Skip to content

Latest commit

 

History

History
109 lines (63 loc) · 3.77 KB

wishlist.md

File metadata and controls

109 lines (63 loc) · 3.77 KB

Wish list for dw-link

List of tasks to work on:

  • produce short youtube video to promote dw-link probe

  • use "blinkmodes.ino" in order to highlight embedded programming, perhaps in a video?

  • Implement STK500v2 protocol

  • debug tiny13 problem, perhaps by reverting back to the version in 2022, when it seemed to work

  • maybe have a variable length break, dependent on the speed of ISP?

  • implement semi-hosting

  • design case for debugger and print it

  • perhaps make conditional/repeating breakpoints faster: less register saving/restoring (would give you perhaps 10 ms out of 40 ms), shorter pauses by GDB (but where to control this?)

List of tasks done:

  • Write a step-by-step explanation on how to restore an UNO after debugging

  • try integration of tool into IDE 2.0 again

  • clean up Gede interface

  • no multiple main breaks oints

  • reloading files in GUI when reloading files

  • write a short debug description for PIO

  • write/design setup for UNO debugging with connected USB cable

  • rewrite compile-time conditionals so that they do not show up as errors in PIO

  • use DEBTX pin (if TXODEBUG is not defined) as a sensing pin to disable automatic DWEN programming, i.e., you have to use mo dw +/- by yourself.

  • optimize ispTransfer so that higher ISP rates are possible (currently, 50 kHz is max)

  • important (3.1.0): When trying to debug the boarduino at 5V with the probe shield, the feedback is: Wrong Wiring. It works flawlessly with a lot of other boards (ATtiny, UNO, ...), with other shields, with 5V, you name it --> turned out to be that the boarduino had a particularly strong pull-down resistor. That is, the cure is to use 1K or stringer pull-ups on the SPI lines!

  • monitor clock command that displays the current clock setting -- well, redesigned all monitor commands!

  • integrate programmer into code so that one also could do ISP programming

  • design new adapter board

  • test prototype of such a board

  • Adding response in dw-link to one special byte in order to be able to identify the debug serial port

  • writing Python function for identifying debug port

  • Erasing MCU automatically when lock bits are set

  • write hook to set UPLOAD_PORT and DEBUG_PORT in extra_script.

  • Explaining how to recover UNO (soldering + Bootloader burning) in Manual

  • remove part about lock bits in Manual

  • remove automatic speed recognition

  • delete the fancy control stuff in dw-link

  • rewrite part in manual about adapter boards

  • rewrite part in Manual about PlatformIO

List of tasks abandoned:

  • try to use seer GUI again

  • integrate into the Arduino VSC plugin

  • a mode where reads & writes are double-checked: "monitor verify/noverify". Could be on or off by default.

  • reorganize BP management: have a list of stored and a list of new BPs, which would save us 3 bytes per BP, i.e., we could easily go from 25 to 30 BPs -- but be careful!

New control regime for entering/leaving DW

AutoDW: Entering and leaving DW mode is something the user can be ignorant about. Works only if the target system is powered by the debugger.

Non-AutoDW: At some points, the user is asked to power-cycle the system. After debugging, the user has to restore the original state.

Possible states are: Normal, Transitional, DebugWIRE

  • target remote ... gdbStartConnect

    • AutoDW: N|T|D -> D
    • Non-AutoDW: D -> D, N|T -> T

  • quit gdbStopConnect

    • AutoDW: D|T|N -> N
    • Non-AutoDW: D -> D, T|N -> N

  • monitor dwire + gdbConnectDW

    • AutoDW: N|T|D -> D
    • Non-AutoDW: N|T -> PC -> D, D -> D

  • monitor dwire - gdbDisconnectDW

    • AutoDW: D|T|N -> N
    • Non-AutoDW: D -> T, T -> T, N -> N

In the end, I noticed that quit did not work as expected, because sometimes the quit is called more than once! So, I added extra control-variables to take of that.