procyberian/gcc
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

combine: Fix up expand_compound_operation [PR99905]

Browse source 
The following testcase is miscompiled on x86_64-linux.
expand_compound_operation is called on
(zero_extract:DI (mem/c:TI (reg/f:DI 16 argp) [3 i+0 S16 A128])
    (const_int 16 [0x10])
    (const_int 63 [0x3f]))
so mode is DImode, inner_mode is TImode, pos 63, len 16 and modewidth 64.

A couple of lines above the problematic spot we have:
  if (modewidth >= pos + len)
    {
      tem = gen_lowpart (mode, XEXP (x, 0));
where the code uses gen_lowpart and then shift left/right to extract it
in mode.  But the guarding condition is false - 64 >= 63 + 16
and so we enter the next condition, where the code shifts XEXP (x, 0)
right by pos and then adds AND.  It does so incorrectly though.
Given the modewidth < pos + len, inner_mode must be necessarily larger
than mode and XEXP (x, 0) has the innermode, but it was calling
simplify_shift_const with mode rather than inner_mode, which meant
inconsistent arguments to simplify_shift_const and in this case made
a DImode MEM shift out of it.

The following patch fixes it, by doing the shift in inner_mode properly
and then after the shift doing the lowpart subreg and masking already
in mode.

2021-04-13  Jakub Jelinek  <jakub@redhat.com>

	PR rtl-optimization/99905
	* combine.c (expand_compound_operation): If pos + len > modewidth,
	perform the right shift by pos in inner_mode and then convert to mode,
	instead of trying to simplify a shift of rtx with inner_mode by pos
	as if it was a shift in mode.

	* gcc.target/i386/pr99905.c: New test.
This commit is contained in:
Jakub Jelinek 2021-04-13 01:01:45 +02:00
parent 9c1c8ad833
commit ffc4155b6e
2 changed files with 42 additions and 5 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

14
gcc/combine.c
View file

@ -7409,11 +7409,15 @@ expand_compound_operation (rtx x)
mode, tem, modewidth - len);
}
else if (unsignedp && len < HOST_BITS_PER_WIDE_INT)
tem = simplify_and_const_int (NULL_RTX, mode,
simplify_shift_const (NULL_RTX, LSHIFTRT,
mode, XEXP (x, 0),
pos),
(HOST_WIDE_INT_1U << len) - 1);
{
tem = simplify_shift_const (NULL_RTX, LSHIFTRT, inner_mode,
XEXP (x, 0), pos);
tem = gen_lowpart (mode, tem);
if (!tem || GET_CODE (tem) == CLOBBER)
return x;
tem = simplify_and_const_int (NULL_RTX, mode, tem,
(HOST_WIDE_INT_1U << len) - 1);
}
else
/* Any other cases we can't handle. */
return x;

33
gcc/testsuite/gcc.target/i386/pr99905.c Normal file
View file

@ -0,0 +1,33 @@
/* PR rtl-optimization/99905 */
/* { dg-do run { target int128 } } */
/* { dg-options "-Os -mno-mmx -mno-sse" } */
typedef unsigned char U;
typedef unsigned char __attribute__((__vector_size__ (8))) A;
typedef unsigned char __attribute__((__vector_size__ (16))) B;
typedef unsigned char __attribute__((__vector_size__ (32))) C;
typedef unsigned int __attribute__((__vector_size__ (8))) D;
typedef unsigned long long __attribute__((__vector_size__ (8))) E;
typedef unsigned __int128 I;
typedef unsigned long long L;
D gv;
I gi;
L __attribute__((__noipa__))
foo (int ua, int ub, int uc, int ud, E ue, I i)
{
D d = (U) __builtin_bswap16 (i >> 63) + gv;
B y = ((union { C a; B b[2];}) (C){ }).b[0] + (B) gi;
A z = ((union { B a; A b[2];}) y).b[0] + (A) d;
return (L)z;
}
int
main ()
{
L x = foo (0, 0, 0, 0, (E) { }, (I) 0x100 << 63);
if (x != 0x100000001)
__builtin_abort ();
return 0;
}