luke
/
iverilog
mirror of https://github.com/steveicarus/iverilog.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Multiply of real values has width of 1. 

When a multiply expression is IVL_VT_REAL, then the default width is
the width of the operands, and not the sum of the operands. In fact, it
should turn out to be 1.

Also, the width of the multiply expression when we get to the ivl_target
API, need not be the sum of the widths of its arguments. In fact, if the
arguments are signed, that would be completely wrong. So adjust the
stub target to properly test this constraint.
This commit is contained in:
Stephen Williams 2008-09-25 20:33:29 -07:00
parent ce7dd6b4ff
commit 873ed60ff8
2 changed files with 18 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
net_expr.cc
View File

@ -242,7 +242,11 @@ ivl_variable_type_t NetEBMinMax::expr_type() const
NetEBMult::NetEBMult(char op, NetExpr*l, NetExpr*r) NetEBMult::NetEBMult(char op, NetExpr*l, NetExpr*r)
: NetEBinary(op, l, r) : NetEBinary(op, l, r)
{ {
expr_width(l->expr_width() + r->expr_width()); if (expr_type() == IVL_VT_REAL)
expr_width(1);
else
expr_width(l->expr_width() + r->expr_width());
cast_signed(l->has_sign() && r->has_sign()); cast_signed(l->has_sign() && r->has_sign());
/* If it turns out that this is not a signed expression, then /* If it turns out that this is not a signed expression, then

22
tgt-stub/expression.c
View File

@ -88,15 +88,19 @@ static void show_binary_expression(ivl_expr_t net, unsigned ind)
stub_errors += 1; stub_errors += 1;
} }
} else { } else {
/* The width of multiply expressions is the sum of the /* The width of a multiply may be any width. The
widths of the operands. This is slightly different implicit assumption is that the multiply
from the way the Verilog standard does it, but allows returns a width that is the sum of the widths
us to keep operands smaller. */ of the arguments, that is then truncated to the
width = ivl_expr_width(ivl_expr_oper1(net)); desired width, never padded. The compiler will
width += ivl_expr_width(ivl_expr_oper2(net)); automatically take care of sign extensions of
if (ivl_expr_width(net) != width) { arguments, so that the code generator need only
fprintf(out, "%*sERROR: Result width incorrect. Expecting %u, got %u\n", generate an UNSIGNED multiply, and the result
ind+3, "", width, ivl_expr_width(net)); will come out right. */
unsigned max_width = ivl_expr_width(oper1) + ivl_expr_width(oper2);
if (ivl_expr_width(net) > max_width) {
fprintf(out, "%*sERROR: Result width to width. Expecting <= %u, got %u\n",
ind+3, "", max_width, ivl_expr_width(net));
stub_errors += 1; stub_errors += 1;
} }
} }