[Bug tree-optimization/113783] New: ICE: in lower_stmt, at gimple-lower-bitint.cc:5455 with -O -mavx512f and _BitInt() in memcpy()

[Bug tree-optimization/113783] New: ICE: in lower_stmt, at gimple-lower-bitint.cc:5455 with -O -mavx512f and _BitInt() in memcpy()

[zsojka at seznam dot cz]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113783

            Bug ID: 113783
           Summary: ICE: in lower_stmt, at gimple-lower-bitint.cc:5455
                    with -O -mavx512f and _BitInt() in memcpy()
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Keywords: ice-on-valid-code
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: zsojka at seznam dot cz
                CC: jakub at gcc dot gnu.org
  Target Milestone: ---
              Host: x86_64-pc-linux-gnu
            Target: x86_64-pc-linux-gnu

Created attachment 57340
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=57340&action=edit
reduced testcase

Compiler output:
$ x86_64-pc-linux-gnu-gcc -O -mavx512f testcase.c
during GIMPLE pass: bitintlower
testcase.c: In function 'foo':
testcase.c:2:1: internal compiler error: in lower_stmt, at
gimple-lower-bitint.cc:5455
    2 | foo(void *p, _BitInt(246) x)
      | ^~~
0xd8cedf lower_stmt
        /repo/gcc-trunk/gcc/gimple-lower-bitint.cc:5455
0x2721439 gimple_lower_bitint
        /repo/gcc-trunk/gcc/gimple-lower-bitint.cc:6582
Please submit a full bug report, with preprocessed source (by using
-freport-bug).
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.

$ x86_64-pc-linux-gnu-gcc -v
Using built-in specs.
COLLECT_GCC=/repo/gcc-trunk/binary-latest-amd64/bin/x86_64-pc-linux-gnu-gcc
COLLECT_LTO_WRAPPER=/repo/gcc-trunk/binary-trunk-r14-8817-20240205212943-gc5d34912ad5-checking-yes-rtl-df-extra-nobootstrap-amd64/bin/../libexec/gcc/x86_64-pc-linux-gnu/14.0.1/lto-wrapper
Target: x86_64-pc-linux-gnu
Configured with: /repo/gcc-trunk//configure --enable-languages=c,c++
--enable-valgrind-annotations --disable-nls --enable-checking=yes,rtl,df,extra
--disable-bootstrap --with-cloog --with-ppl --with-isl
--build=x86_64-pc-linux-gnu --host=x86_64-pc-linux-gnu
--target=x86_64-pc-linux-gnu --with-ld=/usr/bin/x86_64-pc-linux-gnu-ld
--with-as=/usr/bin/x86_64-pc-linux-gnu-as --disable-libstdcxx-pch
--prefix=/repo/gcc-trunk//binary-trunk-r14-8817-20240205212943-gc5d34912ad5-checking-yes-rtl-df-extra-nobootstrap-amd64
Thread model: posix
Supported LTO compression algorithms: zlib zstd
gcc version 14.0.1 20240206 (experimental) (GCC)

[Bug tree-optimization/113783] ICE: in lower_stmt, at gimple-lower-bitint.cc:5455 with -O -mavx512f and _BitInt() in memcpy()

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113783

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Ugh, folding turns this into
  uint256_t D.3758;

  <bb 2> :
  D.3758 = MEM <uint256_t> [(char * {ref-all})&x];
  MEM <uint256_t> [(char * {ref-all})p] = D.3758;
and later ssa turns it into
  uint256_t _2;

  <bb 2> :
  _2 = VIEW_CONVERT_EXPR<uint256_t>(x_1(D));
  MEM <uint256_t> [(char * {ref-all})p_4(D)] = _2;
where presumably uint256_t is some 256-bit unsigned INTEGER_TYPE.
Except it isn't really supported except perhaps for moves.

Similarly
void
bar (void *p, _BitInt(502) x)
{
  __builtin_memcpy (p, &x, sizeof x);
}
into
  uint512_t D.3758;

  D.3758 = MEM <uint512_t> [(char * {ref-all})&x];
  MEM <uint512_t> [(char * {ref-all})p] = D.3758;
and later
  uint512_t _2;

  <bb 2> :
  _2 = VIEW_CONVERT_EXPR<uint512_t>(x_1(D));
  MEM <uint512_t> [(char * {ref-all})p_4(D)] = _2;


I don't see how I could handle that in bitint lowering when shifts,
BIT_IOR_EXPR etc.
will not work on such types.

