14:42 <tnt> gatecat: mmm ... there seem to be a lot of atomic ops before it even setups the kernel traps.

14:56 <gatecat> tnt: yeah, this was on top of antmicros heavily patched tree, I suspect those may have been patched out too

14:56 <gatecat> tbh emulating them in M-mode is a better option, but I don't have a good reference for that

14:57 <gatecat> someone else here might

14:58 <tnt> So the kernel by default runs in S-mode ?

15:03 <gatecat> yeah

15:32 <david-sawatzke[m> It's probably something to do with https://github.com/enjoy-digital/litex/pull/965, but I'm not sure

19:50 <tnt> mmm ... does the kernel except to be able to do misaligned loads ? I get a "load address misaligned" exception.

19:51 <gatecat> I think it expects those to be emulated in M-mode too

19:51 <tnt> But how does linux-on-litex handles that ?

19:52 <gatecat> it used to have its own emulator, now it uses opensbi by the looks of things

19:53 <tnt> was that done in the bios ?

19:53 <gatecat> no, it goes bios->opensbi (which remains resident in m-mode)->linux

19:54 <gatecat> the emulator used to be in the place of opensbi

19:54 <tnt> right, I mean before it used opensbi.

19:56 <tnt> I'm trying to find related code in the litex git history

19:56 <gatecat> https://github.com/litex-hub/linux-on-litex-vexriscv/commit/382fe2ac67b2bb2fbc738d32b2054f95f1422780

19:56 <gatecat> was the commit that removed it and switched to opensbi

19:59 <tnt> Thanks