https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114331
--- Comment #6 from Aldy Hernandez <aldyh at gcc dot gnu.org> ---
(In reply to Andrew Macleod from comment #5)
> (In reply to Jakub Jelinek from comment #4)
> > Actually, looking at value-range.h, irange_bitmask doesn't have just the
> > mask but also value, so I wonder why it isn't able to figure this out.
> > I'd think m_mask should be ~0xffff and value 111.
>
> _1 = (short int) num_5(D);
> _2 = (int) _1;
> switch (_1)
>
> globally we know
> _2 : [irange] int [-32768, 32767]
> and coming into bb 3:
> _2 : [irange] int [-32768, 32767]
> 2->3 _1 : [irange] short int [111, 111]
> 2->3 _2 : [irange] int [111, 111]
> 2->3 num_5(D) : [irange] int [-INF, -65537][-65425, -65425][111,
> 111][65536, +INF]
>
> I guess what you are looking for is to add known bits to the range produced
> for num_5 on the outgoing edge.
>
> This would have to be done in operator_cast::op1_range where we are
> reducing it to known range of the switch. I do not believe it makes any
> attempts to set bitmasks based on casts like that currently.
>
> so, GORI working backwards has _1 = [irange] short int [111, 111] , so it
> would be satisfying the expression:
> short int [111, 111] = (short int) num_5(D);
>
> when evaluating num_5 in operator_cast::op1_range, in the truncating_cast_p
> section we could also see if there is a bitmask pattern for the LHS that
> could be applied to the range.. I think that basically what you are asking
> for. Simple enough for a constant I suppose for starters. Or maybe Aldy
> has a routine that picks bitmasks and values from ranges somewhere?
You may want to look at:
// Return the bitmask inherent in the range.
irange_bitmask
irange::get_bitmask_from_range () const
{
}
IIRC, this code was originally authored by Jakub and was embedded in the
tree-ssanames.cc code setting nonzero bits every time we set a global range. I
just abstracted it and adapted it to work within our framework.
