>>On Wednesday 30 October 2002 09:23 am, you is done writ: >>> I have a 486 embedded target, and a Celeron-based development host. >> >>> Installation on the 486 target is complicated by the fact >>that the target >>> bios doesn't recognize, let alone boot from, a CD. >> >>> Installation on a Celeron development host is complicated >>by the fact >>> that 686-class libraries are installed, not 486. >>>The kernel (built for a 486) boots and runs until it frees >>memory, that's >>> the last thing it says. >>> Normally, Init is next. I never hear from Init. I never >>see a point in the >>> RH installation path where I can state what the real target >>architecture is. >> >>Hmmm...when you configured the kernel before building, did >>you specify the >>target archetecture?
Yup. Da kernel is 486'd. Da kernel boots. Things stop and hang right after "freeing memory", da next thing printed should be "Init Version x.yz" but it never shows. >>if there are *any* Pentium-class optimizations, including any >>that are included or statically linked in, you'll be toast on >>a [34]86. That's what I think is happening with init. >>compiling your own libraries This is what seems necessary. >>it does *not* have a FPP, since, AFAIK, celery chips <g> all >>have the FPP brain-deaded. On *my* pulpit, the whole x86 instruction architecture is a few brain-cells short of "well thought out". >>AFAIK, that's a compile-time command, not a run-time one. I was hoping it was an *installation* time parameter, with which anaconda would decide to ignore the i686 chip and install i486 stuff. Evidently the anaconda code specifically looks at uname() output to decide what architecture to install. -- redhat-list mailing list unsubscribe mailto:redhat-list-request@;redhat.com?subject=unsubscribe https://listman.redhat.com/mailman/listinfo/redhat-list
