[Bug tree-optimization/97555] [11 Regression] wrong code at -Os and above on x86_64-pc-linux-gnu since r11-3685
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97555 Aldy Hernandez changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #6 from Aldy Hernandez --- fixed
[Bug tree-optimization/97555] [11 Regression] wrong code at -Os and above on x86_64-pc-linux-gnu since r11-3685
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97555 --- Comment #5 from CVS Commits --- The master branch has been updated by Aldy Hernandez : https://gcc.gnu.org/g:2118438f49f0c193abe3fa3def350a8129045746 commit r11-4390-g2118438f49f0c193abe3fa3def350a8129045746 Author: Aldy Hernandez Date: Mon Oct 26 17:50:37 2020 +0100 Handle signed 1-bit ranges in irange::invert. The problem here is we are trying to add 1 to a -1 in a signed 1-bit field and coming up with UNDEFINED because of the overflow. Signed 1-bits are annoying because you can't really add or subtract one, because the one is unrepresentable. For invert() we have a special subtract_one() function that handles 1-bit signed fields. This patch implements the analogous add_one() function so that invert works. gcc/ChangeLog: PR tree-optimization/97555 * range-op.cc (range_tests): Test 1-bit signed invert. * value-range.cc (subtract_one): Adjust comment. (add_one): New. (irange::invert): Call add_one. gcc/testsuite/ChangeLog: * gcc.dg/pr97555.c: New test.
[Bug tree-optimization/97555] [11 Regression] wrong code at -Os and above on x86_64-pc-linux-gnu since r11-3685
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97555
--- Comment #4 from Aldy Hernandez ---
The problem here is we're trying to add 1 to a -1 in a signed 1-bit field.
Signed 1-bits are annoying because you can't really add or subtract one,
because the one is unrepresentable. For invert() we have a special
subtract_one() function that handles 1-bit signed fields. We need a
corresponding add_one() here.
The untested patch below should do it.
diff --git a/gcc/range-op.cc b/gcc/range-op.cc
index ee62f103598..74ab2e57fde 100644
--- a/gcc/range-op.cc
+++ b/gcc/range-op.cc
@@ -3680,15 +3680,28 @@ range_tests ()
// Test 1-bit signed integer union.
// [-1,-1] U [0,0] = VARYING.
tree one_bit_type = build_nonstandard_integer_type (1, 0);
+ tree one_bit_min = vrp_val_min (one_bit_type);
+ tree one_bit_max = vrp_val_max (one_bit_type);
{
-tree one_bit_min = vrp_val_min (one_bit_type);
-tree one_bit_max = vrp_val_max (one_bit_type);
int_range<2> min (one_bit_min, one_bit_min);
int_range<2> max (one_bit_max, one_bit_max);
max.union_ (min);
ASSERT_TRUE (max.varying_p ());
}
+ // Test inversion of 1-bit signed integers.
+ {
+int_range<2> min (one_bit_min, one_bit_min);
+int_range<2> max (one_bit_max, one_bit_max);
+int_range<2> t;
+t = min;
+t.invert ();
+ASSERT_TRUE (t == max);
+t = max;
+t.invert ();
+ASSERT_TRUE (t == min);
+ }
+
// Test that NOT(255) is [0..254] in 8-bit land.
int_range<1> not_255 (UCHAR (255), UCHAR (255), VR_ANTI_RANGE);
ASSERT_TRUE (not_255 == int_range<1> (UCHAR (0), UCHAR (254)));
diff --git a/gcc/value-range.cc b/gcc/value-range.cc
index 7847104050c..f45a342605a 100644
--- a/gcc/value-range.cc
+++ b/gcc/value-range.cc
@@ -1772,18 +1772,29 @@ irange::irange_intersect (const irange )
verify_range ();
}
+// Signed 1-bits are strange. You can't subtract 1, because you can't
+// represent the number 1. This works around that for the invert routine.
