Subject: Re: MP patches for i386 vs. -current
To: Michael Graff <firstname.lastname@example.org>
From: Stefan Grefen <email@example.com>
Date: 07/18/1999 23:46:52
In message <firstname.lastname@example.org> Michael Graff wrote:
> Stefan Grefen <email@example.com> writes:
> > The patches are against a pre 1.4 system, which system did you
> > compile?
> 1.4G, aka -current.
If I get some time (maybe tuesday), I'll try to compile a current see
whats going on.
> > I haven't tested againt the 'current' current yet (time contraints),
> > but my last cursory look showed that there is a little more work
> > todo.
> If you want the diffs, let me know. The only real changes are trivial
> ones, but for the lack of spinning up :)
There were some changes where I'm not sure my hack (esp. for mapping the
local apic) still works.
> > I think for some reason sending the IPI to processor 1 failed. As it
> > didn't crash the mapping of the local APIC may have failed.
> I assume this is different from apic* in the config file?
> Also, the mapping of the apic did not fail, at least not in the calls
> to i386_init() or i386_ipi() -- they both returned 0.
They would return 0 if the (re)mapping of the APIC failed (because busy
wouldn't be set :-)) )
I think that maybe the remapping didn't work and there is still the old
RAM-page behind this.
Stefan Grefen Tantau Software International Inc.
--- Hacking's just another word for nothing left to kludge. ---