Viktor Vilhelm Sonesten

Results 170 comments of Viktor Vilhelm Sonesten

ARM DDI 0403E.d, p. C1-716: > If multi-source trace is in use, the debugger must write a non-zero value to this field. C-718: > If a system supports multiple trace...

This refactored implementation of `ITM::configure` works as expected on the `atsame51n`. I've sent an email to ARM in hopes of some clarification on TraceBusID; eventual fixes for which can be...

`lock` and `unlock` has been streamlined a bit and in the process `has_software_lock` and `locked` have been hidden. `lock` and `unlock` could potentially be hidden later when #393 and similar...

Latest changes work as expected on hardware. Also tested on the `atsame51n`.

On TraceBusID: after a brief back-and-forth with ARM it *seems* to be valid to write non-zero. If only ITM is supported on the system, TraceBusID is potentially RAZ (in ARMv8-M,...

The re-locking sequence currently incorrectly always assumes that a software lock is implemented. Amendmend will follow.

openocd warns the user if a trace port frequency cannot be obtained from a clock input: ``` (gdb) monitor tpiu config external uart on 120000000 115200 Can not obtain 115200...

I'm guessing a bit here (Button4 is bound to k*scrollup*), but I think this can be done by appending ```c { Button4, ShiftMask, externalpipe, {.v = "" } }, ```...