USB Flow Control for STM32 2

[jlpoltrack]

Title. Now can have reasonable buffers. Few other tidies as well.

[olliw42]

do we have evidence that this would work also on other SMT32 than G4?
(I assume you only have tested on G4?)
we had the F072 with usb, could be worth a test

[jlpoltrack]

(I assume you only have tested on G4?)
we had the F072 with usb, could be worth a test

Yes, I only tested on G4. Re: F0 Is this a funny joke :)? I wouldn't be able to anyways as I swapped the MCU on my FRM303 to use STM32F1...

[olliw42]

ok, so only G4
all this stm lib stuff comes actually from my STorM32 project, so I could test this - at some point in time - for F1 and H7. For now, I guess, we should add a #if def checking for G4 and throwing a warning otherwise.
It's a bit confusing otherwise since in usbd_conf.h we do have ifs for F1 and F0, and thus the lib pretends to work (be tested) for these cases

[jlpoltrack]

all this stm lib stuff comes actually from my STorM32 project, so I could test this - at some point in time - for F1 and H7. For now, I guess, we should add a #if def checking for G4 and throwing a warning otherwise.

Warning added.

[olliw42]

very nice! many thx.
you've tested that this till works as expected?

[jlpoltrack]

you've tested that this till works as expected?

Yea, I used your latest DB30 branch to flash ESP32 with these changes. Worked fine flashing at 921600 baud.

RAM usage with this branch, 39 KB:

RAM usage previously, 59 KB:

Note - you can also cut these buffers back to 2 KB as well:

https://github.com/olliw42/mLRS/blob/dev-matek-mtx/mLRS/Common/hal/matek/tx-hal-matek-mtx-db30-g474ce.h#L43-L49

That saves another 12 KB of RAM, so 28 KB used / 32 KB saved.