RP2xxx PIO and Reset fixes #633
Conversation
…hen adding the program
…prove api for 48 pin RP2350B variant
| }; | ||
| } | ||
|
|
||
| const ReleaseExceptionHandler = struct { |
There was a problem hiding this comment.
is this just to save on code + data size on release builds?
| for (program.instructions, program.relocations, offset..) |insn, reloc, i| | ||
| instruction_memory[i] = switch (reloc) { | ||
| .none => insn, | ||
| .jmpslot => blk: { |
There was a problem hiding this comment.
Would you mind just adding a comment here reminding us how you're doing the reloc
| .jmpslot => blk: { | ||
| const Jmp = packed struct(u16) { address: u5, reset: u11 }; | ||
| var jmp: Jmp = @bitCast(insn); | ||
| jmp.address += offset; |
There was a problem hiding this comment.
In release this won't catch if offset + address overflows. Do we have a guarantee that the offset will be in the correct range? Probably worse a comment asserting that.
| fn pin_to_index(self: EnumType, pin: gpio.Pin) error{InvalidPin}!u5 { | ||
| const index = @intFromEnum(pin); | ||
| const base = (0x10 & self.get_regs().GPIOBASE.raw); | ||
| if (index < base or index >= base + 32) { |
There was a problem hiding this comment.
Is this OK for rp2350? the 32 is the number of pins that be setup in pio, not how many pins the chip has, correct?
| } | ||
| pub fn sm_set_pin_mappings(self: EnumType, sm: StateMachine, options: PinMappingOptions) void { | ||
|
|
||
| pub fn sm_set_pin_mappings(self: EnumType, sm: StateMachine, options: PinMappingOptions) !void { |
There was a problem hiding this comment.
Is InvalidPin the only error we can get here?
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
|
|
||
| const IrqHandlerFn = *const fn () callconv(.c) void; | ||
|
|
||
| fn debugExceptionHandler(comptime name: []const u8) IrqHandlerFn { |
There was a problem hiding this comment.
Please change to debug_exception_handler, in MicroZig we use snake case for function names.
… Adds deadlines to UART.reader() and UART.writer(), adds support for high priority DMA
|
@ikskuh are you going to fix this up? |
|
This'll require a conflict resolution after #700 |
3 participants
No description provided.