search

adafruit / uf2-samdx1

MSC bootloader (based on UF2) for SAMD21
Other
212 stars 184 forks source link

Upstream changes and fixes #204

Open pesco opened 1 year ago

pesco commented 1 year ago

Some regular updates by the looks of it, but notably fixes two serious bugs that could alternatively be cherry-picked: b68f3a1edc36f5a645add26b2fda6e75b87e1156 and 84c32dc901aaa71ba3286b763c1620555247a1bb.

I stumbled over the latter because it makes a device using the bootloader unusable with my OpenBSD machine.

Also includes a drive-by tweak to the README to clarify that we need GNU Make to build.

pesco commented 1 year ago

Other points of note:

[EDIT by @dhalbert - fixed some commit links to point to upstream]

dhalbert commented 1 year ago
  • Upstream includes a wait state fix for SAMD51 (569c1d8) that overlaps/conflicts with 246e826 here. I chose to keep the "automatic wait states" (VMCTRL->CTRLA.bit.AUTOWS = 1) version from here, but please verify - I don't know what's correct/better.

I prefer the auto-wait-state, and was thinking of PR'ing it to upstream. I pinged the upstream PR author to ask what they prefer.

dhalbert commented 1 year ago
  • There was a conflict wrt. the PewPew M4 which got independently added here (6211469) and upstream (c08c68a1), I think? The config_data entry for DISPLAY_CFG0 ended up as 0x20180 here and 0x80 upstream. To my eye the former looks like a typo or something, but I can't be sure - please verify.

@deshipu Could you take a look at this?

deshipu commented 1 year ago
  • There was a conflict wrt. the PewPew M4 which got independently added here (6211469) and upstream (c08c68a1), I think? The config_data entry for DISPLAY_CFG0 ended up as 0x20180 here and 0x80 upstream. To my eye the former looks like a typo or something, but I can't be sure - please verify.

@deshipu Could you take a look at this?

This is because I originally prototyped with a different display than I ended up using in the boards I sold. Both values are correct, depending on which display you have, the difference is in the horizontal offset – the value 0x20180 is what was sold in the end, so I suppose this is what people will want. It can be changed with the uf2patcher tool if needed.

pesco commented 1 year ago

the value 0x20180 is what was sold in the end, so I suppose this is what people will want.

Thanks for clarifying! I've added another commit to change it back, paraphrasing your explanation in the commit message.