IRC channel logs

<wildwestrom>stikonas: So yeah, that's a problem. I don't know the extent to which it's hardware specific other than the UART IO addresses.

<janneke>mwette: thanks, will do!

<wildwestrom>I hear the hardware incompatibility is a big constraint with RISC-V in the context of bootstrapping. For now I'm literally just using QEMU. I wish I could answer your question about SRAM. I don't really know the details yet.

<wildwestrom>How big of a project is it to abstract RISC-V hardware?

<matrix_bridge><Andrius Štikonas> wildwestrom: well, an intermediate step would be to use UEFI

<matrix_bridge><Andrius Štikonas> And assume that e.g. U-boot is installed

<matrix_bridge><Andrius Štikonas> I played with UEFI on amd64: https://codeberg.org/stikonas/stage0-uefi

<matrix_bridge><Andrius Štikonas> But yes, it's not fully baremetal

<aggi>reminds me, i've been complaining x86 16bit real-mode boot is gradually phased out by UEFI systems

<aggi>can the complete bootstrapping chain be spawn from stage0-uefi already?

<matrix_bridge><Andrius Štikonas> No, I have only managed to run POSIX version of mes there, haven't reached tcc

<matrix_bridge><Andrius Štikonas> But there were no progress on uefi for a year

<matrix_bridge><Andrius Štikonas> To run tcc you need to configure MMU

<aggi>Andrius, no need to excuse yourself. it's a miracle almost the x86 real mode one succeeded

<aggi>and it's not your fault Intel(R) is phasing out their systems which offered the only known complete bootstrapping dependency chain (software side at least)

<matrix_bridge><Andrius Štikonas> UEFI is not really that hard to work with either and memory mapping is usually the basic element of any OS

<matrix_bridge><Andrius Štikonas> I just didn't find enough time or motivation to finish it

<matrix_bridge><Andrius Štikonas> And nobody else stepped in either

<aggi>sigh, if memory mappings were the only issue with UEFI and bootstrapping

IRC channel logs

2026-02-04.log