Re: Building current...

[Johnny Billquist <bqt%update.uu.se@localhost>]

On 2015-12-03 01:37, Johnny Billquist wrote:
> On 2015-12-03 01:27, Johnny Billquist wrote:
> > On 2015-12-03 01:09, Christos Zoulas wrote:
> > > On Dec 3, 12:01am, bqt%update.uu.se@localhost (Johnny Billquist) wrote:
> > > -- Subject: Re: Building current...
> > >
> > > | => 0x004455ed <+75>:    extzv $0x1f,$0x1,r10,0xfffffff8(fp)
> > > |
> > > | Hmm, I can't see what's wrong. extzv with those arguments should be ok
> > > | according to VARM. But maybe I'm confused...?
> > >
> > > It
> > >
> > >
> > > http://www2.hmc.edu/www_common/OVMS072-OLD/72final/4515/4515pro_018.html
> > >
> > > says:
> > >
> > > -----
> > > Opcodes
> > >
> > > EE    EXTV    Extract Field
> > > EF    EXTZV    Extract Zero-Extended Field
> > >
> > > Description
> > >
> > > For EXTV, the destination operand is replaced by the sign-extended
> > > field specified by the position, size, and base operands. For EXTZV,
> > > the destination operand is replaced by the zero-extended field
> > > specified by the position, size, and base operands. If the size
> > > operand is zero, the only action is to replace the destination
> > > operand with zero and to modify the condition codes.
> > >
> > > Notes
> > >
> > > 1. A reserved operand fault occurs if:
> > >
> > >     o size GTRU 32
> > >     o pos GTRU 31, size NEQ 0, and the field is contained in the
> > > registers
> > >
> > > On a reserved operand fault, the destination operand is unaffected,
> > > and the condition codes are UNPREDICTABLE.
> > >
> > > -----
> > > Here:
> > >
> > > pos == 31 (0x1f)
> > > size != 0 (0x1)
> > > field = r10
> > >
> > > So, the processor is right :-) Now how to figure out what generates
> > > this...
> >
> > GTRU = Greater Than Unsigned. 31 is not greater than 31... :-)
>
> Actually,
>
> extzv $0x1f, $1, r10, -4(sp)
>
> would pretty much be code to set -4(sp) to -1 if r10 is negative, and
> set -4(sp) to 0 if r10 is positive.

And it's getting late (or at least I pretend that is the source of my 
errors).

Apart from just grabbing -4(sp) instead of -8(fp), since this 
instruction do zero extending, the result is (obviously) that

-4(sp) would be 1 if r10 was negative, and 0 if r10 was positivt.

I'll be quiet now, and someone can try to figure out where I'm thinking 
wrong, or the documentation is wrong, since I do get traps on my 4000/90 
as well as the 8650. I should check if it traps at the same place on the 
8650... It will take a while to check, but I'll get to it tomorrow. Time 
for bed now.

	Johnny

--
Johnny Billquist                  || "I'm on a bus
                                  ||  on a psychedelic trip
email: bqt%softjar.se@localhost             ||  Reading murder books
pdp is alive!                     ||  tryin' to stay hip" - B. Idol