Another regression caused by my recent patch.
This time because define_insn_and_split only requires that the
constant mask is const_int_operand. When it was only SImode,
that wasn't a problem, HImode neither, but for DImode if we need
to and the shift count we might run into a problem that it isn't
a representable signed 32-bit immediate.
But, we don't really care about the upper bits of the mask, so
we can just mask the CONST_INT with the mode mask.
2022-06-13 Jakub Jelinek <jakub@redhat.com>
PR target/105911
* config/i386/i386.md (*ashl<dwi>3_doubleword_mask,
*<insn><dwi>3_doubleword_mask): Use operands[3] masked with
(<MODE_SIZE> * BITS_PER_UNIT) - 1 as AND operand instead of
operands[3] unmodified.
* gcc.dg/pr105911.c: New test.
rtx xops[3];
xops[0] = gen_reg_rtx (GET_MODE (operands[2]));
xops[1] = operands[2];
- xops[2] = operands[3];
+ xops[2] = GEN_INT (INTVAL (operands[3])
+ & ((<MODE_SIZE> * BITS_PER_UNIT) - 1));
ix86_expand_binary_operator (AND, GET_MODE (operands[2]), xops);
operands[2] = xops[0];
}
rtx xops[3];
xops[0] = gen_reg_rtx (GET_MODE (operands[2]));
xops[1] = operands[2];
- xops[2] = operands[3];
+ xops[2] = GEN_INT (INTVAL (operands[3])
+ & ((<MODE_SIZE> * BITS_PER_UNIT) - 1));
ix86_expand_binary_operator (AND, GET_MODE (operands[2]), xops);
operands[2] = xops[0];
}
--- /dev/null
+/* PR target/105911 */
+/* { dg-do compile { target int128 } } */
+/* { dg-options "-O2" } */
+
+__int128 v, x;
+unsigned __int128 w;
+
+void bar (__int128, __int128);
+
+void
+foo (void)
+{
+ bar (v /= v, v >> (v &= 0x100000001));
+ bar (w /= w, w >> (w &= 0x300000003));
+ bar (x /= x, x << (x &= 0x700000007));
+}