On Wed, 05 May 2010 12:20:10 -0700, "Kevin D. Kissell" <kevink@paralogos.com>
wrote:
> > Yes, that's my patch. Though I cannot remember precisely, maybe I
> > just had (mis)thought Cause bits in FCSR are read-only at that time.
> > I have never used real SMP MIPS platforms, so there must be no
> > SMP-related issues.
> >
> The patch was labeled for "preemption and SMP problems", and if you
> weren't working on an SMP system, I'd guess that you were working with
> full preemption and seeing a problem that you might have assumed was
> also relevant to SMP. The FPU emulator *shouldn't* have pre-emption
> issues, since it works off of data structures that are instantiated per
> thread context. The FCSR seen by thread A is logically independent of
> that seen by thread B, so that even if one emulation was preempted in
> the middle by another, they shouldn't be able to interfere with one
> another. That was the concept, anyway.
Yes, I created the patch for full preemption support. At that time,
ieee754_csr was a global variable, not an alias of
current->thread.fpu.soft.fcr31.
The commmit 72402e was mixture of two patch: "just make FPU emulator
fully-preemptive" and "fix FPU ownership issues on preemption". The
wrong code is a part of the first one. So I don't think your fix
causes any regressions on SMP/preemption point of view.
I'm still trying to remember why I masked out some bits on CTC1
emulation, but no success. Maybe I was just confused.
Anyway, thanks Shane and Kevin!
---
Atsushi Nemoto
|