**** BEGIN LOGGING AT Wed Nov 04 02:59:57 2020
Nov 04 15:05:31 <kremlin> m
Nov 04 15:06:17 <Ragnorok> r
Nov 04 15:10:33 <mru> u
Nov 04 15:19:24 <Ragnorok> \o/
Nov 04 20:26:49 <CoffeeBreakfast_> I love this kind of stuff.. if (gdb) print gd->baudrate says "Missing ELF symbol gd", I just do "print ((gd_t*)$r9)->baudrate" and it works
Nov 04 20:48:33 <mm302> are you debugging host instructions, btw do you know if it's possbible to use gdb to debug a PRU step by step?
Nov 04 20:51:16 <CoffeeBreakfast_> using a jtag probe?
Nov 04 20:54:00 <Ragnorok> I think it's technically possible but non-trivial.  I wound up writing a little logger that used shared RAM as a rolling log buffer.  If something crashed it I'd look at shared ram to see where it go to.
Nov 04 20:54:06 <mm302> I see that while the PRU is disable it's possible to set some registers PRU_ICSS_PRU_DEBUG to see into a PRU
Nov 04 20:56:32 <mm302> however this only applies to when the PRU is disabled unfortunately, but there may be other tricks to run one instruction and then disable it immediately maybe
Nov 04 21:01:12 <mm302> one way is using the HALT instruction on each breakpoint, not trivial though
Nov 04 21:03:32 <mm302> actually there is also a SINGLE_STEP mode, still looks like it would require some work to set it up
**** ENDING LOGGING AT Thu Nov 05 02:59:57 2020