Haven't seen a trap 9 yet :) (Re: trap 9 still happening; ata1-slave still repeats; my kernel in uue/tar/bzip2 format with a couple extra files)

Hi again John,

No trap 9s yet with either scheduler.

I was running the kernel built with ULE for about 100 minutes.
Trying to stress it some with mplayer and KDE.  I'd say it
should have popped by then.  ;)

Then I switched to the other kernel built with 4BSD,
been running it for about an hour, and stressing it the same
way.  It should've crashed by now, too.

So, this PC seems to be working ok after backing out these
particular changes, i.e. 
  apic_vector.s is now at rev 1.95,   and
  atpic_vector.s is now at rev 1.40.

> Can you try reverting these changes to see if that fixes your trap 9?
>
> jhb         2004/01/28 12:44:08 PST
>
>   FreeBSD src repository
>
>   Modified files:
>     sys/i386/i386        apic_vector.s
>     sys/i386/isa         atpic_vector.s
>   Log:
>   Optimize the i386 interrupt entry code to not reload the segment registers
>   if they already contain the correct kernel selectors.
>
>   Reviewed by:    peter
>   Suggested by:   peter
>
>   Revision  Changes    Path
>   1.96      +30 -39    src/sys/i386/i386/apic_vector.s
>   1.41      +4 -1      src/sys/i386/isa/atpic_vector.s


Now that a current kernel is working long enough to use it, 
I'm seeing some brand-new problems from other parts of the bootup
phase and normal usage.  Nothing to do with these particular changes.
I'll iron them down and probably start a new discussion for them.

Thank you -- Paul Seniura.