Re: next68k: current issues

[Izumi Tsutsui <tsutsui%ceres.dti.ne.jp@localhost>]

Hi,

> > I've also managed to build the latest (r1291) on my NetBSD/i386 9.3 host
> > (with several tweaks for NetBSD) and also build a NetBSD/next68k 10.99.2
> > disk image using snapshot built on my host locally:
> > http://teokurebsd.org/netbsd/liveimage/20230211-next68k/
> 
> If you send me a patch with the tweaks that were necessary for NetBSD
> I'll add them to Previous' repository.

Changes are trivial. Here is my quick patch:
 https://gist.github.com/tsutsui/87bf0822187613e074da2bb9363b7a65

- src/ditools
 - <dirent.h> is explicitly necessary for 'struct dirent'
 - On most BSDs, inode is 64 bit integer so gcc (c++) complains if 64 bit inode
   (ino_t or daddr_t block number) values are passed to fsv()
   (I'm not sure where/how ditool is used though)

- src/fast_screen.c
 - On NetBSD Xorg support is not enough so it could fail with acceleration
   (it's worth to fallback to no acceleration if SDL_CreateRenderer() fails?)

> I can confirm that your image starts booting all emulated Turbos, but hangs
> after "kern.module.patch=/stand/next68k/10.99.2/modules" (same for all machines,
> including non-Turbo).

Even on Previous, sometimes I got "Enter pathname of shell or RETURN for /bn/sh:"
prompt.  An image with NetBSD/next68k 9.3 userland and 10.99.2 (trunk) kernel
works (i.e. /bin/sh starts properly) even on Previous.

One possibility is lack of entropy (NetBSD doesn't have pcap so Previous on it
has no network input), but I'll check what's going on it.

Thanks,
---
Izumi Tsutsui