07:34 <benh_> somlo: allright, going to look at that .data section issue now

07:55 <benh_> somlo: so your "test" works in sim... I assume the "ROM" is really a RAM ?

07:56 <benh_> ah no I see

10:59 <somlo> benh: don't know what sim used to do before 05d4756e, but on fpga .data really was read-only

10:59 <somlo> but after 05d4756e I added a printout of the initial global .data value (before attempting to change it)

11:00 <somlo> and instead of the actual value you got 0 (both on fpga and in sim), unless you successfully initialized .data by copying it from rom

11:01 <somlo> but I'm guessing you already figured that out, this is just in case anyone else is wondering :)

11:04 <zyp> it doesn't really matter whether it's read only or writable, if you want to be able to reset the cpu without reloading the entire fpga, you need to reload .data with initial values

11:22 <benh_> somlo: nah I misread the test output

11:22 <benh_> it was broken on sim ... I sent a PR for the fix

11:23 <benh_> zyp: oh absolutely, I was wondering mostly why it worked without any fix... it's just that we don't seem to rely much on initialized data so far, we got lucky

11:24 <benh_> we were also missing a .toc in the ld script

11:24 <zyp> yeah