TychoVrahe

I believe this is now ready to review. The UI changes are caused by the different pin callback implementation, the remaining seconds message is shown earlier, which is likely more...

there is actually a problem with this approach. when the updated firmware is first ran, the bootloader is not yet updated and the transition from old bootloader to new firmware...

smooth in this case meaning that the two labels with be partially rendered over each other causing it being illegible for the whole second. i will try some bootloader update...

Actually, that specific usecase was broken already - for MCUs different than STM32F4. So here is more complex proposal: https://github.com/trezor/trezor-firmware/commit/e22248f725968131bc4d34762c2b11dd172c1dd0 Using `core/embed/models/model_XXXX.h` as a source of truth, `makefile` and combine...

the usage requires for every passed binary to specify the name of the binary, so that the layout parser knows what to look for, e.g. `-b BOOTLOADER ./build/bootloader/bootloader.bin`

rebased.

Resetting the stack in all shutdown handlers, and force-clearing the buffers seems like a reasonable thing to do. We might also want to make sure no DMA is running, as...

possibly applies also to wiping storage when changing vendor, downgrading etc.

This is the main branch i use https://github.com/trezor/trezor-firmware/tree/tychovrahe/bluetooth/master Some of the issues above i already attempted to solve, i.e.: - when USB workflow is in progress, and message is received...

- current impl in trezorctl is linux only