deo@eecs.cs.pdx.edu (Steven Furber) (11/12/90)
I have rebuild my MacMinix kernel twice now. Once after applying the ps fix, and then again after making changes in the config.h (robust filesystem, wrap lines after 80, etc..). The only difference that I have been able to actually see is the ps fix, but that fix didn't fix all of the way. A 'ps -U' gives me: Can't get kernel proc table from /dev/kmem: Error 00 For those of you who rebuild the kernel, when you go back to [Multi]Finder and reboot with the net boot application, I've found that holding down on the mouse so you can select the configuration options will result in a system bomb. What I have ended up doing is openning the boot application and then holding me mouse over the `File' menu-item with the button down. Once the menu comes down I select configuration and setup everything that way. As soon as that is done, I select quite. This will enable to you reboot without having to go through the RAM disk being loaded and the other stuff that is *not* what you previously had Minix set up to do. I also recommend renaming the boot application as soon as you get to finder, otherwise when you build a new one it will halt.
wbb@hpclove.HP.COM (Bill Buzbee) (11/12/90)
>deo@eecs.cs.pdx.edu (Steven Furber) writes: >I have rebuild my MacMinix kernel twice now. Once after applying the >ps fix, and then again after making changes in the config.h (robust filesystem, >wrap lines after 80, etc..). > >The only difference that I have been able to actually see is the ps fix, but >that fix didn't fix all of the way. A 'ps -U' gives me: > Can't get kernel proc table from /dev/kmem: Error 00 I've hit the same thing. Thinking perhaps I'd screwed something up, I even rebuilt a system from virgin disks, reapplied the patch and tried again. Same problem. Do have some good news though :-). The clam shell executable I pulled off of atari.archive.umich.edu (141.221.192.10) works great! Of course, I can't recreate it from source :-( (fairly simple source change gave me a successful compile, but I immediately died with a malloc assert). .......bill buzbee
archetyp@uxh.cso.uiuc.edu (Joseph R Pickert) (11/13/90)
deo@eecs.cs.pdx.edu (Steven Furber) writes: >I have rebuild my MacMinix kernel twice now. Once after applying the >ps fix, and then again after making changes in the config.h (robust filesystem, >wrap lines after 80, etc..). Someone has told me that the my posting contained an error. In /usr/src/kernel/memory.c, line 55, the sum (sizes[0] + sizes[1]) should be cast as a long. I will post a diff file.