Subject: Re: k0/k1 register while mulhi/mullo manipulation
To: Michael L. Hitch <mhitch@lightning.msu.montana.edu>
From: Erik Bertelsen <erik@mediator.uni-c.dk>
List: port-pmax
Date: 02/24/2000 12:16:10
On Wed, Feb 23, 2000 at 06:11:46AM -0700, Michael L. Hitch wrote:
> On Wed, 23 Feb 2000, Toru Nishimura wrote:
>
> > Now, I could build another, more likely, reasoning why recent
> > NetBSD/pmax kernels (1.4T) lock up sporadorically; an unexpected
> > interrupt during critical section. Let see following locore_mips1.S
> > change;
> >
> > + lw k0, TF_BASE+TF_REG_EPC(sp)
> > mtc0 a0, MIPS_COP_0_STATUS
> > mtlo t0
> > mthi t1
> > - lw k0, TF_BASE+TF_REG_EPC(sp)
> > ...
> > j k0
> > rfe
> >
On my pmax, I hand-reverted the above cited update yesterday, and now the machine is still running
with this code in a kernel otherwise based on yesterday's sup.
It has now rebuilt libraries and it is busily rebuilding the rest of userland, at this moment
it is in usr.bin/logger.
With the hanging problem, I would sometimes have several hangs/reboots just while building a
kernel, so it seems to be better.
- Erik