Can someone try compiling bootloader and firmware .bin for 401CC?

[Ganesharyal24894]

So I am trying to find the problem as to why my STM32F401CC is not able to be converted as BMP(discussed here issues #1899 ). Initially, I was thinking there could be some problem with the hardware, so I ordered another board and received it today, but even that Blackpill board does not work. So now I am sure that it is not a hardware problem. Can someone try to compile the bootloader and firmware .bin for Blackpill (401CC) and send it here? I just want to verify that it is a problem with compilation on my system.

[ALTracer]

Compiled from current main, v1.10.0-1031-g3a3558d7, and flashed to WeActStudio MiniSTM32F411CE blackpill-f411ce with 25MHz HSE crystal, confirmed booting/enumerating/answering to GDB and BMDA. All targets enabled, RTT enabled. No diagnostic logging or other fancy stuff.

OEM ST bootloader:
$ meson setup build-f401cc_maskrom --cross-file cross-file/blackpill-f401cc.ini -Dbmd_bootloader=false
build-f401cc_maskrom.zip
$ dfu-util -d 0483:df11 -S 375F334C3032 --alt 0 -s 0x08000000:leave -D build-f401cc_maskrom/blackmagic_blackpill_f401cc_firmware.bin

BMD bootloader:
$ meson setup build-f401cc --cross-file cross-file/blackpill-f401cc.ini
build-f401cc.zip
$ dfu-util -d 0483:df11 -S 375F334C3032 --alt 0 -s 0x08000000:leave -D build-f401cc/blackmagic_blackpill_f401cc_bootloader.bin
$ dfu-util -d 1d50:6017 -S 375F334C3032 --alt 0 -s 0x08004000:leave -D build-f401cc/blackmagic_blackpill_f401cc_firmware.bin

In case of troubles, use ELF files for inception debugging your blackpills via an stlink.

[dragonmux]

The bug in question only triggers if a USB Reset occurs under certain conditions - which are possible under Linux, but just not guaranteed to occur. This is why we say we suspect this has just been broken since that PR.

[Ganesharyal24894]

Got it now. Once again, really Thank you.

[ALTracer]

Since you've discovered and fixed a bug of extraneous USBOTG_FS reset in F4 drivers, should I delete my bug-exhibiting tested builds which worked on Linux for me but break on Windows for @Ganesharyal24894 ? Or upload new ones, keeping these for reproducibility with a real USB analyzer? Or CI builds and nightly links are enough?
To be clear, I was seeing a new kernel warning then, Device not responding to setup address., and then it enumerated the second time successfully. Only when booting normally through BMD bootloader, but not when leaving from MaskROM into BMD bootloader or booting with PA0 pressed down. It's gone now.
I would prefer the new developments to be its own accepted answer, but whatever.

[Ganesharyal24894]

@ALTracer I have unmarked that as an answer, just in case anyone comes across that.

[dragonmux]

We'll close the discussion as resolved for now in that you've been able to prove it was not your compilation environment, but rather a bug that's now been fixed in libopencm3. It can be re-opened later if necessary.