m3 -mthumb -Wno-implicit-function-declaration -fno-builtin -DHYDRA -O -I. -o usb_core.o -c driver/usb_core.c
arm-none-eabi-gcc -mcpu=cortex-m3 -mthumb -Wno-implicit-function-declaration -fno-builtin -DHYDRA -O -I. -o usb_dcd.o -c driver/usb_dcd.c
arm-none-eabi-gcc -mcpu=cortex-m3 -mthumb -Wno-implicit-function-declaration -fno-builtin -DHYDRA -O -I. -o usb_dcd_int.o -c driver/usb_dcd_int.c
arm-none-eabi-gcc -mcpu=cortex-m3 -mthumb -Wno-implicit-function-declaration -fno-builtin -DHYDRA -O -I. -o usbd_cdc_core.o -c library/usbd_cdc_core.c
arm-none-eabi-gcc -mcpu=cortex-m3 -mthumb -Wno-implicit-function-declaration -fno-builtin -DHYDRA -O -I. -o usbd_core.o -c library/usbd_core.c
arm-none-eabi-gcc -mcpu=cortex-m3 -mthumb -Wno-implicit-function-declaration -fno-builtin -DHYDRA -O -I. -o usbd_ioreq.o -c library/usbd_ioreq.c
arm-none-eabi-gcc -mcpu=cortex-m3 -mthumb -Wno-implicit-function-declaration -fno-builtin -DHYDRA -O -I. -o usbd_req.o -c library/usbd_req.c

HS device support

The sources include support for OTG HS. Some extra registers are present in the OTG_HS controller, but for the most part it is the same (or we would have an entirely different driver). The STM32F407 and STM32F417 would be an example with two USB OTG ports, one FS and one HS.

And I have 3 Olimex boards with F407 and F405 chips. So I should not eradicate the HS support in this source tree in case I ever want to use it with them.

I have the Olimex STM32-P405 with a F405RGT6 chip. It has 1M flash, 196K sram, two OTG usb (one FS, one HS). I see conflicting information about ethernet - there is no connector on the board. The CPU runs at 196 Mhz.
I also have the Olimex STM32-P407 and an Olimex STM32-E407. Both have the F407ZGT6 chip. Like the F405, it has 1M flash, 196K sram, two OTG usb (one FS, one HS). It runs at the same 196 Mhz clock. These boards do have a connector for the ethernet.

There is also the STM32F429. Somewhere I have a 32F429IDISCOVERY board with a STM32F429ZI chip on it. This has 2M of flash, 256K of ram along with USB OTG. 180 Mhz CPU (Cortex M4). Interestingly, the TRM here is common for the F405, F407, and F429 and they all seem to have the same Synopsys OTG controller. The section in the TRM for USB OTG is still well over 100 pages.

Pruning

• boilerplate - verbose and useless comments from some automatic code management system
• HS code - this is generally, but no always in some identifiable ifdef
• Host code - there are entire subroutines that ought to be gotten rid of.

As for the HS and Host (HCD) code, I will ensure it is in ifdefs, adding USB_HS or USB_HOST myself as necessary. I will also prefix every line with "--" inside of those conditional blocks to keep myself from tripping over them. I have found these conditionals to be important:

USE_HOST_MODE