[Bug target/107243] New: ice in arithmetic_instr, at config/arm/arm.cc:2089

[Bug target/107243] New: ice in arithmetic_instr, at config/arm/arm.cc:2089

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107243

            Bug ID: 107243
           Summary: ice in arithmetic_instr, at config/arm/arm.cc:2089
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: dcb314 at hotmail dot com
  Target Milestone: ---

On machine type arm-linux-gnueabihf, this code 

void gdbstub_write_word(void) {
  int val = 0;
  void *addr = 0;
  unsigned long brr;
  asm volatile("        movgs   gr0,brr \n"
               "        sth%I2  %1,%M2  \n"
               "        movsg   brr,%0  \n"
               : "=r"(brr)
               : "r"(val), "m"(*addr));
}

does this:

during RTL pass: final
q.i: In function 'gdbstub_write_word':
q.i:9:1: internal compiler error: in arithmetic_instr, at
config/arm/arm.c:20612
    9 | }
      | ^

[Bug target/107243] ice in arithmetic_instr, at config/arm/arm.cc:2089

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107243

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2022-10-13
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
static void
arm_print_operand (FILE *stream, rtx x, int code)
{
  switch (code)
    {
...
    case 'I':
      fprintf (stream, "%s", arithmetic_instr (x, 0));
      return;


...
const char *
arithmetic_instr (rtx op, int shift_first_arg)
{
  switch (GET_CODE (op))
    {
...
    default:
      gcc_unreachable ();
    }
}

The gcc_unreachable here should rather be a call to output_operand_lossage

[Bug target/107243] ice in arithmetic_instr with invalid inline-asm and %I

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107243

--- Comment #2 from David Binderman <dcb314 at hotmail dot com> ---
Nothing has happened on this for over a month.

The bug still exists in current sources.