[Bug tree-optimization/113783] ICE: in lower_stmt, at gimple-lower-bitint.cc:5455 with -O -mavx512f and _BitInt() in memcpy()

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113783

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
This comes from
              scalar_int_mode mode;
              if (int_mode_for_size (ilen * 8, 0).exists (&mode)
                  && GET_MODE_SIZE (mode) * BITS_PER_UNIT == ilen * 8
                  /* If the destination pointer is not aligned we must be able
                     to emit an unaligned store.  */
                  && (dest_align >= GET_MODE_ALIGNMENT (mode)
                      || !targetm.slow_unaligned_access (mode, dest_align)
                      || (optab_handler (movmisalign_optab, mode)
                          != CODE_FOR_nothing)))
where we just care if that huge mode exists, not whether it is supported (but
for moves or in misaligned case movmisalign).

[Bug tree-optimization/113783] ICE: in lower_stmt, at gimple-lower-bitint.cc:5455 with -O -mavx512f and _BitInt() in memcpy()

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113783

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
           Assignee|unassigned at gcc dot gnu.org      |jakub at gcc dot gnu.org
   Last reconfirmed|                            |2024-02-08
             Status|UNCONFIRMED                 |ASSIGNED

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 57358
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=57358&action=edit
gcc14-pr113783.patch

Untested fix.

[Bug tree-optimization/113783] ICE: in lower_stmt, at gimple-lower-bitint.cc:5455 with -O -mavx512f and _BitInt() in memcpy()

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113783

--- Comment #4 from GCC Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Jakub Jelinek <jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:c9bdcb0c3433ce09f5bb713a51a14130858578a2

commit r14-8899-gc9bdcb0c3433ce09f5bb713a51a14130858578a2
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Fri Feb 9 16:17:08 2024 +0100

    lower-bitint: Fix handling of VIEW_CONVERT_EXPRs to minimally supported
huge INTEGER_TYPEs [PR113783]

    On the following testcases memcpy lowering folds the calls to
    reading and writing of MEM_REFs with huge INTEGER_TYPEs - uint256_t
    with OImode or uint512_t with XImode.  Further optimization turn
    the load from MEM_REF from the large/huge _BitInt var into
VIEW_CONVERT_EXPR
    from it to the uint256_t/uint512_t.  The backend doesn't really
    support those except for "movoi"/"movxi" insns, so it isn't possible
    to handle it like casts to supportable INTEGER_TYPEs where we can
    construct those from individual limbs - there are no OImode/XImode shifts
    and the like we can use.
    So, the following patch makes sure for such VCEs that the SSA_NAME operand
    of the VCE lives in memory and then turns it into a VIEW_CONVERT_EXPR so
    that we actually load the OImode/XImode integer from memory (i.e. a mov).
    We need to make sure those aren't merged with other
    operations in the gimple_lower_bitint hunks.
    For SSA_NAMEs which have underlying VAR_DECLs that is all we need, those
    VAR_DECL have ARRAY_TYPEs.
    For SSA_NAMEs which have underlying PARM_DECLs or RESULT_DECLs those have
    BITINT_TYPE and I had to tweak expand_expr_real_1 for that so that it
    doesn't try convert_modes on those when one of the modes is BLKmode - we
    want to fall through into the adjust_address on the MEM.

    2024-02-09  Jakub Jelinek  <jakub@redhat.com>

            PR tree-optimization/113783
            * gimple-lower-bitint.cc (bitint_large_huge::lower_stmt): Look
            through VIEW_CONVERT_EXPR for final cast checks.  Handle
            VIEW_CONVERT_EXPRs from large/huge _BitInt to > MAX_FIXED_MODE_SIZE
            INTEGER_TYPEs.
            (gimple_lower_bitint): Don't merge mergeable operations or other
            casts with VIEW_CONVERT_EXPRs to > MAX_FIXED_MODE_SIZE
INTEGER_TYPEs.
            * expr.cc (expand_expr_real_1): Don't use convert_modes if either
            mode is BLKmode.

            * gcc.dg/bitint-88.c: New test.

[Bug tree-optimization/113783] ICE: in lower_stmt, at gimple-lower-bitint.cc:5455 with -O -mavx512f and _BitInt() in memcpy()

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113783

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |FIXED
             Status|ASSIGNED                    |RESOLVED

--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed.

[Bug tree-optimization/113783] ICE: in lower_stmt, at gimple-lower-bitint.cc:5455 with -O -mavx512f and _BitInt() in memcpy()

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113783

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |14.0