Re: Insight with 64bit architectures -- problem

["John R. Moore" <jmoore at redhat dot com>]

All right, I was missing a %llx in gdb_set_bp (I had it in gdb_set_bp_addr)
and now I'm ready for an [RFA]. I'll send it under separate cover since
I need to put together the patches and ChangeLog.

John

On Wed, 4 Apr 2001, John R. Moore wrote:

>
> Well, the fix has lots of longs getting converted to long long's
> and I still don't have it working perfectly... One can set
> the breakpoint in the ASM window and it reflects up on the SRC,
> but not the reverse. Also I note that the break point addresses
> are all 32bit when they should be 64. I think I might have to
> fix that first before everyone (SRC & ASM) is cahoots.
>
> Comments anybody? Seen this before? Advise?
>
> John
>
> On Wed, 4 Apr 2001, John R. Moore wrote:
>
> >
> > Ok, I wrote a strtoull and got the following to disappear,
> > and selecting a line in the ASM code selects corresponding
> > SRC, but not vice versa. Must be a different problem since
> > I don't see the "integer value too large to represent".
> >
> > I'll find and fix the above before posting my [RFA].
> >
> > John
> >
> > On Wed, 4 Apr 2001, John R. Moore wrote:
> >
> > > Insighters,
> > >
> > > I've followed my problem into SetIntFromAny where I'm
> > > passing the address 0xffffffffa0020154 and on line
> > > 1730 of tcl/generic/tclObj.c we try to use
> > >
> > > strtoul()
> > >
> > > and output into a long .... well, we get -1 and the result
> > > is
> > >
> > > "integer value too large to represent" (line 1748).
> > >
> > > The above breaks setting the break points at assembler lines
> > > on 64bit architectures (with sign extension - MIPS) when
> > > we set display of SRC+ASM (or ASM, etc...)
> > >
> > > Any suggestions here?  Time for a stroull() and change to
> > > unsigned long long our value.
> > >
> > > John
> > >
> > >
> > >
> >
> >
>
>