+
static wide_int inline
subtract_one (const wide_int , tree type, wi::overflow_type )
{
- // A signed 1-bit bit-field, has a range of [-1,0] so subtracting +1
- // overflows, since +1 is unrepresentable. This is why we have an
- // addition of -1 here.
if (TYPE_SIGN (type) == SIGNED)
-return wi::add (x, -1 , SIGNED, );
+return wi::add (x, -1, SIGNED, );
else
return wi::sub (x, 1, UNSIGNED, );
}
+// The analogous function for adding 1.
+
+static wide_int inline
+add_one (const wide_int , tree type, wi::overflow_type )
+{
+ if (TYPE_SIGN (type) == SIGNED)
+return wi::sub (x, -1, SIGNED, );
+ else
+return wi::add (x, 1, TYPE_SIGN (type), );
+}
+
/* Return the inverse of a range. */
void
@@ -1881,7 +1892,7 @@ irange::invert ()
// set the overflow bit.
if (type_max != wi::to_wide (orig_range.m_base[i]))
{
- tmp = wi::add (wi::to_wide (orig_range.m_base[i]), 1, sign, );
+ tmp = add_one (wi::to_wide (orig_range.m_base[i]), ttype, ovf);
m_base[nitems++] = wide_int_to_tree (ttype, tmp);
m_base[nitems++] = wide_int_to_tree (ttype, type_max);
if (ovf)
[Bug tree-optimization/97555] [11 Regression] wrong code at -Os and above on x86_64-pc-linux-gnu since r11-3685
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97555 Aldy Hernandez changed: What|Removed |Added Assignee|unassigned at gcc dot gnu.org |aldyh at gcc dot gnu.org --- Comment #3 from Aldy Hernandez --- Mine. Thanks for the analysis.
[Bug tree-optimization/97555] [11 Regression] wrong code at -Os and above on x86_64-pc-linux-gnu since r11-3685
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97555
--- Comment #2 from Andrew Macleod ---
:
f.0_1 = f;
_2 = 1 % f.0_1;
h_24 = (char) _2;
_3 = _2;
c = _3;
_4 = b.a;
_5 = (int) _4;
_6 = ~_5;
f = _6;
if (_4 != -1)
goto ; [INV]
else
goto ; [INV]
when calculating the outgoing_range_p() of edge 3->4,
we know that then range is != -1.
operator_not_equal::op1_range
calculates the range on the true side as:
if (wi::eq_p (op2.lower_bound(), op2.upper_bound()))
{
r = op2;
r.invert ();
}
538 r = op2;
(gdb) p op2.dump(stderr)
[-1, -1]$12 = void
(gdb) n
539 r.invert ();
(gdb) p r.dump(stderr)
[-1, -1]$13 = void
(gdb) n
543 break;
(gdb) p r.dump(stderr)
UNDEFINED$14 = void
when we invert the range [-1, -1]
we should get [0, 0]
but instead its returning UNDEFINED.
which when the post dominated merge happens in bb5, we are unioning [-1,-1] and
undefined, producing [-1, -1] ofr the range of _4..
if we were unioning [-1, -1] and [0, 0] like we we are suppose to get, we'd get
VARYING, and the statement would not be incorrectly folded.
You can put a breakpoint in operator_not_equal::op1_range and it should be the
first time it is hit.
[Bug tree-optimization/97555] [11 Regression] wrong code at -Os and above on x86_64-pc-linux-gnu since r11-3685
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97555 Jakub Jelinek changed: What|Removed |Added Last reconfirmed||2020-10-24 Target Milestone|--- |11.0 Ever confirmed|0 |1 Summary|wrong code at -Os and above |[11 Regression] wrong code |on x86_64-pc-linux-gnu |at -Os and above on ||x86_64-pc-linux-gnu since ||r11-3685 Status|UNCONFIRMED |NEW CC||aldyh at gcc dot gnu.org, ||amacleod at redhat dot com, ||jakub at gcc dot gnu.org Priority|P3 |P1 --- Comment #1 from Jakub Jelinek --- Started with r11-3685-gfcae5121154d1c3382b056bcc2c563cedac28e74
