machine/stm32: add OTG FS USB driver for F4/F7#5404
Conversation
5e806ae to
3aaf8e5
Compare
|
Hello @digitalentity thank you very much for working on this. I would suggest breaking this into a few different PRs so that it is easier to review the changes. For example:
Very importantly, at present we do not have any code in the You should figure out how to isolate your changes so that platform code does not end up in that package or sub-packages. Probably after you split off the non-platform specific changes, this will be a lot easier to achieve. |
|
Hey, thanks for the review.
The HID and MSC descriptor remap stems from the fact that STM32 has only 4 EPs and the TinyGo system currently assumes 8. I see three possible options:
I suggest we do the (2). This comes with some risk of breaking existing applications that use USB though. |
Perhaps some of the serious USB people like @sago35 @mikesmitty and others of course can offer their feedback on this question. If option number 2 is practical on all platforms that sounds like the best choice, but I have not looked into that question yet. |
|
This I think |
|
On Option 2 - EPs are bidirectional, nothing is preventing us from doing mix & match - with 4 EPs (effectively 8 unidirectional streams) we can fit HID (2), MSC (2) and CDC (3) into the same device in principle. |
|
Hello @digitalentity any chance of getting this PR back in action now that the endpoint changes have been merged? Please and thank you! 😸 |
|
Hey @deadprogram, yes, I plan to get back to this, been too busy with other things lately 🧑🏭 |
1d9d344 to
7b5155e
Compare
|
So, this is a rebase which builds, but is completely untested. I still need to address the flash support. |
7b5155e to
f329e11
Compare
|
@digitalentity please see formatting error: |
|
Weird, I definitely ran |
|
Not sure but perhaps you also need to rebase the latest |
This is rebased onto the current |
|
Other Windows runner jobs are passing, so ? |
The windows builds for PR #5511 also fail. Definitely a runner or cached llvm issue. |
I saw that a few minutes ago with another build and just tried deleting some caches. GH fun and games! |
fc909b3 to
e2f6ed0
Compare
|
All 3 common examples work:
|
|
@deadprogram should I factor out the usb-storage example changes (use tiny RAM disk instead of flash) or is this good as-is? |
I think the RAM disk change is probably fine. Now that the CI has been fixed using #5512 can you please rebase just once more so the tests can run? Thanks and sorry about the inconvenience. |
STM32F4 and F7 share the same OTG FS IP but have no USB driver in TinyGo. This adds a full device-mode driver covering CDC, HID, and MSC with working examples. - OTG FS has 4 physical EPs (0–3); virtual indices 4–7 fold onto them via physEP() so the existing machine/usb API is unchanged - F4 bypasses VBUS sensing via GCCFG.NOVBUSSENS; F7 uses the USB voltage regulator + GOTGCTL B-valid override instead - STM32F7 PLL_Q changed 2→9 to produce the 48 MHz clock required by USB/RNG/SDMMC; CK48MSEL cleared to select main PLL as source - HID and MSC descriptors remapped at init() to physical endpoint addresses (EP2/EP1 for HID, EP2/EP3 for MSC) - usb-storage example replaced machine.Flash with a FAT12 RAM disk so the host mounts without reformatting - MSC sendCSW sets queuedBytes before state transition to fix a missed byte-count on the status phase
e2f6ed0 to
6e56f02
Compare
|
@deadprogram done, CI/CD tests passing |
|
Thank you very much for all your work on this @digitalentity now merging! |
|
great work! Thank You. |
3 participants
STM32F4 and F7 share the same OTG FS IP but have no USB driver in TinyGo. This adds a full device-mode driver covering CDC, HID, and MSC with working examples.
Tested on CDC, HID and MSC examples.
CDC example, echo verified to work:
HID example:
MSC example, RAM disk verified to handle write/read correctly:
Fixes #4751